Intel® 64 and IA-32 Architectures
Software Developer’s Manual
Volume 3A:
System Programming Guide, Part 1
NOTE: The Intel® 64 and IA-32 Architectures Software Developer's Manual
consists of five volumes: Basic Architecture, Order Number 253665;
Instruction Set Reference A-M, Order Number 253666; Instruction Set
Reference N-Z, Order Number 253667; System Programming Guide,
Part 1, Order Number 253668; System Programming Guide, Part 2, Order
Number 253669. Refer to all five volumes when evaluating your design
needs.
Order Number: 253668-032US
September 2009
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
CHAPTER 1
ABOUT THIS MANUAL
1.1
PROCESSORS COVERED IN THIS MANUAL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
OVERVIEW OF THE SYSTEM PROGRAMMING GUIDE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-3
NOTATIONAL CONVENTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-6
Bit and Byte Order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-6
Reserved Bits and Software Compatibility. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-7
Instruction Operands. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-8
Hexadecimal and Binary Numbers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-8
Segmented Addressing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-8
Syntax for CPUID, CR, and MSR Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-9
Exceptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-10
RELATED LITERATURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-11
1.2
1.3
1.3.1
1.3.2
1.3.3
1.3.4
1.3.5
1.3.6
1.3.7
1.4
CHAPTER 2
SYSTEM ARCHITECTURE OVERVIEW
2.1
OVERVIEW OF THE SYSTEM-LEVEL ARCHITECTURE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-2
Global and Local Descriptor Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-5
Global and Local Descriptor Tables in IA-32e Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-5
System Segments, Segment Descriptors, and Gates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-5
Gates in IA-32e Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-6
Task-State Segments and Task Gates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-6
Task-State Segments in IA-32e Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-7
Interrupt and Exception Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-7
Interrupt and Exception Handling IA-32e Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-7
Memory Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-8
Memory Management in IA-32e Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-8
System Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-9
System Registers in IA-32e Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-9
Other System Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-10
MODES OF OPERATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-10
SYSTEM FLAGS AND FIELDS IN THE EFLAGS REGISTER. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-12
System Flags and Fields in IA-32e Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-15
MEMORY-MANAGEMENT REGISTERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-15
Global Descriptor Table Register (GDTR). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-16
Local Descriptor Table Register (LDTR). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-16
IDTR Interrupt Descriptor Table Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-17
Task Register (TR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-17
CONTROL REGISTERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-17
CPUID Qualification of Control Register Flags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-26
EXTENDED CONTROL REGISTERS (INCLUDING THE XFEATURE_ENABLED_MASK REGISTER)
2-26
2.1.1
2.1.1.1
2.1.2
2.1.2.1
2.1.3
2.1.3.1
2.1.4
2.1.4.1
2.1.5
2.1.5.1
2.1.6
2.1.6.1
2.1.7
2.2
2.3
2.3.1
2.4
2.4.1
2.4.2
2.4.3
2.4.4
2.5
2.5.1
2.6
2.7
SYSTEM INSTRUCTION SUMMARY. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-27
Loading and Storing System Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-29
Verifying of Access Privileges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-30
Loading and Storing Debug Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-31
Invalidating Caches and TLBs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-31
2.7.1
2.7.2
2.7.3
2.7.4
Vol. 3A iii
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
2.7.5
Controlling the Processor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-31
Reading Performance-Monitoring and Time-Stamp Counters . . . . . . . . . . . . . . . . . . . . . 2-32
Reading Counters in 64-Bit Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-33
Reading and Writing Model-Specific Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-33
Reading and Writing Model-Specific Registers in 64-Bit Mode. . . . . . . . . . . . . . . . . . 2-34
Enabling Processor Extended States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-34
2.7.6
2.7.6.1
2.7.7
2.7.7.1
2.7.8
CHAPTER 3
PROTECTED-MODE MEMORY MANAGEMENT
3.1
MEMORY MANAGEMENT OVERVIEW. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
USING SEGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3
Basic Flat Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3
Protected Flat Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4
Multi-Segment Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5
Segmentation in IA-32e Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-6
Paging and Segmentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7
PHYSICAL ADDRESS SPACE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7
Intel® 64 Processors and Physical Address Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-8
LOGICAL AND LINEAR ADDRESSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-8
Logical Address Translation in IA-32e Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-9
Segment Selectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-9
Segment Registers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-10
Segment Loading Instructions in IA-32e Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-12
Segment Descriptors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-13
Code- and Data-Segment Descriptor Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-16
SYSTEM DESCRIPTOR TYPES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-18
Segment Descriptor Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-20
Segment Descriptor Tables in IA-32e Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-22
3.2
3.2.1
3.2.2
3.2.3
3.2.4
3.2.5
3.3
3.3.1
3.4
3.4.1
3.4.2
3.4.3
3.4.4
3.4.5
3.4.5.1
3.5
3.5.1
3.5.2
CHAPTER 4
PAGING
4.1
PAGING MODES AND CONTROL BITS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1
4.1.1
4.1.2
4.1.3
4.1.4
4.2
Three Paging Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-2
Paging-Mode Enabling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-3
Paging-Mode Modifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-5
Enumeration of Paging Features by CPUID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-6
HIERARCHICAL PAGING STRUCTURES: AN OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-7
32-BIT PAGING. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-8
PAE PAGING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-15
PDPTE Registers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-16
Linear-Address Translation with PAE Paging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-17
IA-32E PAGING. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-23
ACCESS RIGHTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-32
PAGE-FAULT EXCEPTIONS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-34
ACCESSED AND DIRTY FLAGS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-36
PAGING AND MEMORY TYPING. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-37
Paging and Memory Typing When the PAT is Not Supported (Pentium Pro and Pentium II
Processors). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-37
Paging and Memory Typing When the PAT is Supported (Pentium III and More Recent
Processor Families) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-37
4.3
4.4
4.4.1
4.4.2
4.5
4.6
4.7
4.8
4.9
4.9.1
4.9.2
iv Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
4.9.3
Caching Paging-Related Information about Memory Typing . . . . . . . . . . . . . . . . . . . . . . .4-38
CACHING TRANSLATION INFORMATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-38
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Translation Lookaside Buffers (TLBs)4-39
Page Numbers, Page Frames, and Page Offsets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-39
Caching Translations in TLBs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-40
Details of TLB Use. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-40
Global Pages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-41
Paging-Structure Caches. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-41
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Caches for Paging Structures4-41
Using the Paging-Structure Caches to Translate Linear Addresses . . . . . . . . . . . . .4-44
. . . . . . . . . . . . . . . . . . . .Multiple Cached Entries for a Single Paging-Structure Entry4-45
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Invalidation of TLBs and Paging-Structure Caches4-46
Operations that Invalidate TLBs and Paging-Structure Caches . . . . . . . . . . . . . . . . .4-46
Recommended Invalidation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-47
Optional Invalidation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-48
. . . . . . . . . . . . . . . . . Propagation of Paging-Structure Changes to Multiple Processors4-49
INTERACTIONS WITH VIRTUAL-MACHINE EXTENSIONS (VMX). . . . . . . . . . . . . . . . . . . . . . . 4-51
VMX Transitions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-51
VMX Support for Address Translation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-51
USING PAGING FOR VIRTUAL MEMORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-52
MAPPING SEGMENTS TO PAGES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-52
4.10
4.10.1
4.10.1.1
4.10.1.2
4.10.1.3
4.10.1.4
4.10.2
4.10.2.1
4.10.2.2
4.10.2.3
4.10.3
4.10.3.1
4.10.3.2
4.10.3.3
4.10.4
4.11
4.11.1
4.11.2
4.12
4.13
CHAPTER 5
PROTECTION
5.1
5.2
ENABLING AND DISABLING SEGMENT AND PAGE PROTECTION. . . . . . . . . . . . . . . . . . . . . . . 5-1
FIELDS AND FLAGS USED FOR SEGMENT-LEVEL AND
PAGE-LEVEL PROTECTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-2
Code Segment Descriptor in 64-bit Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-5
LIMIT CHECKING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-6
Limit Checking in 64-bit Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-7
TYPE CHECKING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-7
Null Segment Selector Checking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-9
NULL Segment Checking in 64-bit Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-9
PRIVILEGE LEVELS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-9
PRIVILEGE LEVEL CHECKING WHEN ACCESSING DATA SEGMENTS . . . . . . . . . . . . . . . . . . . 5-11
Accessing Data in Code Segments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-14
PRIVILEGE LEVEL CHECKING WHEN LOADING THE SS REGISTER . . . . . . . . . . . . . . . . . . . . . 5-14
PRIVILEGE LEVEL CHECKING WHEN TRANSFERRING PROGRAM CONTROL BETWEEN CODE
SEGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-14
Direct Calls or Jumps to Code Segments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-15
Accessing Nonconforming Code Segments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-16
Accessing Conforming Code Segments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-17
Gate Descriptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-18
Call Gates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-19
IA-32e Mode Call Gates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-20
Accessing a Code Segment Through a Call Gate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-22
Stack Switching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-25
Stack Switching in 64-bit Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-28
Returning from a Called Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-28
Performing Fast Calls to System Procedures with the
5.2.1
5.3
5.3.1
5.4
5.4.1
5.4.1.1
5.5
5.6
5.6.1
5.7
5.8
5.8.1
5.8.1.1
5.8.1.2
5.8.2
5.8.3
5.8.3.1
5.8.4
5.8.5
5.8.5.1
5.8.6
5.8.7
SYSENTER and SYSEXIT Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-30
Vol. 3A v
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
5.8.7.1
5.8.8
SYSENTER and SYSEXIT Instructions in IA-32e Mode. . . . . . . . . . . . . . . . . . . . . . . . . . 5-31
Fast System Calls in 64-bit Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-32
PRIVILEGED INSTRUCTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-33
POINTER VALIDATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-34
Checking Access Rights (LAR Instruction). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-35
Checking Read/Write Rights (VERR and VERW Instructions) . . . . . . . . . . . . . . . . . . . . . . 5-36
Checking That the Pointer Offset Is Within Limits (LSL Instruction). . . . . . . . . . . . . . . . 5-36
Checking Caller Access Privileges (ARPL Instruction) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-37
Checking Alignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-39
PAGE-LEVEL PROTECTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-39
Page-Protection Flags. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-40
Restricting Addressable Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-40
Page Type. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-40
Combining Protection of Both Levels of Page Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-41
Overrides to Page Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-41
COMBINING PAGE AND SEGMENT PROTECTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-41
PAGE-LEVEL PROTECTION AND EXECUTE-DISABLE BIT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-43
Detecting and Enabling the Execute-Disable Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-43
Execute-Disable Page Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-44
Reserved Bit Checking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-45
Exception Handling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-47
5.9
5.10
5.10.1
5.10.2
5.10.3
5.10.4
5.10.5
5.11
5.11.1
5.11.2
5.11.3
5.11.4
5.11.5
5.12
5.13
5.13.1
5.13.2
5.13.3
5.13.4
CHAPTER 6
INTERRUPT AND EXCEPTION HANDLING
6.1
INTERRUPT AND EXCEPTION OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1
EXCEPTION AND INTERRUPT VECTORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-2
SOURCES OF INTERRUPTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-2
External Interrupts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-2
Maskable Hardware Interrupts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-4
Software-Generated Interrupts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-5
SOURCES OF EXCEPTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-5
Program-Error Exceptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-5
Software-Generated Exceptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-6
Machine-Check Exceptions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-6
EXCEPTION CLASSIFICATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-6
PROGRAM OR TASK RESTART . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-7
NONMASKABLE INTERRUPT (NMI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-8
Handling Multiple NMIs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-9
ENABLING AND DISABLING INTERRUPTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-9
Masking Maskable Hardware Interrupts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-9
Masking Instruction Breakpoints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-10
Masking Exceptions and Interrupts When Switching Stacks. . . . . . . . . . . . . . . . . . . . . . . 6-11
PRIORITY AMONG SIMULTANEOUS EXCEPTIONS AND INTERRUPTS . . . . . . . . . . . . . . . . . . 6-11
INTERRUPT DESCRIPTOR TABLE (IDT). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-12
IDT DESCRIPTORS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-14
EXCEPTION AND INTERRUPT HANDLING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-15
Exception- or Interrupt-Handler Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-16
Protection of Exception- and Interrupt-Handler Procedures . . . . . . . . . . . . . . . . . . . 6-18
Flag Usage By Exception- or Interrupt-Handler Procedure. . . . . . . . . . . . . . . . . . . . . 6-19
Interrupt Tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-20
ERROR CODE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-21
6.2
6.3
6.3.1
6.3.2
6.3.3
6.4
6.4.1
6.4.2
6.4.3
6.5
6.6
6.7
6.7.1
6.8
6.8.1
6.8.2
6.8.3
6.9
6.10
6.11
6.12
6.12.1
6.12.1.1
6.12.1.2
6.12.2
6.13
vi Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
6.14
EXCEPTION AND INTERRUPT HANDLING IN 64-BIT MODE. . . . . . . . . . . . . . . . . . . . . . . . . . . 6-22
64-Bit Mode IDT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-23
64-Bit Mode Stack Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-24
IRET in IA-32e Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-25
Stack Switching in IA-32e Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-25
Interrupt Stack Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-26
EXCEPTION AND INTERRUPT REFERENCE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-27
Interrupt 0—Divide Error Exception (#DE). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-28
Interrupt 1—Debug Exception (#DB). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-29
Interrupt 2—NMI Interrupt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-30
Interrupt 3—Breakpoint Exception (#BP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-31
Interrupt 4—Overflow Exception (#OF) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-32
Interrupt 5—BOUND Range Exceeded Exception (#BR) . . . . . . . . . . . . . . . . . . . . . . . . . . .6-33
Interrupt 6—Invalid Opcode Exception (#UD). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-34
Interrupt 7—Device Not Available Exception (#NM). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-36
Interrupt 8—Double Fault Exception (#DF) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-38
Interrupt 9—Coprocessor Segment Overrun. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-41
Interrupt 10—Invalid TSS Exception (#TS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-42
Interrupt 11—Segment Not Present (#NP). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-46
Interrupt 12—Stack Fault Exception (#SS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-48
Interrupt 13—General Protection Exception (#GP). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-50
Interrupt 14—Page-Fault Exception (#PF) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-54
Interrupt 16—x87 FPU Floating-Point Error (#MF). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-58
Interrupt 17—Alignment Check Exception (#AC). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-60
Interrupt 18—Machine-Check Exception (#MC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-62
Interrupt 19—SIMD Floating-Point Exception (#XM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-64
Interrupts 32 to 255—User Defined Interrupts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-67
6.14.1
6.14.2
6.14.3
6.14.4
6.14.5
6.15
CHAPTER 7
TASK MANAGEMENT
7.1
TASK MANAGEMENT OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1
7.1.1
7.1.2
7.1.3
7.2
Task Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1
Task State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-2
Executing a Task . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-3
TASK MANAGEMENT DATA STRUCTURES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-4
Task-State Segment (TSS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-4
TSS Descriptor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-7
TSS Descriptor in 64-bit mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-8
Task Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-9
Task-Gate Descriptor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7-11
TASK SWITCHING. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-12
TASK LINKING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-16
Use of Busy Flag To Prevent Recursive Task Switching. . . . . . . . . . . . . . . . . . . . . . . . . . .7-18
Modifying Task Linkages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7-18
TASK ADDRESS SPACE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-19
Mapping Tasks to the Linear and Physical Address Spaces . . . . . . . . . . . . . . . . . . . . . . . .7-19
Task Logical Address Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7-20
16-BIT TASK-STATE SEGMENT (TSS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-21
TASK MANAGEMENT IN 64-BIT MODE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-22
7.2.1
7.2.2
7.2.3
7.2.4
7.2.5
7.3
7.4
7.4.1
7.4.2
7.5
7.5.1
7.5.2
7.6
7.7
Vol. 3A vii
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
CHAPTER 8
MULTIPLE-PROCESSOR MANAGEMENT
8.1
LOCKED ATOMIC OPERATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-2
8.1.1
8.1.2
8.1.2.1
8.1.2.2
8.1.3
8.1.4
8.2
Guaranteed Atomic Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-3
Bus Locking. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-3
Automatic Locking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-4
Software Controlled Bus Locking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-5
Handling Self- and Cross-Modifying Code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-6
Effects of a LOCK Operation on Internal Processor Caches . . . . . . . . . . . . . . . . . . . . . . . . 8-7
MEMORY ORDERING. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-8
®
®
™
8.2.1
8.2.2
8.2.3
8.2.3.1
8.2.3.2
8.2.3.3
8.2.3.4
8.2.3.5
8.2.3.6
8.2.3.7
8.2.3.8
8.2.3.9
8.2.4
8.2.4.1
8.2.4.2
8.2.5
8.3
Memory Ordering in the Intel Pentium and Intel486 Processors. . . . . . . . . . . . . 8-8
Memory Ordering in P6 and More Recent Processor Families . . . . . . . . . . . . . . . . . . . . . . 8-9
Examples Illustrating the Memory-Ordering Principles. . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-11
Assumptions, Terminology, and Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-11
. . . . . . . . . . . . . . . . Neither Loads Nor Stores Are Reordered with Like Operations8-12
Stores Are Not Reordered With Earlier Loads. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-13
Loads May Be Reordered with Earlier Stores to Different Locations . . . . . . . . . . . 8-13
Intra-Processor Forwarding Is Allowed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-14
Stores Are Transitively Visible. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-15
Stores Are Seen in a Consistent Order by Other Processors . . . . . . . . . . . . . . . . . . . 8-15
Locked Instructions Have a Total Order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-16
. . . . . . . . . . . . . . . . Loads and Stores Are Not Reordered with Locked Instructions8-16
Out-of-Order Stores For String Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-18
Memory-Ordering Model for String Operations on Write-back (WB) Memory . . . . 8-18
Examples Illustrating Memory-Ordering Principles for String Operations. . . . . . . . 8-19
Strengthening or Weakening the Memory-Ordering Model. . . . . . . . . . . . . . . . . . . . . . . . 8-22
SERIALIZING INSTRUCTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-24
MULTIPLE-PROCESSOR (MP) INITIALIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-26
BSP and AP Processors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-27
MP Initialization Protocol Requirements and Restrictions. . . . . . . . . . . . . . . . . . . . . . . . . 8-27
MP Initialization Protocol Algorithm for Intel Xeon Processors . . . . . . . . . . . . . . . . . . . . 8-28
MP Initialization Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-29
Typical BSP Initialization Sequence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-30
Typical AP Initialization Sequence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-32
Identifying Logical Processors in an MP System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-33
8.4
8.4.1
8.4.2
8.4.3
8.4.4
8.4.4.1
8.4.4.2
8.4.5
8.5
®
®
INTEL HYPER-THREADING TECHNOLOGY AND INTEL MULTI-CORE TECHNOLOGY. 8-35
DETECTING HARDWARE MULTI-THREADING SUPPORT AND TOPOLOGY. . . . . . . . . . . . . . 8-35
Initializing Processors Supporting Hyper-Threading Technology . . . . . . . . . . . . . . . . . . 8-36
8.6
8.6.1
8.6.2
8.6.3
Initializing Multi-Core Processors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-37
®
Executing Multiple Threads on an Intel 64 or IA-32 Processor Supporting Hardware
Multi-Threading. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-37
8.6.4
8.7
Handling Interrupts on an IA-32 Processor Supporting Hardware Multi-Threading . 8-37
®
INTEL HYPER-THREADING TECHNOLOGY ARCHITECTURE . . . . . . . . . . . . . . . . . . . . . . . . . 8-38
8.7.1
8.7.2
8.7.3
8.7.4
8.7.5
8.7.6
8.7.7
8.7.8
State of the Logical Processors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-39
APIC Functionality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-40
Memory Type Range Registers (MTRR). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-40
Page Attribute Table (PAT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-41
Machine Check Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-41
Debug Registers and Extensions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-41
Performance Monitoring Counters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-42
IA32_MISC_ENABLE MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-42
viii Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
8.7.9
Memory Ordering. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-42
Serializing Instructions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-42
MICROCODE UPDATE Resources. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-43
Self Modifying Code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-43
Implementation-Specific Intel HT Technology Facilities . . . . . . . . . . . . . . . . . . . . . . . . . . .8-43
Processor Caches. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-43
Processor Translation Lookaside Buffers (TLBs). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-44
Thermal Monitor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-44
External Signal Compatibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-45
MULTI-CORE ARCHITECTURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-46
Logical Processor Support. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-46
Memory Type Range Registers (MTRR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-46
Performance Monitoring Counters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-47
IA32_MISC_ENABLE MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-47
MICROCODE UPDATE Resources. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-47
PROGRAMMING CONSIDERATIONS FOR HARDWARE MULTI-THREADING CAPABLE
PROCESSORS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-47
Hierarchical Mapping of Shared Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-48
Hierarchical Mapping of CPUID Extended Topology Leaf . . . . . . . . . . . . . . . . . . . . . . . . . .8-50
Hierarchical ID of Logical Processors in an MP System . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-51
Hierarchical ID of Logical Processors with x2APIC ID. . . . . . . . . . . . . . . . . . . . . . . . . . .8-53
Algorithm for Three-Level Mappings of APIC_ID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-54
Identifying Topological Relationships in a MP System . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-60
MANAGEMENT OF IDLE AND BLOCKED CONDITIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-64
HLT Instruction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-64
PAUSE Instruction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-65
Detecting Support MONITOR/MWAIT Instruction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-65
MONITOR/MWAIT Instruction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-66
Monitor/Mwait Address Range Determination. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-67
Required Operating System Support. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-68
Use the PAUSE Instruction in Spin-Wait Loops. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-68
Potential Usage of MONITOR/MWAIT in C0 Idle Loops . . . . . . . . . . . . . . . . . . . . . . . . .8-69
Halt Idle Logical Processors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-71
Potential Usage of MONITOR/MWAIT in C1 Idle Loops . . . . . . . . . . . . . . . . . . . . . . . . .8-71
Guidelines for Scheduling Threads on Logical Processors Sharing Execution
8.7.10
8.7.11
8.7.12
8.7.13
8.7.13.1
8.7.13.2
8.7.13.3
8.7.13.4
8.8
8.8.1
8.8.2
8.8.3
8.8.4
8.8.5
8.9
8.9.1
8.9.2
8.9.3
8.9.3.1
8.9.4
8.9.5
8.10
8.10.1
8.10.2
8.10.3
8.10.4
8.10.5
8.10.6
8.10.6.1
8.10.6.2
8.10.6.3
8.10.6.4
8.10.6.5
Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-72
Eliminate Execution-Based Timing Loops. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8-72
Place Locks and Semaphores in Aligned, 128-Byte Blocks of Memory. . . . . . . . . . .8-73
8.10.6.6
8.10.6.7
CHAPTER 9
PROCESSOR MANAGEMENT AND INITIALIZATION
9.1
INITIALIZATION OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-1
Processor State After Reset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-2
Processor Built-In Self-Test (BIST). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-2
Model and Stepping Information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-5
First Instruction Executed. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-6
X87 FPU INITIALIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-6
Configuring the x87 FPU Environment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-6
Setting the Processor for x87 FPU Software Emulation . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-7
CACHE ENABLING. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-8
MODEL-SPECIFIC REGISTERS (MSRS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-9
9.1.1
9.1.2
9.1.3
9.1.4
9.2
9.2.1
9.2.2
9.3
9.4
Vol. 3A ix
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
9.5
MEMORY TYPE RANGE REGISTERS (MTRRS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-9
INITIALIZING SSE/SSE2/SSE3/SSSE3 EXTENSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-10
SOFTWARE INITIALIZATION FOR REAL-ADDRESS MODE OPERATION. . . . . . . . . . . . . . . . . 9-10
Real-Address Mode IDT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-11
NMI Interrupt Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-11
SOFTWARE INITIALIZATION FOR PROTECTED-MODE OPERATION . . . . . . . . . . . . . . . . . . . . 9-11
Protected-Mode System Data Structures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-12
Initializing Protected-Mode Exceptions and Interrupts. . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-13
Initializing Paging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-13
Initializing Multitasking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-14
Initializing IA-32e Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-14
IA-32e Mode System Data Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-15
IA-32e Mode Interrupts and Exceptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-15
64-bit Mode and Compatibility Mode Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-16
Switching Out of IA-32e Mode Operation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-16
MODE SWITCHING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-17
Switching to Protected Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-17
Switching Back to Real-Address Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-18
INITIALIZATION AND MODE SWITCHING EXAMPLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-19
Assembler Usage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-22
STARTUP.ASM Listing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-23
MAIN.ASM Source Code. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-33
Supporting Files. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-34
MICROCODE UPDATE FACILITIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-36
Microcode Update. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-37
Optional Extended Signature Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-41
Processor Identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-41
Platform Identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-42
Microcode Update Checksum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-44
Microcode Update Loader . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-45
Hard Resets in Update Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-46
Update in a Multiprocessor System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-46
Update in a System Supporting Intel Hyper-Threading Technology . . . . . . . . . . . . 9-46
Update in a System Supporting Dual-Core Technology . . . . . . . . . . . . . . . . . . . . . . . . 9-46
Update Loader Enhancements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-47
Update Signature and Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-47
Determining the Signature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-48
Authenticating the Update . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-48
Pentium 4, Intel Xeon, and P6 Family Processor
9.6
9.7
9.7.1
9.7.2
9.8
9.8.1
9.8.2
9.8.3
9.8.4
9.8.5
9.8.5.1
9.8.5.2
9.8.5.3
9.8.5.4
9.9
9.9.1
9.9.2
9.10
9.10.1
9.10.2
9.10.3
9.10.4
9.11
9.11.1
9.11.2
9.11.3
9.11.4
9.11.5
9.11.6
9.11.6.1
9.11.6.2
9.11.6.3
9.11.6.4
9.11.6.5
9.11.7
9.11.7.1
9.11.7.2
9.11.8
Microcode Update Specifications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-49
Responsibilities of the BIOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-49
Responsibilities of the Calling Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-52
Microcode Update Functions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-55
INT 15H-based Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-55
Function 00H—Presence Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-56
Function 01H—Write Microcode Update Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-57
Function 02H—Microcode Update Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-62
Function 03H—Read Microcode Update Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-63
Return Codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-64
9.11.8.1
9.11.8.2
9.11.8.3
9.11.8.4
9.11.8.5
9.11.8.6
9.11.8.7
9.11.8.8
9.11.8.9
x
Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
CHAPTER 10
ADVANCED PROGRAMMABLE
INTERRUPT CONTROLLER (APIC)
10.1
LOCAL AND I/O APIC OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-1
10.2
SYSTEM BUS VS. APIC BUS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-5
®
10.3
10.4
THE INTEL 82489DX EXTERNAL APIC, THE APIC, THE XAPIC, AND THE X2APIC. . . . 10-5
LOCAL APIC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-6
The Local APIC Block Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-6
Presence of the Local APIC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-10
Enabling or Disabling the Local APIC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-10
Local APIC Status and Location. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-11
Relocating the Local APIC Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-12
Local APIC ID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-12
Local APIC State. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-13
Local APIC State After Power-Up or Reset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-14
Local APIC State After It Has Been Software Disabled . . . . . . . . . . . . . . . . . . . . . . . 10-14
Local APIC State After an INIT Reset (“Wait-for-SIPI” State) . . . . . . . . . . . . . . . . . . 10-15
Local APIC State After It Receives an INIT-Deassert IPI . . . . . . . . . . . . . . . . . . . . . . 10-15
Local APIC Version Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-15
EXTENDED XAPIC (X2APIC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-16
DETECTING AND ENABLING x2APIC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-16
Instructions to Access APIC Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-17
APIC Register Address Space. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-18
Reserved Bit Checking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-21
x2APIC Register Availability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-22
MSR Access in x2APIC Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-22
VM-exit Controls for MSRs and x2APIC Registers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-22
Directed EOI with x2APIC Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-23
x2APIC State Transitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-24
x2APIC States. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-24
x2APIC After RESET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-26
x2APIC Transitions From x2APIC Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-27
x2APIC Transitions From Disabled Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-27
State Changes From xAPIC Mode to x2APIC Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-27
System Software Transitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-27
CPUID Extensions And Topology Enumeration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-28
Consistency of APIC IDs and CPUID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-29
HANDLING LOCAL INTERRUPTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-30
Local Vector Table. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-30
Valid Interrupt Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-33
Error Handling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-34
x2APIC Differences in Error Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-35
APIC Timer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-36
Local Interrupt Acceptance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-38
ISSUING INTERPROCESSOR INTERRUPTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-38
Interrupt Command Register (ICR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-38
ICR Operation in x2APIC Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-44
Determining IPI Destination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-46
Physical Destination Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-46
Logical Destination Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-47
Logical Destination Mode in x2APIC Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-49
10.4.1
10.4.2
10.4.3
10.4.4
10.4.5
10.4.6
10.4.7
10.4.7.1
10.4.7.2
10.4.7.3
10.4.7.4
10.4.8
10.5
10.5.1
10.5.1.1
10.5.1.2
10.5.1.3
10.5.2
10.5.3
10.5.4
10.5.5
10.5.6
10.5.6.1
10.5.7
10.5.8
10.5.8.1
10.6
10.6.1
10.6.2
10.6.3
10.6.3.1
10.6.4
10.6.5
10.7
10.7.1
10.7.1.1
10.7.2
10.7.2.1
10.7.2.2
10.7.2.3
Vol. 3A xi
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
10.7.2.4
10.7.2.5
10.7.2.6
10.7.3
10.7.4
10.8
Deriving Logical x2APIC ID from the Local x2APIC ID . . . . . . . . . . . . . . . . . . . . . . . . .10-50
Broadcast/Self Delivery Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-51
Lowest Priority Delivery Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-51
IPI Delivery and Acceptance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-52
SELF IPI Register. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-52
SYSTEM AND APIC BUS ARBITRATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-53
HANDLING INTERRUPTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-54
Interrupt Handling with the Pentium 4 and Intel Xeon Processors . . . . . . . . . . . . . . .10-54
Interrupt Handling with the P6 Family and Pentium Processors. . . . . . . . . . . . . . . . . .10-55
Interrupt, Task, and Processor Priority . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-57
Task and Processor Priorities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-58
Interrupt Acceptance for Fixed Interrupts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-59
Signaling Interrupt Servicing Completion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-61
Signaling Interrupt Servicing Completion in x2APIC Mode. . . . . . . . . . . . . . . . . . . . .10-61
Task Priority in IA-32e Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-61
Interaction of Task Priorities between CR8 and APIC. . . . . . . . . . . . . . . . . . . . . . . . .10-62
10.9
10.9.1
10.9.2
10.9.3
10.9.3.1
10.9.4
10.9.5
10.9.5.1
10.9.6
10.9.6.1
10.10 SPURIOUS INTERRUPT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-63
10.11 APIC BUS MESSAGE PASSING MECHANISM AND
PROTOCOL (P6 FAMILY, PENTIUM PROCESSORS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-64
10.11.1
Bus Message Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-65
10.12 MESSAGE SIGNALLED INTERRUPTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-65
10.12.1
10.12.2
Message Address Register Format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-66
Message Data Register Format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-67
CHAPTER 11
MEMORY CACHE CONTROL
11.1
INTERNAL CACHES, TLBS, AND BUFFERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-1
11.2
CACHING TERMINOLOGY. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-7
METHODS OF CACHING AVAILABLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-8
Buffering of Write Combining Memory Locations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11-11
Choosing a Memory Type. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11-12
Code Fetches in Uncacheable Memory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11-13
CACHE CONTROL PROTOCOL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-13
CACHE CONTROL. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-14
Cache Control Registers and Bits. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11-15
Precedence of Cache Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11-20
Selecting Memory Types for Pentium Pro and Pentium II Processors. . . . . . . . . .11-20
Selecting Memory Types for Pentium III and More Recent Processor Families. .11-22
Writing Values Across Pages with Different Memory Types . . . . . . . . . . . . . . . . . .11-23
Preventing Caching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11-24
Disabling and Enabling the L3 Cache . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11-25
Cache Management Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11-25
L1 Data Cache Context Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11-26
Adaptive Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11-26
Shared Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11-26
SELF-MODIFYING CODE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-27
IMPLICIT CACHING (PENTIUM 4, INTEL XEON,
11.3
11.3.1
11.3.2
11.3.3
11.4
11.5
11.5.1
11.5.2
11.5.2.1
11.5.2.2
11.5.2.3
11.5.3
11.5.4
11.5.5
11.5.6
11.5.6.1
11.5.6.2
11.6
11.7
AND P6 FAMILY PROCESSORS). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-27
EXPLICIT CACHING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-28
INVALIDATING THE TRANSLATION LOOKASIDE BUFFERS (TLBS). . . . . . . . . . . . . . . . . . . 11-29
11.8
11.9
11.10 STORE BUFFER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-29
xii Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
11.11 MEMORY TYPE RANGE REGISTERS (MTRRS). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-30
11.11.1
MTRR Feature Identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-32
Setting Memory Ranges with MTRRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-33
IA32_MTRR_DEF_TYPE MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-33
Fixed Range MTRRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-34
Variable Range MTRRs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-34
System-Management Range Register Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-37
Example Base and Mask Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-38
Base and Mask Calculations for Greater-Than 36-bit Physical Address Support11-40
Range Size and Alignment Requirement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-41
MTRR Precedences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-41
MTRR Initialization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-41
Remapping Memory Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-42
MTRR Maintenance Programming Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-42
MemTypeGet() Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-42
MemTypeSet() Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-44
MTRR Considerations in MP Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-46
Large Page Size Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-47
11.11.2
11.11.2.1
11.11.2.2
11.11.2.3
11.11.2.4
11.11.3
11.11.3.1
11.11.4
11.11.4.1
11.11.5
11.11.6
11.11.7
11.11.7.1
11.11.7.2
11.11.8
11.11.9
11.12 PAGE ATTRIBUTE TABLE (PAT). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-48
11.12.1
11.12.2
11.12.3
11.12.4
11.12.5
Detecting Support for the PAT Feature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-48
IA32_PAT MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-49
Selecting a Memory Type from the PAT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-50
Programming the PAT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-50
PAT Compatibility with Earlier IA-32 Processors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-52
CHAPTER 12
®
™
INTEL MMX TECHNOLOGY SYSTEM PROGRAMMING
12.1
EMULATION OF THE MMX INSTRUCTION SET. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-1
THE MMX STATE AND MMX REGISTER ALIASING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-1
Effect of MMX, x87 FPU, FXSAVE, and FXRSTOR
12.2
12.2.1
Instructions on the x87 FPU Tag Word . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12-3
SAVING AND RESTORING THE MMX STATE AND REGISTERS . . . . . . . . . . . . . . . . . . . . . . . . 12-4
SAVING MMX STATE ON TASK OR CONTEXT SWITCHES . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-5
EXCEPTIONS THAT CAN OCCUR WHEN EXECUTING MMX INSTRUCTIONS . . . . . . . . . . . . 12-5
Effect of MMX Instructions on Pending x87 Floating-Point Exceptions. . . . . . . . . . . . .12-6
DEBUGGING MMX CODE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-6
12.3
12.4
12.5
12.5.1
12.6
CHAPTER 13
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND
PROCESSOR EXTENDED STATES
13.1
PROVIDING OPERATING SYSTEM SUPPORT FOR
SSE/SSE2/SSE3/SSSE3/SSE4 EXTENSIONS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-1
Adding Support to an Operating System for SSE/SSE2/SSE3/SSSE3/SSE4 Extensions . .
13-2
13.1.1
13.1.2
13.1.3
13.1.4
13.1.5
Checking for SSE/SSE2/SSE3/SSSE3/SSE4 Extension Support. . . . . . . . . . . . . . . . . . . . .13-2
Checking for Support for the FXSAVE and FXRSTOR Instructions . . . . . . . . . . . . . . . . .13-3
Initialization of the SSE/SSE2/SSE3/SSSE3/SSE4 Extensions. . . . . . . . . . . . . . . . . . . . . .13-3
Providing Non-Numeric Exception Handlers for Exceptions Generated by the
SSE/SSE2/SSE3/SSSE3/SSE4 Instructions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13-5
Providing an Handler for the SIMD Floating-Point Exception (#XM) . . . . . . . . . . . . . . . .13-7
13.1.6
Vol. 3A xiii
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
13.1.6.1
13.2
Numeric Error flag and IGNNE# . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-8
EMULATION OF SSE/SSE2/SSE3/SSSE3/SSE4 EXTENSIONS. . . . . . . . . . . . . . . . . . . . . . . . . . 13-8
SAVING AND RESTORING THE SSE/SSE2/SSE3/SSSE3/SSE4 STATE. . . . . . . . . . . . . . . . . . 13-8
SAVING THE SSE/SSE2/SSE3/SSSE3/SSE4 STATE ON TASK OR CONTEXT SWITCHES . 13-9
DESIGNING OS FACILITIES FOR AUTOMATICALLY SAVING X87 FPU, MMX, AND
SSE/SSE2/SSE3/SSSE3/SSE4 STATE ON TASK OR CONTEXT SWITCHES. . . . . . . . . . . . . . 13-9
Using the TS Flag to Control the Saving of the
13.3
13.4
13.5
13.5.1
x87 FPU, MMX, SSE, SSE2, SSE3 SSSE3 and SSE4 State . . . . . . . . . . . . . . . . . . . . . . . . .13-10
XSAVE/XRSTOR AND PROCESSOR EXTENDED STATE MANAGEMENT . . . . . . . . . . . . . . 13-12
XSAVE Header. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13-13
INTEROPERABILITY OF XSAVE/XRSTOR AND FXSAVE/FXRSTOR . . . . . . . . . . . . . . . . . . 13-15
DETECTION, ENUMERATION, ENABLING PROCESSOR EXTENDED STATE SUPPORT. . 13-17
Application Programming Model and Processor Extended States. . . . . . . . . . . . . . . . .13-18
13.6
13.6.1
13.7
13.8
13.8.1
CHAPTER 14
POWER AND THERMAL MANAGEMENT
®
14.1
14.1.1
14.2
ENHANCED INTEL SPEEDSTEP TECHNOLOGY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-1
Software Interface For Initiating Performance State Transitions . . . . . . . . . . . . . . . . . 14-1
P-STATE HARDWARE COORDINATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-2
SYSTEM SOFTWARE CONSIDERATIONS AND OPPORTUNISTIC PROCESSOR PERFORMANCE
OPERATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-4
Intel Dynamic Acceleration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-4
System Software Interfaces for Opportunistic Processor Performance Operation . 14-4
Discover Hardware Support and Enabling of Opportunistic Processor Operation 14-5
OS Control of Opportunistic Processor Performance Operation . . . . . . . . . . . . . . . . 14-5
Required Changes to OS Power Management P-state Policy. . . . . . . . . . . . . . . . . . . 14-6
Application Awareness of Opportunistic Processor Operation (Optional). . . . . . . . 14-7
Intel Turbo Boost Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-8
Performance and Energy Bias Hint support. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-8
MWAIT EXTENSIONS FOR ADVANCED POWER MANAGEMENT . . . . . . . . . . . . . . . . . . . . . . . 14-9
THERMAL MONITORING AND PROTECTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-10
Catastrophic Shutdown Detector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-12
Thermal Monitor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-12
Thermal Monitor 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-12
Thermal Monitor 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-12
Two Methods for Enabling TM2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-13
Performance State Transitions and Thermal Monitoring. . . . . . . . . . . . . . . . . . . . . .14-14
Thermal Status Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-14
Adaptive Thermal Monitor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-16
Software Controlled Clock Modulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-16
Detection of Thermal Monitor and Software Controlled
14.3
14.3.1
14.3.2
14.3.2.1
14.3.2.2
14.3.2.3
14.3.2.4
14.3.3
14.3.4
14.4
14.5
14.5.1
14.5.2
14.5.2.1
14.5.2.2
14.5.2.3
14.5.2.4
14.5.2.5
14.5.2.6
14.5.3
14.5.4
Clock Modulation Facilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-18
On Die Digital Thermal Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-18
Digital Thermal Sensor Enumeration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-18
Reading the Digital Sensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-19
14.5.5
14.5.5.1
14.5.5.2
CHAPTER 15
MACHINE-CHECK ARCHITECTURE
15.1
15.2
MACHINE-CHECK ARCHITECTURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-1
COMPATIBILITY WITH PENTIUM PROCESSOR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-2
xiv Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
15.3
MACHINE-CHECK MSRS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-2
Machine-Check Global Control MSRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-3
IA32_MCG_CAP MSR. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-3
IA32_MCG_STATUS MSR. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-5
IA32_MCG_CTL MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-6
Error-Reporting Register Banks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-6
IA32_MCi_CTL MSRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-6
IA32_MCi_STATUS MSRS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-7
IA32_MCi_ADDR MSRs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-11
IA32_MCi_MISC MSRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-12
IA32_MCi_CTL2 MSRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-13
IA32_MCG Extended Machine Check State MSRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-15
Mapping of the Pentium Processor Machine-Check Errors
15.3.1
15.3.1.1
15.3.1.2
15.3.1.3
15.3.2
15.3.2.1
15.3.2.2
15.3.2.3
15.3.2.4
15.3.2.5
15.3.2.6
15.3.3
to the Machine-Check Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-17
ENHANCED CACHE ERROR REPORTING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-18
CORRECTED MACHINE CHECK ERROR INTERRUPT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-18
CMCI Local APIC Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-19
System Software Recommendation for Managing CMCI and Machine Check Resources .
15-21
15.4
15.5
15.5.1
15.5.2
15.5.2.1
15.5.2.2
15.5.2.3
15.6
15.6.1
15.6.2
15.6.3
15.6.4
15.7
15.8
15.9
15.9.1
15.9.2
15.9.2.1
15.9.2.2
15.9.2.3
15.9.2.4
15.9.2.5
15.9.2.6
15.9.3
15.9.3.1
15.9.3.2
15.9.4
15.9.5
CMCI Initialization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-21
CMCI Threshold Management. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-22
CMCI Interrupt Handler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-23
RECOVERY OF UNCORRECTED RECOVERABLE (UCR) ERRORS . . . . . . . . . . . . . . . . . . . . . . 15-23
Detection of Software Error Recovery Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-24
UCR Error Reporting and Logging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-24
UCR Error Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-25
UCR Error Overwrite Rules. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-27
MACHINE-CHECK AVAILABILITY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-28
MACHINE-CHECK INITIALIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-28
INTERPRETING THE MCA ERROR CODES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-29
Simple Error Codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-30
Compound Error Codes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-31
Correction Report Filtering (F) Bit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-31
Transaction Type (TT) Sub-Field. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-32
Level (LL) Sub-Field. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-32
Request (RRRR) Sub-Field. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-32
Bus and Interconnect Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-33
Memory Controller Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-34
Architecturally Defined UCR Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-34
Architecturally Defined SRAO Errors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-34
Architecturally Defined SRAR Errors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-36
Multiple MCA Errors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-38
Machine-Check Error Codes Interpretation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-39
15.10 GUIDELINES FOR WRITING MACHINE-CHECK SOFTWARE . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-39
15.10.1
15.10.2
15.10.3
15.10.4
15.10.4.1
15.10.4.2
Machine-Check Exception Handler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-40
Pentium Processor Machine-Check Exception Handling . . . . . . . . . . . . . . . . . . . . . . . . . 15-41
Logging Correctable Machine-Check Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-42
Machine-Check Software Handler Guidelines for Error Recovery . . . . . . . . . . . . . . . . 15-44
Machine-Check Exception Handler for Error Recovery . . . . . . . . . . . . . . . . . . . . . . . 15-44
Corrected Machine-Check Handler for Error Recovery . . . . . . . . . . . . . . . . . . . . . . . 15-50
Vol. 3A xv
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
CHAPTER 16
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
16.1
OVERVIEW OF DEBUG SUPPORT FACILITIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-1
16.2
DEBUG REGISTERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-2
Debug Address Registers (DR0-DR3). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-4
Debug Registers DR4 and DR5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-4
Debug Status Register (DR6) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-4
Debug Control Register (DR7). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-5
16.2.1
16.2.2
16.2.3
16.2.4
16.2.5
16.2.6
16.3
16.3.1
16.3.1.1
16.3.1.2
16.3.1.3
16.3.1.4
16.3.1.5
16.3.2
16.4
16.4.1
16.4.2
16.4.3
16.4.4
16.4.5
16.4.6
16.4.7
16.4.8
16.4.8.1
16.4.8.2
16.4.8.3
16.4.9
16.4.9.1
16.4.9.2
16.4.9.3
16.4.9.4
16.4.9.5
16.5
Breakpoint Field Recognition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-6
®
Debug Registers and Intel 64 Processors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-8
DEBUG EXCEPTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-9
Debug Exception (#DB)—Interrupt Vector 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-9
Instruction-Breakpoint Exception Condition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-10
Data Memory and I/O Breakpoint Exception Conditions. . . . . . . . . . . . . . . . . . . . . . .16-12
General-Detect Exception Condition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-12
Single-Step Exception Condition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-12
Task-Switch Exception Condition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-13
Breakpoint Exception (#BP)—Interrupt Vector 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-13
LAST BRANCH, INTERRUPT, AND EXCEPTION RECORDING OVERVIEW. . . . . . . . . . . . . . 16-14
IA32_DEBUGCTL MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-14
Monitoring Branches, Exceptions, and Interrupts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-16
Single-Stepping on Branches, Exceptions, and Interrupts . . . . . . . . . . . . . . . . . . . . . . . .16-16
Branch Trace Messages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-17
Branch Trace Store (BTS). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-17
CPL-Qualified Branch Trace Mechanism. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-18
Freezing LBR and Performance Counters on PMI. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-18
LBR Stack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-19
®
LBR Stack and Intel 64 Processors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-20
LBR Stack and IA-32 Processors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-20
Last Exception Records and Intel 64 Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . .16-21
BTS and DS Save Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-21
DS Save Area and IA-32e Mode Operation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-25
Setting Up the DS Save Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-28
Setting Up the BTS Buffer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-29
Setting Up CPL-Qualified BTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-30
Writing the DS Interrupt Service Routine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-31
®
™
LAST BRANCH, INTERRUPT, AND EXCEPTION RECORDING (INTEL CORE 2 DUO AND
®
™
INTEL ATOM PROCESSOR FAMILY). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-32
LBR Stack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-33
16.5.1
16.6
®
™
LAST BRANCH, INTERRUPT, AND EXCEPTION RECORDING (INTEL CORE I7 PROCESSOR
FAMILY) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-33
LBR Stack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-34
Filtering of Last Branch Records . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-35
16.6.1
16.6.2
16.7
LAST BRANCH, INTERRUPT, AND EXCEPTION RECORDING (PROCESSORS BASED ON INTEL
NETBURST MICROARCHITECTURE). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-36
®
16.7.1
16.7.2
16.7.3
16.8
MSR_DEBUGCTLA MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-37
LBR Stack for Processors Based on Intel NetBurst Microarchitecture . . . . . . . . . . . .16-38
Last Exception Records . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-40
®
™
LAST BRANCH, INTERRUPT, AND EXCEPTION RECORDING (INTEL CORE SOLO AND
®
™
INTEL CORE DUO PROCESSORS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-41
xvi Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
16.9
LAST BRANCH, INTERRUPT, AND EXCEPTION
RECORDING (PENTIUM M PROCESSORS). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-43
16.10 LAST BRANCH, INTERRUPT, AND EXCEPTION
RECORDING (P6 FAMILY PROCESSORS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-45
16.10.1
16.10.2
16.10.3
DEBUGCTLMSR Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-45
Last Branch and Last Exception MSRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-46
Monitoring Branches, Exceptions, and Interrupts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-47
16.11 TIME-STAMP COUNTER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-48
16.11.1
16.11.2
Invariant TSC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-49
IA32_TSC_AUX Register and RDTSCP Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-50
CHAPTER 17
8086 EMULATION
17.1
REAL-ADDRESS MODE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-1
17.1.1
17.1.2
17.1.3
17.1.4
17.2
Address Translation in Real-Address Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17-3
Registers Supported in Real-Address Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17-4
Instructions Supported in Real-Address Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17-4
Interrupt and Exception Handling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17-6
VIRTUAL-8086 MODE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-8
Enabling Virtual-8086 Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17-9
Structure of a Virtual-8086 Task. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17-9
Paging of Virtual-8086 Tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-10
Protection within a Virtual-8086 Task . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-11
Entering Virtual-8086 Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-11
Leaving Virtual-8086 Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-14
Sensitive Instructions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-15
Virtual-8086 Mode I/O . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-15
I/O-Port-Mapped I/O . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-15
Memory-Mapped I/O . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-16
Special I/O Buffers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-16
INTERRUPT AND EXCEPTION HANDLING
17.2.1
17.2.2
17.2.3
17.2.4
17.2.5
17.2.6
17.2.7
17.2.8
17.2.8.1
17.2.8.2
17.2.8.3
17.3
IN VIRTUAL-8086 MODE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-16
Class 1—Hardware Interrupt and Exception Handling in Virtual-8086 Mode. . . . . . 17-18
Handling an Interrupt or Exception Through a Protected-Mode Trap or Interrupt Gate
17-18
17.3.1
17.3.1.1
17.3.1.2
Handling an Interrupt or Exception With an 8086 Program Interrupt or Exception
Handler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-20
Handling an Interrupt or Exception Through a Task Gate . . . . . . . . . . . . . . . . . . . . 17-21
Class 2—Maskable Hardware Interrupt Handling in Virtual-8086 Mode Using the Virtual
Interrupt Mechanism. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-22
Class 3—Software Interrupt Handling in Virtual-8086 Mode . . . . . . . . . . . . . . . . . . . . 17-24
Method 1: Software Interrupt Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-27
Methods 2 and 3: Software Interrupt Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-28
Method 4: Software Interrupt Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-28
Method 5: Software Interrupt Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-28
Method 6: Software Interrupt Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-29
PROTECTED-MODE VIRTUAL INTERRUPTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-30
17.3.1.3
17.3.2
17.3.3
17.3.3.1
17.3.3.2
17.3.3.3
17.3.3.4
17.3.3.5
17.4
Vol. 3A xvii
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
CHAPTER 18
MIXING 16-BIT AND 32-BIT CODE
18.1
DEFINING 16-BIT AND 32-BIT PROGRAM MODULES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-2
18.2
MIXING 16-BIT AND 32-BIT OPERATIONS WITHIN A CODE SEGMENT . . . . . . . . . . . . . . . . . 18-2
SHARING DATA AMONG MIXED-SIZE CODE SEGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-4
TRANSFERRING CONTROL AMONG MIXED-SIZE CODE SEGMENTS . . . . . . . . . . . . . . . . . . . . 18-4
Code-Segment Pointer Size. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-5
Stack Management for Control Transfer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-5
Controlling the Operand-Size Attribute For a Call . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-7
Passing Parameters With a Gate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-8
Interrupt Control Transfers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-8
Parameter Translation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-8
Writing Interface Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-9
18.3
18.4
18.4.1
18.4.2
18.4.2.1
18.4.2.2
18.4.3
18.4.4
18.4.5
CHAPTER 19
ARCHITECTURE COMPATIBILITY
19.1
19.2
19.3
19.4
19.5
19.6
19.7
19.8
19.9
PROCESSOR FAMILIES AND CATEGORIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-1
RESERVED BITS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-2
ENABLING NEW FUNCTIONS AND MODES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-2
DETECTING THE PRESENCE OF NEW FEATURES THROUGH SOFTWARE . . . . . . . . . . . . . . 19-3
INTEL MMX TECHNOLOGY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-3
STREAMING SIMD EXTENSIONS (SSE). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-3
STREAMING SIMD EXTENSIONS 2 (SSE2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-4
STREAMING SIMD EXTENSIONS 3 (SSE3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-4
ADDITIONAL STREAMING SIMD EXTENSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-4
19.10 INTEL HYPER-THREADING TECHNOLOGY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-5
19.11 MULTI-CORE TECHNOLOGY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-5
19.12 SPECIFIC FEATURES OF DUAL-CORE PROCESSOR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-5
19.13 NEW INSTRUCTIONS IN THE PENTIUM AND LATER IA-32 PROCESSORS . . . . . . . . . . . . . . 19-5
19.13.1
Instructions Added Prior to the Pentium Processor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-6
19.14 OBSOLETE INSTRUCTIONS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-7
19.15 UNDEFINED OPCODES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-7
19.16 NEW FLAGS IN THE EFLAGS REGISTER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-7
19.16.1
Using EFLAGS Flags to Distinguish Between 32-Bit IA-32 Processors . . . . . . . . . . . . . 19-8
19.17 STACK OPERATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-8
19.17.1
19.17.2
PUSH SP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-8
EFLAGS Pushed on the Stack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-9
19.18 X87 FPU . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-9
19.18.1
19.18.2
Control Register CR0 Flags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-9
x87 FPU Status Word . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-10
Condition Code Flags (C0 through C3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-10
Stack Fault Flag. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-11
x87 FPU Control Word . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-11
x87 FPU Tag Word. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-11
Data Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-12
NaNs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-12
Pseudo-zero, Pseudo-NaN, Pseudo-infinity, and Unnormal Formats . . . . . . . . . . .19-12
Floating-Point Exceptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-13
Denormal Operand Exception (#D) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-13
Numeric Overflow Exception (#O). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-13
19.18.2.1
19.18.2.2
19.18.3
19.18.4
19.18.5
19.18.5.1
19.18.5.2
19.18.6
19.18.6.1
19.18.6.2
xviii Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
19.18.6.3
19.18.6.4
19.18.6.5
19.18.6.6
19.18.6.7
19.18.6.8
19.18.6.9
19.18.6.10
19.18.6.11
19.18.6.12
19.18.6.13
19.18.6.14
19.18.7
Numeric Underflow Exception (#U) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-14
Exception Precedence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-14
CS and EIP For FPU Exceptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-14
FPU Error Signals. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-14
Assertion of the FERR# Pin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-15
Invalid Operation Exception On Denormals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-15
Alignment Check Exceptions (#AC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-16
Segment Not Present Exception During FLDENV . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-16
Device Not Available Exception (#NM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-16
Coprocessor Segment Overrun Exception. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-16
General Protection Exception (#GP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-16
Floating-Point Error Exception (#MF) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-16
Changes to Floating-Point Instructions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-17
FDIV, FPREM, and FSQRT Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-17
FSCALE Instruction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-17
FPREM1 Instruction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-17
FPREM Instruction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-17
FUCOM, FUCOMP, and FUCOMPP Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-17
FPTAN Instruction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-18
Stack Overflow. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-18
FSIN, FCOS, and FSINCOS Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-18
FPATAN Instruction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-18
F2XM1 Instruction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-18
FLD Instruction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-18
FXTRACT Instruction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-19
Load Constant Instructions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-19
FSETPM Instruction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-19
FXAM Instruction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-20
FSAVE and FSTENV Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-20
Transcendental Instructions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-20
Obsolete Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-20
19.18.7.1
19.18.7.2
19.18.7.3
19.18.7.4
19.18.7.5
19.18.7.6
19.18.7.7
19.18.7.8
19.18.7.9
19.18.7.10
19.18.7.11
19.18.7.12
19.18.7.13
19.18.7.14
19.18.7.15
19.18.7.16
19.18.8
19.18.9
19.18.10 WAIT/FWAIT Prefix Differences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-21
19.18.11 Operands Split Across Segments and/or Pages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-21
19.18.12 FPU Instruction Synchronization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-21
19.19 SERIALIZING INSTRUCTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-21
19.20 FPU AND MATH COPROCESSOR INITIALIZATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-22
®
®
19.20.1
19.20.2
Intel 387 and Intel 287 Math Coprocessor Initialization. . . . . . . . . . . . . . . . . . . . . 19-22
Intel486 SX Processor and Intel 487 SX Math Coprocessor Initialization . . . . . . . . . 19-22
19.21 CONTROL REGISTERS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-24
19.22 MEMORY MANAGEMENT FACILITIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-25
19.22.1
New Memory Management Control Flags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-25
Physical Memory Addressing Extension. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-25
Global Pages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-26
Larger Page Sizes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-26
CD and NW Cache Control Flags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-26
Descriptor Types and Contents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-26
Changes in Segment Descriptor Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-27
19.22.1.1
19.22.1.2
19.22.1.3
19.22.2
19.22.3
19.22.4
19.23 DEBUG FACILITIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-27
19.23.1
19.23.2
19.23.3
Differences in Debug Register DR6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-27
Differences in Debug Register DR7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-27
Debug Registers DR4 and DR5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-27
19.24 RECOGNITION OF BREAKPOINTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-28
Vol. 3A xix
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
19.25 EXCEPTIONS AND/OR EXCEPTION CONDITIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-28
19.25.1
19.25.2
Machine-Check Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-30
Priority OF Exceptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-30
19.26 INTERRUPTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-30
19.26.1
19.26.2
19.26.3
Interrupt Propagation Delay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-30
NMI Interrupts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-30
IDT Limit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-31
19.27 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC). . . . . . . . . . . . . . . . . . . . 19-31
19.27.1
19.27.2
Software Visible Differences Between the Local APIC and the 82489DX. . . . . . . . .19-31
New Features Incorporated in the Local APIC for the P6 Family and Pentium Processors
19-32
19.27.3
New Features Incorporated in the Local APIC of the Pentium 4 and Intel Xeon Processors
19-32
19.28 TASK SWITCHING AND TSS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-32
19.28.1
19.28.2
19.28.3
19.28.4
19.28.5
P6 Family and Pentium Processor TSS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-33
TSS Selector Writes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-33
Order of Reads/Writes to the TSS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-33
Using A 16-Bit TSS with 32-Bit Constructs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-33
Differences in I/O Map Base Addresses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-33
19.29 CACHE MANAGEMENT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-34
19.29.1
19.29.2
Self-Modifying Code with Cache Enabled . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-35
Disabling the L3 Cache . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-36
19.30 PAGING. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-36
19.30.1
19.30.2
19.30.3
Large Pages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-36
PCD and PWT Flags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-36
Enabling and Disabling Paging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-37
19.31 STACK OPERATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-37
19.31.1
19.31.2
19.31.3
19.31.4
19.32
Selector Pushes and Pops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-37
Error Code Pushes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-38
Fault Handling Effects on the Stack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-38
Interlevel RET/IRET From a 16-Bit Interrupt or Call Gate . . . . . . . . . . . . . . . . . . . . . . . .19-38
MIXING 16- AND 32-BIT SEGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-39
SEGMENT AND ADDRESS WRAPAROUND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-39
Segment Wraparound . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-40
STORE BUFFERS AND MEMORY ORDERING. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-40
BUS LOCKING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-42
BUS HOLD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-42
MODEL-SPECIFIC EXTENSIONS TO THE IA-32. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-42
Model-Specific Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-43
RDMSR and WRMSR Instructions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-43
Memory Type Range Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-43
Machine-Check Exception and Architecture. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-44
Performance-Monitoring Counters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-44
19.33
19.33.1
19.34
19.35
19.36
19.37
19.37.1
19.37.2
19.37.3
19.37.4
19.37.5
19.38 TWO WAYS TO RUN INTEL 286 PROCESSOR TASKS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-45
CHAPTER 20
INTRODUCTION TO VIRTUAL-MACHINE EXTENSIONS
20.1
20.2
20.3
20.4
OVERVIEW. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-1
VIRTUAL MACHINE ARCHITECTURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-1
INTRODUCTION TO VMX OPERATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-1
LIFE CYCLE OF VMM SOFTWARE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-2
xx Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
20.5
20.6
20.7
20.8
VIRTUAL-MACHINE CONTROL STRUCTURE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-3
DISCOVERING SUPPORT FOR VMX. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-3
ENABLING AND ENTERING VMX OPERATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-4
RESTRICTIONS ON VMX OPERATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-5
CHAPTER 21
VIRTUAL-MACHINE CONTROL STRUCTURES
21.1
OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-1
FORMAT OF THE VMCS REGION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-2
ORGANIZATION OF VMCS DATA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-3
GUEST-STATE AREA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-3
Guest Register State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21-3
Guest Non-Register State. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21-6
HOST-STATE AREA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-9
VM-EXECUTION CONTROL FIELDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-10
Pin-Based VM-Execution Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-10
Processor-Based VM-Execution Controls. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-11
Exception Bitmap. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-15
I/O-Bitmap Addresses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-15
Time-Stamp Counter Offset. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-16
Guest/Host Masks and Read Shadows for CR0 and CR4. . . . . . . . . . . . . . . . . . . . . . . . . 21-16
CR3-Target Controls. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-16
Controls for APIC Accesses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-17
MSR-Bitmap Address . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-18
Executive-VMCS Pointer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-19
Extended-Page-Table Pointer (EPTP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-19
Virtual-Processor Identifier (VPID). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-19
Controls for PAUSE-Loop Exiting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-20
VM-EXIT CONTROL FIELDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-20
VM-Exit Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-20
VM-Exit Controls for MSRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-22
VM-ENTRY CONTROL FIELDS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-23
VM-Entry Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-23
VM-Entry Controls for MSRs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-24
VM-Entry Controls for Event Injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-24
VM-EXIT INFORMATION FIELDS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-26
Basic VM-Exit Information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-26
Information for VM Exits Due to Vectored Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-27
Information for VM Exits That Occur During Event Delivery . . . . . . . . . . . . . . . . . . . . . 21-28
Information for VM Exits Due to Instruction Execution. . . . . . . . . . . . . . . . . . . . . . . . . . 21-29
VM-Instruction Error Field. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-29
SOFTWARE ACCESS TO THE VMCS AND RELATED STRUCTURES . . . . . . . . . . . . . . . . . . 21-30
Software Access to the Virtual-Machine Control Structure. . . . . . . . . . . . . . . . . . . . . . 21-30
VMREAD, VMWRITE, and Encodings of VMCS Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-31
Software Access to Related Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-33
VMXON Region. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-33
21.2
21.3
21.4
21.4.1
21.4.2
21.5
21.6
21.6.1
21.6.2
21.6.3
21.6.4
21.6.5
21.6.6
21.6.7
21.6.8
21.6.9
21.6.10
21.6.11
21.6.12
21.6.13
21.7
21.7.1
21.7.2
21.8
21.8.1
21.8.2
21.8.3
21.9
21.9.1
21.9.2
21.9.3
21.9.4
21.9.5
21.10
21.10.1
21.10.2
21.10.3
21.10.4
21.11 USING VMCLEAR TO INITIALIZE A VMCS REGION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-34
Vol. 3A xxi
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
CHAPTER 22
VMX NON-ROOT OPERATION
22.1
INSTRUCTIONS THAT CAUSE VM EXITS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-1
22.1.1
22.1.2
22.1.3
22.2
Relative Priority of Faults and VM Exits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-1
Instructions That Cause VM Exits Unconditionally. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-2
Instructions That Cause VM Exits Conditionally . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-3
APIC-ACCESS VM EXITS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-7
Linear Accesses to the APIC-Access Page . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-7
Linear Accesses That Cause APIC-Access VM Exits. . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-7
Priority of APIC-Access VM Exits Caused by Linear Accesses . . . . . . . . . . . . . . . . . . 22-9
Instructions That May Cause Page Faults or EPT Violations Without Accessing
22.2.1
22.2.1.1
22.2.1.2
22.2.1.3
Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22-10
Guest-Physical Accesses to the APIC-Access Page . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22-10
Guest-Physical Accesses That Might Not Cause APIC-Access VM Exits . . . . . . . .22-11
Priority of APIC-Access VM Exits Caused by Guest-Physical Accesses . . . . . . . . .22-12
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Physical Accesses to the APIC-Access Page22-12
VTPR Accesses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22-13
OTHER CAUSES OF VM EXITS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-14
CHANGES TO INSTRUCTION BEHAVIOR IN VMX NON-ROOT OPERATION . . . . . . . . . . . 22-16
APIC ACCESSES THAT DO NOT CAUSE VM EXITS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-21
Linear Accesses to the APIC-Access Page Using Large-Page Translations . . . . . . . .22-22
Physical Accesses to the APIC-Access Page. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22-22
VTPR Accesses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22-22
Treatment of Individual VTPR Accesses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22-23
Operations with Multiple Accesses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22-23
TPR-Shadow Updates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22-25
OTHER CHANGES IN VMX NON-ROOT OPERATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-25
Event Blocking. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22-25
Treatment of Task Switches. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22-26
FEATURES SPECIFIC TO VMX NON-ROOT OPERATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-27
VMX-Preemption Timer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22-27
Monitor Trap Flag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22-28
Translation of Guest-Physical Addresses Using EPT. . . . . . . . . . . . . . . . . . . . . . . . . . . . .22-29
UNRESTRICTED GUESTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-29
22.2.2
22.2.2.1
22.2.2.2
22.2.3
22.2.4
22.3
22.4
22.5
22.5.1
22.5.2
22.5.3
22.5.3.1
22.5.3.2
22.5.3.3
22.6
22.6.1
22.6.2
22.7
22.7.1
22.7.2
22.7.3
22.8
CHAPTER 23
VM ENTRIES
23.1
BASIC VM-ENTRY CHECKS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-2
23.2
CHECKS ON VMX CONTROLS AND HOST-STATE AREA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-3
Checks on VMX Controls. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-3
VM-Execution Control Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-3
VM-Exit Control Fields. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-6
VM-Entry Control Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-7
Checks on Host Control Registers and MSRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-8
Checks on Host Segment and Descriptor-Table Registers. . . . . . . . . . . . . . . . . . . . . . . . . 23-9
Checks Related to Address-Space Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-9
CHECKING AND LOADING GUEST STATE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-10
Checks on the Guest State Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23-10
Checks on Guest Control Registers, Debug Registers, and MSRs . . . . . . . . . . . . . .23-10
Checks on Guest Segment Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23-12
23.2.1
23.2.1.1
23.2.1.2
23.2.1.3
23.2.2
23.2.3
23.2.4
23.3
23.3.1
23.3.1.1
23.3.1.2
xxii Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
23.3.1.3
23.3.1.4
23.3.1.5
23.3.1.6
23.3.2
23.3.2.1
23.3.2.2
23.3.2.3
23.3.2.4
23.3.2.5
23.3.3
23.4
Checks on Guest Descriptor-Table Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-15
Checks on Guest RIP and RFLAGS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-15
Checks on Guest Non-Register State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-16
Checks on Guest Page-Directory-Pointer-Table Entries . . . . . . . . . . . . . . . . . . . . . . 23-18
Loading Guest State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-19
Loading Guest Control Registers, Debug Registers, and MSRs . . . . . . . . . . . . . . . . 23-19
Loading Guest Segment Registers and Descriptor-Table Registers . . . . . . . . . . . 23-21
Loading Guest RIP, RSP, and RFLAGS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-22
Loading Page-Directory-Pointer-Table Entries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-22
Updating Non-Register State. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-23
Clearing Address-Range Monitoring. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-23
LOADING MSRS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-23
EVENT INJECTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-24
Vectored-Event Injection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-24
Details of Vectored-Event Injection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-25
VM Exits During Event Injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-27
Event Injection for VM Entries to Real-Address Mode. . . . . . . . . . . . . . . . . . . . . . . . 23-28
Injection of Pending MTF VM Exits. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-28
SPECIAL FEATURES OF VM ENTRY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-28
Interruptibility State. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-29
Activity State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-30
Delivery of Pending Debug Exceptions after VM Entry. . . . . . . . . . . . . . . . . . . . . . . . . . 23-31
VMX-Preemption Timer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-32
Interrupt-Window Exiting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-32
NMI-Window Exiting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-32
VM Exits Induced by the TPR Shadow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-33
Pending MTF VM Exits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-34
VM Entries and Advanced Debugging Features. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-34
VM-ENTRY FAILURES DURING OR AFTER LOADING GUEST STATE . . . . . . . . . . . . . . . . . . 23-34
MACHINE CHECKS DURING VM ENTRY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-35
23.5
23.5.1
23.5.1.1
23.5.1.2
23.5.1.3
23.5.2
23.6
23.6.1
23.6.2
23.6.3
23.6.4
23.6.5
23.6.6
23.6.7
23.6.8
23.6.9
23.7
23.8
CHAPTER 24
VM EXITS
24.1
ARCHITECTURAL STATE BEFORE A VM EXIT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-1
24.2
RECORDING VM-EXIT INFORMATION AND UPDATING VM-ENTRY CONTROL FIELDS. . . 24-5
Basic VM-Exit Information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24-5
Information for VM Exits Due to Vectored Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-14
Information for VM Exits During Event Delivery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-15
Information for VM Exits Due to Instruction Execution. . . . . . . . . . . . . . . . . . . . . . . . . . 24-17
SAVING GUEST STATE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-26
Saving Control Registers, Debug Registers, and MSRs . . . . . . . . . . . . . . . . . . . . . . . . . . 24-27
Saving Segment Registers and Descriptor-Table Registers. . . . . . . . . . . . . . . . . . . . . . 24-27
Saving RIP, RSP, and RFLAGS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-28
Saving Non-Register State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-30
SAVING MSRS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-32
LOADING HOST STATE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-33
Loading Host Control Registers, Debug Registers, MSRs . . . . . . . . . . . . . . . . . . . . . . . . 24-33
Loading Host Segment and Descriptor-Table Registers . . . . . . . . . . . . . . . . . . . . . . . . . 24-34
Loading Host RIP, RSP, and RFLAGS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-36
Checking and Loading Host Page-Directory-Pointer-Table Entries . . . . . . . . . . . . . . . 24-36
Updating Non-Register State. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-37
24.2.1
24.2.2
24.2.3
24.2.4
24.3
24.3.1
24.3.2
24.3.3
24.3.4
24.4
24.5
24.5.1
24.5.2
24.5.3
24.5.4
24.5.5
Vol. 3A xxiii
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
24.5.6
24.6
24.7
24.8
Clearing Address-Range Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24-37
LOADING MSRS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-38
VMX ABORTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-38
MACHINE CHECK DURING VM EXIT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-40
CHAPTER 25
VMX SUPPORT FOR ADDRESS TRANSLATION
25.1
VIRTUAL PROCESSOR IDENTIFIERS (VPIDS). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25-1
THE EXTENDED PAGE TABLE MECHANISM (EPT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25-2
EPT Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25-2
EPT Translation Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25-4
EPT-Induced VM Exits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25-9
EPT Misconfigurations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25-10
EPT Violations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25-11
Prioritization of EPT-Induced VM Exits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25-12
EPT and Memory Typing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25-13
Memory Type Used for Accessing EPT Paging Structures . . . . . . . . . . . . . . . . . . . .25-14
Memory Type Used for Translated Guest-Physical Addresses . . . . . . . . . . . . . . . .25-14
CACHING TRANSLATION INFORMATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25-15
Information That May Be Cached . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25-15
Creating and Using Cached Translation Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25-16
Invalidating Cached Translation Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25-18
Operations that Invalidate Cached Mappings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25-18
Operations that Need Not Invalidate Cached Mappings . . . . . . . . . . . . . . . . . . . . . . .25-19
Guidelines for Use of the INVVPID Instruction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25-20
Guidelines for Use of the INVEPT Instruction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25-21
25.2
25.2.1
25.2.2
25.2.3
25.2.3.1
25.2.3.2
25.2.3.3
25.2.4
25.2.4.1
25.2.4.2
25.3
25.3.1
25.3.2
25.3.3
25.3.3.1
25.3.3.2
25.3.3.3
25.3.3.4
CHAPTER 26
SYSTEM MANAGEMENT MODE
26.1
SYSTEM MANAGEMENT MODE OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-1
System Management Mode and VMX Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-2
SYSTEM MANAGEMENT INTERRUPT (SMI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-3
SWITCHING BETWEEN SMM AND THE OTHER
26.1.1
26.2
26.3
PROCESSOR OPERATING MODES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-3
Entering SMM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-3
Exiting From SMM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-4
SMRAM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-5
SMRAM State Save Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-6
SMRAM State Save Map and Intel 64 Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-8
SMRAM Caching. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26-11
System Management Range Registers (SMRR). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26-12
SMI HANDLER EXECUTION ENVIRONMENT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-12
EXCEPTIONS AND INTERRUPTS WITHIN SMM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-14
MANAGING SYNCHRONOUS AND ASYNCHRONOUS
26.3.1
26.3.2
26.4
26.4.1
26.4.1.1
26.4.2
26.4.2.1
26.5
26.6
26.7
SYSTEM MANAGEMENT INTERRUPTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-15
I/O State Implementation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26-16
NMI HANDLING WHILE IN SMM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-17
SMM REVISION IDENTIFIER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-18
26.7.1
26.8
26.9
26.10 AUTO HALT RESTART. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-18
26.10.1
Executing the HLT Instruction in SMM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26-19
xxiv Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
26.11 SMBASE RELOCATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-19
26.11.1
Relocating SMRAM to an Address Above 1 MByte . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-20
26.12 I/O INSTRUCTION RESTART . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-20
26.12.1 Back-to-Back SMI Interrupts When I/O Instruction Restart Is Being Used. . . . . . . . . 26-22
26.13 SMM MULTIPLE-PROCESSOR CONSIDERATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-22
26.14 DEFAULT TREATMENT OF SMIS AND SMM WITH VMX OPERATION AND SMX OPERATION .
26-23
26.14.1
26.14.2
26.14.3
Default Treatment of SMI Delivery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-23
Default Treatment of RSM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-24
Protection of CR4.VMXE in SMM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-26
26.15 DUAL-MONITOR TREATMENT OF SMIs AND SMM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-26
26.15.1
26.15.2
Dual-Monitor Treatment Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-26
SMM VM Exits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-27
Architectural State Before a VM Exit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-27
Updating the Current-VMCS and Executive-VMCS Pointers. . . . . . . . . . . . . . . . . . . 26-27
Recording VM-Exit Information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-28
Saving Guest State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-29
Updating Non-Register State. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-29
Operation of an SMM Monitor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-30
VM Entries that Return from SMM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-30
Checks on the Executive-VMCS Pointer Field. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-30
Checks on VM-Execution Control Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-31
Checks on VM-Entry Control Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-31
Checks on the Guest State Area. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-32
Loading Guest State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-32
VMX-Preemption Timer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-32
Updating the Current-VMCS and SMM-Transfer VMCS Pointers. . . . . . . . . . . . . . . 26-33
VM Exits Induced by VM Entry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-33
SMI Blocking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-33
Failures of VM Entries That Return from SMM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-34
Enabling the Dual-Monitor Treatment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-34
Activating the Dual-Monitor Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-36
Initial Checks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-36
MSEG Checking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-38
Updating the Current-VMCS and Executive-VMCS Pointers. . . . . . . . . . . . . . . . . . . 26-38
Loading Host State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-38
Loading MSRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-40
Deactivating the Dual-Monitor Treatment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-40
26.15.2.1
26.15.2.2
26.15.2.3
26.15.2.4
26.15.2.5
26.15.3
26.15.4
26.15.4.1
26.15.4.2
26.15.4.3
26.15.4.4
26.15.4.5
26.15.4.6
26.15.4.7
26.15.4.8
26.15.4.9
26.15.4.10
26.15.5
26.15.6
26.15.6.1
26.15.6.2
26.15.6.3
26.15.6.4
26.15.6.5
26.15.7
26.16 SMI AND PROCESSOR EXTENDED STATE MANAGEMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-41
CHAPTER 27
VIRTUAL-MACHINE MONITOR PROGRAMMING CONSIDERATIONS
27.1
27.2
27.2.1
27.3
27.4
27.5
27.5.1
27.6
27.7
VMX SYSTEM PROGRAMMING OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27-1
SUPPORTING PROCESSOR OPERATING MODES IN GUEST ENVIRONMENTS. . . . . . . . . . . 27-1
Emulating Guest Execution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-2
MANAGING VMCS REGIONS AND POINTERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27-2
USING VMX INSTRUCTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27-5
VMM SETUP & TEAR DOWN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27-5
Algorithms for Determining VMX Capabilities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-6
PREPARATION AND LAUNCHING A VIRTUAL MACHINE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27-9
HANDLING OF VM EXITS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27-11
Vol. 3A xxv
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
27.7.1
27.7.1.1
27.7.1.2
27.8
27.8.1
27.8.2
27.8.3
27.8.4
27.8.5
27.9
27.9.1
27.9.2
27.9.2.1
27.9.2.2
27.9.3
27.9.4
27.9.5
Handling VM Exits Due to Exceptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-11
Reflecting Exceptions to Guest Software. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-11
Resuming Guest Software after Handling an Exception . . . . . . . . . . . . . . . . . . . . . .27-13
MULTI-PROCESSOR CONSIDERATIONS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27-15
Initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-15
Moving a VMCS Between Processors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-16
Paired Index-Data Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-16
External Data Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-17
CPUID Emulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-17
32-BIT AND 64-BIT GUEST ENVIRONMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27-17
Operating Modes of Guest Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-17
Handling Widths of VMCS Fields. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-18
Natural-Width VMCS Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-18
64-Bit VMCS Fields. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-18
IA-32e Mode Hosts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-19
IA-32e Mode Guests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-20
32-Bit Guests. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-21
27.10 HANDLING MODEL SPECIFIC REGISTERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27-21
27.10.1
Using VM-Execution Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-21
Using VM-Exit Controls for MSRs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-22
Using VM-Entry Controls for MSRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-22
Handling Special-Case MSRs and Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-23
Handling IA32_EFER MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-23
Handling the SYSENTER and SYSEXIT Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . .27-23
Handling the SYSCALL and SYSRET Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-23
Handling the SWAPGS Instruction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-24
Implementation Specific Behavior on Writing to Certain MSRs . . . . . . . . . . . . . . . .27-24
Handling Accesses to Reserved MSR Addresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27-24
27.10.2
27.10.3
27.10.4
27.10.4.1
27.10.4.2
27.10.4.3
27.10.4.4
27.10.4.5
27.10.5
27.11 HANDLING ACCESSES TO CONTROL REGISTERS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27-25
27.12 PERFORMANCE CONSIDERATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27-25
CHAPTER 28
VIRTUALIZATION OF SYSTEM RESOURCES
28.1
OVERVIEW. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-1
VIRTUALIZATION SUPPORT FOR DEBUGGING FACILITIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-1
Debug Exceptions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-2
MEMORY VIRTUALIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-3
Processor Operating Modes & Memory Virtualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-3
Guest & Host Physical Address Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-3
Virtualizing Virtual Memory by Brute Force. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-4
Alternate Approach to Memory Virtualization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-4
Details of Virtual TLB Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-6
Initialization of Virtual TLB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-7
Response to Page Faults . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-8
Response to Uses of INVLPG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .28-11
Response to CR3 Writes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .28-11
MICROCODE UPDATE FACILITY. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-11
Early Load of Microcode Updates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .28-12
Late Load of Microcode Updates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .28-12
28.2
28.2.1
28.3
28.3.1
28.3.2
28.3.3
28.3.4
28.3.5
28.3.5.1
28.3.5.2
28.3.5.3
28.3.5.4
28.4
28.4.1
28.4.2
xxvi Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
CHAPTER 29
HANDLING BOUNDARY CONDITIONS IN A VIRTUAL MACHINE MONITOR
29.1
OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29-1
INTERRUPT HANDLING IN VMX OPERATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29-1
EXTERNAL INTERRUPT VIRTUALIZATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29-3
Virtualization of Interrupt Vector Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29-4
Control of Platform Interrupts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29-5
PIC Virtualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29-6
xAPIC Virtualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29-6
Local APIC Virtualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29-6
I/O APIC Virtualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29-7
Virtualization of Message Signaled Interrupts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29-8
Examples of Handling of External Interrupts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29-8
Guest Setup. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29-8
Processor Treatment of External Interrupt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29-9
Processing of External Interrupts by VMM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29-9
Generation of Virtual Interrupt Events by VMM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29-10
ERROR HANDLING BY VMM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29-11
VM-Exit Failures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29-11
Machine Check Considerations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29-12
MCA Error Handling Guidelines for VMM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29-13
VMM Error Handling Strategies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29-13
Basic VMM MCA error recovery handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29-14
Implementation Considerations for the Basic Model . . . . . . . . . . . . . . . . . . . . . . . . . 29-14
MCA Virtualization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29-15
Implementation Considerations for the MCA Virtualization Model. . . . . . . . . . . . . 29-15
HANDLING ACTIVITY STATES BY VMM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29-15
29.2
29.3
29.3.1
29.3.2
29.3.2.1
29.3.2.2
29.3.2.3
29.3.2.4
29.3.2.5
29.3.3
29.3.3.1
29.3.3.2
29.3.3.3
29.3.3.4
29.4
29.4.1
29.4.2
29.4.3
29.4.3.1
29.4.3.2
29.4.3.3
29.4.3.4
29.4.3.5
29.5
CHAPTER 30
PERFORMANCE MONITORING
30.1
PERFORMANCE MONITORING OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-1
ARCHITECTURAL PERFORMANCE MONITORING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-2
Architectural Performance Monitoring Version 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-3
Architectural Performance Monitoring Version 1 Facilities . . . . . . . . . . . . . . . . . . . . .30-4
Additional Architectural Performance Monitoring Extensions. . . . . . . . . . . . . . . . . . . . . .30-6
Architectural Performance Monitoring Version 2 Facilities . . . . . . . . . . . . . . . . . . . . .30-6
Architectural Performance Monitoring Version 3 Facilities . . . . . . . . . . . . . . . . . . . 30-10
Pre-defined Architectural Performance Events. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-12
30.2
30.2.1
30.2.1.1
30.2.2
30.2.2.1
30.2.2.2
30.2.3
30.3
®
™
®
™
PERFORMANCE MONITORING (INTEL CORE SOLO AND INTEL CORE DUO
PROCESSORS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-14
®
™
30.4
PERFORMANCE MONITORING (PROCESSORS BASED ON INTEL CORE
MICROARCHITECTURE). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-16
Fixed-function Performance Counters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-18
Global Counter Control Facilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-19
At-Retirement Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-21
Precise Event Based Sampling (PEBS). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-22
Setting up the PEBS Buffer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-23
PEBS Record Format. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-23
Writing a PEBS Interrupt Service Routine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-23
30.4.1
30.4.2
30.4.3
30.4.4
30.4.4.1
30.4.4.2
30.4.4.3
Vol. 3A xxvii
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
®
™
30.5
30.6
PERFORMANCE MONITORING (PROCESSORS BASED ON INTEL ATOM
MICROARCHITECTURE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-25
PERFORMANCE MONITORING FOR PROCESSORS BASED ON INTEL MICROARCHITECTURE
®
(NEHALEM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-26
Enhancements of Performance Monitoring in the Processor Core . . . . . . . . . . . . . . . .30-27
Precise Event Based Sampling (PEBS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-27
Load Latency Performance Monitoring Facility. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-32
Off-core Response Performance Monitoring in the Processor Core. . . . . . . . . . . .30-34
Performance Monitoring Facility in the Uncore. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-37
Uncore Performance Monitoring Management Facility. . . . . . . . . . . . . . . . . . . . . . . .30-37
Uncore Performance Event Configuration Facility. . . . . . . . . . . . . . . . . . . . . . . . . . . .30-40
30.6.1
30.6.1.1
30.6.1.2
30.6.1.3
30.6.2
30.6.2.1
30.6.2.2
30.6.2.3
30.7
Uncore Address/Opcode Match MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-42
®
PERFORMANCE MONITORING FOR PROCESSORS BASED ON NEXT GENERATION INTEL
PROCESSOR (CODENAMED WESTMERE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-43
PERFORMANCE MONITORING (PROCESSORS
30.8
BASED ON INTEL NETBURST MICROARCHITECTURE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-44
ESCR MSRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-48
Performance Counters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-50
CCCR MSRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-51
Debug Store (DS) Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-53
Programming the Performance Counters
30.8.1
30.8.2
30.8.3
30.8.4
30.8.5
for Non-Retirement Events. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-54
Selecting Events to Count . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-54
Filtering Events. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-56
Starting Event Counting. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-58
Reading a Performance Counter’s Count . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-58
Halting Event Counting. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-59
Cascading Counters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-59
EXTENDED CASCADING. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-60
Generating an Interrupt on Overflow. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-62
Counter Usage Guideline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-62
At-Retirement Counting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-63
Using At-Retirement Counting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-64
Tagging Mechanism for Front_end_event . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-65
Tagging Mechanism For Execution_event . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-65
Tagging Mechanism for Replay_event. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-66
Precise Event-Based Sampling (PEBS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-66
Detection of the Availability of the PEBS Facilities . . . . . . . . . . . . . . . . . . . . . . . . . . .30-67
Setting Up the DS Save Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-67
Setting Up the PEBS Buffer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-67
Writing a PEBS Interrupt Service Routine. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-67
Other DS Mechanism Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-68
Operating System Implications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-68
PERFORMANCE MONITORING AND INTEL HYPER-THREADING TECHNOLOGY IN
PROCESSORS BASED ON INTEL NETBURST MICROARCHITECTURE . . . . . . . . . . . . . . . . . 30-68
ESCR MSRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-69
CCCR MSRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-70
IA32_PEBS_ENABLE MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-72
Performance Monitoring Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-72
30.8.5.1
30.8.5.2
30.8.5.3
30.8.5.4
30.8.5.5
30.8.5.6
30.8.5.7
30.8.5.8
30.8.5.9
30.8.6
30.8.6.1
30.8.6.2
30.8.6.3
30.8.6.4
30.8.7
30.8.7.1
30.8.7.2
30.8.7.3
30.8.7.4
30.8.7.5
30.8.8
30.9
30.9.1
30.9.2
30.9.3
30.9.4
30.10 COUNTING CLOCKS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-74
30.10.1
30.10.2
Non-Halted Clockticks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-75
Non-Sleep Clockticks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-76
xxviii Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
30.10.3
30.10.4
30.10.5
Incrementing the Time-Stamp Counter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-77
Non-Halted Reference Clockticks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-77
Cycle Counting and Opportunistic Processor Operation . . . . . . . . . . . . . . . . . . . . . . . . . 30-77
30.11 PERFORMANCE MONITORING, BRANCH PROFILING AND SYSTEM EVENTS . . . . . . . . . . 30-78
30.12 PERFORMANCE MONITORING AND DUAL-CORE TECHNOLOGY. . . . . . . . . . . . . . . . . . . . . . 30-79
30.13 PERFORMANCE MONITORING ON 64-BIT INTEL XEON PROCESSOR MP WITH UP TO 8-
MBYTE L3 CACHE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-79
30.14 PERFORMANCE MONITORING ON L3 AND CACHING BUS CONTROLLER SUB-SYSTEMS . 30-
84
30.14.1
30.14.2
30.14.3
30.14.4
30.14.4.1
30.14.5
Overview of Performance Monitoring with L3/Caching Bus Controller . . . . . . . . . . . 30-86
GBSQ Event Interface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-87
GSNPQ Event Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-89
FSB Event Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-91
FSB Sub-Event Mask Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-92
Common Event Control Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-93
30.15 PERFORMANCE MONITORING (P6 FAMILY PROCESSOR). . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-93
30.15.1
30.15.2
30.15.3
30.15.4
30.15.5
PerfEvtSel0 and PerfEvtSel1 MSRs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-94
PerfCtr0 and PerfCtr1 MSRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-96
Starting and Stopping the Performance-Monitoring Counters . . . . . . . . . . . . . . . . . . . 30-96
Event and Time-Stamp Monitoring Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-97
Monitoring Counter Overflow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-97
30.16 PERFORMANCE MONITORING (PENTIUM PROCESSORS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-98
30.16.1
30.16.2
30.16.3
Control and Event Select Register (CESR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-98
Use of the Performance-Monitoring Pins. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-100
Events Counted . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-100
APPENDIX A
PERFORMANCE-MONITORING EVENTS
A.1
A.2
A.3
ARCHITECTURAL PERFORMANCE-MONITORING EVENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-1
®
™
PERFORMANCE MONITORING EVENTS FOR INTEL CORE I7 PROCESSOR FAMILY. . A-2
®
PERFORMANCE MONITORING EVENTS FOR NEXT GENERATION INTEL PROCESSOR
(CODENAMED WESTMERE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-57
®
®
A.4
A.5
PERFORMANCE MONITORING EVENTS FOR INTEL XEON PROCESSOR 5200, 5400
®
™
SERIES AND INTEL CORE 2 EXTREME PROCESSORS QX 9000 SERIES . . . . . . . . . . . . A-87
®
®
PERFORMANCE MONITORING EVENTS FOR INTEL XEON PROCESSOR 3000, 3200,
®
™
5100, 5300 SERIES AND INTEL CORE 2 DUO PROCESSORS. . . . . . . . . . . . . . . . . . . . . . A-88
®
®
™
™
A.6
A.7
PERFORMANCE MONITORING EVENTS FOR INTEL ATOM PROCESSORS. . . . . . . . . A-133
®
™
PERFORMANCE MONITORING EVENTS FOR INTEL CORE SOLO AND INTEL CORE
DUO PROCESSORS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-156
PENTIUM 4 AND INTEL XEON PROCESSOR PERFORMANCE-MONITORING EVENTS. . . A-165
PERFORMANCE MONITORING EVENTS FOR
A.8
A.9
®
®
INTEL PENTIUM M PROCESSORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-214
P6 FAMILY PROCESSOR PERFORMANCE-MONITORING EVENTS . . . . . . . . . . . . . . . . . . . . A-217
PENTIUM PROCESSOR PERFORMANCE-
A.10
A.11
MONITORING EVENTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-235
APPENDIX B
MODEL-SPECIFIC REGISTERS (MSRS)
B.1
B.2
ARCHITECTURAL MSRS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-2
®
™
MSRS IN THE INTEL CORE 2 PROCESSOR FAMILY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-37
Vol. 3A xxix
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
®
™
B.3
B.4
B.5
B.5.1
B.6
B.7
B.8
B.9
MSRS IN THE INTEL ATOM PROCESSOR FAMILY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-58
®
MSRS IN THE INTEL MICROARCHITECTURE (NEHALEM). . . . . . . . . . . . . . . . . . . . . . . . . . . . B-73
MSRS IN THE PENTIUM 4 AND INTEL XEON PROCESSORS . . . . . . . . . . . . . . . . . . . . B-96
MSRs Unique to Intel Xeon Processor MP with L3 Cache . . . . . . . . . . . . . . . . . . . . . . . .B-136
MSRS IN INTEL CORE SOLO AND INTEL CORE DUO PROCESSORS . . . . . . . . . . B-139
MSRS IN THE PENTIUM M PROCESSOR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-152
MSRS IN THE P6 FAMILY PROCESSORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-162
MSRS IN PENTIUM PROCESSORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-174
®
®
®
®
™
®
™
APPENDIX C
MP INITIALIZATION FOR P6 FAMILY PROCESSORS
C.1
OVERVIEW OF THE MP INITIALIZATION PROCESS FOR P6 FAMILY PROCESSORS . . . . . . . C-1
MP INITIALIZATION PROTOCOL ALGORITHM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C-2
Error Detection and Handling During the MP Initialization Protocol . . . . . . . . . . . . . . . . . .C-4
C.2
C.2.1
APPENDIX D
PROGRAMMING THE LINT0 AND LINT1 INPUTS
D.1
D.2
CONSTANTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D-1
LINT[0:1] PINS PROGRAMMING PROCEDURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D-1
APPENDIX E
INTERPRETING MACHINE-CHECK
ERROR CODES
E.1
INCREMENTAL DECODING INFORMATION: PROCESSOR FAMILY 06H MACHINE ERROR
CODES FOR MACHINE CHECK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-1
INCREMENTAL DECODING INFORMATION: INTEL CORE 2 PROCESSOR FAMILY MACHINE
ERROR CODES FOR MACHINE CHECK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-5
Model-Specific Machine Check Error Codes for Intel Xeon Processor 7400 Series . . . .E-9
Processor Machine Check Status Register
E.2
E.2.1
E.2.1.1
Incremental MCA Error Code Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .E-9
Intel Xeon Processor 7400 Model Specific Error Code Field. . . . . . . . . . . . . . . . . . . . . . . E-10
Processor Model Specific Error Code Field
E.2.2
E.2.2.1
Type B: Bus and Interconnect Error. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-10
Processor Model Specific Error Code Field
E.2.2.2
E.3
Type C: Cache Bus Controller Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-10
INCREMENTAL DECODING INFORMATION: PROCESSOR FAMILY WITH CPUID
DISPLAYFAMILY_DISPLAYMODEL SIGNATURE 06_1AH, MACHINE ERROR CODES FOR
MACHINE CHECK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-11
QPI Machine Check Errors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-12
Internal Machine Check Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-13
Memory Controller Errors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-14
INCREMENTAL DECODING INFORMATION: PROCESSOR FAMILY 0FH MACHINE ERROR CODES
FOR MACHINE CHECK. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-15
Model-Specific Machine Check Error Codes for Intel Xeon Processor MP 7100 Series. . E-
16
E.3.1
E.3.2
E.3.3
E.4
E.4.1
E.4.1.1
E.4.2
Processor Machine Check Status Register
MCA Error Code Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-18
Other_Info Field (all MCA Error Types). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-19
xxx Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
E.4.3
Processor Model Specific Error Code Field. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .E-21
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . MCA Error Type A: L3 ErrorE-21
Processor Model Specific Error Code Field
E.4.3.1
E.4.3.2
Type B: Bus and Interconnect Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .E-21
Processor Model Specific Error Code Field
E.4.3.3
Type C: Cache Bus Controller Error. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .E-23
APPENDIX F
APIC BUS MESSAGE FORMATS
F.1
BUS MESSAGE FORMATS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F-1
EOI MESSAGE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F-1
Short Message . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F-2
Non-focused Lowest Priority Message. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F-3
APIC Bus Status Cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F-5
F.2
F.2.1
F.2.2
F.2.3
APPENDIX G
VMX CAPABILITY REPORTING FACILITY
G.1
BASIC VMX INFORMATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G-1
RESERVED CONTROLS AND DEFAULT SETTINGS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G-2
VM-EXECUTION CONTROLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G-3
Pin-Based VM-Execution Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G-3
Primary Processor-Based VM-Execution Controls. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G-4
Secondary Processor-Based VM-Execution Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G-5
VM-EXIT CONTROLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G-6
VM-ENTRY CONTROLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G-7
MISCELLANEOUS DATA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G-7
VMX-FIXED BITS IN CR0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G-8
VMX-FIXED BITS IN CR4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G-9
VMCS ENUMERATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G-9
VPID AND EPT CAPABILITIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G-9
G.2
G.3
G.3.1
G.3.2
G.3.3
G.4
G.5
G.6
G.7
G.8
G.9
G.10
APPENDIX H
FIELD ENCODING IN VMCS
H.1
16-BIT FIELDS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H-1
H.1.1
H.1.2
H.1.3
H.2
16-Bit Control Field. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .H-1
16-Bit Guest-State Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .H-1
16-Bit Host-State Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .H-2
64-BIT FIELDS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H-2
64-Bit Control Fields. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .H-3
64-Bit Read-Only Data Field. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .H-4
64-Bit Guest-State Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .H-4
64-Bit Host-State Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .H-5
32-BIT FIELDS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H-6
32-Bit Control Fields. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .H-6
32-Bit Read-Only Data Fields. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .H-7
32-Bit Guest-State Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .H-7
32-Bit Host-State Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .H-9
NATURAL-WIDTH FIELDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H-9
Natural-Width Control Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .H-9
H.2.1
H.2.2
H.2.3
H.2.4
H.3
H.3.1
H.3.2
H.3.3
H.3.4
H.4
H.4.1
Vol. 3A xxxi
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
H.4.2
H.4.3
H.4.4
Natural-Width Read-Only Data Fields. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H-10
Natural-Width Guest-State Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H-10
Natural-Width Host-State Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H-11
APPENDIX I
VMX BASIC EXIT REASONS
xxxii Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
FIGURES
Figure 1-1.
Figure 1-2.
Figure 2-1.
Figure 2-2.
Figure 2-3.
Figure 2-4.
Figure 2-5.
Figure 2-6.
Figure 2-7.
Figure 3-1.
Figure 3-2.
Figure 3-3.
Figure 3-4.
Figure 3-5.
Figure 3-6.
Figure 3-7.
Figure 3-8.
Figure 3-9.
Bit and Byte Order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-7
Syntax for CPUID, CR, and MSR Data Presentation. . . . . . . . . . . . . . . . . . . . . . . . . . . . .1-10
IA-32 System-Level Registers and Data Structures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3
System-Level Registers and Data Structures in IA-32e Mode . . . . . . . . . . . . . . . . . . . 2-4
Transitions Among the Processor’s Operating Modes . . . . . . . . . . . . . . . . . . . . . . . . . .2-11
System Flags in the EFLAGS Register. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-13
Memory Management Registers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-16
Control Registers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-19
XFEATURE_ENABLED_MASK Register (XCR0). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-26
Segmentation and Paging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-2
Flat Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4
Protected Flat Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4
Multi-Segment Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-6
Logical Address to Linear Address Translation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-9
Segment Selector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-10
Segment Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-11
Segment Descriptor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-13
Segment Descriptor When Segment-Present Flag Is Clear. . . . . . . . . . . . . . . . . . . . . .3-15
Figure 3-10. Global and Local Descriptor Tables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-20
Figure 3-11. Pseudo-Descriptor Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-22
Figure 4-1.
Figure 4-2.
Figure 4-3.
Figure 4-4.
Figure 4-5.
Figure 4-6.
Figure 4-7.
Figure 4-8.
Figure 4-9.
Enabling and Changing Paging Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-4
Linear-Address Translation to a 4-KByte Page using 32-Bit Paging. . . . . . . . . . . . .4-10
Linear-Address Translation to a 4-MByte Page using 32-Bit Paging . . . . . . . . . . . .4-11
Formats of CR3 and Paging-Structure Entries with 32-Bit Paging . . . . . . . . . . . . . .4-15
Linear-Address Translation to a 4-KByte Page using PAE Paging . . . . . . . . . . . . . . .4-18
Linear-Address Translation to a 2-MByte Page using PAE Paging. . . . . . . . . . . . . . .4-18
Formats of CR3 and Paging-Structure Entries with PAE Paging . . . . . . . . . . . . . . . .4-23
Linear-Address Translation to a 4-KByte Page using IA-32e Paging . . . . . . . . . . . .4-25
Linear-Address Translation to a 2-MByte Page using IA-32e Paging . . . . . . . . . . . .4-26
Figure 4-10. Formats of CR3 and Paging-Structure Entries with IA-32e Paging. . . . . . . . . . . . . .4-33
Figure 4-11. Page-Fault Error Code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-35
Figure 4-12. Memory Management Convention That Assigns a Page Table
to Each Segment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-53
Figure 5-1.
Figure 5-2.
Figure 5-3.
Figure 5-4.
Figure 5-5.
Figure 5-6.
Figure 5-7.
Descriptor Fields Used for Protection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-4
Descriptor Fields with Flags used in IA-32e Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-6
Protection Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-10
Privilege Check for Data Access . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-12
Examples of Accessing Data Segments From Various Privilege Levels . . . . . . . . . .5-13
Privilege Check for Control Transfer Without Using a Gate . . . . . . . . . . . . . . . . . . . . .5-15
Examples of Accessing Conforming and Nonconforming Code Segments From Various
Privilege Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-17
Call-Gate Descriptor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-19
Call-Gate Descriptor in IA-32e Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-21
Figure 5-8.
Figure 5-9.
Figure 5-10. Call-Gate Mechanism. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-22
Figure 5-11. Privilege Check for Control Transfer with Call Gate . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-23
Figure 5-12. Example of Accessing Call Gates At Various Privilege Levels . . . . . . . . . . . . . . . . . . .5-25
Figure 5-13. Stack Switching During an Interprivilege-Level Call . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-27
Figure 5-14. MSRs Used by SYSCALL and SYSRET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-33
Figure 5-15. Use of RPL to Weaken Privilege Level of Called Procedure. . . . . . . . . . . . . . . . . . . . .5-38
Figure 6-1.
Relationship of the IDTR and IDT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-14
Vol. 3A xxxiii
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
Figure 6-2.
Figure 6-3.
Figure 6-4.
Figure 6-5.
Figure 6-6.
Figure 6-7.
Figure 6-8.
Figure 6-9.
Figure 7-1.
Figure 7-2.
Figure 7-3.
Figure 7-4.
Figure 7-5.
Figure 7-6.
Figure 7-7.
Figure 7-8.
Figure 7-9.
IDT Gate Descriptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-15
Interrupt Procedure Call . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-16
Stack Usage on Transfers to Interrupt and Exception-Handling Routines. . . . . . . 6-18
Interrupt Task Switch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-21
Error Code. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-22
64-Bit IDT Gate Descriptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-23
IA-32e Mode Stack Usage After Privilege Level Change . . . . . . . . . . . . . . . . . . . . . . . 6-26
Page-Fault Error Code. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-55
Structure of a Task . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-2
32-Bit Task-State Segment (TSS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-5
TSS Descriptor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-7
Format of TSS and LDT Descriptors in 64-bit Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-9
Task Register. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-10
Task-Gate Descriptor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-11
Task Gates Referencing the Same Task . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-12
Nested Tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-17
Overlapping Linear-to-Physical Mappings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-20
Figure 7-10. 16-Bit TSS Format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-22
Figure 7-11. 64-Bit TSS Format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-24
Figure 8-1.
Figure 8-2.
Figure 8-3.
Figure 8-4.
Figure 8-5.
Figure 8-6.
Figure 8-7.
Example of Write Ordering in Multiple-Processor Systems. . . . . . . . . . . . . . . . . . . . . 8-10
Interpretation of APIC ID in Early MP Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-34
Local APICs and I/O APIC in MP System Supporting Intel HT Technology. . . . . . . . 8-38
IA-32 Processor with Two Logical Processors Supporting Intel HT Technology . 8-39
Generalized Four level Interpretation of the APIC ID . . . . . . . . . . . . . . . . . . . . . . . . . . 8-49
Conceptual Five-level Topology and 32-bit APIC ID Composition . . . . . . . . . . . . . . . 8-49
Topological Relationships between Hierarchical IDs in a Hypothetical MP Platform.8-
52
Figure 9-1.
Figure 9-2.
Figure 9-3.
Figure 9-4.
Contents of CR0 Register after Reset. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-5
Version Information in the EDX Register after Reset. . . . . . . . . . . . . . . . . . . . . . . . . . . 9-5
Processor State After Reset. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-21
Constructing Temporary GDT and Switching to Protected Mode (Lines 162-172 of
List File) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-31
Moving the GDT, IDT, and TSS from ROM to RAM (Lines 196-261 of List File). . . 9-32
Task Switching (Lines 282-296 of List File). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-33
Applying Microcode Updates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-37
Microcode Update Write Operation Flow [1] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-60
Microcode Update Write Operation Flow [2] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-61
Figure 9-5.
Figure 9-6.
Figure 9-7.
Figure 9-8.
Figure 9-9.
Figure 10-1. Relationship of Local APIC and I/O APIC In Single-Processor Systems. . . . . . . . . . . 10-3
Figure 10-2. Local APICs and I/O APIC When Intel Xeon Processors Are Used in Multiple-
Processor Systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-4
Figure 10-3. Local APICs and I/O APIC When P6 Family Processors Are Used in Multiple-Processor
Systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-4
Figure 10-4. Local APIC Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-7
Figure 10-5. IA32_APIC_BASE MSR (APIC_BASE_MSR in P6 Family) . . . . . . . . . . . . . . . . . . . . . . .10-12
Figure 10-6. Local APIC ID Register. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-13
Figure 10-7. Local APIC Version Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-16
Figure 10-8. IA32_APIC_BASE MSR Supporting x2APIC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-17
Figure 10-9. Spurious Interrupt Vector Register (SVR) of x2APIC . . . . . . . . . . . . . . . . . . . . . . . . .10-24
Figure 10-10. Local APIC Version Register of x2APIC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-24
Figure 10-11. Local x2APIC State Transitions with IA32_APIC_BASE, INIT, and RESET. . . . . . .10-26
Figure 10-12. Local Vector Table (LVT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-31
Figure 10-13. Error Status Register (ESR). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-34
xxxiv Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
Figure 10-14. Error Status Register (ESR) in x2APIC Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-36
Figure 10-15. Divide Configuration Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-37
Figure 10-16. Initial Count and Current Count Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-37
Figure 10-17. Interrupt Command Register (ICR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-39
Figure 10-18. Interrupt Command Register (ICR) in x2APIC Mode . . . . . . . . . . . . . . . . . . . . . . . . . . 10-45
Figure 10-19. Logical Destination Register (LDR). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-47
Figure 10-20. Destination Format Register (DFR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-48
Figure 10-21. Logical Destination Register in x2APIC Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-49
Figure 10-22. Arbitration Priority Register (APR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-51
Figure 10-23. SELF IPI register. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-53
Figure 10-24. Interrupt Acceptance Flow Chart for the Local APIC (Pentium 4 and Intel Xeon
Processors) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-54
Figure 10-25. Interrupt Acceptance Flow Chart for the Local APIC (P6 Family and Pentium
Processors) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-56
Figure 10-26. Task Priority Register (TPR). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-58
Figure 10-27. Processor Priority Register (PPR). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-59
Figure 10-28. IRR, ISR and TMR Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-60
Figure 10-29. EOI Register. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-61
Figure 10-30. CR8 Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-62
Figure 10-31. Spurious-Interrupt Vector Register (SVR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-64
Figure 10-32. Layout of the MSI Message Address Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-66
Figure 10-33. Layout of the MSI Message Data Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-67
Figure 11-1. Cache Structure of the Pentium 4 and Intel Xeon Processors . . . . . . . . . . . . . . . . . .11-1
Figure 11-2. Cache Structure of the Intel Core i7 Processors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11-2
Figure 11-3. Cache-Control Registers and Bits Available in Intel 64 and IA-32 Processors . . 11-16
Figure 11-4. Mapping Physical Memory With MTRRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-31
Figure 11-5. IA32_MTRRCAP Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-32
Figure 11-6. IA32_MTRR_DEF_TYPE MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-33
Figure 11-7. IA32_MTRR_PHYSBASEn and IA32_MTRR_PHYSMASKn Variable-Range Register
Pair. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-36
Figure 11-8. IA32_SMRR_PHYSBASE and IA32_SMRR_PHYSMASK SMRR Pair. . . . . . . . . . . . . 11-38
Figure 11-9. IA32_PAT MSR. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-49
Figure 12-1. Mapping of MMX Registers to Floating-Point Registers . . . . . . . . . . . . . . . . . . . . . . . .12-2
Figure 12-2. Mapping of MMX Registers to x87 FPU Data Register Stack . . . . . . . . . . . . . . . . . . .12-7
Figure 13-1. Example of Saving the x87 FPU, MMX, SSE, SSE2, SSE3, and SSSE3 State During an
Operating-System Controlled Task Switch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-11
Figure 13-2. Future Layout of XSAVE/XRSTOR Area and XSTATE_BV with Five Sets of Processor
State Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-14
Figure 13-3. OS Enabling of Processor Extended State Support . . . . . . . . . . . . . . . . . . . . . . . . . . 13-17
Figure 13-4. Application Detection of New Instruction Extensions and Processor Extended State
13-19
Figure 14-1. IA32_MPERF MSR and IA32_APERF MSR for P-state Coordination . . . . . . . . . . . . .14-2
Figure 14-2. IA32_PERF_CTL Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-6
Figure 14-3. Periodic Query of Activity Ratio of Opportunistic Processor Operation . . . . . . . . .14-7
Figure 14-4. IA32_ENERGY_PERF_BIAS Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-9
Figure 14-5. Processor Modulation Through Stop-Clock Mechanism . . . . . . . . . . . . . . . . . . . . . . . 14-11
Figure 14-6. MSR_THERM2_CTL Register On Processors with CPUID Family/Model/Stepping
Signature Encoded as 0x69n or 0x6Dn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-13
Figure 14-7. MSR_THERM2_CTL Register for Supporting TM2. . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-14
Figure 14-8. IA32_THERM_STATUS MSR. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-15
Figure 14-9. IA32_THERM_INTERRUPT MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-15
Figure 14-10. IA32_CLOCK_MODULATION MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-17
Vol. 3A xxxv
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
Figure 14-11. IA32_THERM_STATUS Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-19
Figure 14-12. IA32_THERM_INTERRUPT Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-21
Figure 15-1. Machine-Check MSRs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-3
Figure 15-2. IA32_MCG_CAP Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-4
Figure 15-3. IA32_MCG_STATUS Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-5
Figure 15-4. IA32_MCi_CTL Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-6
Figure 15-5. IA32_MCi_STATUS Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-8
Figure 15-6. IA32_MCi_ADDR MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-12
Figure 15-7. UCR Support in IA32_MCi_MISC Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-13
Figure 15-8. IA32_MCi_CTL2 Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-14
Figure 15-9. CMCI Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-19
Figure 15-10. Local APIC CMCI LVT Register. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-20
Figure 16-1. Debug Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-3
Figure 16-2. DR6/DR7 Layout on Processors Supporting Intel 64 Technology . . . . . . . . . . . . . . 16-9
Figure 16-3. IA32_DEBUGCTL MSR for Processors based
on Intel Core microarchitecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-15
Figure 16-4. 64-bit Address Layout of LBR MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-20
Figure 16-5. DS Save Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-23
Figure 16-6. 32-bit Branch Trace Record Format. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-24
Figure 16-7. PEBS Record Format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-25
Figure 16-8. IA-32e Mode DS Save Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-26
Figure 16-9. 64-bit Branch Trace Record Format. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-27
Figure 16-10. 64-bit PEBS Record Format. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-27
Figure 16-11. IA32_DEBUGCTL MSR for Processors based
on Intel microarchitecture (Nehalem). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-34
Figure 16-12. MSR_DEBUGCTLA MSR for Pentium 4 and Intel Xeon Processors . . . . . . . . . . . . .16-38
Figure 16-13. LBR MSR Branch Record Layout for the Pentium 4
and Intel Xeon Processor Family . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-40
Figure 16-14. IA32_DEBUGCTL MSR for Intel Core Solo
and Intel Core Duo Processors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-42
Figure 16-15. LBR Branch Record Layout for the Intel Core Solo
and Intel Core Duo Processor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-43
Figure 16-16. MSR_DEBUGCTLB MSR for Pentium M Processors. . . . . . . . . . . . . . . . . . . . . . . . . . . .16-44
Figure 16-17. LBR Branch Record Layout for the Pentium M Processor . . . . . . . . . . . . . . . . . . . . .16-45
Figure 16-18. DEBUGCTLMSR Register (P6 Family Processors) . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-46
Figure 17-1. Real-Address Mode Address Translation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-4
Figure 17-2. Interrupt Vector Table in Real-Address Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-7
Figure 17-3. Entering and Leaving Virtual-8086 Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17-13
Figure 17-4. Privilege Level 0 Stack After Interrupt or
Exception in Virtual-8086 Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17-19
Figure 17-5. Software Interrupt Redirection Bit Map in TSS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17-27
Figure 18-1. Stack after Far 16- and 32-Bit Calls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-6
Figure 19-1. I/O Map Base Address Differences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-34
Figure 20-1. Interaction of a Virtual-Machine Monitor and Guests . . . . . . . . . . . . . . . . . . . . . . . . . . 20-3
Figure 25-1. Formats of EPTP and EPT Paging-Structure Entries. . . . . . . . . . . . . . . . . . . . . . . . . .25-10
Figure 26-1. SMRAM Usage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-6
Figure 26-2. SMM Revision Identifier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26-18
Figure 26-3. Auto HALT Restart Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26-19
Figure 26-4. SMBASE Relocation Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26-20
Figure 26-5. I/O Instruction Restart Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26-21
Figure 27-1. VMX Transitions and States of VMCS in a Logical Processor . . . . . . . . . . . . . . . . . . . 27-4
Figure 28-1. Virtual TLB Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-7
xxxvi Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
Figure 29-1. Host External Interrupts and Guest Virtual Interrupts . . . . . . . . . . . . . . . . . . . . . . . . .29-5
Figure 30-1. Layout of IA32_PERFEVTSELx MSRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-4
Figure 30-2. Layout of IA32_FIXED_CTR_CTRL MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-7
Figure 30-3. Layout of IA32_PERF_GLOBAL_CTRL MSR. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-8
Figure 30-4. Layout of IA32_PERF_GLOBAL_STATUS MSR. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-9
Figure 30-5. Layout of IA32_PERF_GLOBAL_OVF_CTRL MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-9
Figure 30-6. Layout of IA32_PERFEVTSELx MSRs Supporting Architectural Performance
Monitoring Version 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-10
Figure 30-7. Layout of IA32_FIXED_CTR_CTRL MSR Supporting Architectural Performance
Monitoring Version 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-11
Figure 30-8. Layout of Global Performance Monitoring Control MSR . . . . . . . . . . . . . . . . . . . . . . 30-12
Figure 30-9. Layout of MSR_PERF_FIXED_CTR_CTRL MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-19
Figure 30-10. Layout of MSR_PERF_GLOBAL_CTRL MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-20
Figure 30-11. Layout of MSR_PERF_GLOBAL_STATUS MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-20
Figure 30-12. Layout of MSR_PERF_GLOBAL_OVF_CTRL MSR. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-21
Figure 30-13. Layout of IA32_PEBS_ENABLE MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-28
Figure 30-14. PEBS Programming Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-30
Figure 30-15. Layout of MSR_PEBS_LD_LAT MSR. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-34
Figure 30-16. Layout of MSR_OFFCORE_RSP_0 and MSR_OFFCORE_RSP_1 to Configure Off-core
Response Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-35
Figure 30-17. Layout of MSR_UNCORE_PERF_GLOBAL_CTRL MSR. . . . . . . . . . . . . . . . . . . . . . . . . 30-38
Figure 30-18. Layout of MSR_UNCORE_PERF_GLOBAL_STATUS MSR. . . . . . . . . . . . . . . . . . . . . . 30-39
Figure 30-19. Layout of MSR_UNCORE_PERF_GLOBAL_OVF_CTRL MSR . . . . . . . . . . . . . . . . . . . 30-39
Figure 30-20. Layout of MSR_UNCORE_PERFEVTSELx MSRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-40
Figure 30-21. Layout of MSR_UNCORE_FIXED_CTR_CTRL MSR . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-41
Figure 30-22. Layout of MSR_UNCORE_ADDR_OPCODE_MATCH MSR . . . . . . . . . . . . . . . . . . . . . . 30-42
Figure 30-23. Event Selection Control Register (ESCR) for Pentium 4
and Intel Xeon Processors without Intel HT Technology Support . . . . . . . . . . . . . 30-49
Figure 30-24. Performance Counter (Pentium 4 and Intel Xeon Processors) . . . . . . . . . . . . . . . . 30-51
Figure 30-25. Counter Configuration Control Register (CCCR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-52
Figure 30-26. Effects of Edge Filtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-58
Figure 30-27. Event Selection Control Register (ESCR) for the Pentium 4 Processor, Intel Xeon
Processor and Intel Xeon Processor MP Supporting Hyper-Threading Technology30-
69
Figure 30-28. Counter Configuration Control Register (CCCR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-71
Figure 30-29. Layout of IA32_PERF_CAPABILITIES MSR. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-79
Figure 30-30. Block Diagram of 64-bit Intel Xeon Processor MP with 8-MByte L3. . . . . . . . . . . 30-80
Figure 30-31. MSR_IFSB_IBUSQx, Addresses: 107CCH and 107CDH. . . . . . . . . . . . . . . . . . . . . . . . 30-81
Figure 30-32. MSR_IFSB_ISNPQx, Addresses: 107CEH and 107CFH. . . . . . . . . . . . . . . . . . . . . . . . 30-82
Figure 30-33. MSR_EFSB_DRDYx, Addresses: 107D0H and 107D1H . . . . . . . . . . . . . . . . . . . . . . . 30-83
Figure 30-34. MSR_IFSB_CTL6, Address: 107D2H;
MSR_IFSB_CNTR7, Address: 107D3H. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-84
Figure 30-35. Block Diagram of Intel Xeon Processor 7400 Series . . . . . . . . . . . . . . . . . . . . . . . . . 30-85
Figure 30-36. Block Diagram of Intel Xeon Processor 7100 Series . . . . . . . . . . . . . . . . . . . . . . . . . 30-86
Figure 30-37. MSR_EMON_L3_CTR_CTL0/1, Addresses: 107CCH/107CDH . . . . . . . . . . . . . . . . . 30-88
Figure 30-38. MSR_EMON_L3_CTR_CTL2/3, Addresses: 107CEH/107CFH. . . . . . . . . . . . . . . . . . 30-91
Figure 30-39. MSR_EMON_L3_CTR_CTL4/5/6/7, Addresses: 107D0H-107D3H. . . . . . . . . . . . . 30-92
Figure 30-40. PerfEvtSel0 and PerfEvtSel1 MSRs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-95
Figure 30-41. CESR MSR (Pentium Processor Only). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-99
Figure C-1.
MP System With Multiple Pentium III Processors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C-3
Vol. 3A xxxvii
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
TABLES
Table 2-1.
Action Taken By x87 FPU Instructions for Different
Combinations of EM, MP, and TS2-21
Table 2-2.
Table 3-1.
Table 3-2.
Table 4-1.
Table 4-3.
Table 4-2.
Table 4-4.
Table 4-5.
Table 4-6.
Table 4-7.
Table 4-8.
Table 4-9.
Table 4-10.
Table 4-11.
Table 4-12.
Table 4-13.
Summary of System Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-27
Code- and Data-Segment Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-17
System-Segment and Gate-Descriptor Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-19
Properties of Different Paging Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-3
Use of CR3 with 32-Bit Paging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-9
Paging Structures in the Different Paging Modes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-9
Format of a 32-Bit Page-Directory Entry that Maps a 4-MByte Page. . . . . . . . . . . 4-12
Format of a 32-Bit Page-Directory Entry that References a Page Table. . . . . . . . 4-13
Format of a 32-Bit Page-Table Entry that Maps a 4-KByte Page . . . . . . . . . . . . . . . 4-14
Use of CR3 with PAE Paging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-16
Format of an PAE Page-Directory-Pointer-Table Entry (PDPTE). . . . . . . . . . . . . . . . 4-17
Format of a PAE Page-Directory Entry that Maps a 2-MByte Page . . . . . . . . . . . . . 4-20
Format of a PAE Page-Directory Entry that References a Page Table . . . . . . . . . . 4-21
Format of a PAE Page-Table Entry that Maps a 4-KByte Page . . . . . . . . . . . . . . . . . 4-22
Use of CR3 with IA-32e Paging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-24
Format of an IA-32e PML4 Entry (PML4E) that References a Page-Directory-Pointer
Table4-27
Table 4-15.
Table 4-14.
Format of an IA-32e Page-Directory Entry that Maps a 2-MByte Page . . . . . . . . . 4-28
Format of an IA-32e Page-Directory-Pointer-Table Entry (PDPTE) that References a
Page Directory4-28
Table 4-16.
Table 4-17.
Table 5-1.
Table 5-2.
Table 5-3.
Table 5-4.
Table 5-5.
Format of an IA-32e Page-Directory Entry that References a Page Table . . . . . . 4-30
Format of an IA-32e Page-Table Entry that Maps a 4-KByte Page . . . . . . . . . . . . . 4-31
Privilege Check Rules for Call Gates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-23
64-Bit-Mode Stack Layout After CALLF with CPL Change. . . . . . . . . . . . . . . . . . . . . . 5-28
Combined Page-Directory and Page-Table Protection . . . . . . . . . . . . . . . . . . . . . . . . . 5-42
Extended Feature Enable MSR (IA32_EFER) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-43
IA-32e Mode Page Level Protection Matrix
with Execute-Disable Bit Capability5-44
Table 5-6.
Table 5-7.
Table 5-8.
Legacy PAE-Enabled 4-KByte Page Level Protection Matrix
with Execute-Disable Bit Capability5-45
Legacy PAE-Enabled 2-MByte Page Level Protection
with Execute-Disable Bit Capability5-45
IA-32e Mode Page Level Protection Matrix with Execute-Disable Bit Capability
Enabled5-46
Table 5-9.
Table 6-1.
Table 6-2.
Table 6-3.
Table 6-4.
Table 6-5.
Table 6-6.
Table 6-7.
Table 6-8.
Table 7-1.
Table 7-2.
Reserved Bit Checking WIth Execute-Disable Bit Capability Not Enabled. . . . . . . . 5-47
Protected-Mode Exceptions and Interrupts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-3
Priority Among Simultaneous Exceptions and Interrupts . . . . . . . . . . . . . . . . . . . . . . 6-11
Debug Exception Conditions and Corresponding Exception Classes. . . . . . . . . . . . . 6-29
Interrupt and Exception Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-38
Conditions for Generating a Double Fault. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-39
Invalid TSS Conditions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-42
Alignment Requirements by Data Type. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-60
SIMD Floating-Point Exceptions Priority . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-65
Exception Conditions Checked During a Task Switch . . . . . . . . . . . . . . . . . . . . . . . . . . 7-15
Effect of a Task Switch on Busy Flag, NT Flag,
Previous Task Link Field, and TS Flag7-17
Table 8-1.
Initial APIC IDs for the Logical Processors in a System that has Four Intel Xeon MP
18-52
Processors Supporting Intel Hyper-Threading Technology
xxxviii Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
Table 8-2.
Table 8-3.
Initial APIC IDs for the Logical Processors in a System that has Two Physical
Processors Supporting Dual-Core and Intel Hyper-Threading Technology8-53
Example of Possible x2APIC ID Assignment in a System that has Two Physical
Processors Supporting x2APIC and Intel Hyper-Threading Technology8-53
IA-32 Processor States Following Power-up, Reset, or INIT . . . . . . . . . . . . . . . . . . . . . 9-2
Recommended Settings of EM and MP Flags on IA-32 Processors . . . . . . . . . . . . . . . 9-7
Software Emulation Settings of EM, MP, and NE Flags . . . . . . . . . . . . . . . . . . . . . . . . . . 9-8
Main Initialization Steps in STARTUP.ASM Source Listing . . . . . . . . . . . . . . . . . . . . . .9-21
Relationship Between BLD Item and ASM Source File. . . . . . . . . . . . . . . . . . . . . . . . . .9-35
Microcode Update Field Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9-38
Microcode Update Format. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9-40
Extended Processor Signature Table Header Structure . . . . . . . . . . . . . . . . . . . . . . . .9-41
Processor Signature Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9-41
Processor Flags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9-43
Microcode Update Signature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9-48
Microcode Update Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9-55
Parameters for the Presence Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9-56
Parameters for the Write Update Data Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9-57
Parameters for the Control Update Sub-function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9-62
Parameters for the Read Microcode Update Data Function . . . . . . . . . . . . . . . . . . . .9-63
Mnemonic Values. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9-63
Return Code Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9-65
Local APIC Register Address Map. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-8
x2APIC Operating Mode Configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-17
Local APIC Register Address Map Supported by x2APIC. . . . . . . . . . . . . . . . . . . . . . 10-18
MSR/MMIO Interface of a Local x2APIC in Different Modes of Operation . . . . . . 10-22
ESR Flags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-35
Valid Combinations for the Pentium 4 and Intel Xeon Processors’
Table 9-1.
Table 9-2.
Table 9-3.
Table 9-4.
Table 9-5.
Table 9-6.
Table 9-7.
Table 9-8.
Table 9-9.
Table 9-10.
Table 9-11.
Table 9-12.
Table 9-13.
Table 9-14.
Table 9-15.
Table 9-17.
Table 9-16.
Table 9-18.
Table 10-1
Table 10-2.
Table 10-3.
Table 10-4.
Table 10-5.
Table 10-6
Local xAPIC Interrupt Command Register10-42
Table 10-7
Table 11-1.
Valid Combinations for the P6 Family Processors’
Local APIC Interrupt Command Register10-43
Characteristics of the Caches, TLBs, Store Buffer, and
Write Combining Buffer in Intel 64 and IA-32 Processors11-2
Table 11-2.
Table 11-3.
Memory Types and Their Properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11-9
Methods of Caching Available in Intel Core 2 Duo, Intel Atom, Intel Core Duo, Pentium
M, Pentium 4, Intel Xeon, P6 Family, and Pentium Processors11-10
Table 11-4.
Table 11-5.
Table 11-6.
MESI Cache Line States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-14
Cache Operating Modes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-17
Effective Page-Level Memory Type for Pentium Pro and
Pentium II Processors11-21
Table 11-7.
Effective Page-Level Memory Types for Pentium III and More Recent Processor
Families11-22
Memory Types That Can Be Encoded in MTRRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-30
Address Mapping for Fixed-Range MTRRs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-35
Table 11-8.
Table 11-9.
Table 11-10. Memory Types That Can Be Encoded With PAT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-49
Table 11-11. Selection of PAT Entries with PAT, PCD, and PWT Flags . . . . . . . . . . . . . . . . . . . . . 11-50
Table 11-12. Memory Type Setting of PAT Entries Following a Power-up or Reset. . . . . . . . . 11-50
Table 12-1.
Action Taken By MMX Instructions
for Different Combinations of EM, MP and TS12-1
Effects of MMX Instructions on x87 FPU State. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12-3
Effect of the MMX, x87 FPU, and FXSAVE/FXRSTOR Instructions on the
x87 FPU Tag Word12-4
Table 12-2.
Table 12-3.
Vol. 3A xxxix
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
Table 13-1.
Action Taken for Combinations of OSFXSR, OSXMMEXCPT, SSE, SSE2, SSE3, EM, MP,
and TS113-4
Table 13-2.
Table 13-3.
Table 13-4.
Table 13-5.
Table 14-1.
Table 15-1.
Action Taken for Combinations of OSFXSR, SSSE3, SSE4, EM, and TS . . . . . . . . . . 13-5
XSAVE Header Format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13-14
XRSTOR Action on MXCSR, x87 FPU, XMM Register. . . . . . . . . . . . . . . . . . . . . . . . . .13-16
XSAVE Action on MXCSR, x87 FPU, XMM Register . . . . . . . . . . . . . . . . . . . . . . . . . . .13-16
On-Demand Clock Modulation Duty Cycle Field Encoding. . . . . . . . . . . . . . . . . . . . . .14-17
Bits 54:53 in IA32_MCi_STATUS MSRs
when IA32_MCG_CAP[11] = 1 and UC = 015-9
Table 15-2.
Table 15-3.
Table 15-4.
Overwrite Rules for Enabled Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-11
Address Mode in IA32_MCi_MISC[8:6] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-13
Extended Machine Check State MSRs
in Processors Without Support for Intel 64 Architecture15-15
Table 15-5.
Extended Machine Check State MSRs
In Processors With Support For Intel 64 Architecture15-16
Table 15-6.
Table 15-7.
Table 15-8.
Table 15-9.
MC Error Classifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-26
Overwrite Rules for UC, CE, and UCR Errors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-27
IA32_MCi_Status [15:0] Simple Error Code Encoding . . . . . . . . . . . . . . . . . . . . . . . . .15-30
IA32_MCi_Status [15:0] Compound Error Code Encoding . . . . . . . . . . . . . . . . . . . . .15-31
Table 15-10. Encoding for TT (Transaction Type) Sub-Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-32
Table 15-11. Level Encoding for LL (Memory Hierarchy Level) Sub-Field . . . . . . . . . . . . . . . . . . .15-32
Table 15-12. Encoding of Request (RRRR) Sub-Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-33
Table 15-13. Encodings of PP, T, and II Sub-Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-33
Table 15-14. Encodings of MMM and CCCC Sub-Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-34
Table 15-15. MCA Compound Error Code Encoding for SRAO Errors . . . . . . . . . . . . . . . . . . . . . . . .15-35
Table 15-16. IA32_MCi_STATUS Values for SRAO Errors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-35
Table 15-17. IA32_MCG_STATUS Flag Indication for SRAO Errors . . . . . . . . . . . . . . . . . . . . . . . . .15-36
Table 15-18. MCA Compound Error Code Encoding for SRAR Errors . . . . . . . . . . . . . . . . . . . . . . . .15-36
Table 15-19. IA32_MCi_STATUS Values for SRAR Errors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-37
Table 15-20. IA32_MCG_STATUS Flag Indication for SRAR Errors. . . . . . . . . . . . . . . . . . . . . . . . . .15-37
Table 16-1.
Table 16-2.
Table 16-3.
Table 16-4.
Table 16-5.
Table 16-6.
Table 16-7.
Table 16-8.
Table 16-9.
Breakpoint Examples. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-7
Debug Exception Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-10
LBR Stack Size and TOS Pointer Range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-19
IA32_DEBUGCTL Flag Encodings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-29
CPL-Qualified Branch Trace Store Encodings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-30
IA32_LASTBRACH_x_FROM_IP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-35
IA32_LASTBRACH_x_TO_IP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-35
LBR Stack Size and TOS Pointer Range. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-35
MSR_LBR_SELECT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-35
Table 16-10. LBR MSR Stack Size and TOS Pointer Range for the Pentium® 4 and the
Intel® Xeon® Processor Family16-39
Table 17-1.
Table 17-2.
Table 18-1.
Table 19-1.
Real-Address Mode Exceptions and Interrupts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-8
Software Interrupt Handling Methods While in Virtual-8086 Mode. . . . . . . . . . . .17-26
Characteristics of 16-Bit and 32-Bit Program Modules . . . . . . . . . . . . . . . . . . . . . . . . 18-1
New Instruction in the Pentium Processor and
Later IA-32 Processors19-6
Table 19-2.
Recommended Values of the EM, MP, and NE Flags for Intel486 SX
Microprocessor/Intel 487 SX Math Coprocessor System19-22
Table 19-3.
Table 21-1.
Table 21-2.
Table 21-3.
EM and MP Flag Interpretation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19-23
Format of the VMCS Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-2
Format of Access Rights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-5
Format of Interruptibility State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-7
xl Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
Table 21-4.
Table 21-5.
Table 21-6.
Table 21-7.
Table 21-8.
Table 21-9.
Format of Pending-Debug-Exceptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21-8
Definitions of Pin-Based VM-Execution Controls. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-11
Definitions of Primary Processor-Based VM-Execution Controls . . . . . . . . . . . . . . 21-12
Definitions of Secondary Processor-Based VM-Execution Controls . . . . . . . . . . . 21-14
Format of Extended-Page-Table Pointer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-19
Definitions of VM-Exit Controls. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-20
Table 21-10. Format of an MSR Entry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-22
Table 21-11. Definitions of VM-Entry Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-23
Table 21-12. Format of the VM-Entry Interruption-Information Field . . . . . . . . . . . . . . . . . . . . . . 21-24
Table 21-13. Format of Exit Reason. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-26
Table 21-14. Format of the VM-Exit Interruption-Information Field. . . . . . . . . . . . . . . . . . . . . . . . 21-27
Table 21-15. Format of the IDT-Vectoring Information Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-28
Table 21-16. Structure of VMCS Component Encoding. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-31
Table 24-1.
Table 24-2.
Table 24-3.
Table 24-4.
Table 24-5.
Table 24-6.
Exit Qualification for Debug Exceptions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24-6
Exit Qualification for Task Switch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24-6
Exit Qualification for Control-Register Accesses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24-8
Exit Qualification for MOV DR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24-9
Exit Qualification for I/O Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24-9
Exit Qualification for APIC-Access VM Exits from Linear Accesses and Guest-Physical
Accesses24-10
Exit Qualification for EPT Violations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-11
Format of the VM-Exit Instruction-Information Field as Used for INS and OUTS . . 24-
18
Format of the VM-Exit Instruction-Information Field as Used for LIDT, LGDT, SIDT, or
SGDT24-19
Table 24-7.
Table 24-8.
Table 24-9.
Table 24-10. Format of the VM-Exit Instruction-Information Field as Used for LLDT, LTR, SLDT, and
STR24-21
Table 24-11. Format of the VM-Exit Instruction-Information Field as Used for VMCLEAR, VMPTRLD,
VMPTRST, and VMXON24-22
Table 24-12. Format of the VM-Exit Instruction-Information Field as Used for VMREAD and
VMWRITE24-23
Table 24-13. Format of the VM-Exit Instruction-Information Field as Used for INVEPT and INVVPID
24-25
Table 25-1.
Table 25-2.
Format of an EPT PML4 Entry (PML4E) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25-5
Format of an EPT Page-Directory-Pointer-Table Entry (PDPTE) that References an
EPT Page Directory25-6
Format of an EPT Page-Directory Entry (PDE) that Maps a 2-MByte Page. . . . . . .25-7
Format of an EPT Page-Directory Entry (PDE) that References an EPT Page Table . .
25-8
Table 25-3.
Table 25-4.
Table 25-5.
Table 26-1.
Table 26-2.
Table 26-3.
Table 26-4.
Table 26-5.
Table 26-6.
Table 26-7.
Table 26-8.
Table 26-9.
Format of an EPT Page-Table Entry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25-9
SMRAM State Save Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26-6
Processor Signatures and 64-bit SMRAM State Save Map Format. . . . . . . . . . . . . .26-9
SMRAM State Save Map for Intel 64 Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26-9
Processor Register Initialization in SMM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-13
I/O Instruction Information in the SMM State Save Map . . . . . . . . . . . . . . . . . . . . . . 26-16
I/O Instruction Type Encodings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-17
Auto HALT Restart Flag Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-19
I/O Instruction Restart Field Values. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-21
Exit Qualification for SMIs That Arrive Immediately
After the Retirement of an I/O Instruction26-28
Table 26-10. Format of MSEG Header . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-35
Table 27-1.
Operating Modes for Host and Guest Environments . . . . . . . . . . . . . . . . . . . . . . . . . 27-18
Vol. 3A xli
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
Table 30-1.
UMask and Event Select Encodings for Pre-Defined
Architectural Performance Events30-13
Table 30-2.
Table 30-3.
Table 30-4.
Table 30-5.
Table 30-6.
Table 30-7.
Table 30-8.
Core Specificity Encoding within a Non-Architectural Umask. . . . . . . . . . . . . . . . . .30-15
Agent Specificity Encoding within a Non-Architectural Umask . . . . . . . . . . . . . . . .30-15
HW Prefetch Qualification Encoding within a Non-Architectural Umask. . . . . . . .30-16
MESI Qualification Definitions within a Non-Architectural Umask. . . . . . . . . . . . . .30-16
Bus Snoop Qualification Definitions within a Non-Architectural Umask . . . . . . . .30-17
Snoop Type Qualification Definitions within a Non-Architectural Umask. . . . . . .30-17
Association of Fixed-Function Performance Counters with
Architectural Performance Events30-18
Table 30-10. PEBS Performance Events for Intel Core Microarchitecture. . . . . . . . . . . . . . . . . . .30-22
Table 30-9.
At-Retirement Performance Events for Intel Core Microarchitecture. . . . . . . . . .30-22
Table 30-11. Requirements to Program PEBS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-24
Table 30-12. PEBS Record Format for Intel Core i7 Processor Family . . . . . . . . . . . . . . . . . . . . . .30-28
Table 30-13. Data Source Encoding for Load Latency Record. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-33
Table 30-14. Off-Core Response Event Encoding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-35
Table 30-15. MSR_OFFCORE_RSP_Z Bit Field Definition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-35
Table 30-16. Opcode Field Encoding for MSR_UNCORE_ADDR_OPCODE_MATCH. . . . . . . . . . . .30-42
Table 30-17. Performance Counter MSRs and Associated CCCR and
ESCR MSRs (Pentium 4 and Intel Xeon Processors)30-45
Table 30-18. Event Example. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-54
Table 30-19. CCR Names and Bit Positions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30-60
Table 30-20. Effect of Logical Processor and CPL Qualification
for Logical-Processor-Specific (TS) Events30-73
Table 30-21. Effect of Logical Processor and CPL Qualification
for Non-logical-Processor-specific (TI) Events30-74
Table A-1.
Table A-2.
Architectural Performance Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-2
Non-Architectural Performance Events In the Processor Core for Intel Core i7
Processor and Intel Xeon Processor 5500 SeriesA-2
Non-Architectural Performance Events In the Processor Uncore for Intel Core i7
Processor and Intel Xeon Processor 5500 SeriesA-35
Non-Architectural Performance Events In Next Generation Processor Core
(Codenamed Westmere)A-58
Table A-3.
Table A-4.
Table A-5.
Table A-6.
Table A-7.
Non-Architectural Performance Events for Processors based on Enhanced Intel Core
MicroarchitectureA-88
Fixed-Function Performance Counter
and Pre-defined Performance EventsA-89
Non-Architectural Performance Events
in Processors Based on Intel Core MicroarchitectureA-90
Non-Architectural Performance Events for Intel Atom Processors . . . . . . . . . . . .A-133
Non-Architectural Performance Events
Table A-8.
Table A-9.
in Intel Core Solo and Intel Core Duo ProcessorsA-156
Performance Monitoring Events Supported by Intel NetBurst Microarchitecture for
Non-Retirement CountingA-165
Table A-10.
Table A-11.
Table A-12.
Table A-14.
Table A-13.
Performance Monitoring Events For Intel NetBurst
Microarchitecture for At-Retirement CountingA-197
Intel NetBurst Microarchitecture Model-Specific Performance Monitoring Events (For
Model Encoding 3, 4 or 6)A-204
List of Metrics Available for Execution Tagging
(For Execution Event Only)A-205
List of Metrics Available for Front_end Tagging
(For Front_end Event Only)A-205
xlii Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
Table A-15.
List of Metrics Available for Replay Tagging
(For Replay Event Only)A-206
Table A-16.
Table A-17.
Event Mask Qualification for Logical Processors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-208
Performance Monitoring Events on Intel® Pentium®
ProcessorsA-214
M
Table A-18.
Table A-19.
Performance Monitoring Events Modified on Intel® Pentium® M Processors . . A-216
Events That Can Be Counted with the P6 Family Performance-
Monitoring CountersA-218
Table A-20.
Events That Can Be Counted with Pentium Processor
Performance-Monitoring CountersA-235
Table B-1.
Table B-2.
Table B-3.
Table B-4.
Table B-5.
Table B-6.
Table B-7.
CPUID Signature Values of DisplayFamily_DisplayModel. . . . . . . . . . . . . . . . . . . . . . . . B-1
IA-32 Architectural MSRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-3
MSRs in Processors Based on Intel Core Microarchitecture . . . . . . . . . . . . . . . . . . . . .B-38
MSRs in Intel Atom Processor Family . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .B-58
MSRs in Processors Based on Intel Microarchitecture (Nehalem) . . . . . . . . . . . . . . .B-73
MSRs in the Pentium 4 and Intel Xeon Processors. . . . . . . . . . . . . . . . . . . . . . . . . . . . .B-96
MSRs Unique to 64-bit Intel Xeon Processor MP with
Up to an 8 MB L3 CacheB-136
Table B-8.
Table B-9.
MSRs Unique to Intel Xeon Processor 7100 Series . . . . . . . . . . . . . . . . . . . . . . . . . . B-138
MSRs in Intel Core Solo, Intel Core Duo Processors, and Dual-Core Intel Xeon
Processor LVB-139
Table B-10.
Table B-11.
Table B-12.
Table C-1.
Table E-1.
Table E-2.
MSRs in Pentium M Processors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-153
MSRs in the P6 Family Processors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-162
MSRs in the Pentium Processor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-174
Boot Phase IPI Message Format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C-2
CPUID DisplayFamily_DisplayModel Signatures for Family 6 . . . . . . . . . . . . . . . . . . . . E-1
Incremental Decoding Information: Processor Family 06H
Machine Error Codes For Machine CheckE-2
Table E-3.
Table E-4.
CPUID DisplayFamily_DisplayModel Signatures for Processors Based on Intel Core
MicroarchitectureE-5
Incremental Bus Error Codes of Machine Check for Processors Based on Intel Core
MicroarchitectureE-6
Table E-5.
Table E-6.
Table E-7.
Table E-8.
Table E-9.
Table E-10.
Table E-11.
Incremental MCA Error Code Types for Intel Xeon Processor 7400 . . . . . . . . . . . . . . E-9
Type B Bus and Interconnect Error Codes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .E-10
Type C Cache Bus Controller Error Codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .E-10
QPI Machine Check Error codes for IA32_MC0_STATUS and IA32_MC1_STATUSE-12
QPI Machine Check Error codes for IA32_MC0_MISC and IA32_MC1_MISC. . . . . . .E-13
Machine Check Error codes for IA32_MC7_STATUS. . . . . . . . . . . . . . . . . . . . . . . . . . . .E-13
Incremental Memory Controller Error Codes of Machine Check for IA32_MC8_STATUS
E-14
Table E-12.
Table E-13.
Incremental Memory Controller Error Codes of Machine Check for IA32_MC8_MISC E-
15
Incremental Decoding Information: Processor Family 0FH
Machine Error Codes For Machine CheckE-15
Table E-14.
Table E-15.
Table E-16.
Table E-17.
Table E-18.
Table E-19.
Table E-20.
Table F-1.
MCi_STATUS Register Bit Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .E-17
Incremental MCA Error Code for Intel Xeon Processor MP 7100 . . . . . . . . . . . . . . . .E-18
Other Information Field Bit Definition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .E-20
Type A: L3 Error Codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .E-21
Type B Bus and Interconnect Error Codes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .E-22
Type C Cache Bus Controller Error Codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .E-23
Decoding Family 0FH Machine Check Codes for Cache Hierarchy Errors . . . . . . . . .E-24
EOI Message (14 Cycles) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F-1
Vol. 3A xliii
Download from Www.Somanuals.com. All Manuals Search And Download.
CONTENTS
PAGE
Table F-2.
Table F-3.
Table F-4.
Table G-1.
Table H-1.
Table H-2.
Table H-3.
Table H-4.
Table H-5.
Table H-6.
Table H-7.
Table H-8.
Table H-9.
Table H-10.
Short Message (21 Cycles) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .F-2
Non-Focused Lowest Priority Message (34 Cycles). . . . . . . . . . . . . . . . . . . . . . . . . . . . . .F-3
APIC Bus Status Cycles Interpretation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .F-5
Memory Types Used For VMCS Access . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G-2
Encoding for 32-Bit Control Fields (0000_00xx_xxxx_xxx0B) . . . . . . . . . . . . . . . . . H-1
Encodings for 16-Bit Guest-State Fields (0000_10xx_xxxx_xxx0B). . . . . . . . . . . . H-1
Encodings for 16-Bit Host-State Fields (0000_11xx_xxxx_xxx0B) . . . . . . . . . . . . . H-2
Encodings for 64-Bit Control Fields (0010_00xx_xxxx_xxxAb). . . . . . . . . . . . . . . . . H-3
Encodings for 64-Bit Read-Only Data Field (0010_01xx_xxxx_xxxAb) . . . . . . . . . H-4
Encodings for 64-Bit Guest-State Fields (0010_10xx_xxxx_xxxAb) . . . . . . . . . . . . H-4
Encodings for 64-Bit Host-State Fields (0010_11xx_xxxx_xxxAb) . . . . . . . . . . . . . H-5
Encodings for 32-Bit Control Fields (0100_00xx_xxxx_xxx0B) . . . . . . . . . . . . . . . . H-6
Encodings for 32-Bit Read-Only Data Fields (0100_01xx_xxxx_xxx0B) . . . . . . . . H-7
Encodings for 32-Bit Guest-State Fields
(0100_10xx_xxxx_xxx0B)H-7
Table H-11.
Table H-12.
Table H-13.
Table H-14.
Table H-15.
Table I-1.
Encoding for 32-Bit Host-State Field (0100_11xx_xxxx_xxx0B) . . . . . . . . . . . . . . . H-9
Encodings for Natural-Width Control Fields (0110_00xx_xxxx_xxx0B) . . . . . . . . . H-9
Encodings for Natural-Width Read-Only Data Fields (0110_01xx_xxxx_xxx0B)H-10
Encodings for Natural-Width Guest-State Fields (0110_10xx_xxxx_xxx0B) . . . H-10
Encodings for Natural-Width Host-State Fields (0110_11xx_xxxx_xxx0B) . . . . H-11
Basic Exit Reasons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I-1
xliv Vol. 3A
Download from Www.Somanuals.com. All Manuals Search And Download.
CHAPTER 1
ABOUT THIS MANUAL
The Intel® 64 and IA-32 Architectures Software Developer’s Manual, Volume 3A:
System Programming Guide, Part 1 (order number 253668) and the Intel® 64 and
IA-32 Architectures Software Developer’s Manual, Volume 3B: System Programming
Guide, Part 2 (order number 253669) are part of a set that describes the architecture
and programming environment of Intel 64 and IA-32 Architecture processors. The
other volumes in this set are:
• Intel® 64 and IA-32 Architectures Software Developer’s Manual, Volume 1: Basic
Architecture (order number 253665).
• Intel® 64 and IA-32 Architectures Software Developer’s Manual, Volumes
2A & 2B: Instruction Set Reference (order numbers 253666 and 253667).
The Intel® 64 and IA-32 Architectures Software Developer’s Manual, Volume 1,
describes the basic architecture and programming environment of Intel 64 and IA-32
processors. The Intel® 64 and IA-32 Architectures Software Developer’s Manual,
Volumes 2A & 2B, describe the instruction set of the processor and the opcode struc-
ture. These volumes apply to application programmers and to programmers who
write operating systems or executives. The Intel® 64 and IA-32 Architectures Soft-
ware Developer’s Manual, Volumes 3A & 3B, describe the operating-system support
environment of Intel 64 and IA-32 processors. These volumes target operating-
system and BIOS designers. In addition, Intel® 64 and IA-32 Architectures Software
Developer’s Manual, Volume 3B, addresses the programming environment for
classes of software that host operating systems.
1.1
PROCESSORS COVERED IN THIS MANUAL
®
This manual set includes information pertaining primarily to the most recent Intel
64 and IA-32 processors, which include:
®
• Pentium processors
• P6 family processors
®
• Pentium 4 processors
®
• Pentium M processors
®
®
• Intel Xeon processors
®
• Pentium D processors
• Pentium processor Extreme Editions
®
®
®
• 64-bit Intel Xeon processors
®
• Intel Core™ Duo processor
®
• Intel Core™ Solo processor
Vol. 3 1-1
Download from Www.Somanuals.com. All Manuals Search And Download.
ABOUT THIS MANUAL
®
®
• Dual-Core Intel Xeon processor LV
®
• Intel Core™2 Duo processor
• Intel Core™2 Quad processor Q6000 series
• Intel Xeon processor 3000, 3200 series
• Intel Xeon processor 5000 series
• Intel Xeon processor 5100, 5300 series
• Intel Core™2 Extreme processor X7000 and X6800 series
• Intel Core™2 Extreme QX6000 series
®
®
®
®
®
®
®
®
®
®
®
• Intel Xeon processor 7100 series
®
®
• Intel Pentium Dual-Core processor
®
®
• Intel Xeon processor 7200, 7300 series
• Intel Core™2 Extreme QX9000 series
®
®
®
• Intel Xeon processor 5200, 5400, 7400 series
®
• Intel CoreTM2 Extreme processor QX9000 and X9000 series
®
• Intel CoreTM2 Quad processor Q9000 series
®
• Intel CoreTM2 Duo processor E8000, T9000 series
®
• Intel AtomTM processor family
®
• Intel CoreTM i7 processor
®
• Intel CoreTM i5 processor
P6 family processors are IA-32 processors based on the P6 family microarchitecture.
®
®
®
®
®
This includes the Pentium Pro, Pentium II, Pentium III, and Pentium III Xeon
processors.
®
®
®
The Pentium 4, Pentium D, and Pentium processor Extreme Editions are based
®
®
®
on the Intel NetBurst microarchitecture. Most early Intel Xeon processors are
based on the Intel NetBurst microarchitecture. Intel Xeon processor 5000, 7100
series are based on the Intel NetBurst microarchitecture.
®
®
®
®
®
®
The Intel Core™ Duo, Intel Core™ Solo and dual-core Intel Xeon processor LV
®
are based on an improved Pentium M processor microarchitecture.
®
®
®
The Intel Xeon processor 3000, 3200, 5100, 5300, 7200, and 7300 series, Intel
®
®
®
®
Pentium dual-core, Intel Core™2 Duo, Intel Core™2 Quad and Intel Core™2
®
Extreme processors are based on Intel Core™ microarchitecture.
®
®
®
The Intel Xeon processor 5200, 5400, 7400 series, Intel CoreTM2 Quad processor
®
®
Q9000 series, and Intel CoreTM2 Extreme processors QX9000, X9000 series, Intel
®
CoreTM2 processor E8000 series are based on Enhanced Intel CoreTM microarchitec-
ture.
®
®
The Intel AtomTM processor family is based on the Intel AtomTM microarchitecture
and supports Intel 64 architecture.
1-2 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ABOUT THIS MANUAL
®
®
The Intel CoreTM i7 processor and the Intel CoreTM i5 processor are based on the
®
Intel microarchitecture (Nehalem) and support Intel 64 architecture.
Processors based on the Next Generation Intel Processor, codenamed Westmere,
support Intel 64 architecture.
®
®
®
P6 family, Pentium M, Intel Core™ Solo, Intel Core™ Duo processors, dual-core
®
®
Intel Xeon processor LV, and early generations of Pentium 4 and Intel Xeon
®
processors support IA-32 architecture. The Intel AtomTM processor Z5xx series
support IA-32 architecture.
®
®
The Intel Xeon processor 3000, 3200, 5000, 5100, 5200, 5300, 5400, 7100,
®
®
7200, 7300, 7400 series, Intel Core™2 Duo, Intel Core™2 Extreme processors,
Intel Core 2 Quad processors, Pentium D processors, Pentium Dual-Core
®
®
processor, newer generations of Pentium 4 and Intel Xeon processor family support
®
Intel 64 architecture.
IA-32 architecture is the instruction set architecture and programming environment
®
for Intel's 32-bit microprocessors. Intel 64 architecture is the instruction set archi-
tecture and programming environment which is a superset of and compatible with
IA-32 architecture.
1.2
OVERVIEW OF THE SYSTEM PROGRAMMING GUIDE
A description of this manual’s content follows:
Chapter 1 — About This Manual. Gives an overview of all five volumes of the
Intel® 64 and IA-32 Architectures Software Developer’s Manual. It also describes
the notational conventions in these manuals and lists related Intel manuals and
documentation of interest to programmers and hardware designers.
Chapter 2 — System Architecture Overview. Describes the modes of operation
used by Intel 64 and IA-32 processors and the mechanisms provided by the architec-
tures to support operating systems and executives, including the system-oriented
registers and data structures and the system-oriented instructions. The steps neces-
sary for switching between real-address and protected modes are also identified.
Chapter 3 — Protected-Mode Memory Management. Describes the data struc-
tures, registers, and instructions that support segmentation and paging. The chapter
explains how they can be used to implement a “flat” (unsegmented) memory model
or a segmented memory model.
Chapter 4 — Paging. Describes the paging modes supported by Intel 64 and IA-32
processors.
Chapter 5 — Protection. Describes the support for page and segment protection
provided in the Intel 64 and IA-32 architectures. This chapter also explains the
implementation of privilege rules, stack switching, pointer validation, user and
supervisor modes.
Vol. 3 1-3
Download from Www.Somanuals.com. All Manuals Search And Download.
ABOUT THIS MANUAL
Chapter 6 — Interrupt and Exception Handling. Describes the basic interrupt
mechanisms defined in the Intel 64 and IA-32 architectures, shows how interrupts
and exceptions relate to protection, and describes how the architecture handles each
exception type. Reference information for each exception is given at the end of this
chapter.
Chapter 7 — Task Management. Describes mechanisms the Intel 64 and IA-32
architectures provide to support multitasking and inter-task protection.
Chapter 8 — Multiple-Processor Management. Describes the instructions and
flags that support multiple processors with shared memory, memory ordering, and
®
Intel Hyper-Threading Technology.
Chapter 9 — Processor Management and Initialization. Defines the state of an
Intel 64 or IA-32 processor after reset initialization. This chapter also explains how to
set up an Intel 64 or IA-32 processor for real-address mode operation and protected-
mode operation, and how to switch between modes.
Chapter 10 — Advanced Programmable Interrupt Controller (APIC).
Describes the programming interface to the local APIC and gives an overview of the
interface between the local APIC and the I/O APIC.
Chapter 11 — Memory Cache Control. Describes the general concept of caching
and the caching mechanisms supported by the Intel 64 or IA-32 architectures. This
chapter also describes the memory type range registers (MTRRs) and how they can
be used to map memory types of physical memory. Information on using the new
cache control and memory streaming instructions introduced with the Pentium III,
Pentium 4, and Intel Xeon processors is also given.
®
Chapter 12 — Intel MMX™ Technology System Programming. Describes
®
those aspects of the Intel MMX™ technology that must be handled and considered
at the system programming level, including: task switching, exception handling, and
compatibility with existing system environments.
Chapter 13 — System Programming For Instruction Set Extensions And
Processor Extended States. Describes the operating system requirements to
support SSE/SSE2/SSE3/SSSE3/SSE4 extensions, including task switching, excep-
tion handling, and compatibility with existing system environments. The latter part of
this chapter describes the extensible framework of operating system requirements to
support processor extended states. Processor extended state may be required by
instruction set extensions beyond those of SSE/SSE2/SSE3/SSSE3/SSE4 extensions.
Chapter 14 — Power and Thermal Management. Describes facilities of Intel 64
and IA-32 architecture used for power management and thermal monitoring.
Chapter 15 — Machine-Check Architecture. Describes the machine-check
architecture and machine-check exception mechanism found in the Pentium
4, Intel Xeon, and P6 family processors. Additionally, a signaling mechanism
for software to respond to hardware corrected machine check error is
covered.
1-4 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ABOUT THIS MANUAL
Chapter 16 — Debugging, Branch Profiles and Time-Stamp Counter.
Describes the debugging registers and other debug mechanism provided in Intel 64
or IA-32 processors. This chapter also describes the time-stamp counter.
Chapter 17 — 8086 Emulation. Describes the real-address and virtual-8086
modes of the IA-32 architecture.
Chapter 18 — Mixing 16-Bit and 32-Bit Code. Describes how to mix 16-bit and
32-bit code modules within the same program or task.
Chapter 19 — IA-32 Architecture Compatibility. Describes architectural
compatibility among IA-32 processors.
Chapter 20 — Introduction to Virtual-Machine Extensions. Describes the basic
elements of virtual machine architecture and the virtual-machine extensions for
Intel 64 and IA-32 Architectures.
Chapter 21 — Virtual-Machine Control Structures. Describes components that
manage VMX operation. These include the working-VMCS pointer and the control-
ling-VMCS pointer.
Chapter 22— VMX Non-Root Operation. Describes the operation of a VMX non-
root operation. Processor operation in VMX non-root mode can be restricted
programmatically such that certain operations, events or conditions can cause the
processor to transfer control from the guest (running in VMX non-root mode) to the
monitor software (running in VMX root mode).
Chapter 23 — VM Entries. Describes VM entries. VM entry transitions the processor
from the VMM running in VMX root-mode to a VM running in VMX non-root mode.
VM-Entry is performed by the execution of VMLAUNCH or VMRESUME instructions.
Chapter 24 — VM Exits. Describes VM exits. Certain events, operations or situa-
tions while the processor is in VMX non-root operation may cause VM-exit transitions.
In addition, VM exits can also occur on failed VM entries.
Chapter 25 — VMX Support for Address Translation. Describes virtual-machine
extensions that support address translation and the virtualization of physical
memory.
Chapter 26 — System Management Mode. Describes Intel 64 and IA-32 architec-
tures’ system management mode (SMM) facilities.
Chapter 27 — Virtual-Machine Monitoring Programming Considerations.
Describes programming considerations for VMMs. VMMs manage virtual machines
(VMs).
Chapter 28 — Virtualization of System Resources. Describes the virtualization
of the system resources. These include: debugging facilities, address translation,
physical memory, and microcode update facilities.
Chapter 29 — Handling Boundary Conditions in a Virtual Machine Monitor.
Describes what a VMM must consider when handling exceptions, interrupts, error
conditions, and transitions between activity states.
Vol. 3 1-5
Download from Www.Somanuals.com. All Manuals Search And Download.
ABOUT THIS MANUAL
Chapter 30 — Performance Monitoring. Describes the Intel 64 and IA-32 archi-
tectures’ facilities for monitoring performance.
Appendix A — Performance-Monitoring Events. Lists architectural performance
events. Non-architectural performance events (i.e. model-specific events) are listed
for each generation of microarchitecture.
Appendix B — Model-Specific Registers (MSRs). Lists the MSRs available in the
Pentium processors, the P6 family processors, the Pentium 4, Intel Xeon, Intel Core
Solo, Intel Core Duo processors, and Intel Core 2 processor family and describes
their functions.
Appendix C — MP Initialization For P6 Family Processors. Gives an example of
how to use of the MP protocol to boot P6 family processors in n MP system.
Appendix D — Programming the LINT0 and LINT1 Inputs. Gives an example of
how to program the LINT0 and LINT1 pins for specific interrupt vectors.
Appendix E — Interpreting Machine-Check Error Codes. Gives an example of
how to interpret the error codes for a machine-check error that occurred on a P6
family processor.
Appendix F — APIC Bus Message Formats. Describes the message formats for
messages transmitted on the APIC bus for P6 family and Pentium processors.
Appendix G — VMX Capability Reporting Facility. Describes the VMX capability
MSRs. Support for specific VMX features is determined by reading capability MSRs.
Appendix H — Field Encoding in VMCS. Enumerates all fields in the VMCS and
their encodings. Fields are grouped by width (16-bit, 32-bit, etc.) and type (guest-
state, host-state, etc.).
Appendix I — VM Basic Exit Reasons. Describes the 32-bit fields that encode
reasons for a VM exit. Examples of exit reasons include, but are not limited to: soft-
ware interrupts, processor exceptions, software traps, NMIs, external interrupts, and
triple faults.
1.3
NOTATIONAL CONVENTIONS
This manual uses specific notation for data-structure formats, for symbolic represen-
tation of instructions, and for hexadecimal and binary numbers. A review of this
notation makes the manual easier to read.
1.3.1
Bit and Byte Order
In illustrations of data structures in memory, smaller addresses appear toward the
bottom of the figure; addresses increase toward the top. Bit positions are numbered
from right to left. The numerical value of a set bit is equal to two raised to the power
of the bit position. Intel 64 and IA-32 processors are “little endian” machines; this
1-6 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ABOUT THIS MANUAL
means the bytes of a word are numbered starting from the least significant byte.
Figure 1-1 illustrates these conventions.
1.3.2
Reserved Bits and Software Compatibility
In many register and memory layout descriptions, certain bits are marked as
reserved. When bits are marked as reserved, it is essential for compatibility with
future processors that software treat these bits as having a future, though unknown,
effect. The behavior of reserved bits should be regarded as not only undefined, but
unpredictable. Software should follow these guidelines in dealing with reserved bits:
• Do not depend on the states of any reserved bits when testing the values of
registers which contain such bits. Mask out the reserved bits before testing.
• Do not depend on the states of any reserved bits when storing to memory or to a
register.
• Do not depend on the ability to retain information written into any reserved bits.
• When loading a register, always load the reserved bits with the values indicated
in the documentation, if any, or reload them with values previously read from the
same register.
NOTE
Avoid any software dependence upon the state of reserved bits in
Intel 64 and IA-32 registers. Depending upon the values of reserved
register bits will make software dependent upon the unspecified
manner in which the processor handles these bits. Programs that
depend upon reserved values risk incompatibility with future
processors.
Data Structure
Highest
Address
24 23
16 15
8 7
0
Bit offset
31
28
24
20
16
12
8
4
Lowest
Address
Byte 3
Byte 2
Byte 1
0
Byte 0
Byte Offset
Figure 1-1. Bit and Byte Order
Vol. 3 1-7
Download from Www.Somanuals.com. All Manuals Search And Download.
ABOUT THIS MANUAL
1.3.3
Instruction Operands
When instructions are represented symbolically, a subset of assembly language is
used. In this subset, an instruction has the following format:
label: mnemonic argument1, argument2, argument3
where:
• A label is an identifier which is followed by a colon.
• A mnemonic is a reserved name for a class of instruction opcodes which have
the same function.
• The operands argument1, argument2, and argument3 are optional. There
may be from zero to three operands, depending on the opcode. When present,
they take the form of either literals or identifiers for data items. Operand
identifiers are either reserved names of registers or are assumed to be assigned
to data items declared in another part of the program (which may not be shown
in the example).
When two operands are present in an arithmetic or logical instruction, the right
operand is the source and the left operand is the destination.
For example:
LOADREG: MOV EAX, SUBTOTAL
In this example LOADREG is a label, MOV is the mnemonic identifier of an opcode,
EAX is the destination operand, and SUBTOTAL is the source operand. Some
assembly languages put the source and destination in reverse order.
1.3.4
Hexadecimal and Binary Numbers
Base 16 (hexadecimal) numbers are represented by a string of hexadecimal digits
followed by the character H (for example, F82EH). A hexadecimal digit is a character
from the following set: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, A, B, C, D, E, and F.
Base 2 (binary) numbers are represented by a string of 1s and 0s, sometimes
followed by the character B (for example, 1010B). The “B” designation is only used in
situations where confusion as to the type of number might arise.
1.3.5
Segmented Addressing
The processor uses byte addressing. This means memory is organized and accessed
as a sequence of bytes. Whether one or more bytes are being accessed, a byte
address is used to locate the byte or bytes memory. The range of memory that can
be addressed is called an address space.
The processor also supports segmented addressing. This is a form of addressing
where a program may have many independent address spaces, called segments.
1-8 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ABOUT THIS MANUAL
For example, a program can keep its code (instructions) and stack in separate
segments. Code addresses would always refer to the code space, and stack
addresses would always refer to the stack space. The following notation is used to
specify a byte address within a segment:
Segment-register:Byte-address
For example, the following segment address identifies the byte at address FF79H in
the segment pointed by the DS register:
DS:FF79H
The following segment address identifies an instruction address in the code segment.
The CS register points to the code segment and the EIP register contains the address
of the instruction.
CS:EIP
1.3.6
Syntax for CPUID, CR, and MSR Values
Obtain feature flags, status, and system information by using the CPUID instruction,
by checking control register bits, and by reading model-specific registers. We are
moving toward a single syntax to represent this type of information. See Figure 1-2.
Vol. 3 1-9
Download from Www.Somanuals.com. All Manuals Search And Download.
ABOUT THIS MANUAL
6\QWD[ꢀ5HSUHVHQWDWLRQꢀIRUꢀ&38,'ꢀ,QSXWꢀDQGꢀ2XWSXW
&38,'ꢀꢁꢂ+ꢃꢂꢀ(&;ꢀ66(ꢃ>ELWꢃꢄꢅ@ꢃ ꢃꢂ
,QSXWꢃYDOXHꢃIRUꢃ($;ꢃGHILQHVꢃRXWSXW
ꢆ127(ꢌꢃ6RPHꢃOHDYHVꢃUHTXLUHꢃLQSXWꢃYDOXHVꢃIRU
($;ꢃDQGꢃ(&;ꢀꢃ,IꢃRQO\ꢃRQHꢃYDOXHꢃLVꢃSUHVHQWꢍꢃ
($;ꢃLVꢃLPSOLHGꢀꢇ
2XWSXWꢃUHJLVWHUꢃDQGꢃIHDWXUHꢃIODJꢃRUꢃ
ILHOGꢃQDPHꢃZLWKꢃELWꢃSRVLWLRQꢆVꢇ
9DOXHꢃꢆRUꢃUDQJHꢇꢃRIꢃRXWSXW
)RUꢀ&RQWUROꢀ5HJLVWHUꢀ9DOXHV
&5ꢈꢀ26);65>ELWꢃꢉ@ꢃ ꢃꢂ
([DPSOHꢃ&5ꢃQDPH
)HDWXUHꢃIODJꢃRUꢃILHOGꢃQDPHꢃ
ZLWKꢃELWꢃSRVLWLRQꢆVꢇ
9DOXHꢃꢆRUꢃUDQJHꢇꢃRIꢃRXWSXW
)RUꢀ0RGHOꢁ6SHFLILFꢀ5HJLVWHUꢀ9DOXHV
,$ꢊꢄB0,6&B(1$%/(6ꢀ(1$%/()23&2'(>ELWꢃꢄ@ꢃ ꢃꢂ
([DPSOHꢃ065ꢃQDPH
)HDWXUHꢃIODJꢃRUꢃILHOGꢃQDPHꢃZLWKꢃELWꢃSRVLWLRQꢆVꢇ
9DOXHꢃꢆRUꢃUDQJHꢇꢃRIꢃRXWSXW
20ꢂꢋꢋꢊꢄ
Figure 1-2. Syntax for CPUID, CR, and MSR Data Presentation
1.3.7
Exceptions
An exception is an event that typically occurs when an instruction causes an error.
For example, an attempt to divide by zero generates an exception. However, some
exceptions, such as breakpoints, occur under other conditions. Some types of excep-
tions may provide error codes. An error code reports additional information about the
error. An example of the notation used to show an exception and error code is shown
below:
#PF(fault code)
1-10 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ABOUT THIS MANUAL
This example refers to a page-fault exception under conditions where an error code
naming a type of fault is reported. Under some conditions, exceptions which produce
error codes may not be able to report an accurate code. In this case, the error code
is zero, as shown below for a general-protection exception:
#GP(0)
1.4
RELATED LITERATURE
Literature related to Intel 64 and IA-32 processors is listed on-line at:
Some of the documents listed at this web site can be viewed on-line; others can be
ordered. The literature available is listed by Intel processor and then by the following
literature types: applications notes, data sheets, manuals, papers, and specification
updates.
See also:
• The data sheet for a particular Intel 64 or IA-32 processor
• The specification update for a particular Intel 64 or IA-32 processor
®
http://www.intel.com/cd/software/products/asmo-na/eng/index.htm
®
http://www.intel.com/cd/software/products/asmo-na/eng/index.htm
®
http://www.intel.com/cd/software/products/asmo-na/eng/index.htm
®
®
http://developer.intel.com/products/processor/manuals/index.htm
®
http://www.intel.com/design/processor/applnots/241618.htm
®
• Intel 64 Architecture Memory Ordering White Paper,
http://developer.intel.com/products/processor/manuals/index.htm
®
• Intel 64 Architecture x2APIC Specification:
®
• Intel Virtualization Technology for Directed I/O, Rev 1.2 specification
http://download.intel.com/technology/computing/vptech/Intel(r)_VT_for_Direct_I
O.pdf
Vol. 3 1-11
Download from Www.Somanuals.com. All Manuals Search And Download.
ABOUT THIS MANUAL
®
• Intel 64 Architecture Processor Topology Enumeration:
®
• Intel Trusted Execution Technology Measured Launched Environment
Programming Guide, http://www.intel.com/technology/security/index.htm
• Developing Multi-threaded Applications: A Platform Consistent Approach
http://cache-
www.intel.com/cd/00/00/05/15/51534_developing_multithreaded_applications.pdf
http://www3.intel.com/cd/ids/developer/asmo-
na/eng/dc/threading/knowledgebase/19083.htm
More relevant links are:
• Software network link:
http://softwarecommunity.intel.com/isn/home/
• Developer centers:
• Processor support general link:
• Software products and packages:
®
• Intel 64 and IA-32 processor manuals (printed or PDF downloads):
®
• Intel multi-core technology:
http://developer.intel.com/multi-core/index.htm
®
®
• Intel Hyper-Threading Technology (Intel HT Technology):
http://developer.intel.com/technology/hyperthread/
1-12 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
CHAPTER 2
SYSTEM ARCHITECTURE OVERVIEW
IA-32 architecture (beginning with the Intel386 processor family) provides extensive
support for operating-system and system-development software. This support offers
multiple modes of operation, which include:
• Real mode, protected mode, virtual 8086 mode, and system management mode.
These are sometimes referred to as legacy modes.
Intel 64 architecture supports almost all the system programming facilities available
in IA-32 architecture and extends them to a new operating mode (IA-32e mode) that
supports a 64-bit programming environment. IA-32e mode allows software to
operate in one of two sub-modes:
• 64-bit mode supports 64-bit OS and 64-bit applications
• Compatibility mode allows most legacy software to run; it co-exists with 64-bit
applications under a 64-bit OS.
The IA-32 system-level architecture and includes features to assist in the following
operations:
• Memory management
• Protection of software modules
• Multitasking
• Exception and interrupt handling
• Multiprocessing
• Cache management
• Hardware resource and power management
• Debugging and performance monitoring
This chapter provides a description of each part of this architecture. It also describes
the system registers that are used to set up and control the processor at the system
level and gives a brief overview of the processor’s system-level (operating system)
instructions.
Many features of the system-level architectural are used only by system program-
mers. However, application programmers may need to read this chapter and the
following chapters in order to create a reliable and secure environment for applica-
tion programs.
This overview and most subsequent chapters of this book focus on protected-mode
operation of the IA-32 architecture. IA-32e mode operation of the Intel 64 architec-
ture, as it differs from protected mode operation, is also described.
All Intel 64 and IA-32 processors enter real-address mode following a power-up or
reset (see Chapter 9, “Processor Management and Initialization”). Software then
Vol. 3 2-1
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
initiates the switch from real-address mode to protected mode. If IA-32e mode oper-
ation is desired, software also initiates a switch from protected mode to IA-32e
mode.
2.1
OVERVIEW OF THE SYSTEM-LEVEL ARCHITECTURE
tions designed to support basic system-level operations such as memory manage-
ment, interrupt and exception handling, task management, and control of multiple
processors.
Figure 2-1 provides a summary of system registers and data structures that applies
to 32-bit modes. System registers and data structures that apply to IA-32e mode are
shown in Figure 2-2.
2-2 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
Physical Address
Linear Address
EFLAGS Register
Control Registers
Code, Data or
Stack Segment
Task-State
CR4
Segment Selector
Register
Segment (TSS)
Task
CR3
CR2
CR1
CR0
Code
Data
Stack
Global Descriptor
Table (GDT)
Task Register
Interrupt Handler
Segment Sel.
TSS Seg. Sel.
Seg. Desc.
TSS Desc.
Seg. Desc.
TSS Desc.
LDT Desc.
Code
Current
Stack
TSS
Interrupt
Vector
Task-State
Interrupt Descriptor
Segment (TSS)
Table (IDT)
Task
Code
Data
Interrupt Gate
Task Gate
Stack
GDTR
Trap Gate
Exception Handler
Local Descriptor
Table (LDT)
Code
Current
Stack
TSS
Call-Gate
Segment Selector
Seg. Desc.
IDTR
Call Gate
Protected Procedure
Code
XCR0 (XFEM)
Current
TSS
LDTR
Stack
Linear Address Space
Linear Addr.
Linear Address
Table
Dir
Offset
Page Directory
Pg. Dir. Entry
Page Table
Page
Physical Addr.
Pg. Tbl. Entry
0
This page mapping example is for 4-KByte pages
and the normal 32-bit physical address size.
CR3*
*Physical Address
Figure 2-1. IA-32 System-Level Registers and Data Structures
Vol. 3 2-3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
RFLAGS
Physical Address
Linear Address
Code, Data or Stack
Segment (Base =0)
Control Register
CR8
CR4
CR3
CR2
CR1
Task-State
Segment (TSS)
Segment Selector
Register
CR0
Global Descriptor
Table (GDT)
Task Register
Interrupt Handler
Code
Segment Sel.
Seg. Desc.
TSS Desc.
Seg. Desc.
Seg. Desc.
LDT Desc.
NULL
Stack
TR
Interrupt
Vector
Interrupt Descriptor
Table (IDT)
Interr. Handler
Code
Current TSS
Interrupt Gate
Interrupt Gate
Trap Gate
Stack
GDTR
IST
Exception Handler
Code
Local Descriptor
Table (LDT)
NULL
NULL
Stack
Call-Gate
Segment Selector
Seg. Desc.
Call Gate
IDTR
Protected Procedure
Code
Stack
XCR0 (XFEM)
LDTR
Linear Address Space
Linear Addr.
Linear Address
Dir. Pointer Directory Table
PML4
Offset
PML4
Pg. Dir. Ptr. Page Dir.
Page
Page Table
Physical
Addr.
Pg. Dir.
Entry
PML4.
Entry
Page Tbl
Entry
0
This page mapping example is for 4-KByte pages
and 40-bit physical address size.
CR3*
*Physical Address
Figure 2-2. System-Level Registers and Data Structures in IA-32e Mode
2-4 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
2.1.1
Global and Local Descriptor Tables
When operating in protected mode, all memory accesses pass through either the
global descriptor table (GDT) or an optional local descriptor table (LDT) as shown in
Figure 2-1. These tables contain entries called segment descriptors. Segment
descriptors provide the base address of segments well as access rights, type, and
usage information.
Each segment descriptor has an associated segment selector. A segment selector
provides the software that uses it with an index into the GDT or LDT (the offset of its
associated segment descriptor), a global/local flag (determines whether the selector
points to the GDT or the LDT), and access rights information.
To access a byte in a segment, a segment selector and an offset must be supplied.
The segment selector provides access to the segment descriptor for the segment (in
address of the segment in the linear address space. The offset then provides the
location of the byte relative to the base address. This mechanism can be used to
access any valid code, data, or stack segment, provided the segment is accessible
from the current privilege level (CPL) at which the processor is operating. The CPL is
defined as the protection level of the currently executing code segment.
See Figure 2-1. The solid arrows in the figure indicate a linear address, dashed lines
indicate a segment selector, and the dotted arrows indicate a physical address. For
simplicity, many of the segment selectors are shown as direct pointers to a segment.
However, the actual path from a segment selector to its associated segment is always
through a GDT or LDT.
The linear address of the base of the GDT is contained in the GDT register (GDTR);
2.1.1.1
Global and Local Descriptor Tables in IA-32e Mode
GDTR and LDTR registers are expanded to 64-bits wide in both IA-32e sub-modes
(64-bit mode and compatibility mode). For more information: see Section 3.5.2,
“Segment Descriptor Tables in IA-32e Mode.”
Global and local descriptor tables are expanded in 64-bit mode to support 64-bit base
addresses, (16-byte LDT descriptors hold a 64-bit base address and various
attributes). In compatibility mode, descriptors are not expanded.
2.1.2
System Segments, Segment Descriptors, and Gates
Besides code, data, and stack segments that make up the execution environment of
a program or procedure, the architecture defines two system segments: the task-
state segment (TSS) and the LDT. The GDT is not considered a segment because it is
not accessed by means of a segment selector and segment descriptor. TSSs and LDTs
have segment descriptors defined for them.
Vol. 3 2-5
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
The architecture also defines a set of special descriptors called gates (call gates,
interrupt gates, trap gates, and task gates). These provide protected gateways to
system procedures and handlers that may operate at a different privilege level than
application programs and most procedures. For example, a CALL to a call gate can
provide access to a procedure in a code segment that is at the same or a numerically
lower privilege level (more privileged) than the current code segment. To access a
1
procedure through a call gate, the calling procedure supplies the selector for the call
gate. The processor then performs an access rights check on the call gate, comparing
the CPL with the privilege level of the call gate and the destination code segment
pointed to by the call gate.
If access to the destination code segment is allowed, the processor gets the segment
selector for the destination code segment and an offset into that code segment from
the call gate. If the call requires a change in privilege level, the processor also
switches to the stack for the targeted privilege level. The segment selector for the
new stack is obtained from the TSS for the currently running task. Gates also facili-
tate transitions between 16-bit and 32-bit code segments, and vice versa.
2.1.2.1
Gates in IA-32e Mode
In IA-32e mode, the following descriptors are 16-byte descriptors (expanded to allow
a 64-bit base): LDT descriptors, 64-bit TSSs, call gates, interrupt gates, and trap
gates.
gates are not supported in IA-32e mode. On privilege level changes, stack segment
selectors are not read from the TSS. Instead, they are set to NULL.
2.1.3
Task-State Segments and Task Gates
The TSS (see Figure 2-1) defines the state of the execution environment for a task.
It includes the state of general-purpose registers, segment registers, the EFLAGS
register, the EIP register, and segment selectors with stack pointers for three stack
segments (one stack for each privilege level). The TSS also includes the segment
selector for the LDT associated with the task and the base address of the paging-
structure hierarchy.
All program execution in protected mode happens within the context of a task (called
the current task). The segment selector for the TSS for the current task is stored in
the task register. The simplest method for switching to a task is to make a call or
jump to the new task. Here, the segment selector for the TSS of the new task is given
in the CALL or JMP instruction. In switching tasks, the processor performs the
following actions:
1. Stores the state of the current task in the current TSS.
1. The word “procedure” is commonly used in this document as a general term for a logical unit or
block of code (such as a program, procedure, function, or routine).
2-6 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
2. Loads the task register with the segment selector for the new task.
3. Accesses the new TSS through a segment descriptor in the GDT.
4. Loads the state of the new task from the new TSS into the general-purpose
registers, the segment registers, the LDTR, control register CR3 (base address of
the paging-structure hierarchy), the EFLAGS register, and the EIP register.
5. Begins execution of the new task.
A task can also be accessed through a task gate. A task gate is similar to a call gate,
except that it provides access (through a segment selector) to a TSS rather than a
code segment.
2.1.3.1
Task-State Segments in IA-32e Mode
Hardware task switches are not supported in IA-32e mode. However, TSSs continue
to exist. The base address of a TSS is specified by its descriptor.
• Stack pointer addresses for each privilege level
• Pointer addresses for the interrupt stack table
• Offset address of the IO-permission bitmap (from the TSS base)
The task register is expanded to hold 64-bit base addresses in IA-32e mode. See
also: Section 7.7, “Task Management in 64-bit Mode.”
2.1.4
Interrupt and Exception Handling
External interrupts, software interrupts and exceptions are handled through the
interrupt descriptor table (IDT). The IDT stores a collection of gate descriptors that
provide access to interrupt and exception handlers. Like the GDT, the IDT is not a
segment. The linear address for the base of the IDT is contained in the IDT register
(IDTR).
Gate descriptors in the IDT can be interrupt, trap, or task gate descriptors. To access
an interrupt or exception handler, the processor first receives an interrupt vector
(interrupt number) from internal hardware, an external interrupt controller, or from
software by means of an INT, INTO, INT 3, or BOUND instruction. The interrupt
vector provides an index into the IDT. If the selected gate descriptor is an interrupt
gate or a trap gate, the associated handler procedure is accessed in a manner similar
to calling a procedure through a call gate. If the descriptor is a task gate, the handler
is accessed through a task switch.
2.1.4.1
Interrupt and Exception Handling IA-32e Mode
In IA-32e mode, interrupt descriptors are expanded to 16 bytes to support 64-bit
base addresses. This is true for 64-bit mode and compatibility mode.
Vol. 3 2-7
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
The IDTR register is expanded to hold a 64-bit base address. Task gates are not
supported.
2.1.5
Memory Management
System architecture supports either direct physical addressing of memory or virtual
memory (through paging). When physical addressing is used, a linear address is
treated as a physical address. When paging is used: all code, data, stack, and system
segments (including the GDT and IDT) can be paged with only the most recently
accessed pages being held in physical memory.
The location of pages (sometimes called page frames) in physical memory is
contained in the paging structures. These structures reside in physical memory (see
Figure 2-1 for the case of 32-bit paging).
The base physical address of the paging-structure hierarchy is contained in control
register CR3. The entries in the paging structures determine the physical address of
the base of a page frame, access rights and memory management information.
To use this paging mechanism, a linear address is broken into parts. The parts
provide separate offsets into the paging structures and the page frame. A system can
have a single hierarchy of paging structures or several. For example, each task can
have its own hierarchy.
2.1.5.1
Memory Management in IA-32e Mode
In IA-32e mode, physical memory pages are managed by a set of system data struc-
tures. In compatibility mode and 64-bit mode, four levels of system data structures
are used. These include:
• The page map level 4 (PML4) — An entry in a PML4 table contains the physical
address of the base of a page directory pointer table, access rights, and memory
management information. The base physical address of the PML4 is stored in
CR3.
• A set of page directory pointer tables — An entry in a page directory pointer
table contains the physical address of the base of a page directory table, access
rights, and memory management information.
• Sets of page directories — An entry in a page directory table contains the
physical address of the base of a page table, access rights, and memory
management information.
• Sets of page tables — An entry in a page table contains the physical address of
a page frame, access rights, and memory management information.
2-8 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
2.1.6
To assist in initializing the processor and controlling system operations, the system
architecture provides system flags in the EFLAGS register and several system
registers:
Section 2.3, “System Flags and Fields in the EFLAGS Register.”
data fields for controlling system-level operations. Other flags in these registers
are used to indicate support for specific processor capabilities within the
• The debug registers (not shown in Figure 2-1) allow the setting of breakpoints for
use in debugging programs and systems software. See also: Chapter 16,
“Debugging, Profiling Branches and Time-Stamp Counter.”
• The GDTR, LDTR, and IDTR registers contain the linear addresses and sizes
(limits) of their respective tables. See also: Section 2.4, “Memory-Management
Registers.”
• The task register contains the linear address and size of the TSS for the current
task. See also: Section 2.4, “Memory-Management Registers.”
• Model-specific registers (not shown in Figure 2-1).
The model-specific registers (MSRs) are a group of registers available primarily to
These registers control items such as the debug extensions, the performance-moni-
toring counters, the machine- check architecture, and the memory type ranges
(MTRRs).
The number and function of these registers varies among different members of the
Intel 64 and IA-32 processor families. See also: Section 9.4, “Model-Specific Regis-
ters (MSRs),” and Appendix B, “Model-Specific Registers (MSRs).”
Most systems restrict access to system registers (other than the EFLAGS register) by
application programs. Systems can be designed, however, where all programs and
procedures run at the most privileged level (privilege level 0). In such a case, appli-
cation programs would be allowed to modify the system registers.
2.1.6.1
System Registers in IA-32e Mode
In IA-32e mode, the four system-descriptor-table registers (GDTR, IDTR, LDTR, and
TR) are expanded in hardware to hold 64-bit base addresses. EFLAGS becomes the
64-bit RFLAGS register. CR0-CR4 are expanded to 64 bits. CR8 becomes available.
CR8 provides read-write access to the task priority register (TPR) so that the oper-
ating system can control the priority classes of external interrupts.
In 64-bit mode, debug registers DR0–DR7 are 64 bits. In compatibility mode,
address-matching in DR0-DR3 is also done at 64-bit granularity.
Vol. 3 2-9
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
On systems that support IA-32e mode, the extended feature enable register
(IA32_EFER) is available. This model-specific register controls activation of IA-32e
mode and other IA-32e mode operations. In addition, there are several model-
specific registers that govern IA-32e mode instructions:
• IA32_KernelGSbase — Used by SWAPGS instruction.
• IA32_LSTAR — Used by SYSCALL instruction.
• IA32_SYSCALL_FLAG_MASK — Used by SYSCALL instruction.
• IA32_STAR_CS — Used by SYSCALL and SYSRET instruction.
2.1.7
Other System Resources
Besides the system registers and data structures described in the previous sections,
system architecture provides the following additional resources:
• Operating system instructions (see also: Section 2.7, “System Instruction
Summary”).
• Performance-monitoring counters (not shown in Figure 2-1).
• Internal caches and buffers (not shown in Figure 2-1).
Performance-monitoring counters are event counters that can be programmed to
interrupts received, or the number of cache loads. See also: Section 20, “Introduc-
tion to Virtual-Machine Extensions.”
The processor provides several internal caches and buffers. The caches are used to
store both data and instructions. The buffers are used to store things like decoded
addresses to system and application segments and write operations waiting to be
performed. See also: Chapter 11, “Memory Cache Control.”
2.2
MODES OF OPERATION
The IA-32 supports three operating modes and one quasi-operating mode:
• Protected mode — This is the native operating mode of the processor. It
provides a rich set of architectural features, flexibility, high performance and
backward compatibility to existing software base.
• Real-address mode — This operating mode provides the programming
environment of the Intel 8086 processor, with a few extensions (such as the
ability to switch to protected or system management mode).
• System management mode (SMM) — SMM is a standard architectural feature
in all IA-32 processors, beginning with the Intel386 SL processor. This mode
provides an operating system or executive with a transparent mechanism for
implementing power management and OEM differentiation features. SMM is
entered through activation of an external system interrupt pin (SMI#), which
generates a system management interrupt (SMI). In SMM, the processor
switches to a separate address space while saving the context of the currently
2-10 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
running program or task. SMM-specific code may then be executed transparently.
Upon returning from SMM, the processor is placed back into its state prior to the
SMI.
• Virtual-8086 mode — In protected mode, the processor supports a quasi-
operating mode known as virtual-8086 mode. This mode allows the processor
execute 8086 software in a protected, multitasking environment.
Intel 64 architecture supports all operating modes of IA-32 architecture and IA-32e
modes:
• IA-32e mode — In IA-32e mode, the processor supports two sub-modes:
compatibility mode and 64-bit mode. 64-bit mode provides 64-bit linear
addressing and support for physical address space larger than 64 GBytes.
Compatibility mode allows most legacy protected-mode applications to run
unchanged.
Figure 2-3 shows how the processor moves between operating modes.
SMI#
Real-Address
Mode
Reset
or
RSM
Reset or
PE=0
PE=1
SMI#
Reset
Protected Mode
System
RSM
Management
LME=1, CR0.PG=1*
Mode
SMI#
RSM
See**
IA-32e
Mode
VM=0
VM=1
* See Section 9.8.5
** See Section 9.8.5.4
SMI#
RSM
Virtual-8086
Mode
Figure 2-3. Transitions Among the Processor’s Operating Modes
The processor is placed in real-address mode following power-up or a reset. The PE
flag in control register CR0 then controls whether the processor is operating in real-
address or protected mode. See also: Section 9.9, “Mode Switching.” and Section
4.1.2, “Paging-Mode Enabling.”
Vol. 3 2-11
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
The VM flag in the EFLAGS register determines whether the processor is operating in
protected mode or virtual-8086 mode. Transitions between protected mode and
virtual-8086 mode are generally carried out as part of a task switch or a return from
Mode.”
The LMA bit (IA32_EFER.LMA.LMA[bit 10]) determines whether the processor is
operating in IA-32e mode. When running in IA-32e mode, 64-bit or compatibility
sub-mode operation is determined by CS.L bit of the code segment. The processor
enters into IA-32e mode from protected mode by enabling paging and setting the
LME bit (IA32_EFER.LME[bit 8]). See also: Chapter 9, “Processor Management and
Initialization.”
The processor switches to SMM whenever it receives an SMI while the processor is in
real-address, protected, virtual-8086, or IA-32e modes. Upon execution of the RSM
instruction, the processor always returns to the mode it was in when the SMI
occurred.
2.3
SYSTEM FLAGS AND FIELDS IN THE EFLAGS
REGISTER
The system flags and IOPL field of the EFLAGS register control I/O, maskable hard-
ware interrupts, debugging, task switching, and the virtual-8086 mode (see
Figure 2-4). Only privileged code (typically operating system or executive code)
should be allowed to modify these bits.
The system flags and IOPL are:
TF
Trap (bit 8) — Set to enable single-step mode for debugging; clear to
disable single-step mode. In single-step mode, the processor generates a
debug exception after each instruction. This allows the execution state of a
program to be inspected after each instruction. If an application program
sets the TF flag using a POPF, POPFD, or IRET instruction, a debug exception
is generated after the instruction that follows the POPF, POPFD, or IRET.
2-12 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
31
21 20 19
17 16
18
22
15 14 13 12 11 10
I
9
7
5
0
4
3
0
1
1
0
8
6
2
V
I
V
I
O
P
L
O
F
A
C
R
F
N
T
D
F
I
F
T
F
S
F
Z
F
A
F
P
F
C
F
I
D
V
M
0
Reserved (set to 0)
P
F
ID — Identification Flag
VIP — Virtual Interrupt Pending
VIF — Virtual Interrupt Flag
AC — Alignment Check
VM — Virtual-8086 Mode
RF — Resume Flag
NT — Nested Task Flag
IOPL— I/O Privilege Level
IF — Interrupt Enable Flag
TF — Trap Flag
Figure 2-4. System Flags in the EFLAGS Register
IF
Interrupt enable (bit 9) — Controls the response of the processor to
maskable hardware interrupt requests (see also: Section 6.3.2, “Maskable
Hardware Interrupts”). The flag is set to respond to maskable hardware
interrupts; cleared to inhibit maskable hardware interrupts. The IF flag does
not affect the generation of exceptions or nonmaskable interrupts (NMI
interrupts). The CPL, IOPL, and the state of the VME flag in control register
CR4 determine whether the IF flag can be modified by the CLI, STI, POPF,
POPFD, and IRET.
IOPL
I/O privilege level field (bits 12 and 13) — Indicates the I/O privilege
level (IOPL) of the currently running program or task. The CPL of the
currently running program or task must be less than or equal to the IOPL to
access the I/O address space. This field can only be modified by the POPF
and IRET instructions when operating at a CPL of 0.
The IOPL is also one of the mechanisms that controls the modification of the
IF flag and the handling of interrupts in virtual-8086 mode when virtual
mode extensions are in effect (when CR4.VME = 1). See also: Chapter 13,
“Input/Output,” in the Intel® 64 and IA-32 Architectures Software Devel-
oper’s Manual, Volume 1.
NT
Nested task (bit 14) — Controls the chaining of interrupted and called
tasks. The processor sets this flag on calls to a task initiated with a CALL
instruction, an interrupt, or an exception. It examines and modifies this flag
on returns from a task initiated with the IRET instruction. The flag can be
explicitly set or cleared with the POPF/POPFD instructions; however,
Vol. 3 2-13
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
changing to the state of this flag can generate unexpected exceptions in
application programs.
See also: Section 7.4, “Task Linking.”
RF
Resume (bit 16) — Controls the processor’s response to instruction-break-
point conditions. When set, this flag temporarily disables debug exceptions
(#DB) from being generated for instruction breakpoints (although other
exception conditions can cause an exception to be generated). When clear,
instruction breakpoints will generate debug exceptions.
The primary function of the RF flag is to allow the restarting of an instruction
following a debug exception that was caused by an instruction breakpoint
condition. Here, debug software must set this flag in the EFLAGS image on
the stack just prior to returning to the interrupted program with IRETD (to
prevent the instruction breakpoint from causing another debug exception).
The processor then automatically clears this flag after the instruction
returned to has been successfully executed, enabling instruction breakpoint
faults again.
See also: Section 16.3.1.1, “Instruction-Breakpoint Exception Condition.”
VM
AC
Virtual-8086 mode (bit 17) — Set to enable virtual-8086 mode; clear to
return to protected mode.
See also: Section 17.2.1, “Enabling Virtual-8086 Mode.”
Alignment check (bit 18) — Set this flag and the AM flag in control register
CR0 to enable alignment checking of memory references; clear the AC flag
and/or the AM flag to disable alignment checking. An alignment-check
exception is generated when reference is made to an unaligned operand,
such as a word at an odd byte address or a doubleword at an address which
is not an integral multiple of four. Alignment-check exceptions are generated
only in user mode (privilege level 3). Memory references that default to priv-
ilege level 0, such as segment descriptor loads, do not generate this excep-
tion even when caused by instructions executed in user-mode.
The alignment-check exception can be used to check alignment of data. This
is useful when exchanging data with processors which require all data to be
aligned. The alignment-check exception can also be used by interpreters to
flag some pointers as special by misaligning the pointer. This eliminates
overhead of checking each pointer and only handles the special pointer when
used.
VIF
flag is used in conjunction with the VIP flag. The processor only recognizes
the VIF flag when either the VME flag or the PVI flag in control register CR4 is
set and the IOPL is less than 3. (The VME flag enables the virtual-8086 mode
extensions; the PVI flag enables the protected-mode virtual interrupts.)
See also: Section 17.3.3.5, “Method 6: Software Interrupt Handling,” and
Section 17.4, “Protected-Mode Virtual Interrupts.”
2-14 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
VIP
Virtual interrupt pending (bit 20) — Set by software to indicate that an
never modifies it. The processor only recognizes the VIP flag when either the
VME flag or the PVI flag in control register CR4 is set and the IOPL is less than
3. The VME flag enables the virtual-8086 mode extensions; the PVI flag
enables the protected-mode virtual interrupts.
See Section 17.3.3.5, “Method 6: Software Interrupt Handling,” and Section
17.4, “Protected-Mode Virtual Interrupts.”
ID
Identification (bit 21). — The ability of a program or procedure to set or
2.3.1
System Flags and Fields in IA-32e Mode
In 64-bit mode, the RFLAGS register expands to 64 bits with the upper 32 bits
reserved. System flags in RFLAGS (64-bit mode) or EFLAGS (compatibility mode)
are shown in Figure 2-4.
In IA-32e mode, the processor does not allow the VM bit to be set because virtual-
8086 mode is not supported (attempts to set the bit are ignored). Also, the processor
will not set the NT bit. The processor does, however, allow software to set the NT bit
(note that an IRET causes a general protection fault in IA-32e mode if the NT bit is
set).
In IA-32e mode, the SYSCALL/SYSRET instructions have a programmable method of
specifying which bits are cleared in RFLAGS/EFLAGS. These instructions save/restore
EFLAGS/RFLAGS.
2.4
MEMORY-MANAGEMENT REGISTERS
The processor provides four memory-management registers (GDTR, LDTR, IDTR,
and TR) that specify the locations of the data structures which control segmented
memory management (see Figure 2-5). Special instructions are provided for loading
and storing these registers.
Vol. 3 2-15
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
System Table Registers
16 15
47(79)
32(64)-bit Linear Base Address
32(64)-bit Linear Base Address
0
16-Bit Table Limit
16-Bit Table Limit
GDTR
IDTR
System Segment
Registers
Segment Descriptor Registers (Automatically Loaded)
15
Attributes
0
Task
Register
LDTR
Seg. Sel.
Seg. Sel.
32(64)-bit Linear Base Address
32(64)-bit Linear Base Address
Segment Limit
Segment Limit
Figure 2-5. Memory Management Registers
2.4.1
Global Descriptor Table Register (GDTR)
The GDTR register holds the base address (32 bits in protected mode; 64 bits in
IA-32e mode) and the 16-bit table limit for the GDT. The base address specifies the
the table.
The LGDT and SGDT instructions load and store the GDTR register, respectively. On
power up or reset of the processor, the base address is set to the default value of 0
and the limit is set to 0FFFFH. A new base address must be loaded into the GDTR as
part of the processor initialization process for protected-mode operation.
See also: Section 3.5.1, “Segment Descriptor Tables.”
2.4.2
Local Descriptor Table Register (LDTR)
The LDTR register holds the 16-bit segment selector, base address (32 bits in
protected mode; 64 bits in IA-32e mode), segment limit, and descriptor attributes
for the LDT. The base address specifies the linear address of byte 0 of the LDT
segment; the segment limit specifies the number of bytes in the segment. See also:
Section 3.5.1, “Segment Descriptor Tables.”
The LLDT and SLDT instructions load and store the segment selector part of the LDTR
register, respectively. The segment that contains the LDT must have a segment
descriptor in the GDT. When the LLDT instruction loads a segment selector in the
LDTR: the base address, limit, and descriptor attributes from the LDT descriptor are
automatically loaded in the LDTR.
When a task switch occurs, the LDTR is automatically loaded with the segment
selector and descriptor for the LDT for the new task. The contents of the LDTR are not
automatically saved prior to writing the new LDT information into the register.
On power up or reset of the processor, the segment selector and base address are set
to the default value of 0 and the limit is set to 0FFFFH.
2-16 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
2.4.3
IDTR Interrupt Descriptor Table Register
The IDTR register holds the base address (32 bits in protected mode; 64 bits in
address of byte 0 of the IDT; the table limit specifies the number of bytes in the table.
The LIDT and SIDT instructions load and store the IDTR register, respectively. On
power up or reset of the processor, the base address is set to the default value of 0
and the limit is set to 0FFFFH. The base address and limit in the register can then be
changed as part of the processor initialization process.
See also: Section 6.10, “Interrupt Descriptor Table (IDT).”
2.4.4
Task Register (TR)
The task register holds the 16-bit segment selector, base address (32 bits in
protected mode; 64 bits in IA-32e mode), segment limit, and descriptor attributes
for the TSS of the current task. The selector references the TSS descriptor in the GDT.
The base address specifies the linear address of byte 0 of the TSS; the segment limit
specifies the number of bytes in the TSS. See also: Section 7.2.4, “Task Register.”
The LTR and STR instructions load and store the segment selector part of the task
register, respectively. When the LTR instruction loads a segment selector in the task
register, the base address, limit, and descriptor attributes from the TSS descriptor
are automatically loaded into the task register. On power up or reset of the processor,
the base address is set to the default value of 0 and the limit is set to 0FFFFH.
When a task switch occurs, the task register is automatically loaded with the
segment selector and descriptor for the TSS for the new task. The contents of the
task register are not automatically saved prior to writing the new TSS information
into the register.
2.5
CONTROL REGISTERS
Control registers (CR0, CR1, CR2, CR3, and CR4; see Figure 2-6) determine oper-
ating mode of the processor and the characteristics of the currently executing task.
These registers are 32 bits in all 32-bit modes and compatibility mode.
In 64-bit mode, control registers are expanded to 64 bits. The MOV CRn instructions
are used to manipulate the register bits. Operand-size prefixes for these instructions
are ignored. The following is also true:
• Bits 63:32 of CR0 and CR4 are reserved and must be written with zeros. Writing
a nonzero value to any of the upper 32 bits results in a general-protection
exception, #GP(0).
• All 64 bits of CR2 are writable by software.
• Bits 51:40 of CR3 are reserved and must be 0.
Vol. 3 2-17
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
• The MOV CRn instructions do not check that addresses written to CR2 and CR3
are within the linear-address or physical-address limitations of the implemen-
tation.
• Register CR8 is available in 64-bit mode only.
The control registers are summarized below, and each architecturally defined control
field in these control registers are described individually. In Figure 2-6, the width of
the register in 64-bit mode is indicated in parenthesis (except for CR0).
• CR0 — Contains system control flags that control operating mode and states of
the processor.
• CR1 — Reserved.
• CR2 — Contains the page-fault linear address (the linear address that caused a
page fault).
• CR3 — Contains the physical address of the base of the paging-structure
hierarchy and two flags (PCD and PWT). Only the most-significant bits (less the
lower 12 bits) of the base address are specified; the lower 12 bits of the address
are assumed to be 0. The first paging structure must thus be aligned to a page
structure in the processor’s internal data caches (they do not control TLB caching
of page-directory information).
When using the physical address extension, the CR3 register contains the base
address of the page-directory-pointer table In IA-32e mode, the CR3 register
contains the base address of the PML4 table.
See also: Chapter 4, “Paging.”
• CR4 — Contains a group of flags that enable several architectural extensions,
and indicate operating system or executive support for specific processor capabil-
ities. The control registers can be read and loaded (or modified) using the move-
to-or-from-control-registers forms of the MOV instruction. In protected mode,
the MOV instructions allow the control registers to be read or loaded (at privilege
level 0 only). This restriction means that application programs or operating-
system procedures (running at privilege levels 1, 2, or 3) are prevented from
reading or loading the control registers.
• CR8 — Provides read and write access to the Task Priority Register (TPR). It
specifies the priority threshold value that operating systems use to control the
priority class of external interrupts allowed to interrupt the processor. This
register is available only in 64-bit mode. However, interrupt filtering continues to
apply in compatibility mode.
2-18 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
31(63)
18
13 12 11 10
V
9
8
7
6
5
4
3
2
1
0
14
S
M
X
E
T
S
D
P
V
I
V
M
E
P
C
E
P
A
E
P
S
E
P
G
E
M
C
E
D
E
M
X
E
0
0
Reserved (set to 0)
CR4
OSXSAVE
OSXMMEXCPT
OSFXSR
31(63)
31(63)
12 11
5
4
3
2
P
C W
D
P
CR3
(PDBR)
Page-Directory Base
T
0
0
Page-Fault Linear Address
CR2
CR1
31(63)
31 30 29 28
19
17 16
18
15
5
4
3
1
0
6
2
P
G
C
D
N
W
A
M
W
P
N
E
E
T
T
S
E
M
M
P
P
E
CR0
Reserved
Figure 2-6. Control Registers
previously read. The flags in control registers are:
PG
Paging (bit 31 of CR0) — Enables paging when set; disables paging when
clear. When paging is disabled, all linear addresses are treated as physical
addresses. The PG flag has no effect if the PE flag (bit 0 of register CR0) is
not also set; setting the PG flag when the PE flag is clear causes a general-
On Intel 64 processors, enabling and disabling IA-32e mode operation also
requires modifying CR0.PG.
CD
Cache Disable (bit 30 of CR0) — When the CD and NW flags are clear,
caching of memory locations for the whole of physical memory in the
processor’s internal (and external) caches is enabled. When the CD flag is
set, caching is restricted as described in Table 11-5. To prevent the processor
from accessing and updating its caches, the CD flag must be set and the
caches must be invalidated so that no cache hits can occur.
Vol. 3 2-19
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
See also: Section 11.5.3, “Preventing Caching,” and Section 11.5, “Cache
Control.”
NW
Not Write-through (bit 29 of CR0) — When the NW and CD flags are
clear, write-back (for Pentium 4, Intel Xeon, P6 family, and Pentium proces-
sors) or write-through (for Intel486 processors) is enabled for writes that hit
the cache and invalidation cycles are enabled. See Table 11-5 for detailed
information about the affect of the NW flag on caching for other settings of
the CD and NW flags.
AM
WP
Alignment Mask (bit 18 of CR0) — Enables automatic alignment checking
when set; disables alignment checking when clear. Alignment checking is
set, CPL is 3, and the processor is operating in either protected or virtual-
8086 mode.
Write Protect (bit 16 of CR0) — When set, inhibits supervisor-level proce-
dures from writing into read-only pages; when clear, allows supervisor-level
procedures to write into read-only pages (regardless of the U/S bit setting;
see Section 4.1.3 and Section 4.6). This flag facilitates implementation of the
copy-on-write method of creating a new process (forking) used by operating
systems such as UNIX.
NE
Numeric Error (bit 5 of CR0) — Enables the native (internal) mechanism
for reporting x87 FPU errors when set; enables the PC-style x87 FPU error
reporting mechanism when clear. When the NE flag is clear and the IGNNE#
input is asserted, x87 FPU errors are ignored. When the NE flag is clear and
the IGNNE# input is deasserted, an unmasked x87 FPU error causes the
processor to assert the FERR# pin to generate an external interrupt and to
stop instruction execution immediately before executing the next waiting
floating-point instruction or WAIT/FWAIT instruction.
The FERR# pin is intended to drive an input to an external interrupt
controller (the FERR# pin emulates the ERROR# pin of the Intel 287 and
Intel 387 DX math coprocessors). The NE flag, IGNNE# pin, and FERR# pin
are used with external logic to implement PC-style error reporting. Using
FERR# and IGNNE# to handle floating-point exceptions is deprecated by
modern operating systems; this non-native approach also limits newer
processors to operate with one logical processor active.
See also: “Software Exception Handling” in Chapter 8, “Programming with
the x87 FPU,” and Appendix A, “EFLAGS Cross-Reference,” in the Intel® 64
and IA-32 Architectures Software Developer’s Manual, Volume 1.
ET
TS
Extension Type (bit 4 of CR0) — Reserved in the Pentium 4, Intel Xeon, P6
family, and Pentium processors. In the Pentium 4, Intel Xeon, and P6 family
processors, this flag is hardcoded to 1. In the Intel386 and Intel486 proces-
sors, this flag indicates support of Intel 387 DX math coprocessor instruc-
tions when set.
Task Switched (bit 3 of CR0) — Allows the saving of the x87
FPU/MMX/SSE/SSE2/SSE3/SSSE3/SSE4 context on a task switch to be
2-20 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
delayed until an x87 FPU/MMX/SSE/SSE2/SSE3/SSSE3/SSE4 instruction is
actually executed by the new task. The processor sets this flag on every task
switch and tests it when executing x87
FPU/MMX/SSE/SSE2/SSE3/SSSE3/SSE4 instructions.
•
If the TS flag is set and the EM flag (bit 2 of CR0) is clear, a device-not-
available exception (#NM) is raised prior to the execution of any x87
FPU/MMX/SSE/ SSE2/SSE3/SSSE3/SSE4 instruction; with the exception
of PAUSE, PREFETCHh, SFENCE, LFENCE, MFENCE, MOVNTI, CLFLUSH,
CRC32, and POPCNT. See the paragraph below for the special case of the
WAIT/FWAIT instructions.
•
•
If the TS flag is set and the MP flag (bit 1 of CR0) and EM flag are clear, an
WAIT/FWAIT instruction.
If the EM flag is set, the setting of the TS flag has no affect on the
execution of x87 FPU/MMX/SSE/SSE2/SSE3/SSSE3/SSE4 instructions.
Table 2-1 shows the actions taken when the processor encounters an x87
FPU instruction based on the settings of the TS, EM, and MP flags. Table 12-1
and 13-1 show the actions taken when the processor encounters an
MMX/SSE/SSE2/SSE3/SSSE3/SSE4 instruction.
The processor does not automatically save the context of the x87 FPU, XMM,
and MXCSR registers on a task switch. Instead, it sets the TS flag, which
causes the processor to raise an #NM exception whenever it encounters an
x87 FPU/MMX/SSE /SSE2/SSE3/SSSE3/SSE4 instruction in the instruction
stream for the new task (with the exception of the instructions listed above).
The fault handler for the #NM exception can then be used to clear the TS flag (with
the CLTS instruction) and save the context of the x87 FPU, XMM, and MXCSR regis-
ters. If the task never encounters an x87 FPU/MMX/SSE/SSE2/SSE3//SSSE3/SSE4
instruction; the x87 FPU/MMX/SSE/SSE2/ SSE3/SSSE3/SSE4 context is never saved.
Table 2-1. Action Taken By x87 FPU Instructions for Different
Combinations of EM, MP, and TS
CR0 Flags
x87 FPU Instruction Type
Floating-Point WAIT/FWAIT
EM
0
MP
0
TS
0
Execute
Execute.
0
0
1
#NM Exception
Execute
Execute.
0
1
0
Execute.
0
1
1
#NM Exception
#NM Exception
#NM Exception
#NM Exception
#NM exception.
Execute.
1
0
0
1
0
1
Execute.
1
1
0
Execute.
Vol. 3 2-21
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
Table 2-1. Action Taken By x87 FPU Instructions for Different
Combinations of EM, MP, and TS
x87 FPU Instruction Type
#NM exception.
CR0 Flags
1
1
1
#NM Exception
EM
Emulation (bit 2 of CR0) — Indicates that the processor does not have an
internal or external x87 FPU when set; indicates an x87 FPU is present when
clear. This flag also affects the execution of
MMX/SSE/SSE2/SSE3/SSSE3/SSE4 instructions.
device-not-available exception (#NM). This flag must be set when the
processor does not have an internal x87 FPU or is not connected to an
tions to be handled by software emulation. Table 9-2 shows the recom-
mended setting of this flag, depending on the IA-32 processor and x87 FPU
or math coprocessor present in the system. Table 2-1 shows the interaction
of the EM, MP, and TS flags.
Also, when the EM flag is set, execution of an MMX instruction causes an
invalid-opcode exception (#UD) to be generated (see Table 12-1). Thus, if an
IA-32 or Intel 64 processor incorporates MMX technology, the EM flag must
be set to 0 to enable execution of MMX instructions.
Similarly for SSE/SSE2/SSE3/SSSE3/SSE4 extensions, when the EM flag is
set, execution of most SSE/SSE2/SSE3/SSSE3/SSE4 instructions causes an
invalid opcode exception (#UD) to be generated (see Table 13-1). If an IA-32
or Intel 64 processor incorporates the SSE/SSE2/SSE3/SSSE3/SSE4 exten-
sions, the EM flag must be set to 0 to enable execution of these extensions.
SSE/SSE2/SSE3/SSSE3/SSE4 instructions not affected by the EM flag
CRC32, and POPCNT.
MP
PE
Monitor Coprocessor (bit 1 of CR0). — Controls the interaction of the
WAIT (or FWAIT) instruction with the TS flag (bit 3 of CR0). If the MP flag is
set, a WAIT instruction generates a device-not-available exception (#NM) if
the TS flag is also set. If the MP flag is clear, the WAIT instruction ignores the
setting of the TS flag. Table 9-2 shows the recommended setting of this flag,
in the system. Table 2-1 shows the interaction of the MP, EM, and TS flags.
Protection Enable (bit 0 of CR0) — Enables protected mode when set;
enables real-address mode when clear. This flag does not enable paging
directly. It only enables segment-level protection. To enable paging, both the
PE and PG flags must be set.
See also: Section 9.9, “Mode Switching.”
PCD
Page-level Cache Disable (bit 4 of CR3) — Controls caching of the first
paging structure of the current paging-structure hierarchy. When the PCD
2-22 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
the page-directory can be cached. This flag affects only the processor’s
internal caches (both L1 and L2, when present). The processor ignores this
flag if paging is not used (the PG flag in register CR0 is clear) or the CD
(cache disable) flag in CR0 is set.
See also: Chapter 11, “Memory Cache Control” (for more about the use of
the PCD flag) and Section 4.9, “Paging and Memory Typing” (for a discussion
of a companion PCD flag in page-directory and page-table entries).
PWT
Page-level Write-Through (bit 3 of CR3) — Controls the write-through or
structure hierarchy. When the PWT flag is set, write-through caching is
enabled; when the flag is clear, write-back caching is enabled. This flag
affects only internal caches (both L1 and L2, when present). The processor
ignores this flag if paging is not used (the PG flag in register CR0 is clear) or
the CD (cache disable) flag in CR0 is set.
See also: Section 11.5, “Cache Control” (for more information about the use
of this flag), and Section 4.9, “Paging and Memory Typing” (for a discussion
of a companion PCD flag in the page-directory and page-table entries).
VME
Virtual-8086 Mode Extensions (bit 0 of CR4) — Enables interrupt- and
exception-handling extensions in virtual-8086 mode when set; disables the
extensions when clear. Use of the virtual mode extensions can improve the
performance of virtual-8086 applications by eliminating the overhead of
calling the virtual-8086 monitor to handle interrupts and exceptions that
occur while executing an 8086 program and, instead, redirecting the inter-
rupts and exceptions back to the 8086 program’s handlers. It also provides
hardware support for a virtual interrupt flag (VIF) to improve reliability of
running 8086 programs in multitasking and multiple-processor environ-
ments.
See also: Section 17.3, “Interrupt and Exception Handling in Virtual-8086
Mode.”
PVI
Protected-Mode Virtual Interrupts (bit 1 of CR4) — Enables hardware
support for a virtual interrupt flag (VIF) in protected mode when set; disables
the VIF flag in protected mode when clear.
See also: Section 17.4, “Protected-Mode Virtual Interrupts.”
TSD
Time Stamp Disable (bit 2 of CR4) — Restricts the execution of the
RDTSC instruction (including RDTSCP instruction if
CPUID.80000001H:EDX[27] = 1) to procedures running at privilege level 0
when set; allows RDTSC instruction (including RDTSCP instruction if
CPUID.80000001H:EDX[27] = 1) to be executed at any privilege level when
clear.
DE
Debugging Extensions (bit 3 of CR4) — References to debug registers
DR4 and DR5 cause an undefined opcode (#UD) exception to be generated
Vol. 3 2-23
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
for compatibility with software written to run on earlier IA-32 processors.
See also: Section 16.2.2, “Debug Registers DR4 and DR5.”
PSE
PAE
Page Size Extensions (bit 4 of CR4) — Enables 4-MByte pages with 32-bit
See also: Section 4.3, “32-Bit Paging.”
Physical Address Extension (bit 5 of CR4) — When set, enables paging
physical addresses to 32 bits. PAE must be set before entering IA-32e mode.
See also: Chapter 4, “Paging.”
MCE
PGE
Machine-Check Enable (bit 6 of CR4) — Enables the machine-check
exception when set; disables the machine-check exception when clear.
See also: Chapter 15, “Machine-Check Architecture.”
Page Global Enable (bit 7 of CR4) — (Introduced in the P6 family proces-
sors.) Enables the global page feature when set; disables the global page
feature when clear. The global page feature allows frequently used or shared
pages to be marked as global to all users (done with the global flag, bit 8, in
a page-directory or page-table entry). Global pages are not flushed from the
translation-lookaside buffer (TLB) on a task switch or a write to register CR3.
When enabling the global page feature, paging must be enabled (by setting
the PG flag in control register CR0) before the PGE flag is set. Reversing this
sequence may affect program correctness, and processor performance will
be impacted.
See also: Section 4.10, “Caching Translation Information.”
PCE
Performance-Monitoring Counter Enable (bit 8 of CR4) — Enables
execution of the RDPMC instruction for programs or procedures running at
any protection level when set; RDPMC instruction can be executed only at
protection level 0 when clear.
OSFXSR
Operating System Support for FXSAVE and FXRSTOR instructions
(bit 9 of CR4) — When set, this flag: (1) indicates to software that the oper-
ating system supports the use of the FXSAVE and FXRSTOR instructions, (2)
enables the FXSAVE and FXRSTOR instructions to save and restore the
contents of the XMM and MXCSR registers along with the contents of the x87
FPU and MMX registers, and (3) enables the processor to execute
SSE/SSE2/SSE3/SSSE3/SSE4 instructions, with the exception of the PAUSE,
PREFETCHh, SFENCE, LFENCE, MFENCE, MOVNTI, CLFLUSH, CRC32, and
POPCNT.
If this flag is clear, the FXSAVE and FXRSTOR instructions will save and
restore the contents of the x87 FPU and MMX instructions, but they may not
save and restore the contents of the XMM and MXCSR registers. Also, the
2-24 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
processor will generate an invalid opcode exception (#UD) if it attempts to
execute any SSE/SSE2/SSE3and instruction, with the exception of PAUSE,
PREFETCHh, SFENCE, LFENCE, MFENCE, MOVNTI, CLFLUSH, CRC32, and
POPCNT. The operating system or executive must explicitly set this flag.
NOTE
CPUID feature flags FXSR indicates availability of the
FXSAVE/FXRSTOR instructions. The OSFXSR bit provides operating
system software with a means of enabling FXSAVE/FXRSTOR to
save/restore the contents of the X87 FPU, XMM and MXCSR registers.
Consequently OSFXSR bit indicates that the operating system
provides context switch support for SSE/SSE2/SSE3/SSSE3/SSE4.
OSXMMEXCPT
Operating System Support for Unmasked SIMD Floating-Point Excep-
tions (bit 10 of CR4) — When set, indicates that the operating system
supports the handling of unmasked SIMD floating-point exceptions through
an exception handler that is invoked when a SIMD floating-point exception
(#XF) is generated. SIMD floating-point exceptions are only generated by
SSE/SSE2/SSE3/SSE4.1 SIMD floating-point instructions.
The operating system or executive must explicitly set this flag. If this flag is
not set, the processor will generate an invalid opcode exception (#UD)
whenever it detects an unmasked SIMD floating-point exception.
VMXE
SMXE
VMX-Enable Bit (bit 13 of CR4) — Enables VMX operation when set. See
Chapter 20, “Introduction to Virtual-Machine Extensions.”
SMX-Enable Bit (bit 14 of CR4) — Enables SMX operation when set. See
Chapter 6, “Safer Mode Extensions Reference” of Intel® 64 and IA-32 Archi-
tectures Software Developer’s Manual, Volume 2B.
OSXSAVE
XSAVE and Processor Extended States-Enable Bit (bit 18 of CR4) —
When set, this flag: (1) indicates (via CPUID.01H:ECX.OSXSAVE[bit 27])
XRSTOR instructions by general software; (2) enables the XSAVE and
XRSTOR instructions to save and restore the x87 FPU state (including MMX
registers), the SSE state (XMM registers and MXCSR), along with other
processor extended states enabled in the XFEATURE_ENABLED_MASK
register (XCR0); (3) enables the processor to execute XGETBV and XSETBV
instructions in order to read and write XCR0. See Section 2.6 and Chapter
13, “System Programming for Instruction Set Extensions and Processor
Extended States”.
TPL
Task Priority Level (bit 3:0 of CR8) — This sets the threshold value corre-
sponding to the highest-priority interrupt to be blocked. A value of 0 means
Vol. 3 2-25
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
all interrupts are enabled. This field is available in 64-bit mode. A value of 15
means all interrupts will be disabled.
2.5.1
CPUID Qualification of Control Register Flags
The VME, PVI, TSD, DE, PSE, PAE, MCE, PGE, PCE, OSFXSR, and OSXMMEXCPT flags
in control register CR4 are model specific. All of these flags (except the PCE flag) can
be qualified with the CPUID instruction to determine if they are implemented on the
processor before they are used.
The CR8 register is available on processors that support Intel 64 architecture.
2.6
EXTENDED CONTROL REGISTERS (INCLUDING THE
XFEATURE_ENABLED_MASK REGISTER)
If CPUID.01H:ECX.XSAVE[bit 26] is 1, the processor supports one or more
extended control registers (XCRs). Currently, the only such register defined is
XCR0, the XFEATURE_ENABLED_MASK register. This register specifies the set of
processor states that the operating system enables on that processor, e.g. x87 FPU
States, SSE states, and other processor extended states that Intel 64 architecture
may introduce in the future. The OS programs XCR0 to reflect the features it
supports.
63
1
0
1
2
Reserved for XCR0 bit vector expansion
Reserved / Future processor extended states
SSE state
x87 FPU/MMX state (must be 1)
Reserved (must be 0)
Figure 2-7. XFEATURE_ENABLED_MASK Register (XCR0)
Software can access XCR0 only if CR4.OSXSAVE[bit 18] = 1. (This bit is also readable
as CPUID.01H:ECX.OSXSAVE[bit 27].) The layout of XCR0 is architected to allow
software to use CPUID leaf function 0DH to enumerate the set of bits that the
processor supports in XCR0 (see CPUID instruction in Intel® 64 and IA-32 Architec-
tures Software Developer’s Manual, Volume 2A). Each processor state (X87 FPU
2-26 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
state, SSE state, or a future processor extended state) is represented by a bit in
XCR0. The OS can enable future processor extended states in a forward manner by
specifying the appropriate bit mask value using the XSETBV instruction according to
the results of the CPUID leaf 0DH.
With the exception of bit 63, each bit in the XFEATURE_ENABLED_MASK register
(XCR0) corresponds to a subset of the processor states. XCR0 thus provides space
for up to 63 sets of processor state extensions. Bit 63 of XCR0 is reserved for future
expansion and will not represent a processor extended state.
Currently, the XFEATURE_ENABLED_MASK register (XCR0) has two processor states
defined, with up to 61 bits reserved for future processor extended states:
• XCR0.X87 (bit 0): If 1, indicates x87 FPU state (including MMX register states) is
supported in the processor. Bit 0 must be 1. An attempt to write 0 causes a #GP
exception.
• XCR0.SSE (bit 1): If 1, indicates MXCSR and XMM registers (XMM0-XMM15 in 64-
bit mode, otherwise XMM0-XMM7) are supported by XSAVE/XRESTOR in the
processor.
Any attempt to set a reserved bit (as determined by the contents of EAX and EDX
after executing CPUID with EAX=0DH, ECX= 0H) in the XFEATURE_ENABLED_MASK
register for a given processor will result in a #GP exception. An attempt to write 0 to
XFEATURE_ENABLED_MASK.x87 (bit 0) will result in a #GP exception.
If a bit in the XFEATURE_ENABLED_MASK register is 1, XSAVE instruction can selec-
tively (in conjunction with a save mask) save a partial or full set of processor states
to memory (See XSAVE instruction in Intel® 64 and IA-32 Architectures Software
Developer’s Manual, Volume 2B).
After reset all bits (except bit 0) in the XFEATURE_ENABLED_MASK register (XCR0)
are cleared to zero. XCR0[0] is set to 1.
2.7
SYSTEM INSTRUCTION SUMMARY
System instructions handle system-level functions such as loading system registers,
managing the cache, managing interrupts, or setting up the debug registers. Many of
these instructions can be executed only by operating-system or executive proce-
dures (that is, procedures running at privilege level 0). Others can be executed at
any privilege level and are thus available to application programs.
Table 2-2 lists the system instructions and indicates whether they are available and
useful for application programs. These instructions are described in the Intel® 64
and IA-32 Architectures Software Developer’s Manual, Volumes 2A & 2B.
Table 2-2. Summary of System Instructions
Useful to
Application?
Protected from
Application?
Instruction
Description
LLDT
Load LDT Register
No
Yes
Vol. 3 2-27
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
Table 2-2. Summary of System Instructions (Contd.)
Useful to
Protected from
Application?
Instruction
SLDT
Description
Application?
Store LDT Register
Load GDT Register
Store GDT Register
Load Task Register
Store Task Register
Load IDT Register
Store IDT Register
Load and store control registers
Store MSW
No
No
No
Yes
No
LGDT
SGDT
No
LTR
No
Yes
No
STR
No
LIDT
No
Yes
No
SIDT
No
MOV CRn
SMSW
LMSW
CLTS
No
Yes
No
Yes
No
Load MSW
Yes
Yes
No
Clear TS flag in CR0
Adjust RPL
No
ARPL
Yes1, 5
Yes
Yes
Yes
Yes
No
LAR
Load Access Rights
Load Segment Limit
Verify for Reading
Verify for Writing
Load and store debug registers
Invalidate cache, no writeback
Invalidate cache, with writeback
Invalidate TLB entry
Halt Processor
No
LSL
No
VERR
No
VERW
MOV DRn
INVD
No
Yes
Yes
Yes
Yes
Yes
No
No
WBINVD
INVLPG
HLT
No
No
No
LOCK (Prefix)
RSM
Bus Lock
Yes
No
Return from system management
mode
Yes
3
RDMSR
Read Model-Specific Registers
Write Model-Specific Registers
No
No
Yes
Yes
3
WRMSR
4
RDPMC
Read Performance-Monitoring
Counter
Yes
Yes2
3
RDTSC
Read Time-Stamp Counter
Yes
Yes
Yes2
Yes2
7
RDTSCP
Read Serialized Time-Stamp Counter
2-28 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
Table 2-2. Summary of System Instructions (Contd.)
Useful to
Application?
Protected from
Application?
Instruction
Description
XGETBV
Return the state of the the
XFEATURE_ENABLED_MASK register
Yes
No
XSETBV
Enable one or more processor
extended states
No6
Yes
NOTES:
1. Useful to application programs running at a CPL of 1 or 2.
2. The TSD and PCE flags in control register CR4 control access to these instructions by application
programs running at a CPL of 3.
3. These instructions were introduced into the IA-32 Architecture with the Pentium processor.
4. This instruction was introduced into the IA-32 Architecture with the Pentium Pro processor and
the Pentium processor with MMX technology.
5. This instruction is not supported in 64-bit mode.
6. Application uses XGETBV to query which set of processor extended states are enabled.
7. RDTSCP is introduced in Intel Core i7 processor.
2.7.1
Loading and Storing System Registers
The GDTR, LDTR, IDTR, and TR registers each have a load and store instruction for
loading data into and storing data from the register:
• LGDT (Load GDTR Register) — Loads the GDT base address and limit from
memory into the GDTR register.
• SGDT (Store GDTR Register) — Stores the GDT base address and limit from
the GDTR register into memory.
• LIDT (Load IDTR Register) — Loads the IDT base address and limit from
memory into the IDTR register.
• SIDT (Load IDTR Register — Stores the IDT base address and limit from the
IDTR register into memory.
• LLDT (Load LDT Register) — Loads the LDT segment selector and segment
descriptor from memory into the LDTR. (The segment selector operand can also
be located in a general-purpose register.)
• SLDT (Store LDT Register) — Stores the LDT segment selector from the LDTR
register into memory or a general-purpose register.
• LTR (Load Task Register) — Loads segment selector and segment descriptor
for a TSS from memory into the task register. (The segment selector operand can
also be located in a general-purpose register.)
• STR (Store Task Register) — Stores the segment selector for the current task
TSS from the task register into memory or a general-purpose register.
Vol. 3 2-29
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
The LMSW (load machine status word) and SMSW (store machine status word)
instructions operate on bits 0 through 15 of control register CR0. These instructions
are provided for compatibility with the 16-bit Intel 286 processor. Programs written
should access the control register CR0 using the MOV instruction.
The CLTS (clear TS flag in CR0) instruction is provided for use in handling a
device-not-available exception (#NM) that occurs when the processor attempts to
execute a floating-point instruction when the TS flag is set. This instruction allows
the TS flag to be cleared after the x87 FPU context has been saved, preventing
further #NM exceptions. See Section 2.5, “Control Registers,” for more information
on the TS flag.
The control registers (CR0, CR1, CR2, CR3, CR4, and CR8) are loaded using the MOV
instruction. The instruction loads a control register from a general-purpose register
or stores the content of a control register in a general-purpose register.
2.7.2
Verifying of Access Privileges
The processor provides several instructions for examining segment selectors
and segment descriptors to determine if access to their associated segments
and type checking done by the processor, thus allowing operating-system or
executive software to prevent exceptions from being generated.
The ARPL (adjust RPL) instruction adjusts the RPL (requestor privilege level)
of a segment selector to match that of the program or procedure that
supplied the segment selector. See Section 5.10.4, “Checking Caller Access
The LAR (load access rights) instruction verifies the accessibility of a speci-
fied segment and loads access rights information from the segment’s
segment descriptor into a general-purpose register. Software can then
examine the access rights to determine if the segment type is compatible
with its intended use. See Section 5.10.1, “Checking Access Rights (LAR
instruction.
The LSL (load segment limit) instruction verifies the accessibility of a speci-
fied segment and loads the segment limit from the segment’s segment
descriptor into a general-purpose register. Software can then compare the
segment limit with an offset into the segment to determine whether the
offset lies within the segment. See Section 5.10.3, “Checking That the
Pointer Offset Is Within Limits (LSL Instruction),” for a detailed explanation
of the function and use of this instruction.
The VERR (verify for reading) and VERW (verify for writing) instructions
verify if a selected segment is readable or writable, respectively, at a given
CPL. See Section 5.10.2, “Checking Read/Write Rights (VERR and VERW
2-30 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
Instructions),” for a detailed explanation of the function and use of this
instruction.
2.7.3
Loading and Storing Debug Registers
Internal debugging facilities in the processor are controlled by a set of 8 debug regis-
ters (DR0-DR7). The MOV instruction allows setup data to be loaded to and stored
from these registers.
On processors that support Intel 64 architecture, debug registers DR0-DR7 are 64
bits. In 32-bit modes and compatibility mode, writes to a debug register fill the upper
32 bits with zeros. Reads return the lower 32 bits. In 64-bit mode, the upper 32 bits
of DR6-DR7 are reserved and must be written with zeros. Writing one to any of the
upper 32 bits causes an exception, #GP(0).
In 64-bit mode, MOV DRn instructions read or write all 64 bits of a debug register
(operand-size prefixes are ignored). All 64 bits of DR0-DR3 are writable by software.
However, MOV DRn instructions do not check that addresses written to DR0-DR3 are
in the limits of the implementation. Address matching is supported only on valid
addresses generated by the processor implementation.
2.7.4
Invalidating Caches and TLBs
The processor provides several instructions for use in explicitly invalidating its caches
and TLB entries. The INVD (invalidate cache with no writeback) instruction invali-
dates all data and instruction entries in the internal caches and sends a signal to the
external caches indicating that they should be also be invalidated.
The WBINVD (invalidate cache with writeback) instruction performs the same func-
tion as the INVD instruction, except that it writes back modified lines in its internal
caches to memory before it invalidates the caches. After invalidating the internal
caches, WBINVD signals external caches to write back modified data and invalidate
their contents.
The INVLPG (invalidate TLB entry) instruction invalidates (flushes) the TLB entry for
a specified page.
2.7.5
Controlling the Processor
The HLT (halt processor) instruction stops the processor until an enabled interrupt
(such as NMI or SMI, which are normally enabled), a debug exception, the BINIT#
signal, the INIT# signal, or the RESET# signal is received. The processor generates a
special bus cycle to indicate that the halt mode has been entered.
Hardware may respond to this signal in a number of ways. An indicator light on the
front panel may be turned on. An NMI interrupt for recording diagnostic information
may be generated. Reset initialization may be invoked (note that the BINIT# pin was
Vol. 3 2-31
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
introduced with the Pentium Pro processor). If any non-wake events are pending
during shutdown, they will be handled after the wake event from shutdown is
processed (for example, A20M# interrupts).
The LOCK prefix invokes a locked (atomic) read-modify-write operation when modi-
fying a memory operand. This mechanism is used to allow reliable communications
between processors in multiprocessor systems, as described below:
• In the Pentium processor and earlier IA-32 processors, the LOCK prefix causes
the processor to assert the LOCK# signal during the instruction. This always
causes an explicit bus lock to occur.
• In the Pentium 4, Intel Xeon, and P6 family processors, the locking operation is
handled with either a cache lock or bus lock. If a memory access is cacheable and
affects only a single cache line, a cache lock is invoked and the system bus and
the actual memory location in system memory are not locked during the
operation. Here, other Pentium 4, Intel Xeon, or P6 family processors on the bus
write-back any modified data and invalidate their caches as necessary to
maintain system memory coherency. If the memory access is not cacheable
and/or it crosses a cache line boundary, the processor’s LOCK# signal is asserted
and the processor does not respond to requests for bus control during the locked
operation.
The RSM (return from SMM) instruction restores the processor (from a context
dump) to the state it was in prior to an system management mode (SMM) interrupt.
2.7.6
Reading Performance-Monitoring and Time-Stamp Counters
The RDPMC (read performance-monitoring counter) and RDTSC (read time-stamp
counter) instructions allow application programs to read the processor’s perfor-
mance-monitoring and time-stamp counters, respectively. Processors based on Intel
NetBurst microarchitecture have eighteen 40-bit performance-monitoring counters;
P6 family processors have two 40-bit counters. Intel Atom processors and most of
the processors based on the Intel Core microarchitecture support two types of
performance monitoring counters: two programmable performance counters similar
to those available in the P6 family, and three fixed-function performance monitoring
counters.
The programmable performance counters can support counting either the occurrence
or duration of events. Events that can be monitored on programmable counters
generally are model specific (except for architectural performance events enumer-
ated by CPUID leaf 0AH); they may include the number of instructions decoded,
interrupts received, or the number of cache loads. Individual counters can be set up
to monitor different events. Use the system instruction WRMSR to set up values in
IA32_PERFEVTSEL0/1 (for Intel Atom, Intel Core 2, Intel Core Duo, and Intel
Pentium M processors), in one of the 45 ESCRs and one of the 18 CCCR MSRs (for
Pentium 4 and Intel Xeon processors); or in the PerfEvtSel0 or the PerfEvtSel1 MSR
(for the P6 family processors). The RDPMC instruction loads the current count from
the selected counter into the EDX:EAX registers.
2-32 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
Fixed-function performance counters record only specific events that are defined in
Chapter 20, “Introduction to Virtual-Machine Extensions”, and the width/number of
fixed-function counters are enumerated by CPUID leaf 0AH.
The time-stamp counter is a model-specific 64-bit counter that is reset to zero each
16
times per year when the processor is operating at a clock rate of 3GHz. At this
clock frequency, it would take over 190 years for the counter to wrap around. The
RDTSC instruction loads the current count of the time-stamp counter into the
EDX:EAX registers.
See Section 30.1, “Performance Monitoring Overview,” and Section 16.11, “Time-
Stamp Counter,” for more information about the performance monitoring and time-
stamp counters.
The RDTSC instruction was introduced into the IA-32 architecture with the Pentium
processor. The RDPMC instruction was introduced into the IA-32 architecture with the
Pentium Pro processor and the Pentium processor with MMX technology. Earlier
Pentium processors have two performance-monitoring counters, but they can be
read only with the RDMSR instruction, and only at privilege level 0.
2.7.6.1
Reading Counters in 64-Bit Mode
In 64-bit mode, RDTSC operates the same as in protected mode. The count in the
time-stamp counter is stored in EDX:EAX (or RDX[31:0]:RAX[31:0] with
RDX[63:32]:RAX[63:32] cleared).
RDPMC requires an index to specify the offset of the performance-monitoring
counter. In 64-bit mode for Pentium 4 or Intel Xeon processor families, the index is
specified in ECX[30:0]. The current count of the performance-monitoring counter is
stored in EDX:EAX (or RDX[31:0]:RAX[31:0] with RDX[63:32]:RAX[63:32]
cleared).
2.7.7
Reading and Writing Model-Specific Registers
The RDMSR (read model-specific register) and WRMSR (write model-specific
register) instructions allow a processor’s 64-bit model-specific registers (MSRs) to be
in the ECX register.
RDMSR reads the value from the specified MSR to the EDX:EAX registers; WRMSR
writes the value in the EDX:EAX registers to the specified MSR. RDMSR and WRMSR
were introduced into the IA-32 architecture with the Pentium processor.
See Section 9.4, “Model-Specific Registers (MSRs),” for more information.
Vol. 3 2-33
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM ARCHITECTURE OVERVIEW
2.7.7.1
Reading and Writing Model-Specific Registers in 64-Bit Mode
RDMSR and WRMSR require an index to specify the address of an MSR. In 64-bit
mode, the index is 32 bits; it is specified using ECX.
2.7.8
Enabling Processor Extended States
The XSETBV instruction is required to enable OS support of individual processor
extended states in the XFEATURE_ENABLED_MASK register (see Section 2.6).
2-34 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
CHAPTER 3
This chapter describes the Intel 64 and IA-32 architecture’s protected-mode memory
management facilities, including the physical memory requirements, segmentation
mechanism, and paging mechanism.
See also: Chapter 5, “Protection” (for a description of the processor’s protection
mechanism) and Chapter 17, “8086 Emulation” (for a description of memory
addressing protection in real-address and virtual-8086 modes).
3.1
MEMORY MANAGEMENT OVERVIEW
The memory management facilities of the IA-32 architecture are divided into two
parts: segmentation and paging. Segmentation provides a mechanism of isolating
individual code, data, and stack modules so that multiple programs (or tasks) can
run on the same processor without interfering with one another. Paging provides a
mechanism for implementing a conventional demand-paged, virtual-memory system
where sections of a program’s execution environment are mapped into physical
memory as needed. Paging can also be used to provide isolation between multiple
There is no mode bit to disable segmentation. The use of paging, however, is
optional.
These two mechanisms (segmentation and paging) can be configured to support
simple single-program (or single-task) systems, multitasking systems, or multiple-
processor systems that used shared memory.
As shown in Figure 3-1, segmentation provides a mechanism for dividing the
processor’s addressable memory space (called the linear address space) into
smaller protected address spaces called segments. Segments can be used to hold
the code, data, and stack for a program or to hold system data structures (such as a
TSS or LDT). If more than one program (or task) is running on a processor, each
program can be assigned its own set of segments. The processor then enforces the
boundaries between these segments and insures that one program does not interfere
with the execution of another program by writing into the other program’s segments.
The segmentation mechanism also allows typing of segments so that the operations
that may be performed on a particular type of segment can be restricted.
All the segments in a system are contained in the processor’s linear address space.
To locate a byte in a particular segment, a logical address (also called a far pointer)
must be provided. A logical address consists of a segment selector and an offset. The
segment selector is a unique identifier for a segment. Among other things it provides
an offset into a descriptor table (such as the global descriptor table, GDT) to a data
structure called a segment descriptor. Each segment has a segment descriptor, which
specifies the size of the segment, the access rights and privilege level for the
Vol. 3 3-1
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
segment, the segment type, and the location of the first byte of the segment in the
linear address space (called the base address of the segment). The offset part of the
logical address is added to the base address for the segment to locate a byte within
the segment. The base address plus the offset thus forms a linear address in the
processor’s linear address space.
Logical Address
(or Far Pointer)
Segment
Selector
Offset
Linear Address
Space
Linear Address
Table Offset
Global Descriptor
Table (GDT)
Physical
Address
Space
Dir
Segment
Lin. Addr.
Page Table
Entry
Page
Segment
Descriptor
Page Directory
Entry
Phy. Addr.
Segment
Base Address
Page
Paging
Segmentation
Figure 3-1. Segmentation and Paging
If paging is not used, the linear address space of the processor is mapped directly
into the physical address space of processor. The physical address space is defined as
the range of addresses that the processor can generate on its address bus.
Because multitasking computing systems commonly define a linear address space
much larger than it is economically feasible to contain all at once in physical memory,
some method of “virtualizing” the linear address space is needed. This virtualization
of the linear address space is handled through the processor’s paging mechanism.
Paging supports a “virtual memory” environment where a large linear address space
is simulated with a small amount of physical memory (RAM and ROM) and some disk
3-2 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
storage. When using paging, each segment is divided into pages (typically 4 KBytes
each in size), which are stored either in physical memory or on the disk. The oper-
ating system or executive maintains a page directory and a set of page tables to keep
track of the pages. When a program (or task) attempts to access an address location
in the linear address space, the processor uses the page directory and page tables to
translate the linear address into a physical address and then performs the requested
operation (read or write) on the memory location.
If the page being accessed is not currently in physical memory, the processor inter-
rupts execution of the program (by generating a page-fault exception). The oper-
ating system or executive then reads the page into physical memory from the disk
and continues executing the program.
When paging is implemented properly in the operating-system or executive, the
swapping of pages between physical memory and the disk is transparent to the
correct execution of a program. Even programs written for 16-bit IA-32 processors
can be paged (transparently) when they are run in virtual-8086 mode.
3.2
USING SEGMENTS
The segmentation mechanism supported by the IA-32 architecture can be used to
implement a wide variety of system designs. These designs range from flat models
that make only minimal use of segmentation to protect programs to multi-
segmented models that employ segmentation to create a robust operating environ-
ment in which multiple programs and tasks can be executed reliably.
The following sections give several examples of how segmentation can be employed
in a system to improve memory management performance and reliability.
3.2.1
Basic Flat Model
The simplest memory model for a system is the basic “flat model,” in which the oper-
address space. To the greatest extent possible, this basic flat model hides the
segmentation mechanism of the architecture from both the system designer and the
application programmer.
To implement a basic flat memory model with the IA-32 architecture, at least two
segment descriptors must be created, one for referencing a code segment and one
for referencing a data segment (see Figure 3-2). Both of these segments, however,
are mapped to the entire linear address space: that is, both segment descriptors
have the same base address value of 0 and the same segment limit of 4 GBytes. By
setting the segment limit to 4 GBytes, the segmentation mechanism is kept from
generating exceptions for out of limit memory references, even if no physical
memory resides at a particular address. ROM (EPROM) is generally located at the top
of the physical address space, because the processor begins execution at
Vol. 3 3-3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
FFFF_FFF0H. RAM (DRAM) is placed at the bottom of the address space because the
3.2.2
Protected Flat Model
The protected flat model is similar to the basic flat model, except the segment limits
are set to include only the range of addresses for which physical memory actually
exists (see Figure 3-3). A general-protection exception (#GP) is then generated on
any attempt to access nonexistent memory. This model provides a minimum level of
hardware protection against some kinds of program bugs.
Linear Address Space
(or Physical Memory)
Segment
Registers
FFFFFFFFH
Code
CS
Code- and Data-Segment
Descriptors
Not Present
SS
DS
ES
FS
GS
Access
Limit
Data and
Stack
Base Address
0
Figure 3-2. Flat Model
Segment
Descriptors
Linear Address Space
(or Physical Memory)
Segment
Registers
Access
Limit
FFFFFFFFH
Code
Base Address
CS
Not Present
Memory I/O
ES
SS
DS
FS
GS
Access
Limit
Base Address
Data and
Stack
0
Figure 3-3. Protected Flat Model
3-4 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
More complexity can be added to this protected flat model to provide more protec-
tion. For example, for the paging mechanism to provide isolation between user and
supervisor code and data, four segments need to be defined: code and data
segments at privilege level 3 for the user, and code and data segments at privilege
level 0 for the supervisor. Usually these segments all overlay each other and start at
address 0 in the linear address space. This flat segmentation model along with a
simple paging structure can protect the operating system from applications, and by
cations from each other. Similar designs are used by several popular multitasking
operating systems.
3.2.3
Multi-Segment Model
A multi-segment model (such as the one shown in Figure 3-4) uses the full capabili-
ties of the segmentation mechanism to provided hardware enforced protection of
code, data structures, and programs and tasks. Here, each program (or task) is given
its own table of segment descriptors and its own segments. The segments can be
completely private to their assigned programs or shared among programs. Access to
all segments and to the execution environments of individual programs running on
the system is controlled by hardware.
Vol. 3 3-5
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
Segment
Descriptors
Linear Address Space
(or Physical Memory)
Segment
Registers
Access Limit
Base Address
Access Limit
Base Address
Access Limit
Base Address
Access Limit
Base Address
Access Limit
Base Address
Access Limit
Base Address
Access Limit
Base Address
Access Limit
Base Address
Access Limit
Base Address
Access Limit
CS
SS
DS
ES
FS
GS
Stack
Code
Data
Data
Data
Data
Base Address
Figure 3-4. Multi-Segment Model
Access checks can be used to protect not only against referencing an address outside
the limit of a segment, but also against performing disallowed operations in certain
segments. For example, since code segments are designated as read-only segments,
hardware can be used to prevent writes into code segments. The access rights infor-
mation created for segments can also be used to set up protection rings or levels.
Protection levels can be used to protect operating-system procedures from unautho-
rized access by application programs.
3.2.4
Segmentation in IA-32e Mode
In IA-32e mode of Intel 64 architecture, the effects of segmentation depend on
whether the processor is running in compatibility mode or 64-bit mode. In compati-
bility mode, segmentation functions just as it does using legacy 16-bit or 32-bit
protected mode semantics.
3-6 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
In 64-bit mode, segmentation is generally (but not completely) disabled, creating a
flat 64-bit linear-address space. The processor treats the segment base of CS, DS,
ES, SS as zero, creating a linear address that is equal to the effective address. The FS
and GS segments are exceptions. These segment registers (which hold the segment
base) can be used as an additional base registers in linear address calculations. They
facilitate addressing local data and certain operating system data structures.
mode.
3.2.5
Paging and Segmentation
Paging can be used with any of the segmentation models described in Figures 3-2,
3-3, and 3-4. The processor’s paging mechanism divides the linear address space
(into which segments are mapped) into pages (as shown in Figure 3-1). These linear-
address-space pages are then mapped to pages in the physical address space. The
paging mechanism offers several page-level protection facilities that can be used
with or instead of the segment-protection facilities. For example, it lets read-write
protection be enforced on a page-by-page basis. The paging mechanism also
provides two-level user-supervisor protection that can also be specified on a page-
by-page basis.
3.3
PHYSICAL ADDRESS SPACE
In protected mode, the IA-32 architecture provides a normal physical address space
32
of 4 GBytes (2 bytes). This is the address space that the processor can address on
its address bus. This address space is flat (unsegmented), with addresses ranging
continuously from 0 to FFFFFFFFH. This physical address space can be mapped to
read-write memory, read-only memory, and memory mapped I/O. The memory
mapping facilities described in this chapter can be used to divide this physical
memory up into segments and/or pages.
Starting with the Pentium Pro processor, the IA-32 architecture also supports an
36
extension of the physical address space to 2 bytes (64 GBytes); with a maximum
physical address of FFFFFFFFFH. This extension is invoked in either of two ways:
• Using the physical address extension (PAE) flag, located in bit 5 of control
register CR4.
• Using the 36-bit page size extension (PSE-36) feature (introduced in the Pentium
III processors).
Physical address support has since been extended beyond 36 bits. See Chapter 4,
“Paging” for more information about 36-bit physical addressing.
Vol. 3 3-7
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
3.3.1
On processors that support Intel 64 architecture (CPUID.80000001:EDX[29] = 1),
the size of the physical address range is implementation-specific and indicated by
CPUID.80000008H:EAX[bits 7-0].
For the format of information returned in EAX, see “CPUID—CPU Identification” in
Chapter 3 of the Intel® 64 and IA-32 Architectures Software Developer’s Manual,
Volume 2A. See also: Chapter 4, “Paging.”
3.4
LOGICAL AND LINEAR ADDRESSES
At the system-architecture level in protected mode, the processor uses two stages of
address translation to arrive at a physical address: logical-address translation and
linear address space paging.
Even with the minimum use of segments, every byte in the processor’s address
space is accessed with a logical address. A logical address consists of a 16-bit
segment selector and a 32-bit offset (see Figure 3-5). The segment selector identi-
fies the segment the byte is located in and the offset specifies the location of the byte
in the segment relative to the base address of the segment.
The processor translates every logical address into a linear address. A linear address
is a 32-bit address in the processor’s linear address space. Like the physical address
32
space, the linear address space is a flat (unsegmented), 2 -byte address space,
with addresses ranging from 0 to FFFFFFFFH. The linear address space contains all
the segments and system tables defined for a system.
To translate a logical address into a linear address, the processor does the following:
1. Uses the offset in the segment selector to locate the segment descriptor for the
segment in the GDT or LDT and reads it into the processor. (This step is needed
only when a new segment selector is loaded into a segment register.)
2. Examines the segment descriptor to check the access rights and range of the
segment to insure that the segment is accessible and that the offset is within the
limits of the segment.
3. Adds the base address of the segment from the segment descriptor to the offset
to form a linear address.
3-8 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
15
0
31(63)
Offset (Effective Address)
0
Logical
Address
Seg. Selector
Descriptor Table
Base Address
Segment
Descriptor
+
31(63)
0
Linear Address
Figure 3-5. Logical Address to Linear Address Translation
If paging is not used, the processor maps the linear address directly to a physical
address (that is, the linear address goes out on the processor’s address bus). If the
linear address space is paged, a second level of address translation is used to trans-
late the linear address into a physical address.
See also: Chapter 4, “Paging.”
3.4.1
Logical Address Translation in IA-32e Mode
In IA-32e mode, an Intel 64 processor uses the steps described above to translate a
logical address to a linear address. In 64-bit mode, the offset and base address of the
segment are 64-bits instead of 32 bits. The linear address format is also 64 bits wide
and is subject to the canonical form requirement.
Each code segment descriptor provides an L bit. This bit allows a code segment to
execute 64-bit code or legacy 32-bit code by code segment.
3.4.2
Segment Selectors
A segment selector is a 16-bit identifier for a segment (see Figure 3-6). It does not
point directly to the segment, but instead points to the segment descriptor that
defines the segment. A segment selector contains the following items:
Index
(Bits 3 through 15) — Selects one of 8192 descriptors in the GDT or
LDT. The processor multiplies the index value by 8 (the number of
bytes in a segment descriptor) and adds the result to the base address
of the GDT or LDT (from the GDTR or LDTR register, respectively).
Vol. 3 3-9
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
TI (table indicator) flag
(Bit 2) — Specifies the descriptor table to use: clearing this flag
selects the GDT; setting this flag selects the current LDT.
15
3
1
0
2
T
I
Index
RPL
Table Indicator
0 = GDT
1 = LDT
Requested Privilege Level (RPL)
Figure 3-6. Segment Selector
Requested Privilege Level (RPL)
(Bits 0 and 1) — Specifies the privilege level of the selector. The priv-
ilege level can range from 0 to 3, with 0 being the most privileged
level. See Section 5.5, “Privilege Levels”, for a description of the rela-
tionship of the RPL to the CPL of the executing program (or task) and
the descriptor privilege level (DPL) of the descriptor the segment
selector points to.
The first entry of the GDT is not used by the processor. A segment selector that points
to this entry of the GDT (that is, a segment selector with an index of 0 and the TI flag
set to 0) is used as a “null segment selector.” The processor does not generate an
exception when a segment register (other than the CS or SS registers) is loaded with
a null selector. It does, however, generate an exception when a segment register
holding a null selector is used to access memory. A null selector can be used to
initialize unused segment registers. Loading the CS or SS register with a null
segment selector causes a general-protection exception (#GP) to be generated.
Segment selectors are visible to application programs as part of a pointer variable,
but the values of selectors are usually assigned or modified by link editors or linking
loaders, not application programs.
3.4.3
Segment Registers
To reduce address translation time and coding complexity, the processor provides
registers for holding up to 6 segment selectors (see Figure 3-7). Each of these
segment registers support a specific kind of memory reference (code, stack, or
data). For virtually any kind of program execution to take place, at least the code-
segment (CS), data-segment (DS), and stack-segment (SS) registers must be
loaded with valid segment selectors. The processor also provides three additional
data-segment registers (ES, FS, and GS), which can be used to make additional data
segments available to the currently executing program (or task).
3-10 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
For a program to access a segment, the segment selector for the segment must have
been loaded in one of the segment registers. So, although a system can define thou-
sands of segments, only 6 can be available for immediate use. Other segments can
be made available by loading their segment selectors into these registers during
program execution.
Visible Part
Hidden Part
Segment Selector
Base Address, Limit, Access Information
CS
SS
DS
ES
FS
GS
Figure 3-7. Segment Registers
Every segment register has a “visible” part and a “hidden” part. (The hidden part is
sometimes referred to as a “descriptor cache” or a “shadow register.”) When a
segment selector is loaded into the visible part of a segment register, the processor
also loads the hidden part of the segment register with the base address, segment
limit, and access control information from the segment descriptor pointed to by the
segment selector. The information cached in the segment register (visible and
hidden) allows the processor to translate addresses without taking extra bus cycles
to read the base address and limit from the segment descriptor. In systems in which
multiple processors have access to the same descriptor tables, it is the responsibility
of software to reload the segment registers when the descriptor tables are modified.
If this is not done, an old segment descriptor cached in a segment register might be
used after its memory-resident version has been modified.
Two kinds of load instructions are provided for loading the segment registers:
1. Direct load instructions such as the MOV, POP, LDS, LES, LSS, LGS, and LFS
instructions. These instructions explicitly reference the segment registers.
2. Implied load instructions such as the far pointer versions of the CALL, JMP, and
RET instructions, the SYSENTER and SYSEXIT instructions, and the IRET, INTn,
INTO and INT3 instructions. These instructions change the contents of the CS
register (and sometimes other segment registers) as an incidental part of their
operation.
The MOV instruction can also be used to store visible part of a segment register in a
general-purpose register.
Vol. 3 3-11
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
3.4.4
Segment Loading Instructions in IA-32e Mode
Because ES, DS, and SS segment registers are not used in 64-bit mode, their fields
(base, limit, and attribute) in segment descriptor registers are ignored. Some forms
of segment load instructions are also invalid (for example, LDS, POP ES). Address
calculations that reference the ES, DS, or SS segments are treated as if the segment
base is zero.
The processor checks that all linear-address references are in canonical form instead
of performing limit checks. Mode switching does not change the contents of the
segment registers or the associated descriptor registers. These registers are also not
changed during 64-bit mode execution, unless explicit segment loads are performed.
In order to set up compatibility mode for an application, segment-load instructions
(MOV to Sreg, POP Sreg) work normally in 64-bit mode. An entry is read from the
system descriptor table (GDT or LDT) and is loaded in the hidden portion of the
segment descriptor register. The descriptor-register base, limit, and attribute fields
are all loaded. However, the contents of the data and stack segment selector and the
descriptor registers are ignored.
When FS and GS segment overrides are used in 64-bit mode, their respective base
addresses are used in the linear address calculation: (FS or GS).base + index +
displacement. FS.base and GS.base are then expanded to the full linear-address size
supported by the implementation. The resulting effective address calculation can
wrap across positive and negative addresses; the resulting linear address must be
canonical.
In 64-bit mode, memory accesses using FS-segment and GS-segment overrides are
not checked for a runtime limit nor subjected to attribute-checking. Normal segment
loads (MOV to Sreg and POP Sreg) into FS and GS load a standard 32-bit base value
in the hidden portion of the segment descriptor register. The base address bits above
the standard 32 bits are cleared to 0 to allow consistency for implementations that
use less than 64 bits.
The hidden descriptor register fields for FS.base and GS.base are physically mapped
to MSRs in order to load all address bits supported by a 64-bit implementation. Soft-
ware with CPL = 0 (privileged software) can load all supported linear-address bits
into FS.base or GS.base using WRMSR. Addresses written into the 64-bit FS.base and
GS.base registers must be in canonical form. A WRMSR instruction that attempts to
write a non-canonical address to those registers causes a #GP fault.
When in compatibility mode, FS and GS overrides operate as defined by 32-bit mode
behavior regardless of the value loaded into the upper 32 linear-address bits of the
hidden descriptor register base field. Compatibility mode ignores the upper 32 bits
when calculating an effective address.
A new 64-bit mode instruction, SWAPGS, can be used to load GS base. SWAPGS
exchanges the kernel data structure pointer from the IA32_KernelGSbase MSR with
the GS base register. The kernel can then use the GS prefix on normal memory refer-
ences to access the kernel data structures. An attempt to write a non-canonical value
(using WRMSR) to the IA32_KernelGSBase MSR causes a #GP fault.
3-12 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
3.4.5
Segment Descriptors
A segment descriptor is a data structure in a GDT or LDT that provides the processor
with the size and location of a segment, as well as access control and status informa-
tion. Segment descriptors are typically created by compilers, linkers, loaders, or the
operating system or executive, but not application programs. Figure 3-8 illustrates
the general descriptor format for all types of segment descriptors.
31
31
23
G
21 20 19
A
V
L
16
24
22
15 14 13 12 11
D
0
8
7
Seg.
Limit
19:16
D
/
B
Base 31:24
L
P
S
Type
Base 23:16
P
L
4
0
16
15
0
Base Address 15:00
Segment Limit 15:00
L
— 64-bit code segment (IA-32e mode only)
AVL — Available for use by system software
BASE — Segment base address
D/B — Default operation size (0 = 16-bit segment; 1 = 32-bit segment)
DPL — Descriptor privilege level
G
— Granularity
LIMIT — Segment Limit
P
S
— Segment present
— Descriptor type (0 = system; 1 = code or data)
TYPE — Segment type
Figure 3-8. Segment Descriptor
The flags and fields in a segment descriptor are as follows:
Segment limit field
Specifies the size of the segment. The processor puts together the
two segment limit fields to form a 20-bit value. The processor inter-
prets the segment limit in one of two ways, depending on the setting
of the G (granularity) flag:
•
1 byte to 1 MByte, in byte increments.
•
If the granularity flag is set, the segment size can range from
4 KBytes to 4 GBytes, in 4-KByte increments.
The processor uses the segment limit in two different ways,
depending on whether the segment is an expand-up or an expand-
down segment. See Section 3.4.5.1, “Code- and Data-Segment
Descriptor Types”, for more information about segment types. For
expand-up segments, the offset in a logical address can range from 0
Vol. 3 3-13
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
to the segment limit. Offsets greater than the segment limit generate
general-protection exceptions (#GP). For expand-down segments,
the segment limit has the reverse function; the offset can range from
the segment limit to FFFFFFFFH or FFFFH, depending on the setting of
the B flag. Offsets less than the segment limit generate general-
protection exceptions. Decreasing the value in the segment limit field
for an expand-down segment allocates new memory at the bottom of
the segment's address space, rather than at the top. IA-32 architec-
ture stacks always grow downwards, making this mechanism conve-
nient for expandable stacks.
Base address fields
Defines the location of byte 0 of the segment within the 4-GByte
linear address space. The processor puts together the three base
address fields to form a single 32-bit value. Segment base addresses
should be aligned to 16-byte boundaries. Although 16-byte alignment
is not required, this alignment allows programs to maximize perfor-
mance by aligning code and data on 16-byte boundaries.
Type field
Indicates the segment or gate type and specifies the kinds of access
that can be made to the segment and the direction of growth. The
interpretation of this field depends on whether the descriptor type flag
specifies an application (code or data) descriptor or a system
descriptor. The encoding of the type field is different for code, data,
and system descriptors (see Figure 5-1). See Section 3.4.5.1, “Code-
and Data-Segment Descriptor Types”, for a description of how this
field is used to specify code and data-segment types.
S (descriptor type) flag
Specifies whether the segment descriptor is for a system segment
(S flag is clear) or a code or data segment (S flag is set).
DPL (descriptor privilege level) field
Specifies the privilege level of the segment. The privilege level can
range from 0 to 3, with 0 being the most privileged level. The DPL is
used to control access to the segment. See Section 5.5, “Privilege
Levels”, for a description of the relationship of the DPL to the CPL of
the executing code segment and the RPL of a segment selector.
P (segment-present) flag
Indicates whether the segment is present in memory (set) or not
segment-not-present exception (#NP) when a segment selector that
points to the segment descriptor is loaded into a segment register.
Memory management software can use this flag to control which
segments are actually loaded into physical memory at a given time. It
offers a control in addition to paging for managing virtual memory.
Figure 3-9 shows the format of a segment descriptor when the
segment-present flag is clear. When this flag is clear, the operating
system or executive is free to use the locations marked “Available” to
3-14 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
store its own data, such as information regarding the whereabouts of
the missing segment.
D/B (default operation size/default stack pointer size and/or upper bound)
flag
Performs different functions depending on whether the segment
descriptor is an executable code segment, an expand-down data
segment, or a stack segment. (This flag should always be set to 1 for
32-bit code and data segments and to 0 for 16-bit code and data
segments.)
•
Executable code segment. The flag is called the D flag and it
indicates the default length for effective addresses and operands
referenced by instructions in the segment. If the flag is set, 32-bit
addresses and 32-bit or 8-bit operands are assumed; if it is clear,
16-bit addresses and 16-bit or 8-bit operands are assumed.
The instruction prefix 66H can be used to select an operand size
other than the default, and the prefix 67H can be used select an
address size other than the default.
•
Stack segment (data segment pointed to by the SS
register). The flag is called the B (big) flag and it specifies the
size of the stack pointer used for implicit stack operations (such as
pushes, pops, and calls). If the flag is set, a 32-bit stack pointer is
used, which is stored in the 32-bit ESP register; if the flag is clear,
a 16-bit stack pointer is used, which is stored in the 16-bit SP
register. If the stack segment is set up to be an expand-down data
segment (described in the next paragraph), the B flag also
specifies the upper bound of the stack segment.
•
Expand-down data segment. The flag is called the B flag and it
specifies the upper bound of the segment. If the flag is set, the
upper bound is FFFFFFFFH (4 GBytes); if the flag is clear, the
upper bound is FFFFH (64 KBytes).
31
31
16
15 14 13 12 11
D
0
8
7
Available
0
S
Type
Available
P
L
4
0
0
Available
Figure 3-9. Segment Descriptor When Segment-Present Flag Is Clear
Vol. 3 3-15
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
G (granularity) flag
Determines the scaling of the segment limit field. When the
granularity flag is clear, the segment limit is interpreted in byte
units; when flag is set, the segment limit is interpreted in
4-KByte units. (This flag does not affect the granularity of the
base address; it is always byte granular.) When the granularity
flag is set, the twelve least significant bits of an offset are not
tested when checking the offset against the segment limit. For
example, when the granularity flag is set, a limit of 0 results in
valid offsets from 0 to 4095.
L (64-bit code segment) flag
In IA-32e mode, bit 21 of the second doubleword of the segment
descriptor indicates whether a code segment contains native 64-bit
code. A value of 1 indicates instructions in this code segment are
executed in 64-bit mode. A value of 0 indicates the instructions in this
code segment are executed in compatibility mode. If L-bit is set, then
D-bit must be cleared. When not in IA-32e mode or for non-code
segments, bit 21 is reserved and should always be set to 0.
Available and reserved bits
Bit 20 of the second doubleword of the segment descriptor is available
for use by system software.
3.4.5.1
Code- and Data-Segment Descriptor Types
either a code or a data segment. The highest order bit of the type field (bit 11 of the
second double word of the segment descriptor) then determines whether the
descriptor is for a data segment (clear) or a code segment (set).
For data segments, the three low-order bits of the type field (bits 8, 9, and 10) are
interpreted as accessed (A), write-enable (W), and expansion-direction (E). See
Table 3-1 for a description of the encoding of the bits in the type field for code and
data segments. Data segments can be read-only or read/write segments, depending
on the setting of the write-enable bit.
3-16 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
Table 3-1. Code- and Data-Segment Types
Type Field
Descriptor
Type
Description
Decimal
11
10
E
9
W
8
A
0
1
2
3
4
5
6
7
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
C
0
0
0
0
1
1
1
1
0
0
1
1
0
0
1
1
R
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
A
0
1
0
1
0
1
0
1
Data
Data
Data
Data
Data
Data
Data
Data
Read-Only
Read-Only, accessed
Read/Write
Read/Write, accessed
Read-Only, expand-down
Read-Only, expand-down, accessed
Read/Write, expand-down
Read/Write, expand-down, accessed
8
1
1
1
1
1
1
1
1
Code
Code
Code
Code
Code
Code
Code
Code
Execute-Only
9
Execute-Only, accessed
Execute/Read
10
11
12
13
14
15
Execute/Read, accessed
Execute-Only, conforming
Execute-Only, conforming, accessed
Execute/Read, conforming
Execute/Read, conforming, accessed
Stack segments are data segments which must be read/write segments. Loading the
SS register with a segment selector for a nonwritable data segment generates a
general-protection exception (#GP). If the size of a stack segment needs to be
changed dynamically, the stack segment can be an expand-down data segment
(expansion-direction flag set). Here, dynamically changing the segment limit causes
stack space to be added to the bottom of the stack. If the size of a stack segment is
intended to remain static, the stack segment may be either an expand-up or expand-
down type.
The accessed bit indicates whether the segment has been accessed since the last
time the operating-system or executive cleared the bit. The processor sets this bit
whenever it loads a segment selector for the segment into a segment register,
assuming that the type of memory that contains the segment descriptor supports
processor writes. The bit remains set until explicitly cleared. This bit can be used both
for virtual memory management and for debugging.
Vol. 3 3-17
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
For code segments, the three low-order bits of the type field are interpreted as
accessed (A), read enable (R), and conforming (C). Code segments can be execute-
only or execute/read, depending on the setting of the read-enable bit. An
execute/read segment might be used when constants or other static data have been
placed with instruction code in a ROM. Here, data can be read from the code segment
either by using an instruction with a CS override prefix or by loading a segment
selector for the code segment in a data-segment register (the DS, ES, FS, or GS
registers). In protected mode, code segments are not writable.
Code segments can be either conforming or nonconforming. A transfer of execution
into a more-privileged conforming segment allows execution to continue at the
current privilege level. A transfer into a nonconforming segment at a different privi-
lege level results in a general-protection exception (#GP), unless a call gate or task
gate is used (see Section 5.8.1, “Direct Calls or Jumps to Code Segments”, for more
information on conforming and nonconforming code segments). System utilities that
do not access protected facilities and handlers for some types of exceptions (such as,
divide error or overflow) may be loaded in conforming code segments. Utilities that
need to be protected from less privileged programs and procedures should be placed
in nonconforming code segments.
NOTE
Execution cannot be transferred by a call or a jump to a less-
privileged (numerically higher privilege level) code segment,
regardless of whether the target segment is a conforming or noncon-
forming code segment. Attempting such an execution transfer will
result in a general-protection exception.
All data segments are nonconforming, meaning that they cannot be accessed by less
privileged programs or procedures (code executing at numerically high privilege
levels). Unlike code segments, however, data segments can be accessed by more
privileged programs or procedures (code executing at numerically lower privilege
levels) without using a special access gate.
If the segment descriptors in the GDT or an LDT are placed in ROM, the processor can
enter an indefinite loop if software or the processor attempts to update (write to) the
ROM-based segment descriptors. To prevent this problem, set the accessed bits for
all segment descriptors placed in a ROM. Also, remove operating-system or executive
code that attempts to modify segment descriptors located in ROM.
3.5
SYSTEM DESCRIPTOR TYPES
When the S (descriptor type) flag in a segment descriptor is clear, the descriptor type
is a system descriptor. The processor recognizes the following types of system
descriptors:
• Local descriptor-table (LDT) segment descriptor.
3-18 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
• Task-state segment (TSS) descriptor.
• Call-gate descriptor.
• Interrupt-gate descriptor.
• Trap-gate descriptor.
• Task-gate descriptor.
These descriptor types fall into two categories: system-segment descriptors and gate
descriptors. System-segment descriptors point to system segments (LDT and TSS
segments). Gate descriptors are in themselves “gates,” which hold pointers to proce-
dure entry points in code segments (call, interrupt, and trap gates) or which hold
segment selectors for TSS’s (task gates).
Table 3-2 shows the encoding of the type field for system-segment descriptors and
gate descriptors. Note that system descriptors in IA-32e mode are 16 bytes instead
of 8 bytes.
Table 3-2. System-Segment and Gate-Descriptor Types
Type Field
Description
Decimal
11
10
9
8
32-Bit Mode
Reserved
IA-32e Mode
0
0
0
0
0
Upper 8 byte of an 16-
byte descriptor
1
2
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
16-bit TSS (Available)
LDT
Reserved
LDT
3
16-bit TSS (Busy)
16-bit Call Gate
Task Gate
Reserved
4
Reserved
5
Reserved
6
16-bit Interrupt Gate
16-bit Trap Gate
Reserved
Reserved
7
Reserved
8
Reserved
9
32-bit TSS (Available)
Reserved
64-bit TSS (Available)
Reserved
10
11
12
13
14
15
32-bit TSS (Busy)
32-bit Call Gate
Reserved
64-bit TSS (Busy)
64-bit Call Gate
Reserved
32-bit Interrupt Gate
32-bit Trap Gate
64-bit Interrupt Gate
64-bit Trap Gate
Vol. 3 3-19
Download from Www.Somanuals.com. All Manuals Search And Download.
See also: Section 3.5.1, “Segment Descriptor Tables”, and Section 7.2.2, “TSS
Descriptor” (for more information on the system-segment descriptors); see Section
5.8.3, “Call Gates”, Section 6.11, “IDT Descriptors”, and Section 7.2.5, “Task-Gate
Descriptor” (for more information on the gate descriptors).
3.5.1
Segment Descriptor Tables
A segment descriptor table is an array of segment descriptors (see Figure 3-10). A
13
descriptor table is variable in length and can contain up to 8192 (2 ) 8-byte descrip-
tors. There are two kinds of descriptor tables:
• The global descriptor table (GDT)
• The local descriptor tables (LDT)
Global
Local
Descriptor
Table (GDT)
Descriptor
Table (LDT)
T
I
TI = 0
TI = 1
Segment
Selector
56
48
40
32
24
56
48
40
32
24
16
8
16
8
First Descriptor in
GDT is Not Used
0
0
GDTR Register
LDTR Register
Limit
Limit
Base Address
Base Address
Seg. Sel.
Figure 3-10. Global and Local Descriptor Tables
3-20 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
tasks in the system. Optionally, one or more LDTs can be defined. For example, an
LDT can be defined for each separate task being run, or some or all tasks can share
the same LDT.
The GDT is not a segment itself; instead, it is a data structure in linear address space.
The base linear address and limit of the GDT must be loaded into the GDTR register
(see Section 2.4, “Memory-Management Registers”). The base addresses of the GDT
should be aligned on an eight-byte boundary to yield the best processor perfor-
mance. The limit value for the GDT is expressed in bytes. As with segments, the limit
value is added to the base address to get the address of the last valid byte. A limit
value of 0 results in exactly one valid byte. Because segment descriptors are always
8 bytes long, the GDT limit should always be one less than an integral multiple of
eight (that is, 8N – 1).
The first descriptor in the GDT is not used by the processor. A segment selector to
this “null descriptor” does not generate an exception when loaded into a data-
segment register (DS, ES, FS, or GS), but it always generates a general-protection
exception (#GP) when an attempt is made to access memory using the descriptor. By
unused segment registers can be guaranteed to generate an exception.
The LDT is located in a system segment of the LDT type. The GDT must contain a
segment descriptor for the LDT segment. If the system supports multiple LDTs, each
segment descriptor for an LDT can be located anywhere in the GDT. See Section 3.5,
An LDT is accessed with its segment selector. To eliminate address translations when
accessing the LDT, the segment selector, base linear address, limit, and access rights
of the LDT are stored in the LDTR register (see Section 2.4, “Memory-Management
Registers”).
When the GDTR register is stored (using the SGDT instruction), a 48-bit “pseudo-
descriptor” is stored in memory (see top diagram in Figure 3-11). To avoid alignment
check faults in user mode (privilege level 3), the pseudo-descriptor should be located
at an odd word address (that is, address MOD 4 is equal to 2). This causes the
processor to store an aligned word, followed by an aligned doubleword. User-mode
programs normally do not store pseudo-descriptors, but the possibility of generating
an alignment check fault can be avoided by aligning pseudo-descriptors in this way.
The same alignment should be used when storing the IDTR register using the SIDT
instruction. When storing the LDTR or task register (using the SLTR or STR instruc-
tion, respectively), the pseudo-descriptor should be located at a doubleword address
(that is, address MOD 4 is equal to 0).
Vol. 3 3-21
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTED-MODE MEMORY MANAGEMENT
47
16 15
16 15
0
0
Limit
Limit
32-bit Base Address
79
64-bit Base Address
Figure 3-11. Pseudo-Descriptor Formats
3.5.2
Segment Descriptor Tables in IA-32e Mode
13
In IA-32e mode, a segment descriptor table can contain up to 8192 (2 ) 8-byte
descriptors. An entry in the segment descriptor table can be 8 bytes. System descrip-
tors are expanded to 16 bytes (occupying the space of two entries).
GDTR and LDTR registers are expanded to hold 64-bit base address. The corre-
The following system descriptors expand to 16 bytes:
— Call gate descriptors (see Section 5.8.3.1, “IA-32e Mode Call Gates”)
— IDT gate descriptors (see Section 6.14.1, “64-Bit Mode IDT”)
— LDT and TSS descriptors (see Section 7.2.3, “TSS Descriptor in 64-bit
mode”).
3-22 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
CHAPTER 4
PAGING
Chapter 3 explains how segmentation converts logical addresses to linear addresses.
linear address to a physical address and determines, for each translation, what
type of caching used for such accesses (the address’s memory type).
fied and defined in Section 4.1. Section 4.2 gives an overview of the translation
mechanism that is used in all modes. Section 4.3, Section 4.4, and Section 4.5
discuss the three paging modes in detail.
discusses exceptions that may be generated by paging (page-fault exceptions).
Section 4.8 considers data which the processor writes in response to linear-address
accesses (accessed and dirty flags).
Section 4.9 describes how paging determines the memory types used for accesses to
linear addresses. Section 4.10 provides details of how a processor may cache infor-
mation about linear-address translation. Section 4.11 outlines interactions between
paging and certain VMX features. Section 4.12 gives an overview of how paging can
be used to implement virtual memory.
4.1
PAGING MODES AND CONTROL BITS
Paging behavior is controlled by the following control bits:
• The WP and PG flags in control register CR0 (bit 16 and bit 31, respectively).
respectively).
• The LME and NXE flags in the IA32_EFER MSR (bit 8 and bit 11, respectively).
doing so, software should ensure that control register CR3 contains the physical
translation (see Section 4.2) and that structure is initialized as desired. See
Table 4-3, Table 4-7, and Table 4-12 for the use of CR3 in the different paging
modes.
Section 4.1.1 describes how the values of CR0.PG, CR4.PAE, and IA32_EFER.LME
determine whether paging is in use and, if so, which of three paging modes is in use.
Section 4.1.2 explains how to manage these bits to establish or make changes in
Vol. 3 4-1
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
paging modes. Section 4.1.3 discusses how CR0.WP, CR4.PSE, CR4.PGE, and
IA32_EFER.NXE modify the operation of the different paging modes.
4.1.1
Three Paging Modes
If CR0.PG = 0, paging is not used. The logical processor treats all linear addresses as
Paging is enabled if CR0.PG = 1. Paging can be enabled only if protection is enabled
CR4.PAE and IA32_EFER.LME determine which paging mode is used:
• If CR0.PG = 1 and CR4.PAE = 0, 32-bit paging is used. 32-bit paging is detailed
in Section 4.1.3.
• If CR0.PG = 1, CR4.PAE = 1, and IA32_EFER.LME = 0, PAE paging is used. PAE
paging is detailed in Section 4.4. PAE paging uses CR0.WP, CR4.PGE, and
IA32_EFER.NXE as described in Section 4.1.3.
1
• If CR0.PG = 1, CR4.PAE = 1, and IA32_EFER.LME = 1, IA-32e paging is used.
IA-32e paging is detailed in Section 4.5. IA-32e paging uses CR0.WP, CR4.PGE,
and IA32_EFER.NXE as described in Section 4.1.3. IA-32e paging is available
only on processors that support the Intel 64 architecture.
The three paging modes differ with regard to the following details:
• Linear-address width. The size of the linear addresses that can be translated.
• Physical-address width. The size of the physical addresses produced by paging.
• Page size. The granularity at which linear addresses are translated. Linear
addresses on the same page are translated to corresponding physical addresses
on the same page.
• Support for execute-disable access rights. In some paging modes, software can
be prevented from fetching instructions from pages that are otherwise readable.
Table 4-1 illustrates the key differences between the three paging modes.
Because they are used only if IA32_EFER.LME = 0, 32-bit paging and PAE paging is
used only in legacy protected mode. Because legacy protected mode cannot produce
1. The LMA flag in the IA32_EFER MSR (bit 10) is a status bit that indicates whether the logical pro-
cessor is in IA-32e mode (and thus using IA-32e paging). The processor always sets
IA32_EFER.LMA to CR0.PG & IA32_EFER.LME. Software cannot directly modify IA32_EFER.LMA;
an execution of WRMSR to the IA32_EFER MSR ignores bit 10 of its source operand.
4-2 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
Table 4-1. Properties of Different Paging Modes
Linear-
Physical-
Supports
Execute-
Disable?
Paging
Mode
LME in
IA32_EFER
Page
Size(s)
CR0.PG CR4.PAE
Address Address
1
Width
Width
None
0
1
N/A
0
N/A
32
32
N/A
No
No
4-KByte
4-MByte
2
3
32-bit
0
32
32
48
Up to 40
Up to 52
Up to 52
4
4-KByte
2-MByte
PAE
1
1
1
1
0
2
4-KByte
5
IA-32e
NOTES:
1. The physical-address width is always bounded by MAXPHYADDR; see Section 4.1.4.
2. The processor ensures that IA32_EFER.LME must be 0 if CR0.PG = 1 and CR4.PAE = 0.
3. 32-bit paging supports physical-address widths of more than 32 bits only for 4-MByte pages and
only if the PSE-36 mechanism is supported; see Section 4.1.4 and Section 4.3.
4. 4-MByte pages are used with 32-bit paging only if CR4.PSE = 1; see Section 4.3.
5. Execute-disable access rights are applied only if IA32_EFER.NXE = 1; see Section 4.6.
linear addresses larger than 32 bits, 32-bit paging and PAE paging translate 32-bit
linear addresses.
Because it is used only if IA32_EFER.LME = 1, IA-32e paging is used only in IA-32e
mode. (In fact, it is the use of IA-32e paging that defines IA-32e mode.) IA-32e
mode has two sub-modes:
• Compatibility mode. This mode uses only 32-bit linear addresses. IA-32e paging
treats bits 47:32 of such an address as all 0.
• 64-bit mode. While this mode produces 64-bit linear addresses, the processor
1
ensures that bits 63:47 of such an address are identical. IA-32e paging does not
use bits 63:48 of such addresses.
4.1.2
Paging-Mode Enabling
If CR0.PG = 1, a logical processor is in one of three paging modes, depending on the
values of CR4.PAE and IA32_EFER.LME. Figure 4-1 illustrates how software can
1. Such an address is called canonical. Use of a non-canonical linear address in 64-bit mode pro-
duces a general-protection exception (#GP(0)); the processor does not attempt to translate non-
canonical linear addresses using IA-32e paging.
Vol. 3 4-3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
enable these modes and make transitions between them. The following items identify
certain limitations and other details:
#GP
#GP
Set LME
Set LME
No Paging
Set PG
Set PAE
32-bit Paging
PAE Paging
PG = 0
PAE = 0
LME = 0
PG = 1
PAE = 0
LME = 0
PG = 1
PAE = 1
LME = 0
Clear PG
Clear PAE
#GP
Clear PAE
Set PG
Clear PAE
Set PAE
Clear PG
No Paging
No Paging
IA-32e Paging
PG = 0
PAE = 0
LME = 1
PG = 0
PAE = 1
LME = 0
PG = 1
PAE = 1
LME = 1
Set PG
Set PG
Clear PG
Clear PAE
Set PAE
#GP
#GP
No Paging
PG = 0
PAE = 1
LME = 1
Figure 4-1. Enabling and Changing Paging Modes
• IA32_EFER.LME cannot be modified while paging is enabled (CR0.PG = 1).
Attempts to do so using WRMSR cause a general-protection exception (#GP(0)).
• Paging cannot be enabled (by setting CR0.PG to 1) while CR4.PAE = 0 and
IA32_EFER.LME = 1. Attempts to do so using MOV to CR0 cause a general-
protection exception (#GP(0)).
• CR4.PAE cannot be cleared while IA-32e paging is active (CR0.PG = 1 and
IA32_EFER.LME = 1). Attempts to do so using MOV to CR4 cause a general-
protection exception (#GP(0)).
4-4 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
• Software can always disable paging by clearing CR0.PG with MOV to CR0.
• Software can make transitions between 32-bit paging and PAE paging by
changing the value of CR4.PAE with MOV to CR4.
• Software cannot make transitions directly between IA-32e paging and either of
the other two paging modes. It must first disable paging (by clearing CR0.PG with
MOV to CR4 and WRMSR), and then re-enable paging (by setting CR0.PG with
MOV to CR0). As noted earlier, an attempt to clear either CR4.PAE or
IA32_EFER.LME cause a general-protection exception (#GP(0)).
• VMX transitions allow transitions between paging modes that are not possible
using MOV to CR or WRMSR. This is because VMX transitions can load CR0, CR4,
and IA32_EFER in one operation. See Section 4.11.1.
4.1.3
Paging-Mode Modifiers
Details of how each paging mode operates are determined by the following control
bits:
• The WP flag in CR0 (bit 16).
• The NXE flag in the IA32_EFER MSR (bit 11).
CR0.WP allows pages to be protected from supervisor-mode writes. If CR0.WP = 0,
software operating with CPL < 3 (supervisor mode) can write to linear addresses with
— user mode — cannot write to linear addresses with read-only access rights,
regardless of the value of CR0.WP.) Section 4.6 explains how access rights are deter-
mined.
use only 4-KByte pages; if CR4.PSE = 1, 32-bit paging can use both 4-KByte pages
and 4-MByte pages. See Section 4.3 for more information. (PAE paging and IA-32e
paging can use multiple page sizes regardless of the value of CR4.PSE.)
CR4.PGE enables global pages. If CR4.PGE = 0, no translations are shared across
spaces. See Section 4.10.1.4 for more information.
IA32_EFER.NXE enables execute-disable access rights for PAE paging and IA-32e
paging. If IA32_EFER.NXE = 0, software may fetch instructions from any linear
address that paging allows the software to read; if IA32_EFER.NXE = 1, instructions
fetches can be prevented from specified linear addresses (even if data reads from the
addresses are allowed). Section 4.6 explains how access rights are determined. (32-
bit paging always allows software to fetch instructions from any linear address that
may be read; IA32_EFER.NXE has no effect with 32-bit paging. Software that wants
to limit instruction fetches from readable pages must use either PAE paging or IA-32e
paging.)
Vol. 3 4-5
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
4.1.4
Enumeration of Paging Features by CPUID
Software can discover support for different paging features using the CPUID instruc-
tion:
• PSE: page-size extensions for 32-bit paging.
If CPUID.01H:EDX.PSE [bit 3] = 1, CR4.PSE may be set to 1, enabling support
for 4-MByte pages with 32-bit paging (see Section 4.3).
• PAE: physical-address extension.
If CPUID.01H:EDX.PAE [bit 6] = 1, CR4.PAE may be set to 1, enabling PAE
paging (this setting is also required for IA-32e paging).
• PGE: global-page support.
global-page feature (see Section 4.10.1.4).
• PAT: page-attribute table.
If CPUID.01H:EDX.PAT [bit 16] = 1, the 8-entry page-attribute table (PAT) is
supported. When the PAT is supported, three bits in certain paging-structure
entries select a memory type (used to determine type of caching used) from the
PAT (see Section 4.9).
• PSE-36: 36-Bit page size extension.
indicating that translations using 4-MByte pages with 32-bit paging may produce
physical addresses with more than 32 bits (see Section 4.3).
• NX: execute disable.
If CPUID.80000001H:EDX.NX [bit 20] = 1, IA32_EFER.NXE may be set to 1,
allowing PAE paging and IA-32e paging to disable execute access to selected
pages (see Section 4.6). (Processors that do not support CPUID function
80000001H do not allow IA32_EFER.NXE to be set to 1.)
• LM: IA-32e mode support.
If CPUID.80000001H:EDX.LM [bit 29] = 1, IA32_EFER.LME may be set to 1,
enabling IA-32e paging. (Processors that do not support CPUID function
80000001H do not allow IA32_EFER.LME to be set to 1.)
• CPUID.80000008H:EAX[7:0] reports the physical-address width supported by
the processor. (For processors that do not support CPUID function 80000008H,
the width is generally 36 if CPUID.01H:EDX.PAE [bit 6] = 1 and 32 otherwise.)
This width is referred to as MAXPHYADDR. MAXPHYADDR is at most 52.
• CPUID.80000008H:EAX[15:8] reports the linear-address width supported by the
processor. Generally, this value is 48 if CPUID.80000001H:EDX.LM [bit 29] = 1
and 32 otherwise. (Processors that do not support CPUID function 80000008H,
support a linear-address width of 32.)
4-6 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
4.2
HIERARCHICAL PAGING STRUCTURES: AN OVERVIEW
All three paging modes translate linear addresses use hierarchical paging struc-
tures. This section provides an overview of their operation. Section 4.3, Section 4.4,
and Section 4.5 provide details for the three paging modes.
Every paging structure is 4096 Bytes in size and comprises a number of individual
entries. With 32-bit paging, each entry is 32 bits (4 bytes); there are thus 1024
entries in each structure. With PAE paging and IA-32e paging, each entry is 64 bits
(8 bytes); there are thus 512 entries in each structure. (PAE paging includes one
exception, a paging structure that is 32 bytes in size, containing 4 64-bit entries.)
The processor uses the upper portion of a linear address to identify a series of
paging-structure entries. The last of these entries identifies the physical address of
the region to which the linear address translates (called the page frame). The lower
portion of the linear address (called the page offset) identifies the specific address
within that region to which the linear address translates.
Each paging-structure entry contains a physical address, which is either the address
of another paging structure or the address of a page frame. In the first case, the
entry is said to reference the other paging structure; in the latter, the entry is said
to map a page.
The first paging structure used for any translation is located at the physical address
in CR3. A linear address is translated using the following iterative procedure. A
portion of the linear address (initially the uppermost bits) select an entry in a paging
structure (initially the one located using CR3). If that entry references another
paging structure, the process continues with that paging structure and with the
portion of the linear address immediately below that just used. If instead the entry
maps a page, the process completes: the physical address in the entry is that of the
page frame and the remaining lower portion of the linear address is the page offset.
The following items give an example for each of the three paging modes (each
example locates a 4-KByte page frame):
10
• With 32-bit paging, each paging structure comprises 1024 = 2 entries. For this
reason, the translation process uses 10 bits at a time from a 32-bit linear
address. Bits 31:22 identify the first paging-structure entry and bits 21:12
identify a second. The latter identifies the page frame. Bits 11:0 of the linear
address are the page offset within the 4-KByte page frame. (See Figure 4-2 for
an illustration.)
2
• With PAE paging, the first paging structure comprises only 4 = 2 entries.
Translation thus begins by using bits 31:30 from a 32-bit linear address to
identify the first paging-structure entry. Other paging structures comprise
9
512 =2 entries, so the process continues by using 9 bits at a time. Bits 29:21
identify a second paging-structure entry and bits 20:12 identify a third. This last
identifies the page frame. (See Figure 4-5 for an illustration.)
9
• With IA-32e paging, each paging structure comprises 512 = 2 entries and
translation uses 9 bits at a time from a 48-bit linear address. Bits 47:39 identify
the first paging-structure entry, bits 38:30 identify a second, bits 29:21 a third,
Vol. 3 4-7
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
and bits 20:12 identify a fourth. Again, the last identifies the page frame. (See
Figure 4-8 for an illustration.)
The translation process in each of the examples above completes by identifying a
page frame. However, the paging structures may be configured so that translation
terminates before doing so. This occurs if process encounters a paging-structure
entry that is marked “not present” (because its P flag — bit 0 — is clear) or in which
a reserved bit is set. In this case, there is no translation for the linear address; an
access to that address causes a page-fault exception (see Section 4.7).
In the examples above, a paging-structure entry maps a page with 4-KByte page
frame when only 12 bits remain in the linear address; entries identified earlier always
reference other paging structures. That may not apply in other cases. The following
items identify when an entry maps a page and when it references another paging
structure:
• If more than 12 bits remain in the linear address, bit 7 (PS — page size) of the
current paging-structure entry is consulted. If the bit is 0, the entry references
another paging structure; if the bit is 1, the entry maps a page.
• If only 12 bits remain in the linear address, the current paging-structure entry
always maps a page (bit 7 is used for other purposes).
If a paging-structure entry maps a page when more than 12 bits remain in the linear
paging uses the upper 10 bits of a linear address to locate the first paging-structure
22
entry; 22 bits remain. If that entry maps a page, the page frame is 2 Bytes = 4
MBytes. 32-bit paging supports 4-MByte pages if CR4.PSE = 1. PAE paging and
IA-32e paging support 2-MByte pages (regardless of the value of CR4.PSE).
Paging structures are given different names based their uses in the translation
process. Table 4-2 gives the names of the different paging structures. It also
provides, for each structure, the source of the physical address used to locate it (CR3
or a different paging-structure entry); the bits in the linear address used to select an
entry from the structure; and details of about whether and how such an entry can
map a page.
4.3
32-BIT PAGING
A logical processor uses 32-bit paging if CR0.PG = 1 and CR4.PAE = 0. 32-bit paging
1
translates 32-bit linear addresses to 40-bit physical addresses. Although 40 bits
1. Bits in the range 39:32 are 0 in any physical address used by 32-bit paging except those used to
map 4-MByte pages. If the processor does not support the PSE-36 mechanism, this is true also
for physical addresses used to map 4-MByte pages. If the processor does support the PSE-36
mechanism and MAXPHYADDR < 40, bits in the range 39:MAXPHYADDR are 0 in any physical
address used to map a 4-MByte page. (The corresponding bits are reserved in PDEs.) See Section
4.1.4 for how to determine MAXPHYADDR and whether the PSE-36 mechanism is supported.
4-8 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
Table 4-2. Paging Structures in the Different Paging Modes
Physical
Paging Mode Address of
Structure
Bits
Paging
Structure
Entry
Name
Selecting Page Mapping
Entry
32-bit, PAE
N/A
PML4 table
PML4E
PDPTE
IA-32e
32-bit
CR3
47:39
N/A (PS must be 0)
N/A
Page-directory-
pointer table
PAE
CR3
31:30
38:30
31:22
29:21
21:12
20:12
N/A (PS must be 0)
IA-32e
32-bit
PML4E
CR3
1
4-MByte page if PS=1
2-MByte page if PS=1
4-KByte page
Page directory
PDE
PTE
PAE, IA-32e
32-bit
PDPTE
Page table
PDE
PAE, IA-32e
4-KByte page
NOTES:
1. 32-bit paging ignores the PS flag in a PDE (and uses the entry to reference a page table) unless
CR4.PSE = 1. Not all processors allow CR4.PSE to be 1; see Section 4.1.4 for how to determine
whether 4-MByte pages are supported with 32-bit paging.
corresponds to 1 TByte, linear addresses are limited to 32 bits; at most 4 GBytes of
linear-address space may be accessed at any given time.
32-bit paging uses a hierarchy of paging structures to produce a translation for a
linear address. CR3 is used to locate the first paging-structure, the page directory.
Table 4-3 illustrates how CR3 is used with 32-bit paging.
Table 4-3. Use of CR3 with 32-Bit Paging
Bit
Position(s)
Contents
2:0
Ignored
3 (PWT)
Page-level write-through; indirectly determines the memory type used to access
the page directory during linear-address translation (see Section 4.9)
4 (PCD)
11:5
Page-level cache disable; indirectly determines the memory type used to access
the page directory during linear-address translation (see Section 4.9)
Ignored
Vol. 3 4-9
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
Table 4-3. Use of CR3 with 32-Bit Paging (Contd.)
Contents
Bit
Position(s)
31:12
63:32
Physical address of the 4-KByte aligned page directory used for linear-address
translation
Ignored (these bits exist only on processors supporting the Intel-64 architecture)
32-bit paging may map linear addresses to either 4-KByte pages or 4-MByte pages.
Figure 4-2 illustrates the translation process when it uses a 4-KByte page; Figure 4-3
covers the case of a 4-MByte page. The following items describe the 32-bit paging
process in more detail as well has how the page size is determined:
Linear Address
31
22 21
12 11
0
Table
Offset
12
Directory
4-KByte Page
Page Table
Physical Address
10
10
Page Directory
PTE
20
PDE with PS=0
CR3
20
32
Figure 4-2. Linear-Address Translation to a 4-KByte Page using 32-Bit Paging
• A 4-KByte naturally aligned page directory is located at the physical address
specified in bits 31:12 of CR3 (see Table 4-3). A page directory comprises 1024
32-bit entries (PDEs). A PDE is selected using the physical address defined as
follows:
— Bits 39:32 are all 0.
— Bits 31:12 are from CR3.
— Bits 11:2 are bits 31:22 of the linear address.
— Bits 1:0 are 0.
4-10 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
Linear Address
22 21
Directory
31
0
Offset
22
4-MByte Page
Page Directory
10
Physical Address
PDE with PS=1
CR3
18
32
Figure 4-3. Linear-Address Translation to a 4-MByte Page using 32-Bit Paging
Because a PDE is identified using bits 31:22 of the linear address, it controls access
to a 4-Mbyte region of the linear-address space. Use of the PDE depends on CR.PSE
and the PDE’s PS flag (bit 7):
• If CR4.PSE = 1 and the PDE’s PS flag is 1, the PDE maps a 4-MByte page (see
Table 4-4). The final physical address is computed as follows:
— Bits 39:32 are bits 20:13 of the PDE.
1
— Bits 31:22 are bits 31:22 of the PDE.
— Bits 21:0 are from the original linear address.
• If CR4.PSE = 0 or the PDE’s PS flag is 0, a 4-KByte naturally aligned page table is
located at the physical address specified in bits 31:12 of the PDE (see Table 4-5).
A page table comprises 1024 32-bit entries (PTEs). A PTE is selected using the
physical address defined as follows:
— Bits 39:32 are all 0.
— Bits 31:12 are from the PDE.
— Bits 11:2 are bits 21:12 of the linear address.
— Bits 1:0 are 0.
• Because a PTE is identified using bits 31:12 of the linear address, every PTE
maps a 4-KByte page (see Table 4-6). The final physical address is computed as
follows:
— Bits 39:32 are all 0.
1. The upper bits in the final physical address do not all come from corresponding positions in the
PDE; the physical-address bits in the PDE are not all contiguous.
Vol. 3 4-11
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
Table 4-4. Format of a 32-Bit Page-Directory Entry that Maps a 4-MByte Page
Bit
Position(s)
Contents
0 (P)
Present; must be 1 to map a 4-MByte page
1 (R/W)
Read/write; if 0, writes may not be allowed to the 4-MByte page referenced by
this entry (depends on CPL and CR0.WP; see Section 4.6)
2 (U/S)
3 (PWT)
4 (PCD)
5 (A)
referenced by this entry (see Section 4.6)
the 4-MByte page referenced by this entry (see Section 4.9)
the 4-MByte page referenced by this entry (see Section 4.9)
Accessed; indicates whether software has accessed the 4-MByte page referenced
by this entry (see Section 4.8)
6 (D)
Dirty; indicates whether software has written to the 4-MByte page referenced by
this entry (see Section 4.8)
7 (PS)
8 (G)
Page size; must be 1 (otherwise, this entry references a page table; see Table 4-5)
Global; if CR4.PGE = 1, determines whether the translation is global (see Section
4.10); ignored otherwise
11:9
Ignored
12 (PAT)
If the PAT is supported, indirectly determines the memory type used to access the
4-MByte page referenced by this entry (see Section 4.9); otherwise, reserved
(must be 0)
1
2
M–20:13
21:M–19
31:22
Reserved (must be 0)
NOTES:
1. See Section 4.1.4 for how to determine whether the PAT is supported.
2. If the PSE-36 mechanism is not supported, M is 32, and this row does not apply. If the PSE-36
mechanism is supported, M is the minimum of 40 and MAXPHYADDR (this row does not apply if
MAXPHYADDR = 32). See Section 4.1.4 for how to determine MAXPHYADDR and whether the
PSE-36 mechanism is supported.
— Bits 31:12 are from the PTE.
4-12 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
Table 4-5. Format of a 32-Bit Page-Directory Entry that References a Page Table
Bit
Position(s)
Contents
0 (P)
Present; must be 1 to reference a page table
1 (R/W)
this entry (depends on CPL and CR0.WP; see Section 4.6)
2 (U/S)
3 (PWT)
4 (PCD)
5 (A)
controlled by this entry (see Section 4.6)
the page table referenced by this entry (see Section 4.9)
Page-level cache disable; indirectly determines the memory type used to access
the page table referenced by this entry (see Section 4.9)
Accessed; indicates whether this entry has been used for linear-address
translation (see Section 4.8)
6
Ignored
7 (PS)
If CR4.PSE = 1, must be 0 (otherwise, this entry maps a 4-MByte page; see
Table 4-4); otherwise, ignored
11:8
Ignored
31:12
Physical address of 4-KByte aligned page table referenced by this entry
— Bits 11:0 are from the original linear address.
If a paging-structure entry’s P flag (bit 0) is 0 or if the entry sets any reserved bit, the
entry is used neither to reference another paging-structure entry nor to map a page.
A reference using a linear address whose translation would use such a paging-struc-
ture entry causes a page-fault exception (see Section 4.7).
With 32-bit paging, there are reserved bits only if CR4.PSE = 1:
• If the P flag and the PS flag (bit 7) of a PDE are both 1, the bits reserved depend
1
on MAXPHYADDR whether the PSE-36 mechanism is supported:
— If the PSE-36 mechanism is not supported, bits 21:13 are reserved.
is the minimum of 40 and MAXPHYADDR.
2
• If the PAT is not supported:
1. See Section 1.1.5 for how to determine MAXPHYADDR and whether the PSE-36 mechanism is
supported.
2. See Section 4.1.4 for how to determine whether the PAT is supported.
Vol. 3 4-13
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
Table 4-6. Format of a 32-Bit Page-Table Entry that Maps a 4-KByte Page
Bit
Position(s)
Contents
0 (P)
Present; must be 1 to map a 4-KByte page
1 (R/W)
entry (depends on CPL and CR0.WP; see Section 4.6)
2 (U/S)
3 (PWT)
4 (PCD)
5 (A)
referenced by this entry (see Section 4.6)
the 4-KByte page referenced by this entry (see Section 4.9)
the 4-KByte page referenced by this entry (see Section 4.9)
by this entry (see Section 4.8)
6 (D)
Dirty; indicates whether software has written to the 4-KByte page referenced by
this entry (see Section 4.8)
7 (PAT)
If the PAT is supported, indirectly determines the memory type used to access the
4-KByte page referenced by this entry (see Section 4.9); otherwise, reserved
(must be 0)
1
8 (G)
Global; if CR4.PGE = 1, determines whether the translation is global (see Section
11:9
31:12
NOTES:
Ignored
Physical address of the 4-KByte page referenced by this entry
1. See Section 4.1.4 for how to determine whether the PAT is supported.
— If the P flag and the PS flag of a PDE are both 1, bit 12 is reserved.
(If CR4.PSE = 0, no bits are reserved with 32-bit paging.)
A reference using a linear address that is successfully translated to a physical
address is performed only if allowed by the access rights of the translation; see
Section 4.6.
Figure 4-4 gives a summary of the formats of CR3 and the paging-structure entries
with 32-bit paging. For the paging structure entries, it identifies separately the
format of entries that map pages, those that reference other paging structures, and
4-14 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
those that do neither because they are “not present”; bit 0 (P) and bit 7 (PS) are
highlighted because they determine how such an entry is used.
31302928272625242322212019181716151413121110 9
8
7
6
5
4
3
2
1
0
P P
1
Address of page directory
Ignored
C W Ignored CR3
D T
Bits 39:32 P
P P U R
PDE:
Bits 31:22 of address
of 2MB page frame
Reserved
of
A Ignored G 1 D A C W / / 1 4MB
(must be 0)
2
address
T
D T S W
page
I
P P U R
PDE:
Address of page table
Ignored 0 g A C W / / 1 page
n
D T S W
table
PDE:
not
Ignored
0
present
P
P P U R
PTE:
Address of 4KB page frame
Ignored G A D A C W / / 1 4KB
T
D T S W
page
PTE:
not
Ignored
0
present
Figure 4-4. Formats of CR3 and Paging-Structure Entries with 32-Bit Paging
NOTES:
1. CR3 has 64 bits on processors supporting the Intel-64 architecture. These bits are ignored with
32-bit paging.
2. This example illustrates a processor in which MAXPHYADDR is 36. If this value is larger or smaller,
the number of bits reserved in positions 20:13 of a PDE mapping a 4-MByte will change.
4.4
PAE PAGING
A logical processor uses PAE paging if CR0.PG = 1, CR4.PAE = 1, and
IA32_EFER.LME = 0. PAE paging translates 32-bit linear addresses to 52-bit physical
1
addresses. Although 52 bits corresponds to 4 PBytes, linear addresses are limited to
32 bits; at most 4 GBytes of linear-address space may be accessed at any given time.
which are loaded from an address in CR3. Linear address are translated using 4 hier-
archies of in-memory paging structures, each located using one of the PDPTE regis-
1. If MAXPHYADDR < 52, bits in the range 51:MAXPHYADDR will be 0 in any physical address used
by PAE paging. (The corresponding bits are reserved in the paging-structure entries.) See Section
4.1.4 for how to determine MAXPHYADDR.
Vol. 3 4-15
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
ters. (This is different from the other paging modes, in which there is one hierarchy
referenced by CR3.)
translation with PAE paging.
4.4.1
PDPTE Registers
When PAE paging is used, CR3 references the base of a 32-Byte page-directory-
pointer table. Table 4-7 illustrates how CR3 is used with PAE paging.
Table 4-7. Use of CR3 with PAE Paging
Bit
Position(s)
Contents
4:0
Ignored
31:5
Physical address of the 32-Byte aligned page-directory-pointer table used for
linear-address translation
63:32
Ignored (these bits exist only on processors supporting the Intel-64 architecture)
The page-directory-pointer-table comprises four (4) 64-bit entries called PDPTEs.
Each PDPTE controls access to a 1-GByte region of the linear-address space. Corre-
non-architectural PDPTE registers, called PDPTE0, PDPTE1, PDPTE2, and PDPTE3.
The logical processor loads these registers from the PDPTEs in memory as part of
certain executions the MOV to CR instruction:
• If an execution MOV to CR0 or MOV to CR4 causes the logical processor to
transition from either no paging or 32-bit paging to PAE paging (see Section
4.1.2), the PDPTEs are loaded from the address in CR3.
• If MOV to CR3 is executed while the logical processor is using PAE paging, the
PDPTEs are loaded from the address being loaded into CR3.
• If PAE paging is in use and a task switch changes the value of CR3, the PDPTEs
are loaded from the address in the new CR3 value.
• Certain VMX transitions load the PDPTE registers. See Section 4.11.1.
Unless the caches are disabled, the processor uses the WB memory type to load the
1
PDPTEs from memory.
1. Older IA-32 processors used the UC memory type when loading the PDPTEs. This behavior is
model-specific and not architectural.
4-16 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
Table 4-8 gives the format of a PDPTE. If any of the PDPTEs sets both the P flag
Table 4-8. Format of an PAE Page-Directory-Pointer-Table Entry (PDPTE)
Contents
Bit
Position(s)
0 (P)
Present; must be 1 to reference a page directory
Reserved (must be 0)
2:1
3 (PWT)
Page-level write-through; indirectly determines the memory type used to access
the page directory referenced by this entry (see Section 4.9)
4 (PCD)
Page-level cache disable; indirectly determines the memory type used to access
the page directory referenced by this entry (see Section 4.9)
8:5
Reserved (must be 0)
Ignored
11:9
1
M–1:12
63:M
Physical address of 4-KByte aligned page directory referenced by this entry
Reserved (must be 0)
NOTES:
1. M is an abbreviation for MAXPHYADDR, which is at most 52; see Section 4.1.4.
(bit 0) and any reserved bit, the MOV to CR instruction causes a general-protection
1
exception (#GP(0)) and the PDPTEs are not loaded. As show in Table 4-8, bits 2:1,
4.4.2
Linear-Address Translation with PAE Paging
PAE paging may map linear addresses to either 4-KByte pages or 2-MByte pages.
Figure 4-5 illustrates the translation process when it produces a 4-KByte page;
Figure 4-6 covers the case of a 2-MByte page. The following items describe the PAE
paging process in more detail as well has how the page size is determined:
• Bits 31:30 of the linear address select a PDPTE register (see Section 4.4.1); this
2
is PDPTEi, where i is the value of bits 31:30. Because a PDPTE register is
identified using bits 31:30 of the linear address, it controls access to a 1-GByte
region of the linear-address space. If the P flag (bit 0) of PDPTEi is 0, the
1. On some processors, reserved bits are checked even in PDPTEs in which the P flag (bit 0) is 0.
2. With PAE paging, the processor does not use CR3 when translating a linear address (as it does
the other paging modes). It does not access the PDPTEs in the page-directory-pointer table dur-
ing linear-address translation.
Vol. 3 4-17
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
Linear Address
21 20 12 11
31 30 29
Directory
0
Directory Pointer
Table
Offset
12
4-KByte Page
Page Table
Physical Address
9
Page Directory
PDE with PS=0
PTE
9
40
40
2
PDPTE Registers
PDPTE value
40
Figure 4-5. Linear-Address Translation to a 4-KByte Page using PAE Paging
Linear Address
31 30 29
21 20
0
Directory
Pointer
Offset
Directory
2-MByte Page
21
9
Page Directory
Physical Address
PDPTE Registers
2
PDE with PS=1
31
PDPTE value
40
Figure 4-6. Linear-Address Translation to a 2-MByte Page using PAE Paging
processor ignores bits 63:1, and there is no mapping for the 1-GByte region
controlled by PDPTEi. A reference using a linear address in this region causes a
page-fault exception (see Section 4.7).
• If the P flag of PDPTEi is 1, 4-KByte naturally aligned page directory is located at
the physical address specified in bits 51:12 of PDPTEi (see Table 4-8 in Section
4-18 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
4.4.1) A page directory comprises 512 64-bit entries (PDEs). A PDE is selected
using the physical address defined as follows:
— Bits 51:12 are from PDPTEi.
— Bits 11:3 are bits 29:21 of the linear address.
— Bits 2:0 are 0.
Because a PDE is identified using bits 31:21 of the linear address, it controls access
to a 2-Mbyte region of the linear-address space. Use of the PDE depends on its PS
flag (bit 7):
physical address is computed as follows:
— Bits 51:21 are from the PDE.
— Bits 20:0 are from the original linear address.
• If the PDE’s PS flag is 0, a 4-KByte naturally aligned page table is located at the
physical address specified in bits 51:12 of the PDE (see Table 4-10). A page
directory comprises 512 64-bit entries (PTEs). A PTE is selected using the
physical address defined as follows:
— Bits 51:12 are from the PDE.
— Bits 11:3 are bits 20:12 of the linear address.
— Bits 2:0 are 0.
• Because a PTE is identified using bits 31:12 of the linear address, every PTE maps
a 4-KByte page (see Table 4-11). The final physical address is computed as
follows:
— Bits 51:12 are from the PTE.
— Bits 11:0 are from the original linear address.
If the P flag (bit 0) of a PDE or a PTE is 0 or if a PDE or a PTE sets any reserved bit,
the entry is used neither to reference another paging-structure entry nor to map a
page. A reference using a linear address whose translation would use such a paging-
structure entry causes a page-fault exception (see Section 4.7).
The following bits are reserved with PAE paging:
• If the P flag (bit 0) of a PDE or a PTE is 1, bits 62:MAXPHYADDR are reserved.
• If the P flag and the PS flag (bit 7) of a PDE are both 1, bits 20:13 are reserved.
• If IA32_EFER.NXE = 0 and the P flag of a PDE or a PTE is 1, the XD flag (bit 63)
is reserved.
1
• If the PAT is not supported:
— If the P flag of a PTE is 1, bit 7 is reserved.
— If the P flag and the PS flag of a PDE are both 1, bit 12 is reserved.
1. See Section 4.1.4 for how to determine whether the PAT is supported.
Vol. 3 4-19
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
Table 4-9. Format of a PAE Page-Directory Entry that Maps a 2-MByte Page
Bit
Position(s)
Contents
0 (P)
Present; must be 1 to map a 2-MByte page
1 (R/W)
Read/write; if 0, writes may not be allowed to the 2-MByte page referenced by
this entry (depends on CPL and CR0.WP; see Section 4.6)
2 (U/S)
3 (PWT)
4 (PCD)
5 (A)
referenced by this entry (see Section 4.6)
the 2-MByte page referenced by this entry (see Section 4.9)
the 2-MByte page referenced by this entry (see Section 4.9)
by this entry (see Section 4.8)
6 (D)
this entry (see Section 4.8)
7 (PS)
8 (G)
Page size; must be 1 (otherwise, this entry references a page table; see
Table 4-10)
Global; if CR4.PGE = 1, determines whether the translation is global (see Section
4.10); ignored otherwise
11:9
Ignored
12 (PAT)
If the PAT is supported, indirectly determines the memory type used to access the
2-MByte page referenced by this entry (see Section 4.9); otherwise, reserved
(must be 0)
1
20:13
M–1:21
62:M
Reserved (must be 0)
Physical address of the 2-MByte page referenced by this entry
Reserved (must be 0)
63 (XD)
If IA32_EFER.NXE = 1, execute-disable (if 1, instruction fetches are not allowed
from the 2-MByte page controlled by this entry; see Section 4.6); otherwise,
reserved (must be 0)
NOTES:
1. See Section 4.1.4 for how to determine whether the PAT is supported.
4-20 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
Table 4-10. Format of a PAE Page-Directory Entry that References a Page Table
Bit
Position(s)
Contents
0 (P)
Present; must be 1 to reference a page table
1 (R/W)
this entry (depends on CPL and CR0.WP; see Section 4.6)
2 (U/S)
3 (PWT)
4 (PCD)
5 (A)
controlled by this entry (see Section 4.6)
the page table referenced by this entry (see Section 4.9)
Page-level cache disable; indirectly determines the memory type used to access
the page table referenced by this entry (see Section 4.9)
Accessed; indicates whether this entry has been used for linear-address
translation (see Section 4.8)
6
Ignored
7 (PS)
11:8
Page size; must be 0 (otherwise, this entry maps a 2-MByte page; see Table 4-9)
Ignored
M–1:12
62:M
63 (XD)
Physical address of 4-KByte aligned page table referenced by this entry
Reserved (must be 0)
from the 2-MByte region controlled by this entry; see Section 4.6); otherwise,
reserved (must be 0)
A reference using a linear address that is successfully translated to a physical
address is performed only if allowed by the access rights of the translation; see
Section 4.6.
Figure 4-7 gives a summary of the formats of CR3 and the paging-structure entries
with PAE paging. For the paging structure entries, it identifies separately the format
of entries that map pages, those that reference other paging structures, and those
Vol. 3 4-21
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
Table 4-11. Format of a PAE Page-Table Entry that Maps a 4-KByte Page
Bit
Position(s)
Contents
0 (P)
Present; must be 1 to map a 4-KByte page
1 (R/W)
entry (depends on CPL and CR0.WP; see Section 4.6)
2 (U/S)
3 (PWT)
4 (PCD)
5 (A)
referenced by this entry (see Section 4.6)
the 4-KByte page referenced by this entry (see Section 4.9)
the 4-KByte page referenced by this entry (see Section 4.9)
by this entry (see Section 4.8)
6 (D)
Dirty; indicates whether software has written to the 4-KByte page referenced by
this entry (see Section 4.8)
7 (PAT)
If the PAT is supported, indirectly determines the memory type used to access the
4-KByte page referenced by this entry (see Section 4.9); otherwise, reserved
(must be 0)
1
8 (G)
Global; if CR4.PGE = 1, determines whether the translation is global (see Section
4.10); ignored otherwise
11:9
Ignored
M–1:12
62:M
Physical address of the 4-KByte page referenced by this entry
Reserved (must be 0)
63 (XD)
If IA32_EFER.NXE = 1, execute-disable (if 1, instruction fetches are not allowed
from the 4-KByte page controlled by this entry; see Section 4.6); otherwise,
reserved (must be 0)
NOTES:
1. See Section 4.1.4 for how to determine whether the PAT is supported.
4-22 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
that do neither because they are “not present”; bit 0 (P) and bit 7 (PS) are high-
lighted because they determine how a paging-structure entry is used.
6 6 6 6 5 5 5 5 5 5 5 5 5
3 2 1 0 9 8 7 6 5 4 3 2 1
3 3 3 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1
2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0
M1 M-1
Ignored2
Ignored
CR3
Address of page-directory-pointer table
P P
Rs
PDPTE:
present
Reserved3
Address of page directory
Ign. Rsvd. CW
D T
1
0
vd
PDTPE:
not
present
Ignored
P
P P U R
PDE:
X
D
Address of
Ignored
Ignored
Rsvd.
Reserved A Ign. G 1 D A CW / / 1 2MB
2MB page frame
T
D T SW
page
I
P P U R
PDE:
X
D
Rsvd.
Address of page table
Ign. 0 g A CW / / 1 page
n
D T SW
table
PDE:
not
present
Ignored
0
P
P P U R
PTE:
X
D
Ignored
Rsvd.
Address of 4KB page frame
Ignored
Ign. G A D A CW / / 1 4KB
T
D T SW
page
PTE:
not
0
present
Figure 4-7. Formats of CR3 and Paging-Structure Entries with PAE Paging
NOTES:
1. M is an abbreviation for MAXPHYADDR.
2. CR3 has 64 bits only on processors supporting the Intel-64 architecture. These bits are ignored with
PAE paging.
3. Reserved fields must be 0.
4.5
IA-32E PAGING
A logical processor uses IA-32e paging if CR0.PG = 1, CR4.PAE = 1, and
IA32_EFER.LME = 1. With IA-32e paging, linear address are translated using a hier-
archy of in-memory paging structures located using the contents of CR3. IA-32e
1
paging translates 48-bit linear addresses to 52-bit physical addresses. Although 52
Vol. 3 4-23
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
bits corresponds to 4 PBytes, linear addresses are limited to 48 bits; at most 256
TBytes of linear-address space may be accessed at any given time.
IA-32e paging uses a hierarchy of paging structures to produce a translation for a
linear address. CR3 is used to locate the first paging-structure, the PML4 table.
Table 4-12 illustrates how CR3 is used with IA-32e paging.
Table 4-12. Use of CR3 with IA-32e Paging
Bit
Position(s)
Contents
2:0
Ignored
3 (PWT)
Page-level write-through; indirectly determines the memory type used to access
the PML4 table during linear-address translation (see Section 4.9)
4 (PCD)
Page-level cache disable; indirectly determines the memory type used to access
the PML4 table during linear-address translation (see Section 4.9)
11:5
Ignored
M–1:12
Physical address of the 4-KByte aligned PML4 table used for linear-address
translation
1
63:M
Reserved (must be 0)
NOTES:
1. M is an abbreviation for MAXPHYADDR, which is at most 52; see Section 4.1.4.
IA-32e paging may map linear addresses to either 4-KByte pages or 2-MByte pages.
Figure 4-8 illustrates the translation process when it produces a 4-KByte page;
Figure 4-9 covers the case of a 2-MByte page. The following items describe the
IA-32e paging process in more detail as well has how the page size is determined:
• A 4-KByte naturally aligned PML4 table is located at the physical address
specified in bits 51:12 of CR3 (see Table 4-12). A PML4 table comprises 512 64-
bit entries (PML4Es). A PML4E is selected using the physical address defined as
follows:
— Bits 51:12 are from CR3.
— Bits 11:3 are bits 47:39 of the linear address.
Because a PML4E is identified using bits 47:39 of the linear address, it controls
access to a 512-GByte region of the linear-address space.
1. If MAXPHYADDR < 52, bits in the range 51:MAXPHYADDR will be 0 in any physical address used
by IA-32e paging. (The corresponding bits are reserved in the paging-structure entries.) See Sec-
tion 4.1.4 for how to determine MAXPHYADDR.
4-24 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
Linear Address
30 29
39 38
47
21 20
12 11
0
PML4
9
Directory
Table
9
Offset
Directory Ptr
9
4-KByte Page
Physical Addr
12
PTE
Page Table
40
Page-Directory-
Pointer Table
PDE with PS=0
40
Page-Directory
40
PDPTE
9
40
PML4E
CR3
40
Figure 4-8. Linear-Address Translation to a 4-KByte Page using IA-32e Paging
• A 4-KByte naturally aligned page-directory-pointer table is located at the
physical address specified in bits 51:12 of the PML4E (see Table 4-13). A page-
directory-pointer table comprises 512 64-bit entries (PDPTEs). A PDPTE is
selected using the physical address defined as follows:
— Bits 51:12 are from the PML4E.
— Bits 11:3 are bits 38:30 of the linear address.
— Bits 2:0 are all 0.
Because a PDPTE is identified using bits 47:30 of the linear address, it controls
access to a 1-GByte region of the linear-address space.
• A 4-KByte naturally aligned page directory is located at the physical address
specified in bits 51:12 of the PDPTE (see Table 4-14). A page directory comprises
512 64-bit entries (PDEs). A PDE is selected using the physical address defined as
follows:
— Bits 51:12 are from the PDPTE.
— Bits 11:3 are bits 29:21 of the linear address.
— Bits 2:0 are all 0.
Vol. 3 4-25
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
Linear Address
30 29
39 38
47
21 20
0
PML4
9
Directory
Offset
21
Directory Ptr
9
2-MByte Page
Physical Addr
Page-Directory-
Pointer Table
PDE with PS=1
31
Page-Directory
PDPTE
40
9
40
PML4E
CR3
40
Figure 4-9. Linear-Address Translation to a 2-MByte Page using IA-32e Paging
Because a PDE is identified using bits 47:21 of the linear address, it controls access
to a 2-MByte region of the linear-address space. Use of the PDE depends on its PS
flag (bit 7):
4-26 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
Table 4-13. Format of an IA-32e PML4 Entry (PML4E) that References a Page-
Directory-Pointer Table
Bit
Position(s)
Contents
0 (P)
Present; must be 1 to reference a page-directory-pointer table
1 (R/W)
Read/write; if 0, writes may not be allowed to the 512-GByte region controlled by
this entry (depends on CPL and CR0.WP; see Section 4.6)
2 (U/S)
3 (PWT)
4 (PCD)
5 (A)
User/supervisor; if 0, accesses with CPL=3 are not allowed to the 512-GByte
region controlled by this entry (see Section 4.6)
the page-directory-pointer table referenced by this entry (see Section 4.9)
Page-level cache disable; indirectly determines the memory type used to access
the page-directory-pointer table referenced by this entry (see Section 4.9)
Accessed; indicates whether this entry has been used for linear-address
translation (see Section 4.8)
6
Ignored
7 (PS)
11:8
M–1:12
Reserved (must be 0)
Ignored
Physical address of 4-KByte aligned page-directory-pointer table referenced by
this entry
51:M
Reserved (must be 0)
Ignored
62:52
63 (XD)
If IA32_EFER.NXE = 1, execute-disable (if 1, instruction fetches are not allowed
from the 512-GByte region controlled by this entry; see Section 4.6); otherwise,
reserved (must be 0)
Vol. 3 4-27
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
Table 4-14. Format of an IA-32e Page-Directory-Pointer-Table Entry (PDPTE) that
References a Page Directory
Bit
Position(s)
Contents
0 (P)
Present; must be 1 to reference a page directory
1 (R/W)
Read/write; if 0, writes may not be allowed to the 1-GByte region controlled by
this entry (depends on CPL and CR0.WP; see Section 4.6)
2 (U/S)
3 (PWT)
4 (PCD)
5 (A)
User/supervisor; if 0, accesses with CPL=3 are not allowed to the 1-GByte region
controlled by this entry (see Section 4.6)
the page directory referenced by this entry (see Section 4.9)
Page-level cache disable; indirectly determines the memory type used to access
the page directory referenced by this entry (see Section 4.9)
Accessed; indicates whether this entry has been used for linear-address
translation (see Section 4.8)
6
Ignored
7 (PS)
11:8
Reserved (must be 0)
Ignored
M–1:12
51:M
62:52
63 (XD)
Physical address of 4-KByte aligned page directory referenced by this entry
Reserved (must be 0)
Ignored
If IA32_EFER.NXE = 1, execute-disable (if 1, instruction fetches are not allowed
from the 1-GByte region controlled by this entry; see Section 4.6); otherwise,
reserved (must be 0)
• If the PDE’s PS flag is 1, the PDE maps a 2-MByte page (see Table 4-15). The final
physical address is computed as follows:
Table 4-15. Format of an IA-32e Page-Directory Entry that Maps a 2-MByte Page
Bit
Position(s)
Contents
0 (P)
Present; must be 1 to map a 2-MByte page
1 (R/W)
Read/write; if 0, writes may not be allowed to the 2-MByte page referenced by
this entry (depends on CPL and CR0.WP; see Section 4.6)
4-28 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
Table 4-15. Format of an IA-32e Page-Directory Entry that Maps a 2-MByte Page
Bit
Position(s)
Contents
2 (U/S)
3 (PWT)
4 (PCD)
5 (A)
referenced by this entry (see Section 4.6)
the 2-MByte page referenced by this entry (see Section 4.9)
the 2-MByte page referenced by this entry (see Section 4.9)
by this entry (see Section 4.8)
6 (D)
this entry (see Section 4.8)
7 (PS)
8 (G)
Page size; must be 1 (otherwise, this entry references a page table; see
Table 4-16)
Global; if CR4.PGE = 1, determines whether the translation is global (see Section
4.10); ignored otherwise
11:9
Ignored
12 (PAT)
Indirectly determines the memory type used to access the 2-MByte page
referenced by this entry (see Section 4.9)
1
20:13
M–1:21
51:M
Reserved (must be 0)
Physical address of the 2-MByte page referenced by this entry
Reserved (must be 0)
Ignored
62:52
63 (XD)
If IA32_EFER.NXE = 1, execute-disable (if 1, instruction fetches are not allowed
from the 2-MByte page controlled by this entry; see Section 4.6); otherwise,
reserved (must be 0)
NOTES:
1. The PAT is supported on all processors that support IA-32e paging.
— Bits 51:21 are from the PDE.
— Bits 20:0 are from the original linear address.
• If the PDE’s PS flag is 0, a 4-KByte naturally aligned page table is located at the
physical address specified in bits 51:12 of the PDE (see Table 4-16). A page table
Vol. 3 4-29
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
comprises 512 64-bit entries (PTEs). A PTE is selected using the physical address
defined as follows:
Table 4-16. Format of an IA-32e Page-Directory Entry that References a Page Table
Bit
Position(s)
Contents
0 (P)
Present; must be 1 to reference a page table
1 (R/W)
this entry (depends on CPL and CR0.WP; see Section 4.6)
2 (U/S)
3 (PWT)
4 (PCD)
5 (A)
controlled by this entry (see Section 4.6)
the page table referenced by this entry (see Section 4.9)
Page-level cache disable; indirectly determines the memory type used to access
the page table referenced by this entry (see Section 4.9)
Accessed; indicates whether this entry has been used for linear-address
translation (see Section 4.8)
6
Ignored
7 (PS)
11:8
Page size; must be 0 (otherwise, this entry maps a 2-MByte page; see Table 4-15)
Ignored
M–1:12
51:M
62:52
63 (XD)
Physical address of 4-KByte aligned page table referenced by this entry
Reserved (must be 0)
Ignored
If IA32_EFER.NXE = 1, execute-disable (if 1, instruction fetches are not allowed
from the 2-MByte region controlled by this entry; see Section 4.6); otherwise,
reserved (must be 0)
— Bits 51:12 are from the PDE.
— Bits 11:3 are bits 20:12 of the linear address.
— Bits 2:0 are all 0.
• Because a PTE is identified using bits 47:12 of the linear address, every PTE
maps a 4-KByte page (see Table 4-17). The final physical address is computed as
follows:
— Bits 51:12 are from the PTE.
4-30 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
Table 4-17. Format of an IA-32e Page-Table Entry that Maps a 4-KByte Page
Bit
Position(s)
Contents
0 (P)
Present; must be 1 to map a 4-KByte page
1 (R/W)
entry (depends on CPL and CR0.WP; see Section 4.6)
2 (U/S)
3 (PWT)
4 (PCD)
5 (A)
referenced by this entry (see Section 4.6)
the 4-KByte page referenced by this entry (see Section 4.9)
the 4-KByte page referenced by this entry (see Section 4.9)
by this entry (see Section 4.8)
6 (D)
this entry (see Section 4.8)
7 (PAT)
8 (G)
Indirectly determines the memory type used to access the 2-MByte page
referenced by this entry (see Section 4.9)
Global; if CR4.PGE = 1, determines whether the translation is global (see Section
4.10); ignored otherwise
11:9
Ignored
M–1:12
51:M
Physical address of the 4-KByte page referenced by this entry
Reserved (must be 0)
Ignored
62:52
63 (XD)
If IA32_EFER.NXE = 1, execute-disable (if 1, instruction fetches are not allowed
from the 4-KByte page controlled by this entry; see Section 4.6); otherwise,
reserved (must be 0)
— Bits 11:0 are from the original linear address.
If a paging-structure entry’s P flag (bit 0) is 0 or if the entry sets any reserved bit, the
entry is used neither to reference another paging-structure entry nor to map a page.
A reference using a linear address whose translation would use such a paging-struc-
ture entry causes a page-fault exception (see Section 4.7).
The following bits are reserved with IA-32e paging:
• If the P flag of a paging-structure entry is 1, bits 51:MAXPHYADDR are reserved.
Vol. 3 4-31
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
• If the P flag of a PML4E or a PDPTE is 1, the PS flag is reserved.
• If the P flag and the PS flag of a PDE are both 1, bits 20:13 are reserved.
• If IA32_EFER.NXE = 0 and the P flag of a paging-structure entry is 1, the XD flag
(bit 63) is reserved.
A reference using a linear address that is successfully translated to a physical
address is performed only if allowed by the access rights of the translation; see
Section 4.6.
Figure 4-10 gives a summary of the formats of CR3 and the IA-32e paging-structure
entries. For the paging structure entries, it identifies separately the format of entries
that map pages, those that reference other paging structures, and those that do
neither because they are “not present”; bit 0 (P) and bit 7 (PS) are highlighted
4.6
ACCESS RIGHTS
There is a translation for a linear address if the processes described in Section 4.3,
Section 4.4.2, and Section 4.5 (depending upon the paging mode) completes and
produces a physical address. The accesses permitted by a translation is determined
by the access rights specified by the paging-structure entries controlling the transla-
1
tion. The following items detail how paging determines access rights:
• For accesses in supervisor mode (CPL < 3):
— Data reads.
Data may be read from any linear address with a valid translation.
— Data writes.
•
If CR0.WP = 0, data may be written to any linear address with a valid
translation.
•
If CR0.WP = 1, data may be written to any linear address with a valid
translation for which the R/W flag (bit 1) is 1 in every paging-structure
entry controlling the translation.
— Instruction fetches.
•
For 32-bit paging or if IA32_EFER.NXE = 0, instructions may be fetched
from any linear address with a valid translation.
•
For PAE paging or IA-32e paging with IA32_EFER.NXE = 1, instructions
may be fetched from any linear address with a valid translation for which
the XD flag (bit 63) is 0 in every paging-structure entry controlling the
translation.
• For accesses in user mode (CPL = 3):
1. With PAE paging, the PDPTEs do not determine access rights.
4-32 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
CR3
6 6 6 6 5 5 5 5 5 5 5 5 5
3 2 1 0 9 8 7 6 5 4 3 2 1
3 3 3 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1
2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0
M1 M-1
P P
Reserved2
Address of PML4 table
Address of page-directory-pointer table
Ignored
Ignored CW Ign.
D T
R
I
P P U R
X
s
v
d
PML4E:
present
Ignored
D
Rsvd.
Ign.
Ign.
g A CW / / 1
n
D T SW
PML4E:
not
0
present
R
s
v
d
I
P P U R
X
PDPTE:
present
Ignored
D
Rsvd.
Address of page directory
g A CW / / 1
n
D T SW
PDTPE:
not
Ignored
0
present
P
P P U R
PDE:
X
Address of
Ignored
D
Rsvd.
Rsvd.
Reserved A Ign. G 1 D A CW / / 1 2MB
2MB page frame
T
D T SW
page
I
P P U R
PDE:
X
Ignored
D
Address of page table
Ign. 0 g A CW / / 1 page
n
D T SW
table
PDE:
not
present
Ignored
0
P
P P U R
PTE:
X
Ignored
D
Rsvd.
Address of 4KB page frame
Ignored
Ign. G A D A CW / / 1 4KB
T
D T SW
page
PTE:
not
present
0
Figure 4-10. Formats of CR3 and Paging-Structure Entries with IA-32e Paging
NOTES:
1. M is an abbreviation for MAXPHYADDR.
2. Reserved fields must be 0.
— Data reads.
Data may be read from any linear address with a valid translation for which
the U/S flag (bit 2) is 1 in every paging-structure entry controlling the trans-
lation.
— Data writes.
Data may be written to any linear address with a valid translation for which
Vol. 3 4-33
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
both the R/W flag and the U/S flag are 1 in every paging-structure entry
controlling the translation.
— Instruction fetches.
•
For 32-bit paging or if IA32_EFER.NXE = 0, instructions may be fetched
from any linear address with a valid translation for which the U/S flag is 1
•
For PAE paging or IA-32e paging with IA32_EFER.NXE = 1, instructions
may be fetched from any linear address with a valid translation for which
the U/S flag is 1 and the XD flag is 0 in every paging-structure entry
controlling the translation.
paging-structure caches (see Section 4.10). These structures may include informa-
tion about access rights. The processor may enforce access rights based on the TLBs
and paging-structure caches instead of on the paging structures in memory.
This fact implies that, if software modifies a paging-structure entry to change access
rights, the processor might not use that change for a subsequent access to an
affected linear address (see Section 4.10.3.3). See Section 4.10.3.2 for how soft-
ware can ensure that the processor uses the modified access rights.
4.7
Accesses using linear addresses may cause page-fault exceptions (#PF; exception
14). An access to a linear address may cause page-fault exception for either of two
As noted in Section 4.3, Section 4.4.2, and Section 4.5, there is no valid translation
for a linear address if the translation process for that address would use a paging-
structure entry in which the P flag (bit 0) is 0 or one that sets a reserved bit. If there
is a valid translation for a linear address, its access rights are determined as specified
in Section 4.6.
Figure 4-11 illustrates the error code that the processor provides on delivery of a
page-fault exception. The following items explain how the bits in the error code
describe the nature of the page-fault exception:
• P flag (bit 0).
This flag is 0 if there is no valid translation for the linear address because the P
flag was 0 in one of the paging-structure entries used to translate that address.
• W/R (bit 1).
If the access causing the page-fault exception was a write, this flag is 1;
otherwise, it is 0. This flag describes the access causing the page-fault exception,
not the access rights specified by paging.
• U/S (bit 2).
If a supervisor-mode (CPL < 3) access caused the page-fault exception, this flag
4-34 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
31
3
2
1
0
4
Reserved
0 The fault was caused by a non-present page.
1 The fault was caused by a page-level protection violation.
P
W/R
U/S
0 The access causing the fault was a read.
1
The access causing the fault was a write.
The access causing the fault originated when the processor
was executing in supervisor mode.
The access causing the fault originated when the processor
was executing in user mode.
0
1
0
The fault was not caused by reserved bit violation.
RSVD
I/D
1 The fault was caused by a reserved bit set to 1 in some
paging-structure entry.
0
1
The fault was not caused by an instruction fetch.
The fault was caused by an instruction fetch.
Figure 4-11. Page-Fault Error Code
is 1; it is 0 if a user-mode (CPL = 3) access did so. This flag describes the access
causing the page-fault exception, not the access rights specified by paging.
• RSVD flag (bit 3).
This flag is 1 if there is no valid translation for the linear address because a
reserved bit was set in one of the paging-structure entries used to translate that
address. (Because reserved bits are not checked in a paging-structure entry
whose P flag is 0, bit 3 of the error code can be set only if bit 0 is also set.)
Bits reserved in the paging-structure entries are reserved for future functionality.
Software developers should be aware that such bits may be used in the future
and that a paging-structure entry that causes a page-fault exception on one
processor might not do so in the future.
• I/D flag (bit 4).
Use of this flag depends on the settings of CR4.PAE and IA32_EFER.NXE:
— CR4.PAE = 0 (32-bit paging is in use) or IA32_EFER.NXE= 0.
This flag is 0.
— CR4.PAE = 1 (either PAE paging or IA-32e paging is in use) and
IA32_EFER.NXE= 1.
If the access causing the page-fault exception was an instruction fetch, this
flag is 1; otherwise, it is 0. This flag describes the access causing the page-
fault exception, not the access rights specified by paging.
Vol. 3 4-35
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
Page-fault exceptions occur only due to an attempt to use a linear address. Failures
to load the PDPTE registers with PAE paging (see Section 4.4.1) cause general-
protection exceptions (#GP(0)) and not page-fault exceptions.
4.8
ACCESSED AND DIRTY FLAGS
For any paging-structure entry that is used during linear-address translation, bit 5 is
1
the accessed flag. For paging-structure entries that map a page (as opposed to
referencing another paging structure), bit 6 is the dirty flag. These flags are
provided for use by memory-management software to manage the transfer of pages
and paging structures into and out of physical memory.
Whenever the processor uses a paging-structure entry as part of linear-address
translation, it sets the accessed flag in that entry (if it is not already set).
Whenever there is a write to a linear address, the processor sets the dirty flag (if it is
address for the linear address (either a PTE or a PDE in which the PS flag is 1).
Memory-management software may clear these flags when a page or a paging struc-
once set, the processor does not clear them; only software can clear them.
A processor may cache information from the paging-structure entries in TLBs and
paging-structure caches (see Section 4.10). This fact implies that, if software
changes an accessed flag or a dirty flag from 1 to 0, the processor might not set the
corresponding bit in memory on a subsequent access using an affected linear
address (see Section 4.10.3.3). See Section 4.10.3.2 for how software can ensure
that these bits are updated as desired.
NOTE
The accesses used by the processor to set these flags may or may not
be exposed to the processor’s self-modifying code detection logic. If
the processor is executing code from the same memory area that is
being used for the paging structures, the setting of these flags may
or may not result in an immediate change to the executing code
stream.
1. With PAE paging, the PDPTEs are not used during linear-address translation but only to load the
PDPTE registers for some executions of the MOV CR instruction (see Section 4.4.1). For this rea-
son, the PDPTEs do not contain accessed flags with PAE paging.
4-36 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
4.9
PAGING AND MEMORY TYPING
access. Chapter 11, “Memory Cache Control” provides many details regarding
memory typing in the Intel-64 and IA-32 architectures. This section describes how
paging contributes to the determination of memory typing.
The way in which paging contributes to memory typing depends on whether the
1
processor supports the Page Attribute Table (PAT; see Section 11.12). Section
4.9.1 and Section 4.9.2 explain how paging contributes to memory typing depending
on whether the PAT is supported.
4.9.1
Paging and Memory Typing When the PAT is Not Supported
(Pentium Pro and Pentium II Processors)
NOTE
The PAT is supported on all processors that support IA-32e paging.
Thus, this section applies only to 32-bit paging and PAE paging.
If the PAT is not supported, paging contributes to memory typing in conjunction with
the memory-type range registers (MTRRs) as specified in Table 11-6 in Section
11.5.2.1.
For any access to a physical address, the table combines the memory type specified
for that physical address by the MTRRs with a PCD value and a PWT value. The latter
two values are determined as follows:
• For an access to a PDE with 32-bit paging, the PCD and PWT values come from
CR3.
• For an access to a PDE with PAE paging, the PCD and PWT values come from the
relevant PDPTE register.
• For an access to a PTE, the PCD and PWT values come from the relevant PDE.
• For an access to the physical address that is the translation of a linear address,
the PCD and PWT values come from the relevant PTE (if the translation uses a 4-
KByte page) or the relevant PDE (otherwise).
4.9.2
Paging and Memory Typing When the PAT is Supported
(Pentium III and More Recent Processor Families)
If the PAT is supported, paging contributes to memory typing in conjunction with the
PAT and the memory-type range registers (MTRRs) as specified in Table 11-7 in
Section 11.5.2.2.
1. The PAT is supported on Pentium III and more recent processor families. See Section 4.1.4 for
how to determine whether the PAT is supported.
Vol. 3 4-37
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
The PAT is a 64-bit MSR (IA32_PAT; MSR index 277H) comprising eight (8) 8-bit
entries (entry i comprises bits 8i+7:8i of the MSR).
For any access to a physical address, the table combines the memory type specified
for that physical address by the MTRRs with a memory type selected from the PAT.
Table 11-11 in Section 11.12.3 specifies how a memory type is selected from the PAT.
Specifically, it comes from entry i of the PAT, where i is defined as follows:
• For an access to an entry in a paging structure whose address is in CR3 (e.g., the
PML4 table with IA-32e paging), i = 2*PCD+PWT, where the PCD and PWT values
come from CR3.
• For an access to a PDE with PAE paging, i = 2*PCD+PWT, where the PCD and PWT
values come from the relevant PDPTE register.
• For an access to a paging-structure entry X whose address is in another paging-
structure entry Y, i = 2*PCD+PWT, where the PCD and PWT values come from Y.
• For an access to the physical address that is the translation of a linear address,
i = 4*PAT+2*PCD+PWT, where the PAT, PCD, and PWT values come from the
translation uses a 2-MByte page or a 4-MByte page).
4.9.3
Caching Paging-Related Information about Memory Typing
A processor may cache information from the paging-structure entries in TLBs and
paging-structure caches (see Section 4.10). These structures may include informa-
tion about memory typing. The processor may memory-typing information from the
TLBs and paging-structure caches instead of from the paging structures in memory.
This fact implies that, if software modifies a paging-structure entry to change the
memory-typing bits, the processor might not use that change for a subsequent
translation using that entry or for access to an affected linear address. See Section
4.10.3.2 for how software can ensure that the processor uses the modified memory
typing.
4.10
CACHING TRANSLATION INFORMATION
The Intel-64 and IA-32 architectures may accelerate the address-translation process
by caching data from the paging structures on the processor. Because the processor
in memory, it is important for software developers to understand how and when the
processor may cache such data. They should also understand what actions software
can take to remove cached data that may be inconsistent and when it should do so.
This section provides software developers information about the relevant processor
operation.
Section 4.10.1 and Section 4.10.2 describe how the processor may cache informa-
tion in translation lookaside buffers (TLBs) and paging-structure caches, respec-
4-38 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
tively. Section 4.10.3 explains how software can remove inconsistent cached
information by invalidating portions of the TLBs and paging-structure caches. Section
4.10.4 describes special considerations for multiprocessor systems.
4.10.1
Translation Lookaside Buffers (TLBs)
A processor may cache information about the translation of linear addresses in trans-
lation lookaside buffers (TLBs). In general, TLBs contain entries that map page
numbers to page frames; these terms are defined in Section 4.10.1.1. Section
4.10.1.2 describes how information may be cached in TLBs, and Section 4.10.1.3
allows software to indicate that certain translations should receive special treatment
when cached in the TLBs.
4.10.1.1 Page Numbers, Page Frames, and Page Offsets
Section 4.3, Section 4.4.2, and Section 4.5 give details of how the different paging
modes translate linear addresses to physical addresses. Specifically, the upper bits of
a linear address (called the page number) determine the upper bits of the physical
address (called the page frame); the lower bits of the linear address (called the
page offset) determine the lower bits of the physical address. The boundary
between the page number and the page offset is determined by the page size.
Specifically:
• 32-bit paging:
— If the translation does not use a PTE (because CR4.PSE = 1 and the PS flag is
1 in the PDE used), the page size is 4 MBytes and the page number comprises
bits 31:22 of the linear address.
— If the translation does use a PTE, the page size is 4 KBytes and the page
number comprises bits 31:12 of the linear address.
• PAE paging:
— If the translation does not use a PTE (because the PS flag is 1 in the PDE
used), the page size is 2 MBytes and the page number comprises bits 31:21
of the linear address.
— If the translation does uses a PTE, the page size is 4 KBytes and the page
number comprises bits 31:12 of the linear address.
• IA-32e paging:
— If the translation does not uses a PTE (because the PS flag is 1 in the PDE
used), the page size is 2 MBytes and the page number comprises bits 47:21
of the linear address.
— If the translation does use a PTE, the page size is 4 KBytes and the page
number comprises bits 47:12 of the linear address.
Vol. 3 4-39
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
4.10.1.2 Caching Translations in TLBs
The processor may accelerate the paging process by caching individual translations
lation. Each translation is referenced by a page number. It contains the following
information from the paging-structure entries used to translate linear addresses with
the page number:
• The physical address corresponding to the page number (the page frame).
• The access rights from the paging-structure entries used to translate linear
addresses with the page number (see Section 4.6):
— The logical-AND of the R/W flags.
— The logical-AND of the U/S flags.
— The logical-OR of the XD flags (necessary only if IA32_EFER.NXE = 1).
• Attributes from a paging-structure entry that identifies the final page frame for
the page number (either a PTE or a PDE in which the PS flag is 1):
— The dirty flag (see Section 4.8).
— The memory type (see Section 4.9).
(TLB entries may contain other information as well. A processor may implement
multiple TLBs, and some of these may be for special purposes, e.g., only for instruc-
tion fetches. Such special-purpose TLBs may not contain some of this information if
it is not necessary. For example, a TLB used only for instruction fetches need not
contain information about the R/W and dirty flags.)
Processors need not implement any TLBs. Processors that do implement TLBs may
invalidate any TLB entry at any time. Software should not rely on the existence of
TLBs or on the retention of TLB entries.
4.10.1.3 Details of TLB Use
Because the TLBs cache only valid translations, there can be a TLB entry for a page
number only if the P flag is 1 and the reserved bits are 0 in each of the paging-struc-
ture entries used to translate that page number. In addition, the processor does not
cache a translation for a page number unless the accessed flag is 1 in each of the
paging-structure entries used during translation; before caching a translation, the
processor sets any of these accessed flags that is not already 1.
The processor may cache translations required for prefetches and for accesses that
are a result of speculative execution that would never actually occur in the executed
code path.
If the page number of a linear address corresponds to a TLB entry, the processor may
use that TLB entry to determine the page frame, access rights, and other attributes
for accesses to that linear address. In this case, the processor may not actually
consult the paging structures in memory. The processor may retain a TLB entry
unmodified even if software subsequently modifies the relevant paging-structure
4-40 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
entries in memory. See Section 4.10.3.2 for how software can ensure that the
processor uses the modified paging-structure entries.
If the paging structures specify a translation using a page larger than 4 KBytes, some
processors may choose to cache multiple smaller-page TLB entries for that transla-
tion. Each such TLB entry would be associated with a page number corresponding to
the smaller page size (e.g., bits 47:12 of a linear address with IA-32e paging), even
though part of that page number (e.g., bits 20:12) are part of the offset with respect
to the page specified by the paging structures. The upper bits of the physical address
in such a TLB entry are derived from the physical address in the PDE used to create
the translation, while the lower bits come from the linear address of the access for
which the translation is created. There is no way for software to be aware that
multiple translations for smaller pages have been used for a large page.
If software modifies the paging structures so that the page size used for a 4-KByte
range of linear addresses changes, the TLBs may subsequently contain multiple
translations for the address range (one for each page size). A reference to a linear
address in the address range may use either translation. Which translation is used
may vary from one execution to another, and the choice may be implementation-
specific.
4.10.1.4 Global Pages
The Intel-64 and IA-32 architectures also allow for global pages when the PGE flag
(bit 7) is 1 in CR4. If the G flag (bit 8) is 1 in a paging-structure entry that maps a
page (either a PTE or a PDE in which the PS flag is 1), any TLB entry cached for a
linear address using that paging-structure entry is considered to be global. Because
the G flag is used only in paging-structure entries that map a page, and because
information from such entries are not cached in the paging-structure caches, the
global-page feature does not affect the behavior of the paging-structure caches.
4.10.2
Paging-Structure Caches
In addition to the TLBs, a processor may cache other information about the paging
structures in memory.
4.10.2.1 Caches for Paging Structures
A processor may support any or of all the following paging-structure caches:
• PML4 cache (IA-32e paging only). Each PML4-cache entry is referenced by a 9-
bit value and is used for linear addresses for which bits 47:39 have that value.
The entry contains information from the PML4E used to translate such linear
addresses:
— The physical address from the PML4E (the address of the page-directory-
pointer table).
Vol. 3 4-41
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
— The value of the R/W flag of the PML4E.
— The value of the U/S flag of the PML4E.
— The value of the XD flag of the PML4E.
— The values of the PCD and PWT flags of the PML4E.
The following items detail how a processor may use the PML4 cache:
— If the processor has a PML4-cache entry for a linear address, it may use that
entry when translating the linear address (instead of the PML4E in memory).
— The processor does not create a PML4-cache entry unless the P flag is 1 and
all reserved bits are 0 in the PML4E in memory.
— The processor does not create a PML4-cache entry unless the accessed flag is
1 in the PML4E in memory; before caching a translation, the processor sets
the accessed flag if it is not already 1.
— The processor may create a PML4-cache entry even if there are no transla-
tions for any linear address that might use that entry (e.g., because the P
flags are 0 in all entries in the referenced page-directory-pointer table).
— If the processor creates a PML4-cache entry, the processor may retain it
unmodified even if software subsequently modifies the corresponding PML4E
in memory.
1
• PDPTE cache (IA-32e paging only). Each PDPTE-cache entry is referenced by
an 18-bit value and is used for linear addresses for which bits 47:30 have that
value. The entry contains information from the PML4E and PDPTE used to
translate such linear addresses:
— The physical address from the PDPTE (the address of the page directory).
— The logical-AND of the R/W flags in the PML4E and the PDPTE.
— The logical-AND of the U/S flags in the PML4E and the PDPTE.
— The logical-OR of the XD flags in the PML4E and the PDPTE.
— The values of the PCD and PWT flags of the PDPTE.
The following items detail how a processor may use the PDPTE cache:
— If the processor has a PDPTE-cache entry for a linear address, it may use that
entry when translating the linear address (instead of the PML4E and the
PDPTE in memory).
— The processor does not create a PDPTE-cache entry unless the P flag is 1 and
— The processor does not create a PDPTE-cache entry unless the accessed flags
are 1 in the PML4E and the PDPTE in memory; before caching a translation,
the processor sets any accessed flags that are not already 1.
1. With PAE paging, the PDPTEs are stored in internal, non-architectural registers. The operation of
these registers is described in Section 4.4.1 and differs from that described here.
4-42 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
— The processor may create a PDPTE-cache entry even if there are no transla-
tions for any linear address that might use that entry.
— If the processor creates a PDPTE-cache entry, the processor may retain it
unmodified even if software subsequently modifies the corresponding PML4E
or PDPTE in memory.
• PDE cache. The use of the PDE cache depends on the paging mode:
— For 32-bit paging, each PDE-cache entry is referenced by a 10-bit value and
is used for linear addresses for which bits 31:22 have that value.
— For PAE paging, each PDE-cache entry is referenced by an 11-bit value and is
used for linear addresses for which bits 31:21 have that value.
— For IA-32e paging, each PDE-cache entry is referenced by a 27-bit value and
is used for linear addresses for which bits 47:21 have that value.
A PDE-cache entry contains information from the PML4E, PDPTE, and PDE used to
translate the relevant linear addresses (for 32-bit paging and PAE paging, only
the PDE applies):
— The physical address from the PDE (the address of the page table). (No PDE-
cache entry is created for a PDE that maps a page.)
— The logical-AND of the R/W flags in the PML4E, PDPTE, and PDE.
— The logical-AND of the U/S flags in the PML4E, PDPTE, and PDE.
— The logical-OR of the XD flags in the PML4E, PDPTE, and PDE.
— The values of the PCD and PWT flags of the PDE.
The following items detail how a processor may use the PDE cache (references
below to PML4Es and PDPTEs apply on to IA-32e paging):
— If the processor has a PDE-cache entry for a linear address, it may use that
entry when translating the linear address (instead of the PML4E, the PDPTE,
and the PDE in memory).
— The processor does not create a PDE-cache entry unless the P flag is 1, the PS
flag is 0, and the reserved bits are 0 in the PML4E, the PDPTE, and the PDE in
memory.
— The processor does not create a PDE-cache entry unless the accessed flag is
1 in the PML4E, the PDPTE, and the PDE in memory; before caching a trans-
lation, the processor sets any accessed flags that are not already 1.
— The processor may create a PDE-cache entry even if there are no translations
for any linear address that might use that entry.
— If the processor creates a PDE-cache entry, the processor may retain it
unmodified even if software subsequently modifies the corresponding PML4E,
the PDPTE, or the PDE in memory.
Information from a paging-structure entry can be included in entries in the paging-
structure caches for other paging-structure entries referenced by the original entry.
Vol. 3 4-43
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
For example, if the R/W flag is 0 in a PML4E, then the R/W flag will be 0 in any PDPTE-
cache entry for a PDPTE from the page-directory-pointer table referenced by that
PML4E. This is because the R/W flag of each such PDPTE-cache entry is the logical-
AND of the R/W flags in the appropriate PML4E and PDPTE.
The paging-structure caches contain information only from paging-structure entries
that reference other paging structures (and not those that map pages). Because the
G flag is not used in such paging-structure entries, the global-page feature does not
affect the behavior of the paging-structure caches.
The processor may create entries in paging-structure caches for translations
required for prefetches and for accesses that are a result of speculative execution
that would never actually occur in the executed code path.
A processor may or may not implement any of the paging-structure caches. Software
should rely on neither their presence nor their absence. The processor may invalidate
entries in these caches at any time. Because the processor may create the cache
entries at the time of translation and not update them following subsequent modifi-
cations to the paging structures in memory, software should take care to invalidate
the cache entries appropriately when causing such modifications. The invalidation of
TLBs and the paging-structure caches is described in Section 4.10.3.
4.10.2.2 Using the Paging-Structure Caches to Translate Linear Addresses
When a linear address is accessed, the processor uses a procedure such as the
following to determine the physical address to which it translates and whether the
access should be allowed:
• If the processor finds a TLB entry for the page number of the linear address, it
may use the physical address, access rights, and other attributes from that entry.
• If the processor does not find a TLB entry, it may use the upper bits of the linear
address to select an entry from the PDE cache (Section 4.10.2.1 indicates which
bits are used in each paging mode). It can then use that entry to complete the
translation process (locating a PTE, etc.) as if it had traversed the PDE (and, for
IA-32e paging, the PDPTE and PML4) corresponding to the PDE-cache entry.
• The following items apply when IA-32e paging is used:
— If the processor does not find a TLB entry or a PDE-cache entry, it may use
bits 47:30 of the linear address to select an entry from the PDPTE cache. It
can then use that entry to complete the translation process (locating a PDE,
etc.) as if it had traversed the PDPTE and the PML4 corresponding to the
PDPTE-cache entry.
— If the processor does not find a TLB entry, a PDE-cache entry, or a PDPTE-
cache entry, it may use bits 47:39 of the linear address to select an entry
from the PML4 cache. It can then use that entry to complete the translation
process (locating a PDPTE, etc.) as if it had traversed the corresponding
PML4.
4-44 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
(Any of the above steps would be skipped if the processor does not support the cache
in question.)
If the processor does not find a TLB or paging-structure-cache entry for the linear
address, it uses the linear address to traverse the entire paging-structure hierarchy,
as described in Section 4.3, Section 4.4.2, and Section 4.5.
4.10.2.3 Multiple Cached Entries for a Single Paging-Structure Entry
Note that multiple cached entries (in the paging-structure caches or TLBs) may
contain information derived from a single paging-structure entry. The following items
give some examples for IA-32e paging:
• Suppose that two PML4Es contain the same physical address and thus reference
the same page-directory-pointer table. Any PDPTE in that table may result in two
PDPTE-cache entries, each associated with a different set of linear addresses.
th
th
Specifically, suppose that the n
and n
entries in the PML4 table contain the
1
2
th
same physical address. This implies that the physical address in the m PDPTE in
the page-directory-pointer table would appear in the PDPTE-cache entries
associated with both p and p , where (p » 9) = n , (p » 9) = n , and (p &
1
2
1
1
2
2
1
1FFH) = (p & 1FFH) = m. This is because both PDPTE-cache entries use the
2
th
same PDPTE, one resulting from a reference from the n
PML4E and one from
1
th
the n
PML4E.
2
• Suppose that the first PML4E (i.e., the one in position 0) contains the physical
address X in CR3 (the physical address of the PML4 table). This implies the
following:
— Any PML4-cache entry associated with linear addresses with 0 in bits 47:39
contains address X.
— Any PDPTE-cache entry associated with linear addresses with 0 in bits 47:30
contains address X. This is because the translation for a linear address for
which the value of bits 47:30 is 0 uses the value of bits 47:39 (0) to locate a
page-directory-pointer table at address X (the address of the PML4 table). It
then uses the value of bits 38:30 (also 0) to find address X again and to store
that address in the PDPTE-cache entry.
— Any PDE-cache entry associated with linear addresses with 0 in bits 47:21
contains address X for similar reasons.
— Any TLB entry for page number 0 (associated with linear addresses with 0 in
bits 47:12) translates to page frame X » 12 for similar reasons.
The same PML4E contributes its address X to all these cache entries because the
self-referencing nature of the entry causes it to be used as a PML4E, a PDPTE, a
PDE, and a PTE.
Vol. 3 4-45
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
4.10.3
Invalidation of TLBs and Paging-Structure Caches
As noted in Section 4.10.1 and Section 4.10.2, the processor may create entries in
the TLBs and the paging-structure caches when linear addresses are translated, and
it may retain these entries even after the paging structures used to create them have
been modified. To ensure that linear-address translation uses the modified paging
structures, software should take action to invalidate any cached entries that may
contain information that has since been modified.
4.10.3.1 Operations that Invalidate TLBs and Paging-Structure Caches
It is recommended that software use the following instructions to invalidate entries in
the TLBs and the paging-structure caches:
• INVLPG. This instruction takes a single operand, which is a linear address. The
instruction invalidates any TLB entries with a page number corresponding to the
1
linear address, including those for global pages (see Section 4.10.1.4). INVLPG
also invalidates all entries in all paging-structure caches regardless of the linear
addresses to which they correspond.
• MOV to CR3. This instruction invalidates all TLB entries except those for global
pages. It also invalidates all entries in all paging-structure caches.
• MOV to CR4. If this instruction changes value of the PGE flag (bit 7) of CR4, it
invalidates all TLB entries and all entries in all paging-structure caches. This
were no global TLB entries before the execution; and (2) if CR4.PGE is changing
from 1 to 0, there will be no global TLB entries after the execution.
• Task switch. If a task switch changes the value of CR3, it invalidates all TLB
entries except those for global pages. It also invalidates all entries in all paging-
structure caches.
• VMX transitions. See Section 4.11.1.
The processor is always free to invalidate additional entries in the TLBs and paging-
structure caches. The following are some examples:
• INVLPG may invalidate TLB entries for pages other than the one corresponding to
its linear-address operand.
• MOV to CR3 may invalidate TLB entries for global pages.
• On a processor supporting Hyper-Threading Technology, invalidations performed
on one logical processor may invalidate entries in the TLBs and paging-structure
caches used by other logical processors.
(Other instructions and operations may invalidate entries in the TLBs and the paging-
structure caches, but the instructions identified above are recommended.)
1. Even if the paging structures map the linear address using a page larger than 4 KBytes and there
are multiple TLB entries for that page (see Section 4.10.1), the instruction invalidates all of them.
4-46 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
In addition to the instructions identified above, page faults invalidate entries in the
TLBs and paging-structure caches. In particular, a page-fault exception resulting
from an attempt to use a linear address will invalidate any PML4-cache, PDPTE-
1
any TLB entry for that address's page number. These invalidations ensure that the
page-fault exception will not recur (if the faulting instruction is re-executed) if it
would not be caused by the contents of the paging structures in memory (and if,
therefore, it resulted from cached entries that were not invalidated after the paging
structures were modified in memory).
As noted in Section 4.10.1, some processors may choose to cache multiple smaller-
page TLB entries for a translation specified by the paging structures to use a page
larger than 4 KBytes. There is no way for software to be aware that multiple transla-
tions for smaller pages have been used for a large page. The INVLPG instruction and
page faults provide the same assurances that they provide when a single TLB entry is
used: they invalidate all TLB entries corresponding to the translation specified by the
paging structures.
4.10.3.2 Recommended Invalidation
perform invalidations:
• If software modifies a paging-structure entry that identifies the final page frame
for a page number (either a PTE or a PDE in which the PS flag is 1), it should
execute INVLPG for any linear address with a page number whose translation
2
uses that PTE. (If the paging-structure entry may be used in the translation of
different page numbers — see Section 4.10.2.3 — software should execute
INVLPG for linear addresses with each of those page numbers; alternatively, it
could use MOV to CR3 or MOV to CR4.)
• If software modifies a paging-structure entry that references another paging
structure, it may use one of the following approaches depending upon the types
and number of translations controlled by the modified entry:
— Execute INVLPG for linear addresses with each of the page numbers with
translations that would use the entry. However, if no page numbers that
would use the entry have translations (e.g., because the P flags are 0 in all
entries in the paging structure referenced by the modified entry), it remains
necessary to execute INVLPG at least once.
— Execute MOV to CR3 if the modified entry controls no global pages.
— Execute MOV to CR4 to modify CR4.PGE.
1. Unlike INVLPG, page faults need not invalidate all entries in the paging-structure caches, only
those that would be used to translate the faulting linear address.
2. One execution of INVLPG is sufficient even for a page with size greater than 4 KBytes.
Vol. 3 4-47
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
• If software using PAE paging modifies a PDPTE, it should reload CR3 with the
register’s current value to ensure that the modified PDPTE is loaded into the
corresponding PDPTE register (see Section 4.4.1).
multiple purposes (see Section 4.10.2.3), software should perform invalidations
for all of these purposes. For example, if a single entry might serve as both a PDE
and PTE, it may be necessary to execute INVLPG with two (or more) linear
addresses, one that uses the entry as a PDE and one that uses it as a PTE. (Alter-
natively, software could use MOV to CR3 or MOV to CR4.)
• As noted in Section 4.10.1, the TLBs may subsequently contain multiple transla-
tions for the address range if software modifies the paging structures so that the
page size used for a 4-KByte range of linear addresses changes. A reference to a
Software wishing to prevent this uncertainty should not write to a paging-
structure entry in a way that would change, for any linear address, both the page
size and either the page frame, access rights, or other attributes. It can instead
use the following algorithm: first clear the P flag in the relevant paging-structure
entry (e.g., PDE); then invalidate any translations for the affected linear
addresses (see Section 4.10.3.2); and then modify the relevant paging-structure
entry to set the P flag and establish modified translation(s) for the new page size.
4.10.3.3 Optional Invalidation
The following items describe cases in which software may choose not to invalidate
and the potential consequences of that choice:
• If a paging-structure entry is modified to change the P flag from 0 to 1, no inval-
idation is necessary. This is because no TLB entry or paging-structure cache entry
1
is created with information from a paging-structure entry in which the P flag is 0.
• If a paging-structure entry is modified to change the accessed flag from 0 to 1,
no invalidation is necessary (assuming that an invalidation was performed the
last time the accessed flag was changed from 1 to 0). This is because no TLB
structure entry in which the accessed flag is 0.
• If a paging-structure entry is modified to change the R/W flag from 0 to 1, failure
to perform an invalidation may result in a “spurious” page-fault exception (e.g.,
in response to an attempted write access) but no other adverse behavior. Such
an exception will occur at most once for each affected linear address (see Section
4.10.3.1).
• If a paging-structure entry is modified to change the U/S flag from 0 to 1, failure
to perform an invalidation may result in a “spurious” page-fault exception (e.g.,
1. If it is also the case that no invalidation was performed the last time the P flag was changed
from 1 to 0, the processor may use a TLB entry or paging-structure cache entry that was cre-
ated when the P flag had earlier been 1.
4-48 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
in response to an attempted user-mode access) but no other adverse behavior.
Section 4.10.3.1).
• If a paging-structure entry is modified to change the XD flag from 1 to 0, failure
to perform an invalidation may result in a “spurious” page-fault exception (e.g.,
in response to an attempted instruction fetch) but no other adverse behavior.
Such an exception will occur at most once for each affected linear address (see
Section 4.10.3.1).
• If a paging-structure entry is modified to change the accessed flag from 1 to 0,
failure to perform an invalidation may result in the processor not setting that bit
in response to a subsequent access to a linear address whose translation uses the
entry. Software cannot interpret the bit being clear as an indication that such an
access has not occurred.
• If software modifies a paging-structure entry that identifies the final physical
address for a linear address (either a PTE or a PDE in which the PS flag is 1) to
change the dirty flag from 1 to 0, failure to perform an invalidation may result in
the processor not setting that bit in response to a subsequent write to a linear
• The read of a paging-structure entry in translating an address being used to fetch
an instruction may appear to execute before an earlier write to that paging-
structure entry if there is no serializing instruction between the write and the
instruction fetch. Note that the invalidating instructions identified in Section
4.10.3.1 are all serializing instructions.
• Section 4.10.2.3 describes situations in which a single paging-structure entry
may contain information cached in multiple entries in the paging-structure
caches. Because all entries in these caches are invalidated by any execution of
INVLPG, it is not necessary to follow the modification of such a paging-structure
entry by executing INVLPG multiple times solely for the purpose of invalidating
these multiple cached entries. (It may be necessary to do so to invalidate
multiple TLB entries.)
4.10.4
Propagation of Paging-Structure Changes to Multiple
Processors
As noted in Section 4.10.3, software that modifies a paging-structure entry may
need to invalidate entries in the TLBs and paging-structure caches that were derived
from the modified entry before it was modified. In a system containing more than
one logical processor, software must account for the fact that there may be entries in
the TLBs and paging-structure caches of logical processors other than the one used
to modify the paging-structure entry. The process of propagating the changes to a
paging-structure entry is commonly referred to as “TLB shootdown.”
TLB shootdown can be done using memory-based semaphores and/or interprocessor
interrupts (IPI). The following items describe a simple but inefficient example of a
Vol. 3 4-49
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
TLB shootdown algorithm for processors supporting the Intel-64 and IA-32 architec-
tures:
1. Begin barrier: Stop all but one logical processor; that is, cause all but one to
execute the HLT instruction or to enter a spin loop.
2. Allow the active logical processor to change the necessary paging-structure
entries.
3. Allow all logical processors to perform invalidations appropriate to the modifica-
tions to the paging-structure entries.
4. Allow all logical processors to resume normal operation.
Alternative, performance-optimized, TLB shootdown algorithms may be developed;
however, software developers must take care to ensure that the following conditions
are met:
• All logical processors that are using the paging structures that are being modified
must participate and perform appropriate invalidations after the modifications
are made.
• If the modifications to the paging-structure entries are made before the barrier
or if there is no barrier, the operating system must ensure one of the following:
(1) that the affected linear-address range is not used between the time of modifi-
cation and the time of invalidation; or (2) that it is prepared to deal with the
consequences of the affected linear-address range being used during that period.
For example, if the operating system does not allow pages being freed to be
reallocated for another purpose until after the required invalidations, writes to
those pages by errant software will not unexpectedly modify memory that is in
use.
• Software must be prepared to deal with reads, instruction fetches, and prefetch
requests to the affected linear-address range that are a result of speculative
execution that would never actually occur in the executed code path.
When multiple logical processors are using the same linear-address space at the
same time, they must coordinate before any request to modify the paging-structure
entries that control that linear-address space. In these cases, the barrier in the TLB
shootdown routine may not be required. For example, when freeing a range of linear
addresses, some other mechanism can assure no logical processor is using that
range before the request to free it is made. In this case, a logical processor freeing
the range can clear the P flags in the PTEs associated with the range, free the phys-
ical page frames associated with the range, and then signal the other logical proces-
sors using that linear-address space to perform the necessary invalidations. All the
affected logical processors must complete their invalidations before the linear-
address range and the physical page frames previously associated with that range
can be reallocated.
4-50 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
4.11
INTERACTIONS WITH VIRTUAL-MACHINE
EXTENSIONS (VMX)
The architecture for virtual-machine extensions (VMX) includes features that interact
with paging. Section 4.11.1 discusses ways in which VMX-specific control transfers,
called VMX transitions specially affect paging. Section 4.11.2 gives an overview of
VMX features specifically designed to support address translation.
4.11.1
VMX Transitions
The VMX architecture defines two control transfers called VM entries and VM exits;
collectively, these are called VMX transitions. VM entries and VM exits are
described in detail in Chapter 23 and Chapter 24, respectively, in the Intel® 64 and
IA-32 Architectures Software Developer’s Manual, Volume 3B. The following items
identify paging-related details:
• VMX transitions modify the CR0 and CR4 registers and the IA32_EFER MSR
concurrently. For this reason, they allow transitions between paging modes that
would not otherwise be possible:
— VM entries allow transitions from IA-32e paging directly to either 32-bit
paging or PAE paging.
— VM exits allow transitions from either 32-bit paging or PAE paging directly to
IA-32e paging.
• VMX transitions that result in PAE paging load the PDPTE registers (see Section
4.4.1) as follows:
— VM entries load the PDPTE registers either from the physical address being
loaded into CR3 or from the virtual-machine control structure (VMCS); see
Section 23.3.2.4.
— VM exits load the PDPTE registers from the physical address being loaded into
CR3; see Section 24.5.4.
• VMX transitions invalidate the TLBs and paging-structure caches based on certain
control settings. See Section 23.3.2.5 and Section 24.5.5.
4.11.2
VMX Support for Address Translation
Chapter 25, “VMX Support for Address Translation,” in the Intel® 64 and IA-32 Archi-
tectures Software Developer’s Manual, Volume 3B describe two features of the
virtual-machine extensions (VMX) that interact directly with paging. These are
virtual-processor identifiers (VPIDs) and the extended page table mechanism
(EPT).
VPIDs provide a way for software to identify to the processor the address spaces for
different “virtual processors.” The processor may use this identification to maintain
Vol. 3 4-51
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
concurrently information for multiple address spaces in its TLBs and paging-structure
caches. See Section 25.1 for details.
When EPT is in use, the addresses in the paging-structures are not used as physical
addresses to access memory and memory-mapped I/O. Instead, they are treated as
to produce physical addresses. EPT can also specify its own access rights and
memory typing; these are used on conjunction with those specified in this chapter.
See Section 25.2 for more information.
Both VPIDs and EPT may change the way that a processor maintains information in
TLBs and paging structure caches and the ways in which software can manage that
information. Some of the behaviors documented in Section 4.10 may change. See
Section 25.3 for details.
4.12
USING PAGING FOR VIRTUAL MEMORY
With paging, portions of the linear-address space need not be mapped to the phys-
ical-address space; data for the unmapped addresses can be stored externally (e.g.,
on disk). This method of mapping the linear-address space is referred to as virtual
memory or demand-paged virtual memory.
the physical-address space and/or external storage. When a program (or task) refer-
ences a linear address, the processor uses paging to translate the linear address into
a corresponding physical address if such an address is defined.
If the page containing the linear address is not currently mapped into the physical-
address space, the processor generates a page-fault exception as described in
Section 4.7. The handler for page-fault exceptions typically directs the operating
system or executive to load data for the unmapped page from external storage into
physical memory (perhaps writing a different page from physical memory out to
external storage in the process) and to map it using paging (by updating the paging
structures). When the page has been loaded into physical memory, a return from the
exception handler causes the instruction that generated the exception to be
restarted.
Paging differs from segmentation through its use of fixed-size pages. Unlike
segments, which usually are the same size as the code or data structures they hold,
pages have a fixed size. If segmentation is the only form of address translation used,
a data structure present in physical memory will have all of its parts in memory. If
paging is used, a data structure can be partly in memory and partly in disk storage.
4.13
MAPPING SEGMENTS TO PAGES
The segmentation and paging mechanisms provide in the support a wide variety of
approaches to memory management. When segmentation and paging are combined,
4-52 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
segments can be mapped to pages in several ways. To implement a flat (unseg-
mented) addressing environment, for example, all the code, data, and stack modules
can be mapped to one or more large segments (up to 4-GBytes) that share same
range of linear addresses (see Figure 3-2 in Section 3.2.2). Here, segments are
essentially invisible to applications and the operating-system or executive. If paging
is used, the paging mechanism can map a single linear-address space (contained in
a single segment) into virtual memory. Alternatively, each program (or task) can
have its own large linear-address space (contained in its own segment), which is
mapped into virtual memory through its own paging structures.
Segments can be smaller than the size of a page. If one of these segments is placed
in a page which is not shared with another segment, the extra memory is wasted. For
example, a small data structure, such as a 1-Byte semaphore, occupies 4 KBytes if it
is placed in a page by itself. If many semaphores are used, it is more efficient to pack
them into a single page.
The Intel-64 and IA-32 architectures do not enforce correspondence between the
boundaries of pages and segments. A page can contain the end of one segment and
the beginning of another. Similarly, a segment can contain the end of one page and
the beginning of another.
Memory-management software may be simpler and more efficient if it enforces some
alignment between page and segment boundaries. For example, if a segment which
can fit in one page is placed in two pages, there may be twice as much paging over-
head to support access to that segment.
One approach to combining paging and segmentation that simplifies memory-
management software is to give each segment its own page table, as shown in
Figure 4-12. This convention gives the segment a single entry in the page directory,
and this entry provides the access control information for paging the entire segment.
Page Frames
Page Directory
Page Tables
LDT
PTE
PTE
PTE
Seg. Descript.
Seg. Descript.
PDE
PDE
PTE
PTE
Figure 4-12. Memory Management Convention That Assigns a Page Table
to Each Segment
Vol. 3 4-53
Download from Www.Somanuals.com. All Manuals Search And Download.
PAGING
4-54 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
CHAPTER 5
PROTECTION
In protected mode, the Intel 64 and IA-32 architectures provide a protection mecha-
nism that operates at both the segment level and the page level. This protection
mechanism provides the ability to limit access to certain segments or pages based on
privilege levels (four privilege levels for segments and two privilege levels for pages).
For example, critical operating-system code and data can be protected by placing
them in more privileged segments than those that contain applications code. The
processor’s protection mechanism will then prevent application code from accessing
the operating-system code and data in any but a controlled, defined manner.
Segment and page protection can be used at all stages of software development to
assist in localizing and detecting design problems and bugs. It can also be incorpo-
rated into end-products to offer added robustness to operating systems, utilities soft-
ware, and applications software.
When the protection mechanism is used, each memory reference is checked to verify
that it satisfies various protection checks. All checks are made before the memory
cycle is started; any violation results in an exception. Because checks are performed
in parallel with address translation, there is no performance penalty. The protection
checks that are performed fall into the following categories:
• Limit checks.
• Type checks.
• Restriction of addressable domain.
• Restriction of procedure entry-points.
All protection violation results in an exception being generated. See Chapter 6,
“Interrupt and Exception Handling,” for an explanation of the exception mechanism.
This chapter describes the protection mechanism and the violations which lead to
exceptions.
The following sections describe the protection mechanism available in protected
mode. See Chapter 17, “8086 Emulation,” for information on protection in real-
address and virtual-8086 mode.
5.1
ENABLING AND DISABLING SEGMENT AND PAGE
PROTECTION
Setting the PE flag in register CR0 causes the processor to switch to protected mode,
which in turn enables the segment-protection mechanism. Once in protected mode,
Vol. 3 5-1
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
there is no control bit for turning the protection mechanism on or off. The part of the
segment-protection mechanism that is based on privilege levels can essentially be
disabled while still in protected mode by assigning a privilege level of 0 (most privi-
leged) to all segment selectors and segment descriptors. This action disables the
privilege level protection barriers between segments, but other protection checks
such as limit checking and type checking are still carried out.
Page-level protection is automatically enabled when paging is enabled (by setting the
PG flag in register CR0). Here again there is no mode bit for turning off page-level
protection once paging is enabled. However, page-level protection can be disabled by
performing the following operations:
• Clear the WP flag in control register CR0.
• Set the read/write (R/W) and user/supervisor (U/S) flags for each page-directory
and page-table entry.
This action makes each page a writable, user page, which in effect disables page-
level protection.
5.2
FIELDS AND FLAGS USED FOR SEGMENT-LEVEL AND
PAGE-LEVEL PROTECTION
The processor’s protection mechanism uses the following fields and flags in the
system data structures to control access to segments and pages:
• Descriptor type (S) flag — (Bit 12 in the second doubleword of a segment
descriptor.) Determines if the segment descriptor is for a system segment or a
code or data segment.
• Type field — (Bits 8 through 11 in the second doubleword of a segment
descriptor.) Determines the type of code, data, or system segment.
• Limit field — (Bits 0 through 15 of the first doubleword and bits 16 through 19
of the second doubleword of a segment descriptor.) Determines the size of the
segment, along with the G flag and E flag (for data segments).
• G flag — (Bit 23 in the second doubleword of a segment descriptor.) Determines
the size of the segment, along with the limit field and E flag (for data segments).
• E flag — (Bit 10 in the second doubleword of a data-segment descriptor.)
Determines the size of the segment, along with the limit field and G flag.
• Descriptor privilege level (DPL) field — (Bits 13 and 14 in the second
doubleword of a segment descriptor.) Determines the privilege level of the
segment.
• Requested privilege level (RPL) field — (Bits 0 and 1 of any segment
selector.) Specifies the requested privilege level of a segment selector.
• Current privilege level (CPL) field — (Bits 0 and 1 of the CS segment
register.) Indicates the privilege level of the currently executing program or
5-2 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
procedure. The term current privilege level (CPL) refers to the setting of this
field.
• User/supervisor (U/S) flag — (Bit 2 of paging-structure entries.) Determines
type of access allowed to a page: read-only or read/write.
• Execute-disable (XD) flag — (Bit 63 of certain paging-structure entries.)
Determines the type of access allowed to a page: executable or not-executable.
Figure 5-1 shows the location of the various fields and flags in the data, code, and
system- segment descriptors; Figure 3-6 shows the location of the RPL (or CPL) field
in a segment selector (or the CS register); and Chapter 4 identifies the locations of
the U/S, R/W, and XD flags in the paging-structure entries.
Vol. 3 5-3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
Data-Segment Descriptor
31
31
21 20 19
16
24 23 22
15 14 13 12 11
0
0
8
7
A
V
L
D
P
L
Type
Limit
19:16
G
B
Base 31:24
0
P
Base 23:16
4
0
1
0
E
W
A
16
15
Base Address 15:00
Segment Limit 15:00
Code-Segment Descriptor
31
31
21 20 19
16
24 23 22
15 14 13 12 11
7
0
0
8
A
V
D
P
L
Type
Limit
19:16
G
D
0
P
Base 31:24
Base 23:16
4
0
L
1
1
C
R
A
16
15
Base Address 15:00
Segment Limit 15:00
System-Segment Descriptor
31
31
23
G
21 20 19
16
24
22
15 14 13 12 11
7
0
0
8
D
P
L
Limit
19:16
Base 31:24
0
P
0
Type
Base 23:16
4
0
16
15
Base Address 15:00
Accessed
Segment Limit 15:00
A
E
G
R
Expansion Direction
Granularity
Readable
AVL Available to Sys. Programmer’s
B
C
D
Big
LIMIT Segment Limit
W
P
Conforming
Default
Writable
Present
DPL Descriptor Privilege Level
Reserved
Figure 5-1. Descriptor Fields Used for Protection
Many different styles of protection schemes can be implemented with these fields
and flags. When the operating system creates a descriptor, it places values in these
fields and flags in keeping with the particular protection style chosen for an operating
system or executive. Application program do not generally access or modify these
fields and flags.
5-4 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
The following sections describe how the processor uses these fields and flags to
perform the various categories of checks described in the introduction to this chapter.
5.2.1
Code Segment Descriptor in 64-bit Mode
Code segments continue to exist in 64-bit mode even though, for address calcula-
tions, the segment base is treated as zero. Some code-segment (CS) descriptor
content (the base address and limit fields) is ignored; the remaining fields function
normally (except for the readable bit in the type field).
Code segment descriptors and selectors are needed in IA-32e mode to establish the
processor’s operating mode and execution privilege-level. The usage is as follows:
• IA-32e mode uses a previously unused bit in the CS descriptor. Bit 53 is defined
as the 64-bit (L) flag and is used to select between 64-bit mode and compatibility
mode when IA-32e mode is active (IA32_EFER.LMA = 1). See Figure 5-2.
— If CS.L = 0 and IA-32e mode is active, the processor is running in compati-
bility mode. In this case, CS.D selects the default size for data and addresses.
If CS.D = 0, the default data and address size is 16 bits. If CS.D = 1, the
default data and address size is 32 bits.
— If CS.L = 1 and IA-32e mode is active, the only valid setting is CS.D = 0. This
setting indicates a default operand size of 32 bits and a default address size
of 64 bits. The CS.L = 1 and CS.D = 1 bit combination is reserved for future
use and a #GP fault will be generated on an attempt to use a code segment
with these bits set in IA-32e mode.
• In IA-32e mode, the CS descriptor’s DPL is used for execution privilege checks
(as in legacy 32-bit mode).
Vol. 3 5-5
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
Code-Segment Descriptor
31
23
G
21 20 19
16
24
22
D
15 14 13 12 11
D
7
0
0
8
A
Type
L
P
V
L
P
L
4
0
1
1
C
R
A
31
A
Accessed
G
R
P
Granularity
Readable
Present
AVL Available to Sys. Programmer’s
C
D
Conforming
Default
DPL Descriptor Privilege Level
64-Bit Flag
L
Figure 5-2. Descriptor Fields with Flags used in IA-32e Mode
5.3
LIMIT CHECKING
The limit field of a segment descriptor prevents programs or procedures from
addressing memory locations outside the segment. The effective value of the limit
depends on the setting of the G (granularity) flag (see Figure 5-1). For data
segments, the limit also depends on the E (expansion direction) flag and the B
(default stack pointer size and/or upper bound) flag. The E flag is one of the bits in
the type field when the segment descriptor is for a data-segment type.
When the G flag is clear (byte granularity), the effective limit is the value of the
20-bit limit field in the segment descriptor. Here, the limit ranges from 0 to FFFFFH
(1 MByte). When the G flag is set (4-KByte page granularity), the processor scales
12
the value in the limit field by a factor of 2 (4 KBytes). In this case, the effective
limit ranges from FFFH (4 KBytes) to FFFFFFFFH (4 GBytes). Note that when scaling
is used (G flag is set), the lower 12 bits of a segment offset (address) are not checked
against the limit; for example, note that if the segment limit is 0, offsets 0 through
FFFH are still valid.
For all types of segments except expand-down data segments, the effective limit is
the last address that is allowed to be accessed in the segment, which is one less than
the size, in bytes, of the segment. The processor causes a general-protection excep-
tion any time an attempt is made to access the following addresses in a segment:
• A byte at an offset greater than the effective limit
• A word at an offset greater than the (effective-limit – 1)
5-6 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
• A doubleword at an offset greater than the (effective-limit – 3)
• A quadword at an offset greater than the (effective-limit – 7)
For expand-down data segments, the segment limit has the same function but is
interpreted differently. Here, the effective limit specifies the last address that is not
allowed to be accessed within the segment; the range of valid offsets is from (effec-
tive-limit + 1) to FFFFFFFFH if the B flag is set and from (effective-limit + 1) to FFFFH
if the B flag is clear. An expand-down segment has maximum size when the segment
limit is 0.
Limit checking catches programming errors such as runaway code, runaway
subscripts, and invalid pointer calculations. These errors are detected when they
occur, so identification of the cause is easier. Without limit checking, these errors
could overwrite code or data in another segment.
In addition to checking segment limits, the processor also checks descriptor table
tive descriptor tables. The LDTR and task registers contain 32-bit segment limit value
(read from the segment descriptors for the current LDT and TSS, respectively). The
processor uses these segment limits to prevent accesses beyond the bounds of the
current LDT and TSS. See Section 3.5.1, “Segment Descriptor Tables,” for more infor-
mation on the GDT and LDT limit fields; see Section 6.10, “Interrupt Descriptor Table
(IDT),” for more information on the IDT limit field; and see Section 7.2.4, “Task
Register,” for more information on the TSS segment limit field.
5.3.1
Limit Checking in 64-bit Mode
In 64-bit mode, the processor does not perform runtime limit checking on code or
data segments. However, the processor does check descriptor-table limits.
5.4
TYPE CHECKING
Segment descriptors contain type information in two places:
• The S (descriptor type) flag.
• The type field.
The processor uses this information to detect programming errors that result in an
attempt to use a segment or gate in an incorrect or unintended manner.
The S flag indicates whether a descriptor is a system type or a code or data type. The
type field provides 4 additional bits for use in defining various types of code, data,
and system descriptors. Table 3-1 shows the encoding of the type field for code and
data descriptors; Table 3-2 shows the encoding of the field for system descriptors.
Vol. 3 5-7
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
The processor examines type information at various times while operating on
segment selectors and segment descriptors. The following list gives examples of
typical operations where type checking is performed (this list is not exhaustive):
• When a segment selector is loaded into a segment register — Certain
segment registers can contain only certain descriptor types, for example:
— The CS register only can be loaded with a selector for a code segment.
— Segment selectors for code segments that are not readable or for system
segments cannot be loaded into data-segment registers (DS, ES, FS, and
GS).
— Only segment selectors of writable data segments can be loaded into the SS
register.
• When a segment selector is loaded into the LDTR or task register — For example:
— The LDTR can only be loaded with a selector for an LDT.
— The task register can only be loaded with a segment selector for a TSS.
• When instructions access segments whose descriptors are already
loaded into segment registers — Certain segments can be used by instruc-
tions only in certain predefined ways, for example:
— No instruction may write into an executable segment.
— No instruction may write into a data segment if it is not writable.
— No instruction may read an executable segment unless the readable flag is
set.
• When an instruction operand contains a segment selector — Certain
instructions can access segments or gates of only a particular type, for example:
— A far CALL or far JMP instruction can only access a segment descriptor for a
conforming code segment, nonconforming code segment, call gate, task
gate, or TSS.
— The LLDT instruction must reference a segment descriptor for an LDT.
— The LTR instruction must reference a segment descriptor for a TSS.
— The LAR instruction must reference a segment or gate descriptor for an LDT,
TSS, call gate, task gate, code segment, or data segment.
— The LSL instruction must reference a segment descriptor for a LDT, TSS, code
segment, or data segment.
— IDT entries must be interrupt, trap, or task gates.
• During certain internal operations — For example:
— On a far call or far jump (executed with a far CALL or far JMP instruction), the
processor determines the type of control transfer to be carried out (call or
jump to another code segment, a call or jump through a gate, or a task
switch) by checking the type field in the segment (or gate) descriptor pointed
to by the segment (or gate) selector given as an operand in the CALL or JMP
5-8 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
instruction. If the descriptor type is for a code segment or call gate, a call or
jump to another code segment is indicated; if the descriptor type is for a TSS
or task gate, a task switch is indicated.
— On a call or jump through a call gate (or on an interrupt- or exception-handler
call through a trap or interrupt gate), the processor automatically checks that
the segment descriptor being pointed to by the gate is for a code segment.
— On a call or jump to a new task through a task gate (or on an interrupt- or
exception-handler call to a new task through a task gate), the processor
automatically checks that the segment descriptor being pointed to by the
task gate is for a TSS.
— On a call or jump to a new task by a direct reference to a TSS, the processor
automatically checks that the segment descriptor being pointed to by the
CALL or JMP instruction is for a TSS.
— On return from a nested task (initiated by an IRET instruction), the processor
checks that the previous task link field in the current TSS points to a TSS.
5.4.1
Null Segment Selector Checking
Attempting to load a null segment selector (see Section 3.4.2, “Segment Selectors”)
into the CS or SS segment register generates a general-protection exception (#GP).
A null segment selector can be loaded into the DS, ES, FS, or GS register, but any
attempt to access a segment through one of these registers when it is loaded with a
null segment selector results in a #GP exception being generated. Loading unused
data-segment registers with a null segment selector is a useful method of detecting
accesses to unused segment registers and/or preventing unwanted accesses to data
segments.
5.4.1.1
NULL Segment Checking in 64-bit Mode
In 64-bit mode, the processor does not perform runtime checking on NULL segment
selectors. The processor does not cause a #GP fault when an attempt is made to
5.5
PRIVILEGE LEVELS
The processor’s segment-protection mechanism recognizes 4 privilege levels,
numbered from 0 to 3. The greater numbers mean lesser privileges. Figure 5-3
shows how these levels of privilege can be interpreted as rings of protection.
The center (reserved for the most privileged code, data, and stacks) is used for the
segments containing the critical software, usually the kernel of an operating system.
Outer rings are used for less critical software. (Systems that use only 2 of the 4
possible privilege levels should use levels 0 and 3.)
Vol. 3 5-9
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
Protection Rings
Operating
System
Kernel
Level 0
Level 1
Operating System
Services
Level 2
Level 3
Applications
Figure 5-3. Protection Rings
The processor uses privilege levels to prevent a program or task operating at a lesser
privilege level from accessing a segment with a greater privilege, except under
controlled situations. When the processor detects a privilege level violation, it gener-
ates a general-protection exception (#GP).
To carry out privilege-level checks between code segments and data segments, the
processor recognizes the following three types of privilege levels:
• Current privilege level (CPL) — The CPL is the privilege level of the currently
executing program or task. It is stored in bits 0 and 1 of the CS and SS segment
registers. Normally, the CPL is equal to the privilege level of the code segment
from which instructions are being fetched. The processor changes the CPL when
program control is transferred to a code segment with a different privilege level.
The CPL is treated slightly differently when accessing conforming code segments.
Conforming code segments can be accessed from any privilege level that is equal
to or numerically greater (less privileged) than the DPL of the conforming code
segment. Also, the CPL is not changed when the processor accesses a conforming
code segment that has a different privilege level than the CPL.
• Descriptor privilege level (DPL) — The DPL is the privilege level of a segment
or gate. It is stored in the DPL field of the segment or gate descriptor for the
segment or gate. When the currently executing code segment attempts to access
a segment or gate, the DPL of the segment or gate is compared to the CPL and
RPL of the segment or gate selector (as described later in this section). The DPL
is interpreted differently, depending on the type of segment or gate being
accessed:
— Data segment — The DPL indicates the numerically highest privilege level
that a program or task can have to be allowed to access the segment. For
5-10 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
example, if the DPL of a data segment is 1, only programs running at a CPL of
0 or 1 can access the segment.
— Nonconforming code segment (without using a call gate) — The DPL
indicates the privilege level that a program or task must be at to access the
segment. For example, if the DPL of a nonconforming code segment is 0, only
programs running at a CPL of 0 can access the segment.
— Call gate — The DPL indicates the numerically highest privilege level that the
currently executing program or task can be at and still be able to access the
call gate. (This is the same access rule as for a data segment.)
— Conforming code segment and nonconforming code segment
accessed through a call gate — The DPL indicates the numerically lowest
privilege level that a program or task can have to be allowed to access the
segment. For example, if the DPL of a conforming code segment is 2,
programs running at a CPL of 0 or 1 cannot access the segment.
— TSS — The DPL indicates the numerically highest privilege level that the
currently executing program or task can be at and still be able to access the
TSS. (This is the same access rule as for a data segment.)
• Requested privilege level (RPL) — The RPL is an override privilege level that
is assigned to segment selectors. It is stored in bits 0 and 1 of the segment
selector. The processor checks the RPL along with the CPL to determine if access
segment has sufficient privilege to access the segment, access is denied if the
RPL is not of sufficient privilege level. That is, if the RPL of a segment selector is
numerically greater than the CPL, the RPL overrides the CPL, and vice versa. The
RPL can be used to insure that privileged code does not access a segment on
behalf of an application program unless the program itself has access privileges
for that segment. See Section 5.10.4, “Checking Caller Access Privileges (ARPL
Instruction),” for a detailed description of the purpose and typical use of the RPL.
Privilege levels are checked when the segment selector of a segment descriptor is
loaded into a segment register. The checks used for data access differ from those
used for transfers of program control among code segments; therefore, the two
kinds of accesses are considered separately in the following sections.
5.6
PRIVILEGE LEVEL CHECKING WHEN ACCESSING DATA
SEGMENTS
To access operands in a data segment, the segment selector for the data segment
must be loaded into the data-segment registers (DS, ES, FS, or GS) or into the stack-
segment register (SS). (Segment registers can be loaded with the MOV, POP, LDS,
LES, LFS, LGS, and LSS instructions.) Before the processor loads a segment selector
into a segment register, it performs a privilege check (see Figure 5-4) by comparing
the privilege levels of the currently running program or task (the CPL), the RPL of the
segment selector, and the DPL of the segment’s segment descriptor. The processor
Vol. 3 5-11
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
loads the segment selector into the segment register if the DPL is numerically greater
than or equal to both the CPL and the RPL. Otherwise, a general-protection fault is
generated and the segment register is not loaded.
CS Register
CPL
Segment Selector
For Data Segment
RPL
Privilege
Check
Data-Segment Descriptor
DPL
Figure 5-4. Privilege Check for Data Access
Figure 5-5 shows four procedures (located in codes segments A, B, C, and D), each
running at different privilege levels and each attempting to access the same data
segment.
1. The procedure in code segment A is able to access data segment E using segment
selector E1, because the CPL of code segment A and the RPL of segment selector
E1 are equal to the DPL of data segment E.
2. The procedure in code segment B is able to access data segment E using segment
selector E2, because the CPL of code segment B and the RPL of segment selector
E2 are both numerically lower than (more privileged) than the DPL of data
segment E. A code segment B procedure can also access data segment E using
segment selector E1.
3. The procedure in code segment C is not able to access data segment E using
segment selector E3 (dotted line), because the CPL of code segment C and the
RPL of segment selector E3 are both numerically greater than (less privileged)
than the DPL of data segment E. Even if a code segment C procedure were to use
segment selector E1 or E2, such that the RPL would be acceptable, it still could
not access data segment E because its CPL is not privileged enough.
4. The procedure in code segment D should be able to access data segment E
because code segment D’s CPL is numerically less than the DPL of data segment
E. However, the RPL of segment selector E3 (which the code segment D
procedure is using to access data segment E) is numerically greater than the DPL
of data segment E, so access is not allowed. If the code segment D procedure
were to use segment selector E1 or E2 to access the data segment, access would
be allowed.
5-12 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
Code
Segment Sel. E3
RPL=3
Segment C
CPL=3
3
2
1
0
Lowest Privilege
Code
Segment A
CPL=2
Data
Segment Sel. E1
RPL=2
Segment E
DPL=2
Code
Segment B
Segment Sel. E2
RPL=1
CPL=1
Code
Segment D
CPL=0
Highest Privilege
Figure 5-5. Examples of Accessing Data Segments From Various Privilege Levels
As demonstrated in the previous examples, the addressable domain of a program or
task varies as its CPL changes. When the CPL is 0, data segments at all privilege
levels are accessible; when the CPL is 1, only data segments at privilege levels 1
through 3 are accessible; when the CPL is 3, only data segments at privilege level 3
are accessible.
The RPL of a segment selector can always override the addressable domain of a
program or task. When properly used, RPLs can prevent problems caused by acci-
dental (or intensional) use of segment selectors for privileged data segments by less
privileged programs or procedures.
software control. For example, an application program running at a CPL of 3 can set
the RPL for a data- segment selector to 0. With the RPL set to 0, only the CPL checks,
not the RPL checks, will provide protection against deliberate, direct attempts to
violate privilege-level security for the data segment. To prevent these types of privi-
lege-level-check violations, a program or procedure can check access privileges
whenever it receives a data-segment selector from another procedure (see Section
5.10.4, “Checking Caller Access Privileges (ARPL Instruction)”).
5.6.1
Accessing Data in Code Segments
In some instances it may be desirable to access data structures that are contained in
a code segment. The following methods of accessing data in code segments are
possible:
Vol. 3 5-13
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
• Load a data-segment register with a segment selector for a nonconforming,
readable, code segment.
• Load a data-segment register with a segment selector for a conforming,
readable, code segment.
• Use a code-segment override prefix (CS) to read a readable, code segment
whose selector is already loaded in the CS register.
The same rules for accessing data segments apply to method 1. Method 2 is always
valid because the privilege level of a conforming code segment is effectively the
same as the CPL, regardless of its DPL. Method 3 is always valid because the DPL of
the code segment selected by the CS register is the same as the CPL.
5.7
PRIVILEGE LEVEL CHECKING WHEN LOADING THE SS
REGISTER
Privilege level checking also occurs when the SS register is loaded with the segment
selector for a stack segment. Here all privilege levels related to the stack segment
must match the CPL; that is, the CPL, the RPL of the stack-segment selector, and the
DPL of the stack-segment descriptor must be the same. If the RPL and DPL are not
equal to the CPL, a general-protection exception (#GP) is generated.
5.8
PRIVILEGE LEVEL CHECKING WHEN TRANSFERRING
PROGRAM CONTROL BETWEEN CODE SEGMENTS
To transfer program control from one code segment to another, the segment selector
for the destination code segment must be loaded into the code-segment register
(CS). As part of this loading process, the processor examines the segment descriptor
for the destination code segment and performs various limit, type, and privilege
transferred to the new code segment, and program execution begins at the instruc-
tion pointed to by the EIP register.
Program control transfers are carried out with the JMP, CALL, RET, SYSENTER,
SYSEXIT, INT n, and IRET instructions, as well as by the exception and interrupt
mechanisms. Exceptions, interrupts, and the IRET instruction are special cases
discussed in Chapter 6, “Interrupt and Exception Handling.” This chapter discusses
only the JMP, CALL, RET, SYSENTER, and SYSEXIT instructions.
A JMP or CALL instruction can reference another code segment in any of four ways:
• The target operand contains the segment selector for the target code segment.
• The target operand points to a call-gate descriptor, which contains the segment
selector for the target code segment.
5-14 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
• The target operand points to a TSS, which contains the segment selector for the
target code segment.
• The target operand points to a task gate, which points to a TSS, which in turn
contains the segment selector for the target code segment.
Switching,” for information on transferring program control through a task gate
and/or TSS.
The SYSENTER and SYSEXIT instructions are special instructions for making fast calls
to and returns from operating system or executive procedures. These instructions
are discussed briefly in Section 5.8.7, “Performing Fast Calls to System Procedures
with the SYSENTER and SYSEXIT Instructions.”
5.8.1
Direct Calls or Jumps to Code Segments
within the current code segment, so privilege-level checks are not performed. The far
forms of the JMP, CALL, and RET instructions transfer control to other code segments,
so the processor does perform privilege-level checks.
When transferring program control to another code segment without going through a
call gate, the processor examines four kinds of privilege level and type information
(see Figure 5-6):
• The CPL. (Here, the CPL is the privilege level of the calling code segment; that is,
the code segment that contains the procedure that is making the call or jump.)
CS Register
CPL
Segment Selector
For Code Segment
RPL
Destination Code
Segment Descriptor
Privilege
Check
DPL
C
Figure 5-6. Privilege Check for Control Transfer Without Using a Gate
• The DPL of the segment descriptor for the destination code segment that
contains the called procedure.
Vol. 3 5-15
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
• The RPL of the segment selector of the destination code segment.
• The conforming (C) flag in the segment descriptor for the destination code
segment, which determines whether the segment is a conforming (C flag is set)
or nonconforming (C flag is clear) code segment. See Section 3.4.5.1, “Code-
and Data-Segment Descriptor Types,” for more information about this flag.
The rules that the processor uses to check the CPL, RPL, and DPL depends on the
setting of the C flag, as described in the following sections.
5.8.1.1
Accessing Nonconforming Code Segments
When accessing nonconforming code segments, the CPL of the calling procedure
must be equal to the DPL of the destination code segment; otherwise, the processor
generates a general-protection exception (#GP). For example in Figure 5-7:
• Code segment C is a nonconforming code segment. A procedure in code segment
A can call a procedure in code segment C (using segment selector C1) because
they are at the same privilege level (CPL of code segment A is equal to the DPL of
code segment C).
• A procedure in code segment B cannot call a procedure in code segment C (using
segment selector C2 or C1) because the two code segments are at different
privilege levels.
5-16 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
Segment Sel. D2
RPL=3
Code
Segment B
Segment Sel. C2
RPL=3
CPL=3
3
2
1
0
Lowest Privilege
Code
Segment Sel. C1
RPL=2
Code
Segment A
Segment C
DPL=2
Segment Sel. D1
RPL=2
CPL=2
Nonconforming
Code Segment
Code
Segment D
DPL=1
Conforming
Code Segment
Highest Privilege
From Various Privilege Levels
The RPL of the segment selector that points to a nonconforming code segment has a
limited effect on the privilege check. The RPL must be numerically less than or equal
to the CPL of the calling procedure for a successful control transfer to occur. So, in the
example in Figure 5-7, the RPLs of segment selectors C1 and C2 could legally be set
to 0, 1, or 2, but not to 3.
When the segment selector of a nonconforming code segment is loaded into the CS
register, the privilege level field is not changed; that is, it remains at the CPL (which
is the privilege level of the calling procedure). This is true, even if the RPL of the
segment selector is different from the CPL.
5.8.1.2
Accessing Conforming Code Segments
When accessing conforming code segments, the CPL of the calling procedure may be
numerically equal to or greater than (less privileged) the DPL of the destination code
segment; the processor generates a general-protection exception (#GP) only if the
CPL is less than the DPL. (The segment selector RPL for the destination code segment
is not checked if the segment is a conforming code segment.)
Vol. 3 5-17
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
In the example in Figure 5-7, code segment D is a conforming code segment. There-
fore, calling procedures in both code segment A and B can access code segment D
(using either segment selector D1 or D2, respectively), because they both have CPLs
that are greater than or equal to the DPL of the conforming code segment. For
conforming code segments, the DPL represents the numerically lowest priv-
ilege level that a calling procedure may be at to successfully make a call to
the code segment.
(Note that segments selectors D1 and D2 are identical except for their respective
RPLs. But since RPLs are not checked when accessing conforming code segments,
the two segment selectors are essentially interchangeable.)
When program control is transferred to a conforming code segment, the CPL does not
change, even if the DPL of the destination code segment is less than the CPL. This
situation is the only one where the CPL may be different from the DPL of the current
code segment. Also, since the CPL does not change, no stack switch occurs.
Conforming segments are used for code modules such as math libraries and excep-
tion handlers, which support applications but do not require access to protected
system facilities. These modules are part of the operating system or executive soft-
ware, but they can be executed at numerically higher privilege levels (less privileged
levels). Keeping the CPL at the level of a calling code segment when switching to a
conforming code segment prevents an application program from accessing noncon-
forming code segments while at the privilege level (DPL) of a conforming code
segment and thus prevents it from accessing more privileged data.
Most code segments are nonconforming. For these segments, program control can
be transferred only to code segments at the same level of privilege, unless the
transfer is carried out through a call gate, as described in the following sections.
5.8.2
Gate Descriptors
To provide controlled access to code segments with different privilege levels, the
processor provides special set of descriptors called gate descriptors. There are four
kinds of gate descriptors:
• Call gates
• Trap gates
• Interrupt gates
• Task gates
Task gates are used for task switching and are discussed in Chapter 7, “Task Manage-
ment”. Trap and interrupt gates are special kinds of call gates used for calling excep-
tion and interrupt handlers. The are described in Chapter 6, “Interrupt and Exception
Handling.” This chapter is concerned only with call gates.
5-18 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
5.8.3
Call Gates
Call gates facilitate controlled transfers of program control between different privi-
lege levels. They are typically used only in operating systems or executives that use
the privilege-level protection mechanism. Call gates are also useful for transferring
program control between 16-bit and 32-bit code segments, as described in Section
18.4, “Transferring Control Among Mixed-Size Code Segments.”
Figure 5-8 shows the format of a call-gate descriptor. A call-gate descriptor may
reside in the GDT or in an LDT, but not in the interrupt descriptor table (IDT). It
performs six functions:
• It specifies the code segment to be accessed.
• It defines an entry point for a procedure in the specified code segment.
• It specifies the privilege level required for a caller trying to access the procedure.
31
16
15 14 13 12 11
D
0
8
7
6
0
5
4
Type
Param.
Count
P
0
0
Offset in Segment 31:16
P
L
4
0
0
1
1
0
0
31
16
15
0
Segment Selector
Offset in Segment 15:00
DPL Descriptor Privilege Level
P
Gate Valid
Figure 5-8. Call-Gate Descriptor
• If a stack switch occurs, it specifies the number of optional parameters to be
copied between stacks.
• It defines the size of values to be pushed onto the target stack: 16-bit gates force
16-bit pushes and 32-bit gates force 32-bit pushes.
• It specifies whether the call-gate descriptor is valid.
The segment selector field in a call gate specifies the code segment to be accessed.
The offset field specifies the entry point in the code segment. This entry point is
generally to the first instruction of a specific procedure. The DPL field indicates the
privilege level of the call gate, which in turn is the privilege level required to access
the selected procedure through the gate. The P flag indicates whether the call-gate
descriptor is valid. (The presence of the code segment to which the gate points is
indicated by the P flag in the code segment’s descriptor.) The parameter count field
indicates the number of parameters to copy from the calling procedures stack to the
new stack if a stack switch occurs (see Section 5.8.5, “Stack Switching”). The param-
eter count specifies the number of words for 16-bit call gates and doublewords for
32-bit call gates.
Vol. 3 5-19
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
Note that the P flag in a gate descriptor is normally always set to 1. If it is set to 0, a
not present (#NP) exception is generated when a program attempts to access the
descriptor. The operating system can use the P flag for special purposes. For
example, it could be used to track the number of times the gate is used. Here, the P
flag is initially set to 0 causing a trap to the not-present exception handler. The
exception handler then increments a counter and sets the P flag to 1, so that on
5.8.3.1
IA-32e Mode Call Gates
Call-gate descriptors in 32-bit mode provide a 32-bit offset for the instruction pointer
(EIP); 64-bit extensions double the size of 32-bit mode call gates in order to store
64-bit instruction pointers (RIP). See Figure 5-9:
• The first eight bytes (bytes 7:0) of a 64-bit mode call gate are similar but not
identical to legacy 32-bit mode call gates. The parameter-copy-count field has
been removed.
• Bytes 11:8 hold the upper 32 bits of the target-segment offset in canonical form.
A general-protection exception (#GP) is generated if software attempts to use a
call gate with a target offset that is not in canonical form.
• 16-byte descriptors may reside in the same descriptor table with 16-bit and
32-bit descriptors. A type field, used for consistency checking, is defined in bits
12:8 of the 64-bit descriptor’s highest dword (cleared to zero). A general-
protection exception (#GP) results if an attempt is made to access the upper half
of a 64-bit mode descriptor as a 32-bit mode descriptor.
5-20 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
31
31
13 12 11 10 9 8 7
Type
0
0
Reserved
Reserved
16
8
0
0
0
0
0
Offset in Segment 63:31
31
31
8
7
0
0
16 15 14 13 12 11
D
Type
.
0
P
Offset in Segment 31:16
P
L
4
0
0
1
1
0
0
16
15
Segment Selector
Offset in Segment 15:00
DPL Descriptor Privilege Level
P
Gate Valid
Figure 5-9. Call-Gate Descriptor in IA-32e Mode
• Target code segments referenced by a 64-bit call gate must be 64-bit code
segments (CS.L = 1, CS.D = 0). If not, the reference generates a general-
protection exception, #GP (CS selector).
• Only 64-bit mode call gates can be referenced in IA-32e mode (64-bit mode and
compatibility mode). The legacy 32-bit mode call gate type (0CH) is redefined in
IA-32e mode as a 64-bit call-gate type; no 32-bit call-gate type exists in IA-32e
mode.
• If a far call references a 16-bit call gate type (04H) in IA-32e mode, a general-
protection exception (#GP) is generated.
When a call references a 64-bit mode call gate, actions taken are identical to those
taken in 32-bit mode, with the following exceptions:
• Stack pushes are made in eight-byte increments.
• A 64-bit RIP is pushed onto the stack.
• Parameter copying is not performed.
Use a matching far-return instruction size for correct operation (returns from 64-bit
calls must be performed with a 64-bit operand-size return to process the stack
correctly).
Vol. 3 5-21
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
5.8.4
Accessing a Code Segment Through a Call Gate
To access a call gate, a far pointer to the gate is provided as a target operand in a
CALL or JMP instruction. The segment selector from this pointer identifies the call
gate (see Figure 5-10); the offset from the pointer is required, but not used or
When the processor has accessed the call gate, it uses the segment selector from the
call gate to locate the segment descriptor for the destination code segment. (This
segment descriptor can be in the GDT or the LDT.) It then combines the base address
from the code-segment descriptor with the offset from the call gate to form the linear
address of the procedure entry point in the code segment.
As shown in Figure 5-11, four different privilege levels are used to check the validity
of a program control transfer through a call gate:
• The CPL (current privilege level).
• The RPL (requestor's privilege level) of the call gate’s selector.
• The DPL (descriptor privilege level) of the call gate descriptor.
• The DPL of the segment descriptor of the destination code segment.
The C flag (conforming) in the segment descriptor for the destination code segment
is also checked.
Far Pointer to Call Gate
Segment Selector
Offset
Required but not used by processor
Descriptor Table
Offset
Call-Gate
Descriptor
Segment Selector
Offset
Base
Base
Base
Code-Segment
Descriptor
+
Procedure
Entry Point
Figure 5-10. Call-Gate Mechanism
5-22 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
CS Register
CPL
Call-Gate Selector
RPL
Call Gate (Descriptor)
DPL
Privilege
Check
Destination Code-
Segment Descriptor
DPL
Figure 5-11. Privilege Check for Control Transfer with Call Gate
The privilege checking rules are different depending on whether the control transfer
was initiated with a CALL or a JMP instruction, as shown in Table 5-1.
Table 5-1. Privilege Check Rules for Call Gates
Instruction
Privilege Check Rules
CALL
CPL ≤ call gate DPL; RPL ≤ call gate DPL
Destination conforming code segment DPL ≤ CPL
Destination nonconforming code segment DPL ≤ CPL
CPL ≤ call gate DPL; RPL ≤ call gate DPL
Destination conforming code segment DPL ≤ CPL
Destination nonconforming code segment DPL = CPL
JMP
The DPL field of the call-gate descriptor specifies the numerically highest privilege
level from which a calling procedure can access the call gate; that is, to access a call
gate, the CPL of a calling procedure must be equal to or less than the DPL of the call
gate. For example, in Figure 5-15, call gate A has a DPL of 3. So calling procedures at
all CPLs (0 through 3) can access this call gate, which includes calling procedures in
code segments A, B, and C. Call gate B has a DPL of 2, so only calling procedures at
a CPL or 0, 1, or 2 can access call gate B, which includes calling procedures in code
Vol. 3 5-23
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
segments B and C. The dotted line shows that a calling procedure in code segment A
cannot access call gate B.
The RPL of the segment selector to a call gate must satisfy the same test as the CPL
of the calling procedure; that is, the RPL must be less than or equal to the DPL of the
call gate. In the example in Figure 5-15, a calling procedure in code segment C can
access call gate B using gate selector B2 or B1, but it could not use gate selector B3
to access call gate B.
If the privilege checks between the calling procedure and call gate are successful, the
processor then checks the DPL of the code-segment descriptor against the CPL of the
calling procedure. Here, the privilege check rules vary between CALL and JMP
instructions. Only CALL instructions can use call gates to transfer program control to
more privileged (numerically lower privilege level) nonconforming code segments;
that is, to nonconforming code segments with a DPL less than the CPL. A JMP instruc-
tion can use a call gate only to transfer program control to a nonconforming code
segment with a DPL equal to the CPL. CALL and JMP instruction can both transfer
program control to a more privileged conforming code segment; that is, to a
conforming code segment with a DPL less than or equal to the CPL.
If a call is made to a more privileged (numerically lower privilege level) noncon-
forming destination code segment, the CPL is lowered to the DPL of the destination
code segment and a stack switch occurs (see Section 5.8.5, “Stack Switching”). If a
call or jump is made to a more privileged conforming destination code segment, the
CPL is not changed and no stack switch occurs.
5-24 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
Gate Selector A
RPL=3
Call
Gate A
Code
Segment A
DPL=3
Gate Selector B3
RPL=3
CPL=3
3
2
1
0
Lowest Privilege
Code
Segment B
Call
Gate Selector B1
RPL=2
Gate B
CPL=2
DPL=2
Code
Segment C
Gate Selector B2
RPL=1
CPL=1
No Stack
Switch Occurs
Stack Switch
Occurs
Code
Code
Segment E
Segment D
DPL=0
DPL=0
Conforming
Code Segment
Nonconforming
Code Segment
Highest Privilege
Figure 5-12. Example of Accessing Call Gates At Various Privilege Levels
Call gates allow a single code segment to have procedures that can be accessed at
different privilege levels. For example, an operating system located in a code
segment may have some services which are intended to be used by both the oper-
ating system and application software (such as procedures for handling character
I/O). Call gates for these procedures can be set up that allow access at all privilege
levels (0 through 3). More privileged call gates (with DPLs of 0 or 1) can then be set
up for other operating system services that are intended to be used only by the oper-
ating system (such as procedures that initialize device drivers).
5.8.5
Stack Switching
Whenever a call gate is used to transfer program control to a more privileged
nonconforming code segment (that is, when the DPL of the nonconforming destina-
tion code segment is less than the CPL), the processor automatically switches to the
stack for the destination code segment’s privilege level. This stack switching is
carried out to prevent more privileged procedures from crashing due to insufficient
stack space. It also prevents less privileged procedures from interfering (by accident
or intent) with more privileged procedures through a shared stack.
Vol. 3 5-25
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
Each task must define up to 4 stacks: one for applications code (running at privilege
level 3) and one for each of the privilege levels 2, 1, and 0 that are used. (If only two
privilege levels are used [3 and 0], then only two stacks must be defined.) Each of
selector and an offset into the stack segment (a stack pointer).
The segment selector and stack pointer for the privilege level 3 stack is located in the
SS and ESP registers, respectively, when privilege-level-3 code is being executed and
is automatically stored on the called procedure’s stack when a stack switch occurs.
Pointers to the privilege level 0, 1, and 2 stacks are stored in the TSS for the currently
running task (see Figure 7-2). Each of these pointers consists of a segment selector
and a stack pointer (loaded into the ESP register). These initial pointers are strictly
read-only values. The processor does not change them while the task is running.
They are used only to create new stacks when calls are made to more privileged
levels (numerically lower privilege levels). These stacks are disposed of when a
return is made from the called procedure. The next time the procedure is called, a
new stack is created using the initial stack pointer. (The TSS does not specify a stack
for privilege level 3 because the processor does not allow a transfer of program
control from a procedure running at a CPL of 0, 1, or 2 to a procedure running at a
CPL of 3, except on a return.)
The operating system is responsible for creating stacks and stack-segment descrip-
tors for all the privilege levels to be used and for loading initial pointers for these
stacks into the TSS. Each stack must be read/write accessible (as specified in the
type field of its segment descriptor) and must contain enough space (as specified in
the limit field) to hold the following items:
• The contents of the SS, ESP, CS, and EIP registers for the calling procedure.
• The parameters and temporary variables required by the called procedure.
• The EFLAGS register and error code, when implicit calls are made to an exception
or interrupt handler.
The stack will need to require enough space to contain many frames of these items,
because procedures often call other procedures, and an operating system may
support nesting of multiple interrupts. Each stack should be large enough to allow for
the worst case nesting scenario at its privilege level.
(If the operating system does not use the processor’s multitasking mechanism, it still
must create at least one TSS for this stack-related purpose.)
When a procedure call through a call gate results in a change in privilege level, the
processor performs the following steps to switch stacks and begin execution of the
called procedure at a new privilege level:
1. Uses the DPL of the destination code segment (the new CPL) to select a pointer
to the new stack (segment selector and stack pointer) from the TSS.
2. Reads the segment selector and stack pointer for the stack to be switched to from
the current TSS. Any limit violations detected while reading the stack-segment
selector, stack pointer, or stack-segment descriptor cause an invalid TSS (#TS)
exception to be generated.
5-26 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
3. Checks the stack-segment descriptor for the proper privileges and type and
4. Temporarily saves the current values of the SS and ESP registers.
5. Loads the segment selector and stack pointer for the new stack in the SS and ESP
registers.
6. Pushes the temporarily saved values for the SS and ESP registers (for the calling
procedure) onto the new stack (see Figure 5-13).
7. Copies the number of parameter specified in the parameter count field of the call
gate from the calling procedure’s stack to the new stack. If the count is 0, no
parameters are copied.
8. Pushes the return instruction pointer (the current contents of the CS and EIP
registers) onto the new stack.
9. Loads the segment selector for the new code segment and the new instruction
pointer from the call gate into the CS and EIP registers, respectively, and begins
execution of the called procedure.
See the description of the CALL instruction in Chapter 3, Instruction Set Reference, in
the IA-32 Intel Architecture Software Developer’s Manual, Volume 2, for a detailed
description of the privilege level checks and other protection checks that the
processor performs on a far call through a call gate.
Calling Procedure’s Stack
Called Procedure’s Stack
Calling SS
Calling ESP
Parameter 1
Parameter 2
Parameter 3
Calling CS
Parameter 1
Parameter 2
Parameter 3
ESP
ESP
Calling EIP
Figure 5-13. Stack Switching During an Interprivilege-Level Call
The parameter count field in a call gate specifies the number of data items (up to 31)
that the processor should copy from the calling procedure’s stack to the stack of the
called procedure. If more than 31 data items need to be passed to the called proce-
Vol. 3 5-27
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
dure, one of the parameters can be a pointer to a data structure, or the saved
contents of the SS and ESP registers may be used to access parameters in the old
stack space. The size of the data items passed to the called procedure depends on
the call gate size, as described in Section 5.8.3, “Call Gates.”
5.8.5.1
Stack Switching in 64-bit Mode
Although protection-check rules for call gates are unchanged from 32-bit mode,
stack-switch changes in 64-bit mode are different.
When stacks are switched as part of a 64-bit mode privilege-level change through a
call gate, a new SS (stack segment) descriptor is not loaded; 64-bit mode only loads
an inner-level RSP from the TSS. The new SS is forced to NULL and the SS selector’s
RPL field is forced to the new CPL. The new SS is set to NULL in order to handle
nested far transfers (CALLF, INTn, interrupts and exceptions). The old SS and RSP
are saved on the new stack.
On a subsequent RETF, the old SS is popped from the stack and loaded into the SS
register. See Table 5-2.
Table 5-2. 64-Bit-Mode Stack Layout After CALLF with CPL Change
32-bit Mode
IA-32e mode
Old SS Selector
Old RSP
Old SS Selector
+12
+8
+4
0
+24
+16
+8
Old ESP
CS Selector
EIP
Old CS Selector
RIP
ESP
RSP
0
< 4 Bytes >
< 8 Bytes >
In 64-bit mode, stack operations resulting from a privilege-level-changing far call or
far return are eight-bytes wide and change the RSP by eight. The mode does not
support the automatic parameter-copy feature found in 32-bit mode. The call-gate
count field is ignored. Software can access the old stack, if necessary, by referencing
the old stack-segment selector and stack pointer saved on the new process stack.
In 64-bit mode, RETF is allowed to load a NULL SS under certain conditions. If the
target mode is 64-bit mode and the target CPL< >3, IRET allows SS to be loaded with
a NULL selector. If the called procedure itself is interrupted, the NULL SS is pushed on
the stack frame. On the subsequent RETF, the NULL SS on the stack acts as a flag to
tell the processor not to load a new SS descriptor.
5.8.6
Returning from a Called Procedure
The RET instruction can be used to perform a near return, a far return at the same
privilege level, and a far return to a different privilege level. This instruction is
5-28 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
intended to execute returns from procedures that were called with a CALL instruc-
tion. It does not support returns from a JMP instruction, because the JMP instruction
does not save a return instruction pointer on the stack.
A near return only transfers program control within the current code segment; there-
fore, the processor performs only a limit check. When the processor pops the return
instruction pointer from the stack into the EIP register, it checks that the pointer does
not exceed the limit of the current code segment.
On a far return at the same privilege level, the processor pops both a segment
selector for the code segment being returned to and a return instruction pointer from
the stack. Under normal conditions, these pointers should be valid, because they
privilege checks to detect situations where the current procedure might have altered
the pointer or failed to maintain the stack properly.
A far return that requires a privilege-level change is only allowed when returning to a
less privileged level (that is, the DPL of the return code segment is numerically
greater than the CPL). The processor uses the RPL field from the CS register value
saved for the calling procedure (see Figure 5-13) to determine if a return to a numer-
ically higher privilege level is required. If the RPL is numerically greater (less privi-
leged) than the CPL, a return across privilege levels occurs.
The processor performs the following steps when performing a far return to a calling
procedure (see Figures 6-2 and 6-4 in the Intel® 64 and IA-32 Architectures Soft-
ware Developer’s Manual, Volume 1, for an illustration of the stack contents prior to
and after a return):
1. Checks the RPL field of the saved CS register value to determine if a privilege
level change is required on the return.
2. Loads the CS and EIP registers with the values on the called procedure’s stack.
(Type and privilege level checks are performed on the code-segment descriptor
and RPL of the code- segment selector.)
3. (If the RET instruction includes a parameter count operand and the return
requires a privilege level change.) Adds the parameter count (in bytes obtained
from the RET instruction) to the current ESP register value (after popping the CS
and EIP values), to step past the parameters on the called procedure’s stack. The
resulting value in the ESP register points to the saved SS and ESP values for the
calling procedure’s stack. (Note that the byte count in the RET instruction must
be chosen to match the parameter count in the call gate that the calling
procedure referenced when it made the original call multiplied by the size of the
parameters.)
4. (If the return requires a privilege level change.) Loads the SS and ESP registers
with the saved SS and ESP values and switches back to the calling procedure’s
stack. The SS and ESP values for the called procedure’s stack are discarded. Any
limit violations detected while loading the stack-segment selector or stack
pointer cause a general-protection exception (#GP) to be generated. The new
stack-segment descriptor is also checked for type and privilege violations.
Vol. 3 5-29
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
5. (If the RET instruction includes a parameter count operand.) Adds the parameter
count (in bytes obtained from the RET instruction) to the current ESP register
value, to step past the parameters on the calling procedure’s stack. The resulting
ESP value is not checked against the limit of the stack segment. If the ESP value
is beyond the limit, that fact is not recognized until the next stack operation.
6. (If the return requires a privilege level change.) Checks the contents of the DS,
ES, FS, and GS segment registers. If any of these registers refer to segments
whose DPL is less than the new CPL (excluding conforming code segments), the
segment register is loaded with a null segment selector.
See the description of the RET instruction in Chapter 4 of the Intel® 64 and IA-32
Architectures Software Developer’s Manual, Volume 2B, for a detailed description of
the privilege level checks and other protection checks that the processor performs on
a far return.
5.8.7
Performing Fast Calls to System Procedures with the
SYSENTER and SYSEXIT Instructions
The SYSENTER and SYSEXIT instructions were introduced into the IA-32 architecture
in the Pentium II processors for the purpose of providing a fast (low overhead) mech-
anism for calling operating system or executive procedures. SYSENTER is intended
for use by user code running at privilege level 3 to access operating system or exec-
utive procedures running at privilege level 0. SYSEXIT is intended for use by privilege
level 0 operating system or executive procedures for fast returns to privilege level 3
user code. SYSENTER can be executed from privilege levels 3, 2, 1, or 0; SYSEXIT
can only be executed from privilege level 0.
The SYSENTER and SYSEXIT instructions are companion instructions, but they do not
constitute a call/return pair. This is because SYSENTER does not save any state infor-
mation for use by SYSEXIT on a return.
The target instruction and stack pointer for these instructions are not specified
through instruction operands. Instead, they are specified through parameters
entered in MSRs and general-purpose registers.
For SYSENTER, target fields are generated using the following sources:
• Target code segment — Reads this from IA32_SYSENTER_CS.
• Target instruction — Reads this from IA32_SYSENTER_EIP.
• Stack segment — Computed by adding 8 to the value in IA32_SYSENTER_CS.
• Stack pointer — Reads this from the IA32_SYSENTER_ESP.
For SYSEXIT, target fields are generated using the following sources:
• Target code segment — Computed by adding 16 to the value in the
IA32_SYSENTER_CS.
• Target instruction — Reads this from EDX.
5-30 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
• Stack segment — Computed by adding 24 to the value in IA32_SYSENTER_CS.
• Stack pointer — Reads this from ECX.
The SYSENTER and SYSEXIT instructions preform “fast” calls and returns because
they force the processor into a predefined privilege level 0 state when SYSENTER is
executed and into a predefined privilege level 3 state when SYSEXIT is executed. By
forcing predefined and consistent processor states, the number of privilege checks
ordinarily required to perform a far call to another privilege levels are greatly
reduced. Also, by predefining the target context state in MSRs and general-purpose
registers eliminates all memory accesses except when fetching the target code.
Any additional state that needs to be saved to allow a return to the calling procedure
must be saved explicitly by the calling procedure or be predefined through program-
ming conventions.
5.8.7.1
SYSENTER and SYSEXIT Instructions in IA-32e Mode
For Intel 64 processors, the SYSENTER and SYSEXIT instructions are enhanced to
allow fast system calls from user code running at privilege level 3 (in compatibility
mode or 64-bit mode) to 64-bit executive procedures running at privilege level 0.
IA32_SYSENTER_EIP MSR and IA32_SYSENTER_ESP MSR are expanded to hold
64-bit addresses. If IA-32e mode is inactive, only the lower 32-bit addresses stored
in these MSRs are used. If 64-bit mode is active, addresses stored in
IA32_SYSENTER_EIP and IA32_SYSENTER_ESP must be canonical. Note that, in
64-bit mode, IA32_SYSENTER_CS must not contain a NULL selector.
When SYSENTER transfers control, the following fields are generated and bits set:
• Target code segment — Reads non-NULL selector from IA32_SYSENTER_CS.
• New CS attributes — CS base = 0, CS limit = FFFFFFFFH.
• Target instruction — Reads 64-bit canonical address from
IA32_SYSENTER_EIP.
• Stack segment — Computed by adding 8 to the value from
IA32_SYSENTER_CS.
• Stack pointer — Reads 64-bit canonical address from IA32_SYSENTER_ESP.
• New SS attributes — SS base = 0, SS limit = FFFFFFFFH.
When the SYSEXIT instruction transfers control to 64-bit mode user code using
REX.W, the following fields are generated and bits set:
• Target code segment — Computed by adding 32 to the value in
IA32_SYSENTER_CS.
• New CS attributes — L-bit = 1 (go to 64-bit mode).
• Target instruction — Reads 64-bit canonical address in RDX.
• Stack segment — Computed by adding 40 to the value of IA32_SYSENTER_CS.
• Stack pointer — Update RSP using 64-bit canonical address in RCX.
Vol. 3 5-31
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
When SYSEXIT transfers control to compatibility mode user code when the operand
size attribute is 32 bits, the following fields are generated and bits set:
• Target code segment — Computed by adding 16 to the value in
IA32_SYSENTER_CS.
• New CS attributes — L-bit = 0 (go to compatibility mode).
• Target instruction — Fetch the target instruction from 32-bit address in EDX.
• Stack segment — Computed by adding 24 to the value in IA32_SYSENTER_CS.
• Stack pointer — Update ESP from 32-bit address in ECX.
5.8.8
Fast System Calls in 64-bit Mode
The SYSCALL and SYSRET instructions are designed for operating systems that use a
flat memory model (segmentation is not used). The instructions, along with
SYSENTER and SYSEXIT, are suited for IA-32e mode operation. SYSCALL and
SYSRET, however, are not supported in compatibility mode. Use CPUID to check if
SYSCALL and SYSRET are available (CPUID.80000001H.EDX[bit 11] = 1).
SYSCALL is intended for use by user code running at privilege level 3 to access oper-
ating system or executive procedures running at privilege level 0. SYSRET is
intended for use by privilege level 0 operating system or executive procedures for
fast returns to privilege level 3 user code.
Stack pointers for SYSCALL/SYSRET are not specified through model specific regis-
ters. The clearing of bits in RFLAGS is programmable rather than fixed.
SYSCALL/SYSRET save and restore the RFLAGS register.
For SYSCALL, the processor saves the RIP of the instruction in RCX and gets the priv-
ilege level 0 target instruction and stack pointer from:
• Target code segment — Reads a non-NULL selector from IA32_STAR[47:32].
• Target instruction — Reads a 64-bit canonical address from IA32_LSTAR.
• Stack segment — Computed by adding 8 to the value in IA32_STAR[47:32].
• System flags — The processor uses a mask derived from IA32_FMASK to
perform a logical-AND operation with the lower 32-bits of RFLAGS. The result is
saved into R11. The mask is the complement of the value supplied by privileged
executives using the IA32_FMASK MSR.
When SYSRET transfers control to 64-bit mode user code using REX.W, the processor
gets the privilege level 3 target instruction and stack pointer from:
• Target code segment — Reads a non-NULL selector from IA32_STAR[63:48] +
16.
• Target instruction — Copies the value in RCX into RIP.
• Stack segment — IA32_STAR[63:48] + 8.
• EFLAGS — Loaded from R11.
5-32 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
When SYSRET transfers control to 32-bit mode user code using a 32-bit operand size,
the processor gets the privilege level 3 target instruction and stack pointer from:
• Target code segment — Reads a non-NULL selector from IA32_STAR[63:48].
• Target instruction — Copies the value in ECX into EIP.
• Stack segment — IA32_STAR[63:48] + 8.
• EFLAGS — Loaded from R11.
It is the responsibility of the OS to ensure the descriptors in the GDT/LDT correspond
to the selectors loaded by SYSCALL/SYSRET (consistent with the base, limit, and
attribute values forced by the instructions).
Any address written to IA32_LSTAR is first checked by WRMSR to ensure canonical
form. If an address is not canonical, an exception is generated (#GP).
See Figure 5-14 for the layout of IA32_STAR, IA32_LSTAR and IA32_FMASK.
63
0
32 31
SYSCALL EFLAGS Mask
Reserved
IA32_FMASK
63
0
Target RIP for 64-bit Mode Calling Program
IA32_LSTAR
63
32
31
0
48 47
SYSCALL CS and SS
SYSRET CS and SS
Reserved
IA32_STAR
Figure 5-14. MSRs Used by SYSCALL and SYSRET
5.9
PRIVILEGED INSTRUCTIONS
Some of the system instructions (called “privileged instructions”) are protected from
use by application programs. The privileged instructions control system functions
(such as the loading of system registers). They can be executed only when the CPL is
0 (most privileged). If one of these instructions is executed when the CPL is not 0, a
Vol. 3 5-33
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
general-protection exception (#GP) is generated. The following system instructions
are privileged instructions:
• LGDT — Load GDT register.
• LLDT — Load LDT register.
• LTR — Load task register.
• LIDT — Load IDT register.
• MOV (control registers) — Load and store control registers.
• LMSW — Load machine status word.
• CLTS — Clear task-switched flag in register CR0.
• MOV (debug registers) — Load and store debug registers.
• INVD — Invalidate cache, without writeback.
• WBINVD — Invalidate cache, with writeback.
• INVLPG —Invalidate TLB entry.
• HLT— Halt processor.
• RDMSR — Read Model-Specific Registers.
• RDPMC — Read Performance-Monitoring Counter.
• RDTSC — Read Time-Stamp Counter.
Some of the privileged instructions are available only in the more recent families of
Intel 64 and IA-32 processors (see Section 19.13, “New Instructions In the Pentium
and Later IA-32 Processors”).
The PCE and TSD flags in register CR4 (bits 4 and 2, respectively) enable the RDPMC
and RDTSC instructions, respectively, to be executed at any CPL.
5.10
POINTER VALIDATION
When operating in protected mode, the processor validates all pointers to enforce
protection between segments and maintain isolation between privilege levels.
Pointer validation consists of the following checks:
1. Checking access rights to determine if the segment type is compatible with its
use.
2. Checking read/write rights.
3. Checking if the pointer offset exceeds the segment limit.
4. Checking if the supplier of the pointer is allowed to access the segment.
5. Checking the offset alignment.
5-34 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
The processor automatically performs first, second, and third checks during instruc-
tion execution. Software must explicitly request the fourth check by issuing an ARPL
instruction. The fifth check (offset alignment) is performed automatically at privilege
level 3 if alignment checking is turned on. Offset alignment does not affect isolation
of privilege levels.
5.10.1
Checking Access Rights (LAR Instruction)
When the processor accesses a segment using a far pointer, it performs an access
rights check on the segment descriptor pointed to by the far pointer. This check is
performed to determine if type and privilege level (DPL) of the segment descriptor
are compatible with the operation to be performed. For example, when making a far
call in protected mode, the segment-descriptor type must be for a conforming or
nonconforming code segment, a call gate, a task gate, or a TSS. Then, if the call is to
a nonconforming code segment, the DPL of the code segment must be equal to the
CPL, and the RPL of the code segment’s segment selector must be less than or equal
to the DPL. If type or privilege level are found to be incompatible, the appropriate
exception is generated.
To prevent type incompatibility exceptions from being generated, software can check
the access rights of a segment descriptor using the LAR (load access rights) instruc-
tion. The LAR instruction specifies the segment selector for the segment descriptor
whose access rights are to be checked and a destination register. The instruction then
performs the following operations:
1. Check that the segment selector is not null.
2. Checks that the segment selector points to a segment descriptor that is within
the descriptor table limit (GDT or LDT).
3. Checks that the segment descriptor is a code, data, LDT, call gate, task gate, or
TSS segment-descriptor type.
4. If the segment is not a conforming code segment, checks if the segment
descriptor is visible at the CPL (that is, if the CPL and the RPL of the segment
selector are less than or equal to the DPL).
5. If the privilege level and type checks pass, loads the second doubleword of the
segment descriptor into the destination register (masked by the value
00FXFF00H, where X indicates that the corresponding 4 bits are undefined) and
sets the ZF flag in the EFLAGS register. If the segment selector is not visible at
the current privilege level or is an invalid type for the LAR instruction, the
instruction does not modify the destination register and clears the ZF flag.
Once loaded in the destination register, software can preform additional checks on
the access rights information.
Vol. 3 5-35
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
5.10.2
Checking Read/Write Rights (VERR and VERW Instructions)
When the processor accesses any code or data segment it checks the read/write priv-
ileges assigned to the segment to verify that the intended read or write operation is
allowed. Software can check read/write rights using the VERR (verify for reading)
and VERW (verify for writing) instructions. Both these instructions specify the
segment selector for the segment being checked. The instructions then perform the
following operations:
1. Check that the segment selector is not null.
2. Checks that the segment selector points to a segment descriptor that is within
the descriptor table limit (GDT or LDT).
3. Checks that the segment descriptor is a code or data-segment descriptor type.
4. If the segment is not a conforming code segment, checks if the segment
descriptor is visible at the CPL (that is, if the CPL and the RPL of the segment
selector are less than or equal to the DPL).
5. Checks that the segment is readable (for the VERR instruction) or writable (for
the VERW) instruction.
The VERR instruction sets the ZF flag in the EFLAGS register if the segment is visible
at the CPL and readable; the VERW sets the ZF flag if the segment is visible and writ-
able. (Code segments are never writable.) The ZF flag is cleared if any of these
checks fail.
5.10.3
Checking That the Pointer Offset Is Within Limits (LSL
Instruction)
When the processor accesses any segment it performs a limit check to insure that the
offset is within the limit of the segment. Software can perform this limit check using
the LSL (load segment limit) instruction. Like the LAR instruction, the LSL instruction
specifies the segment selector for the segment descriptor whose limit is to be
checked and a destination register. The instruction then performs the following oper-
ations:
1. Check that the segment selector is not null.
2. Checks that the segment selector points to a segment descriptor that is within
the descriptor table limit (GDT or LDT).
3. Checks that the segment descriptor is a code, data, LDT, or TSS segment-
descriptor type.
4. If the segment is not a conforming code segment, checks if the segment
descriptor is visible at the CPL (that is, if the CPL and the RPL of the segment
selector less than or equal to the DPL).
5. If the privilege level and type checks pass, loads the unscrambled limit (the limit
scaled according to the setting of the G flag in the segment descriptor) into the
5-36 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
destination register and sets the ZF flag in the EFLAGS register. If the segment
selector is not visible at the current privilege level or is an invalid type for the LSL
instruction, the instruction does not modify the destination register and clears
the ZF flag.
Once loaded in the destination register, software can compare the segment limit with
the offset of a pointer.
5.10.4
Checking Caller Access Privileges (ARPL Instruction)
The requestor’s privilege level (RPL) field of a segment selector is intended to carry
the privilege level of a calling procedure (the calling procedure’s CPL) to a called
procedure. The called procedure then uses the RPL to determine if access to a
segment is allowed. The RPL is said to “weaken” the privilege level of the called
procedure to that of the RPL.
Operating-system procedures typically use the RPL to prevent less privileged appli-
cation programs from accessing data located in more privileged segments. When an
operating-system procedure (the called procedure) receives a segment selector from
an application program (the calling procedure), it sets the segment selector’s RPL to
the privilege level of the calling procedure. Then, when the operating system uses
ilege checks using the calling procedure’s privilege level (stored in the RPL) rather
than the numerically lower privilege level (the CPL) of the operating-system proce-
dure. The RPL thus insures that the operating system does not access a segment on
behalf of an application program unless that program itself has access to the
segment.
Figure 5-15 shows an example of how the processor uses the RPL field. In this
example, an application program (located in code segment A) possesses a segment
selector (segment selector D1) that points to a privileged data structure (that is, a
data structure located in a data segment D at privilege level 0).
The application program cannot access data segment D, because it does not have
sufficient privilege, but the operating system (located in code segment C) can. So, in
an attempt to access data segment D, the application program executes a call to the
operating system and passes segment selector D1 to the operating system as a
parameter on the stack. Before passing the segment selector, the (well behaved)
application program sets the RPL of the segment selector to its current privilege level
(which in this example is 3). If the operating system attempts to access data
segment D using segment selector D1, the processor compares the CPL (which is
now 0 following the call), the RPL of segment selector D1, and the DPL of data
segment D (which is 0). Since the RPL is greater than the DPL, access to data
segment D is denied. The processor’s protection mechanism thus protects data
segment D from access by the operating system, because application program’s priv-
ilege level (represented by the RPL of segment selector B) is greater than the DPL of
data segment D.
Vol. 3 5-37
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
Passed as a
parameter on
the stack.
Application Program
Code
Call
Gate Selector B
RPL=3
Segment Sel. D1
RPL=3
Segment A
Gate B
CPL=3
DPL=3
3
2
1
0
Lowest Privilege
Access
not
allowed
Code
Data
Segment D
Operating
System
Segment C
Segment Sel. D2
RPL=0
Access
allowed
DPL=0
DPL=0
Highest Privilege
Figure 5-15. Use of RPL to Weaken Privilege Level of Called Procedure
Now assume that instead of setting the RPL of the segment selector to 3, the appli-
cation program sets the RPL to 0 (segment selector D2). The operating system can
now access data segment D, because its CPL and the RPL of segment selector D2 are
both equal to the DPL of data segment D.
Because the application program is able to change the RPL of a segment selector to
any value, it can potentially use a procedure operating at a numerically lower privi-
lege level to access a protected data structure. This ability to lower the RPL of a
segment selector breaches the processor’s protection mechanism.
Because a called procedure cannot rely on the calling procedure to set the RPL
correctly, operating-system procedures (executing at numerically lower privilege-
levels) that receive segment selectors from numerically higher privilege-level proce-
dures need to test the RPL of the segment selector to determine if it is at the appro-
priate level. The ARPL (adjust requested privilege level) instruction is provided for
this purpose. This instruction adjusts the RPL of one segment selector to match that
of another segment selector.
5-38 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
The example in Figure 5-15 demonstrates how the ARPL instruction is intended to be
used. When the operating-system receives segment selector D2 from the application
program, it uses the ARPL instruction to compare the RPL of the segment selector
with the privilege level of the application program (represented by the code-segment
selector pushed onto the stack). If the RPL is less than application program’s privi-
lege level, the ARPL instruction changes the RPL of the segment selector to match the
privilege level of the application program (segment selector D1). Using this instruc-
tion thus prevents a procedure running at a numerically higher privilege level from
accessing numerically lower privilege-level (more privileged) segments by lowering
the RPL of a segment selector.
Note that the privilege level of the application program can be determined by reading
the RPL field of the segment selector for the application-program’s code segment.
This segment selector is stored on the stack as part of the call to the operating
system. The operating system can copy the segment selector from the stack into a
register for use as an operand for the ARPL instruction.
5.10.5
Checking Alignment
When the CPL is 3, alignment of memory references can be checked by setting the
AM flag in the CR0 register and the AC flag in the EFLAGS register. Unaligned memory
references generate alignment exceptions (#AC). The processor does not generate
alignment exceptions when operating at privilege level 0, 1, or 2. See Table 6-7 for a
description of the alignment requirements when alignment checking is enabled.
5.11
PAGE-LEVEL PROTECTION
Page-level protection can be used alone or applied to segments. When page-level
protection is used with the flat memory model, it allows supervisor code and data
(the operating system or executive) to be protected from user code and data (appli-
cation programs). It also allows pages containing code to be write protected. When
the segment- and page-level protection are combined, page-level read/write protec-
tion allows more protection granularity within segments.
With page-level protection (as with segment-level protection) each memory refer-
ence is checked to verify that protection checks are satisfied. All checks are made
before the memory cycle is started, and any violation prevents the cycle from
starting and results in a page-fault exception being generated. Because checks are
The processor performs two page-level protection checks:
• Restriction of addressable domain (supervisor and user modes).
• Page type (read only or read/write).
Violations of either of these checks results in a page-fault exception being generated.
See Chapter 6, “Interrupt 14—Page-Fault Exception (#PF),” for an explanation of the
Vol. 3 5-39
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
which lead to page-fault exceptions.
5.11.1
Page-Protection Flags
Protection information for pages is contained in two flags in a paging-structure entry
(see Chapter 4): the read/write flag (bit 1) and the user/supervisor flag (bit 2). The
protection checks use the flags in all paging structures.
5.11.2
Restricting Addressable Domain
The page-level protection mechanism allows restricting access to pages based on
two privilege levels:
• Supervisor mode (U/S flag is 0)—(Most privileged) For the operating system or
executive, other system software (such as device drivers), and protected system
data (such as page tables).
The segment privilege levels map to the page privilege levels as follows. If the
processor is currently operating at a CPL of 0, 1, or 2, it is in supervisor mode; if it is
operating at a CPL of 3, it is in user mode. When the processor is in supervisor mode,
it can access all pages; when in user mode, it can access only user-level pages. (Note
that the WP flag in control register CR0 modifies the supervisor permissions, as
described in Section 5.11.3, “Page Type.”)
Note that to use the page-level protection mechanism, code and data segments must
be set up for at least two segment-based privilege levels: level 0 for supervisor code
and data segments and level 3 for user code and data segments. (In this model, the
stacks are placed in the data segments.) To minimize the use of segments, a flat
memory model can be used (see Section 3.2.1, “Basic Flat Model”).
Here, the user and supervisor code and data segments all begin at address zero in
the linear address space and overlay each other. With this arrangement, operating-
system code (running at the supervisor level) and application code (running at the
user level) can execute as if there are no segments. Protection between operating-
system and application code and data is provided by the processor’s page-level
protection mechanism.
5.11.3
Page Type
The page-level protection mechanism recognizes two page types:
• Read-only access (R/W flag is 0).
• Read/write access (R/W flag is 1).
5-40 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
When the processor is in supervisor mode and the WP flag in register CR0 is clear (its
state following reset initialization), all pages are both readable and writable (write-
protection is ignored). When the processor is in user mode, it can write only to user-
mode pages that are read/write accessible. User-mode pages which are read/write or
read-only are readable; supervisor-mode pages are neither readable nor writable
from user mode. A page-fault exception is generated on any attempt to violate the
protection rules.
Starting with the P6 family, Intel processors allow user-mode pages to be write-
protected against supervisor-mode access. Setting CR0.WP = 1 enables supervisor-
mode sensitivity to write protected pages. If CR0.WP = 1, read-only pages are not
writable from any privilege level. This supervisor write-protect feature is useful for
implementing a “copy-on-write” strategy used by some operating systems, such as
UNIX*, for task creation (also called forking or spawning). When a new task is
created, it is possible to copy the entire address space of the parent task. This gives
the child task a complete, duplicate set of the parent's segments and pages. An alter-
native copy-on-write strategy saves memory space and time by mapping the child's
segments and pages to the same segments and pages used by the parent task. A
private copy of a page gets created only when one of the tasks writes to the page. By
using the WP flag and marking the shared pages as read-only, the supervisor can
detect an attempt to write to a page, and can copy the page at that time.
5.11.4
Combining Protection of Both Levels of Page Tables
For any one page, the protection attributes of its page-directory entry (first-level
page table) may differ from those of its page-table entry (second-level page table).
The processor checks the protection for a page in both its page-directory and the
page-table entries. Table 5-3 shows the protection provided by the possible combina-
tions of protection attributes when the WP flag is clear.
5.11.5
Overrides to Page Protection
The following types of memory accesses are checked as if they are privilege-level 0
accesses, regardless of the CPL at which the processor is currently operating:
• Access to segment descriptors in the GDT, LDT, or IDT.
• Access to an inner-privilege-level stack during an inter-privilege-level call or a
call to in exception or interrupt handler, when a change of privilege level occurs.
5.12
COMBINING PAGE AND SEGMENT PROTECTION
When paging is enabled, the processor evaluates segment protection first, then
evaluates page protection. If the processor detects a protection violation at either
the segment level or the page level, the memory access is not carried out and an
Vol. 3 5-41
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
exception is generated. If an exception is generated by segmentation, no paging
exception is generated.
Page-level protections cannot be used to override segment-level protection. For
example, a code segment is by definition not writable. If a code segment is paged,
setting the R/W flag for the pages to read-write does not make the pages writable.
Attempts to write into the pages will be blocked by segment-level protection checks.
Page-level protection can be used to enhance segment-level protection. For
example, if a large read-write data segment is paged, the page-protection mecha-
nism can be used to write-protect individual pages.
Table 5-3. Combined Page-Directory and Page-Table Protection
Page-Directory Entry
Page-Table Entry
Combined Effect
Privilege
User
Access Type
Privilege
User
Access Type
Privilege
User
Access Type
Read-Only
Read-Only
Read-Write
Read-Write
Read-Only
Read-Only
Read-Write
Read-Write
Read-Only
Read-Only
Read-Write
Read-Write
Read-Only
Read-Only
Read-Write
Read-Write
Read-Only
Read-Write
Read-Only
Read-Write
Read-Only
Read-Write
Read-Only
Read-Write
Read-Only
Read-Write
Read-Only
Read-Write
Read-Only
Read-Write
Read-Only
Read-Write
Read-Only
User
User
User
Read-Only
User
User
User
Read-Only
User
User
User
Read/Write
Read/Write*
Read/Write*
Read/Write*
Read/Write
Read/Write*
Read/Write*
Read/Write*
Read/Write
Read/Write*
Read/Write*
Read/Write*
Read/Write
User
Supervisor
Supervisor
Supervisor
Supervisor
User
Supervisor
Supervisor
Supervisor
Supervisor
Supervisor
Supervisor
Supervisor
Supervisor
Supervisor
Supervisor
Supervisor
Supervisor
User
User
User
Supervisor
Supervisor
Supervisor
Supervisor
Supervisor
Supervisor
Supervisor
Supervisor
NOTE:
User
User
User
Supervisor
Supervisor
Supervisor
Supervisor
* If CR0.WP = 1, access type is determined by the R/W flags of the page-directory and page-table
entries. IF CR0.WP = 0, supervisor privilege permits read-write access.
5-42 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
5.13
PAGE-LEVEL PROTECTION AND EXECUTE-DISABLE
BIT
In addition to page-level protection offered by the U/S and R/W flags, paging struc-
tures used with PAE paging and IA-32e paging (see Chapter 4) provide the execute-
disable bit. This bit offers additional protection for data pages.
An Intel 64 or IA-32 processor with the execute-disable bit capability can prevent
data pages from being used by malicious software to execute code. This capability is
provided in:
• 32-bit protected mode with PAE enabled.
• IA-32e mode.
While the execute-disable bit capability does not introduce new instructions, it does
require operating systems to use a PAE-enabled environment and establish a page-
granular protection policy for memory pages.
If the execute-disable bit of a memory page is set, that page can be used only as
data. An attempt to execute code from a memory page with the execute-disable bit
set causes a page-fault exception.
The execute-disable capability is supported only with PAE paging and IA-32e paging.
It is not supported with 32-bit paging. Existing page-level protection mechanisms
(see Section 5.11, “Page-Level Protection”) continue to apply to memory pages inde-
pendent of the execute-disable setting.
5.13.1
Detecting and Enabling the Execute-Disable Capability
instruction. CPUID.80000001H:EDX.NX [bit 20] = 1 indicates the capability is avail-
able.
If the capability is available, software can enable it by setting IA32_EFER.NXE[bit 11]
to 1. IA32_EFER is available if CPUID.80000001H.EDX[bit 20 or 29] = 1.
If the execute-disable capability is not available, a write to set IA32_EFER.NXE
produces a #GP exception. See Table 5-4.
Table 5-4. Extended Feature Enable MSR (IA32_EFER)
63:12
11
10
IA-32e mode Reserve IA-32e mode Reserve SysCallenable
active (LMA) enable (LME) (SCE)
9
8
7:1
0
Reserved Execute-
disable bit
d
d
enable (NXE)
Vol. 3 5-43
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
5.13.2
Execute-Disable Page Protection
The execute-disable bit in the paging structures enhances page protection for data
pages. Instructions cannot be fetched from a memory page if IA32_EFER.NXE =1
and the execute-disable bit is set in any of the paging-structure entries used to map
the page. Table 5-5 lists the valid usage of a page in relation to the value of execute-
disable bit (bit 63) of the corresponding entry in each level of the paging structures.
Execute-disable protection can be activated using the execute-disable bit at any level
of the paging structure, irrespective of the corresponding entry in other levels. When
execute-disable protection is not activated, the page can be used as code or data.
Table 5-5. IA-32e Mode Page Level Protection Matrix
with Execute-Disable Bit Capability
Execute Disable Bit Value (Bit 63)
Valid Usage
PML4
PDP
PDE
PTE
Bit 63 = 1
*
*
*
Data
*
*
*
Bit 63 = 1
*
*
Data
*
*
Bit 63 = 1
*
*
Data
Bit 63 = 1
Bit 63 = 0
Data
Bit 63 = 0 Bit 63 = 0 Bit 63 = 0
Data/Code
NOTES:
* Value not checked.
In legacy PAE-enabled mode, Table 5-6 and Table 5-7 show the effect of setting the
execute-disable bit for code and data pages.
5-44 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
Table 5-6. Legacy PAE-Enabled 4-KByte Page Level Protection Matrix
with Execute-Disable Bit Capability
Execute Disable Bit Value (Bit 63) Valid Usage
PDE
PTE
*
Bit 63 = 1
*
Data
Bit 63 = 1
Bit 63 = 0
Data
Bit 63 = 0
NOTE:
Data/Code
* Value not checked.
Table 5-7. Legacy PAE-Enabled 2-MByte Page Level Protection
with Execute-Disable Bit Capability
Execute Disable Bit Value (Bit 63) Valid Usage
PDE
Bit 63 = 1
Bit 63 = 0
Data
Data/Code
5.13.3
Reserved Bit Checking
The processor enforces reserved bit checking in paging data structure entries. The
bits being checked varies with paging mode and may vary with the size of physical
address space.
Table 5-8 shows the reserved bits that are checked when the execute disable bit
capability is enabled (CR4.PAE = 1 and IA32_EFER.NXE = 1). Table 5-8 and Table
show the following paging modes:
• Non-PAE 4-KByte paging: 4-KByte-page only paging (CR4.PAE = 0,
CR4.PSE = 0).
• PSE36: 4-KByte and 4-MByte pages (CR4.PAE = 0, CR4.PSE = 1).
• PAE: 4-KByte and 2-MByte pages (CR4.PAE = 1, CR4.PSE = X).
The reserved bit checking depends on the physical address size supported by the
implementation, which is reported in CPUID.80000008H. See the table note.
Vol. 3 5-45
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
Table 5-8. IA-32e Mode Page Level Protection Matrix with Execute-Disable Bit
Capability Enabled
Check Bits
Mode
Paging Mode
32-bit
4-KByte paging (non-PAE)
PSE36 - PDE, 4-MByte page
PSE36 - PDE, 4-KByte page
PSE36 - PTE
No reserved bits checked
Bit [21]
No reserved bits checked
No reserved bits checked
Bits [63:MAXPHYADDR] & [8:5] & [2:1] *
Bits [62:MAXPHYADDR] & [20:13] *
Bits [62:MAXPHYADDR] *
Bits [62:MAXPHYADDR] *
Bits [51:MAXPHYADDR] *
Bits [51:MAXPHYADDR] *
Bits [51:MAXPHYADDR] & [20:13] *
Bits [51:MAXPHYADDR] *
Bits [51:MAXPHYADDR] *
PAE - PDP table entry
PAE - PDE, 2-MByte page
PAE - PDE, 4-KByte page
PAE - PTE
64-bit
PML4E
PDPTE
PDE, 2-MByte page
PDE, 4-KByte page
PTE
NOTES:
* MAXPHYADDR is the maximum physical address size and is indicated by
CPUID.80000008H:EAX[bits 7-0].
If execute disable bit capability is not enabled or not available, reserved bit checking
in 64-bit mode includes bit 63 and additional bits. This and reserved bit checking for
legacy 32-bit paging modes are shown in Table 5-10.
5-46 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
Table 5-9. Reserved Bit Checking WIth Execute-Disable Bit Capability Not Enabled
Mode
Paging Mode
Check Bits
32-bit
KByte paging (non-PAE)
PSE36 - PDE, 4-MByte page
PSE36 - PDE, 4-KByte page
PSE36 - PTE
No reserved bits checked
Bit [21]
No reserved bits checked
No reserved bits checked
PAE - PDP table entry
PAE - PDE, 2-MByte page
PAE - PDE, 4-KByte page
PAE - PTE
Bits [63:MAXPHYADDR] & [8:5] & [2:1]*
Bits [63:MAXPHYADDR] & [20:13]*
Bits [63:MAXPHYADDR]*
Bits [63:MAXPHYADDR]*
64-bit
PML4E
Bit [63], bits [51:MAXPHYADDR]*
Bit [63], bits [51:MAXPHYADDR]*
Bit [63], bits [51:MAXPHYADDR] & [20:13]*
Bit [63], bits [51:MAXPHYADDR]*
Bit [63], bits [51:MAXPHYADDR]*
PDPTE
PDE, 2-MByte page
PDE, 4-KByte page
PTE
NOTES:
* MAXPHYADDR is the maximum physical address size and is indicated by
CPUID.80000008H:EAX[bits 7-0].
5.13.4
Exception Handling
When execute disable bit capability is enabled (IA32_EFER.NXE = 1), conditions for
a page fault to occur include the same conditions that apply to an Intel 64 or IA-32
processor without execute disable bit capability plus the following new condition: an
instruction fetch to a linear address that translates to physical address in a memory
page that has the execute-disable bit set.
An Execute Disable Bit page fault can occur at all privilege levels. It can occur on any
instruction fetch, including (but not limited to): near branches, far branches,
CALL/RET/INT/IRET execution, sequential instruction fetches, and task switches. The
execute-disable bit in the page translation mechanism is checked only when:
• IA32_EFER.NXE = 1.
• The instruction translation look-aside buffer (ITLB) is loaded with a page that is
not already present in the ITLB.
Vol. 3 5-47
Download from Www.Somanuals.com. All Manuals Search And Download.
PROTECTION
5-48 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
CHAPTER 6
INTERRUPT AND EXCEPTION HANDLING
This chapter describes the interrupt and exception-handling mechanism when oper-
ating in protected mode on an Intel 64 or IA-32 processor. Most of the information
provided here also applies to interrupt and exception mechanisms used in real-
address, virtual-8086 mode, and 64-bit mode.
Chapter 17, “8086 Emulation,” describes information specific to interrupt and excep-
tion mechanisms in real-address and virtual-8086 mode. Section 6.14, “Exception
and Interrupt Handling in 64-bit Mode,” describes information specific to interrupt
and exception mechanisms in IA-32e mode and 64-bit sub-mode.
6.1
INTERRUPT AND EXCEPTION OVERVIEW
Interrupts and exceptions are events that indicate that a condition exists somewhere
in the system, the processor, or within the currently executing program or task that
requires the attention of a processor. They typically result in a forced transfer of
execution from the currently running program or task to a special software routine or
task called an interrupt handler or an exception handler. The action taken by a
processor in response to an interrupt or exception is referred to as servicing or
handling the interrupt or exception.
Interrupts occur at random times during the execution of a program, in response to
signals from hardware. System hardware uses interrupts to handle events external
to the processor, such as requests to service peripheral devices. Software can also
generate interrupts by executing the INT n instruction.
Exceptions occur when the processor detects an error condition while executing an
instruction, such as division by zero. The processor detects a variety of error condi-
tions including protection violations, page faults, and internal machine faults. The
machine-check architecture of the Pentium 4, Intel Xeon, P6 family, and Pentium
processors also permits a machine-check exception to be generated when internal
hardware errors and bus errors are detected.
When an interrupt is received or an exception is detected, the currently running
procedure or task is suspended while the processor executes an interrupt or excep-
tion handler. When execution of the handler is complete, the processor resumes
execution of the interrupted procedure or task. The resumption of the interrupted
procedure or task happens without loss of program continuity, unless recovery from
an exception was not possible or an interrupt caused the currently running program
to be terminated.
This chapter describes the processor’s interrupt and exception-handling mechanism,
when operating in protected mode. A description of the exceptions and the conditions
that cause them to be generated is given at the end of this chapter.
Vol. 3 6-1
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
6.2
To aid in handling exceptions and interrupts, each architecturally defined exception
and each interrupt condition requiring special handling by the processor is assigned
a unique identification number, called a vector. The processor uses the vector
assigned to an exception or interrupt as an index into the interrupt descriptor table
(IDT). The table provides the entry point to an exception or interrupt handler (see
Section 6.10, “Interrupt Descriptor Table (IDT)”).
The allowable range for vector numbers is 0 to 255. Vectors in the range 0 through
31 are reserved by the Intel 64 and IA-32 architectures for architecture-defined
exceptions and interrupts. Not all of the vectors in this range have a currently defined
function. The unassigned vectors in this range are reserved. Do not use the reserved
vectors.
Vectors in the range 32 to 255 are designated as user-defined interrupts and are not
assigned to external I/O devices to enable those devices to send interrupts to the
processor through one of the external hardware interrupt mechanisms (see Section
6.3, “Sources of Interrupts”).
Table 6-1 shows vector assignments for architecturally defined exceptions and for the
NMI interrupt. This table gives the exception type (see Section 6.5, “Exception Clas-
sifications”) and indicates whether an error code is saved on the stack for the excep-
tion. The source of each predefined exception and the NMI interrupt is also given.
6.3
SOURCES OF INTERRUPTS
The processor receives interrupts from two sources:
• External (hardware generated) interrupts.
• Software-generated interrupts.
6.3.1
External Interrupts
External interrupts are received through pins on the processor or through the local
APIC. The primary interrupt pins on Pentium 4, Intel Xeon, P6 family, and Pentium
processors are the LINT[1:0] pins, which are connected to the local APIC (see
Chapter 10, “Advanced Programmable Interrupt Controller (APIC)”). When the local
APIC is enabled, the LINT[1:0] pins can be programmed through the APIC’s local
vector table (LVT) to be associated with any of the processor’s exception or interrupt
vectors.
When the local APIC is global/hardware disabled, these pins are configured as INTR
and NMI pins, respectively. Asserting the INTR pin signals the processor that an
external interrupt has occurred. The processor reads from the system bus the inter-
rupt vector number provided by an external interrupt controller, such as an 8259A
6-2 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
(see Section 6.2, “Exception and Interrupt Vectors”). Asserting the NMI pin signals a
non-maskable interrupt (NMI), which is assigned to interrupt vector 2.
Table 6-1. Protected-Mode Exceptions and Interrupts
Vector Mne- Description
Type
Error Source
Code
No.
monic
0
#DE
Divide Error
RESERVED
Fault
No
No
DIV and IDIV instructions.
1
#DB
Fault/
Trap
For Intel use only.
2
—
NMI Interrupt
Interrupt No
Nonmaskable external
interrupt.
3
4
5
6
#BP
#OF
#BR
#UD
Breakpoint
Trap
Trap
Fault
No
No
No
No
INT 3 instruction.
Overflow
INTO instruction.
BOUND Range Exceeded
BOUND instruction.
Invalid Opcode (Undefined Fault
Opcode)
UD2 instruction or reserved
1
opcode.
7
8
#NM
#DF
Device Not Available (No
Math Coprocessor)
Fault
No
Floating-point or WAIT/FWAIT
instruction.
Double Fault
Abort
Yes
Any instruction that can
(zero) generate an exception, an NMI,
or an INTR.
2
9
Coprocessor Segment
Overrun (reserved)
Fault
No
Floating-point instruction.
10
11
#TS
#NP
Invalid TSS
Fault
Fault
Yes
Yes
Task switch or TSS access.
Segment Not Present
Loading segment registers or
accessing system segments.
12
13
#SS
#GP
Stack-Segment Fault
General Protection
Page Fault
Fault
Fault
Fault
Yes
Yes
Stack operations and SS
register loads.
Any memory reference and
other protection checks.
14
15
#PF
—
Yes
No
Any memory reference.
(Intel reserved. Do not
use.)
16
17
#MF
#AC
x87 FPU Floating-Point
Error (Math Fault)
Fault
Fault
No
x87 FPU floating-point or
WAIT/FWAIT instruction.
Alignment Check
Yes
Any data reference in
3
(Zero) memory.
Vol. 3 6-3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Table 6-1. Protected-Mode Exceptions and Interrupts (Contd.)
18
#MC
Machine Check
Abort
No
Error codes (if any) and source
are model dependent.
4
19
#XM
SIMD Floating-Point
Exception
Fault
No
SSE/SSE2/SSE3 floating-point
instructions
5
20-31
—
—
Intel reserved. Do not use.
32-
255
User Defined (Non-
reserved) Interrupts
Interrupt
External interrupt or INT n
instruction.
NOTES:
1. The UD2 instruction was introduced in the Pentium Pro processor.
2. Processors after the Intel386 processor do not generate this exception.
3. This exception was introduced in the Intel486 processor.
4. This exception was introduced in the Pentium processor and enhanced in the P6 family proces-
sors.
5. This exception was introduced in the Pentium III processor.
The processor’s local APIC is normally connected to a system-based I/O APIC. Here,
external interrupts received at the I/O APIC’s pins can be directed to the local APIC
through the system bus (Pentium 4, Intel Core Duo, Intel Core 2, Intel Atom, and
Intel Xeon processors) or the APIC serial bus (P6 family and Pentium processors).
The I/O APIC determines the vector number of the interrupt and sends this number
to the local APIC. When a system contains multiple processors, processors can also
send interrupts to one another by means of the system bus (Pentium 4, Intel Core
Duo, Intel Core 2, Intel Atom, and Intel Xeon processors) or the APIC serial bus (P6
family and Pentium processors).
The LINT[1:0] pins are not available on the Intel486 processor and earlier Pentium
processors that do not contain an on-chip local APIC. These processors have dedi-
cated NMI and INTR pins. With these processors, external interrupts are typically
generated by a system-based interrupt controller (8259A), with the interrupts being
signaled through the INTR pin.
Note that several other pins on the processor can cause a processor interrupt to
occur. However, these interrupts are not handled by the interrupt and exception
mechanism described in this chapter. These pins include the RESET#, FLUSH#,
STPCLK#, SMI#, R/S#, and INIT# pins. Whether they are included on a particular
processor is implementation dependent. Pin functions are described in the data
books for the individual processors. The SMI# pin is described in Chapter 26,
“System Management.”
6.3.2
Maskable Hardware Interrupts
Any external interrupt that is delivered to the processor by means of the INTR pin or
through the local APIC is called a maskable hardware interrupt. Maskable hardware
interrupts that can be delivered through the INTR pin include all IA-32 architecture
6-4 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
defined interrupt vectors from 0 through 255; those that can be delivered through
the local APIC include interrupt vectors 16 through 255.
The IF flag in the EFLAGS register permits all maskable hardware interrupts to be
masked as a group (see Section 6.8.1, “Masking Maskable Hardware Interrupts”).
Note that when interrupts 0 through 15 are delivered through the local APIC, the
APIC indicates the receipt of an illegal vector.
6.3.3
Software-Generated Interrupts
The INT n instruction permits interrupts to be generated from within software by
supplying an interrupt vector number as an operand. For example, the INT 35
instruction forces an implicit call to the interrupt handler for interrupt 35.
Any of the interrupt vectors from 0 to 255 can be used as a parameter in this instruc-
tion. If the processor’s predefined NMI vector is used, however, the response of the
processor will not be the same as it would be from an NMI interrupt generated in the
normal manner. If vector number 2 (the NMI vector) is used in this instruction, the
NMI interrupt handler is called, but the processor’s NMI-handling hardware is not
activated.
Interrupts generated in software with the INT n instruction cannot be masked by the
IF flag in the EFLAGS register.
6.4
SOURCES OF EXCEPTIONS
The processor receives exceptions from three sources:
• Processor-detected program-error exceptions.
• Software-generated exceptions.
• Machine-check exceptions.
6.4.1
Program-Error Exceptions
The processor generates one or more exceptions when it detects program errors
during the execution in an application program or the operating system or executive.
Intel 64 and IA-32 architectures define a vector number for each processor-detect-
able exception. Exceptions are classified as faults, traps, and aborts (see Section
6.5, “Exception Classifications”).
Vol. 3 6-5
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
6.4.2
Software-Generated Exceptions
The INTO, INT 3, and BOUND instructions permit exceptions to be generated in soft-
ware. These instructions allow checks for exception conditions to be performed at
points in the instruction stream. For example, INT 3 causes a breakpoint exception to
be generated.
The INT n instruction can be used to emulate exceptions in software; but there is a
limitation. If INT n provides a vector for one of the architecturally-defined excep-
tions, the processor generates an interrupt to the correct vector (to access the
exception handler) but does not push an error code on the stack. This is true even if
the associated hardware-generated exception normally produces an error code. The
exception handler will still attempt to pop an error code from the stack while handling
the exception. Because no error code was pushed, the handler will pop off and
discard the EIP instead (in place of the missing error code). This sends the return to
the wrong location.
check mechanisms for checking the operation of the internal chip hardware and bus
transactions. These mechanisms are implementation dependent. When a machine-
check error is detected, the processor signals a machine-check exception (vector 18)
and returns an error code.
See Chapter 6, “Interrupt 18—Machine-Check Exception (#MC)” and Chapter 15,
“Machine-Check Architecture,” for more information about the machine-check
mechanism.
6.5
EXCEPTION CLASSIFICATIONS
Exceptions are classified as faults, traps, or aborts depending on the way they are
reported and whether the instruction that caused the exception can be restarted
without loss of program or task continuity.
• Faults — A fault is an exception that can generally be corrected and that, once
corrected, allows the program to be restarted with no loss of continuity. When a
fault is reported, the processor restores the machine state to the state prior to
the beginning of execution of the faulting instruction. The return address (saved
contents of the CS and EIP registers) for the fault handler points to the faulting
instruction, rather than to the instruction following the faulting instruction.
• Traps — A trap is an exception that is reported immediately following the
execution of the trapping instruction. Traps allow execution of a program or task
to be continued without loss of program continuity. The return address for the
trap handler points to the instruction to be executed after the trapping
instruction.
6-6 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
• Aborts — An abort is an exception that does not always report the precise
location of the instruction causing the exception and does not allow a restart of
the program or task that caused the exception. Aborts are used to report severe
errors, such as hardware errors and inconsistent or illegal values in system
tables.
NOTE
One exception subset normally reported as a fault is not restartable.
Such exceptions result in loss of some processor state. For example,
executing a POPAD instruction where the stack frame crosses over
the end of the stack segment causes a fault to be reported. In this
situation, the exception handler sees that the instruction pointer
(CS:EIP) has been restored as if the POPAD instruction had not been
executed. However, internal processor state (the general-purpose
registers) will have been modified. Such cases are considered
programming errors. An application causing this class of exceptions
should be terminated by the operating system.
6.6
PROGRAM OR TASK RESTART
To allow the restarting of program or task following the handling of an exception or
an interrupt, all exceptions (except aborts) are guaranteed to report exceptions on
an instruction boundary. All interrupts are guaranteed to be taken on an instruction
boundary.
For fault-class exceptions, the return instruction pointer (saved when the processor
generates an exception) points to the faulting instruction. So, when a program or task
is restarted following the handling of a fault, the faulting instruction is restarted (re-
executed). Restarting the faulting instruction is commonly used to handle exceptions
that are generated when access to an operand is blocked. The most common example
of this type of fault is a page-fault exception (#PF) that occurs when a program or
task references an operand located on a page that is not in memory. When a page-
fault exception occurs, the exception handler can load the page into memory and
resume execution of the program or task by restarting the faulting instruction. To
insure that the restart is handled transparently to the currently executing program or
task, the processor saves the necessary registers and stack pointers to allow a restart
to the state prior to the execution of the faulting instruction.
For trap-class exceptions, the return instruction pointer points to the instruction
following the trapping instruction. If a trap is detected during an instruction which
transfers execution, the return instruction pointer reflects the transfer. For example,
if a trap is detected while executing a JMP instruction, the return instruction pointer
points to the destination of the JMP instruction, not to the next address past the JMP
instruction. All trap exceptions allow program or task restart with no loss of conti-
nuity. For example, the overflow exception is a trap exception. Here, the return
instruction pointer points to the instruction following the INTO instruction that tested
Vol. 3 6-7
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
EFLAGS.OF (overflow) flag. The trap handler for this exception resolves the overflow
condition. Upon return from the trap handler, program or task execution continues at
the instruction following the INTO instruction.
The abort-class exceptions do not support reliable restarting of the program or task.
Abort handlers are designed to collect diagnostic information about the state of the
processor when the abort exception occurred and then shut down the application and
system as gracefully as possible.
Interrupts rigorously support restarting of interrupted programs and tasks without
loss of continuity. The return instruction pointer saved for an interrupt points to the
next instruction to be executed at the instruction boundary where the processor took
the interrupt. If the instruction just executed has a repeat prefix, the interrupt is
taken at the end of the current iteration with the registers set to execute the next
iteration.
The ability of a P6 family processor to speculatively execute instructions does not
affect the taking of interrupts by the processor. Interrupts are taken at instruction
boundaries located during the retirement phase of instruction execution; so they are
always taken in the “in-order” instruction stream. See Chapter 2, “Intel® 64 and IA-
32 Architectures,” in the Intel® 64 and IA-32 Architectures Software Developer’s
Manual, Volume 1, for more information about the P6 family processors’ microarchi-
tecture and its support for out-of-order instruction execution.
Note that the Pentium processor and earlier IA-32 processors also perform varying
amounts of prefetching and preliminary decoding. With these processors as well,
exceptions and interrupts are not signaled until actual “in-order” execution of the
instructions. For a given code sample, the signaling of exceptions occurs uniformly
when the code is executed on any family of IA-32 processors (except where new
exceptions or new opcodes have been defined).
6.7
NONMASKABLE INTERRUPT (NMI)
The nonmaskable interrupt (NMI) can be generated in either of two ways:
• External hardware asserts the NMI pin.
• The processor receives a message on the system bus (Pentium 4, Intel Core Duo,
family and Pentium processors) with a delivery mode NMI.
When the processor receives a NMI from either of these sources, the processor
handles it immediately by calling the NMI handler pointed to by interrupt vector
number 2. The processor also invokes certain hardware conditions to insure that no
other interrupts, including NMI interrupts, are received until the NMI handler has
completed executing (see Section 6.7.1, “Handling Multiple NMIs”).
Also, when an NMI is received from either of the above sources, it cannot be masked
by the IF flag in the EFLAGS register.
6-8 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
It is possible to issue a maskable hardware interrupt (through the INTR pin) to vector
2 to invoke the NMI interrupt handler; however, this interrupt will not truly be an NMI
interrupt. A true NMI interrupt that activates the processor’s NMI-handling hardware
can only be delivered through one of the mechanisms listed above.
6.7.1
Handling Multiple NMIs
While an NMI interrupt handler is executing, the processor disables additional calls to
the NMI handler until the next IRET instruction is executed. This blocking of subse-
quent NMIs prevents stacking up calls to the NMI handler. It is recommended that the
NMI interrupt handler be accessed through an interrupt gate to disable maskable
hardware interrupts (see Section 6.8.1, “Masking Maskable Hardware Interrupts”). If
the NMI handler is a virtual-8086 task with an IOPL of less than 3, an IRET instruction
issued from the handler generates a general-protection exception (see Section
17.2.7, “Sensitive Instructions”). In this case, the NMI is unmasked before the
general-protection exception handler is invoked.
6.8
ENABLING AND DISABLING INTERRUPTS
The processor inhibits the generation of some interrupts, depending on the state of
following sections.
6.8.1
Masking Maskable Hardware Interrupts
The IF flag can disable the servicing of maskable hardware interrupts received on the
processor’s INTR pin or through the local APIC (see Section 6.3.2, “Maskable Hard-
ware Interrupts”). When the IF flag is clear, the processor inhibits interrupts deliv-
ered to the INTR pin or through the local APIC from generating an internal interrupt
request; when the IF flag is set, interrupts delivered to the INTR or through the local
APIC pin are processed as normal external interrupts.
The IF flag does not affect non-maskable interrupts (NMIs) delivered to the NMI pin
or delivery mode NMI messages delivered through the local APIC, nor does it affect
processor generated exceptions. As with the other flags in the EFLAGS register, the
processor clears the IF flag in response to a hardware reset.
The fact that the group of maskable hardware interrupts includes the reserved inter-
rupt and exception vectors 0 through 32 can potentially cause confusion. Architectur-
ally, when the IF flag is set, an interrupt for any of the vectors from 0 through 32 can
be delivered to the processor through the INTR pin and any of the vectors from 16
through 32 can be delivered through the local APIC. The processor will then generate
an interrupt and call the interrupt or exception handler pointed to by the vector
number. So for example, it is possible to invoke the page-fault handler through the
INTR pin (by means of vector 14); however, this is not a true page-fault exception. It
Vol. 3 6-9
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
is an interrupt. As with the INT n instruction (see Section 6.4.2, “Software-Generated
Exceptions”), when an interrupt is generated through the INTR pin to an exception
vector, the processor does not push an error code on the stack, so the exception
handler may not operate correctly.
(clear interrupt-enable flag) instructions, respectively. These instructions may be
executed only if the CPL is equal to or less than the IOPL. A general-protection excep-
tion (#GP) is generated if they are executed when the CPL is greater than the IOPL.
(The effect of the IOPL on these instructions is modified slightly when the virtual
mode extension is enabled by setting the VME flag in control register CR4: see
Section 17.3, “Interrupt and Exception Handling in Virtual-8086 Mode.” Behavior is
also impacted by the PVI flag: see Section 17.4, “Protected-Mode Virtual Interrupts.”
The IF flag is also affected by the following operations:
• The PUSHF instruction stores all flags on the stack, where they can be examined
and modified. The POPF instruction can be used to load the modified flags back
into the EFLAGS register.
• Task switches and the POPF and IRET instructions load the EFLAGS register;
therefore, they can be used to modify the setting of the IF flag.
• When an interrupt is handled through an interrupt gate, the IF flag is automati-
cally cleared, which disables maskable hardware interrupts. (If an interrupt is
handled through a trap gate, the IF flag is not cleared.)
See the descriptions of the CLI, STI, PUSHF, POPF, and IRET instructions in Chapter
3, “Instruction Set Reference, A-M,” in the Intel® 64 and IA-32 Architectures Soft-
ware Developer’s Manual, Volume 2A, for a detailed description of the operations
6.8.2
Masking Instruction Breakpoints
The RF (resume) flag in the EFLAGS register controls the response of the processor
to instruction-breakpoint conditions (see the description of the RF flag in Section 2.3,
“System Flags and Fields in the EFLAGS Register”).
When set, it prevents an instruction breakpoint from generating a debug exception
(#DB); when clear, instruction breakpoints will generate debug exceptions. The
primary function of the RF flag is to prevent the processor from going into a debug
exception loop on an instruction-breakpoint. See Section 16.3.1.1, “Instruction-
Breakpoint Exception Condition,” for more information on the use of this flag.
6-10 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
6.8.3
Masking Exceptions and Interrupts When Switching Stacks
To switch to a different stack segment, software often uses a pair of instructions, for
example:
MOV SS, AX
MOV ESP, StackTop
If an interrupt or exception occurs after the segment selector has been loaded into
the SS register but before the ESP register has been loaded, these two parts of the
logical address into the stack space are inconsistent for the duration of the interrupt
or exception handler.
To prevent this situation, the processor inhibits interrupts, debug exceptions, and
single-step trap exceptions after either a MOV to SS instruction or a POP to SS
instruction, until the instruction boundary following the next instruction is reached.
All other faults may still be generated. If the LSS instruction is used to modify the
contents of the SS register (which is the recommended method of modifying this
register), this problem does not occur.
6.9
PRIORITY AMONG SIMULTANEOUS EXCEPTIONS AND
INTERRUPTS
If more than one exception or interrupt is pending at an instruction boundary, the
processor services them in a predictable order. Table 6-2 shows the priority among
classes of exception and interrupt sources.
Table 6-2. Priority Among Simultaneous Exceptions and Interrupts
Priority
Description
1 (Highest)
Hardware Reset and Machine Checks
- RESET
- Machine Check
2
3
Trap on Task Switch
- T flag in TSS is set
External Hardware Interventions
- FLUSH
- STOPCLK
- SMI
- INIT
4
Traps on the Previous Instruction
- Breakpoints
- Debug Trap Exceptions (TF flag set or data/I-O breakpoint)
Vol. 3 6-11
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Table 6-2. Priority Among Simultaneous Exceptions and Interrupts (Contd.)
1
5
6
7
8
Nonmaskable Interrupts (NMI)
Maskable Hardware Interrupts
Code Breakpoint Fault
1
Faults from Fetching Next Instruction
- Code-Segment Limit Violation
- Code Page Fault
9
Faults from Decoding the Next Instruction
- Instruction length > 15 bytes
- Invalid Opcode
- Coprocessor Not Available
10 (Lowest)
Faults on Executing an Instruction
- Overflow
- Bound error
- Invalid TSS
- Segment Not Present
- Stack fault
- General Protection
- Data Page Fault
- Alignment Check
- x87 FPU Floating-point exception
- SIMD floating-point exception
NOTE:
1. The Intel486™ processor and earlier processors group nonmaskable and maskable interrupts in
the same priority class.
While priority among these classes listed in Table 6-2 is consistent throughout the
architecture, exceptions within each class are implementation-dependent and may
vary from processor to processor. The processor first services a pending exception or
interrupt from the class which has the highest priority, transferring execution to the
first instruction of the handler. Lower priority exceptions are discarded; lower priority
interrupts are held pending. Discarded exceptions are re-generated when the inter-
rupt handler returns execution to the point in the program or task where the excep-
tions and/or interrupts occurred.
6.10
INTERRUPT DESCRIPTOR TABLE (IDT)
The interrupt descriptor table (IDT) associates each exception or interrupt vector
with a gate descriptor for the procedure or task used to service the associated excep-
tion or interrupt. Like the GDT and LDTs, the IDT is an array of 8-byte descriptors (in
6-12 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
protected mode). Unlike the GDT, the first entry of the IDT may contain a descriptor.
To form an index into the IDT, the processor scales the exception or interrupt vector
by eight (the number of bytes in a gate descriptor). Because there are only 256 inter-
rupt or exception vectors, the IDT need not contain more than 256 descriptors. It can
contain fewer than 256 descriptors, because descriptors are required only for the
interrupt and exception vectors that may occur. All empty descriptor slots in the IDT
should have the present flag for the descriptor set to 0.
The base addresses of the IDT should be aligned on an 8-byte boundary to maximize
performance of cache line fills. The limit value is expressed in bytes and is added to
the base address to get the address of the last valid byte. A limit value of 0 results in
exactly 1 valid byte. Because IDT entries are always eight bytes long, the limit should
always be one less than an integral multiple of eight (that is, 8N – 1).
The IDT may reside anywhere in the linear address space. As shown in Figure 6-1,
the processor locates the IDT using the IDTR register. This register holds both a
32-bit base address and 16-bit limit for the IDT.
The LIDT (load IDT register) and SIDT (store IDT register) instructions load and store
the contents of the IDTR register, respectively. The LIDT instruction loads the IDTR
register with the base address and limit held in a memory operand. This instruction
can be executed only when the CPL is 0. It normally is used by the initialization code
of an operating system when creating an IDT. An operating system also may use it to
change from one IDT to another. The SIDT instruction copies the base and limit value
stored in IDTR to memory. This instruction can be executed at any privilege level.
If a vector references a descriptor beyond the limit of the IDT, a general-protection
exception (#GP) is generated.
NOTE
Because interrupts are delivered to the processor core only once, an
incorrectly configured IDT could result in incomplete interrupt
handling and/or the blocking of interrupt delivery.
IA-32 architecture rules need to be followed for setting up IDTR
base/limit/access fields and each field in the gate descriptors. The
same apply for the Intel 64 architecture. This includes implicit
referencing of the destination code segment through the GDT or LDT
and accessing the stack.
Vol. 3 6-13
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
IDTR Register
47
16 15
0
IDT Base Address
IDT Limit
Interrupt
Descriptor Table (IDT)
+
Gate for
Interrupt #n
(n−1)∗8
Gate for
Interrupt #3
16
8
Gate for
Interrupt #2
Gate for
Interrupt #1
0
31
0
Figure 6-1. Relationship of the IDTR and IDT
6.11
IDT DESCRIPTORS
The IDT may contain any of three kinds of gate descriptors:
• Task-gate descriptor
• Interrupt-gate descriptor
• Trap-gate descriptor
descriptors. The format of a task gate used in an IDT is the same as that of a task
gate used in the GDT or an LDT (see Section 7.2.5, “Task-Gate Descriptor”). The task
gate contains the segment selector for a TSS for an exception and/or interrupt
handler task.
Interrupt and trap gates are very similar to call gates (see Section 5.8.3, “Call
Gates”). They contain a far pointer (segment selector and offset) that the processor
uses to transfer program execution to a handler procedure in an exception- or inter-
rupt-handler code segment. These gates differ in the way the processor handles the
IF flag in the EFLAGS register (see Section 6.12.1.2, “Flag Usage By Exception- or
Interrupt-Handler Procedure”).
6-14 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Task Gate
31
31
16
15 14 13 12
7
0
0
8
D
P
L
P
0
0
1
0
1
4
0
16
15
TSS Segment Selector
Interrupt Gate
31
31
16
15 14 13 12
7
5
4
0
0
8
D
P
L
0
D 1
1
0
Offset 31..16
P
0
0
0
4
0
16
15
Segment Selector
Offset 15..0
Trap Gate
31
31
16
15 14 13 12
5
4
0
0
8
7
D
P
L
0
D 1
1
1
Offset 31..16
P
0
0
0
4
0
16
15
Segment Selector
Offset 15..0
DPL
Offset
P
Descriptor Privilege Level
Offset to procedure entry point
Segment Present flag
Selector
D
Segment Selector for destination code segment
Size of gate: 1 = 32 bits; 0 = 16 bits
Reserved
Figure 6-2. IDT Gate Descriptors
6.12
The processor handles calls to exception- and interrupt-handlers similar to the way it
handles calls with a CALL instruction to a procedure or a task. When responding to an
exception or interrupt, the processor uses the exception or interrupt vector as an
index to a descriptor in the IDT. If the index points to an interrupt gate or trap gate,
the processor calls the exception or interrupt handler in a manner similar to a CALL
to a call gate (see Section 5.8.2, “Gate Descriptors,” through Section 5.8.6,
Vol. 3 6-15
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
“Returning from a Called Procedure”). If index points to a task gate, the processor
executes a task switch to the exception- or interrupt-handler task in a manner similar
to a CALL to a task gate (see Section 7.3, “Task Switching”).
6.12.1
Exception- or Interrupt-Handler Procedures
An interrupt gate or trap gate references an exception- or interrupt-handler proce-
dure that runs in the context of the currently executing task (see Figure 6-3). The
segment selector for the gate points to a segment descriptor for an executable code
segment in either the GDT or the current LDT. The offset field of the gate descriptor
points to the beginning of the exception- or interrupt-handling procedure.
Destination
Code Segment
IDT
Interrupt
Procedure
Offset
Interrupt
Vector
Interrupt or
Trap Gate
+
Segment Selector
GDT or LDT
Base
Address
Segment
Descriptor
Figure 6-3. Interrupt Procedure Call
6-16 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
When the processor performs a call to the exception- or interrupt-handler procedure:
• If the handler procedure is going to be executed at a numerically lower privilege
level, a stack switch occurs. When the stack switch occurs:
a. The segment selector and stack pointer for the stack to be used by the
handler are obtained from the TSS for the currently executing task. On this
new stack, the processor pushes the stack segment selector and stack
pointer of the interrupted procedure.
b. The processor then saves the current state of the EFLAGS, CS, and EIP
registers on the new stack (see Figures 6-4).
c. If an exception causes an error code to be saved, it is pushed on the new
stack after the EIP value.
• If the handler procedure is going to be executed at the same privilege level as the
interrupted procedure:
a. The processor saves the current state of the EFLAGS, CS, and EIP registers
on the current stack (see Figures 6-4).
b. If an exception causes an error code to be saved, it is pushed on the current
stack after the EIP value.
Vol. 3 6-17
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Stack Usage with No
Privilege-Level Change
Interrupted Procedure’s
and Handler’s Stack
ESP Before
Transfer to Handler
EFLAGS
CS
EIP
Error Code
ESP After
Transfer to Handler
Stack Usage with
Privilege-Level Change
Interrupted Procedure’s
Stack
Handler’s Stack
ESP Before
Transfer to Handler
SS
ESP
EFLAGS
CS
EIP
ESP After
Error Code
Transfer to Handler
Figure 6-4. Stack Usage on Transfers to Interrupt and Exception-Handling Routines
To return from an exception- or interrupt-handler procedure, the handler must use
the IRET (or IRETD) instruction. The IRET instruction is similar to the RET instruction
except that it restores the saved flags into the EFLAGS register. The IOPL field of the
EFLAGS register is restored only if the CPL is 0. The IF flag is changed only if the CPL
is less than or equal to the IOPL. See Chapter 3, “Instruction Set Reference, A-M,” of
the Intel® 64 and IA-32 Architectures Software Developer’s Manual, Volume 2A, for
a description of the complete operation performed by the IRET instruction.
If a stack switch occurred when calling the handler procedure, the IRET instruction
6.12.1.1 Protection of Exception- and Interrupt-Handler Procedures
The privilege-level protection for exception- and interrupt-handler procedures is
similar to that used for ordinary procedure calls when called through a call gate (see
Section 5.8.4, “Accessing a Code Segment Through a Call Gate”). The processor does
6-18 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
not permit transfer of execution to an exception- or interrupt-handler procedure in a
less privileged code segment (numerically greater privilege level) than the CPL.
An attempt to violate this rule results in a general-protection exception (#GP). The
protection mechanism for exception- and interrupt-handler procedures is different in
the following ways:
• Because interrupt and exception vectors have no RPL, the RPL is not checked on
implicit calls to exception and interrupt handlers.
• The processor checks the DPL of the interrupt or trap gate only if an exception or
interrupt is generated with an INT n, INT 3, or INTO instruction. Here, the CPL
must be less than or equal to the DPL of the gate. This restriction prevents
application programs or procedures running at privilege level 3 from using a
software interrupt to access critical exception handlers, such as the page-fault
handler, providing that those handlers are placed in more privileged code
segments (numerically lower privilege level). For hardware-generated interrupts
and processor-detected exceptions, the processor ignores the DPL of interrupt
and trap gates.
Because exceptions and interrupts generally do not occur at predictable times, these
privilege rules effectively impose restrictions on the privilege levels at which excep-
tion and interrupt- handling procedures can run. Either of the following techniques
can be used to avoid privilege-level violations.
• The exception or interrupt handler can be placed in a conforming code segment.
This technique can be used for handlers that only need to access data available
on the stack (for example, divide error exceptions). If the handler needs data
from a data segment, the data segment needs to be accessible from privilege
level 3, which would make it unprotected.
• The handler can be placed in a nonconforming code segment with privilege level
0. This handler would always run, regardless of the CPL that the interrupted
program or task is running at.
6.12.1.2 Flag Usage By Exception- or Interrupt-Handler Procedure
When accessing an exception or interrupt handler through either an interrupt gate or
a trap gate, the processor clears the TF flag in the EFLAGS register after it saves the
contents of the EFLAGS register on the stack. (On calls to exception and interrupt
handlers, the processor also clears the VM, RF, and NT flags in the EFLAGS register,
after they are saved on the stack.) Clearing the TF flag prevents instruction tracing
from affecting interrupt response. A subsequent IRET instruction restores the TF
(and VM, RF, and NT) flags to the values in the saved contents of the EFLAGS register
on the stack.
The only difference between an interrupt gate and a trap gate is the way the
processor handles the IF flag in the EFLAGS register. When accessing an exception-
or interrupt-handling procedure through an interrupt gate, the processor clears the
IF flag to prevent other interrupts from interfering with the current interrupt handler.
A subsequent IRET instruction restores the IF flag to its value in the saved contents
Vol. 3 6-19
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
of the EFLAGS register on the stack. Accessing a handler procedure through a trap
gate does not affect the IF flag.
6.12.2
Interrupt Tasks
When an exception or interrupt handler is accessed through a task gate in the IDT, a
task switch results. Handling an exception or interrupt with a separate task offers
several advantages:
• The entire context of the interrupted program or task is saved automatically.
• A new TSS permits the handler to use a new privilege level 0 stack when handling
the exception or interrupt. If an exception or interrupt occurs when the current
privilege level 0 stack is corrupted, accessing the handler through a task gate can
prevent a system crash by providing the handler with a new privilege level 0
stack.
• The handler can be further isolated from other tasks by giving it a separate
The disadvantage of handling an interrupt with a separate task is that the amount of
machine state that must be saved on a task switch makes it slower than using an
interrupt gate, resulting in increased interrupt latency.
A task gate in the IDT references a TSS descriptor in the GDT (see Figure 6-5). A
switch to the handler task is handled in the same manner as an ordinary task switch
(see Section 7.3, “Task Switching”). The link back to the interrupted task is stored in
the previous task link field of the handler task’s TSS. If an exception caused an error
code to be generated, this error code is copied to the stack of the new task.
When exception- or interrupt-handler tasks are used in an operating system, there
are actually two mechanisms that can be used to dispatch tasks: the software sched-
uler (part of the operating system) and the hardware scheduler (part of the
processor's interrupt mechanism). The software scheduler needs to accommodate
interrupt tasks that may be dispatched when interrupts are enabled.
NOTE
Because IA-32 architecture tasks are not re-entrant, an interrupt-
handler task must disable interrupts between the time it completes
handling the interrupt and the time it executes the IRET instruction.
This action prevents another interrupt from occurring while the
interrupt task’s TSS is still marked busy, which would cause a
general-protection (#GP) exception.
6-20 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
TSS for Interrupt-
Handling Task
IDT
Interrupt
Vector
Task Gate
TSS
Base
Address
TSS Selector
GDT
TSS Descriptor
Figure 6-5. Interrupt Task Switch
6.13
ERROR CODE
When an exception condition is related to a specific segment, the processor pushes
an error code onto the stack of the exception handler (whether it is a procedure or
task). The error code has the format shown in Figure 6-6. The error code resembles
a segment selector; however, instead of a TI flag and RPL field, the error code
contains 3 flags:
EXT
External event (bit 0) — When set, indicates that an event external
to the program, such as a hardware interrupt, caused the exception.
IDT
Descriptor location (bit 1) — When set, indicates that the index
portion of the error code refers to a gate descriptor in the IDT; when
Vol. 3 6-21
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
clear, indicates that the index refers to a descriptor in the GDT or the
current LDT.
TI
GDT/LDT (bit 2) — Only used when the IDT flag is clear. When set,
the TI flag indicates that the index portion of the error code refers to
a segment or gate descriptor in the LDT; when clear, it indicates that
the index refers to a descriptor in the current GDT.
31
3 2 1
0
I
T
E
X
T
D
Reserved
Segment Selector Index
I
T
Figure 6-6. Error Code
to the segment or gate selector being referenced by the error code. In some cases
the error code is null (that is, all bits in the lower word are clear). A null error code
indicates that the error was not caused by a reference to a specific segment or that a
null segment descriptor was referenced in an operation.
The format of the error code is different for page-fault exceptions (#PF). See the
“Interrupt 14—Page-Fault Exception (#PF)” section in this chapter.
The error code is pushed on the stack as a doubleword or word (depending on the
default interrupt, trap, or task gate size). To keep the stack aligned for doubleword
pushes, the upper half of the error code is reserved. Note that the error code is not
popped when the IRET instruction is executed to return from an exception handler, so
the handler must remove the error code before executing a return.
Error codes are not pushed on the stack for exceptions that are generated externally
(with the INTR or LINT[1:0] pins) or the INT n instruction, even if an error code is
normally produced for those exceptions.
6.14
EXCEPTION AND INTERRUPT HANDLING IN 64-BIT
MODE
In 64-bit mode, interrupt and exception handling is similar to what has been
described for non-64-bit modes. The following are the exceptions:
• All interrupt handlers pointed by the IDT are in 64-bit code (this does not apply to
the SMI handler).
• The size of interrupt-stack pushes is fixed at 64 bits; and the processor uses
8-byte, zero extended stores.
6-22 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
• The stack pointer (SS:RSP) is pushed unconditionally on interrupts. In legacy
modes, this push is conditional and based on a change in current privilege level
(CPL).
• The new SS is set to NULL if there is a change in CPL.
• IRET behavior changes.
• There is a new interrupt stack-switch mechanism.
• The alignment of interrupt stack frame is different.
6.14.1
64-Bit Mode IDT
Interrupt and trap gates are 16 bytes in length to provide a 64-bit offset for the
instruction pointer (RIP). The 64-bit RIP referenced by interrupt-gate descriptors
allows an interrupt service routine to be located anywhere in the linear-address
space. See Figure 6-7.
Interrupt/Trap Gate
31
31
31
31
0
0
Reserved
12
Offset 63..32
8
4
16
15 14 13 12 11
D
8
7
5
4
2
0
IST
Offset 31..16
TYPE
0
0
0
0
0
0
P
L
P
16
15
0
Segment Selector
Offset 15..0
0
DPL
Offset
P
Descriptor Privilege Level
Offset to procedure entry point
Segment Present flag
Selector
IST
Segment Selector for destination code segment
Interrupt Stack Table
Figure 6-7. 64-Bit IDT Gate Descriptors
In 64-bit mode, the IDT index is formed by scaling the interrupt vector by 16. The
first eight bytes (bytes 7:0) of a 64-bit mode interrupt gate are similar but not iden-
tical to legacy 32-bit interrupt gates. The type field (bits 11:8 in bytes 7:4) is
described in Table 3-2. The Interrupt Stack Table (IST) field (bits 4:0 in bytes 7:4) is
used by the stack switching mechanisms described in Section 6.14.5, “Interrupt
Stack Table.” Bytes 11:8 hold the upper 32 bits of the target RIP (interrupt segment
offset) in canonical form. A general-protection exception (#GP) is generated if soft-
Vol. 3 6-23
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
ware attempts to reference an interrupt gate with a target RIP that is not in canonical
form.
The target code segment referenced by the interrupt gate must be a 64-bit code
segment (CS.L = 1, CS.D = 0). If the target is not a 64-bit code segment, a general-
protection exception (#GP) is generated with the IDT vector number reported as the
error code.
Only 64-bit interrupt and trap gates can be referenced in IA-32e mode (64-bit mode
and compatibility mode). Legacy 32-bit interrupt or trap gate types (0EH or 0FH) are
redefined in IA-32e mode as 64-bit interrupt and trap gate types. No 32-bit interrupt
or trap gate type exists in IA-32e mode. If a reference is made to a 16-bit interrupt
or trap gate (06H or 07H), a general-protection exception (#GP(0)) is generated.
6.14.2
64-Bit Mode Stack Frame
In legacy mode, the size of an IDT entry (16 bits or 32 bits) determines the size of
interrupt-stack-frame pushes. SS:ESP is pushed only on a CPL change. In 64-bit
mode, the size of interrupt stack-frame pushes is fixed at eight bytes. This is because
only 64-bit mode gates can be referenced. 64-bit mode also pushes SS:RSP uncon-
ditionally, rather than only on a CPL change.
Aside from error codes, pushing SS:RSP unconditionally presents operating systems
with a consistent interrupt-stackframe size across all interrupts. Interrupt service-
routine entry points that handle interrupts generated by the INTn instruction or
external INTR# signal can push an additional error code place-holder to maintain
consistency.
In legacy mode, the stack pointer may be at any alignment when an interrupt or
exception causes a stack frame to be pushed. This causes the stack frame and
succeeding pushes done by an interrupt handler to be at arbitrary alignments. In
IA-32e mode, the RSP is aligned to a 16-byte boundary before pushing the stack
frame. The stack frame itself is aligned on a 16-byte boundary when the interrupt
handler is called. The processor can arbitrarily realign the new RSP on interrupts
because the previous (possibly unaligned) RSP is unconditionally saved on the newly
aligned stack. The previous RSP will be automatically restored by a subsequent IRET.
Aligning the stack permits exception and interrupt frames to be aligned on a 16-byte
boundary before interrupts are re-enabled. This allows the stack to be formatted for
optimal storage of 16-byte XMM registers, which enables the interrupt handler to use
faster 16-byte aligned loads and stores (MOVAPS rather than MOVUPS) to save and
restore XMM registers.
Although the RSP alignment is always performed when LMA = 1, it is only of conse-
quence for the kernel-mode case where there is no stack switch or IST used. For a
stack switch or IST, the OS would have presumably put suitably aligned RSP values in
the TSS.
6-24 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
6.14.3
IRET in IA-32e Mode
In IA-32e mode, IRET executes with an 8-byte operand size. There is nothing that
forces this requirement. The stack is formatted in such a way that for actions where
IRET is required, the 8-byte IRET operand size works correctly.
Because interrupt stack-frame pushes are always eight bytes in IA-32e mode, an
IRET must pop eight byte items off the stack. This is accomplished by preceding the
IRET with a 64-bit operand-size prefix. The size of the pop is determined by the
address size of the instruction. The SS/ESP/RSP size adjustment is determined by
the stack size.
IRET pops SS:RSP unconditionally off the interrupt stack frame only when it is
executed in 64-bit mode. In compatibility mode, IRET pops SS:RSP off the stack only
if there is a CPL change. This allows legacy applications to execute properly in
compatibility mode when using the IRET instruction. 64-bit interrupt service routines
that exit with an IRET unconditionally pop SS:RSP off of the interrupt stack frame,
even if the target code segment is running in 64-bit mode or at CPL = 0. This is
because the original interrupt always pushes SS:RSP.
In IA-32e mode, IRET is allowed to load a NULL SS under certain conditions. If the
target mode is 64-bit mode and the target CPL <> 3, IRET allows SS to be loaded
with a NULL selector. As part of the stack switch mechanism, an interrupt or excep-
tion sets the new SS to NULL, instead of fetching a new SS selector from the TSS and
loading the corresponding descriptor from the GDT or LDT. The new SS selector is set
to NULL in order to properly handle returns from subsequent nested far transfers. If
the called procedure itself is interrupted, the NULL SS is pushed on the stack frame.
On the subsequent IRET, the NULL SS on the stack acts as a flag to tell the processor
not to load a new SS descriptor.
6.14.4
Stack Switching in IA-32e Mode
The IA-32 architecture provides a mechanism to automatically switch stack frames in
response to an interrupt. The 64-bit extensions of Intel 64 architecture implement a
modified version of the legacy stack-switching mechanism and an alternative stack-
switching mechanism called the interrupt stack table (IST).
In IA-32 modes, the legacy IA-32 stack-switch mechanism is unchanged. In IA-32e
part of a 64-bit mode privilege-level change (resulting from an interrupt), a new SS
descriptor is not loaded. IA-32e mode loads only an inner-level RSP from the TSS.
The new SS selector is forced to NULL and the SS selector’s RPL field is set to the new
CPL. The new SS is set to NULL in order to handle nested far transfers (CALLF, INT,
interrupts and exceptions). The old SS and RSP are saved on the new stack
(Figure 6-8). On the subsequent IRET, the old SS is popped from the stack and
loaded into the SS register.
Vol. 3 6-25
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
In summary, a stack switch in IA-32e mode works like the legacy stack switch,
except that a new SS selector is not loaded from the TSS. Instead, the new SS is
forced to NULL.
Stack Usage with
Privilege-Level Change
IA-32e Mode
Legacy Mode
Handler’s Stack
Handler’s Stack
+40
+32
+24
+16
SS
ESP
EFLAGS
SS
RSP
RFLAGS
+20
+16
+12
+8
CS
CS
+8
0
+4
0
EIP
Error Code
RIP
Error Code
Stack Pointer After
Transfer to Handler
Figure 6-8. IA-32e Mode Stack Usage After Privilege Level Change
6.14.5
Interrupt Stack Table
In IA-32e mode, a new interrupt stack table (IST) mechanism is available as an alter-
native to the modified legacy stack-switching mechanism described above. This
mechanism unconditionally switches stacks when it is enabled. It can be enabled on
an individual interrupt-vector basis using a field in the IDT entry. This means that
some interrupt vectors can use the modified legacy mechanism and others can use
the IST mechanism.
The IST mechanism is only available in IA-32e mode. It is part of the 64-bit mode
TSS. The motivation for the IST mechanism is to provide a method for specific inter-
rupts (such as NMI, double-fault, and machine-check) to always execute on a known
good stack. In legacy mode, interrupts can use the task-switch mechanism to set up
a known-good stack by accessing the interrupt service routine through a task gate
located in the IDT. However, the legacy task-switch mechanism is not supported in
IA-32e mode.
The IST mechanism provides up to seven IST pointers in the TSS. The pointers are
referenced by an interrupt-gate descriptor in the interrupt-descriptor table (IDT);
see Figure 6-7. The gate descriptor contains a 3-bit IST index field that provides an
offset into the IST section of the TSS. Using the IST mechanism, the processor loads
the value pointed by an IST pointer into the RSP.
When an interrupt occurs, the new SS selector is forced to NULL and the SS selector’s
RPL field is set to the new CPL. The old SS, RSP, RFLAGS, CS, and RIP are pushed
onto the new stack. Interrupt processing then proceeds as normal. If the IST index is
zero, the modified legacy stack-switching mechanism described above is used.
6-26 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
6.15
EXCEPTION AND INTERRUPT REFERENCE
The following sections describe conditions which generate exceptions and interrupts.
They are arranged in the order of vector numbers. The information contained in
these sections are as follows:
• Exception Class — Indicates whether the exception class is a fault, trap, or
abort type. Some exceptions can be either a fault or trap type, depending on
when the error condition is detected. (This section is not applicable to interrupts.)
• Description — Gives a general description of the purpose of the exception or
interrupt type. It also describes how the processor handles the exception or
interrupt.
• Exception Error Code — Indicates whether an error code is saved for the
exception. If one is saved, the contents of the error code are described. (This
section is not applicable to interrupts.)
• Saved Instruction Pointer — Describes which instruction the saved (or return)
instruction pointer points to. It also indicates whether the pointer can be used to
restart a faulting instruction.
• Program State Change — Describes the effects of the exception or interrupt on
the state of the currently running program or task and the possibilities of
restarting the program or task without loss of continuity.
Vol. 3 6-27
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupt 0—Divide Error Exception (#DE)
Exception Class Fault.
Description
Indicates the divisor operand for a DIV or IDIV instruction is 0 or that the result
cannot be represented in the number of bits specified for the destination operand.
Exception Error Code
None.
Saved Instruction Pointer
Saved contents of CS and EIP registers point to the instruction that generated the
exception.
Program State Change
A program-state change does not accompany the divide error, because the exception
occurs before the faulting instruction is executed.
6-28 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupt 1—Debug Exception (#DB)
Exception Class
Trap or Fault. The exception handler can distinguish
between traps or faults by examining the contents of DR6
Description
Indicates that one or more of several debug-exception conditions has been detected.
Whether the exception is a fault or a trap depends on the condition (see Table 6-3).
See Chapter 16, “Debugging, Profiling Branches and Time-Stamp Counter,” for
detailed information about the debug exceptions.
Table 6-3. Debug Exception Conditions and Corresponding Exception Classes
Exception Condition
Exception Class
Fault
Instruction fetch breakpoint
Data read or write breakpoint
I/O read or write breakpoint
General detect condition (in conjunction with in-circuit emulation)
Single-step
Trap
Trap
Fault
Trap
Task-switch
Trap
Exception Error Code
None. An exception handler can examine the debug registers to determine which
condition caused the exception.
Saved Instruction Pointer
Fault — Saved contents of CS and EIP registers point to the instruction that gener-
ated the exception.
Trap — Saved contents of CS and EIP registers point to the instruction following the
instruction that generated the exception.
Program State Change
Fault — A program-state change does not accompany the debug exception, because
the exception occurs before the faulting instruction is executed. The program can
resume normal execution upon returning from the debug exception handler.
Trap — A program-state change does accompany the debug exception, because the
instruction or task switch being executed is allowed to complete before the exception
is generated. However, the new state of the program is not corrupted and execution
of the program can continue reliably.
Vol. 3 6-29
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupt 2—NMI Interrupt
Exception Class
Not applicable.
Description
The nonmaskable interrupt (NMI) is generated externally by asserting the
processor’s NMI pin or through an NMI request set by the I/O APIC to the local APIC.
This interrupt causes the NMI interrupt handler to be called.
Exception Error Code
Not applicable.
Saved Instruction Pointer
The processor always takes an NMI interrupt on an instruction boundary. The saved
contents of CS and EIP registers point to the next instruction to be executed at the
point the interrupt is taken. See Section 6.5, “Exception Classifications,” for more
information about when the processor takes NMI interrupts.
Program State Change
The instruction executing when an NMI interrupt is received is completed before the
NMI is generated. A program or task can thus be restarted upon returning from an
interrupt handler without loss of continuity, provided the interrupt handler saves the
state of the processor before handling the interrupt and restores the processor’s
state prior to a return.
6-30 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupt 3—Breakpoint Exception (#BP)
Exception Class
Trap.
Description
Indicates that a breakpoint instruction (INT 3) was executed, causing a breakpoint
trap to be generated. Typically, a debugger sets a breakpoint by replacing the first
opcode byte of an instruction with the opcode for the INT 3 instruction. (The INT 3
instruction is one byte long, which makes it easy to replace an opcode in a code
tool can use a data segment mapped to the same physical address space as the code
segment to place an INT 3 instruction in places where it is desired to call the
debugger.
With the P6 family, Pentium, Intel486, and Intel386 processors, it is more convenient
to set breakpoints with the debug registers. (See Section 16.3.2, “Breakpoint Excep-
tion (#BP)—Interrupt Vector 3,” for information about the breakpoint exception.) If
more breakpoints are needed beyond what the debug registers allow, the INT 3
instruction can be used.
The breakpoint (#BP) exception can also be generated by executing the INT n
instruction with an operand of 3. The action of this instruction (INT 3) is slightly
different than that of the INT 3 instruction (see “INTn/INTO/INT3—Call to Interrupt
Procedure” in Chapter 3 of the Intel® 64 and IA-32 Architectures Software Devel-
oper’s Manual, Volume 2A).
Exception Error Code
None.
Saved Instruction Pointer
Saved contents of CS and EIP registers point to the instruction following the INT 3
instruction.
Program State Change
Even though the EIP points to the instruction following the breakpoint instruction, the
state of the program is essentially unchanged because the INT 3 instruction does not
affect any register or memory locations. The debugger can thus resume the
suspended program by replacing the INT 3 instruction that caused the breakpoint
with the original opcode and decrementing the saved contents of the EIP register.
Upon returning from the debugger, program execution resumes with the replaced
instruction.
Vol. 3 6-31
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupt 4—Overflow Exception (#OF)
Exception Class
Trap.
Description
Indicates that an overflow trap occurred when an INTO instruction was executed. The
INTO instruction checks the state of the OF flag in the EFLAGS register. If the OF flag
is set, an overflow trap is generated.
Some arithmetic instructions (such as the ADD and SUB) perform both signed and
unsigned arithmetic. These instructions set the OF and CF flags in the EFLAGS
register to indicate signed overflow and unsigned overflow, respectively. When
performing arithmetic on signed operands, the OF flag can be tested directly or the
INTO instruction can be used. The benefit of using the INTO instruction is that if the
overflow exception is detected, an exception handler can be called automatically to
handle the overflow condition.
Exception Error Code
None.
Saved Instruction Pointer
The saved contents of CS and EIP registers point to the instruction following the INTO
instruction.
Program State Change
Even though the EIP points to the instruction following the INTO instruction, the state
of the program is essentially unchanged because the INTO instruction does not affect
any register or memory locations. The program can thus resume normal execution
upon returning from the overflow exception handler.
6-32 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupt 5—BOUND Range Exceeded Exception (#BR)
Exception Class
Fault.
Description
Indicates that a BOUND-range-exceeded fault occurred when a BOUND instruction
was executed. The BOUND instruction checks that a signed array index is within the
upper and lower bounds of an array located in memory. If the array index is not
within the bounds of the array, a BOUND-range-exceeded fault is generated.
Exception Error Code
None.
Saved Instruction Pointer
The saved contents of CS and EIP registers point to the BOUND instruction that
generated the exception.
Program State Change
A program-state change does not accompany the bounds-check fault, because the
operands for the BOUND instruction are not modified. Returning from the BOUND-
range-exceeded exception handler causes the BOUND instruction to be restarted.
Vol. 3 6-33
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupt 6—Invalid Opcode Exception (#UD)
Exception Class
Fault.
Description
Indicates that the processor did one of the following things:
• Attempted to execute an invalid or reserved opcode.
• Attempted to execute an instruction with an operand type that is invalid for its
accompanying opcode; for example, the source operand for a LES instruction is
not a memory location.
• Attempted to execute an MMX or SSE/SSE2/SSE3 instruction on an Intel 64 or
IA-32 processor that does not support the MMX technology or
SSE/SSE2/SSE3/SSSE3 extensions, respectively. CPUID feature flags MMX (bit
23), SSE (bit 25), SSE2 (bit 26), SSE3 (ECX, bit 0), SSSE3 (ECX, bit 9) indicate
support for these extensions.
• Attempted to execute an MMX instruction or SSE/SSE2/SSE3/SSSE3 SIMD
instruction (with the exception of the MOVNTI, PAUSE, PREFETCHh, SFENCE,
LFENCE, MFENCE, CLFLUSH, MONITOR, and MWAIT instructions) when the EM
flag in control register CR0 is set (1).
• Attempted to execute an SSE/SE2/SSE3/SSSE3 instruction when the OSFXSR bit
in control register CR4 is clear (0). Note this does not include the following
SSE/SSE2/SSE3 instructions: MASKMOVQ, MOVNTQ, MOVNTI, PREFETCHh,
SFENCE, LFENCE, MFENCE, and CLFLUSH; or the 64-bit versions of the PAVGB,
PAVGW, PEXTRW, PINSRW, PMAXSW, PMAXUB, PMINSW, PMINUB, PMOVMSKB,
PMULHUW, PSADBW, PSHUFW, PADDQ, PSUBQ, PALIGNR, PABSB, PABSD,
PABSW, PHADDD, PHADDSW, PHADDW, PHSUBD, PHSUBSW, PHSUBW,
PMADDUBSM, PMULHRSW, PSHUFB, PSIGNB, PSIGND, and PSIGNW.
• Attempted to execute an SSE/SSE2/SSE3/SSSE3 instruction on an Intel 64 or
IA-32 processor that caused a SIMD floating-point exception when the
OSXMMEXCPT bit in control register CR4 is clear (0).
• Executed a UD2 instruction. Note that even though it is the execution of the UD2
instruction that causes the invalid opcode exception, the saved instruction
pointer will still points at the UD2 instruction.
• Detected a LOCK prefix that precedes an instruction that may not be locked or
one that may be locked but the destination operand is not a memory location.
• Attempted to execute an LLDT, SLDT, LTR, STR, LSL, LAR, VERR, VERW, or ARPL
instruction while in real-address or virtual-8086 mode.
• Attempted to execute the RSM instruction when not in SMM mode.
In Intel 64 and IA-32 processors that implement out-of-order execution microarchi-
tectures, this exception is not generated until an attempt is made to retire the result
of executing an invalid instruction; that is, decoding and speculatively attempting to
execute an invalid opcode does not generate this exception. Likewise, in the Pentium
6-34 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
processor and earlier IA-32 processors, this exception is not generated as the result
of prefetching and preliminary decoding of an invalid instruction. (See Section 6.5,
“Exception Classifications,” for general rules for taking of interrupts and exceptions.)
The opcodes D6 and F1 are undefined opcodes reserved by the Intel 64 and IA-32
architectures. These opcodes, even though undefined, do not generate an invalid
opcode exception.
The UD2 instruction is guaranteed to generate an invalid opcode exception.
Exception Error Code
None.
Saved Instruction Pointer
The saved contents of CS and EIP registers point to the instruction that generated the
exception.
Program State Change
A program-state change does not accompany an invalid-opcode fault, because the
invalid instruction is not executed.
Vol. 3 6-35
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupt 7—Device Not Available Exception (#NM)
Exception Class
Fault.
Description
Indicates one of the following things:
The device-not-available exception is generated by either of three conditions:
• The processor executed an x87 FPU floating-point instruction while the EM flag in
control register CR0 was set (1). See the paragraph below for the special case of
the WAIT/FWAIT instruction.
• The processor executed a WAIT/FWAIT instruction while the MP and TS flags of
register CR0 were set, regardless of the setting of the EM flag.
• The processor executed an x87 FPU, MMX, or SSE/SSE2/SSE3 instruction (with
the exception of MOVNTI, PAUSE, PREFETCHh, SFENCE, LFENCE, MFENCE, and
CLFLUSH) while the TS flag in control register CR0 was set and the EM flag is
clear.
The EM flag is set when the processor does not have an internal x87 FPU floating-
point unit. A device-not-available exception is then generated each time an x87 FPU
floating-point instruction is encountered, allowing an exception handler to call
floating-point instruction emulation routines.
The TS flag indicates that a context switch (task switch) has occurred since the last
that the context of the x87 FPU, XMM, and MXCSR registers were not saved. When
the TS flag is set and the EM flag is clear, the processor generates a device-not-avail-
able exception each time an x87 floating-point, MMX, or SSE/SSE2/SSE3 instruction
is encountered (with the exception of the instructions listed above). The exception
handler can then save the context of the x87 FPU, XMM, and MXCSR registers before
it executes the instruction. See Section 2.5, “Control Registers,” for more information
about the TS flag.
The MP flag in control register CR0 is used along with the TS flag to determine if WAIT
or FWAIT instructions should generate a device-not-available exception. It extends
the function of the TS flag to the WAIT and FWAIT instructions, giving the exception
handler an opportunity to save the context of the x87 FPU before the WAIT or FWAIT
instruction is executed. The MP flag is provided primarily for use with the Intel 286
and Intel386 DX processors. For programs running on the Pentium 4, Intel Xeon, P6
family, Pentium, or Intel486 DX processors, or the Intel 487 SX coprocessors, the MP
flag should always be set; for programs running on the Intel486 SX processor, the MP
flag should be clear.
Exception Error Code
None.
6-36 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Saved Instruction Pointer
The saved contents of CS and EIP registers point to the floating-point instruction or
the WAIT/FWAIT instruction that generated the exception.
Program State Change
A program-state change does not accompany a device-not-available fault, because
the instruction that generated the exception is not executed.
If the EM flag is set, the exception handler can then read the floating-point instruc-
tion pointed to by the EIP and call the appropriate emulation routine.
If the MP and TS flags are set or the TS flag alone is set, the exception handler can
save the context of the x87 FPU, clear the TS flag, and continue execution at the
interrupted floating-point or WAIT/FWAIT instruction.
Vol. 3 6-37
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupt 8—Double Fault Exception (#DF)
Exception Class
Abort.
Description
handler for a prior exception. Normally, when the processor detects another excep-
tion while trying to call an exception handler, the two exceptions can be handled seri-
ally. If, however, the processor cannot handle them serially, it signals the double-fault
exception. To determine when two faults need to be signalled as a double fault, the
processor divides the exceptions into three classes: benign exceptions, contributory
exceptions, and page faults (see Table 6-4).
Table 6-4. Interrupt and Exception Classes
Class
Vector Number
Description
Benign Exceptions and
Interrupts
1
2
3
Debug
NMI Interrupt
Breakpoint
4
5
6
Overflow
BOUND Range Exceeded
Invalid Opcode
7
9
16
17
18
Device Not Available
Coprocessor Segment Overrun
Floating-Point Error
Alignment Check
Machine Check
19
All
All
SIMD floating-point
INT n
INTR
Contributory Exceptions
Page Faults
0
Divide Error
Invalid TSS
Segment Not Present
Stack Fault
General Protection
10
11
12
13
14
Page Fault
Table 6-5 shows the various combinations of exception classes that cause a double
fault to be generated. A double-fault exception falls in the abort class of exceptions.
The program or task cannot be restarted or resumed. The double-fault handler can
be used to collect diagnostic information about the state of the machine and/or, when
possible, to shut the application and/or system down gracefully or restart the
system.
6-38 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
A segment or page fault may be encountered while prefetching instructions;
however, this behavior is outside the domain of Table 6-5. Any further faults gener-
ated while the processor is attempting to transfer control to the appropriate fault
handler could still lead to a double-fault sequence.
Table 6-5. Conditions for Generating a Double Fault
Second Exception
First Exception
Benign
Contributory
Page Fault
Benign
Handle Exceptions
Serially
Handle Exceptions
Serially
Handle Exceptions
Serially
Contributory
Page Fault
Handle Exceptions
Serially
Generate a Double
Fault
Handle Exceptions
Serially
Handle Exceptions
Serially
Generate a Double
Fault
Generate a Double Fault
If another exception occurs while attempting to call the double-fault handler, the
processor enters shutdown mode. This mode is similar to the state following execu-
tion of an HLT instruction. In this mode, the processor stops executing instructions
until an NMI interrupt, SMI interrupt, hardware reset, or INIT# is received. The
processor generates a special bus cycle to indicate that it has entered shutdown
mode. Software designers may need to be aware of the response of hardware when
it goes into shutdown mode. For example, hardware may turn on an indicator light on
the front panel, generate an NMI interrupt to record diagnostic information, invoke
reset initialization, generate an INIT initialization, or generate an SMI. If any events
are pending during shutdown, they will be handled after an wake event from shut-
down is processed (for example, A20M# interrupts).
If a shutdown occurs while the processor is executing an NMI interrupt handler, then
only a hardware reset can restart the processor. Likewise, if the shutdown occurs
while executing in SMM, a hardware reset must be used to restart the processor.
Exception Error Code
Zero. The processor always pushes an error code of 0 onto the stack of the double-
fault handler.
Saved Instruction Pointer
The saved contents of CS and EIP registers are undefined.
Program State Change
A program-state following a double-fault exception is undefined. The program or task
cannot be resumed or restarted. The only available action of the double-fault excep-
tion handler is to collect all possible context information for use in diagnostics and
then close the application and/or shut down or reset the processor.
Vol. 3 6-39
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
If the double fault occurs when any portion of the exception handling machine state
is corrupted, the handler cannot be invoked and the processor must be reset.
6-40 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupt 9—Coprocessor Segment Overrun
Exception Class
Abort. (Intel reserved; do not use. Recent IA-32 processors
do not generate this exception.)
Description
Indicates that an Intel386 CPU-based systems with an Intel 387 math coprocessor
detected a page or segment violation while transferring the middle portion of an
Intel 387 math coprocessor operand. The P6 family, Pentium, and Intel486 proces-
sors do not generate this exception; instead, this condition is detected with a general
protection exception (#GP), interrupt 13.
Exception Error Code
None.
Saved Instruction Pointer
The saved contents of CS and EIP registers point to the instruction that generated the
exception.
Program State Change
A program-state following a coprocessor segment-overrun exception is unde-
fined. The program or task cannot be resumed or restarted. The only available action
of the exception handler is to save the instruction pointer and reinitialize the x87 FPU
using the FNINIT instruction.
Vol. 3 6-41
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupt 10—Invalid TSS Exception (#TS)
Exception Class
Description
Indicates that there was an error related to a TSS. Such an error might be detected
during a task switch or during the execution of instructions that use information from
a TSS. Table 6-6 shows the conditions that cause an invalid TSS exception to be
generated.
Table 6-6. Invalid TSS Conditions
Error Code Index
Invalid Condition
TSS segment selector index The TSS segment limit is less than 67H for 32-bit TSS or less than
2CH for 16-bit TSS.
TSS segment selector index During an IRET task switch, the TI flag in the TSS segment selector
indicates the LDT.
TSS segment selector index During an IRET task switch, the TSS segment selector exceeds
descriptor table limit.
TSS segment selector index During an IRET task switch, the busy flag in the TSS descriptor
indicates an inactive task.
TSS segment selector index During an IRET task switch, an attempt to load the backlink limit
faults.
TSS segment selector index During an IRET task switch, the backlink is a NULL selector.
TSS segment selector index During an IRET task switch, the backlink points to a descriptor
which is not a busy TSS.
TSS segment selector index The new TSS descriptor is beyond the GDT limit.
TSS segment selector index The new TSS descriptor is not writable.
TSS segment selector index Stores to the old TSS encounter a fault condition.
TSS segment selector index The old TSS descriptor is not writable for a jump or IRET task
switch.
TSS segment selector index The new TSS backlink is not writable for a call or exception task
switch.
TSS segment selector index The new TSS selector is null on an attempt to lock the new TSS.
TSS segment selector index The new TSS selector has the TI bit set on an attempt to lock the
new TSS.
TSS segment selector index The new TSS descriptor is not an available TSS descriptor on an
attempt to lock the new TSS.
LDT segment selector index LDT or LDT not present.
6-42 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Table 6-6. Invalid TSS Conditions (Contd.)
Error Code Index
Invalid Condition
Stack segment selector
index
The stack segment selector exceeds descriptor table limit.
Stack segment selector
index
The stack segment selector is NULL.
Stack segment selector
index
The stack segment descriptor is a non-data segment.
The stack segment is not writable.
Stack segment selector
index
Stack segment selector
index
The stack segment DPL != CPL.
Stack segment selector
index
The stack segment selector RPL != CPL.
Code segment selector
index
The code segment selector exceeds descriptor table limit.
The code segment selector is NULL.
Code segment selector
index
Code segment selector
index
The code segment descriptor is not a code segment type.
The nonconforming code segment DPL != CPL.
The conforming code segment DPL is greater than CPL.
Code segment selector
index
Code segment selector
index
Data segment selector index The data segment selector exceeds the descriptor table limit.
Data segment selector index The data segment descriptor is not a readable code or data type.
Data segment selector index The data segment descriptor is a nonconforming code type and RPL
> DPL.
Data segment selector index The data segment descriptor is a nonconforming code type and CPL
> DPL.
TSS segment selector index The TSS segment selector is NULL for LTR.
TSS segment selector index The TSS segment selector has the TI bit set for LTR.
TSS segment selector index The TSS segment descriptor/upper descriptor is beyond the GDT
segment limit.
TSS segment selector index The TSS segment descriptor is not an available TSS type.
TSS segment selector index The TSS segment descriptor is an available 286 TSS type in IA-32e
mode.
Vol. 3 6-43
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Table 6-6. Invalid TSS Conditions (Contd.)
Invalid Condition
Error Code Index
TSS segment selector index The TSS segment upper descriptor is not the correct type.
TSS segment selector index The TSS segment descriptor contains a non-canonical base.
TSS segment selector index There is a limit violation in attempting to load SS selector or ESP
from a TSS on a call or exception which changes privilege levels in
legacy mode.
TSS segment selector index There is a limit violation or canonical fault in attempting to load RSP
or IST from a TSS on a call or exception which changes privilege
levels in IA-32e mode.
This exception can generated either in the context of the original task or in the
context of the new task (see Section 7.3, “Task Switching”). Until the processor has
completely verified the presence of the new TSS, the exception is generated in the
context of the original task. Once the existence of the new TSS is verified, the task
switch is considered complete. Any invalid-TSS conditions detected after this point
are handled in the context of the new task. (A task switch is considered complete
when the task register is loaded with the segment selector for the new TSS and, if the
switch is due to a procedure call or interrupt, the previous task link field of the new
TSS references the old TSS.)
The invalid-TSS handler must be a task called using a task gate. Handling this excep-
tion inside the faulting TSS context is not recommended because the processor state
may not be consistent.
Exception Error Code
An error code containing the segment selector index for the segment descriptor that
caused the violation is pushed onto the stack of the exception handler. If the EXT flag
is set, it indicates that the exception was caused by an event external to the currently
running program (for example, if an external interrupt handler using a task gate
attempted a task switch to an invalid TSS).
Saved Instruction Pointer
If the exception condition was detected before the task switch was carried out, the
saved contents of CS and EIP registers point to the instruction that invoked the task
the saved contents of CS and EIP registers point to the first instruction of the new
task.
Program State Change
The ability of the invalid-TSS handler to recover from the fault depends on the error
condition than causes the fault. See Section 7.3, “Task Switching,” for more informa-
tion on the task switch process and the possible recovery actions that can be taken.
6-44 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
If an invalid TSS exception occurs during a task switch, it can occur before or after
the commit-to-new-task point. If it occurs before the commit point, no program state
change occurs. If it occurs after the commit point (when the segment descriptor
information for the new segment selectors have been loaded in the segment regis-
ters), the processor will load all the state information from the new TSS before it
generates the exception. During a task switch, the processor first loads all the
segment registers with segment selectors from the TSS, then checks their contents
for validity. If an invalid TSS exception is discovered, the remaining segment regis-
ters are loaded but not checked for validity and therefore may not be usable for refer-
encing memory. The invalid TSS handler should not rely on being able to use the
segment selectors found in the CS, SS, DS, ES, FS, and GS registers without causing
another exception. The exception handler should load all segment registers before
trying to resume the new task; otherwise, general-protection exceptions (#GP) may
result later under conditions that make diagnosis more difficult. The Intel recom-
mended way of dealing situation is to use a task for the invalid TSS exception
handler. The task switch back to the interrupted task from the invalid-TSS exception-
handler task will then cause the processor to check the registers as it loads them
from the TSS.
Vol. 3 6-45
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupt 11—Segment Not Present (#NP)
Exception Class
Fault.
Description
Indicates that the present flag of a segment or gate descriptor is clear. The processor
can generate this exception during any of the following operations:
• While attempting to load CS, DS, ES, FS, or GS registers. [Detection of a not-
present segment while loading the SS register causes a stack fault exception
(#SS) to be generated.] This situation can occur while performing a task switch.
• While attempting to load the LDTR using an LLDT instruction. Detection of a not-
present LDT while loading the LDTR during a task switch operation causes an
invalid-TSS exception (#TS) to be generated.
• When executing the LTR instruction and the TSS is marked not present.
• While attempting to use a gate descriptor or TSS that is marked segment-not-
present, but is otherwise valid.
An operating system typically uses the segment-not-present exception to implement
virtual memory at the segment level. If the exception handler loads the segment and
returns, the interrupted program or task resumes execution.
A not-present indication in a gate descriptor, however, does not indicate that a
segment is not present (because gates do not correspond to segments). The oper-
ating system may use the present flag for gate descriptors to trigger exceptions of
special significance to the operating system.
A contributory exception or page fault that subsequently referenced a not-present
segment would cause a double fault (#DF) to be generated instead of #NP.
Exception Error Code
An error code containing the segment selector index for the segment descriptor that
caused the violation is pushed onto the stack of the exception handler. If the EXT flag
is set, it indicates that the exception resulted from either:
• an external event (NMI or INTR) that caused an interrupt, which subsequently
referenced a not-present segment
• a benign exception that subsequently referenced a not-present segment
The IDT flag is set if the error code refers to an IDT entry. This occurs when the IDT
entry for an interrupt being serviced references a not-present gate. Such an event
could be generated by an INT instruction or a hardware interrupt.
Saved Instruction Pointer
The saved contents of CS and EIP registers normally point to the instruction that
generated the exception. If the exception occurred while loading segment descrip-
6-46 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
tors for the segment selectors in a new TSS, the CS and EIP registers point to the first
instruction in the new task. If the exception occurred while accessing a gate
descriptor, the CS and EIP registers point to the instruction that invoked the access
(for example a CALL instruction that references a call gate).
Program State Change
If the segment-not-present exception occurs as the result of loading a register (CS,
DS, SS, ES, FS, GS, or LDTR), a program-state change does accompany the excep-
tion because the register is not loaded. Recovery from this exception is possible by
simply loading the missing segment into memory and setting the present flag in the
segment descriptor.
If the segment-not-present exception occurs while accessing a gate descriptor, a
program-state change does not accompany the exception. Recovery from this excep-
tion is possible merely by setting the present flag in the gate descriptor.
If a segment-not-present exception occurs during a task switch, it can occur before
or after the commit-to-new-task point (see Section 7.3, “Task Switching”). If it
the commit point, the processor will load all the state information from the new TSS
(without performing any additional limit, present, or type checks) before it generates
the exception. The segment-not-present exception handler should not rely on being
able to use the segment selectors found in the CS, SS, DS, ES, FS, and GS registers
without causing another exception. (See the Program State Change description for
“Interrupt 10—Invalid TSS Exception (#TS)” in this chapter for additional information
on how to handle this situation.)
Vol. 3 6-47
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupt 12—Stack Fault Exception (#SS)
Exception Class
Fault.
Description
Indicates that one of the following stack related conditions was detected:
• A limit violation is detected during an operation that refers to the SS register.
Operations that can cause a limit violation include stack-oriented instructions
such as POP, PUSH, CALL, RET, IRET, ENTER, and LEAVE, as well as other memory
references which implicitly or explicitly use the SS register (for example, MOV
AX, [BP+6] or MOV AX, SS:[EAX+6]). The ENTER instruction generates this
exception when there is not enough stack space for allocating local variables.
• A not-present stack segment is detected when attempting to load the SS register.
This violation can occur during the execution of a task switch, a CALL instruction
to a different privilege level, a return to a different privilege level, an LSS
instruction, or a MOV or POP instruction to the SS register.
• A canonical violation is detected in 64-bit mode during an operation that
reference memory using the stack pointer register containing a non-canonical
memory address.
Recovery from this fault is possible by either extending the limit of the stack segment
(in the case of a limit violation) or loading the missing stack segment into memory (in
the case of a not-present violation.
In the case of a canonical violation that was caused intentionally by software,
recovery is possible by loading the correct canonical value into RSP. Otherwise, a
canonical violation of the address in RSP likely reflects some register corruption in
the software.
Exception Error Code
If the exception is caused by a not-present stack segment or by overflow of the new
stack during an inter-privilege-level call, the error code contains a segment selector
for the segment that caused the exception. Here, the exception handler can test the
present flag in the segment descriptor pointed to by the segment selector to deter-
mine the cause of the exception. For a normal limit violation (on a stack segment
already in use) the error code is set to 0.
Saved Instruction Pointer
The saved contents of CS and EIP registers generally point to the instruction that
generated the exception. However, when the exception results from attempting to
load a not-present stack segment during a task switch, the CS and EIP registers point
to the first instruction of the new task.
6-48 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Program State Change
A program-state change does not generally accompany a stack-fault exception,
because the instruction that generated the fault is not executed. Here, the instruction
can be restarted after the exception handler has corrected the stack fault condition.
If a stack fault occurs during a task switch, it occurs after the commit-to-new-task
point (see Section 7.3, “Task Switching”). Here, the processor loads all the state
type checks) before it generates the exception. The stack fault handler should thus
not rely on being able to use the segment selectors found in the CS, SS, DS, ES, FS,
and GS registers without causing another exception. The exception handler should
check all segment registers before trying to resume the new task; otherwise, general
protection faults may result later under conditions that are more difficult to diagnose.
(See the Program State Change description for “Interrupt 10—Invalid TSS Exception
(#TS)” in this chapter for additional information on how to handle this situation.)
Vol. 3 6-49
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupt 13—General Protection Exception (#GP)
Exception Class
Fault.
Description
Indicates that the processor detected one of a class of protection violations called
“general-protection violations.” The conditions that cause this exception to be gener-
ated comprise all the protection violations that do not cause other exceptions to be
generated (such as, invalid-TSS, segment-not-present, stack-fault, or page-fault
exceptions). The following conditions cause general-protection exceptions to be
generated:
• Exceeding the segment limit when accessing the CS, DS, ES, FS, or GS
segments.
• Exceeding the segment limit when referencing a descriptor table (except during a
task switch or a stack switch).
• Transferring execution to a segment that is not executable.
• Writing to a code segment or a read-only data segment.
• Reading from an execute-only code segment.
• Loading the SS register with a segment selector for a read-only segment (unless
the selector comes from a TSS during a task switch, in which case an invalid-TSS
exception occurs).
• Loading the SS, DS, ES, FS, or GS register with a segment selector for a system
segment.
• Loading the DS, ES, FS, or GS register with a segment selector for an execute-
only code segment.
• Loading the SS register with the segment selector of an executable segment or a
null segment selector.
• Loading the CS register with a segment selector for a data segment or a null
segment selector.
• Accessing memory using the DS, ES, FS, or GS register when it contains a null
segment selector.
• Switching to a busy task during a call or jump to a TSS.
• Using a segment selector on a non-IRET task switch that points to a TSS
descriptor in the current LDT. TSS descriptors can only reside in the GDT. This
condition causes a #TS exception during an IRET task switch.
• Violating any of the privilege rules described in Chapter 5, “Protection.”
• Exceeding the instruction length limit of 15 bytes (this only can occur when
redundant prefixes are placed before an instruction).
• Loading the CR0 register with a set PG flag (paging enabled) and a clear PE flag
(protection disabled).
6-50 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
• Loading the CR0 register with a set NW flag and a clear CD flag.
• Referencing an entry in the IDT (following an interrupt or exception) that is not
an interrupt, trap, or task gate.
• Attempting to access an interrupt or exception handler through an interrupt or
trap gate from virtual-8086 mode when the handler’s code segment DPL is
greater than 0.
• Attempting to write a 1 into a reserved bit of CR4.
• Attempting to execute a privileged instruction when the CPL is not equal to 0 (see
Section 5.9, “Privileged Instructions,” for a list of privileged instructions).
• Writing to a reserved bit in an MSR.
• Accessing a gate that contains a null segment selector.
• Executing the INT n instruction when the CPL is greater than the DPL of the
referenced interrupt, trap, or task gate.
• The segment selector in a call, interrupt, or trap gate does not point to a code
segment.
• The segment selector operand in the LLDT instruction is a local type (TI flag is
set) or does not point to a segment descriptor of the LDT type.
• The segment selector operand in the LTR instruction is local or points to a TSS
that is not available.
• The target code-segment selector for a call, jump, or return is null.
• If the PAE and/or PSE flag in control register CR4 is set and the processor detects
any reserved bits in a page-directory-pointer-table entry set to 1. These bits are
checked during a write to control registers CR0, CR3, or CR4 that causes a
reloading of the page-directory-pointer-table entry.
• Attempting to write a non-zero value into the reserved bits of the MXCSR register.
• Executing an SSE/SSE2/SSE3 instruction that attempts to access a 128-bit
memory location that is not aligned on a 16-byte boundary when the instruction
requires 16-byte alignment. This condition also applies to the stack segment.
A program or task can be restarted following any general-protection exception. If the
exception occurs while attempting to call an interrupt handler, the interrupted
program can be restartable, but the interrupt may be lost.
Exception Error Code
The processor pushes an error code onto the exception handler's stack. If the fault
condition was detected while loading a segment descriptor, the error code contains a
segment selector to or IDT vector number for the descriptor; otherwise, the error
code is 0. The source of the selector in an error code may be any of the following:
• An operand of the instruction.
• A selector from a gate which is the operand of the instruction.
Vol. 3 6-51
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
• A selector from a TSS involved in a task switch.
• IDT vector number.
Saved Instruction Pointer
The saved contents of CS and EIP registers point to the instruction that generated the
exception.
Program State Change
In general, a program-state change does not accompany a general-protection excep-
tion, because the invalid instruction or operation is not executed. An exception
handler can be designed to correct all of the conditions that cause general-protection
exceptions and restart the program or task without any loss of program continuity.
If a general-protection exception occurs during a task switch, it can occur before or
after the commit-to-new-task point (see Section 7.3, “Task Switching”). If it occurs
commit point, the processor will load all the state information from the new TSS
(without performing any additional limit, present, or type checks) before it generates
the exception. The general-protection exception handler should thus not rely on
being able to use the segment selectors found in the CS, SS, DS, ES, FS, and GS
registers without causing another exception. (See the Program State Change
description for “Interrupt 10—Invalid TSS Exception (#TS)” in this chapter for addi-
tional information on how to handle this situation.)
General Protection Exception in 64-bit Mode
The following conditions cause general-protection exceptions in 64-bit mode:
• If the memory address is in a non-canonical form.
• If a segment descriptor memory address is in non-canonical form.
• If the target offset in a destination operand of a call or jmp is in a non-canonical
form.
• If a code segment or 64-bit call gate overlaps non-canonical space.
• If the code segment descriptor pointed to by the selector in the 64-bit gate
doesn't have the L-bit set and the D-bit clear.
• If the EFLAGS.NT bit is set in IRET.
• If the stack segment selector of IRET is null when going back to compatibility
mode.
• If the stack segment selector of IRET is null going back to CPL3 and 64-bit mode.
• If a null stack segment selector RPL of IRET is not equal to CPL going back to non-
CPL3 and 64-bit mode.
• If the proposed new code segment descriptor of IRET has both the D-bit and the
L-bit set.
6-52 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
• If the segment descriptor pointed to by the segment selector in the destination
operand is a code segment and it has both the D-bit and the L-bit set.
• If the segment descriptor from a 64-bit call gate is in non-canonical space.
• If the DPL from a 64-bit call-gate is less than the CPL or than the RPL of the 64-bit
call-gate.
• If the upper type field of a 64-bit call gate is not 0x0.
• If an attempt is made to load a null selector in the SS register in compatibility
mode.
• If an attempt is made to load null selector in the SS register in CPL3 and 64-bit
mode.
• If an attempt is made to load a null selector in the SS register in non-CPL3 and
64-bit mode where RPL is not equal to CPL.
• If an attempt is made to clear CR0.PG while IA-32e mode is enabled.
• If an attempt is made to set a reserved bit in CR3, CR4 or CR8.
Vol. 3 6-53
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupt 14—Page-Fault Exception (#PF)
Exception Class
Fault.
Description
Indicates that, with paging enabled (the PG flag in the CR0 register is set), the
processor detected one of the following conditions while using the page-translation
mechanism to translate a linear address to a physical address:
• The P (present) flag in a page-directory or page-table entry needed for the
address translation is clear, indicating that a page table or the page containing
the operand is not present in physical memory.
• The procedure does not have sufficient privilege to access the indicated page
(that is, a procedure running in user mode attempts to access a supervisor-mode
page).
and later processors, if the WP flag is set in CR0, the page fault will also be
triggered by code running in supervisor mode that tries to write to a read-only
page.
• An instruction fetch to a linear address that translates to a physical address in a
memory page with the execute-disable bit set (for information about the
execute-disable bit, see Chapter 4, “Paging”).
below of RSVD error code flag.
The exception handler can recover from page-not-present conditions and restart the
program or task without any loss of program continuity. It can also restart the
program or task after a privilege violation, but the problem that caused the privilege
violation may be uncorrectable.
See also: Section 4.7, “Page-Fault Exceptions.”
Exception Error Code
Yes (special format). The processor provides the page-fault handler with two items of
information to aid in diagnosing the exception and recovering from it:
• An error code on the stack. The error code for a page fault has a format different
from that for other exceptions (see Figure 6-9). The error code tells the
exception handler four things:
— The P flag indicates whether the exception was due to a not-present page (0)
or to either an access rights violation or the use of a reserved bit (1).
— The W/R flag indicates whether the memory access that caused the exception
was a read (0) or write (1).
6-54 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
— The U/S flag indicates whether the processor was executing at user mode (1)
or supervisor mode (0) at the time of the exception.
— The RSVD flag indicates that the processor detected 1s in reserved bits of the
page directory, when the PSE or PAE flags in control register CR4 are set to 1.
Note:
•
•
•
The PSE flag is only available in recent Intel 64 and IA-32 processors
including the Pentium 4, Intel Xeon, P6 family, and Pentium processors.
The PAE flag is only available on recent Intel 64 and IA-32 processors
including the Pentium 4, Intel Xeon, and P6 family processors.
In earlier IA-32 processors, the bit position of the RSVD flag is reserved
and is cleared to 0.
— The I/D flag indicates whether the exception was caused by an instruction
fetch. This flag is reserved and cleared to 0 if CR4.PAE = 0 (32-bit paging is
in use) or IA32_EFER.NXE= 0 (the execute-disable feature is either
unsupported or not enabled). See Section 4.7, “Page-Fault Exceptions,” for
details.
31
3
2
1
0
4
Reserved
0 The fault was caused by a non-present page.
P
1 The fault was caused by a page-level protection violation.
W/R
U/S
0 The access causing the fault was a read.
1
The access causing the fault was a write.
The access causing the fault originated when the processor
was executing in supervisor mode.
The access causing the fault originated when the processor
was executing in user mode.
0
1
0
1
The fault was not caused by reserved bit violation.
The fault was caused by reserved bits set to 1 in a page directory.
RSVD
I/D
0
1
The fault was not caused by an instruction fetch.
The fault was caused by an instruction fetch.
Figure 6-9. Page-Fault Error Code
• The contents of the CR2 register. The processor loads the CR2 register with the
32-bit linear address that generated the exception. The page-fault handler can
use this address to locate the corresponding page directory and page-table
entries. Another page fault can potentially occur during execution of the page-
fault handler; the handler should save the contents of the CR2 register before a
Vol. 3 6-55
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
1
second page fault can occur. If a page fault is caused by a page-level protection
violation, the access flag in the page-directory entry is set when the fault occurs.
The behavior of IA-32 processors regarding the access flag in the corresponding
page-table entry is model specific and not architecturally defined.
Saved Instruction Pointer
The saved contents of CS and EIP registers generally point to the instruction that
generated the exception. If the page-fault exception occurred during a task switch,
the CS and EIP registers may point to the first instruction of the new task (as
described in the following “Program State Change” section).
Program State Change
A program-state change does not normally accompany a page-fault exception,
because the instruction that causes the exception to be generated is not executed.
After the page-fault exception handler has corrected the violation (for example,
loaded the missing page into memory), execution of the program or task can be
resumed.
When a page-fault exception is generated during a task switch, the program-state
may change, as follows. During a task switch, a page-fault exception can occur
during any of following operations:
• While writing the state of the original task into the TSS of that task.
• While reading the GDT to locate the TSS descriptor of the new task.
• While reading the TSS of the new task.
• While reading segment descriptors associated with segment selectors from the
new task.
• While reading the LDT of the new task to verify the segment registers stored in
the new TSS.
In the last two cases the exception occurs in the context of the new task. The instruc-
tion pointer refers to the first instruction of the new task, not to the instruction which
caused the task switch (or the last instruction to be executed, in the case of an inter-
rupt). If the design of the operating system permits page faults to occur during task-
switches, the page-fault handler should be called through a task gate.
If a page fault occurs during a task switch, the processor will load all the state infor-
mation from the new TSS (without performing any additional limit, present, or type
checks) before it generates the exception. The page-fault handler should thus not
rely on being able to use the segment selectors found in the CS, SS, DS, ES, FS, and
GS registers without causing another exception. (See the Program State Change
1. Processors update CR2 whenever a page fault is detected. If a second page fault occurs while an
earlier page fault is being delivered, the faulting linear address of the second fault will overwrite
the contents of CR2 (replacing the previous address). These updates to CR2 occur even if the
page fault results in a double fault or occurs during the delivery of a double fault.
6-56 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
description for “Interrupt 10—Invalid TSS Exception (#TS)” in this chapter for addi-
tional information on how to handle this situation.)
Additional Exception-Handling Information
Special care should be taken to ensure that an exception that occurs during an
explicit stack switch does not cause the processor to use an invalid stack pointer
(SS:ESP). Software written for 16-bit IA-32 processors often use a pair of instruc-
tions to change to a new stack, for example:
MOV SS, AX
MOV SP, StackTop
When executing this code on one of the 32-bit IA-32 processors, it is possible to get
a page fault, general-protection fault (#GP), or alignment check fault (#AC) after the
segment selector has been loaded into the SS register but before the ESP register
has been loaded. At this point, the two parts of the stack pointer (SS and ESP) are
inconsistent. The new stack segment is being used with the old stack pointer.
The processor does not use the inconsistent stack pointer if the exception handler
switches to a well defined stack (that is, the handler is a task or a more privileged
procedure). However, if the exception handler is called at the same privilege level
and from the same task, the processor will attempt to use the inconsistent stack
pointer.
In systems that handle page-fault, general-protection, or alignment check excep-
tions within the faulting task (with trap or interrupt gates), software executing at the
same privilege level as the exception handler should initialize a new stack by using
the LSS instruction rather than a pair of MOV instructions, as described earlier in this
note. When the exception handler is running at privilege level 0 (the normal case),
the problem is limited to procedures or tasks that run at privilege level 0, typically
the kernel of the operating system.
Vol. 3 6-57
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupt 16—x87 FPU Floating-Point Error (#MF)
Exception Class
Fault.
Description
Indicates that the x87 FPU has detected a floating-point error. The NE flag in the
register CR0 must be set for an interrupt 16 (floating-point error exception) to be
generated. (See Section 2.5, “Control Registers,” for a detailed description of the NE
flag.)
NOTE
SIMD floating-point exceptions (#XM) are signaled through interrupt
19.
While executing x87 FPU instructions, the x87 FPU detects and reports six types of
floating-point error conditions:
• Invalid operation (#I)
— Stack overflow or underflow (#IS)
— Invalid arithmetic operation (#IA)
• Divide-by-zero (#Z)
• Denormalized operand (#D)
• Numeric overflow (#O)
• Numeric underflow (#U)
• Inexact result (precision) (#P)
Each of these error conditions represents an x87 FPU exception type, and for each of
exception type, the x87 FPU provides a flag in the x87 FPU status register and a mask
bit in the x87 FPU control register. If the x87 FPU detects a floating-point error and
the mask bit for the exception type is set, the x87 FPU handles the exception auto-
matically by generating a predefined (default) response and continuing program
execution. The default responses have been designed to provide a reasonable result
for most floating-point applications.
If the mask for the exception is clear and the NE flag in register CR0 is set, the x87
FPU does the following:
1. Sets the necessary flag in the FPU status register.
2. Waits until the next “waiting” x87 FPU instruction or WAIT/FWAIT instruction is
encountered in the program’s instruction stream.
3. Generates an internal error signal that cause the processor to generate a
floating-point exception (#MF).
6-58 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Prior to executing a waiting x87 FPU instruction or the WAIT/FWAIT instruction, the
x87 FPU checks for pending x87 FPU floating-point exceptions (as described in step 2
above). Pending x87 FPU floating-point exceptions are ignored for “non-waiting” x87
FPU instructions, which include the FNINIT, FNCLEX, FNSTSW, FNSTSW AX, FNSTCW,
FNSTENV, and FNSAVE instructions. Pending x87 FPU exceptions are also ignored
when executing the state management instructions FXSAVE and FXRSTOR.
All of the x87 FPU floating-point error conditions can be recovered from. The x87 FPU
floating-point-error exception handler can determine the error condition that caused
the exception from the settings of the flags in the x87 FPU status word. See “Soft-
ware Exception Handling” in Chapter 8 of the Intel® 64 and IA-32 Architectures Soft-
ware Developer’s Manual, Volume 1, for more information on handling x87 FPU
floating-point exceptions.
Exception Error Code
None. The x87 FPU provides its own error information.
Saved Instruction Pointer
The saved contents of CS and EIP registers point to the floating-point or WAIT/FWAIT
instruction that was about to be executed when the floating-point-error exception
was generated. This is not the faulting instruction in which the error condition was
detected. The address of the faulting instruction is contained in the x87 FPU instruc-
tion pointer register. See “x87 FPU Instruction and Operand (Data) Pointers” in
Chapter 8 of the Intel® 64 and IA-32 Architectures Software Developer’s Manual,
Volume 1, for more information about information the FPU saves for use in handling
floating-point-error exceptions.
Program State Change
A program-state change generally accompanies an x87 FPU floating-point exception
because the handling of the exception is delayed until the next waiting x87 FPU
floating-point or WAIT/FWAIT instruction following the faulting instruction. The x87
FPU, however, saves sufficient information about the error condition to allow
recovery from the error and re-execution of the faulting instruction if needed.
In situations where non- x87 FPU floating-point instructions depend on the results of
an x87 FPU floating-point instruction, a WAIT or FWAIT instruction can be inserted in
front of a dependent instruction to force a pending x87 FPU floating-point exception
to be handled before the dependent instruction is executed. See “x87 FPU Exception
Synchronization” in Chapter 8 of the Intel® 64 and IA-32 Architectures Software
Developer’s Manual, Volume 1, for more information about synchronization of x87
floating-point-error exceptions.
Vol. 3 6-59
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupt 17—Alignment Check Exception (#AC)
Exception Class
Fault.
Description
Indicates that the processor detected an unaligned memory operand when alignment
checking was enabled. Alignment checks are only carried out in data (or stack)
accesses (not in code fetches or system segment accesses). An example of an align-
ment-check violation is a word stored at an odd byte address, or a doubleword stored
at an address that is not an integer multiple of 4. Table 6-7 lists the alignment
requirements various data types recognized by the processor.
Table 6-7. Alignment Requirements by Data Type
Data Type
Address Must Be Divisible By
Word
2
4
4
8
8
Doubleword
Single-precision floating-point (32-bits)
Double-precision floating-point (64-bits)
Double extended-precision floating-point (80-
bits)
Quadword
8
Double quadword
16
Segment Selector
2
32-bit Far Pointer
2
48-bit Far Pointer
4
32-bit Pointer
4
GDTR, IDTR, LDTR, or Task Register Contents
FSTENV/FLDENV Save Area
FSAVE/FRSTOR Save Area
Bit String
4
4 or 2, depending on operand size
4 or 2, depending on operand size
2 or 4 depending on the operand-size attribute.
Note that the alignment check exception (#AC) is generated only for data types that
must be aligned on word, doubleword, and quadword boundaries. A general-protec-
tion exception (#GP) is generated 128-bit data types that are not aligned on a
16-byte boundary.
To enable alignment checking, the following conditions must be true:
• AM flag in CR0 register is set.
6-60 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
• AC flag in the EFLAGS register is set.
• The CPL is 3 (protected mode or virtual-8086 mode).
Alignment-check exceptions (#AC) are generated only when operating at privilege
level 3 (user mode). Memory references that default to privilege level 0, such as
segment descriptor loads, do not generate alignment-check exceptions, even when
caused by a memory reference made from privilege level 3.
Storing the contents of the GDTR, IDTR, LDTR, or task register in memory while at
privilege level 3 can generate an alignment-check exception. Although application
programs do not normally store these registers, the fault can be avoided by aligning
the information stored on an even word-address.
The FXSAVE and FXRSTOR instructions save and restore a 512-byte data structure,
the first byte of which must be aligned on a 16-byte boundary. If the alignment-check
exception (#AC) is enabled when executing these instructions (and CPL is 3), a
misaligned memory operand can cause either an alignment-check exception or a
general-protection exception (#GP) depending on the processor implementation
(see “FXSAVE-Save x87 FPU, MMX, SSE, and SSE2 State” and “FXRSTOR-Restore
x87 FPU, MMX, SSE, and SSE2 State” in Chapter 3 of the Intel® 64 and IA-32 Archi-
tectures Software Developer’s Manual, Volume 2A).
The MOVUPS and MOVUPD instructions perform 128-bit unaligned loads or stores.
The LDDQU instructions loads 128-bit unaligned data.They do not generate general-
protection exceptions (#GP) when operands are not aligned on a 16-byte boundary.
If alignment checking is enabled, alignment-check exceptions (#AC) may or may not
be generated depending on processor implementation when data addresses are not
aligned on an 8-byte boundary.
FSAVE and FRSTOR instructions can generate unaligned references, which can cause
alignment-check faults. These instructions are rarely needed by application
programs.
Exception Error Code
Yes (always zero).
Saved Instruction Pointer
The saved contents of CS and EIP registers point to the instruction that generated the
exception.
Program State Change
A program-state change does not accompany an alignment-check fault, because the
instruction is not executed.
Vol. 3 6-61
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupt 18—Machine-Check Exception (#MC)
Exception Class
Abort.
Description
Indicates that the processor detected an internal machine error or a bus error, or that
an external agent detected a bus error. The machine-check exception is model-
specific, available on the Pentium and later generations of processors. The imple-
mentation of the machine-check exception is different between different processor
families, and these implementations may not be compatible with future Intel 64 or
IA-32 processors. (Use the CPUID instruction to determine whether this feature is
present.)
pins: the BINIT# and MCERR# pins on the Pentium 4, Intel Xeon, and P6 family
processors and the BUSCHK# pin on the Pentium processor. When one of these pins
is enabled, asserting the pin causes error information to be loaded into machine-
check registers and a machine-check exception is generated.
The machine-check exception and machine-check architecture are discussed in detail
in Chapter 15, “Machine-Check Architecture.” Also, see the data books for the indi-
vidual processors for processor-specific hardware information.
Exception Error Code
Saved Instruction Pointer
For the Pentium 4 and Intel Xeon processors, the saved contents of extended
machine-check state registers are directly associated with the error that caused the
machine-check exception to be generated (see Section 15.3.1.2,
Check State MSRs”).
For the P6 family processors, if the EIPV flag in the MCG_STATUS MSR is set, the
saved contents of CS and EIP registers are directly associated with the error that
caused the machine-check exception to be generated; if the flag is clear, the saved
instruction pointer may not be associated with the error (see Section 15.3.1.2,
“IA32_MCG_STATUS MSR”).
For the Pentium processor, contents of the CS and EIP registers may not be associ-
ated with the error.
Program State Change
The machine-check mechanism is enabled by setting the MCE flag in control register
CR4.
6-62 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
For the Pentium 4, Intel Xeon, P6 family, and Pentium processors, a program-state
change always accompanies a machine-check exception, and an abort class excep-
tion is generated. For abort exceptions, information about the exception can be
collected from the machine-check MSRs, but the program cannot generally be
restarted.
If the machine-check mechanism is not enabled (the MCE flag in control register CR4
is clear), a machine-check exception causes the processor to enter the shutdown
state.
Vol. 3 6-63
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupt 19—SIMD Floating-Point Exception (#XM)
Exception Class
Fault.
Description
Indicates the processor has detected an SSE/SSE2/SSE3 SIMD floating-point excep-
tion. The appropriate status flag in the MXCSR register must be set and the particular
exception unmasked for this interrupt to be generated.
There are six classes of numeric exception conditions that can occur while executing
an SSE/ SSE2/SSE3 SIMD floating-point instruction:
• Invalid operation (#I)
• Divide-by-zero (#Z)
• Denormal operand (#D)
• Numeric overflow (#O)
• Numeric underflow (#U)
• Inexact result (Precision) (#P)
The invalid operation, divide-by-zero, and denormal-operand exceptions are pre-
computation exceptions; that is, they are detected before any arithmetic operation
occurs. The numeric underflow, numeric overflow, and inexact result exceptions are
post-computational exceptions.
See "SIMD Floating-Point Exceptions" in Chapter 11 of the Intel® 64 and IA-32
Architectures Software Developer’s Manual, Volume 1, for additional information
about the SIMD floating-point exception classes.
When a SIMD floating-point exception occurs, the processor does either of the
following things:
• It handles the exception automatically by producing the most reasonable result
and allowing program execution to continue undisturbed. This is the response to
masked exceptions.
• It generates a SIMD floating-point exception, which in turn invokes a software
exception handler. This is the response to unmasked exceptions.
Each of the six SIMD floating-point exception conditions has a corresponding flag bit
and mask bit in the MXCSR register. If an exception is masked (the corresponding
mask bit in the MXCSR register is set), the processor takes an appropriate automatic
default action and continues with the computation. If the exception is unmasked (the
corresponding mask bit is clear) and the operating system supports SIMD floating-
point exceptions (the OSXMMEXCPT flag in control register CR4 is set), a software
exception handler is invoked through a SIMD floating-point exception. If the excep-
tion is unmasked and the OSXMMEXCPT bit is clear (indicating that the operating
system does not support unmasked SIMD floating-point exceptions), an invalid
opcode exception (#UD) is signaled instead of a SIMD floating-point exception.
6-64 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Note that because SIMD floating-point exceptions are precise and occur immediately,
the situation does not arise where an x87 FPU instruction, a WAIT/FWAIT instruction,
or another SSE/SSE2/SSE3 instruction will catch a pending unmasked SIMD floating-
point exception.
In situations where a SIMD floating-point exception occurred while the SIMD
floating-point exceptions were masked (causing the corresponding exception flag to
be set) and the SIMD floating-point exception was subsequently unmasked, then no
exception is generated when the exception is unmasked.
When SSE/SSE2/SSE3 SIMD floating-point instructions operate on packed operands
(made up of two or four sub-operands), multiple SIMD floating-point exception
conditions may be detected. If no more than one exception condition is detected for
one or more sets of sub-operands, the exception flags are set for each exception
condition detected. For example, an invalid exception detected for one sub-operand
will not prevent the reporting of a divide-by-zero exception for another sub-operand.
However, when two or more exceptions conditions are generated for one sub-
operand, only one exception condition is reported, according to the precedences
shown in Table 6-8. This exception precedence sometimes results in the higher
priority exception condition being reported and the lower priority exception condi-
tions being ignored.
Table 6-8. SIMD Floating-Point Exceptions Priority
Priority
Description
1 (Highest)
Invalid operation exception due to SNaN operand (or any NaN operand for
maximum, minimum, or certain compare and convert operations).
1
2
3
QNaN operand .
Any other invalid operation exception not mentioned above or a divide-by-zero
2
exception .
2
4
5
Denormal operand exception .
Numeric overflow and underflow exceptions possibly in conjunction with the
2
inexact result exception .
6 (Lowest)
Inexact result exception.
NOTES:
1. Though a QNaN this is not an exception, the handling of a QNaN operand has precedence over
lower priority exceptions. For example, a QNaN divided by zero results in a QNaN, not a divide-
by-zero- exception.
2. If masked, then instruction execution continues, and a lower priority exception can occur as
well.
Exception Error Code
None.
Vol. 3 6-65
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Saved Instruction Pointer
The saved contents of CS and EIP registers point to the SSE/SSE2/SSE3 instruction
that was executed when the SIMD floating-point exception was generated. This is the
faulting instruction in which the error condition was detected.
Program State Change
A program-state change does not accompany a SIMD floating-point exception
because the handling of the exception is immediate unless the particular exception is
masked. The available state information is often sufficient to allow recovery from the
error and re-execution of the faulting instruction if needed.
6-66 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
Interrupts 32 to 255—User Defined Interrupts
Exception Class
Not applicable.
Description
Indicates that the processor did one of the following things:
• Executed an INT n instruction where the instruction operand is one of the vector
numbers from 32 through 255.
• Responded to an interrupt request at the INTR pin or from the local APIC when
the interrupt vector number associated with the request is from 32 through 255.
Exception Error Code
Not applicable.
Saved Instruction Pointer
The saved contents of CS and EIP registers point to the instruction that follows the
INT n instruction or instruction following the instruction on which the INTR signal
occurred.
Program State Change
A program-state change does not accompany interrupts generated by the INT n
instruction or the INTR signal. The INT n instruction generates the interrupt within
the instruction stream. When the processor receives an INTR signal, it commits all
state changes for all previous instructions before it responds to the interrupt; so,
program execution can resume upon returning from the interrupt handler.
Vol. 3 6-67
Download from Www.Somanuals.com. All Manuals Search And Download.
INTERRUPT AND EXCEPTION HANDLING
6-68 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
CHAPTER 7
TASK MANAGEMENT
This chapter describes the IA-32 architecture’s task management facilities. These
facilities are only available when the processor is running in protected mode.
This chapter focuses on 32-bit tasks and the 32-bit TSS structure. For information on
16-bit tasks and the 16-bit TSS structure, see Section 7.6, “16-Bit Task-State
Segment (TSS).” For information specific to task management in 64-bit mode, see
Section 7.7, “Task Management in 64-bit Mode.”
7.1
TASK MANAGEMENT OVERVIEW
A task is a unit of work that a processor can dispatch, execute, and suspend. It can
be used to execute a program, a task or process, an operating-system service utility,
an interrupt or exception handler, or a kernel or executive utility.
The IA-32 architecture provides a mechanism for saving the state of a task, for
dispatching tasks for execution, and for switching from one task to another. When
operating in protected mode, all processor execution takes place from within a task.
Even simple systems must define at least one task. More complex systems can use
7.1.1
Task Structure
A task is made up of two parts: a task execution space and a task-state segment
(TSS). The task execution space consists of a code segment, a stack segment, and
one or more data segments (see Figure 7-1). If an operating system or executive
uses the processor’s privilege-level protection mechanism, the task execution space
also provides a separate stack for each privilege level.
a storage place for task state information. In multitasking systems, the TSS also
provides a mechanism for linking tasks.
A task is identified by the segment selector for its TSS. When a task is loaded into the
processor for execution, the segment selector, base address, limit, and segment
descriptor attributes for the TSS are loaded into the task register (see Section 2.4.4,
“Task Register (TR)”).
If paging is implemented for the task, the base address of the page directory used by
the task is loaded into control register CR3.
Vol. 3 7-1
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
Code
Segment
Data
Segment
Task-State
Segment
(TSS)
Stack
Segment
(Current Priv.
Level)
Stack Seg.
Priv. Level 0
Stack Seg.
Priv. Level 1
Task Register
CR3
Stack
Segment
(Priv. Level 2)
Figure 7-1. Structure of a Task
7.1.2
Task State
The following items define the state of the currently executing task:
• The task’s current execution space, defined by the segment selectors in the
segment registers (CS, DS, SS, ES, FS, and GS).
• The state of the general-purpose registers.
• The state of the EFLAGS register.
• The state of the EIP register.
• The state of control register CR3.
• The state of the task register.
• The state of the LDTR register.
• The I/O map base address and I/O map (contained in the TSS).
• Stack pointers to the privilege 0, 1, and 2 stacks (contained in the TSS).
• Link to previously executed task (contained in the TSS).
Prior to dispatching a task, all of these items are contained in the task’s TSS, except
the state of the task register. Also, the complete contents of the LDTR register are not
contained in the TSS, only the segment selector for the LDT.
7-2 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
7.1.3
Executing a Task
Software or the processor can dispatch a task for execution in one of the following
ways:
• A explicit call to a task with the CALL instruction.
• A explicit jump to a task with the JMP instruction.
• An implicit call (by the processor) to an interrupt-handler task.
• An implicit call to an exception-handler task.
• A return (initiated with an IRET instruction) when the NT flag in the EFLAGS
register is set.
All of these methods for dispatching a task identify the task to be dispatched with a
segment selector that points to a task gate or the TSS for the task. When dispatching
a task with a CALL or JMP instruction, the selector in the instruction may select the
TSS directly or a task gate that holds the selector for the TSS. When dispatching a
task to handle an interrupt or exception, the IDT entry for the interrupt or exception
must contain a task gate that holds the selector for the interrupt- or exception-
handler TSS.
When a task is dispatched for execution, a task switch occurs between the currently
running task and the dispatched task. During a task switch, the execution environ-
ment of the currently executing task (called the task’s state or context) is saved in
its TSS and execution of the task is suspended. The context for the dispatched task is
then loaded into the processor and execution of that task begins with the instruction
pointed to by the newly loaded EIP register. If the task has not been run since the
system was last initialized, the EIP will point to the first instruction of the task’s code;
otherwise, it will point to the next instruction after the last instruction that the task
executed when it was last active.
If the currently executing task (the calling task) called the task being dispatched (the
called task), the TSS segment selector for the calling task is stored in the TSS of the
called task to provide a link back to the calling task.
For all IA-32 processors, tasks are not recursive. A task cannot call or jump to itself.
Interrupts and exceptions can be handled with a task switch to a handler task. Here,
the processor performs a task switch to handle the interrupt or exception and auto-
matically switches back to the interrupted task upon returning from the interrupt-
handler task or exception-handler task. This mechanism can also handle interrupts
that occur during interrupt tasks.
As part of a task switch, the processor can also switch to another LDT, allowing each
task to have a different logical-to-physical address mapping for LDT-based segments.
The page-directory base register (CR3) also is reloaded on a task switch, allowing
each task to have its own set of page tables. These protection facilities help isolate
tasks and prevent them from interfering with one another.
If protection mechanisms are not used, the processor provides no protection
between tasks. This is true even with operating systems that use multiple privilege
levels for protection. A task running at privilege level 3 that uses the same LDT and
Vol. 3 7-3
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
page tables as other privilege-level-3 tasks can access code and corrupt data and the
stack of other tasks.
Use of task management facilities for handling multitasking applications is optional.
Multitasking can be handled in software, with each software defined task executed in
the context of a single IA-32 architecture task.
7.2
TASK MANAGEMENT DATA STRUCTURES
The processor defines five data structures for handling task-related activities:
• Task-state segment (TSS).
• Task-gate descriptor.
• TSS descriptor.
• Task register.
• NT flag in the EFLAGS register.
When operating in protected mode, a TSS and TSS descriptor must be created for at
register (using the LTR instruction).
7.2.1
The processor state information needed to restore a task is saved in a system
segment called the task-state segment (TSS). Figure 7-2 shows the format of a TSS
for tasks designed for 32-bit CPUs. The fields of a TSS are divided into two main cate-
gories: dynamic fields and static fields.
For information about 16-bit Intel 286 processor task structures, see Section 7.6,
“16-Bit Task-State Segment (TSS).” For information about 64-bit mode task struc-
tures, see Section 7.7, “Task Management in 64-bit Mode.”
7-4 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
31
15
0
Reserved
I/O Map Base Address
T
100
96
92
88
84
80
76
72
68
64
60
56
52
48
LDT Segment Selector
Reserved
Reserved
Reserved
Reserved
Reserved
Reserved
GS
FS
DS
SS
CS
ES
Reserved
EDI
ESI
EBP
ESP
EBX
EDX
ECX
EAX
44
40
36
32
28
24
20
EFLAGS
EIP
CR3 (PDBR)
Reserved
Reserved
Reserved
SS2
SS1
ESP2
ESP1
ESP0
16
12
8
SS0
4
Reserved
Previous Task Link
0
Reserved bits. Set to 0.
Figure 7-2. 32-Bit Task-State Segment (TSS)
The processor updates dynamic fields when a task is suspended during a task switch.
The following are dynamic fields:
• General-purpose register fields — State of the EAX, ECX, EDX, EBX, ESP, EBP,
ESI, and EDI registers prior to the task switch.
• Segment selector fields — Segment selectors stored in the ES, CS, SS, DS, FS,
and GS registers prior to the task switch.
• EFLAGS register field — State of the EFAGS register prior to the task switch.
Vol. 3 7-5
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
• EIP (instruction pointer) field — State of the EIP register prior to the task
switch.
• Previous task link field — Contains the segment selector for the TSS of the
previous task (updated on a task switch that was initiated by a call, interrupt, or
exception). This field (which is sometimes called the back link field) permits a
task switch back to the previous task by using the IRET instruction.
The processor reads the static fields, but does not normally change them. These
fields are set up when a task is created. The following are static fields:
• LDT segment selector field — Contains the segment selector for the task's
LDT.
• CR3 control register field — Contains the base physical address of the page
directory to be used by the task. Control register CR3 is also known as the page-
directory base register (PDBR).
• Privilege level-0, -1, and -2 stack pointer fields — These stack pointers
consist of a logical address made up of the segment selector for the stack
ESP2). Note that the values in these fields are static for a particular task;
whereas, the SS and ESP values will change if stack switching occurs within the
task.
• T (debug trap) flag (byte 100, bit 0) — When set, the T flag causes the
processor to raise a debug exception when a task switch to this task occurs (see
Section 16.3.1.5, “Task-Switch Exception Condition”).
TSS to the I/O permission bit map and interrupt redirection bitmap. When
present, these maps are stored in the TSS at higher addresses. The I/O map base
address points to the beginning of the I/O permission bit map and the end of the
interrupt redirection bit map. See Chapter 13, “Input/Output,” in the Intel® 64
and IA-32 Architectures Software Developer’s Manual, Volume 1, for more
information about the I/O permission bit map. See Section 17.3, “Interrupt and
Exception Handling in Virtual-8086 Mode,” for a detailed description of the
interrupt redirection bit map.
If paging is used:
• Avoid placing a page boundary in the part of the TSS that the processor reads
during a task switch (the first 104 bytes). The processor may not correctly
perform address translations if a boundary occurs in this area. During a task
switch, the processor reads and writes into the first 104 bytes of each TSS (using
contiguous physical addresses beginning with the physical address of the first
byte of the TSS). So, after TSS access begins, if part of the 104 bytes is not
physically contiguous, the processor will access incorrect information without
generating a page-fault exception.
• Pages corresponding to the previous task’s TSS, the current task’s TSS, and the
descriptor table entries for each all should be marked as read/write.
7-6 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
• Task switches are carried out faster if the pages containing these structures are
present in memory before the task switch is initiated.
7.2.2
TSS Descriptor
The TSS, like all other segments, is defined by a segment descriptor. Figure 7-3
shows the format of a TSS descriptor. TSS descriptors may only be placed in the GDT;
they cannot be placed in an LDT or the IDT.
An attempt to access a TSS using a segment selector with its TI flag set (which indi-
cates the current LDT) causes a general-protection exception (#GP) to be generated
during CALLs and JMPs; it causes an invalid TSS exception (#TS) during IRETs. A
general-protection exception is also generated if an attempt is made to load a
segment selector for a TSS into a segment register.
The busy flag (B) in the type field indicates whether the task is busy. A busy task is
currently running or suspended. A type field with a value of 1001B indicates an inac-
tive task; a value of 1011B indicates a busy task. Tasks are not recursive. The
processor uses the busy flag to detect an attempt to call a task whose execution has
been interrupted. To insure that there is only one busy flag is associated with a task,
each TSS should have only one TSS descriptor that points to it.
TSS Descriptor
31
31
2120 19
16
24 23 22
15 14 13 12 11
0
0
8
7
A
V
L
D
P
L
Type
Limit
19:16
4
0
G
0
Base 31:24
0
P
Base 23:16
0
1
0
B
1
16
15
Base Address 15:00
Segment Limit 15:00
AVL
B
Available for use by system software
Busy flag
BASE Segment Base Address
DPL
G
Descriptor Privilege Level
Granularity
LIMIT Segment Limit
Segment Present
TYPE Segment Type
P
Figure 7-3. TSS Descriptor
The base, limit, and DPL fields and the granularity and present flags have functions
similar to their use in data-segment descriptors (see Section 3.4.5, “Segment
Descriptors”). When the G flag is 0 in a TSS descriptor for a 32-bit TSS, the limit field
must have a value equal to or greater than 67H, one byte less than the minimum size
Vol. 3 7-7
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
of a TSS. Attempting to switch to a task whose TSS descriptor has a limit less than
67H generates an invalid-TSS exception (#TS). A larger limit is required if an I/O
permission bit map is included or if the operating system stores additional data. The
processor does not check for a limit greater than 67H on a task switch; however, it
does check when accessing the I/O permission bit map or interrupt redirection bit
map.
Any program or procedure with access to a TSS descriptor (that is, whose CPL is
numerically equal to or less than the DPL of the TSS descriptor) can dispatch the task
with a call or a jump.
In most systems, the DPLs of TSS descriptors are set to values less than 3, so that
only privileged software can perform task switching. However, in multitasking appli-
application (or user) privilege level.
7.2.3
TSS Descriptor in 64-bit mode
In 64-bit mode, task switching is not supported, but TSS descriptors still exist. The
format of a 64-bit TSS is described in Section 7.7.
In 64-bit mode, the TSS descriptor is expanded to 16 bytes (see Figure 7-4). This
expansion also applies to an LDT descriptor in 64-bit mode. Table 3-2 provides the
encoding information for the segment type field.
7-8 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
TSS (or LDT) Descriptor
31
31
13 12
8
7
0
Reserved
0
12
Reserved
0
Base Address 63:32
8
31
31
23
G
21 20 19
A
V
L
16
16
24
22
0
15 14 13 12 11
7
0
8
D
P
L
Type
Limit
19:16
4
Base 31:24
0
P
Base 23:16
0
15
0
Base Address 15:00
Segment Limit 15:00
0
AVL
B
Available for use by system software
Busy flag
BASE Segment Base Address
DPL
G
Descriptor Privilege Level
Granularity
LIMIT Segment Limit
P
TYPE
Segment Present
Segment Type
7.2.4
Task Register
The task register holds the 16-bit segment selector and the entire segment
descriptor (32-bit base address, 16-bit segment limit, and descriptor attributes) for
the TSS of the current task (see Figure 2-5). This information is copied from the TSS
descriptor in the GDT for the current task. Figure 7-5 shows the path the processor
uses to access the TSS (using the information in the task register).
The task register has a visible part (that can be read and changed by software) and
an invisible part (maintained by the processor and is inaccessible by software). The
segment selector in the visible portion points to a TSS descriptor in the GDT. The
processor uses the invisible portion of the task register to cache the segment
descriptor for the TSS. Caching these values in a register makes execution of the task
more efficient. The LTR (load task register) and STR (store task register) instructions
load and read the visible portion of the task register:
Vol. 3 7-9
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
The LTR instruction loads a segment selector (source operand) into the task register
that points to a TSS descriptor in the GDT. It then loads the invisible portion of the
task register with information from the TSS descriptor. LTR is a privileged instruction
that may be executed only when the CPL is 0. It’s used during system initialization to
put an initial value in the task register. Afterwards, the contents of the task register
are changed implicitly when a task switch occurs.
The STR (store task register) instruction stores the visible portion of the task register
in a general-purpose register or memory. This instruction can be executed by code
running at any privilege level in order to identify the currently running task. However,
it is normally used only by operating system software.
On power up or reset of the processor, segment selector and base address are set to
the default value of 0; the limit is set to FFFFH.
TSS
+
Visible Part
Selector
Invisible Part
Base Address Segment Limit
Task
Register
GDT
TSS Descriptor
0
Figure 7-5. Task Register
7-10 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
7.2.5
Task-Gate Descriptor
A task-gate descriptor provides an indirect, protected reference to a task (see
Figure 7-6). It can be placed in the GDT, an LDT, or the IDT. The TSS segment
selector field in a task-gate descriptor points to a TSS descriptor in the GDT. The RPL
in this segment selector is not used.
The DPL of a task-gate descriptor controls access to the TSS descriptor during a task
switch. When a program or procedure makes a call or jump to a task through a task
gate, the CPL and the RPL field of the gate selector pointing to the task gate must be
less than or equal to the DPL of the task-gate descriptor. Note that when a task gate
is used, the DPL of the destination TSS descriptor is not used.
31
31
16
16
15 14 13 12 11
D
0
0
8
7
Type
Reserved
Reserved
P
4
0
P
L
0
0
1
0
1
15
TSS Segment Selector
Reserved
DPL
P
Descriptor Privilege Level
Segment Present
TYPE Segment Type
Figure 7-6. Task-Gate Descriptor
A task can be accessed either through a task-gate descriptor or a TSS descriptor.
Both of these structures satisfy the following needs:
• Need for a task to have only one busy flag — Because the busy flag for a task
is stored in the TSS descriptor, each task should have only one TSS descriptor.
There may, however, be several task gates that reference the same TSS
descriptor.
• Need to provide selective access to tasks — Task gates fill this need, because
they can reside in an LDT and can have a DPL that is different from the TSS
descriptor's DPL. A program or procedure that does not have sufficient privilege
to access the TSS descriptor for a task in the GDT (which usually has a DPL of 0)
may be allowed access to the task through a task gate with a higher DPL. Task
gates give the operating system greater latitude for limiting access to specific
tasks.
• Need for an interrupt or exception to be handled by an independent task
— Task gates may also reside in the IDT, which allows interrupts and exceptions
Vol. 3 7-11
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
to be handled by handler tasks. When an interrupt or exception vector points to
a task gate, the processor switches to the specified task.
Figure 7-7 illustrates how a task gate in an LDT, a task gate in the GDT, and a task
gate in the IDT can all point to the same task.
LDT
GDT
TSS
Task Gate
Task Gate
TSS Descriptor
IDT
Task Gate
Figure 7-7. Task Gates Referencing the Same Task
7.3
TASK SWITCHING
The processor transfers execution to another task in one of four cases:
• The current program, task, or procedure executes a JMP or CALL instruction to a
TSS descriptor in the GDT.
• The current program, task, or procedure executes a JMP or CALL instruction to a
task-gate descriptor in the GDT or the current LDT.
7-12 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
• An interrupt or exception vector points to a task-gate descriptor in the IDT.
• The current task executes an IRET when the NT flag in the EFLAGS register is set.
JMP, CALL, and IRET instructions, as well as interrupts and exceptions, are all mech-
anisms for redirecting a program. The referencing of a TSS descriptor or a task gate
(when calling or jumping to a task) or the state of the NT flag (when executing an
IRET instruction) determines whether a task switch occurs.
The processor performs the following operations when switching to a new task:
1. Obtains the TSS segment selector for the new task as the operand of the JMP or
CALL instruction, from a task gate, or from the previous task link field (for a task
switch initiated with an IRET instruction).
2. Checks that the current (old) task is allowed to switch to the new task. Data-
access privilege rules apply to JMP and CALL instructions. The CPL of the current
(old) task and the RPL of the segment selector for the new task must be less than
or equal to the DPL of the TSS descriptor or task gate being referenced.
Exceptions, interrupts (except for interrupts generated by the INT n instruction),
and the IRET instruction are permitted to switch tasks regardless of the DPL of
the destination task-gate or TSS descriptor. For interrupts generated by the INT n
instruction, the DPL is checked.
3. Checks that the TSS descriptor of the new task is marked present and has a valid
limit (greater than or equal to 67H).
4. Checks that the new task is available (call, jump, exception, or interrupt) or busy
(IRET return).
5. Checks that the current (old) TSS, new TSS, and all segment descriptors used in
the task switch are paged into system memory.
6. If the task switch was initiated with a JMP or IRET instruction, the processor
clears the busy (B) flag in the current (old) task’s TSS descriptor; if initiated with
a CALL instruction, an exception, or an interrupt: the busy (B) flag is left set.
(See Table 7-2.)
7. If the task switch was initiated with an IRET instruction, the processor clears the
NT flag in a temporarily saved image of the EFLAGS register; if initiated with a
CALL or JMP instruction, an exception, or an interrupt, the NT flag is left
unchanged in the saved EFLAGS image.
8. Saves the state of the current (old) task in the current task’s TSS. The processor
finds the base address of the current TSS in the task register and then copies the
states of the following registers into the current TSS: all the general-purpose
registers, segment selectors from the segment registers, the temporarily saved
image of the EFLAGS register, and the instruction pointer register (EIP).
9. If the task switch was initiated with a CALL instruction, an exception, or an
interrupt, the processor will set the NT flag in the EFLAGS loaded from the new
task. If initiated with an IRET instruction or JMP instruction, the NT flag will reflect
the state of NT in the EFLAGS loaded from the new task (see Table 7-2).
Vol. 3 7-13
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
10. If the task switch was initiated with a CALL instruction, JMP instruction, an
exception, or an interrupt, the processor sets the busy (B) flag in the new task’s
TSS descriptor; if initiated with an IRET instruction, the busy (B) flag is left set.
11. Loads the task register with the segment selector and descriptor for the new
task's TSS.
12. The TSS state is loaded into the processor. This includes the LDTR register, the
PDBR (control register CR3), the EFLAGS registers, the EIP register, the general-
purpose registers, and the segment selectors. Note that a fault during the load of
this state may corrupt architectural state.
13. The descriptors associated with the segment selectors are loaded and qualified.
Any errors associated with this loading and qualification occur in the context of
the new task.
NOTES
If all checks and saves have been carried out successfully, the
processor commits to the task switch. If an unrecoverable error
occurs in steps 1 through 11, the processor does not complete the
task switch and insures that the processor is returned to its state
prior to the execution of the instruction that initiated the task switch.
If an unrecoverable error occurs in step 12, architectural state may
be corrupted, but an attempt will be made to handle the error in the
prior execution environment. If an unrecoverable error occurs after
the commit point (in step 13), the processor completes the task
ability checks) and generates the appropriate exception prior to
beginning execution of the new task.
If exceptions occur after the commit point, the exception handler
must finish the task switch itself before allowing the processor to
begin executing the new task. See Chapter 6, “Interrupt 10—Invalid
TSS Exception (#TS),” for more information about the affect of
exceptions on a task when they occur after the commit point of a task
switch.
14. Begins executing the new task. (To an exception handler, the first instruction of
the new task appears not to have been executed.)
The state of the currently executing task is always saved when a successful task
switch occurs. If the task is resumed, execution starts with the instruction pointed to
by the saved EIP value, and the registers are restored to the values they held when
the task was suspended.
When switching tasks, the privilege level of the new task does not inherit its privilege
level from the suspended task. The new task begins executing at the privilege level
specified in the CPL field of the CS register, which is loaded from the TSS. Because
tasks are isolated by their separate address spaces and TSSs and because privilege
7-14 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
rules control access to a TSS, software does not need to perform explicit privilege
checks on a task switch.
Table 7-1 shows the exception conditions that the processor checks for when
switching tasks. It also shows the exception that is generated for each check if an
error is detected and the segment that the error code references. (The order of the
checks in the table is the order used in the P6 family processors. The exact order is
model specific and may be different for other IA-32 processors.) Exception handlers
designed to handle these exceptions may be subject to recursive calls if they attempt
to reload the segment selector that generated the exception. The cause of the excep-
tion (or the first of multiple causes) should be fixed before reloading the selector.
Table 7-1. Exception Conditions Checked During a Task Switch
Condition Checked
Exception1
Error Code
Reference2
Segment selector for a TSS descriptor references
the GDT and is within the limits of the table.
#GP
New Task’s TSS
#TS (for IRET)
#NP
TSS descriptor is present in memory.
New Task’s TSS
TSS descriptor is not busy (for task switch initiated #GP (for JMP, CALL,
by a call, interrupt, or exception). INT)
Task’s back-link TSS
TSS descriptor is not busy (for task switch initiated #TS (for IRET)
by an IRET instruction).
New Task’s TSS
New Task’s TSS
TSS segment limit greater than or equal to 108 (for #TS
32-bit TSS) or 44 (for 16-bit TSS).
Registers are loaded from the values in the TSS.
LDT segment selector of new task is valid 3.
#TS
New Task’s LDT
Code segment DPL matches segment selector RPL. #TS
New Code Segment
New Stack Segment
New Stack Segment
New stack segment
New Task’s LDT
SS segment selector is valid 2
.
#TS
#SS
#TS
#TS
#TS
#NP
#TS
#TS
#TS
Stack segment is present in memory.
Stack segment DPL matches CPL.
LDT of new task is present in memory.
CS segment selector is valid 3
.
New Code Segment
New Code Segment
New Stack Segment
New Data Segment
New Data Segment
Code segment is present in memory.
Stack segment DPL matches selector RPL.
DS, ES, FS, and GS segment selectors are valid 3.
DS, ES, FS, and GS segments are readable.
Vol. 3 7-15
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
Table 7-1. Exception Conditions Checked During a Task Switch (Contd.)
Condition Checked
Exception1
Error Code
Reference2
DS, ES, FS, and GS segments are present in memory. #NP
New Data Segment
New Data Segment
DS, ES, FS, and GS segment DPL greater than or
equal to CPL (unless these are
#TS
conforming segments).
NOTES:
1. #NP is segment-not-present exception, #GP is general-protection exception, #TS is invalid-TSS
exception, and #SS is stack-fault exception.
2. The error code contains an index to the segment descriptor referenced in this column.
3. A segment selector is valid if it is in a compatible type of table (GDT or LDT), occupies an address
within the table's segment limit, and refers to a compatible type of descriptor (for example, a seg-
The TS (task switched) flag in the control register CR0 is set every time a task switch
occurs. System software uses the TS flag to coordinate the actions of floating-point
unit when generating floating-point exceptions with the rest of the processor. The TS
flag indicates that the context of the floating-point unit may be different from that of
the current task. See Section 2.5, “Control Registers”, for a detailed description of
the function and use of the TS flag.
7.4
TASK LINKING
The previous task link field of the TSS (sometimes called the “backlink”) and the NT
flag in the EFLAGS register are used to return execution to the previous task.
EFLAGS.NT = 1 indicates that the currently executing task is nested within the
execution of another task.
When a CALL instruction, an interrupt, or an exception causes a task switch: the
processor copies the segment selector for the current TSS to the previous task link
field of the TSS for the new task; it then sets EFLAGS.NT = 1. If software uses an
IRET instruction to suspend the new task, the processor checks for EFLAGS.NT = 1;
it then uses the value in the previous task link field to return to the previous task. See
Figures 7-8.
When a JMP instruction causes a task switch, the new task is not nested. The
previous task link field is not used and EFLAGS.NT = 0. Use a JMP instruction to
dispatch a new task when nesting is not desired.
7-16 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
Top Level
Task
Nested
Task
More Deeply
Nested Task
Currently Executing
Task
TSS
TSS
TSS
EFLAGS
NT=1
NT=0
NT=1
NT=1
Previous
Previous
Task Link
Previous
Task Link
Task Register
Figure 7-8. Nested Tasks
Table 7-2 shows the busy flag (in the TSS segment descriptor), the NT flag, the
previous task link field, and TS flag (in control register CR0) during a task switch.
The NT flag may be modified by software executing at any privilege level. It is
possible for a program to set the NT flag and execute an IRET instruction. This might
randomly invoke the task specified in the previous link field of the current task's TSS.
To keep such spurious task switches from succeeding, the operating system should
initialize the previous task link field in every TSS that it creates to 0.
Table 7-2. Effect of a Task Switch on Busy Flag, NT Flag,
Previous Task Link Field, and TS Flag
Flag or Field
Effect of JMP
instruction
Effect of CALL
Instruction or
Interrupt
Effect of IRET
Instruction
Busy (B) flag of new
task.
Flag is set. Must have
been clear before.
Flag is set. Must have No change. Must have
been clear before.
been set.
Busy flag of old task.
NT flag of new task.
NT flag of old task.
Flag is cleared.
No change. Flag is
currently set.
Flag is cleared.
Set to value from TSS Flag is set.
of new task.
Set to value from TSS
of new task.
No change.
No change.
Flag is cleared.
Previous task link field No change.
of new task.
Loaded with selector No change.
for old task’s TSS.
Previous task link field No change.
of old task.
No change.
No change.
TS flag in control
register CR0.
Flag is set.
Flag is set.
Flag is set.
Vol. 3 7-17
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
7.4.1
Use of Busy Flag To Prevent Recursive Task Switching
A TSS allows only one context to be saved for a task; therefore, once a task is called
(dispatched), a recursive (or re-entrant) call to the task would cause the current
state of the task to be lost. The busy flag in the TSS segment descriptor is provided
to prevent re-entrant task switching and a subsequent loss of task state information.
The processor manages the busy flag as follows:
1. When dispatching a task, the processor sets the busy flag of the new task.
2. If during a task switch, the current task is placed in a nested chain (the task
switch is being generated by a CALL instruction, an interrupt, or an exception),
the busy flag for the current task remains set.
3. When switching to the new task (initiated by a CALL instruction, interrupt, or
exception), the processor generates a general-protection exception (#GP) if the
busy flag of the new task is already set. If the task switch is initiated with an IRET
instruction, the exception is not raised because the processor expects the busy
flag to be set.
4. When a task is terminated by a jump to a new task (initiated with a JMP
instruction in the task code) or by an IRET instruction in the task code, the
processor clears the busy flag, returning the task to the “not busy” state.
The processor prevents recursive task switching by preventing a task from switching
may grow to any length, due to multiple calls, interrupts, or exceptions. The busy
flag prevents a task from being invoked if it is in this chain.
The busy flag may be used in multiprocessor configurations, because the processor
follows a LOCK protocol (on the bus or in the cache) when it sets or clears the busy
flag. This lock keeps two processors from invoking the same task at the same time.
See Section 8.1.2.1, “Automatic Locking,” for more information about setting the
busy flag in a multiprocessor applications.
7.4.2
Modifying Task Linkages
In a uniprocessor system, in situations where it is necessary to remove a task from a
chain of linked tasks, use the following procedure to remove the task:
1. Disable interrupts.
2. Change the previous task link field in the TSS of the pre-empting task (the task
that suspended the task to be removed). It is assumed that the pre-empting task
is the next task (newer task) in the chain from the task to be removed. Change
the previous task link field to point to the TSS of the next oldest task in the chain
or to an even older task in the chain.
3. Clear the busy (B) flag in the TSS segment descriptor for the task being removed
from the chain. If more than one task is being removed from the chain, the busy
flag for each task being remove must be cleared.
4. Enable interrupts.
7-18 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
In a multiprocessing system, additional synchronization and serialization operations
must be added to this procedure to insure that the TSS and its segment descriptor
are both locked when the previous task link field is changed and the busy flag is
cleared.
7.5
TASK ADDRESS SPACE
The address space for a task consists of the segments that the task can access.
These segments include the code, data, stack, and system segments referenced in
the TSS and any other segments accessed by the task code. The segments are
mapped into the processor’s linear address space, which is in turn mapped into the
processor’s physical address space (either directly or through paging).
The LDT segment field in the TSS can be used to give each task its own LDT. Giving a
task its own LDT allows the task address space to be isolated from other tasks by
placing the segment descriptors for all the segments associated with the task in the
task’s LDT.
It also is possible for several tasks to use the same LDT. This is a memory-efficient
way to allow specific tasks to communicate with or control each other, without drop-
ping the protection barriers for the entire system.
Because all tasks have access to the GDT, it also is possible to create shared
segments accessed through segment descriptors in this table.
If paging is enabled, the CR3 register (PDBR) field in the TSS allows each task to
have its own set of page tables for mapping linear addresses to physical addresses.
Or, several tasks can share the same set of page tables.
7.5.1
Mapping Tasks to the Linear and Physical Address Spaces
Tasks can be mapped to the linear address space and physical address space in one
of two ways:
• One linear-to-physical address space mapping is shared among all tasks.
— When paging is not enabled, this is the only choice. Without paging, all linear
addresses map to the same physical addresses. When paging is enabled, this
form of linear-to-physical address space mapping is obtained by using one page
directory for all tasks. The linear address space may exceed the available
physical space if demand-paged virtual memory is supported.
• Each task has its own linear address space that is mapped to the physical
address space. — This form of mapping is accomplished by using a different
page directory for each task. Because the PDBR (control register CR3) is loaded
on task switches, each task may have a different page directory.
The linear address spaces of different tasks may map to completely distinct physical
addresses. If the entries of different page directories point to different page tables
Vol. 3 7-19
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
and the page tables point to different pages of physical memory, then the tasks do
not share physical addresses.
With either method of mapping task linear address spaces, the TSSs for all tasks
must lie in a shared area of the physical space, which is accessible to all tasks. This
mapping is required so that the mapping of TSS addresses does not change while the
processor is reading and updating the TSSs during a task switch. The linear address
space mapped by the GDT also should be mapped to a shared area of the physical
space; otherwise, the purpose of the GDT is defeated. Figure 7-9 shows how the
linear address spaces of two tasks can overlap in the physical space by sharing page
tables.
TSS
Page Directories
Page Tables
Page Frames
Task A
Task A TSS
Task A
Task A
Shared
Shared
Task B
Task B
PTE
PTE
PTE
PDBR
PDE
PDE
Shared PT
PTE
PTE
Task B TSS
PDBR
PDE
PDE
PTE
PTE
Figure 7-9. Overlapping Linear-to-Physical Mappings
7.5.2
Task Logical Address Space
To allow the sharing of data among tasks, use the following techniques to create
shared logical-to-physical address-space mappings for data segments:
• Through the segment descriptors in the GDT — All tasks must have access
to the segment descriptors in the GDT. If some segment descriptors in the GDT
point to segments in the linear-address space that are mapped into an area of the
physical-address space common to all tasks, then all tasks can share the data
and code in those segments.
• Through a shared LDT — Two or more tasks can use the same LDT if the LDT
fields in their TSSs point to the same LDT. If some segment descriptors in a
7-20 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
shared LDT point to segments that are mapped to a common area of the physical
address space, the data and code in those segments can be shared among the
tasks that share the LDT. This method of sharing is more selective than sharing
through the GDT, because the sharing can be limited to specific tasks. Other
tasks in the system may have different LDTs that do not give them access to the
shared segments.
• Through segment descriptors in distinct LDTs that are mapped to
common addresses in linear address space — If this common area of the
linear address space is mapped to the same area of the physical address space
for each task, these segment descriptors permit the tasks to share segments.
Such segment descriptors are commonly called aliases. This method of sharing is
even more selective than those listed above, because, other segment descriptors
7.6
16-BIT TASK-STATE SEGMENT (TSS)
The 32-bit IA-32 processors also recognize a 16-bit TSS format like the one used in
Intel 286 processors (see Figure 7-10). This format is supported for compatibility
with software written to run on earlier IA-32 processors.
The following information is important to know about the 16-bit TSS.
• Do not use a 16-bit TSS to implement a virtual-8086 task.
• The valid segment limit for a 16-bit TSS is 2CH.
• The 16-bit TSS does not contain a field for the base address of the page directory,
which is loaded into control register CR3. A separate set of page tables for each
task is not supported for 16-bit tasks. If a 16-bit task is dispatched, the page-
table structure for the previous task is used.
• The I/O base address is not included in the 16-bit TSS. None of the functions of
the I/O map are supported.
• When task state is saved in a 16-bit TSS, the upper 16 bits of the EFLAGS register
and the EIP register are lost.
• When the general-purpose registers are loaded or saved from a 16-bit TSS, the
upper 16 bits of the registers are modified and not maintained.
Vol. 3 7-21
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
15
0
Task LDT Selector
42
40
DS Selector
SS Selector
CS Selector
38
36
34
ES Selector
DI
32
30
28
26
24
22
SI
BP
SP
BX
DX
CX
20
18
16
14
12
AX
FLAG Word
IP (Entry Point)
SS2
10
8
SP2
SS1
6
SP1
SS0
4
2
SP0
Previous Task Link
0
Figure 7-10. 16-Bit TSS Format
7.7
TASK MANAGEMENT IN 64-BIT MODE
In 64-bit mode, task structure and task state are similar to those in protected mode.
However, the task switching mechanism available in protected mode is not supported
in 64-bit mode. Task management and switching must be performed by software.
The processor issues a general-protection exception (#GP) if the following is
attempted in 64-bit mode:
• Control transfer to a TSS or a task gate using JMP, CALL, INTn, or interrupt.
• An IRET with EFLAGS.NT (nested task) set to 1.
7-22 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
Although hardware task-switching is not supported in 64-bit mode, a 64-bit task
state segment (TSS) must exist. Figure 7-11 shows the format of a 64-bit TSS. The
TSS holds information important to 64-bit mode and that is not directly related to the
task-switch mechanism. This information includes:
• RSPn — The full 64-bit canonical forms of the stack pointers (RSP) for privilege
levels 0-2.
• ISTn — The full 64-bit canonical forms of the interrupt stack table (IST) pointers.
• I/O map base address — The 16-bit offset to the I/O permission bit map from
the 64-bit TSS base.
The operating system must create at least one 64-bit TSS after activating IA-32e
mode. It must execute the LTR instruction (in 64-bit mode) to load the TR register
with a pointer to the 64-bit TSS responsible for both 64-bit-mode programs and
compatibility-mode programs.
Vol. 3 7-23
Download from Www.Somanuals.com. All Manuals Search And Download.
TASK MANAGEMENT
31
15
0
Reserved
I/O Map Base Address
100
96
92
88
84
80
76
72
68
64
60
56
52
48
Reserved
Reserved
IST7 (upper 32 bits)
IST7 (lower 32 bits)
IST6 (upper 32 bits)
IST6 (lower 32 bits)
IST5 (upper 32 bits)
IST5 (lower 32 bits)
IST4 (upper 32 bits)
IST4 (lower 32 bits)
IST3 (upper 32 bits)
IST3 (lower 32 bits)
IST2 (upper 32 bits)
IST2 (lower 32 bits)
44
40
36
32
28
24
20
IST1 (upper 32 bits)
IST1 (lower 32 bits)
Reserved
Reserved
RSP2 (upper 32 bits)
RSP2 (lower 32 bits)
RSP1 (upper 32 bits)
16
12
8
RSP1 (lower 32 bits)
RSP0 (upper 32 bits)
RSP0 (lower 32 bits)
Reserved
4
0
Reserved bits. Set to 0.
Figure 7-11. 64-Bit TSS Format
7-24 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
CHAPTER 8
MULTIPLE-PROCESSOR MANAGEMENT
The Intel 64 and IA-32 architectures provide mechanisms for managing and
improving the performance of multiple processors connected to the same system
bus. These include:
• Bus locking and/or cache coherency management for performing atomic
operations on system memory.
• Serializing instructions. These instructions apply only to the Pentium 4, Intel
Xeon, P6 family, and Pentium processors.
• An advance programmable interrupt controller (APIC) located on the processor
chip (see Chapter 10, “Advanced Programmable Interrupt Controller (APIC)”).
This feature was introduced by the Pentium processor.
• A second-level cache (level 2, L2). For the Pentium 4, Intel Xeon, and P6 family
processors, the L2 cache is included in the processor package and is tightly
provided to support an external L2 cache.
• A third-level cache (level 3, L3). For Intel Xeon processors, the L3 cache is
included in the processor package and is tightly coupled to the processor.
• Intel Hyper-Threading Technology. This extension to the Intel 64 and IA-32 archi-
tectures enables a single processor core to execute two or more threads concur-
®
®
rently (see Section 8.5, “Intel Hyper-Threading Technology and Intel Multi-
Core Technology”).
These mechanisms are particularly useful in symmetric-multiprocessing (SMP)
systems. However, they can also be used when an Intel 64 or IA-32 processor and a
special-purpose processor (such as a communications, graphics, or video processor)
share the system bus.
These multiprocessing mechanisms have the following characteristics:
• To maintain system memory coherency — When two or more processors are
attempting simultaneously to access the same address in system memory, some
communication mechanism or memory access protocol must be available to
promote data coherency and, in some instances, to allow one processor to
temporarily lock a memory location.
• To maintain cache consistency — When one processor accesses data cached on
another processor, it must not receive incorrect data. If it modifies data, all other
processors that access that data must receive the modified data.
• To allow predictable ordering of writes to memory — In some circumstances, it is
important that memory writes be observed externally in precisely the same order
as programmed.
Vol. 3 8-1
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
• To distribute interrupt handling among a group of processors — When several
mechanism for receiving interrupts and distributing them to available processors
for servicing.
• To increase system performance by exploiting the multi-threaded and multi-
process nature of contemporary operating systems and applications.
The caching mechanism and cache consistency of Intel 64 and IA-32 processors are
discussed in Chapter 11. The APIC architecture is described in Chapter 10. Bus and
memory locking, serializing instructions, memory ordering, and Intel Hyper-
Threading Technology are discussed in the following sections.
8.1
LOCKED ATOMIC OPERATIONS
The 32-bit IA-32 processors support locked atomic operations on locations in system
memory. These operations are typically used to manage shared data structures (such
as semaphores, segment descriptors, system segments, or page tables) in which two
or more processors may try simultaneously to modify the same field or flag. The
processor uses three interdependent mechanisms for carrying out locked atomic
operations:
• Guaranteed atomic operations
• Bus locking, using the LOCK# signal and the LOCK instruction prefix
• Cache coherency protocols that insure that atomic operations can be carried out
on cached data structures (cache lock); this mechanism is present in the
Pentium 4, Intel Xeon, and P6 family processors
These mechanisms are interdependent in the following ways. Certain basic memory
transactions (such as reading or writing a byte in system memory) are always guar-
anteed to be handled atomically. That is, once started, the processor guarantees that
the operation will be completed before another processor or bus agent is allowed
access to the memory location. The processor also supports bus locking for
performing selected memory operations (such as a read-modify-write operation in a
shared area of memory) that typically need to be handled atomically, but are not
automatically handled this way. Because frequently used memory locations are often
cached in a processor’s L1 or L2 caches, atomic operations can often be carried out
inside a processor’s caches without asserting the bus lock. Here the processor’s
cache coherency protocols insure that other processors that are caching the same
memory locations are managed properly while atomic operations are performed on
cached memory locations.
NOTE
Where there are contested lock accesses, software may need to
implement algorithms that ensure fair access to resources in order to
prevent lock starvation. The hardware provides no resource that
guarantees fairness to participating agents. It is the responsibility of
8-2 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
software to manage the fairness of semaphores and exclusive locking
functions.
The mechanisms for handling locked atomic operations have evolved with the
complexity of IA-32 processors. More recent IA-32 processors (such as the
Pentium 4, Intel Xeon, and P6 family processors) and Intel 64 provide a more refined
locking mechanism than earlier processors. These mechanisms are described in the
following sections.
8.1.1
Guaranteed Atomic Operations
The Intel486 processor (and newer processors since) guarantees that the following
basic memory operations will always be carried out atomically:
• Reading or writing a byte
• Reading or writing a word aligned on a 16-bit boundary
• Reading or writing a doubleword aligned on a 32-bit boundary
The Pentium processor (and newer processors since) guarantees that the following
additional memory operations will always be carried out atomically:
• Reading or writing a quadword aligned on a 64-bit boundary
• 16-bit accesses to uncached memory locations that fit within a 32-bit data bus
The P6 family processors (and newer processors since) guarantee that the following
additional memory operation will always be carried out atomically:
• Unaligned 16-, 32-, and 64-bit accesses to cached memory that fit within a cache
line
Accesses to cacheable memory that are split across bus widths, cache lines, and
page boundaries are not guaranteed to be atomic by the Intel Core 2 Duo, Intel
Atom, Intel Core Duo, Pentium M, Pentium 4, Intel Xeon, P6 family, Pentium, and
Intel486 processors. The Intel Core 2 Duo, Intel Atom, Intel Core Duo, Pentium M,
Pentium 4, Intel Xeon, and P6 family processors provide bus control signals that
permit external memory subsystems to make split accesses atomic; however,
nonaligned data accesses will seriously impact the performance of the processor and
should be avoided.
8.1.2
Bus Locking
Intel 64 and IA-32 processors provide a LOCK# signal that is asserted automatically
during certain critical memory operations to lock the system bus or equivalent link.
While this output signal is asserted, requests from other processors or bus agents for
control of the bus are blocked. Software can specify other occasions when the LOCK
semantics are to be followed by prepending the LOCK prefix to an instruction.
In the case of the Intel386, Intel486, and Pentium processors, explicitly locked
instructions will result in the assertion of the LOCK# signal. It is the responsibility of
Vol. 3 8-3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
the hardware designer to make the LOCK# signal available in system hardware to
control memory accesses among processors.
For the P6 and more recent processor families, if the memory area being accessed is
cached internally in the processor, the LOCK# signal is generally not asserted;
instead, locking is only applied to the processor’s caches (see Section 8.1.4, “Effects
of a LOCK Operation on Internal Processor Caches”).
8.1.2.1
Automatic Locking
The operations on which the processor automatically follows the LOCK semantics are
as follows:
• When executing an XCHG instruction that references memory.
• When setting the B (busy) flag of a TSS descriptor — The processor tests
and sets the busy flag in the type field of the TSS descriptor when switching to a
task. To insure that two processors do not switch to the same task simulta-
neously, the processor follows the LOCK semantics while testing and setting this
flag.
• When updating segment descriptors — When loading a segment descriptor,
the processor will set the accessed flag in the segment descriptor if the flag is
clear. During this operation, the processor follows the LOCK semantics so that the
descriptor will not be modified by another processor while it is being updated. For
this action to be effective, operating-system procedures that update descriptors
should use the following steps:
— Use a locked operation to modify the access-rights byte to indicate that the
segment descriptor is not-present, and specify a value for the type field that
indicates that the descriptor is being updated.
— Update the fields of the segment descriptor. (This operation may require
several memory accesses; therefore, locked operations cannot be used.)
— Use a locked operation to modify the access-rights byte to indicate that the
segment descriptor is valid and present.
• The Intel386 processor always updates the accessed flag in the segment
descriptor, whether it is clear or not. The Pentium 4, Intel Xeon, P6 family,
Pentium, and Intel486 processors only update this flag if it is not already set.
• When updating page-directory and page-table entries — When updating
page-directory and page-table entries, the processor uses locked cycles to set
the accessed and dirty flag in the page-directory and page-table entries.
• Acknowledging interrupts — After an interrupt request, an interrupt controller
may use the data bus to send the interrupt vector for the interrupt to the
processor. The processor follows the LOCK semantics during this time to ensure
that no other data appears on the data bus when the interrupt vector is being
transmitted.
8-4 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
8.1.2.2
Software Controlled Bus Locking
To explicitly force the LOCK semantics, software can use the LOCK prefix with the
following instructions when they are used to modify a memory location. An invalid-
opcode exception (#UD) is generated when the LOCK prefix is used with any other
instruction or when no write operation is made to memory (that is, when the destina-
tion operand is in a register).
• The bit test and modify instructions (BTS, BTR, and BTC).
• The exchange instructions (XADD, CMPXCHG, and CMPXCHG8B).
• The LOCK prefix is automatically assumed for XCHG instruction.
• The following single-operand arithmetic and logical instructions: INC, DEC, NOT,
and NEG.
• The following two-operand arithmetic and logical instructions: ADD, ADC, SUB,
SBB, AND, OR, and XOR.
A locked instruction is guaranteed to lock only the area of memory defined by the
destination operand, but may be interpreted by the system as a lock for a larger
memory area.
Software should access semaphores (shared memory used for signalling between
multiple processors) using identical addresses and operand lengths. For example, if
one processor accesses a semaphore using a word access, other processors should
not access the semaphore using a byte access.
NOTE
Do not implement semaphores using the WC memory type. Do not
perform non-temporal stores to a cache line containing a location
used to implement a semaphore.
The integrity of a bus lock is not affected by the alignment of the memory field. The
LOCK semantics are followed for as many bus cycles as necessary to update the
entire operand. However, it is recommend that locked accesses be aligned on their
natural boundaries for better system performance:
• Any boundary for an 8-bit access (locked or otherwise).
• 16-bit boundary for locked word accesses.
• 32-bit boundary for locked doubleword accesses.
• 64-bit boundary for locked quadword accesses.
Locked operations are atomic with respect to all other memory operations and all
externally visible events. Only instruction fetch and page table accesses can pass
locked instructions. Locked instructions can be used to synchronize data written by
one processor and read by another processor.
For the P6 family processors, locked operations serialize all outstanding load and
store operations (that is, wait for them to complete). This rule is also true for the
Pentium 4 and Intel Xeon processors, with one exception. Load operations that refer-
Vol. 3 8-5
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
ence weakly ordered memory types (such as the WC memory type) may not be seri-
alized.
Locked instructions should not be used to insure that data written can be fetched as
instructions.
NOTE
The locked instructions for the current versions of the Pentium 4,
Intel Xeon, P6 family, Pentium, and Intel486 processors allow data
written to be fetched as instructions. However, Intel recommends
that developers who require the use of self-modifying code use a
different synchronizing mechanism, described in the following
sections.
8.1.3
Handling Self- and Cross-Modifying Code
The act of a processor writing data into a currently executing code segment with
the intent of executing that data as code is called self-modifying code. IA-32
processors exhibit model-specific behavior when executing self-modified code,
depending upon how far ahead of the current execution pointer the code has been
modified.
As processor microarchitectures become more complex and start to speculatively
execute code ahead of the retirement point (as in P6 and more recent processor
families), the rules regarding which code should execute, pre- or post-modification,
become blurred. To write self-modifying code and ensure that it is compliant with
current and future versions of the IA-32 architectures, use one of the following
coding options:
(* OPTION 1 *)
Store modified code (as data) into code segment;
Jump to new code or an intermediate location;
Execute new code;
(* OPTION 2 *)
Store modified code (as data) into code segment;
Execute a serializing instruction; (* For example, CPUID instruction *)
Execute new code;
The use of one of these options is not required for programs intended to run on the
Pentium or Intel486 processors, but are recommended to insure compatibility with
the P6 and more recent processor families.
Self-modifying code will execute at a lower level of performance than non-self-modi-
fying or normal code. The degree of the performance deterioration will depend upon
the frequency of modification and specific characteristics of the code.
8-6 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
The act of one processor writing data into the currently executing code segment of a
second processor with the intent of having the second processor execute that data as
code is called cross-modifying code. As with self-modifying code, IA-32 processors
exhibit model-specific behavior when executing cross-modifying code, depending
upon how far ahead of the executing processors current execution pointer the code
has been modified.
To write cross-modifying code and insure that it is compliant with current and future
versions of the IA-32 architecture, the following processor synchronization algorithm
must be implemented:
(* Action of Modifying Processor *)
Memory_Flag ← 0; (* Set Memory_Flag to value other than 1 *)
Store modified code (as data) into code segment;
Memory_Flag ← 1;
(* Action of Executing Processor *)
WHILE (Memory_Flag ≠ 1)
Wait for code to update;
ELIHW;
Execute serializing instruction; (* For example, CPUID instruction *)
Begin executing modified code;
(The use of this option is not required for programs intended to run on the Intel486
processor, but is recommended to insure compatibility with the Pentium 4, Intel
Xeon, P6 family, and Pentium processors.)
Like self-modifying code, cross-modifying code will execute at a lower level of perfor-
mance than non-cross-modifying (normal) code, depending upon the frequency of
modification and specific characteristics of the code.
The restrictions on self-modifying code and cross-modifying code also apply to the
Intel 64 architecture.
8.1.4
Effects of a LOCK Operation on Internal Processor Caches
For the Intel486 and Pentium processors, the LOCK# signal is always asserted on the
bus during a LOCK operation, even if the area of memory being locked is cached in
the processor.
For the P6 and more recent processor families, if the area of memory being locked
during a LOCK operation is cached in the processor that is performing the LOCK oper-
ation as write-back memory and is completely contained in a cache line, the
processor may not assert the LOCK# signal on the bus. Instead, it will modify the
memory location internally and allow it’s cache coherency mechanism to insure that
the operation is carried out atomically. This operation is called “cache locking.” The
cache coherency mechanism automatically prevents two or more processors that
Vol. 3 8-7
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
have cached the same area of memory from simultaneously modifying data in that
area.
8.2
MEMORY ORDERING
The term memory ordering refers to the order in which the processor issues reads
(loads) and writes (stores) through the system bus to system memory. The Intel 64
and IA-32 architectures support several memory-ordering models depending on the
implementation of the architecture. For example, the Intel386 processor enforces
program ordering (generally referred to as strong ordering), where reads and
writes are issued on the system bus in the order they occur in the instruction stream
under all circumstances.
To allow performance optimization of instruction execution, the IA-32 architecture
Pentium 4, Intel Xeon, and P6 family processors. These processor-ordering varia-
operations such as allowing reads to go ahead of buffered writes. The goal of any of
memory coherency, even in multiple-processor systems.
Section 8.2.1 and Section 8.2.2 describe the memory-ordering implemented by
Intel486, Pentium, Intel Core 2 Duo, Intel Atom, Intel Core Duo, Pentium 4, Intel
Xeon, and P6 family processors. Section 8.2.3 gives examples illustrating the
behavior of the memory-ordering model on IA-32 and Intel-64 processors. Section
8.2.4 considers the special treatment of stores for string operations and Section
8.2.5 discusses how memory-ordering behavior may be modified through the use of
specific instructions.
8.2.1
Memory Ordering in the Intel® Pentium® and Intel486™
Processors
The Pentium and Intel486 processors follow the processor-ordered memory model;
however, they operate as strongly-ordered processors under most circumstances.
Reads and writes always appear in programmed order at the system bus—except for
the following situation where processor ordering is exhibited. Read misses are
permitted to go ahead of buffered writes on the system bus when all the buffered
writes are cache hits and, therefore, are not directed to the same address being
accessed by the read miss.
In the case of I/O operations, both reads and writes always appear in programmed
order.
Software intended to operate correctly in processor-ordered processors (such as the
Pentium 4, Intel Xeon, and P6 family processors) should not depend on the relatively
strong ordering of the Pentium or Intel486 processors. Instead, it should insure that
accesses to shared variables that are intended to control concurrent execution
8-8 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
among processors are explicitly required to obey program ordering through the use
of appropriate locking or serializing operations (see Section 8.2.5, “Strengthening or
Weakening the Memory-Ordering Model”).
8.2.2
Memory Ordering in P6 and More Recent Processor Families
The Intel Core 2 Duo, Intel Atom, Intel Core Duo, Pentium 4, and P6 family proces-
sors also use a processor-ordered memory-ordering model that can be further
defined as “write ordered with store-buffer forwarding.” This model can be character-
ized as follows.
In a single-processor system for memory regions defined as write-back cacheable,
the memory-ordering model respects the following principles (Note the memory-
ordering principles for single-processor and multiple-processor systems are written
from the perspective of software executing on the processor, where the term
“processor“ refers to a logical processor. For example, a physical processor
supporting multiple cores and/or HyperThreading Technology is treated as a multi-
processor systems.):
• Reads are not reordered with other reads.
• Writes to memory are not reordered with other writes, with the exception of
— writes executed with the CLFLUSH instruction,
— streaming stores (writes) executed with the non-temporal move instructions
(MOVNTI, MOVNTQ, MOVNTDQ, MOVNTPS, and MOVNTPD),
— string operations (see Section 8.2.4.1).
• Reads may be reordered with older writes to different locations but not with older
writes to the same location.
• Reads or writes cannot be reordered with I/O instructions, locked instructions, or
serializing instructions.
• Reads cannot pass LFENCE and MFENCE instructions.
• Writes cannot pass SFENCE and MFENCE instructions.
In a multiple-processor system, the following ordering principles apply:
• Individual processors use the same ordering principles as in a single-processor
system.
• Writes by a single processor are observed in the same order by all processors.
• Writes from an individual processor are NOT ordered with respect to the writes
from other processors.
• Memory ordering obeys causality (memory ordering respects transitive
visibility).
• Any two stores are seen in a consistent order by processors other than those
performing the stores
Vol. 3 8-9
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
• Locked instructions have a total order.
See the example in Figure 8-1. Consider three processors in a system and each
processor performs three writes, one to each of three defined locations (A, B, and C).
Individually, the processors perform the writes in the same program order, but
because of bus arbitration and other memory access mechanisms, the order that the
three processors write the individual memory locations can differ each time the
respective code sequences are executed on the processors. The final values in loca-
tion A, B, and C would possibly vary on each execution of the write sequence.
The processor-ordering model described in this section is virtually identical to that
used by the Pentium and Intel486 processors. The only enhancements in the Pentium
4, Intel Xeon, and P6 family processors are:
• Added support for speculative reads, while still adhering to the ordering
principles above.
• Store-buffer forwarding, when a read passes a write to the same memory
location.
• Out of order store from long string store and string move operations (see Section
8.2.4, “Out-of-Order Stores For String Operations,” below).
Order of Writes From Individual Processors
Processor #1
Processor #2
Processor #3
Each processor
is guaranteed to
perform writes in
program order.
Write A.2
Write B.2
Write C.2
Write A.1
Write B.1
Write C.1
Write A.3
Write B.3
Write C.3
Example of order of actual writes
from all processors to memory
Writes are in order
Write A.1
Write B.1
with respect to
individual processes.
Write A.2
Write A.3
Write C.1
Write B.2
Write C.2
Write B.3
Write C.3
Writes from all
processors are
not guaranteed
to occur in a
particular order.
Figure 8-1. Example of Write Ordering in Multiple-Processor Systems
NOTE
In P6 processor family, store-buffer forwarding to reads of WC memory from
streaming stores to the same address does not occur due to errata.
8-10 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
8.2.3
This section provides a set of examples that illustrate the behavior of the memory-
ordering principles introduced in Section 8.2.2. They are designed to give software
writers an understanding of how memory ordering may affect the results of different
sequences of instructions.
These examples are limited to accesses to memory regions defined as write-back
cacheable (WB). (Section 8.2.3.1 describes other limitations on the generality of the
examples.) The reader should understand that they describe only software-visible
behavior. A logical processor may reorder two accesses even if one of examples indi-
cates that they may not be reordered. Such an example states only that software
cannot detect that such a reordering occurred. Similarly, a logical processor may
execute a memory access more than once as long as the behavior visible to software
is consistent with a single execution of the memory access.
8.2.3.1
Assumptions, Terminology, and Notation
As noted above, the examples in this section are limited to accesses to memory
regions defined as write-back cacheable (WB). They apply only to ordinary loads
stores and to locked read-modify-write instructions. They do not necessarily apply to
any of the following: out-of-order stores for string instructions (see Section 8.2.4);
accesses with a non-temporal hint; reads from memory by the processor as part of
address translation (e.g., page walks); and updates to segmentation and paging
structures by the processor (e.g., to update “accessed” bits).
The principles underlying the examples in this section apply to individual memory
accesses and to locked read-modify-write instructions. The Intel-64 memory-
ordering model guarantees that, for each of the following memory-access instruc-
tions, the constituent memory operation appears to execute as a single memory
access:
• Instructions that read or write a single byte.
• Instructions that read or write a word (2 bytes) whose address is aligned on a 2
byte boundary.
• Instructions that read or write a doubleword (4 bytes) whose address is aligned
on a 4 byte boundary.
• Instructions that read or write a quadword (8 bytes) whose address is aligned on
an 8 byte boundary.
Any locked instruction (either the XCHG instruction or another read-modify-write
instruction with a LOCK prefix) appears to execute as an indivisible and uninterrupt-
ible sequence of load(s) followed by store(s) regardless of alignment.
Other instructions may be implemented with multiple memory accesses. From a
memory-ordering point of view, there are no guarantees regarding the relative order
in which the constituent memory accesses are made. There is also no guarantee that
the constituent operations of a store are executed in the same order as the constit-
uent operations of a load.
Vol. 3 8-11
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
Section 8.2.3.2 through Section 8.2.3.7 give examples using the MOV instruction.
The principles that underlie these examples apply to load and store accesses in
general and to other instructions that load from or store to memory. Section 8.2.3.8
and Section 8.2.3.9 give examples using the XCHG instruction. The principles that
underlie these examples apply to other locked read-modify-write instructions.
This section uses the term “processor” is to refer to a logical processor. The examples
are written using Intel-64 assembly-language syntax and use the following nota-
tional conventions:
• Arguments beginning with an “r”, such as r1 or r2 refer to registers (e.g., EAX)
visible only to the processor being considered.
• Memory locations are denoted with x, y, z.
• Stores are written as mov [ _x], val, which implies that val is being stored into
the memory location x.
• Loads are written as mov r, [ _x], which implies that the contents of the memory
location x are being loaded into the register r.
As noted earlier, the examples refer only to software visible behavior. When the
succeeding sections make statement such as “the two stores are reordered,” the
implication is only that “the two stores appear to be reordered from the point of view
of software.”
8.2.3.2
Neither Loads Nor Stores Are Reordered with Like Operations
The Intel-64 memory-ordering model allows neither loads nor stores to be reordered
with the same kind of operation. That is, it ensures that loads are seen in program
order and that stores are seen in program order. This is illustrated by the following
example:
Example 8-1. Stores Are Not Reordered with Other Stores
Processor 0
Processor 1
mov [ _x], 1
mov [ _y], 1
mov r1, [ _y]
mov r2, [ _x]
Initially x == y == 0
r1 == 1 and r2 == 0 is not allowed
The disallowed return values could be exhibited only if processor 0’s two stores are
reordered (with the two loads occurring between them) or if processor 1’s two loads
are reordered (with the two stores occurring between them).
If r1 == 1, the store to y occurs before the load from y. Because the Intel-64
memory-ordering model does not allow stores to be reordered, the earlier store to x
occurs before the load from y. Because the Intel-64 memory-ordering model does
not allow loads to be reordered, the store to x also occurs before the later load from
x. This r2 == 1.
8-12 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
8.2.3.3
Stores Are Not Reordered With Earlier Loads
The Intel-64 memory-ordering model ensures that a store by a processor may not
occur before a previous load by the same processor. This is illustrated by the
following example:
Example 8-2. Stores Are Not Reordered with Older Loads
Processor 0
Processor 1
mov r1, [ _x]
mov [ _y], 1
mov r2, [ _y]
mov [ _x], 1
Initially x == y == 0
r1 == 1 and r2 == 1 is not allowed
Assume r1 == 1.
• Because r1 == 1, processor 1’s store to x occurs before processor 0’s load from
x.
• Because the Intel-64 memory-ordering model prevents each store from being
reordered with the earlier load by the same processor, processor 1’s load from y
occurs before its store to x.
• Similarly, processor 0’s load from x occurs before its store to y.
• Thus, processor 1’s load from y occurs before processor 0’s store to y, implying
r2 == 0.
8.2.3.4
Loads May Be Reordered with Earlier Stores to Different
Locations
The Intel-64 memory-ordering model allows a load to be reordered with an earlier
store to a different location. However, loads are not reordered with stores to the
same location.
The fact that a load may be reordered with an earlier store to a different location is
illustrated by the following example:
Example 8-3. Loads May be Reordered with Older Stores
Processor 0
Processor 1
mov [ _x], 1
mov r1, [ _y]
mov [ _y], 1
mov r2, [ _x]
Initially x == y == 0
r1 == 0 and r2 == 0 is allowed
At each processor, the load and the store are to different locations and hence may be
reordered. Any interleaving of the operations is thus allowed. One such interleaving
Vol. 3 8-13
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
has the two loads occurring before the two stores. This would result in each load
returning value 0.
The fact that a load may not be reordered with an earlier store to the same location
is illustrated by the following example:
Example 8-4. Loads Are not Reordered with Older Stores to the Same Location
Processor 0
mov [ _x], 1
mov r1, [ _x]
Initially x == 0
r1 == 0 is not allowed
The Intel-64 memory-ordering model does not allow the load to be reordered with
the earlier store because the accesses are to the same location. Therefore, r1 == 1
must hold.
8.2.3.5
Intra-Processor Forwarding Is Allowed
The memory-ordering model allows concurrent stores by two processors to be seen
in different orders by those two processors; specifically, each processor may perceive
its own store occurring before that of the other. This is illustrated by the following
example:
Example 8-5. Intra-Processor Forwarding is Allowed
Processor 0
Processor 1
mov [ _x], 1
mov r1, [ _x]
mov r2, [ _y]
mov [ _y], 1
mov r3, [ _y]
mov r4, [ _x]
Initially x == y == 0
r2 == 0 and r4 == 0 is allowed
The memory-ordering model imposes no constraints on the order in which the two
stores appear to execute by the two processors. This fact allows processor 0 to see
its store before seeing processor 1's, while processor 1 sees its store before seeing
processor 0's. (Each processor is self consistent.) This allows r2 == 0 and r4 == 0.
In practice, the reordering in this example can arise as a result of store-buffer
forwarding. While a store is temporarily held in a processor's store buffer, it can
satisfy the processor's own loads but is not visible to (and cannot satisfy) loads by
other processors.
8-14 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
8.2.3.6
Stores Are Transitively Visible
The memory-ordering model ensures transitive visibility of stores; stores that are
causally related appear to all processors to occur in an order consistent with the
causality relation. This is illustrated by the following example:
Example 8-6. Stores Are Transitively Visible
Processor 0
Processor 1
mov r1, [ _x]
mov [ _y], 1
Processor 2
mov [ _x], 1
mov r2, [ _y]
mov r3, [_x]
Initially x == y == 0
r1 == 1, r2 == 1, r3 == 0 is not allowed
Assume that r1 == 1 and r2 == 1.
• Because r1 == 1, processor 0’s store occurs before processor 1’s load.
• Because the memory-ordering model prevents a store from being reordered with
an earlier load (see Section 8.2.3.3), processor 1’s load occurs before its store.
Thus, processor 0’s store causally precedes processor 1’s store.
• Because processor 0’s store causally precedes processor 1’s store, the memory-
ordering model ensures that processor 0’s store appears to occur before
processor 1’s store from the point of view of all processors.
• Because r2 == 1, processor 1’s store occurs before processor 2’s load.
• Because the Intel-64 memory-ordering model prevents loads from being
• The above items imply that processor 0’s store to x occurs before processor 2’s
load from x. This implies that r3 == 1.
8.2.3.7
Stores Are Seen in a Consistent Order by Other Processors
As noted in Section 8.2.3.5, the memory-ordering model allows stores by two
processors to be seen in different orders by those two processors. However, any two
stores must appear to execute in the same order to all processors other than those
performing the stores. This is illustrated by the following example:
Example 8-7. Stores Are Seen in a Consistent Order by Other Processors
Processor 0
Processor 1
Processor 2
mov r1, [ _x]
mov r2, [ _y]
Processor 3
mov r3, [_y]
mov r4, [_x]
mov [ _x], 1
mov [ _y], 1
Initially x == y ==0
r1 == 1, r2 == 0, r3 == 1, r4 == 0is not allowed
Vol. 3 8-15
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
By the principles discussed in Section 8.2.3.2,
• processor 2’s first and second load cannot be reordered,
• processor 3’s first and second load cannot be reordered.
• If r1 == 1 and r2 == 0, processor 0’s store appears to precede processor 1’s
store with respect to processor 2.
• Similarly, r3 == 1 and r4 == 0 imply that processor 1’s store appears to precede
processor 0’s store with respect to processor 1.
Because the memory-ordering model ensures that any two stores appear to execute
in the same order to all processors (other than those performing the stores), this set
of return values is not allowed
8.2.3.8
Locked Instructions Have a Total Order
The memory-ordering model ensures that all processors agree on a single execution
order of all locked instructions, including those that are larger than 8 bytes or are not
naturally aligned. This is illustrated by the following example:
Example 8-8. Locked Instructions Have a Total Order
Processor 0
Processor 1
Processor 2
Processor 3
xchg [ _x], r1
xchg [ _y], r2
mov r3, [ _x]
mov r4, [ _y]
mov r5, [_y]
mov r6, [_x]
Initially r1 == r2 == 1, x == y == 0
r3 == 1, r4 == 0, r5 == 1, r6 == 0 is not allowed
Processor 2 and processor 3 must agree on the order of the two executions of XCHG.
Without loss of generality, suppose that processor 0’s XCHG occurs first.
• If r5 == 1, processor 1’s XCHG into y occurs before processor 3’s load from y.
• Because the Intel-64 memory-ordering model prevents loads from being
reordered (see Section 8.2.3.2), processor 3’s loads occur in order and,
therefore, processor 1’s XCHG occurs before processor 3’s load from x.
• Since processor 0’s XCHG into x occurs before processor 1’s XCHG (by
assumption), it occurs before processor 3’s load from x. Thus, r6 == 1.
A similar argument (referring instead to processor 2’s loads) applies if processor 1’s
XCHG occurs before processor 0’s XCHG.
8.2.3.9
Loads and Stores Are Not Reordered with Locked Instructions
The memory-ordering model prevents loads and stores from being reordered with
locked instructions that execute earlier or later. The examples in this section illustrate
only cases in which a locked instruction is executed before a load or a store. The
8-16 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
reader should note that reordering is prevented also if the locked instruction is
executed after a load or a store.
The first example illustrates that loads may not be reordered with earlier locked
instructions:
Example 8-9. Loads Are not Reordered with Locks
Processor 0
Processor 1
xchg [ _x], r1
mov r2, [ _y]
mov r4, [ _x]
Initially x == y == 0, r1 == r3 == 1
r2 == 0 and r4 == 0 is not allowed
As explained in Section 8.2.3.8, there is a total order of the executions of locked
instructions. Without loss of generality, suppose that processor 0’s XCHG occurs first.
Because the Intel-64 memory-ordering model prevents processor 1’s load from
being reordered with its earlier XCHG, processor 0’s XCHG occurs before
processor 1’s load. This implies r4 == 1.
A similar argument (referring instead to processor 2’s accesses) applies if
processor 1’s XCHG occurs before processor 0’s XCHG.
The second example illustrates that a store may not be reordered with an earlier
locked instruction:
Example 8-10. Stores Are not Reordered with Locks
Processor 0
Processor 1
xchg [ _x], r1
mov [ _y], 1
mov r2, [ _y]
mov r3, [ _x]
Initially x == y == 0, r1 == 1
r2 == 1 and r3 == 0 is not allowed
Assume r2 == 1.
y.
• Because the memory-ordering model prevents a store from being reordered with
an earlier locked instruction, processor 0’s XCHG into x occurs before its store to
y. Thus, processor 0’s XCHG into x occurs before processor 1’s load from y.
• Because the memory-ordering model prevents loads from being reordered (see
Section 8.2.3.2), processor 1’s loads occur in order and, therefore, processor 1’s
XCHG into x occurs before processor 1’s load from x. Thus, r3 == 1.
Vol. 3 8-17
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
8.2.4
Out-of-Order Stores For String Operations
The Intel Core 2 Duo, Intel Core, Pentium 4, and P6 family processors modify the
processors operation during the string store operations (initiated with the MOVS and
STOS instructions) to maximize performance. Once the “fast string” operations initial
conditions are met (as described below), the processor will essentially operate on,
from an external perspective, the string in a cache line by cache line mode. This
results in the processor looping on issuing a cache-line read for the source address
and an invalidation on the external bus for the destination address, knowing that all
bytes in the destination cache line will be modified, for the length of the string. In this
mode interrupts will only be accepted by the processor on cache line boundaries. It is
possible in this mode that the destination line invalidations, and therefore stores, will
be issued on the external bus out of order.
Code dependent upon sequential store ordering should not use the string operations
for the entire data structure to be stored. Data and semaphores should be separated.
Order dependent code should use a discrete semaphore uniquely stored to after any
string operations to allow correctly ordered data to be seen by all processors.
“Fast string” operation can be disabled by clearing the fast-string-enable bit (bit 0) of
IA32_MISC_ENABLES MSR.
Initial conditions for “fast string” operations are implementation specific. Example
conditions include:
• EDI and ESI must be 8-byte aligned for the Pentium III processor. EDI must be 8-
byte aligned for the Pentium 4 processor.
• String operation must be performed in ascending address order.
• The initial operation counter (ECX) must be equal to or greater than 64.
• Source and destination must not overlap by less than a cache line (64 bytes, for
Intel Core 2 Duo, Intel Core, Pentium M, and Pentium 4 processors; 32 bytes P6
family and Pentium processors).
• The memory type for both source and destination addresses must be either WB
or WC.
NOTE
Initial conditions for “fast string“ operation in future Intel 64 or IA-32 processor fami-
lies may differ from above.
8.2.4.1
Memory-Ordering Model for String Operations on Write-back (WB)
Memory
This section deals with the memory-ordering model for string operations on write-
back (WB) memory for the Intel 64 architecture.
The memory-ordering model respects the follow principles:
1. Stores within a single string operation may be executed out of order.
8-18 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
2. Stores from separate string operations (for example, stores from consecutive
string operations) do not execute out of order. All the stores from an earlier string
operation will complete before any store from a later string operation.
3. String operations are not reordered with other store operations.
Fast string operations (e.g. string operations initiated with the MOVS/STOS instruc-
tions and the REP prefix) may be interrupted by exceptions or interrupts. The inter-
rupts are precise but may be delayed - for example, the interruptions may be taken
at cache line boundaries, after every few iterations of the loop, or after operating on
every few bytes. Different implementations may choose different options, or may
even choose not to delay interrupt handling, so software should not rely on the delay.
When the interrupt/trap handler is reached, the source/destination registers point to
the next string element to be operated on, while the EIP stored in the stack points to
the string instruction, and the ECX register has the value it held following the last
successful iteration. The return from that trap/interrupt handler should cause the
string instruction to be resumed from the point where it was interrupted.
The string operation memory-ordering principles, (item 2 and 3 above) should be
interpreted by taking the incorruptibility of fast string operations into account. For
example, if a fast string operation gets interrupted after k iterations, then stores
performed by the interrupt handler will become visible after the fast string stores
from iteration 0 to k, and before the fast string stores from the (k+1)th iteration
onward.
Stores within a single string operation may execute out of order (item 1 above) only
if fast string operation is enabled. Fast string operations are enabled/disabled
8.2.4.2
Examples Illustrating Memory-Ordering Principles for String
Operations
The following examples uses the same notation and convention as described in
Section 8.2.3.1.
In Example 8-11, processor 0 does one round of (128 iterations) doubleword string
store operation via rep:stosd, writing the value 1 (value in EAX) into a block of 512
bytes from location _x (kept in ES:EDI) in ascending order. Since each operation
stores a doubleword (4 bytes), the operation is repeated 128 times (value in ECX).
The block of memory initially contained 0. Processor 1 is reading two memory loca-
tions that are part of the memory block being updated by processor 0, i.e, reading
locations in the range _x to (_x+511).
Example 8-11. Stores Within a String Operation May be Reordered
Processor 0
Processor 1
rep:stosd [ _x]
mov r1, [ _z]
mov r2, [ _y]
Vol. 3 8-19
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
Example 8-11. Stores Within a String Operation May be Reordered
Processor 0
Processor 1
Initially on processor 0: EAX == 1, ECX==128, ES:EDI ==_x
Initially [_x] to 511[_x]== 0, _x <= _y < _z < _x+512
r1 == 1 and r2 == 0 is allowed
It is possible for processor 1 to perceive that the repeated string stores in processor
0 are happening out of order. We assume that fast string operations are enabled on
processor 0.
In Example 8-12, processor 0 does two separate rounds of rep stosd operation of 128
doubleword stores, writing the value 1 (value in EAX) into the first block of 512 bytes
from location _x (kept in ES:EDI) in ascending order. It then writes 1 into a second
block of memory from (_x+512) to (_x+1023). All of the memory locations initially
contain 0. The block of memory initially contained 0. Processor 1 performs two load
operations from the two blocks of memory.
Example 8-12. Stores Across String Operations Are not Reordered
Processor 0
Processor 1
rep:stosd [ _x]
mov r1, [ _z]
mov r2, [ _y]
mov ecx, $128
rep:stosd 512[ _x]
Initially on processor 0: EAX == 1, ECX==128, ES:EDI ==_x
Initially [_x] to 1023[_x]== 0, _x <= _y < _x+512 < _z < _x+1024
r1 == 1 and r2 == 0 is not allowed
It is not possible in the above example for processor 1 to perceive any of the stores
from the later string operation (to the second 512 block) in processor 0 before seeing
the stores from the earlier string operation to the first 512 block.
does not get executed while processor 0’s string operation to the first block has been
interrupted. If the string operation to the first block by processor 0 is interrupted,
and a write to the second memory block is executed by the interrupt handler, then
that change in the second memory block will be visible before the string operation to
the first memory block resumes.
In Example 8-13, processor 0 does one round of (128 iterations) doubleword string
store operation via rep:stosd, writing the value 1 (value in EAX) into a block of 512
bytes from location _x (kept in ES:EDI) in ascending order. It then writes to a second
memory location outside the memory block of the previous string operation.
8-20 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
Processor 1 performs two read operations, the first read is from an address outside
the 512-byte block but to be updated by processor 0, the second ready is from inside
the block of memory of string operation.
Example 8-13. String Operations Are not Reordered with later Stores
Processor 0
Processor 1
rep:stosd [ _x]
mov [_z], $1
mov r1, [ _z]
mov r2, [ _y]
Initially on processor 0: EAX == 1, ECX==128, ES:EDI ==_x
Initially [_y] == [_z] == 0, [_x] to 511[_x]== 0, _x <= _y < _x+512, _z is a separate memory
location
r1 == 1 and r2 == 0 is not allowed
Processor 1 cannot perceive the later store by processor 0 until it sees all the stores
from the string operation. Example 8-13 assumes that processor 0’s store to [_z] is
not executed while the string operation has been interrupted. If the string operation
is interrupted and the store to [_z] by processor 0 is executed by the interrupt
handler, then changes to [_z] will become visible before the string operation
resumes.
Example 8-14 illustrates the visibility principle when a string operation is interrupted.
Example 8-14. Interrupted String Operation
Processor 0
Processor 1
rep:stosd [ _x] // interrupted before es:edi reach mov r1, [ _z]
_y
mov [_z], $1 // interrupt handler
mov r2, [ _y]
Initially [_y] == [_z] == 0, [_x] to 511[_x]== 0, _x <= _y < _x+512, _z is a separate memory
location
r1 == 1 and r2 == 0 is allowed
In Example 8-14, processor 0 started a string operation to write to a memory block
store operations. The address _y has not yet been updated by processor 0 when
processor 0 got interrupted. The interrupt handler that took control on processor 0
writes to the address _z. Processor 1 may see the store to _z from the interrupt
handler, before seeing the remaining stores to the 512-byte memory block that are
executed when the string operation resumes.
Example 8-15 illustrates the ordering of string operations with earlier stores. No
store from a string operation can be visible before all prior stores are visible.
Vol. 3 8-21
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
Example 8-15. String Operations Are not Reordered with Earlier Stores
Processor 0
Processor 1
mov [_z], $1
mov r1, [ _y]
mov r2, [ _z]
rep:stosd [ _x]
Initially on processor 0: EAX == 1, ECX==128, ES:EDI ==_x
Initially [_y] == [_z] == 0, [_x] to 511[_x]== 0, _x <= _y < _x+512, _z is a separate memory
location
r1 == 1 and r2 == 0 is not allowed
8.2.5
Strengthening or Weakening the Memory-Ordering Model
The Intel 64 and IA-32 architectures provide several mechanisms for strengthening
or weakening the memory-ordering model to handle special programming situations.
These mechanisms include:
• The I/O instructions, locking instructions, the LOCK prefix, and serializing
instructions force stronger ordering on the processor.
processor) and the LFENCE and MFENCE instructions (introduced in the Pentium
4 processor) provide memory-ordering and serialization capabilities for specific
types of memory operations.
• The memory type range registers (MTRRs) can be used to strengthen or weaken
memory ordering for specific area of physical memory (see Section 11.11,
“Memory Type Range Registers (MTRRs)”). MTRRs are available only in the
Pentium 4, Intel Xeon, and P6 family processors.
• The page attribute table (PAT) can be used to strengthen memory ordering for a
specific page or group of pages (see Section 11.12, “Page Attribute Table (PAT)”).
The PAT is available only in the Pentium 4, Intel Xeon, and Pentium III processors.
These mechanisms can be used as follows:
Memory mapped devices and other I/O devices on the bus are often sensitive to the
order of writes to their I/O buffers. I/O instructions can be used to (the IN and OUT
instructions) impose strong write ordering on such accesses as follows. Prior to
executing an I/O instruction, the processor waits for all previous instructions in the
program to complete and for all buffered writes to drain to memory. Only instruction
fetch and page tables walks can pass I/O instructions. Execution of subsequent
instructions do not begin until the processor determines that the I/O instruction has
been completed.
Synchronization mechanisms in multiple-processor systems may depend upon a
strong memory-ordering model. Here, a program can use a locking instruction such
8-22 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
as the XCHG instruction or the LOCK prefix to insure that a read-modify-write opera-
tion on memory is carried out atomically. Locking operations typically operate like
I/O operations in that they wait for all previous instructions to complete and for all
buffered writes to drain to memory (see Section 8.1.2, “Bus Locking”).
Program synchronization can also be carried out with serializing instructions (see
Section 8.3). These instructions are typically used at critical procedure or task
boundaries to force completion of all previous instructions before a jump to a new
section of code or a context switch occurs. Like the I/O and locking instructions, the
processor waits until all previous instructions have been completed and all buffered
writes have been drained to memory before executing the serializing instruction.
The SFENCE, LFENCE, and MFENCE instructions provide a performance-efficient way
of insuring load and store memory ordering between routines that produce weakly-
ordered results and routines that consume that data. The functions of these instruc-
tions are as follows:
• SFENCE — Serializes all store (write) operations that occurred prior to the
SFENCE instruction in the program instruction stream, but does not affect load
operations.
• LFENCE — Serializes all load (read) operations that occurred prior to the LFENCE
instruction in the program instruction stream, but does not affect store
operations.
• MFENCE — Serializes all store and load operations that occurred prior to the
MFENCE instruction in the program instruction stream.
Note that the SFENCE, LFENCE, and MFENCE instructions provide a more efficient
method of controlling memory ordering than the CPUID instruction.
The MTRRs were introduced in the P6 family processors to define the cache charac-
teristics for specified areas of physical memory. The following are two examples of
how memory types set up with MTRRs can be used strengthen or weaken memory
ordering for the Pentium 4, Intel Xeon, and P6 family processors:
• The strong uncached (UC) memory type forces a strong-ordering model on
memory accesses. Here, all reads and writes to the UC memory region appear on
the bus and out-of-order or speculative accesses are not performed. This
memory type can be applied to an address range dedicated to memory mapped
I/O devices to force strong memory ordering.
• For areas of memory where weak ordering is acceptable, the write back (WB)
memory type can be chosen. Here, reads can be performed speculatively and
writes can be buffered and combined. For this type of memory, cache locking is
performed on atomic (locked) operations that do not split across cache lines,
which helps to reduce the performance penalty associated with the use of the
typical synchronization instructions, such as XCHG, that lock the bus during the
entire read-modify-write operation. With the WB memory type, the XCHG
instruction locks the cache instead of the bus if the memory access is contained
within a cache line.
Vol. 3 8-23
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
The PAT was introduced in the Pentium III processor to enhance the caching charac-
teristics that can be assigned to pages or groups of pages. The PAT mechanism typi-
cally used to strengthen caching characteristics at the page level with respect to the
caching characteristics established by the MTRRs. Table 11-7 shows the interaction of
the PAT with the MTRRs.
We recommended that software written to run on Intel Core 2 Duo, Intel Atom, Intel
Core Duo, Pentium 4, Intel Xeon, and P6 family processors assume the processor-
ordering model or a weaker memory-ordering model. The Intel Core 2 Duo, Intel
Atom, Intel Core Duo, Pentium 4, Intel Xeon, and P6 family processors do not imple-
ment a strong memory-ordering model, except when using the UC memory type.
Despite the fact that Pentium 4, Intel Xeon, and P6 family processors support
processor ordering, Intel does not guarantee that future processors will support this
model. To make software portable to future processors, it is recommended that oper-
ating systems provide critical region and resource control constructs and API’s (appli-
cation program interfaces) based on I/O, locking, and/or serializing instructions be
used to synchronize access to shared areas of memory in multiple-processor
systems. Also, software should not depend on processor ordering in situations where
the system hardware does not support this memory-ordering model.
8.3
SERIALIZING INSTRUCTIONS
The Intel 64 and IA-32 architectures define several serializing instructions. These
instructions force the processor to complete all modifications to flags, registers, and
memory by previous instructions and to drain all buffered writes to memory before
the next instruction is fetched and executed. For example, when a MOV to control
register instruction is used to load a new value into control register CR0 to enable
protected mode, the processor must perform a serializing operation before it enters
protected mode. This serializing operation insures that all operations that were
started while the processor was in real-address mode are completed before the
switch to protected mode is made.
The concept of serializing instructions was introduced into the IA-32 architecture
with the Pentium processor to support parallel instruction execution. Serializing
instructions have no meaning for the Intel486 and earlier processors that do not
implement parallel instruction execution.
It is important to note that executing of serializing instructions on P6 and more
recent processor families constrain speculative execution because the results of
speculatively executed instructions are discarded. The following instructions are seri-
alizing instructions:
• Privileged serializing instructions — INVD, INVEPT, INVLPG, INVVPID, LGDT,
1
LIDT, LLDT, LTR, MOV (to control register, with the exception of MOV CR8 ), MOV
(to debug register), WBINVD, and WRMSR.
1. MOV CR8 is not defined architecturally as a serializing instruction.
8-24 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
• Non-privileged serializing instructions — CPUID, IRET, and RSM.
When the processor serializes instruction execution, it ensures that all pending
memory transactions are completed (including writes stored in its store buffer)
before it executes the next instruction. Nothing can pass a serializing instruction and
a serializing instruction cannot pass any other instruction (read, write, instruction
fetch, or I/O). For example, CPUID can be executed at any privilege level to serialize
instruction execution with no effect on program flow, except that the EAX, EBX, ECX,
and EDX registers are modified.
The following instructions are memory-ordering instructions, not serializing instruc-
execution stream:
• Non-privileged memory-ordering instructions — SFENCE, LFENCE, and
MFENCE.
The SFENCE, LFENCE, and MFENCE instructions provide more granularity in control-
ling the serialization of memory loads and stores (see Section 8.2.5, “Strengthening
or Weakening the Memory-Ordering Model”).
The following additional information is worth noting regarding serializing instruc-
tions:
• The processor does not writeback the contents of modified data in its data cache
to external memory when it serializes instruction execution. Software can force
modified data to be written back by executing the WBINVD instruction, which is a
serializing instruction. The amount of time or cycles for WBINVD to complete will
vary due to the size of different cache hierarchies and other factors. As a conse-
quence, the use of the WBINVD instruction can have an impact on
interrupt/event response time.
• When an instruction is executed that enables or disables paging (that is, changes
the PG flag in control register CR0), the instruction should be followed by a jump
instruction. The target instruction of the jump instruction is fetched with the new
setting of the PG flag (that is, paging is enabled or disabled), but the jump
instruction itself is fetched with the previous setting. The Pentium 4, Intel Xeon,
and P6 family processors do not require the jump operation following the move to
register CR0 (because any use of the MOV instruction in a Pentium 4, Intel Xeon,
or P6 family processor to write to CR0 is completely serializing). However, to
maintain backwards and forward compatibility with code written to run on other
• Whenever an instruction is executed to change the contents of CR3 while paging
is enabled, the next instruction is fetched using the translation tables that
correspond to the new value of CR3. Therefore the next instruction and the
sequentially following instructions should have a mapping based upon the new
value of CR3. (Global entries in the TLBs are not invalidated, see Section 4.10.3,
“Invalidation of TLBs and Paging-Structure Caches.”)
• The Pentium processor and more recent processor families use branch-prediction
techniques to improve performance by prefetching the destination of a branch
instruction before the branch instruction is executed. Consequently, instruction
Vol. 3 8-25
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
execution is not deterministically serialized when a branch instruction is
executed.
8.4
MULTIPLE-PROCESSOR (MP) INITIALIZATION
The IA-32 architecture (beginning with the P6 family processors) defines a multiple-
processor (MP) initialization protocol called the Multiprocessor Specification Version
1.4. This specification defines the boot protocol to be used by IA-32 processors in
multiple-processor systems. (Here, multiple processors is defined as two or more
processors.) The MP initialization protocol has the following important features:
• It supports controlled booting of multiple processors without requiring dedicated
system hardware.
• It allows hardware to initiate the booting of a system without the need for a
dedicated signal or a predefined boot processor.
• It allows all IA-32 processors to be booted in the same manner, including those
supporting Intel Hyper-Threading Technology.
• The MP initialization protocol also applies to MP systems using Intel 64
processors.
The mechanism for carrying out the MP initialization protocol differs depending on
the IA-32 processor family, as follows:
8.4.1, “BSP and AP Processors”) is handled through arbitration on the APIC bus,
using BIPI and FIPI messages. See Appendix C, “MP Initialization For P6 Family
processors.
• Intel Xeon processors with family, model, and stepping IDs up to F09H —
The selection of the BSP and APs (see Section 8.4.1, “BSP and AP Processors”) is
handled through arbitration on the system bus, using BIPI and FIPI messages
Intel Xeon Processors”).
• Intel Xeon processors with family, model, and stepping IDs of F0AH and
beyond, 6E0H and beyond, 6F0H and beyond — The selection of the BSP and
APs is handled through a special system bus cycle, without using BIPI and FIPI
message arbitration (see Section 8.4.3, “MP Initialization Protocol Algorithm for
Intel Xeon Processors”).
The family, model, and stepping ID for a processor is given in the EAX register when
the CPUID instruction is executed with a value of 1 in the EAX register.
8-26 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
8.4.1
The MP initialization protocol defines two classes of processors: the bootstrap
processor (BSP) and the application processors (APs). Following a power-up or
RESET of an MP system, system hardware dynamically selects one of the processors
on the system bus as the BSP. The remaining processors are designated as APs.
As part of the BSP selection mechanism, the BSP flag is set in the IA32_APIC_BASE
MSR (see Figure 10-5) of the BSP, indicating that it is the BSP. This flag is cleared for
all other processors.
The BSP executes the BIOS’s boot-strap code to configure the APIC environment,
sets up system-wide data structures, and starts and initializes the APs. When the BSP
and APs are initialized, the BSP then begins executing the operating-system initial-
ization code.
Following a power-up or reset, the APs complete a minimal self-configuration, then
wait for a startup signal (a SIPI message) from the BSP processor. Upon receiving a
SIPI message, an AP executes the BIOS AP configuration code, which ends with the
AP being placed in halt state.
For Intel 64 and IA-32 processors supporting Intel Hyper-Threading Technology, the
MP initialization protocol treats each of the logical processors on the system bus or
coherent link domain as a separate processor (with a unique APIC ID). During boot-
up, one of the logical processors is selected as the BSP and the remainder of the
logical processors are designated as APs.
8.4.2
MP Initialization Protocol Requirements and Restrictions
The MP initialization protocol imposes the following requirements and restrictions on
the system:
• The MP protocol is executed only after a power-up or RESET. If the MP protocol
has completed and a BSP is chosen, subsequent INITs (either to a specific
processor or system wide) do not cause the MP protocol to be repeated. Instead,
each logical processor examines its BSP flag (in the IA32_APIC_BASE MSR) to
determine whether it should execute the BIOS boot-strap code (if it is the BSP) or
enter a wait-for-SIPI state (if it is an AP).
• All devices in the system that are capable of delivering interrupts to the
processors must be inhibited from doing so for the duration of the MP initial-
ization protocol. The time during which interrupts must be inhibited includes the
window between when the BSP issues an INIT-SIPI-SIPI sequence to an AP and
when the AP responds to the last SIPI in the sequence.
Vol. 3 8-27
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
8.4.3
MP Initialization Protocol Algorithm for
Intel Xeon Processors
Following a power-up or RESET of an MP system, the processors in the system
sors on the system bus or coherent link domain. In the course of executing this algo-
rithm, the following boot-up and initialization operations are carried out:
1. Each logical processor is assigned a unique APIC ID, based on system topology.
The unique ID is a 32-bit value if the processor supports CPUID leaf 0BH,
otherwise the unique ID is an 8-bit value. (see Section 8.4.5, “Identifying Logical
Processors in an MP System”). This ID is written into the local APIC ID register for
each processor.
2. Each logical processor is assigned a unique arbitration priority based on its
APIC ID.
3. Each logical processor executes its internal BIST simultaneously with the other
logical processors on the system bus.
4. Upon completion of the BIST, the logical processors use a hardware-defined
selection mechanism to select the BSP and the APs from the available logical
processors on the system bus. The BSP selection mechanism differs depending
on the family, model, and stepping IDs of the processors, as follows:
— Family, model, and stepping IDs of F0AH and onwards:
•
The logical processors begin monitoring the BNR# signal, which is
toggling. When the BNR# pin stops toggling, each processor attempts to
issue a NOP special cycle on the system bus.
•
The logical processor with the highest arbitration priority succeeds in
issuing a NOP special cycle and is nominated the BSP. This processor sets
the BSP flag in its IA32_APIC_BASE MSR, then fetches and begins
executing BIOS boot-strap code, beginning at the reset vector (physical
address FFFF FFF0H).
•
The remaining logical processors (that failed in issuing a NOP special
cycle) are designated as APs. They leave their BSP flags in the clear state
and enter a “wait-for-SIPI state.”
— Family, model, and stepping IDs up to F09H:
•
Each processor broadcasts a BIPI to “all including self.” The first processor
that broadcasts a BIPI (and thus receives its own BIPI vector), selects
itself as the BSP and sets the BSP flag in its IA32_APIC_BASE MSR. (See
Appendix C.1, “Overview of the MP Initialization Process For P6 Family
Processors,” for a description of the BIPI, FIPI, and SIPI messages.)
•
The remainder of the processors (which were not selected as the BSP) are
designated as APs. They leave their BSP flags in the clear state and enter
a “wait-for-SIPI state.”
8-28 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
•
The newly established BSP broadcasts an FIPI message to “all including
self,” which the BSP and APs treat as an end of MP initialization signal.
Only the processor with its BSP flag set responds to the FIPI message. It
responds by fetching and executing the BIOS boot-strap code, beginning
at the reset vector (physical address FFFF FFF0H).
5. As part of the boot-strap code, the BSP creates an ACPI table and an MP table and
adds its initial APIC ID to these tables as appropriate.
then broadcasts a SIPI message to all the APs in the system. Here, the SIPI
message contains a vector to the BIOS AP initialization code (at 000VV000H,
where VV is the vector contained in the SIPI message).
7. The first action of the AP initialization code is to set up a race (among the APs) to
a BIOS initialization semaphore. The first AP to the semaphore begins executing
the initialization code. (See Section 8.4.4, “MP Initialization Example,” for
semaphore implementation details.) As part of the AP initialization procedure,
the AP adds its APIC ID number to the ACPI and MP tables as appropriate and
increments the processor counter by 1. At the completion of the initialization
procedure, the AP executes a CLI instruction and halts itself.
8. When each of the APs has gained access to the semaphore and executed the AP
initialization code, the BSP establishes a count for the number of processors
connected to the system bus, completes executing the BIOS boot-strap code,
and then begins executing operating-system boot-strap and start-up code.
9. While the BSP is executing operating-system boot-strap and start-up code, the
APs remain in the halted state. In this state they will respond only to INITs, NMIs,
and SMIs. They will also respond to snoops and to assertions of the STPCLK# pin.
The following section gives an example (with code) of the MP initialization protocol
for multiple Intel Xeon processors operating in an MP configuration.
Appendix B, “Model-Specific Registers (MSRs),” describes how to program the
LINT[0:1] pins of the processor’s local APICs after an MP configuration has been
completed.
8.4.4
MP Initialization Example
The following example illustrates the use of the MP initialization protocol used to
The code runs on Intel 64 or IA-32 processors that use a protocol. This includes P6
Family processors, Pentium 4 processors, Intel Core Duo, Intel Core 2 Duo and Intel
Xeon processors.
The following constants and data definitions are used in the accompanying
code examples. They are based on the addresses of the APIC registers defined in
Table 10-1.
ICR_LOW
EQU 0FEE00300H
Vol. 3 8-29
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
SVR
EQU 0FEE000F0H
APIC_ID
LVT3
APIC_ENABLED
BOOT_ID
COUNT
EQU 0FEE00020H
EQU 0FEE00370H
EQU 0100H
DD ?
VACANT
EQU 00H
8.4.4.1
Typical BSP Initialization Sequence
After the BSP and APs have been selected (by means of a hardware protocol, see
Section 8.4.3, “MP Initialization Protocol Algorithm for Intel Xeon Processors”), the
BSP begins executing BIOS boot-strap code (POST) at the normal IA-32 architecture
starting address (FFFF FFF0H). The boot-strap code typically performs the following
operations:
1. Initializes memory.
2. Loads the microcode update into the processor.
3. Initializes the MTRRs.
4. Enables the caches.
5. Executes the CPUID instruction with a value of 0H in the EAX register, then reads
the EBX, ECX, and EDX registers to determine if the BSP is “GenuineIntel.”
6. Executes the CPUID instruction with a value of 1H in the EAX register, then saves
the values in the EAX, ECX, and EDX registers in a system configuration space in
RAM for use later.
7. Loads start-up code for the AP to execute into a 4-KByte page in the lower 1
MByte of memory.
8. Switches to protected mode and insures that the APIC address space is mapped
to the strong uncacheable (UC) memory type.
9. Determine the BSP’s APIC ID from the local APIC ID register (default is 0), the
code snippet below is an example that applies to logical processors in a system
whose local APIC units operate in xAPIC mode that APIC registers are accessed
using memory mapped interface:
MOV ESI, APIC_ID; Address of local APIC ID register
MOV EAX, [ESI];
AND EAX, 0FF000000H; Zero out all other bits except APIC ID
MOV BOOT_ID, EAX; Save in memory
Saves the APIC ID in the ACPI and MP tables and optionally in the system config-
uration space in RAM.
10. Converts the base address of the 4-KByte page for the AP’s bootup code into 8-bit
vector. The 8-bit vector defines the address of a 4-KByte page in the real-address
8-30 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
mode address space (1-MByte space). For example, a vector of 0BDH specifies a
start-up memory address of 000BD000H.
11. Enables the local APIC by setting bit 8 of the APIC spurious vector register (SVR).
MOV ESI, SVR; Address of SVR
MOV EAX, [ESI];
OR EAX, APIC_ENABLED; Set bit 8 to enable (0 on reset)
MOV [ESI], EAX;
12. Sets up the LVT error handling entry by establishing an 8-bit vector for the APIC
error handler.
MOV ESI, LVT3;
MOV EAX, [ESI];
AND EAX, FFFFFF00H; Clear out previous vector.
OR EAX, 000000xxH; xx is the 8-bit vector the APIC error handler.
MOV [ESI], EAX;
13. Initializes the Lock Semaphore variable VACANT to 00H. The APs use this
semaphore to determine the order in which they execute BIOS AP initialization
code.
14. Performs the following operation to set up the BSP to detect the presence of APs
in the system and the number of processors:
— Sets the value of the COUNT variable to 1.
— Starts a timer (set for an approximate interval of 100 milliseconds). In the AP
BIOS initialization code, the AP will increment the COUNT variable to indicate
its presence. When the timer expires, the BSP checks the value of the COUNT
variable. If the timer expires and the COUNT variable has not been incre-
mented, no APs are present or some error has occurred.
15. Broadcasts an INIT-SIPI-SIPI IPI sequence to the APs to wake them up and
initialize them:
MOV ESI, ICR_LOW; Load address of ICR low dword into ESI.
MOV EAX, 000C4500H; Load ICR encoding for broadcast INIT IPI
; to all APs into EAX.
MOV [ESI], EAX; Broadcast INIT IPI to all APs
; 10-millisecond delay loop.
MOV EAX, 000C46XXH; Load ICR encoding for broadcast SIPI IP
; to all APs into EAX, where xx is the vector computed in step 10.
MOV [ESI], EAX; Broadcast SIPI IPI to all APs
; 200-microsecond delay loop
MOV [ESI], EAX; Broadcast second SIPI IPI to all APs
; 200-microsecond delay loop
Step 15:
Vol. 3 8-31
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
MOV EAX, 000C46XXH; Load ICR encoding from broadcast SIPI IP
; to all APs into EAX where xx is the vector computed in step 8.
16. Waits for the timer interrupt.
17. Reads and evaluates the COUNT variable and establishes a processor count.
18. If necessary, reconfigures the APIC and continues with the remaining system
diagnostics as appropriate.
8.4.4.2
Typical AP Initialization Sequence
When an AP receives the SIPI, it begins executing BIOS AP initialization code at the
vector encoded in the SIPI. The AP initialization code typically performs the following
operations:
1. Waits on the BIOS initialization Lock Semaphore. When control of the semaphore
is attained, initialization continues.
2. Loads the microcode update into the processor.
3. Initializes the MTRRs (using the same mapping that was used for the BSP).
4. Enables the cache.
5. Executes the CPUID instruction with a value of 0H in the EAX register, then reads
the EBX, ECX, and EDX registers to determine if the AP is “GenuineIntel.”
6. Executes the CPUID instruction with a value of 1H in the EAX register, then saves
the values in the EAX, ECX, and EDX registers in a system configuration space in
RAM for use later.
to the strong uncacheable (UC) memory type.
8. Determines the AP’s APIC ID from the local APIC ID register, and adds it to the MP
and ACPI tables and optionally to the system configuration space in RAM.
9. Initializes and configures the local APIC by setting bit 8 in the SVR register and
setting up the LVT3 (error LVT) for error handling (as described in steps 9 and 10
in Section 8.4.4.1, “Typical BSP Initialization Sequence”).
10. Configures the APs SMI execution environment. (Each AP and the BSP must have
a different SMBASE address.)
11. Increments the COUNT variable by 1.
12. Releases the semaphore.
13. Executes the CLI and HLT instructions.
14. Waits for an INIT IPI.
8-32 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
8.4.5
Identifying Logical Processors in an MP System
After the BIOS has completed the MP initialization protocol, each logical processor
can be uniquely identified by its local APIC ID. Software can access these APIC IDs in
either of the following ways:
• Read APIC ID for a local APIC — Code running on a logical processor can read
APIC ID in one of two ways depending on the local APIC unit is operating in
x2APIC mode (see Intel® 64 Architecture x2APIC Specification)or in xAPIC
mode:
bit APIC ID can be read by executing a RDMSR instruction to read the
processor’s x2APIC ID register. This method is equivalent to executing CPUID
leaf 0BH described below.
— If the local APIC unit is operating in xAPIC mode, 8-bit APIC ID can be read by
executing a MOV instruction to read the processor’s local APIC ID register
(see Section 10.4.6, “Local APIC ID”). This is the ID to use for directing
physical destination mode interrupts to the processor.
• Read ACPI or MP table — As part of the MP initialization protocol, the BIOS
creates an ACPI table and an MP table. These tables are defined in the Multipro-
cessor Specification Version 1.4 and provide software with a list of the processors
in the system and their local APIC IDs. The format of the ACPI table is derived
from the ACPI specification, which is an industry standard power management
and platform configuration specification for MP systems.
• Read Initial APIC ID (If the process does not support CPUID leaf 0BH) — An
APIC ID is assigned to a logical processor during power up. This is the initial APIC
ID reported by CPUID.1:EBX[31:24] and may be different from the current value
topological relationship between logical processors for multi-processor systems
that do not support CPUID leaf 0BH.
Bits in the 8-bit initial APIC ID can be interpreted using several bit masks. Each
bit mask can be used to extract an identifier to represent a hierarchical level of
the multi-threading resource topology in an MP system (See Section 8.9.1,
“Hierarchical Mapping of Shared Resources”). The initial APIC ID may consist of
up to four bit-fields. In a non-clustered MP system, the field consists of up to
three bit fields.
This APIC ID is reported by CPUID.0BH:EDX[31:0] as a 32-bit value. Use the 32-
bit APIC ID and CPUID leaf 0BH to determine the topological relationship between
logical processors if the processor supports CPUID leaf 0BH.
Bits in the 32-bit x2APIC ID can be extracted into sub-fields using CPUID leaf 0BH
parameters. (See Section 8.9.1, “Hierarchical Mapping of Shared Resources”).
Figure 8-2 shows two examples of APIC ID bit fields in earlier single-core processors.
In single-core Intel Xeon processors, the APIC ID assigned to a logical processor
Vol. 3 8-33
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
during power-up and initialization is 8 bits. Bits 2:1 form a 2-bit physical package
identifier (which can also be thought of as a socket identifier). In systems that
configure physical processors in clusters, bits 4:3 form a 2-bit cluster ID. Bit 0 is used
package (see Section 8.9.3, “Hierarchical ID of Logical Processors in an MP System”).
For Intel Xeon processors that do not support Intel Hyper-Threading Technology, bit
0 is always set to 0; for Intel Xeon processors supporting Intel Hyper-Threading
Technology, bit 0 performs the same function as it does for Intel Xeon processor MP.
For more recent multi-core processors, see Section 8.9.1, “Hierarchical Mapping of
Shared Resources” for a complete description of the topological relationships
between logical processors and bit field locations within an initial APIC ID across Intel
64 and IA-32 processor families.
Note the number of bit fields and the width of bit-fields are dependent on processor
and platform hardware capabilities. Software should determine these at runtime.
When initial APIC IDs are assigned to logical processors, the value of APIC ID
assigned to a logical processor will respect the bit-field boundaries corresponding
core, physical package, etc. Additional examples of the bit fields in the initial APIC ID
of multi-threading capable systems are shown in Section 8.9.
APIC ID Format for Intel Xeon Processors that
do not Support Intel Hyper-Threading Technology
7
5
4
3
2
1
0
0
Reserved
Cluster
Processor ID
APIC ID Format for P6 Family Processors
7
4
3
2
1
0
Reserved
Cluster
Processor ID
Figure 8-2. Interpretation of APIC ID in Early MP Systems
For P6 family processors, the APIC ID that is assigned to a processor during power-
up and initialization is 4 bits (see Figure 8-2). Here, bits 0 and 1 form a 2-bit
processor (or socket) identifier and bits 2 and 3 form a 2-bit cluster ID.
8-34 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
8.5
INTEL® HYPER-THREADING TECHNOLOGY AND
Intel Hyper-Threading Technology and Intel multi-core technology are extensions to
Intel 64 and IA-32 architectures that enable a single physical processor to execute
two or more separate code streams (called threads) concurrently. In Intel Hyper-
Threading Technology, a single processor core provides two logical processors that
®
two or more processor cores. Both configurations require chipsets and a BIOS that
support the technologies.
Software should not rely on processor names to determine whether a processor
supports Intel Hyper-Threading Technology or Intel multi-core technology. Use the
CPUID instruction to determine processor capability (see Section 8.6.2, “Initializing
Multi-Core Processors”).
8.6
DETECTING HARDWARE MULTI-THREADING
SUPPORT AND TOPOLOGY
Use the CPUID instruction to detect the presence of hardware multi-threading
support in a physical processor. Hardware multi-threading can support several vari-
eties of multigrade and/or Intel Hyper-Threading Technology. CPUID instruction
provides several sets of parameter information to aid software enumerating topology
information. The relevant topology enumeration parameters provided by CPUID
include:
• Hardware Multi-Threading feature flag (CPUID.1:EDX[28] = 1) —
Indicates when set that the physical package is capable of supporting Intel
Hyper-Threading Technology and/or multiple cores.
• Processor topology enumeration parameters for 8-bit APIC ID:
— Addressable IDs for Logical processors in the same Package
(CPUID.1:EBX[23:16]) — Indicates the maximum number of addressable
ID for logical processors in a physical package. Within a physical package,
there may be addressable IDs that are not occupied by any logical
processors. This parameter does not represents the hardware capability of
2
the physical processor.
3
• Addressable IDs for processor cores in the same Package
4
(CPUID.(EAX=4, ECX=0 ):EAX[31:26] + 1 = Y) — Indicates the maximum
2. Operating system and BIOS may implement features that reduce the number of logical proces-
sors available in a platform to applications at runtime to less than the number of physical pack-
ages times the number of hardware-capable logical processors per package.
Vol. 3 8-35
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
number of addressable IDs attributable to processor cores (Y) in the physical
package.
• Extended Processor Topology Enumeration parameters for 32-bit APIC
ID: Intel 64 processors supporting CPUID leaf 0BH will assign unique APIC IDs to
each logical processor in the system. CPUID leaf 0BH reports the 32-bit APIC ID
and provide topology enumeration parameters. See CPUID instruction reference
pages in Intel® 64 and IA-32 Architectures Software Developer’s Manual,
Volume 2A.
The CPUID feature flag may indicate support for hardware multi-threading when only
one logical processor available in the package. In this case, the decimal value repre-
sented by bits 16 through 23 in the EBX register will have a value of 1.
Software should note that the number of logical processors enabled by system soft-
ware may be less than the value of “Addressable IDs for Logical processors”. Simi-
larly, the number of cores enabled by system software may be less than the value of
“Addressable IDs for processor cores”.
Software can detect the availability of the CPUID extended topology enumeration leaf
(0BH) by performing two steps:
• Check maximum input value for basic CPUID information by executing CPUID
with EAX= 0. If CPUID.0H:EAX is greater than or equal or 11 (0BH), then proceed
to next step,
• Check CPUID.EAX=0BH, ECX=0H:EBX is non-zero.
If both of the above conditions are true, extended topology enumeration leaf is avail-
able. Note the presence of CPUID leaf 0BH in a processor does not guarantee support
that the local APIC supports x2APIC. If CPUID.(EAX=0BH, ECX=0H):EBX returns
zero and maximum input value for basic CPUID information is greater than 0BH, then
CPUID.0BH leaf is not supported on that processor.
8.6.1
Initializing Processors
The initialization process for an MP system that contains processors supporting Intel
Hyper-Threading Technology is the same as for conventional MP systems (see
Section 8.4, “Multiple-Processor (MP) Initialization”). One logical processor in the
system is selected as the BSP and other processors (or logical processors) are desig-
nated as APs. The initialization process is identical to that described in Section 8.4.3,
“MP Initialization Protocol Algorithm for Intel Xeon Processors,” and Section 8.4.4,
“MP Initialization Example.”
3. Software must check CPUID for its support of leaf 4 when implementing support for multi-core. If
CPUID leaf 4 is not available at runtime, software should handle the situation as if there is only
one core per package.
4. Maximum number of cores in the physical package must be queried by executing CPUID with
EAX=4 and a valid ECX input value. Valid ECX input values start from 0.
8-36 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
During initialization, each logical processor is assigned an APIC ID that is stored in
the local APIC ID register for each logical processor. If two or more processors
supporting Intel Hyper-Threading Technology are present, each logical processor on
the system bus is assigned a unique ID (see Section 8.9.3, “Hierarchical ID of Logical
Processors in an MP System”). Once logical processors have APIC IDs, software
8.6.2
Initializing Multi-Core Processors
The initialization process for an MP system that contains multi-core Intel 64 or IA-32
processors is the same as for conventional MP systems (see Section 8.4, “Multiple-
Processor (MP) Initialization”). A logical processor in one core is selected as the BSP;
other logical processors are designated as APs.
During initialization, each logical processor is assigned an APIC ID. Once logical
processors have APIC IDs, software may communicate with them by sending APIC
IPI messages.
8.6.3
Executing Multiple Threads on an Intel® 64 or IA-32
Processor Supporting Hardware Multi-Threading
Upon completing the operating system boot-up procedure, the bootstrap processor
(BSP) executes operating system code. Other logical processors are placed in the
halt state. To execute a code stream (thread) on a halted logical processor, the oper-
ating system issues an interprocessor interrupt (IPI) addressed to the halted logical
processor. In response to the IPI, the processor wakes up and begins executing the
thread identified by the interrupt vector received as part of the IPI.
To manage execution of multiple threads on logical processors, an operating system
can use conventional symmetric multiprocessing (SMP) techniques. For example, the
operating-system can use a time-slice or load balancing mechanism to periodically
interrupt each of the active logical processors. Upon interrupting a logical processor,
the operating system checks its run queue for a thread waiting to be executed and
dispatches the thread to the interrupted logical processor.
8.6.4
Handling Interrupts on an IA-32 Processor Supporting
Hardware Multi-Threading
Interrupts are handled on processors supporting Intel Hyper-Threading Technology
as they are on conventional MP systems. External interrupts are received by the I/O
APIC, which distributes them as interrupt messages to specific logical processors
(see Figure 8-3).
Logical processors can also send IPIs to other logical processors by writing to the ICR
register of its local APIC (see Section 10.7, “Issuing Interprocessor Interrupts”). This
also applies to dual-core processors.
Vol. 3 8-37
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
Intel Processor with Intel
Intel Processor with Intel
Hyper-Threading Technology Hyper-Threading Technology
Logical
Logical
Logical
Logical
Processor 0 Processor 1
Processor 0 Processor 1
Processor Core
Processor Core
Local APIC
Local APIC
Local APIC
IPIs
Local APIC
Bus Interface
Bus Interface
Interrupt
Messages
Interrupt
Messages
IPIs
Interrupt Messages
Bridge
PCI
External
Interrupts
I/O APIC
System Chip Set
8.7
INTEL® HYPER-THREADING TECHNOLOGY
ARCHITECTURE
Figure 8-4 shows a generalized view of an Intel processor supporting Intel Hyper-
Threading Technology, using the original Intel Xeon processor MP as an example.
This implementation of the Intel Hyper-Threading Technology consists of two logical
processors (each represented by a separate architectural state) which share the
processor’s execution engine and the bus interface. Each logical processor also has
its own advanced programmable interrupt controller (APIC).
8-38 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
Logical
Logical
Processor 0 Processor 1
Architectural Architectural
State
State
Execution Engine
Local APIC Local APIC
Bus Interface
System Bus
Figure 8-4. IA-32 Processor with Two Logical Processors Supporting Intel HT
Technology
8.7.1
State of the Logical Processors
The following features are part of the architectural state of logical processors within
Intel 64 or IA-32 processors supporting Intel Hyper-Threading Technology. The
features can be subdivided into three groups:
• Duplicated for each logical processor
• Shared by logical processors in a physical processor
• Shared or duplicated, depending on the implementation
The following features are duplicated for each logical processor:
• General purpose registers (EAX, EBX, ECX, EDX, ESI, EDI, ESP, and EBP)
• Segment registers (CS, DS, SS, ES, FS, and GS)
• EFLAGS and EIP registers. Note that the CS and EIP/RIP registers for each logical
processor point to the instruction stream for the thread being executed by the
logical processor.
• x87 FPU registers (ST0 through ST7, status word, control word, tag word, data
operand pointer, and instruction pointer)
• MMX registers (MM0 through MM7)
• XMM registers (XMM0 through XMM7) and the MXCSR register
• Control registers and system table pointer registers (GDTR, LDTR, IDTR, task
register)
Vol. 3 8-39
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
• Debug registers (DR0, DR1, DR2, DR3, DR6, DR7) and the debug control MSRs
• Machine check global status (IA32_MCG_STATUS) and machine check capability
(IA32_MCG_CAP) MSRs
• Thermal clock modulation and ACPI Power management control MSRs
• Time stamp counter MSRs
• Most of the other MSR registers, including the page attribute table (PAT). See the
exceptions below.
• Local APIC registers.
• Additional general purpose registers (R8-R15), XMM registers (XMM8-XMM15),
control register, IA32_EFER on Intel 64 processors.
The following features are shared by logical processors:
• Memory type range registers (MTRRs)
Whether the following features are shared or duplicated is implementation-specific:
• IA32_MISC_ENABLE MSR (MSR address 1A0H)
• Machine check architecture (MCA) MSRs (except for the IA32_MCG_STATUS and
IA32_MCG_CAP MSRs)
• Performance monitoring control and counter MSRs
8.7.2
APIC Functionality
When a processor supporting Intel Hyper-Threading Technology support is initialized,
each logical processor is assigned a local APIC ID (see Table 10-1). The local APIC ID
serves as an ID for the logical processor and is stored in the logical processor’s APIC
ID register. If two or more processors supporting Intel Hyper-Threading Technology
system bus is assigned a unique local APIC ID (see Section 8.9.3, “Hierarchical ID of
Logical Processors in an MP System”).
Software communicates with local processors using the APIC’s interprocessor inter-
rupt (IPI) messaging facility. Setup and programming for APICs is identical in proces-
sors that support and do not support Intel Hyper-Threading Technology. See Chapter
10, “Advanced Programmable Interrupt Controller (APIC),” for a detailed discussion.
8.7.3
Memory Type Range Registers (MTRR)
MTRRs in a processor supporting Intel Hyper-Threading Technology are shared by
logical processors. When one logical processor updates the setting of the MTRRs,
settings are automatically shared with the other logical processors in the same phys-
ical package.
The architectures require that all MP systems based on Intel 64 and IA-32 processors
(this includes logical processors) must use an identical MTRR memory map. This
8-40 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
gives software a consistent view of memory, independent of the processor on which
mation on setting up MTRRs.
8.7.4
Page Attribute Table (PAT)
Each logical processor has its own PAT MSR (IA32_PAT). However, as described in
Section 11.12, “Page Attribute Table (PAT),” the PAT MSR settings must be the same
for all processors in a system, including the logical processors.
8.7.5
Machine Check Architecture
NetBurst microarchitecture, all of the machine check architecture (MCA) MSRs
(except for the IA32_MCG_STATUS and IA32_MCG_CAP MSRs) are duplicated for
each logical processor. This permits logical processors to initialize, configure, query,
and handle machine-check exceptions simultaneously within the same physical
processor. The design is compatible with machine check exception handlers that
follow the guidelines given in Chapter 15, “Machine-Check Architecture.”
The IA32_MCG_STATUS MSR is duplicated for each logical processor so that its
machine check in progress bit field (MCIP) can be used to detect recursion on the
part of MCA handlers. In addition, the MSR allows each logical processor to deter-
mine that a machine-check exception is in progress independent of the actions of
another logical processor in the same physical package.
Because the logical processors within a physical package are tightly coupled with
respect to shared hardware resources, both logical processors are notified of
machine check errors that occur within a given physical processor. If machine-check
exceptions are enabled when a fatal error is reported, all the logical processors within
a physical package are dispatched to the machine-check exception handler. If
machine-check exceptions are disabled, the logical processors enter the shutdown
state and assert the IERR# signal.
When enabling machine-check exceptions, the MCE flag in control register CR4
should be set for each logical processor.
On Intel Atom family processors that support Intel Hyper-Threading Technology, the
MCA facilities are shared between all logical processors on the same processor core.
8.7.6
Debug Registers and Extensions
Each logical processor has its own set of debug registers (DR0, DR1, DR2, DR3, DR6,
DR7) and its own debug control MSR. These can be set to control and record debug
information for each logical processor independently. Each logical processor also has
its own last branch records (LBR) stack.
Vol. 3 8-41
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
8.7.7
Performance Monitoring Counters
Performance counters and their companion control MSRs are shared between the
logical processors within a processor core for processors based on Intel NetBurst
microarchitecture. As a result, software must manage the use of these resources.
The performance counter interrupts, events, and precise event monitoring support
can be set up and allocated on a per thread (per logical processor) basis.
See Section 30.9, “Performance Monitoring and Intel Hyper-Threading Technology in
Processors Based on Intel NetBurst Microarchitecture,” for a discussion of perfor-
mance monitoring in the Intel Xeon processor MP.
In Intel Atom processor family that support Intel Hyper-Threading Technology, the
performance counters (general-purpose and fixed-function counters) and their
companion control MSRs are duplicated for each logical processor.
8.7.8
IA32_MISC_ENABLE MSR
The IA32_MISC_ENABLE MSR (MSR address 1A0H) is generally shared between the
logical processors in a processor core supporting Intel Hyper-Threading Technology.
However, some bit fields within IA32_MISC_ENABLES MSR may be duplicated per
logical processor. The partition of shared or duplicated bit fields within
IA32_MISC_ENABLES is implementation dependent. Software should program dupli-
cated fields carefully on all logical processors in the system to ensure consistent
behavior.
8.7.9
Memory Ordering
The logical processors in an Intel 64 or IA-32 processor supporting Intel Hyper-
Threading Technology obey the same rules for memory ordering as Intel 64 or IA-32
processors without Intel HT Technology (see Section 8.2, “Memory Ordering”). Each
logical processor uses a processor-ordered memory model that can be further
defined as “write-ordered with store buffer forwarding.” All mechanisms for strength-
ening or weakening the memory-ordering model to handle special programming situ-
ations apply to each logical processor.
8.7.10
Serializing Instructions
As a general rule, when a logical processor in a processor supporting Intel Hyper-
Threading Technology executes a serializing instruction, only that logical processor is
affected by the operation. An exception to this rule is the execution of the WBINVD,
INVD, and WRMSR instructions; and the MOV CR instruction when the state of the CD
flag in control register CR0 is modified. Here, both logical processors are serialized.
8-42 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
8.7.11
MICROCODE UPDATE Resources
In an Intel processor supporting Intel Hyper-Threading Technology, the microcode
update facilities are shared between the logical processors; either logical processor
can initiate an update. Each logical processor has its own BIOS signature MSR
(IA32_BIOS_SIGN_ID at MSR address 8BH). When a logical processor performs an
update for the physical processor, the IA32_BIOS_SIGN_ID MSRs for resident logical
processors are updated with identical information. If logical processors initiate an
update simultaneously, the processor core provides the necessary synchronization
needed to insure that only one update is performed at a time.
Operating system microcode update drivers that adhere to Intel’s guidelines do not
need to be modified to run on processors supporting Intel Hyper-Threading Tech-
nology.
8.7.12
Self Modifying Code
Intel processors supporting Intel Hyper-Threading Technology support self-modifying
code, where data writes modify instructions cached or currently in flight. They also
support cross-modifying code, where on an MP system writes generated by one
processor modify instructions cached or currently in flight on another. See Section
8.1.3, “Handling Self- and Cross-Modifying Code,” for a description of the require-
ments for self- and cross-modifying code in an IA-32 processor.
8.7.13
Implementation-Specific Intel HT Technology Facilities
The following non-architectural facilities are implementation-specific in IA-32 proces-
sors supporting Intel Hyper-Threading Technology:
• Caches
• Translation lookaside buffers (TLBs)
• Thermal monitoring facilities
The Intel Xeon processor MP implementation is described in the following sections.
8.7.13.1 Processor Caches
For processors supporting Intel Hyper-Threading Technology, the caches are shared.
Any cache manipulation instruction that is executed on one logical processor has a
global effect on the cache hierarchy of the physical processor. Note the following:
• WBINVD instruction — The entire cache hierarchy is invalidated after modified
data is written back to memory. All logical processors are stopped from executing
until after the write-back and invalidate operation is completed. A special bus
cycle is sent to all caching agents. The amount of time or cycles for WBINVD to
complete will vary due to the size of different cache hierarchies and other factors.
Vol. 3 8-43
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
As a consequence, the use of the WBINVD instruction can have an impact on
interrupt/event response time.
• INVD instruction — The entire cache hierarchy is invalidated without writing
back modified data to memory. All logical processors are stopped from executing
until after the invalidate operation is completed. A special bus cycle is sent to all
caching agents.
• CLFLUSH instruction — The specified cache line is invalidated from the cache
hierarchy after any modified data is written back to memory and a bus cycle is
sent to all caching agents, regardless of which logical processor caused the cache
line to be filled.
• CD flag in control register CR0 — Each logical processor has its own CR0
control register, and thus its own CD flag in CR0. The CD flags for the two logical
processors are ORed together, such that when any logical processor sets its CD
flag, the entire cache is nominally disabled.
8.7.13.2 Processor Translation Lookaside Buffers (TLBs)
shared. The instruction cache TLB may be duplicated or shared in each logical
processor, depending on implementation specifics of different processor families.
Entries in the TLBs are tagged with an ID that indicates the logical processor that
initiated the translation. This tag applies even for translations that are marked global
using the page-global feature for memory paging. See Section 4.10, “Caching Trans-
lation Information,” for information about global translations.
When a logical processor performs a TLB invalidation operation, only the TLB entries
that are tagged for that logical processor are guaranteed to be flushed. This protocol
applies to all TLB invalidation operations, including writes to control registers CR3
and CR4 and uses of the INVLPG instruction.
8.7.13.3 Thermal Monitor
In a processor that supports Intel Hyper-Threading Technology, logical processors
share the catastrophic shutdown detector and the automatic thermal monitoring
mechanism (see Section 14.5, “Thermal Monitoring and Protection”). Sharing results
in the following behavior:
• If the processor’s core temperature rises above the preset catastrophic shutdown
temperature, the processor core halts execution, which causes both logical
processors to stop execution.
• When the processor’s core temperature rises above the preset automatic thermal
monitor trip temperature, the clock speed of the processor core is automatically
modulated, which effects the execution speed of both logical processors.
For software controlled clock modulation, each logical processor has its own
IA32_CLOCK_MODULATION MSR, allowing clock modulation to be enabled or
8-44 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
disabled on a logical processor basis. Typically, if software controlled clock modula-
tion is going to be used, the feature must be enabled for all the logical processors
within a physical processor and the modulation duty cycle must be set to the same
value for each logical processor. If the duty cycle values differ between the logical
processors, the processor clock will be modulated at the highest duty cycle selected.
8.7.13.4 External Signal Compatibility
This section describes the constraints on external signals received through the pins
of a processor supporting Intel Hyper-Threading Technology and how these signals
are shared between its logical processors.
• STPCLK# — A single STPCLK# pin is provided on the physical package of the
Intel Xeon processor MP. External control logic uses this pin for power
management within the system. When the STPCLK# signal is asserted, the
processor core transitions to the stop-grant state, where instruction execution is
halted but the processor core continues to respond to snoop transactions.
Regardless of whether the logical processors are active or halted when the
STPCLK# signal is asserted, execution is stopped on both logical processors and
neither will respond to interrupts.
In MP systems, the STPCLK# pins on all physical processors are generally tied
together. As a result this signal affects all the logical processors within the system
simultaneously.
• LINT0 and LINT1 pins — A processor supporting Intel Hyper-Threading
Technology has only one set of LINT0 and LINT1 pins, which are shared between
the logical processors. When one of these pins is asserted, both logical
processors respond unless the pin has been masked in the APIC local vector
tables for one or both of the logical processors.
Typically in MP systems, the LINT0 and LINT1 pins are not used to deliver
interrupts to the logical processors. Instead all interrupts are delivered to the
local processors through the I/O APIC.
• A20M# pin — On an IA-32 processor, the A20M# pin is typically provided for
compatibility with the Intel 286 processor. Asserting this pin causes bit 20 of the
physical address to be masked (forced to zero) for all external bus memory
accesses. Processors supporting Intel Hyper-Threading Technology provide one
A20M# pin, which affects the operation of both logical processors within the
physical processor.
The functionality of A20M# is used primarily by older operating systems and not
used by modern operating systems. On newer Intel 64 processors, A20M# may
be absent.
Vol. 3 8-45
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
8.8
MULTI-CORE ARCHITECTURE
This section describes the architecture of Intel 64 and IA-32 processors supporting
dual-core and quad-core technology. The discussion is applicable to the Intel Pentium
processor Extreme Edition, Pentium D, Intel Core Duo, Intel Core 2 Duo, Dual-core
Intel Xeon processor, Intel Core 2 Quad processors, and quad-core Intel Xeon
processors. Features vary across different microarchitectures and are detectable
using CPUID.
In general, each processor core has dedicated microarchitectural resources identical
to a single-processor implementation of the underlying microarchitecture without
hardware multi-threading capability. Each logical processor in a dual-core processor
(whether supporting Intel Hyper-Threading Technology or not) has its own APIC
functionality, PAT, machine check architecture, debug registers and extensions. Each
logical processor handles serialization instructions or self-modifying code on its own.
Memory order is handled the same way as in Intel Hyper-Threading Technology.
The topology of the cache hierarchy (with respect to whether a given cache level is
shared by one or more processor cores or by all logical processors in the physical
package) depends on the processor implementation. Software must use the deter-
ministic cache parameter leaf of CPUID instruction to discover the cache-sharing
topology between the logical processors in a multi-threading environment.
8.8.1
The topological composition of processor cores and logical processors in a multi-core
processor can be discovered using CPUID. Within each processor core, one or more
System software must follow the requirement MP initialization sequences (see
Section 8.4, “Multiple-Processor (MP) Initialization”) to recognize and enable logical
processors. At runtime, software can enumerate those logical processors enabled by
system software to identify the topological relationships between these logical
processors. (See Section 8.9.5, “Identifying Topological Relationships in a MP
System”).
8.8.2
Memory Type Range Registers (MTRR)
ical processor supports Intel Hyper-Threading Technology. MTRR is not shared
between logical processors located in different cores or different physical packages.
The Intel 64 and IA-32 architectures require that all logical processors in an MP
system use an identical MTRR memory map. This gives software a consistent view of
memory, independent of the processor on which it is running.
See Section 11.11, “Memory Type Range Registers (MTRRs).”
8-46 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
8.8.3
Performance Monitoring Counters
Performance counters and their companion control MSRs are shared between two
logical processors sharing a processor core if the processor core supports Intel
Hyper-Threading Technology and is based on Intel NetBurst microarchitecture. They
are not shared between logical processors in different cores or different physical
packages. As a result, software must manage the use of these resources, based on
the topology of performance monitoring resources. Performance counter interrupts,
events, and precise event monitoring support can be set up and allocated on a per
thread (per logical processor) basis.
See Section 30.9, “Performance Monitoring and Intel Hyper-Threading Technology in
Processors Based on Intel NetBurst Microarchitecture.”
8.8.4
IA32_MISC_ENABLE MSR
Some bit fields in IA32_MISC_ENABLE MSR (MSR address 1A0H) may be shared
between two logical processors sharing a processor core, or may be shared between
different cores in a physical processor. See Appendix B, “Model-Specific Registers
(MSRs)”.
8.8.5
MICROCODE UPDATE Resources
Microcode update facilities are shared between two logical processors sharing a
processor core if the physical package supports Intel Hyper-Threading Technology.
They are not shared between logical processors in different cores or different phys-
ical packages. Either logical processor that has access to the microcode update
facility can initiate an update.
Each logical processor has its own BIOS signature MSR (IA32_BIOS_SIGN_ID at MSR
address 8BH). When a logical processor performs an update for the physical
processor, the IA32_BIOS_SIGN_ID MSRs for resident logical processors are
updated with identical information. If logical processors initiate an update simulta-
neously, the processor core provides the synchronization needed to ensure that only
one update is performed at a time.
8.9
PROGRAMMING CONSIDERATIONS FOR HARDWARE
MULTI-THREADING CAPABLE PROCESSORS
In a multi-threading environment, there may be certain hardware resources that are
physically shared at some level of the hardware topology. In the multi-processor
systems, typically bus and memory sub-systems are physically shared between
multiple sockets. Within a hardware multi-threading capable processors, certain
resources are provided for each processor core, while other resources may be
Vol. 3 8-47
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
®
provided for each logical processors (see Section 8.7, “Intel Hyper-Threading Tech-
nology Architecture,” and Section 8.8, “Multi-Core Architecture”).
From a software programming perspective, control transfer of processor operation is
managed at the granularity of logical processor (operating systems dispatch a
runnable task by allocating an available logical processor on the platform). To
manage the topology of shared resources in a multi-threading environment, it may
be useful for software to understand and manage resources that are shared by more
8.9.1
Hierarchical Mapping of Shared Resources
The APIC_ID value associated with each logical processor in a multi-processor
system is unique (see Section 8.6, “Detecting Hardware Multi-Threading Support and
Topology”). This 8-bit or 32-bit value can be decomposed into sub-fields, where each
sub-field corresponds a hierarchical level of the topological mapping of hardware
resources.
The decomposition of an APIC_ID may consist of several sub fields representing the
topology within a physical processor package, the higher-order bits of an APIC ID
may also be used by cluster vendors to represent the topology of cluster nodes of
each coherent multiprocessor systems. If the processor does not support CPUID leaf
0BH, the 8-bit initial APIC ID can represent 4 levels of hierarchy:
• Cluster — Some multi-threading environments consists of multiple clusters of
multi-processor systems. The CLUSTER_ID sub-field is usually supported by
vendor firmware to distinguish different clusters. For non-clustered systems,
CLUSTER_ID is usually 0 and system topology is reduced to three levels of
hierarchy.
• Package — A multi-processor system consists of two or more sockets, each
mates with a physical processor package. The PACKAGE_ID sub-field distin-
guishes different physical packages within a cluster.
• Core — A physical processor package consists of one or more processor cores.
The CORE_ID sub-field distinguishes processor cores in a package. For a single-
• SMT — A processor core provides one or more logical processors sharing
execution resources. The SMT_ID sub-field distinguishes logical processors in a
core. The width of this bit field is non-zero if a processor core provides more than
one logical processors.
SMT and CORE sub-fields are bit-wise contiguous in the APIC_ID field (see
Figure 8-5).
8-48 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
0
X
X=31 if x2APIC is supported
Otherwise X= 7
Reserved
Cluster ID
Package ID
Core ID
SMT ID
Figure 8-5. Generalized Four level Interpretation of the APIC ID
plus several levels of topology within the physical processor package. The exact
number of hierarchical levels within a physical processor package must be enumer-
similar to that represented by 8-bit Initial APIC ID. In general, CPUID leaf 0BH can
support topology enumeration algorithm that decompose a 32-bit APIC ID into more
than four sub-fields (see Figure 8-6).
The width of each sub-field depends on hardware and software configurations. Field
widths can be determined at runtime using the algorithm discussed below (Example
8-16 through Example 8-20).
Figure 7-6 depicts the relationships of three of the hierarchical sub-fields in a hypo-
thetical MP system. The value of valid APIC_IDs need not be contiguous across
package boundary or core boundaries.
0
31
Package
Reserved
R
Cluster ID
Package ID
Q ID
SMT
Q
R ID
SMT ID
Physical Processor Topology
32-bit APIC ID Composition
Figure 8-6. Conceptual Five-level Topology and 32-bit APIC ID Composition
Vol. 3 8-49
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
8.9.2
Hierarchical Mapping of CPUID Extended Topology Leaf
CPUID leaf 0BH provides enumeration parameters for software to identify each hier-
archy of the processor topology in a deterministic manner. Each hierarchical level of
the topology starting from the SMT level is represented numerically by a sub-leaf
index within the CPUID 0BH leaf. Each level of the topology is mapped to a sub-field
in the APIC ID, following the general relationship depicted in Figure 8-6. This mech-
processor package and the bit-width of each sub-field of x2APIC ID directly. For
example,
• Starting from sub-leaf index 0 and incrementing ECX until CPUID.(EAX=0BH,
ECX=N):ECX[15:8] returns an invalid “level type“ encoding. The number of
levels within the physical processor package is “N“ (excluding PACKAGE). Using
Figure 8-6 as an example, CPUID.(EAX=0BH, ECX=3):ECX[15:8] will report
00H, indicating sub leaf 03H is invalid. This is also depicted by a pseudo code
example:
Example 8-16. Number of Levels Below the Physical Processor Package
Byte type = 1;
s = 0;
While ( type ) {
EAX = 0BH; // query each sub leaf of CPUID leaf 0BH
ECX = s;
CPUID;
type = ECX[15:8]; // examine level type encoding
s ++;
}
N = ECX[7:0];
• Sub-leaf index 0 (ECX= 0 as input) provides enumeration parameters to extract
the SMT sub-field of x2APIC ID. If EAX = 0BH, and ECX =0 is specified as input
when executing CPUID, CPUID.(EAX=0BH, ECX=0):EAX[4:0] reports a value (a
right-shift count) that allow software to extract part of x2APIC ID to distinguish
the next higher topological entities above the SMT level. This value also
corresponds to the bit-width of the sub-field of x2APIC ID corresponding the
hierarchical level with sub-leaf index 0.
• For each subsequent higher sub-leaf index m, CPUID.(EAX=0BH,
ECX=m):EAX[4:0] reports the right-shift count that will allow software to extract
part of x2APIC ID to distinguish higher-level topological entities. This means the
right-shift value at of sub-leaf m, corresponds to the least significant (m+1)
subfields of the 32-bit x2APIC ID.
Example 8-17. BitWidth Determination of x2APIC ID Subfields
8-50 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
For m = 0, m < N, m ++;
cumulative_width[m] = CPUID.(EAX=0BH, ECX= m): EAX[4:0]; }
{
BitWidth[0] = cumulative_width[0];
BitWidth[m] = cumulative_width[m] - cumulative_width[m-1];
Currently, only the following encoding of hierarchical level type are defined: 0
(invalid), 1 (SMT), and 2 (core). Software must not assume any “level type“ encoding
value to be related to any sub-leaf index, except sub-leaf 0.
Example 8-16 and Example 8-17 represent the general technique for using CPUID
leaf 0BH to enumerate processor topology of more than two levels of hierarchy inside
a physical package. Most processor families to date requires only “SMT” and “CORE”
levels within a physical package. The examples in later sections will focus on these
three-level topology only.
8.9.3
Hierarchical ID of Logical Processors in an MP System
For Intel 64 and IA-32 processors, system hardware establishes an 8-bit initial APIC
ID (or 32-bit APIC ID if the processor supports CPUID leaf 0BH) that is unique for
each logical processor following power-up or RESET (see Section 8.6.1). Each logical
processor on the system is allocated an initial APIC ID. BIOS may implement features
system bus. Those logical processors that are not available to applications at runtime
are halted during the OS boot process. As a result, the number valid local APIC_IDs
that can be queried by affinitizing-current-thread-context (See Example 8-22) is
limited to the number of logical processors enabled at runtime by the OS boot
process.
Table 8-1 shows an example of the 8-bit APIC IDs that are initially reported for logical
processors in a system with four Intel Xeon MP processors that support Intel Hyper-
Threading Technology (a total of 8 logical processors, each physical package has two
processor cores and supports Intel Hyper-Threading Technology). Of the two logical
processors within a Intel Xeon processor MP, logical processor 0 is designated the
primary logical processor and logical processor 1 as the secondary logical processor.
Vol. 3 8-51
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
SMT_ID
T0 T1
Core 0
T0
T1
Core1
T0 T1
Core 0
T0 T1
Core1
Core ID
Package ID
Package 1
Package 0
Figure 8-7. Topological Relationships between Hierarchical IDs in a Hypothetical MP
Platform
Table 8-1. Initial APIC IDs for the Logical Processors in a System that has Four Intel
Xeon MP Processors Supporting Intel Hyper-Threading Technology1
Initial APIC ID
Package ID
Core ID
0H
SMT ID
0H
0H
1H
2H
3H
4H
5H
6H
7H
0H
0H
1H
1H
2H
2H
3H
3H
0H
1H
0H
0H
0H
1H
0H
0H
0H
1H
0H
0H
0H
1H
NOTE:
1. Because information on the number of processor cores in a physical package was not available
in early single-core processors supporting Intel Hyper-Threading Technology, the core ID can be
treated as 0.
Table 8-2 shows the initial APIC IDs for a hypothetical situation with a dual processor
system. Each physical package providing two processor cores, and each processor
core also supporting Intel Hyper-Threading Technology.
8-52 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
Table 8-2. Initial APIC IDs for the Logical Processors in a System that has Two
Physical Processors Supporting Dual-Core and Intel Hyper-Threading Technology
Initial APIC ID
Package ID
Core ID
0H
SMT ID
0H
0H
1H
2H
3H
4H
5H
6H
7H
0H
0H
0H
0H
1H
1H
1H
1H
0H
1H
1H
0H
1H
1H
0H
0H
0H
1H
1H
0H
1H
1H
8.9.3.1
Hierarchical ID of Logical Processors with x2APIC ID
Table 8-3 shows an example of possible x2APIC ID assignments for a dual processor
system that support x2APIC. Each physical package providing four processor cores,
and each processor core also supporting Intel Hyper-Threading Technology. Note that
the x2APIC ID need not be contiguous in the system.
Table 8-3. Example of Possible x2APIC ID Assignment in a System that has Two
Physical Processors Supporting x2APIC and Intel Hyper-Threading Technology
x2APIC ID
0H
Package ID
0H
Core ID
0H
SMT ID
0H
1H
0H
0H
1H
2H
0H
1H
0H
3H
0H
1H
1H
4H
0H
2H
0H
5H
0H
2H
1H
6H
0H
3H
0H
7H
0H
3H
1H
10H
11H
12H
13H
14H
1H
0H
0H
1H
0H
1H
1H
1H
0H
1H
1H
1H
1H
2H
0H
Vol. 3 8-53
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
Table 8-3. Example of Possible x2APIC ID Assignment in a System that has Two
Physical Processors Supporting x2APIC and Intel Hyper-Threading Technology
x2APIC ID
15H
Package ID
Core ID
2H
SMT ID
1H
1H
1H
1H
16H
3H
0H
17H
3H
1H
8.9.4
Algorithm for Three-Level Mappings of APIC_ID
Software can gather the initial APIC_IDs for each logical processor supported by the
5
operating system at runtime and extract identifiers corresponding to the three
levels of sharing topology (package, core, and SMT). The three-level algorithms
below focus on a non-clustered MP system for simplicity. They do not assume APIC
IDs are contiguous or that all logical processors on the platform are enabled.
Intel supports multi-threading systems where all physical processors report identical
6
values in CPUID leaf 0BH, CPUID.1:EBX[23:16]), CPUID.4 :EAX[31:26], and
7
CPUID.4 :EAX[25:14]. The algorithms below assume the target system has
symmetry across physical package boundaries with respect to the number of logical
processors per package, number of cores per package, and cache topology within a
package.
The extraction algorithm (for three-level mappings from an APIC ID) uses the
general procedure depicted in Example 8-18, and is supplemented by more detailed
descriptions on the derivation of topology enumeration parameters for extraction bit
masks:
1. Detect hardware multi-threading support in the processor.
2. Derive a set of bit masks that can extract the sub ID of each hierarchical level of
the topology. The algorithm to derive extraction bit masks for
SMT_ID/CORE_ID/PACKAGE_ID differs based on APIC ID is 32-bit (see step 3
below) or 8-bit (see step 4 below):
3. If the processor supports CPUID leaf 0BH, each APIC ID contains a 32-bit value,
the topology enumeration parameters needed to derive three-level extraction bit
masks are:
5. As noted in Section 8.6 and Section 8.9.3, the number of logical processors supported by the OS
at runtime may be less than the total number logical processors available in the platform hard-
ware.
6. Maximum number of addressable ID for processor cores in a physical processor is obtained by
executing CPUID with EAX=4 and a valid ECX index, The ECX index start at 0.
7. Maximum number addressable ID for processor cores sharing the target cache level is obtained
by executing CPUID with EAX = 4 and the ECX index corresponding to the target cache level.
8-54 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
a. Query the right-shift value for the SMT level of the topology using CPUID leaf
0BH with ECX =0H as input. The number of bits to shift-right on x2APIC ID
processor cores) in the same physical package. This is also the width of the
bit mask to extract the SMT_ID.
b. Query CPUID leaf 0BH for the amount of bit shift to distinguish next higher-
level entities (e.g. physical processor packages) in the system. This describes
an explicit three-level-topology situation for commonly available processors.
Consult Example 8-17 to adapt to situations beyond three-level topology of a
physical processor. The width of the extraction bit mask can be used to derive
the cumulative extraction bitmask to extract the sub IDs of logical processors
(including different processor cores) in the same physical package. The
extraction bit mask to distinguish merely different processor cores can be
derived by xor’ing the SMT extraction bit mask from the cumulative
extraction bit mask.
c. Query the 32-bit x2APIC ID for the logical processor where the current thread
is executing.
d. Derive the extraction bit masks corresponding to SMT_ID, CORE_ID, and
PACKAGE_ID, starting from SMT_ID.
e. Apply each extraction bit mask to the 32-bit x2APIC ID to extract sub-field
IDs.
4. If the processor does not support CPUID leaf 0BH, each initial APIC ID contains
an 8-bit value, the topology enumeration parameters needed to derive extraction
bit masks are:
a. Query the size of address space for sub IDs that can accommodate logical
processors in a physical processor package. This size parameters
(CPUID.1:EBX[23:16]) can be used to derive the width of an extraction
bitmask to enumerate the sub IDs of different logical processors in the same
physical package.
b. Query the size of address space for sub IDs that can accommodate processor
cores in a physical processor package. This size parameters can be used to
derive the width of an extraction bitmask to enumerate the sub IDs of
processor cores in the same physical package.
c. Query the 8-bit initial APIC ID for the logical processor where the current
thread is executing.
d. Derive the extraction bit masks using respective address sizes corresponding
to SMT_ID, CORE_ID, and PACKAGE_ID, starting from SMT_ID.
e. Apply each extraction bit mask to the 8-bit initial APIC ID to extract sub-field
IDs.
Vol. 3 8-55
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
Example 8-18. Support Routines for Detecting Hardware Multi-Threading and Identifying the
Relationships Between Package, Core and Logical Processors
1.
Detect support for Hardware Multi-Threading Support in a processor.
// Returns a non-zero value if CPUID reports the presence of hardware multi-threading
// support in the physical package where the current logical processor is located.
// This does not guarantee BIOS or OS will enable all logical processors in the physical
// package and make them available to applications.
// Returns zero if hardware multi-threading is not present.
#define HWMT_BIT 0x10000000
unsigned int HWMTSupported(void)
{
// ensure cpuid instruction is supported
execute cpuid with eax = 0 to get vendor string
execute cpuid with eax = 1 to get feature flag and signature
// Check to see if this a Genuine Intel Processor
if (vendor string EQ GenuineIntel) {
return (feature_flag_edx & HWMT_BIT); // bit 28
}
return 0;
}
Example 8-19. Support Routines for Identifying Package, Core and Logical Processors from
32-bit x2APIC ID
a.
Derive the extraction bitmask for logical processors in a processor core and
associated mask offset for different cores.
int DeriveSMT_Mask_Offsets (void)
{
if (!HWMTSupported()) return -1;
execute cpuid with eax = 11, ECX = 0;
If (returned level type encoding in ECX[15:8] does not match SMT) return -1;
Mask_SMT_shift = EAX[4:0]; // # bits shift right of APIC ID to distinguish different cores
SMT_MASK = ~( (-1) << Mask_SMT_shift); // shift left to derive extraction bitmask for SMT_ID
return 0;
}
b.
Derive the extraction bitmask for processor cores in a physical processor package
and associated mask offset for different packages.
8-56 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
int DeriveCore_Mask_Offsets (void)
{
if (!HWMTSupported()) return -1;
execute cpuid with eax = 11, ECX = 0;
while( ECX[15:8] ) { // level type encoding is valid
If (returned level type encoding in ECX[15:8] matches CORE) {
Mask_Core_shift = EAX[4:0]; // needed to distinguish different physical packages
COREPlusSMT_MASK = ~( (-1) << Mask_Core_shift);
CORE_MASK = COREPlusSMT_MASK ^ SMT_MASK;
PACKAGE_MASK = (-1) << Mask_Core_shift;
return 0
}
ECX ++;
execute cpuid with eax = 11;
}
return -1;
}
c.
Query the x2APIC ID of a logical processor.
APIC_IDs for each logical processor.
unsigned char Getx2APIC_ID (void)
{
unsigned reg_edx = 0;
execute cpuid with eax = 11, ECX = 0
store returned value of edx
return (unsigned) (reg_edx) ;
}
Example 8-20. Support Routines for Identifying Package, Core and Logical Processors from 8-
bit Initial APIC ID
a.
Find the size of address space for logical processors in a physical processor
package.
#define NUM_LOGICAL_BITS 0x00FF0000
// Use the mask above and CPUID.1.EBX[23:16] to obtain the max number of addressable IDs
// for logical processors in a physical package,
//Returns the size of address space of logical processors in a physical processor package;
// Software should not assume the value to be a power of 2.
Vol. 3 8-57
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
unsigned char MaxLPIDsPerPackage(void)
{
if (!HWMTSupported()) return 1;
execute cpuid with eax = 1
store returned value of ebx
return (unsigned char) ((reg_ebx & NUM_LOGICAL_BITS) >> 16);
}
b.
Find the size of address space for processor cores in a physical processor package.
// Returns the max number of addressable IDs for processor cores in a physical processor package;
// Software should not assume cpuid reports this value to be a power of 2.
unsigned MaxCoreIDsPerPackage(void)
{
if (!HWMTSupported()) return (unsigned char) 1;
if cpuid supports leaf number 4
{ // we can retrieve multi-core topology info using leaf 4
execute cpuid with eax = 4, ecx = 0
store returned value of eax
return (unsigned) ((reg_eax >> 26) +1);
}
else // must be a single-core processor
return 1;
}
c.
Query the initial APIC ID of a logical processor.
#define INITIAL_APIC_ID_BITS 0xFF000000 // CPUID.1.EBX[31:24] initial APIC ID
// Returns the 8-bit unique initial APIC ID for the processor running the code.
// Software can use OS services to affinitize the current thread to each logical processor
// available under the OS to gather the initial APIC_IDs for each logical processor.
unsigned GetInitAPIC_ID (void)
{
unsigned int reg_ebx = 0;
execute cpuid with eax = 1
store returned value of ebx
return (unsigned) ((reg_ebx & INITIAL_APIC_ID_BITS) >> 24;
}
d.
Find the width of an extraction bitmask from the maximum count of the bit-field
(address size).
8-58 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
// Returns the mask bit width of a bit field from the maximum count that bit field can represent.
// This algorithm does not assume ‘address size’ to have a value equal to power of 2.
// Address size for SMT_ID can be calculated from MaxLPIDsPerPackage()/MaxCoreIDsPerPackage()
// Then use the routine below to derive the corresponding width of SMT extraction bitmask
// Address size for CORE_ID is MaxCoreIDsPerPackage(),
// Derive the bitwidth for CORE extraction mask similarly
unsigned FindMaskWidth(Unsigned Max_Count)
{unsigned int mask_width, cnt = Max_Count;
__asm {
mov eax, cnt
mov ecx, 0
mov mask_width, ecx
dec eax
bsr cx, ax
jz next
inc cx
mov mask_width, ecx
next:
mov eax, mask_width
}
return mask_width;
}
e.
Extract a sub ID from an 8-bit full ID, using address size of the sub ID and shift
count.
// The routine below can extract SMT_ID, CORE_ID, and PACKAGE_ID respectively from the init
APIC_ID
// To extract SMT_ID, MaxSubIDvalue is set to the address size of SMT_ID, Shift_Count = 0
// To extract CORE_ID, MaxSubIDvalue is the address size of CORE_ID, Shift_Count is width of SMT
extraction bitmask.
// Returns the value of the sub ID, this is not a zero-based value
Unsigned char GetSubID(unsigned char Full_ID, unsigned char MaxSubIDvalue, unsigned char
Shift_Count)
{
MaskWidth = FindMaskWidth(MaxSubIDValue);
MaskBits = ((uchar) (0xff << Shift_Count)) ^ ((uchar) (0xff << Shift_Count + MaskWidth)) ;
SubID = Full_ID & MaskBits;
Return SubID;
}
Vol. 3 8-59
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
Software must not assume local APIC_ID values in an MP system are consecutive.
Non-consecutive local APIC_IDs may be the result of hardware configurations or
debug features implemented in the BIOS or OS.
An identifier for each hierarchical level can be extracted from an 8-bit APIC_ID using
the support routines illustrated in Example 8-20. The appropriate bit mask and shift
value to construct the appropriate bit mask for each level must be determined
dynamically at runtime.
8.9.5
Identifying Topological Relationships in a MP System
To detect the number of physical packages, processor cores, or other topological
• Extract the three-level identifiers from the APIC ID of each logical processor
enabled by system software. The sequence is as follows (See the pseudo code
shown in Example 8-21 and support routines shown in Example 8-18):
•
The extraction start from the right-most bit field, corresponding to
SMT_ID, the innermost hierarchy in a three-level topology (See Figure
8-7). For the right-most bit field, the shift value of the working mask is
zero. The width of the bit field is determined dynamically using the
maximum number of logical processor per core, which can be derived
from information provided from CPUID.
•
•
To extract the next bit-field, the shift value of the working mask is
determined from the width of the bit mask of the previous step. The width
of the bit field is determined dynamically using the maximum number of
cores per package.
To extract the remaining bit-field, the shift value of the working mask is
determined from the maximum number of logical processor per package.
So the remaining bits in the APIC ID (excluding those bits already
extracted in the two previous steps) are extracted as the third identifier.
This applies to a non-clustered MP system, or if there is no need to
distinguish between PACKAGE_ID and CLUSTER_ID.
If there is need to distinguish between PACKAGE_ID and CLUSTER_ID,
PACKAGE_ID can be extracted using an algorithm similar to the
extraction of CORE_ID, assuming the number of physical packages in
each node of a clustered system is symmetric.
• Assemble the three-level identifiers of SMT_ID, CORE_ID, PACKAGE_IDs into
• To detect the number of physical packages: use PACKAGE_ID to identify those
logical processors that reside in the same physical package. This is shown in
Example 8-22b. This example also depicts a technique to construct a mask to
represent the logical processors that reside in the same package.
• To detect the number of processor cores: use CORE_ID to identify those logical
processors that reside in the same core. This is shown in Example 8-22. This
8-60 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
example also depicts a technique to construct a mask to represent the logical
processors that reside in the same core.
In Example 8-21, the numerical ID value can be obtained from the value extracted
with the mask by shifting it right by shift count. Algorithms below do not shift the
value. The assumption is that the SubID values can be compared for equivalence
without the need to shift.
Example 8-21. Pseudo Code Depicting Three-level Extraction Algorithm
For Each local_APIC_ID{
// Calculate SMT_MASK, the bit mask pattern to extract SMT_ID,
// SMT_MASK is determined using topology enumertaion parameters
// from CPUID leaf 0BH (Example 8-19);
// otherwise, SMT_MASK is determined using CPUID leaf 01H and leaf 04H (Example 8-20).
// This algorithm assumes there is symmetry across core boundary, i.e. each core within a
// package has the same number of logical processors
// SMT_ID always starts from bit 0, corresponding to the right-most bit-field
SMT_ID = APIC_ID & SMT_MASK;
// Extract CORE_ID:
// CORE_MASK is determined in Example 8-19 or Example 8-20
CORE_ID = (APIC_ID & CORE_MASK) ;
// Extract PACKAGE_ID:
// Assume single cluster.
// Shift out the mask width for maximum logical processors per package
// PACKAGE_MASK is determined in Example 8-19 or Example 8-20
PACKAGE_ID = (APIC_ID & PACKAGE_MASK) ;
}
Example 8-22. Compute the Number of Packages, Cores, and Processor Relationships in a MP
System
a) Assemble lists of PACKAGE_ID, CORE_ID, and SMT_ID of each enabled logical processors
//The BIOS and/or OS may limit the number of logical processors available to applications
// after system boot. The below algorithm will compute topology for the processors visible
// to the thread that is computing it.
// Extract the 3-levels of IDs on every processor
// SystemAffinity is a bitmask of all the processors started by the OS. Use OS specific APIs to
// obtain it.
// ThreadAffinityMask is used to affinitize the topology enumeration thread to each processor
Vol. 3 8-61
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
using OS specific APIs.
// Allocate per processor arrays to store the Package_ID, Core_ID and SMT_ID for every started
// processor.
ThreadAffinityMask = 1;
ProcessorNum = 0;
while (ThreadAffinityMask != 0 && ThreadAffinityMask <= SystemAffinity) {
// Check to make sure we can utilize this processor first.
if (ThreadAffinityMask & SystemAffinity){
Set thread to run on the processor specified in ThreadAffinityMask
Wait if necessary and ensure thread is running on specified processor
APIC_ID = GetAPIC_ID(); // 32 bit ID in Example 8-19 or 8-bit ID in Example
8-20
Extract the Package_ID, Core_ID and SMT_ID as explained in three level extraction
algorithm of Example 8-21
PackageID[ProcessorNUM] = PACKAGE_ID;
CoreID[ProcessorNum] = CORE_ID;
SmtID[ProcessorNum] = SMT_ID;
ProcessorNum++;
}
ThreadAffinityMask <<= 1;
}
NumStartedLPs = ProcessorNum;
b) Using the list of PACKAGE_ID to count the number of physical packages in a MP system and
construct, for each package, a multi-bit mask corresponding to those logical processors residing in
the same package.
// Compute the number of packages by counting the number of processors
// with unique PACKAGE_IDs in the PackageID array.
// Compute the mask of processors in each package.
PackageIDBucket is an array of unique PACKAGE_ID values. Allocate an array of
NumStartedLPs count of entries in this array.
PackageProcessorMask is a corresponding array of the bit mask of processors belonging to
the same package, these are processors with the same PACKAGE_ID
The algorithm below assumes there is symmetry across package boundary if more than
one socket is populated in an MP system.
// Bucket Package IDs and compute processor mask for every package.
PackageNum = 1;
PackageIDBucket[0] = PackageID[0];
ProcessorMask = 1;
8-62 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
PackageProcessorMask[0] = ProcessorMask;
For (ProcessorNum = 1; ProcessorNum < NumStartedLPs; ProcessorNum++) {
ProcessorMask << = 1;
For (i=0; i < PackageNum; i++) {
// we may be comparing bit-fields of logical processors residing in different
// packages, the code below assume package symmetry
If (PackageID[ProcessorNum] == PackageIDBucket[i]) {
PackageProcessorMask[i] |= ProcessorMask;
Break; // found in existing bucket, skip to next iteration
}
}
if (i ==PackageNum) {
//PACKAGE_ID did not match any bucket, start new bucket
PackageIDBucket[i] = PackageID[ProcessorNum];
PackageProcessorMask[i] = ProcessorMask;
PackageNum++;
}
}
// PackageNum has the number of Packages started in OS
// PackageProcessorMask[] array has the processor set of each package
c) Using the list of CORE_ID to count the number of cores in a MP system and construct, for each
core, a multi-bit mask corresponding to those logical processors residing in the same core.
Processors in the same core can be determined by bucketing the processors with the same
PACKAGE_ID and CORE_ID. Note that code below can BIT OR the values of PACKGE and CORE ID
because they have not been shifted right.
The algorithm below assumes there is symmetry across package boundary if more than one socket
is populated in an MP system.
//Bucketing PACKAGE and CORE IDs and computing processor mask for every core
CoreNum = 1;
CoreIDBucket[0] = PackageID[0] | CoreID[0];
ProcessorMask = 1;
CoreProcessorMask[0] = ProcessorMask;
For (ProcessorNum = 1; ProcessorNum < NumStartedLPs; ProcessorNum++) {
ProcessorMask << = 1;
For (i=0; i < CoreNum; i++) {
// we may be comparing bit-fields of logical processors residing in different
// packages, the code below assume package symmetry
If ((PackageID[ProcessorNum] | CoreID[ProcessorNum]) == CoreIDBucket[i]) {
CoreProcessorMask[i] |= ProcessorMask;
Break; // found in existing bucket, skip to next iteration
}
Vol. 3 8-63
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
}
if (i == CoreNum) {
//Did not match any bucket, start new bucket
CoreIDBucket[i] = PackageID[ProcessorNum] | CoreID[ProcessorNum];
CoreProcessorMask[i] = ProcessorMask;
CoreNum++;
}
}
// CoreNum has the number of cores started in the OS
// CoreProcessorMask[] array has the processor set of each core
Other processor relationships such as processor mask of sibling cores can be
computed from set operations of the PackageProcessorMask[] and CoreProcessor-
Mask[].
The algorithm shown above can be adapted to work with earlier generations of
single-core IA-32 processors that support Intel Hyper-Threading Technology and in
situations that the deterministic cache parameter leaf is not supported (provided
CPUID supports initial APIC ID). A reference code example is available (see Intel® 64
Architecture Processor Topology Enumeration).
8.10
MANAGEMENT OF IDLE AND BLOCKED CONDITIONS
When a logical processor in an MP system (including multi-core processor or proces-
sors supporting Intel Hyper-Threading Technology) is idle (no work to do) or blocked
(on a lock or semaphore), additional management of the core execution engine
resource can be accomplished by using the HLT (halt), PAUSE, or the
MONITOR/MWAIT instructions.
8.10.1
HLT Instruction
The HLT instruction stops the execution of the logical processor on which it is
executed and places it in a halted state until further notice (see the description of the
Developer’s Manual, Volume 2A). When a logical processor is halted, active logical
processors continue to have full access to the shared resources within the physical
package. Here shared resources that were being used by the halted logical processor
become available to active logical processors, allowing them to execute at greater
efficiency. When the halted logical processor resumes execution, shared resources
are again shared among all active logical processors. (See Section 8.10.6.3, “Halt
Idle Logical Processors,” for more information about using the HLT instruction with
processors supporting Intel Hyper-Threading Technology.)
8-64 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
8.10.2
PAUSE Instruction
The PAUSE instruction can improves the performance of processors supporting Intel
Hyper-Threading Technology when executing “spin-wait loops” and other routines
where one thread is accessing a shared lock or semaphore in a tight polling loop.
penalty when exiting the loop because it detects a possible memory order violation
and flushes the core processor’s pipeline. The PAUSE instruction provides a hint to
the processor that the code sequence is a spin-wait loop. The processor uses this hint
to avoid the memory order violation and prevent the pipeline flush. In addition, the
PAUSE instruction de-pipelines the spin-wait loop to prevent it from consuming
execution resources excessively and consume power needlessly. (See Section
8.10.6.1, “Use the PAUSE Instruction in Spin-Wait Loops,” for more information
about using the PAUSE instruction with IA-32 processors supporting Intel Hyper-
Threading Technology.)
8.10.3
Detecting Support MONITOR/MWAIT Instruction
Streaming SIMD Extensions 3 introduced two instructions (MONITOR and MWAIT) to
help multithreaded software improve thread synchronization. In the initial imple-
mentation, MONITOR and MWAIT are available to software at ring 0. The instructions
are conditionally available at levels greater than 0. Use the following steps to detect
the availability of MONITOR and MWAIT:
• Use CPUID to query the MONITOR bit (CPUID.1.ECX[3] = 1).
• If CPUID indicates support, execute MONITOR inside a TRY/EXCEPT exception
handler and trap for an exception. If an exception occurs, MONITOR and MWAIT
are not supported at a privilege level greater than 0. See Example 8-23.
Example 8-23. Verifying MONITOR/MWAIT Support
boolean MONITOR_MWAIT_works = TRUE;
try {
_asm {
xor ecx, ecx
xor edx, edx
mov eax, MemArea
monitor
}
// Use monitor
} except (UNWIND) {
// if we get here, MONITOR/MWAIT is not supported
MONITOR_MWAIT_works = FALSE;
}
Vol. 3 8-65
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
8.10.4
MONITOR/MWAIT Instruction
Operating systems usually implement idle loops to handle thread synchronization. In
a typical idle-loop scenario, there could be several “busy loops” and they would use a
set of memory locations. An impacted processor waits in a loop and poll a memory
location to determine if there is available work to execute. The posting of work is
typically a write to memory (the work-queue of the waiting processor). The time for
initiating a work request and getting it scheduled is on the order of a few bus cycles.
From a resource sharing perspective (logical processors sharing execution
resources), use of the HLT instruction in an OS idle loop is desirable but has implica-
tions. Executing the HLT instruction on a idle logical processor puts the targeted
processor in a non-execution state. This requires another processor (when posting
work for the halted logical processor) to wake up the halted processor using an inter-
processor interrupt. The posting and servicing of such an interrupt introduces a delay
in the servicing of new work requests.
In a shared memory configuration, exits from busy loops usually occur because of a
state change applicable to a specific memory location; such a change tends to be
triggered by writes to the memory location by another agent (typically a processor).
MONITOR/MWAIT complement the use of HLT and PAUSE to allow for efficient parti-
tioning and un-partitioning of shared resources among logical processors sharing
physical resources. MONITOR sets up an effective address range that is monitored for
write-to-memory activities; MWAIT places the processor in an optimized state (this
may vary between different implementations) until a write to the monitored address
range occurs.
In the initial implementation of MONITOR and MWAIT, they are available at CPL = 0
only.
Both instructions rely on the state of the processor’s monitor hardware. The monitor
hardware can be either armed (by executing the MONITOR instruction) or triggered
(due to a variety of events, including a store to the monitored memory region). If
upon execution of MWAIT, monitor hardware is in a triggered state: MWAIT behaves
as a NOP and execution continues at the next instruction in the execution stream.
The state of monitor hardware is not architecturally visible except through the
behavior of MWAIT.
Multiple events other than a write to the triggering address range can cause a
processor that executed MWAIT to wake up. These include events that would lead to
voluntary or involuntary context switches, such as:
• External interrupts, including NMI, SMI, INIT, BINIT, MCERR, A20M#
• Faults, Aborts (including Machine Check)
• Architectural TLB invalidations including writes to CR0, CR3, CR4 and certain MSR
writes; execution of LMSW (occurring prior to issuing MWAIT but after setting the
monitor)
• Voluntary transitions due to fast system call and far calls (occurring prior to
issuing MWAIT but after setting the monitor)
8-66 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
Power management related events (such as Thermal Monitor 2 or chipset driven
STPCLK# assertion) will not cause the monitor event pending flag to be cleared.
Faults will not cause the monitor event pending flag to be cleared.
Software should not allow for voluntary context switches in between
MONITOR/MWAIT in the instruction flow. Note that execution of MWAIT does not re-
arm the monitor hardware. This means that MONITOR/MWAIT need to be executed in
a loop. Also note that exits from the MWAIT state could be due to a condition other
than a write to the triggering address; software should explicitly check the triggering
data location to determine if the write occurred. Software should also check the value
of the triggering address following the execution of the monitor instruction (and prior
to the execution of the MWAIT instruction). This check is to identify any writes to the
triggering address that occurred during the course of MONITOR execution.
The address range provided to the MONITOR instruction must be of write-back
caching type. Only write-back memory type stores to the monitored address range
will trigger the monitor hardware. If the address range is not in memory of write-
back type, the address monitor hardware may not be set up properly or the monitor
hardware may not be armed. Software is also responsible for ensuring that
• Writes that are not intended to cause the exit of a busy loop do not write to a
location within the address region being monitored by the monitor hardware,
• Writes intended to cause the exit of a busy loop are written to locations within the
monitored address region.
Not doing so will lead to more false wakeups (an exit from the MWAIT state not due
to a write to the intended data location). These have negative performance implica-
tions. It might be necessary for software to use padding to prevent false wakeups.
CPUID provides a mechanism for determining the size data locations for monitoring
as well as a mechanism for determining the size of a the pad.
8.10.5
Monitor/Mwait Address Range Determination
To use the MONITOR/MWAIT instructions, software should know the length of the
region monitored by the MONITOR/MWAIT instructions and the size of the coherence
line size for cache-snoop traffic in a multiprocessor system. This information can be
queried using the CPUID monitor leaf function (EAX = 05H). You will need the
smallest and largest monitor line size:
• To avoid missed wake-ups: make sure that the data structure used to monitor
writes fits within the smallest monitor line-size. Otherwise, the processor may
not wake up after a write intended to trigger an exit from MWAIT.
• To avoid false wake-ups; use the largest monitor line size to pad the data
structure used to monitor writes. Software must make sure that beyond the data
structure, no unrelated data variable exists in the triggering area for MWAIT. A
pad may be needed to avoid this situation.
These above two values bear no relationship to cache line size in the system and soft-
ware should not make any assumptions to that effect. Within a single-cluster system,
Vol. 3 8-67
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
the two parameters should default to be the same (the size of the monitor triggering
area is the same as the system coherence line size).
Based on the monitor line sizes returned by the CPUID, the OS should dynamically
allocate structures with appropriate padding. If static data structures must be used
by an OS, attempt to adapt the data structure and use a dynamically allocated data
buffer for thread synchronization. When the latter technique is not possible, consider
not using MONITOR/MWAIT when using static data structures.
To set up the data structure correctly for MONITOR/MWAIT on multi-clustered
systems: interaction between processors, chipsets, and the BIOS is required (system
coherence line size may depend on the chipset used in the system; the size could be
different from the processor’s monitor triggering area). The BIOS is responsible to
set the correct value for system coherence line size using the
IA32_MONITOR_FILTER_LINE_SIZE MSR. Depending on the relative magnitude of
the size of the monitor triggering area versus the value written into the
IA32_MONITOR_FILTER_LINE_SIZE MSR, the smaller of the parameters will be
reported as the Smallest Monitor Line Size. The larger of the parameters will be
reported as the Largest Monitor Line Size.
8.10.6
Required Operating System Support
This section describes changes that must be made to an operating system to run on
processors supporting Intel Hyper-Threading Technology. It also describes optimiza-
tions that can help an operating system make more efficient use of the logical
processors sharing execution resources. The required changes and suggested opti-
mizations are representative of the types of modifications that appear in Windows*
XP and Linux* kernel 2.4.0 operating systems for Intel processors supporting Intel
Hyper-Threading Technology. Additional optimizations for processors supporting
Intel Hyper-Threading Technology are described in the Intel® 64 and IA-32 Architec-
tures Optimization Reference Manual.
8.10.6.1 Use the PAUSE Instruction in Spin-Wait Loops
Intel recommends that a PAUSE instruction be placed in all spin-wait loops that run
processors.
Software routines that use spin-wait loops include multiprocessor synchronization
primitives (spin-locks, semaphores, and mutex variables) and idle loops. Such
routines keep the processor core busy executing a load-compare-branch loop while a
thread waits for a resource to become available. Including a PAUSE instruction in such
a loop greatly improves efficiency (see Section 8.10.2, “PAUSE Instruction”). The
following routine gives an example of a spin-wait loop that uses a PAUSE instruction:
Spin_Lock:
CMP lockvar, 0
;Check if lock is free
8-68 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
JE Get_Lock
PAUSE
;Short delay
JMP Spin_Lock
Get_Lock:
MOV EAX, 1
XCHG EAX, lockvar ;Try to get lock
CMP EAX, 0
;Test if successful
JNE Spin_Lock
Critical_Section:
<critical section code>
MOV lockvar, 0
...
Continue:
The spin-wait loop above uses a “test, test-and-set” technique for determining the
availability of the synchronization variable. This technique is recommended when
writing spin-wait loops.
instruction is treated as a NOP instruction.
8.10.6.2 Potential Usage of MONITOR/MWAIT in C0 Idle Loops
An operating system may implement different handlers for different idle states. A
typical OS idle loop on an ACPI-compatible OS is shown in Example 8-24:
Example 8-24. A Typical OS Idle Loop
// WorkQueue is a memory location indicating there is a thread
// ready to run. A non-zero value for WorkQueue is assumed to
// indicate the presence of work to be scheduled on the processor.
// The idle loop is entered with interrupts disabled.
WHILE (1) {
IF (WorkQueue) THEN {
// Schedule work at WorkQueue.
}
ELSE {
// No work to do - wait in appropriate C-state handler depending
// on Idle time accumulated
IF (IdleTime >= IdleTimeThreshhold) THEN {
// Call appropriate C1, C2, C3 state handler, C1 handler
// shown below
}
}
}
Vol. 3 8-69
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
// C1 handler uses a Halt instruction
VOID C1Handler()
{
STI
HLT
}
The MONITOR and MWAIT instructions may be considered for use in the C0 idle state loops, if
MONITOR and MWAIT are supported.
Example 8-25. An OS Idle Loop with MONITOR/MWAIT in the C0 Idle Loop
// WorkQueue is a memory location indicating there is a thread
// ready to run. A non-zero value for WorkQueue is assumed to
// indicate the presence of work to be scheduled on the processor.
// The following example assumes that the necessary padding has been
// added surrounding WorkQueue to eliminate false wakeups
// The idle loop is entered with interrupts disabled.
WHILE (1) {
IF (WorkQueue) THEN {
// Schedule work at WorkQueue.
}
ELSE {
// No work to do - wait in appropriate C-state handler depending
// on Idle time accumulated.
IF (IdleTime >= IdleTimeThreshhold) THEN {
// Call appropriate C1, C2, C3 state handler, C1
// handler shown below
MONITOR WorkQueue // Setup of eax with WorkQueue
// LinearAddress,
// ECX, EDX = 0
IF (WorkQueue != 0) THEN {
MWAIT
}
}
}
}
// C1 handler uses a Halt instruction.
VOID C1Handler()
{
STI
HLT
8-70 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
}
8.10.6.3 Halt Idle Logical Processors
If one of two logical processors is idle or in a spin-wait loop of long duration, explicitly
halt that processor by means of a HLT instruction.
In an MP system, operating systems can place idle processors into a loop that contin-
uously checks the run queue for runnable software tasks. Logical processors that
execute idle loops consume a significant amount of core’s execution resources that
might otherwise be used by the other logical processors in the physical package. For
8
this reason, halting idle logical processors optimizes the performance. If all logical
saving state.
8.10.6.4 Potential Usage of MONITOR/MWAIT in C1 Idle Loops
An operating system may also consider replacing HLT with MONITOR/MWAIT in its C1
idle loop. An example is shown in Example 8-26:
Example 8-26. An OS Idle Loop with MONITOR/MWAIT in the C1 Idle Loop
// WorkQueue is a memory location indicating there is a thread
// ready to run. A non-zero value for WorkQueue is assumed to
// indicate the presence of work to be scheduled on the processor.
// The following example assumes that the necessary padding has been
// added surrounding WorkQueue to eliminate false wakeups
// The idle loop is entered with interrupts disabled.
WHILE (1) {
IF (WorkQueue) THEN {
// Schedule work at WorkQueue
}
ELSE {
// No work to do - wait in appropriate C-state handler depending
// on Idle time accumulated
IF (IdleTime >= IdleTimeThreshhold) THEN {
// Call appropriate C1, C2, C3 state handler, C1
// handler shown below
}
}
}
// C1 handler uses a Halt instruction
VOID C1Handler()
8. Excessive transitions into and out of the HALT state could also incur performance penalties.
Operating systems should evaluate the performance trade-offs for their operating system.
Vol. 3 8-71
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
{
MONITOR WorkQueue // Setup of eax with WorkQueue LinearAddress,
// ECX, EDX = 0
IF (WorkQueue != 0) THEN {
STI
MWAIT
// EAX, ECX = 0
}
}
8.10.6.5 Guidelines for Scheduling Threads on Logical Processors Sharing
Execution Resources
Because the logical processors, the order in which threads are dispatched to logical
processors for execution can affect the overall efficiency of a system. The following
guidelines are recommended for scheduling threads for execution.
• Dispatch threads to one logical processor per processor core before dispatching
threads to the other logical processor sharing execution resources in the same
processor core.
• In an MP system with two or more physical packages, distribute threads out over
all the physical processors, rather than concentrate them in one or two physical
processors.
• Use processor affinity to assign a thread to a specific processor core or package,
depending on the cache-sharing topology. The practice increases the chance that
the processor’s caches will contain some of the thread’s code and data when it is
dispatched for execution after being suspended.
8.10.6.6 Eliminate Execution-Based Timing Loops
Intel discourages the use of timing loops that depend on a processor’s execution
speed to measure time. There are several reasons:
• Timing loops cause problems when they are calibrated on a IA-32 processor
running at one clock speed and then executed on a processor running at another
clock speed.
• Routines for calibrating execution-based timing loops produce unpredictable
results when run on an IA-32 processor supporting Intel Hyper-Threading
Technology. This is due to the sharing of execution resources between the logical
processors within a physical package.
To avoid the problems described, timing loop routines must use a timing mechanism
for the loop that does not depend on the execution speed of the logical processors in
the system. The following sources are generally available:
• A high resolution system timer (for example, an Intel 8254).
8-72 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
• A high resolution timer within the processor (such as, the local APIC timer or the
time-stamp counter).
For additional information, see the Intel® 64 and IA-32 Architectures Optimization
Reference Manual.
8.10.6.7 Place Locks and Semaphores in Aligned, 128-Byte Blocks of
Memory
When software uses locks or semaphores to synchronize processes, threads, or other
code sections; Intel recommends that only one lock or semaphore be present within
a cache line (or 128 byte sector, if 128-byte sector is supported). In processors based
on Intel NetBurst microarchitecture (which support 128-byte sector consisting of two
cache lines), following this recommendation means that each lock or semaphore
should be contained in a 128-byte block of memory that begins on a 128-byte
boundary. The practice minimizes the bus traffic required to service locks.
Vol. 3 8-73
Download from Www.Somanuals.com. All Manuals Search And Download.
MULTIPLE-PROCESSOR MANAGEMENT
8-74 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
CHAPTER 9
PROCESSOR MANAGEMENT AND INITIALIZATION
This chapter describes the facilities provided for managing processor wide functions
and for initializing the processor. The subjects covered include: processor initializa-
tion, x87 FPU initialization, processor configuration, feature determination, mode
switching, the MSRs (in the Pentium, P6 family, Pentium 4, and Intel Xeon proces-
sors), and the MTRRs (in the P6 family, Pentium 4, and Intel Xeon processors).
9.1
INITIALIZATION OVERVIEW
Following power-up or an assertion of the RESET# pin, each processor on the system
bus performs a hardware initialization of the processor (known as a hardware reset)
and an optional built-in self-test (BIST). A hardware reset sets each processor’s
registers to a known state and places the processor in real-address mode. It also
invalidates the internal caches, translation lookaside buffers (TLBs) and the branch
target buffer (BTB). At this point, the action taken depends on the processor family:
(including a single processor in a uniprocessor system) execute the multiple
processor (MP) initialization protocol. The processor that is selected through this
protocol as the bootstrap processor (BSP) then immediately starts executing
software-initialization code in the current code segment beginning at the offset in
the EIP register. The application (non-BSP) processors (APs) go into a Wait For
Startup IPI (SIPI) state while the BSP is executing initialization code. See Section
8.4, “Multiple-Processor (MP) Initialization,” for more details. Note that in a
uniprocessor system, the single Pentium 4 or Intel Xeon processor automatically
becomes the BSP.
• P6 family processors — The action taken is the same as for the Pentium 4 and
Intel Xeon processors (as described in the previous paragraph).
• Pentium processors — In either a single- or dual- processor system, a single
Pentium processor is always pre-designated as the primary processor. Following
a reset, the primary processor behaves as follows in both single- and dual-
processor systems. Using the dual-processor (DP) ready initialization protocol,
the primary processor immediately starts executing software-initialization code
in the current code segment beginning at the offset in the EIP register. The
secondary processor (if there is one) goes into a halt state.
• Intel486 processor — The primary processor (or single processor in a unipro-
cessor system) immediately starts executing software-initialization code in the
current code segment beginning at the offset in the EIP register. (The Intel486
does not automatically execute a DP or MP initialization protocol to determine
which processor is the primary processor.)
Vol. 3 9-1
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
The software-initialization code performs all system-specific initialization of the BSP
or primary processor and the system logic.
At this point, for MP (or DP) systems, the BSP (or primary) processor wakes up each
AP (or secondary) processor to enable those processors to execute self-configuration
code.
When all processors are initialized, configured, and synchronized, the BSP or primary
processor begins executing an initial operating-system or executive task.
The x87 FPU is also initialized to a known state during hardware reset. x87 FPU soft-
ware initialization code can then be executed to perform operations such as setting
the precision of the x87 FPU and the exception masks. No special initialization of the
x87 FPU is required to switch operating modes.
Asserting the INIT# pin on the processor invokes a similar response to a hardware
reset. The major difference is that during an INIT, the internal caches, MSRs, MTRRs,
with a hardware reset). An INIT provides a method for switching from protected to
9.1.1
Processor State After Reset
Table 9-1 shows the state of the flags and other registers following power-up for the
Pentium 4, Intel Xeon, P6 family, and Pentium processors. The state of control
register CR0 is 60000010H (see Figure 9-1). This places the processor is in real-
address mode with paging disabled.
9.1.2
Processor Built-In Self-Test (BIST)
Hardware may request that the BIST be performed at power-up. The EAX register is
cleared (0H) if the processor passes the BIST. A nonzero value in the EAX register
after the BIST indicates that a processor fault was detected. If the BIST is not
requested, the contents of the EAX register after a hardware reset is 0H.
The overhead for performing a BIST varies between processor families. For example,
the BIST takes approximately 30 million processor clock periods to execute on the
Pentium 4 processor. This clock count is model-specific; Intel reserves the right to
change the number of periods for any Intel 64 or IA-32 processor, without notification.
Table 9-1. IA-32 Processor States Following Power-up, Reset, or INIT
Register
Pentium 4 and Intel
Xeon Processor
P6 Family Processor
Pentium Processor
EFLAGS1
00000002H
0000FFF0H
60000010H2
00000002H
0000FFF0H
60000010H2
00000002H
0000FFF0H
60000010H2
EIP
CR0
9-2 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
Table 9-1. IA-32 Processor States Following Power-up, Reset, or INIT (Contd.)
Register
Pentium 4 and Intel
Xeon Processor
P6 Family Processor
Pentium Processor
CR2, CR3, CR4
CS
00000000H
00000000H
00000000H
Selector = F000H
Base = FFFF0000H
Limit = FFFFH
Selector = F000H
Base = FFFF0000H
Limit = FFFFH
Selector = F000H
Base = FFFF0000H
Limit = FFFFH
AR = Present, R/W,
Accessed
AR = Present, R/W,
Accessed
AR = Present, R/W,
Accessed
SS, DS, ES, FS, GS
Selector = 0000H
Base = 00000000H
Limit = FFFFH
AR = Present, R/W,
Accessed
Selector = 0000H
Base = 00000000H
Limit = FFFFH
AR = Present, R/W,
Accessed
Selector = 0000H
Base = 00000000H
Limit = FFFFH
AR = Present, R/W,
Accessed
EDX
EAX
00000FxxH
04
000n06xxH3
04
000005xxH
04
EBX, ECX, ESI, EDI, 00000000H
EBP, ESP
00000000H
00000000H
ST0 through ST75 Pwr up or Reset: +0.0
FINIT/FNINIT: Unchanged
Pwr up or Reset: +0.0
FINIT/FNINIT: Unchanged
Pwr up or Reset: +0.0
FINIT/FNINIT: Unchanged
x87 FPU Control
Word5
Pwr up or Reset: 0040H
FINIT/FNINIT: 037FH
Pwr up or Reset: 0040H
FINIT/FNINIT: 037FH
Pwr up or Reset: 0040H
FINIT/FNINIT: 037FH
x87 FPU Status
Word5
Pwr up or Reset: 0000H
FINIT/FNINIT: 0000H
Pwr up or Reset: 0000H
FINIT/FNINIT: 0000H
Pwr up or Reset: 0000H
FINIT/FNINIT: 0000H
x87 FPU Tag
Word5
Pwr up or Reset: 5555H
FINIT/FNINIT: FFFFH
Pwr up or Reset: 5555H
FINIT/FNINIT: FFFFH
Pwr up or Reset: 5555H
FINIT/FNINIT: FFFFH
x87 FPU Data
Operand and CS
Seg. Selectors5
Pwr up or Reset: 0000H
FINIT/FNINIT: 0000H
Pwr up or Reset: 0000H
FINIT/FNINIT: 0000H
Pwr up or Reset: 0000H
FINIT/FNINIT: 0000H
x87 FPU Data
Operand and Inst.
Pointers5
Pwr up or Reset:
Pwr up or Reset:
Pwr up or Reset:
00000000H
00000000H
00000000H
FINIT/FNINIT: 00000000H
FINIT/FNINIT: 00000000H
FINIT/FNINIT: 00000000H
MM0 through
MM75
Pwr up or Reset:
0000000000000000H
INIT or FINIT/FNINIT:
Unchanged
Pentium II and Pentium III
Processors Only—
Pwr up or Reset:
Pentium with MMX
Technology Only—
Pwr up or Reset:
0000000000000000H
INIT or FINIT/FNINIT:
Unchanged
0000000000000000H
INIT or FINIT/FNINIT:
Unchanged
XMM0 through
XMM7
Pwr up or Reset:
Pentium III processor Only— NA
Pwr up or Reset:
0000000000000000H
INIT: Unchanged
0000000000000000H
INIT: Unchanged
MXCSR
Pwr up or Reset: 1F80H
INIT: Unchanged
Pentium III processor only-
Pwr up or Reset: 1F80H
INIT: Unchanged
NA
GDTR, IDTR
Base = 00000000H
Limit = FFFFH
Base = 00000000H
Limit = FFFFH
Base = 00000000H
Limit = FFFFH
AR = Present, R/W
AR = Present, R/W
AR = Present, R/W
Vol. 3 9-3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
Table 9-1. IA-32 Processor States Following Power-up, Reset, or INIT (Contd.)
Register
Pentium 4 and Intel
Xeon Processor
P6 Family Processor
Pentium Processor
LDTR, Task
Register
Selector = 0000H
Base = 00000000H
Limit = FFFFH
Selector = 0000H
Base = 00000000H
Limit = FFFFH
Selector = 0000H
Base = 00000000H
Limit = FFFFH
AR = Present, R/W
AR = Present, R/W
AR = Present, R/W
DR0, DR1, DR2,
DR3
00000000H
00000000H
00000000H
DR6
DR7
FFFF0FF0H
00000400H
FFFF0FF0H
00000400H
FFFF0FF0H
00000400H
Time-Stamp
Counter
Power up or Reset: 0H
INIT: Unchanged
Power up or Reset: 0H
INIT: Unchanged
Power up or Reset: 0H
INIT: Unchanged
Perf. Counters and Power up or Reset: 0H
Power up or Reset: 0H
INIT: Unchanged
Power up or Reset: 0H
INIT: Unchanged
Event Select
INIT: Unchanged
All Other MSRs
Pwr up or Reset:
Undefined
Pwr up or Reset:
Undefined
Pwr up or Reset:
Undefined
INIT: Unchanged
INIT: Unchanged
INIT: Unchanged
Data and Code
Cache, TLBs
Invalid
Invalid
Invalid
Fixed MTRRs
Pwr up or Reset: Disabled
INIT: Unchanged
Pwr up or Reset: Disabled
INIT: Unchanged
Not Implemented
Not Implemented
Not Implemented
Variable MTRRs
Pwr up or Reset: Disabled
INIT: Unchanged
Pwr up or Reset: Disabled
INIT: Unchanged
Machine-Check
Architecture
Pwr up or Reset:
Undefined
INIT: Unchanged
Pwr up or Reset:
Undefined
INIT: Unchanged
APIC
Pwr up or Reset: Enabled
INIT: Unchanged
Pwr up or Reset: Enabled
INIT: Unchanged
Pwr up or Reset: Enabled
INIT: Unchanged
NOTES:
1. The 10 most-significant bits of the EFLAGS register are undefined following a reset. Software
should not depend on the states of any of these bits.
2. The CD and NW flags are unchanged, bit 4 is set to 1, all other bits are cleared.
3. Where “n” is the Extended Model Value for the respective processor.
4. If Built-In Self-Test (BIST) is invoked on power up or reset, EAX is 0 only if all tests passed. (BIST
cannot be invoked during an INIT.)
5. The state of the x87 FPU and MMX registers is not changed by the execution of an INIT.
9-4 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
Paging disabled: 0
Caching disabled: 1
Not write-through disabled: 1
Alignment check disabled: 0
Write-protect disabled: 0
31
19 18 17 16
15
30 29 28
6
5
4
3
2
1
0
P C N
G D W
A
M
W
P
N
E
T E M P
S M P E
Reserved
Reserved
1
External x87 FPU error reporting: 0
(Not used): 1
No task switch: 0
x87 FPU instructions not trapped: 0
WAIT/FWAIT instructions not trapped: 0
Real-address mode: 0
Figure 9-1. Contents of CR0 Register after Reset
9.1.3
Model and Stepping Information
Following a hardware reset, the EDX register contains component identification and
revision information (see Figure 9-2). For example, the model, family, and processor
type returned for the first processor in the Intel Pentium 4 family is as follows: model
(0000B), family (1111B), and processor type (00B).
31
24 23
20 19
1615 14 13 12 11
8 7
4 3
0
Extended
Family
Stepping
ID
Extended
Model
EDX
Model
Family
Processor Type
Family (1111B for the Pentium 4 Processor Family)
Model (Beginning with 0000B)
Reserved
Figure 9-2. Version Information in the EDX Register after Reset
The stepping ID field contains a unique identifier for the processor’s stepping ID or
revision level. The extended family and extended model fields were added to the
IA-32 architecture in the Pentium 4 processors.
Vol. 3 9-5
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
9.1.4
First Instruction Executed
The first instruction that is fetched and executed following a hardware reset is
located at physical address FFFFFFF0H. This address is 16 bytes below the
processor’s uppermost physical address. The EPROM containing the software-
initialization code must be located at this address.
The address FFFFFFF0H is beyond the 1-MByte addressable range of the processor
while in real-address mode. The processor is initialized to this starting address as
follows. The CS register has two parts: the visible segment selector part and the
hidden base address part. In real-address mode, the base address is normally
formed by shifting the 16-bit segment selector value 4 bits to the left to produce a
20-bit base address. However, during a hardware reset, the segment selector in the
CS register is loaded with F000H and the base address is loaded with FFFF0000H. The
starting address is thus formed by adding the base address to the value in the EIP
register (that is, FFFF0000 + FFF0H = FFFFFFF0H).
The first time the CS register is loaded with a new value after a hardware reset, the
processor will follow the normal rule for address translation in real-address mode
(that is, [CS base address = CS segment selector * 16]). To insure that the base
address in the CS register remains unchanged until the EPROM based software-
initialization code is completed, the code must not contain a far jump or far call or
allow an interrupt to occur (which would cause the CS selector value to be changed).
9.2
X87 FPU INITIALIZATION
Software-initialization code can determine the whether the processor contains an
x87 FPU by using the CPUID instruction. The code must then initialize the x87 FPU
and set flags in control register CR0 to reflect the state of the x87 FPU environment.
A hardware reset places the x87 FPU in the state shown in Table 9-1. This state is
different from the state the x87 FPU is placed in following the execution of an FINIT
or FNINIT instruction (also shown in Table 9-1). If the x87 FPU is to be used, the soft-
ware-initialization code should execute an FINIT/FNINIT instruction following a hard-
ware reset. These instructions, tag all data registers as empty, clear all the exception
masks, set the TOP-of-stack value to 0, and select the default rounding and precision
controls setting (round to nearest and 64-bit precision).
If the processor is reset by asserting the INIT# pin, the x87 FPU state is not changed.
9.2.1
Configuring the x87 FPU Environment
Initialization code must load the appropriate values into the MP, EM, and NE flags of
control register CR0. These bits are cleared on hardware reset of the processor.
Figure 9-2 shows the suggested settings for these flags, depending on the IA-32
processor being initialized. Initialization code can test for the type of processor
present before setting or clearing these flags.
9-6 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
Table 9-2. Recommended Settings of EM and MP Flags on IA-32 Processors
EM
MP
NE
IA-32 processor
1
0
1
Intel486™ SX, Intel386™ DX, and Intel386™ SX processors
only, without the presence of a math coprocessor.
*
*
0
1
1
1 or 0
1 or 0
Pentium 4, Intel Xeon, P6 family, Pentium, Intel486™ DX, and
Intel 487 SX processors, and Intel386 DX and Intel386 SX
processors when a companion math coprocessor is present.
0
More recent Intel 64 or IA-32 processors
NOTE:
* The setting of the NE flag depends on the operating system being used.
The EM flag determines whether floating-point instructions are executed by the x87
FPU (EM is cleared) or a device-not-available exception (#NM) is generated for all
floating-point instructions so that an exception handler can emulate the floating-
point operation (EM = 1). Ordinarily, the EM flag is cleared when an x87 FPU or math
coprocessor is present and set if they are not present. If the EM flag is set and no x87
FPU, math coprocessor, or floating-point emulator is present, the processor will hang
when a floating-point instruction is executed.
The MP flag determines whether WAIT/FWAIT instructions react to the setting of the
TS flag. If the MP flag is clear, WAIT/FWAIT instructions ignore the setting of the TS
if the TS flag is set. Generally, the MP flag should be set for processors with an inte-
grated x87 FPU and clear for processors without an integrated x87 FPU and without a
math coprocessor present. However, an operating system can choose to save the
floating-point context at every context switch, in which case there would be no need
to set the MP bit.
Table 2-1 shows the actions taken for floating-point and WAIT/FWAIT instructions
based on the settings of the EM, MP, and TS flags.
The NE flag determines whether unmasked floating-point exceptions are handled by
generating a floating-point error exception internally (NE is set, native mode) or
through an external interrupt (NE is cleared). In systems where an external interrupt
controller is used to invoke numeric exception handlers (such as MS-DOS-based
systems), the NE bit should be cleared.
9.2.2
Setting the Processor for x87 FPU Software Emulation
Setting the EM flag causes the processor to generate a device-not-available excep-
tion (#NM) and trap to a software exception handler whenever it encounters a
floating-point instruction. (Table 9-2 shows when it is appropriate to use this flag.)
Setting this flag has two functions:
Vol. 3 9-7
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
• It allows x87 FPU code to run on an IA-32 processor that has neither an
floating-point emulator.
• It allows floating-point code to be executed using a special or nonstandard
floating-point emulator, selected for a particular application, regardless of
whether an x87 FPU or math coprocessor is present.
To emulate floating-point instructions, the EM, MP, and NE flag in control register CR0
should be set as shown in Table 9-3.
Table 9-3. Software Emulation Settings of EM, MP, and NE Flags
CR0 Bit
EM
Value
1
0
1
MP
NE
Regardless of the value of the EM bit, the Intel486 SX processor generates a device-
not-available exception (#NM) upon encountering any floating-point instruction.
9.3
CACHE ENABLING
IA-32 processors (beginning with the Intel486 processor) and Intel 64 processors
contain internal instruction and data caches. These caches are enabled by clearing
the CD and NW flags in control register CR0. (They are set during a hardware reset.)
Because all internal cache lines are invalid following reset initialization, it is not
necessary to invalidate the cache before enabling caching. Any external caches may
require initialization and invalidation using a system-specific initialization and invali-
dation code sequence.
Depending on the hardware and operating system or executive requirements, addi-
tional configuration of the processor’s caching facilities will probably be required.
Beginning with the Intel486 processor, page-level caching can be controlled with the
PCD and PWT flags in page-directory and page-table entries. Beginning with the P6
family processors, the memory type range registers (MTRRs) control the caching
characteristics of the regions of physical memory. (For the Intel486 and Pentium
processors, external hardware can be used to control the caching characteristics of
regions of physical memory.) See Chapter 11, “Memory Cache Control,” for detailed
information on configuration of the caching facilities in the Pentium 4, Intel Xeon, and
P6 family processors and system memory.
9-8 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
9.4
MODEL-SPECIFIC REGISTERS (MSRS)
Most IA-32 processors (starting from Pentium processors) and Intel 64 processors
contain a model-specific registers (MSRs). A given MSR may not be supported across
all families and models for Intel 64 and IA-32 processors. Some MSRs are designated
as architectural to simplify software programming; a feature introduced by an archi-
tectural MSR is expected to be supported in future processors. Non-architectural
MSRs are not guaranteed to be supported or to have the same functions on future
processors.
MSRs that provide control for a number of hardware and software-related features,
include:
• Performance-monitoring counters (see Chapter 20, “Introduction to Virtual-
Machine Extensions”).
• Debug extensions (see Chapter 20, “Introduction to Virtual-Machine Exten-
sions.”).
• Machine-check exception capability and its accompanying machine-check archi-
tecture (see Chapter 15, “Machine-Check Architecture”).
• MTRRs (see Section 11.11, “Memory Type Range Registers (MTRRs)”).
• Thermal and power management.
• Instruction-specific support (for example: SYSENTER, SYSEXIT, SWAPGS, etc.).
• Processor feature/mode support (for example: IA32_EFER,
IA32_FEATURE_CONTROL).
The MSRs can be read and written to using the RDMSR and WRMSR instructions,
respectively.
When performing software initialization of an IA-32 or Intel 64 processor, many of
the MSRs will need to be initialized to set up things like performance-monitoring
events, run-time machine checks, and memory types for physical memory.
Lists of available performance-monitoring events are given in Appendix A, “Perfor-
mance Monitoring Events”, and lists of available MSRs are given in Appendix B,
“Model-Specific Registers (MSRs)” The references earlier in this section show where
the functions of the various groups of MSRs are described in this manual.
9.5
MEMORY TYPE RANGE REGISTERS (MTRRS)
Memory type range registers (MTRRs) were introduced into the IA-32 architecture
with the Pentium Pro processor. They allow the type of caching (or no caching) to be
specified in system memory for selected physical address ranges. They allow
memory accesses to be optimized for various types of memory such as RAM, ROM,
frame buffer memory, and memory-mapped I/O devices.
In general, initializing the MTRRs is normally handled by the software initialization
code or BIOS and is not an operating system or executive function. At the very least,
Vol. 3 9-9
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
all the MTRRs must be cleared to 0, which selects the uncached (UC) memory type.
See Section 11.11, “Memory Type Range Registers (MTRRs),” for detailed informa-
tion on the MTRRs.
9.6
INITIALIZING SSE/SSE2/SSE3/SSSE3 EXTENSIONS
For processors that contain SSE/SSE2/SSE3/SSSE3 extensions, steps must be taken
when initializing the processor to allow execution of these instructions.
1. Check the CPUID feature flags for the presence of the SSE/SSE2/SSE3/SSSE3
extensions (respectively: EDX bits 25 and 26, ECX bit 0 and 9) and support for
the FXSAVE and FXRSTOR instructions (EDX bit 24). Also check for support for
EDX and ECX registers when the CPUID instruction is executed with a 1 in the
EAX register.
2. Set the OSFXSR flag (bit 9 in control register CR4) to indicate that the operating
system supports saving and restoring the SSE/SSE2/SSE3/SSSE3 execution
environment (XXM and MXCSR registers) with the FXSAVE and FXRSTOR instruc-
tions, respectively. See Section 2.5, “Control Registers,” for a description of the
OSFXSR flag.
3. Set the OSXMMEXCPT flag (bit 10 in control register CR4) to indicate that the
operating system supports the handling of SSE/SSE2/SSE3 SIMD floating-point
exceptions (#XF). See Section 2.5, “Control Registers,” for a description of the
OSXMMEXCPT flag.
4. Set the mask bits and flags in the MXCSR register according to the mode of
operation desired for SSE/SSE2/SSE3 SIMD floating-point instructions. See
“MXCSR Control and Status Register” in Chapter 10, “Programming with
Streaming SIMD Extensions (SSE),” of the Intel® 64 and IA-32 Architectures
Software Developer’s Manual, Volume 1, for a detailed description of the bits and
flags in the MXCSR register.
9.7
SOFTWARE INITIALIZATION FOR REAL-ADDRESS
MODE OPERATION
Following a hardware reset (either through a power-up or the assertion of the
RESET# pin) the processor is placed in real-address mode and begins executing soft-
ware initialization code from physical address FFFFFFF0H. Software initialization code
must first set up the necessary data structures for handling basic system functions,
such as a real-mode IDT for handling interrupts and exceptions. If the processor is to
remain in real-address mode, software must then load additional operating-system
or executive code modules and data structures to allow reliable execution of applica-
tion programs in real-address mode.
If the processor is going to operate in protected mode, software must load the neces-
sary data structures to operate in protected mode and then switch to protected
9-10 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
mode. The protected-mode data structures that must be loaded are described in
Section 9.8, “Software Initialization for Protected-Mode Operation.”
9.7.1
Real-Address Mode IDT
In real-address mode, the only system data structure that must be loaded into
memory is the IDT (also called the “interrupt vector table”). By default, the address
of the base of the IDT is physical address 0H. This address can be changed by using
the LIDT instruction to change the base address value in the IDTR. Software initial-
ization code needs to load interrupt- and exception-handler pointers into the IDT
before interrupts can be enabled.
The actual interrupt- and exception-handler code can be contained either in EPROM
or RAM; however, the code must be located within the 1-MByte addressable range of
the processor in real-address mode. If the handler code is to be stored in RAM, it
must be loaded along with the IDT.
9.7.2
NMI Interrupt Handling
The NMI interrupt is always enabled (except when multiple NMIs are nested). If the
IDT and the NMI interrupt handler need to be loaded into RAM, there will be a period
of time following hardware reset when an NMI interrupt cannot be handled. During
this time, hardware must provide a mechanism to prevent an NMI interrupt from
halting code execution until the IDT and the necessary NMI handler software is
loaded. Here are two examples of how NMIs can be handled during the initial states
of processor initialization:
• A simple IDT and NMI interrupt handler can be provided in EPROM. This allows an
NMI interrupt to be handled immediately after reset initialization.
• The system hardware can provide a mechanism to enable and disable NMIs by
passing the NMI# signal through an AND gate controlled by a flag in an I/O port.
Hardware can clear the flag when the processor is reset, and software can set the
flag when it is ready to handle NMI interrupts.
9.8
SOFTWARE INITIALIZATION FOR PROTECTED-MODE
OPERATION
The processor is placed in real-address mode following a hardware reset. At this
point in the initialization process, some basic data structures and code modules must
be loaded into physical memory to support further initialization of the processor, as
described in Section 9.7, “Software Initialization for Real-Address Mode Operation.”
Before the processor can be switched to protected mode, the software initialization
code must load a minimum number of protected mode data structures and code
Vol. 3 9-11
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
modules into memory to support reliable operation of the processor in protected
mode. These data structures include the following:
• A IDT.
• A GDT.
• A TSS.
• (Optional) An LDT.
• If paging is to be used, at least one page directory and one page table.
• A code segment that contains the code to be executed when the processor
switches to protected mode.
• One or more code modules that contain the necessary interrupt and exception
handlers.
Software initialization code must also initialize the following system registers before
the processor can be switched to protected mode:
• The GDTR.
• (Optional.) The IDTR. This register can also be initialized immediately after
switching to protected mode, prior to enabling interrupts.
• Control registers CR1 through CR4.
• (Pentium 4, Intel Xeon, and P6 family processors only.) The memory type range
registers (MTRRs).
With these data structures, code modules, and system registers initialized, the
processor can be switched to protected mode by loading control register CR0 with a
value that sets the PE flag (bit 0).
9.8.1
Protected-Mode System Data Structures
The contents of the protected-mode system data structures loaded into memory
during software initialization, depend largely on the type of memory management
the protected-mode operating-system or executive is going to support: flat, flat with
paging, segmented, or segmented with paging.
at a minimum load a GDT with one code and one data-segment descriptor. A null
descriptor in the first GDT entry is also required. The stack can be placed in a normal
read/write data segment, so no dedicated descriptor for the stack is required. A flat
memory model with paging also requires a page directory and at least one page table
(unless all pages are 4 MBytes in which case only a page directory is required). See
Section 9.8.3, “Initializing Paging.”
Before the GDT can be used, the base address and limit for the GDT must be loaded
into the GDTR register using an LGDT instruction.
A multi-segmented model may require additional segments for the operating system,
as well as segments and LDTs for each application program. LDTs require segment
9-12 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
descriptors in the GDT. Some operating systems allocate new segments and LDTs as
they are needed. This provides maximum flexibility for handling a dynamic program-
ming environment. However, many operating systems use a single LDT for all tasks,
allocating GDT entries in advance. An embedded system, such as a process
controller, might pre-allocate a fixed number of segments and LDTs for a fixed
number of application programs. This would be a simple and efficient way to struc-
ture the software environment of a real-time system.
9.8.2
Initializing Protected-Mode Exceptions and Interrupts
Software initialization code must at a minimum load a protected-mode IDT with gate
descriptor for each exception vector that the processor can generate. If interrupt or
trap gates are used, the gate descriptors can all point to the same code segment,
which contains the necessary exception handlers. If task gates are used, one TSS
and accompanying code, data, and task segments are required for each exception
handler called with a task gate.
If hardware allows interrupts to be generated, gate descriptors must be provided in
the IDT for one or more interrupt handlers.
Before the IDT can be used, the base address and limit for the IDT must be loaded
into the IDTR register using an LIDT instruction. This operation is typically carried out
immediately after switching to protected mode.
9.8.3
Initializing Paging
Paging is controlled by the PG flag in control register CR0. When this flag is clear (its
state following a hardware reset), the paging mechanism is turned off; when it is set,
paging is enabled. Before setting the PG flag, the following data structures and regis-
ters must be initialized:
• Software must load at least one page directory and one page table into physical
memory. The page table can be eliminated if the page directory contains a
directory entry pointing to itself (here, the page directory and page table reside
in the same page), or if only 4-MByte pages are used.
• Control register CR3 (also called the PDBR register) is loaded with the physical
base address of the page directory.
• (Optional) Software may provide one set of code and data descriptors in the GDT
or in an LDT for supervisor mode and another set for user mode.
With this paging initialization complete, paging is enabled and the processor is
switched to protected mode at the same time by loading control register CR0 with an
image in which the PG and PE flags are set. (Paging cannot be enabled before the
processor is switched to protected mode.)
Vol. 3 9-13
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
9.8.4
Initializing Multitasking
If the multitasking mechanism is not going to be used and changes between privilege
levels are not allowed, it is not necessary load a TSS into memory or to initialize the
task register.
If the multitasking mechanism is going to be used and/or changes between privilege
levels are allowed, software initialization code must load at least one TSS and an
accompanying TSS descriptor. (A TSS is required to change privilege levels because
pointers to the privileged-level 0, 1, and 2 stack segments and the stack pointers for
these stacks are obtained from the TSS.) TSS descriptors must not be marked as
busy when they are created; they should be marked busy by the processor only as a
side-effect of performing a task switch. As with descriptors for LDTs, TSS descriptors
reside in the GDT.
After the processor has switched to protected mode, the LTR instruction can be used
to load a segment selector for a TSS descriptor into the task register. This instruction
marks the TSS descriptor as busy, but does not perform a task switch. The processor
can, however, use the TSS to locate pointers to privilege-level 0, 1, and 2 stacks. The
segment selector for the TSS must be loaded before software performs its first task
switch in protected mode, because a task switch copies the current task state into
the TSS.
After the LTR instruction has been executed, further operations on the task register
are performed by task switching. As with other segments and LDTs, TSSs and TSS
descriptors can be either pre-allocated or allocated as needed.
9.8.5
Initializing IA-32e Mode
On Intel 64 processors, the IA32_EFER MSR is cleared on system reset. The oper-
ating system must be in protected mode with paging enabled before attempting to
initialize IA-32e mode. IA-32e mode operation also requires physical-address exten-
sions with four levels of enhanced paging structures (see Section 4.5, “IA-32e
Paging”).
Operating systems should follow this sequence to initialize IA-32e mode:
1. Starting from protected mode, disable paging by setting CR0.PG = 0. Use the
MOV CR0 instruction to disable paging (the instruction must be located in an
identity-mapped page).
2. Enable physical-address extensions (PAE) by setting CR4.PAE = 1. Failure to
enable PAE will result in a #GP fault when an attempt is made to initialize IA-32e
mode.
3. Load CR3 with the physical base address of the Level 4 page map table (PML4).
4. Enable IA-32e mode by setting IA32_EFER.LME = 1.
5. Enable paging by setting CR0.PG = 1. This causes the processor to set the
IA32_EFER.LMA bit to 1. The MOV CR0 instruction that enables paging and the
9-14 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
following instructions must be located in an identity-mapped page (until such
time that a branch to non-identity mapped pages can be effected).
64-bit mode paging tables must be located in the first 4 GBytes of physical-address
space prior to activating IA-32e mode. This is necessary because the MOV CR3
instruction used to initialize the page-directory base must be executed in legacy
mode prior to activating IA-32e mode (setting CR0.PG = 1 to enable paging).
Because MOV CR3 is executed in protected mode, only the lower 32 bits of the
register are written, limiting the table location to the low 4 GBytes of memory. Soft-
ware can relocate the page tables anywhere in physical memory after IA-32e mode
is activated.
The processor performs 64-bit mode consistency checks whenever software
attempts to modify any of the enable bits directly involved in activating IA-32e mode
(IA32_EFER.LME, CR0.PG, and CR4.PAE). It will generate a general protection fault
(#GP) if consistency checks fail. 64-bit mode consistency checks ensure that the
processor does not enter an undefined mode or state with unpredictable behavior.
64-bit mode consistency checks fail in the following circumstances:
• An attempt is made to enable or disable IA-32e mode while paging is enabled.
• IA-32e mode is enabled and an attempt is made to enable paging prior to
enabling physical-address extensions (PAE).
• IA-32e mode is active and an attempt is made to disable physical-address
extensions (PAE).
• If the current CS has the L-bit set on an attempt to activate IA-32e mode.
• If the TR contains a 16-bit TSS.
9.8.5.1
IA-32e Mode System Data Structures
After activating IA-32e mode, the system-descriptor-table registers (GDTR, LDTR,
IDTR, TR) continue to reference legacy protected-mode descriptor tables. Tables
referenced by the descriptors all reside in the lower 4 GBytes of linear-address space.
After activating IA-32e mode, 64-bit operating-systems should use the LGDT, LLDT,
LIDT, and LTR instructions to load the system-descriptor-table registers with refer-
ences to 64-bit descriptor tables.
9.8.5.2
IA-32e Mode Interrupts and Exceptions
Software must not allow exceptions or interrupts to occur between the time IA-32e
mode is activated and the update of the interrupt-descriptor-table register (IDTR)
that establishes references to a 64-bit interrupt-descriptor table (IDT). This is
because the IDT remains in legacy form immediately after IA-32e mode is activated.
If an interrupt or exception occurs prior to updating the IDTR, a legacy 32-bit inter-
rupt gate will be referenced and interpreted as a 64-bit interrupt gate with unpredict-
able results. External interrupts can be disabled by using the CLI instruction.
Non-maskable interrupts (NMI) must be disabled using external hardware.
Vol. 3 9-15
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
9.8.5.3
64-bit Mode and Compatibility Mode Operation
IA-32e mode uses two code segment-descriptor bits (CS.L and CS.D, see Figure 3-8)
to control the operating modes after IA-32e mode is initialized. If CS.L = 1 and CS.D =
0, the processor is running in 64-bit mode. With this encoding, the default operand
size is 32 bits and default address size is 64 bits. Using instruction prefixes, operand
size can be changed to 64 bits or 16 bits; address size can be changed to 32 bits.
When IA-32e mode is active and CS.L = 0, the processor operates in compatibility
mode. In this mode, CS.D controls default operand and address sizes exactly as it
does in the IA-32 architecture. Setting CS.D = 1 specifies default operand and
address size as 32 bits. Clearing CS.D to 0 specifies default operand and address size
as 16 bits (the CS.L = 1, CS.D = 1 bit combination is reserved).
Compatibility mode execution is selected on a code-segment basis. This mode allows
legacy applications to coexist with 64-bit applications running in 64-bit mode. An
operating system running in IA-32e mode can execute existing 16-bit and 32-bit
applications by clearing their code-segment descriptor’s CS.L bit to 0.
In compatibility mode, the following system-level mechanisms continue to operate
using the IA-32e-mode architectural semantics:
• Linear-to-physical address translation uses the 64-bit mode extended page-
translation mechanism.
• Interrupts and exceptions are handled using the 64-bit mode mechanisms.
• System calls (calls through call gates and SYSENTER/SYSEXIT) are handled using
the IA-32e mode mechanisms.
9.8.5.4
Switching Out of IA-32e Mode Operation
To return from IA-32e mode to paged-protected mode operation. Operating systems
must use the following sequence:
1. Switch to compatibility mode.
2. Deactivate IA-32e mode by clearing CR0.PG = 0. This causes the processor to set
IA32_EFER.LMA = 0. The MOV CR0 instruction used to disable paging and
subsequent instructions must be located in an identity-mapped page.
3. Load CR3 with the physical base address of the legacy page-table-directory base
address.
4. Disable IA-32e mode by setting IA32_EFER.LME = 0.
5. Enable legacy paged-protected mode by setting CR0.PG = 1
6. A branch instruction must follow the MOV CR0 that enables paging. Both the MOV
CR0 and the branch instruction must be located in an identity-mapped page.
Registers only available in 64-bit mode (R8-R15 and XMM8-XMM15) are preserved
across transitions from 64-bit mode into compatibility mode then back into 64-bit
mode. However, values of R8-R15 and XMM8-XMM15 are undefined after transitions
9-16 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
from 64-bit mode through compatibility mode to legacy or real mode and then back
through compatibility mode to 64-bit mode.
9.9
MODE SWITCHING
To use the processor in protected mode after hardware or software reset, a mode
switch must be performed from real-address mode. Once in protected mode, soft-
ware generally does not need to return to real-address mode. To run software written
to run in real-address mode (8086 mode), it is generally more convenient to run the
9.9.1
Switching to Protected Mode
Before switching to protected mode from real mode, a minimum set of system data
structures and code modules must be loaded into memory, as described in Section
9.8, “Software Initialization for Protected-Mode Operation.” Once these tables are
created, software initialization code can switch into protected mode.
Protected mode is entered by executing a MOV CR0 instruction that sets the PE flag
in the CR0 register. (In the same instruction, the PG flag in register CR0 can be set to
enable paging.) Execution in protected mode begins with a CPL of 0.
Intel 64 and IA-32 processors have slightly different requirements for switching to
protected mode. To insure upwards and downwards code compatibility with Intel 64
and IA-32 processors, we recommend that you follow these steps:
1. Disable interrupts. A CLI instruction disables maskable hardware interrupts. NMI
interrupts can be disabled with external circuitry. (Software must guarantee that
no exceptions or interrupts are generated during the mode switching operation.)
2. Execute the LGDT instruction to load the GDTR register with the base address of
the GDT.
3. Execute a MOV CR0 instruction that sets the PE flag (and optionally the PG flag)
in control register CR0.
4. Immediately following the MOV CR0 instruction, execute a far JMP or far CALL
instruction. (This operation is typically a far jump or call to the next instruction in
the instruction stream.)
5. The JMP or CALL instruction immediately after the MOV CR0 instruction changes
the flow of execution and serializes the processor.
6. If paging is enabled, the code for the MOV CR0 instruction and the JMP or CALL
instruction must come from a page that is identity mapped (that is, the linear
address before the jump is the same as the physical address after paging and
protected mode is enabled). The target instruction for the JMP or CALL instruction
does not need to be identity mapped.
Vol. 3 9-17
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
7. If a local descriptor table is going to be used, execute the LLDT instruction to load
the segment selector for the LDT in the LDTR register.
8. Execute the LTR instruction to load the task register with a segment selector to
the initial protected-mode task or to a writable area of memory that can be used
to store TSS information on a task switch.
9. After entering protected mode, the segment registers continue to hold the
contents they had in real-address mode. The JMP or CALL instruction in step 4
resets the CS register. Perform one of the following operations to update the
contents of the remaining segment registers.
— Reload segment registers DS, SS, ES, FS, and GS. If the ES, FS, and/or GS
registers are not going to be used, load them with a null selector.
— Perform a JMP or CALL instruction to a new task, which automatically resets
the values of the segment registers and branches to a new code segment.
10. Execute the LIDT instruction to load the IDTR register with the address and limit
of the protected-mode IDT.
11. Execute the STI instruction to enable maskable hardware interrupts and perform
the necessary hardware operation to enable NMI interrupts.
Random failures can occur if other instructions exist between steps 3 and 4 above.
Failures will be readily seen in some situations, such as when instructions that refer-
ence memory are inserted between steps 3 and 4 while in system management
mode.
9.9.2
Switching Back to Real-Address Mode
The processor switches from protected mode back to real-address mode if software
clears the PE bit in the CR0 register with a MOV CR0 instruction. A procedure that re-
enters real-address mode should perform the following steps:
1. Disable interrupts. A CLI instruction disables maskable hardware interrupts. NMI
interrupts can be disabled with external circuitry.
2. If paging is enabled, perform the following operations:
— Transfer program control to linear addresses that are identity mapped to
physical addresses (that is, linear addresses equal physical addresses).
— Insure that the GDT and IDT are in identity mapped pages.
— Clear the PG bit in the CR0 register.
— Move 0H into the CR3 register to flush the TLB.
3. Transfer program control to a readable segment that has a limit of 64 KBytes
(FFFFH). This operation loads the CS register with the segment limit required in
real-address mode.
9-18 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
4. Load segment registers SS, DS, ES, FS, and GS with a selector for a descriptor
containing the following values, which are appropriate for real-address mode:
— Limit = 64 KBytes (0FFFFH)
— Byte granular (G = 0)
— Expand up (E = 0)
— Writable (W = 1)
— Present (P = 1)
— Base = any value
5. The segment registers must be loaded with non-null segment selectors or the
segment registers will be unusable in real-address mode. Note that if the
segment registers are not reloaded, execution continues using the descriptor
attributes loaded during protected mode.
6. Execute an LIDT instruction to point to a real-address mode interrupt table that is
within the 1-MByte real-address mode address range.
7. Clear the PE flag in the CR0 register to switch to real-address mode.
8. Execute a far JMP instruction to jump to a real-address mode program. This
operation flushes the instruction queue and loads the appropriate base and
access rights values in the CS register.
9. Load the SS, DS, ES, FS, and GS registers as needed by the real-address mode
code. If any of the registers are not going to be used in real-address mode, write
0s to them.
10. Execute the STI instruction to enable maskable hardware interrupts and perform
the necessary hardware operation to enable NMI interrupts.
NOTE
All the code that is executed in steps 1 through 9 must be in a single
page and the linear addresses in that page must be identity mapped
to physical addresses.
9.10
INITIALIZATION AND MODE SWITCHING EXAMPLE
This section provides an initialization and mode switching example that can be incor-
porated into an application. This code was originally written to initialize the Intel386
processor, but it will execute successfully on the Pentium 4, Intel Xeon, P6 family,
Pentium, and Intel486 processors. The code in this example is intended to reside in
EPROM and to run following a hardware reset of the processor. The function of the
code is to do the following:
• Establish a basic real-address mode operating environment.
• Load the necessary protected-mode system data structures into RAM.
Vol. 3 9-19
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
• Load the system registers with the necessary pointers to the data structures and
the appropriate flag settings for protected-mode operation.
• Switch the processor to protected mode.
Figure 9-3 shows the physical memory layout for the processor following a hardware
reset and the starting point of this example. The EPROM that contains the initializa-
starting at address FFFFFFFFH and going down from there. The address of the first
instruction to be executed is at FFFFFFF0H, the default starting address for the
processor following a hardware reset.
The main steps carried out in this example are summarized in Table 9-4. The source
listing for the example (with the filename STARTUP.ASM) is given in Example 9-1.
The line numbers given in Table 9-4 refer to the source listing.
The following are some additional notes concerning this example:
• When the processor is switched into protected mode, the original code segment
base-address value of FFFF0000H (located in the hidden part of the CS register)
is retained and execution continues from the current offset in the EIP register.
The processor will thus continue to execute code in the EPROM until a far jump or
call is made to a new code segment, at which time, the base address in the CS
register will be changed.
• Maskable hardware interrupts are disabled after a hardware reset and should
remain disabled until the necessary interrupt handlers have been installed. The
NMI interrupt is not disabled following a reset. The NMI# pin must thus be
inhibited from being asserted until an NMI handler has been loaded and made
available to the processor.
• The use of a temporary GDT allows simple transfer of tables from the EPROM to
anywhere in the RAM area. A GDT entry is constructed with its base pointing to
address 0 and a limit of 4 GBytes. When the DS and ES registers are loaded with
this descriptor, the temporary GDT is no longer needed and can be replaced by
the application GDT.
• This code loads one TSS and no LDTs. If more TSSs exist in the application, they
must be loaded into RAM. If there are LDTs they may be loaded as well.
9-20 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
After Reset
FFFF FFFFH
FFFF FFF0H
64K EPROM
FFFF 0000H
[CS.BASE+EIP]
EIP = 0000 FFF0H
CS.BASE = FFFF 0000H
DS.BASE = 0H
ES.BASE = 0H
SS.BASE = 0H
ESP = 0H
0
[SP, DS, SS, ES]
Figure 9-3. Processor State After Reset
Table 9-4. Main Initialization Steps in STARTUP.ASM Source Listing
STARTUP.ASM Line
Description
Numbers
From
157
To
157
169
Jump (short) to the entry code in the EPROM
162
Construct a temporary GDT in RAM with one entry:
0 - null
1 - R/W data segment, base = 0, limit = 4 GBytes
171
174
179
184
172
177
181
186
Load the GDTR to point to the temporary GDT
Load CR0 with PE flag set to switch to protected mode
Jump near to clear real mode instruction queue
Load DS, ES registers with GDT[1] descriptor, so both point to the
entire physical memory space
Vol. 3 9-21
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
Table 9-4. Main Initialization Steps in STARTUP.ASM Source Listing (Contd.)
STARTUP.ASM Line
Numbers
Description
From
To
188
195
Perform specific board initialization that is imposed by the new
protected mode
196
220
241
244
247
263
218
238
243
245
261
267
Copy the application's GDT from ROM into RAM
Copy the application's IDT from ROM into RAM
Load application's GDTR
Load application's IDTR
Copy the application's TSS from ROM into RAM
Update TSS descriptor and other aliases in GDT (GDT alias or IDT
alias)
277
282
287
288
289
290
296
277
286
287
288
289
293
296
Load the task register (without task switch) using LTR instruction
Load SS, ESP with the value found in the application's TSS
Push EFLAGS value found in the application's TSS
Push CS value found in the application's TSS
Push EIP value found in the application's TSS
Load DS, ES with the value found in the application's TSS
Perform IRET; pop the above values and enter the application code
9.10.1
Assembler Usage
In this example, the Intel assembler ASM386 and build tools BLD386 are used to
assemble and build the initialization code module. The following assumptions are
used when using the Intel ASM386 and BLD386 tools.
• The ASM386 will generate the right operand size opcodes according to the code-
segment attribute. The attribute is assigned either by the ASM386 invocation
controls or in the code-segment definition.
• If a code segment that is going to run in real-address mode is defined, it must be
set to a USE 16 attribute. If a 32-bit operand is used in an instruction in this code
segment (for example, MOV EAX, EBX), the assembler automatically generates
an operand prefix for the instruction that forces the processor to execute a 32-bit
operation, even though its default code-segment attribute is 16-bit.
• Intel's ASM386 assembler allows specific use of the 16- or 32-bit instructions, for
example, LGDTW, LGDTD, IRETD. If the generic instruction LGDT is used, the
default- segment attribute will be used to generate the right opcode.
9-22 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
9.10.2
STARTUP.ASM Listing
Example 9-1 provides high-level sample code designed to move the processor into
protected mode. This listing does not include any opcode and offset information.
Example 9-1. STARTUP.ASM
MS-DOS* 5.0(045-N) 386(TM) MACRO ASSEMBLER STARTUP 09:44:51 08/19/92
PAGE 1
MS-DOS 5.0(045-N) 386(TM) MACRO ASSEMBLER V4.0, ASSEMBLY OF MODULE
STARTUP
OBJECT MODULE PLACED IN startup.obj
ASSEMBLER INVOKED BY: f:\386tools\ASM386.EXE startup.a58 pw (132 )
LINE
SOURCE
NAME
1
2
STARTUP
3 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
4 ;
5 ; ASSUMPTIONS:
6 ;
7 ;
8 ;
1. The bottom 64K of memory is ram, and can be used for
scratch space by this module.
9 ;
10 ;
11 ;
12 ;
2. The system has sufficient free usable ram to copy the
initial GDT, IDT, and TSS
13 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
14
15 ; configuration data - must match with build definition
16
17 CS_BASE
18
EQU
0FFFF0000H
19 ; CS_BASE is the linear address of the segment STARTUP_CODE
20 ; - this is specified in the build language file
21
22 RAM_START
23
EQU
400H
24 ; RAM_START is the start of free, usable ram in the linear
25 ; memory space. The GDT, IDT, and initial TSS will be
26 ; copied above this space, and a small data segment will be
27 ; discarded at this linear address. The 32-bit word at
Vol. 3 9-23
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
28 ; RAM_START will contain the linear address of the first
29 ; free byte above the copied tables - this may be useful if
30 ; a memory manager is used.
31
32 TSS_INDEX
33
EQU
10
34 ; TSS_INDEX is the index of the TSS of the first task to
35 ; run after startup
36
37
38 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
39
40 ; ------------------------- STRUCTURES and EQU ---------------
41 ; structures for system data
42
43 ; TSS structure
44 TASK_STATE STRUC
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
link
link_h
ESP0
SS0
SS0_h
ESP1
SS1
SS1_h
ESP2
SS2
DW ?
DW ?
DD ?
DW ?
DW ?
DD ?
DW ?
DW ?
DD ?
DW ?
DW ?
SS2_h
CR3_reg DD ?
EIP_reg DD ?
EFLAGS_regDD ?
EAX_reg DD ?
ECX_reg DD ?
EDX_reg DD ?
EBX_reg DD ?
ESP_reg DD ?
EBP_reg DD ?
ESI_reg DD ?
EDI_reg DD ?
ES_reg
ES_h
CS_reg
CS_h
DW ?
DW ?
DW ?
DW ?
9-24 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
71
72
73
74
75
76
77
78
79
80
81
82
SS_reg
SS_h
DS_reg
DS_h
FS_reg
FS_h
GS_reg
GS_h
DW ?
DW ?
DW ?
DW ?
DW ?
DW ?
DW ?
DW ?
LDT_reg DW ?
LDT_h DW ?
TRAP_reg DW ?
IO_map_baseDW ?
83 TASK_STATE ENDS
84
85 ; basic structure of a descriptor
86 DESC
STRUC
87
88
89
lim_0_15 DW ?
bas_0_15 DW ?
bas_16_23DB ?
90
91
access
gran
DB ?
DB ?
92
bas_24_31DB ?
ENDS
93 DESC
94
95 ; structure for use with LGDT and LIDT instructions
96 TABLE_REG STRUC
97
98
table_limDW ?
table_linearDD ?
99 TABLE_REG ENDS
100
101 ; offset of GDT and IDT descriptors in builder generated GDT
102 GDT_DESC_OFF
103 IDT_DESC_OFF
104
EQU 1*SIZE(DESC)
EQU 2*SIZE(DESC)
105 ; equates for building temporary GDT in RAM
106 LINEAR_SEL
107 LINEAR_PROTO_LO
108 LINEAR_PROTO_HI
109
EQU
EQU
EQU
1*SIZE (DESC)
00000FFFFH ; LINEAR_ALIAS
000CF9200H
110 ; Protection Enable Bit in CR0
111 PE_BIT EQU 1B
112
113 ; ------------------------------------------------------------
Vol. 3 9-25
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
114
115 ; ------------------------- DATA SEGMENT----------------------
116
117 ; Initially, this data segment starts at linear 0, according
118 ; to the processor’s power-up state.
119
120 STARTUP_DATA
121
SEGMENT RW
122 free_mem_linear_base
123 TEMP_GDT
LABEL DWORD
LABEL BYTE ; must be first in segment
124 TEMP_GDT_NULL_DESC DESC
125 TEMP_GDT_LINEAR_DESC DESC
126
<>
<>
127 ; scratch areas for LGDT and LIDT instructions
128 TEMP_GDT_SCRATCH TABLE_REG <>
129 APP_GDT_RAM
130 APP_IDT_RAM
TABLE_REG
TABLE_REG
<>
<>
131
132 fill
133
; align end_data
DW
?
134 ; last thing in this segment - should be on a dword boundary
135 end_data
136
LABEL BYTE
137 STARTUP_DATA
ENDS
138 ; ------------------------------------------------------------
139
140
141 ; ------------------------- CODE SEGMENT----------------------
142 STARTUP_CODE SEGMENT ER PUBLIC USE16
143
144 ; filled in by builder
145
PUBLIC GDT_EPROM
146 GDT_EPROM TABLE_REG <>
147
148 ; filled in by builder
149
PUBLIC IDT_EPROM
150 IDT_EPROM TABLE_REG <>
151
152 ; entry point into startup code - the bootstrap will vector
153 ; here with a near JMP generated by the builder. This
154 ; label must be in the top 64K of linear memory.
155
156
PUBLIC STARTUP
157 STARTUP:
158
9-26 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
159 ; DS,ES address the bottom 64K of flat linear memory
160 ASSUME DS:STARTUP_DATA, ES:STARTUP_DATA
161 ; See Figure 9-4
162 ; load GDTR with temporary GDT
163
164
165
166
167
168
169
170
171
172
173
LEA
MOV
MOV
MOV
MOV
MOV
MOV
EBX,TEMP_GDT ; build the TEMP_GDT in low ram,
DWORD PTR [EBX],0 ; where we can address
DWORD PTR [EBX]+4,0
DWORD PTR [EBX]+8, LINEAR_PROTO_LO
DWORD PTR [EBX]+12, LINEAR_PROTO_HI
TEMP_GDT_scratch.table_linear,EBX
TEMP_GDT_scratch.table_lim,15
DB 66H; execute a 32 bit LGDT
LGDT TEMP_GDT_scratch
174 ; enter protected mode
175
176
177
178
MOV
OR
MOV
EBX,CR0
EBX,PE_BIT
CR0,EBX
179 ; clear prefetch queue
180 JMP CLEAR_LABEL
181 CLEAR_LABEL:
182
183 ; make DS and ES address 4G of linear memory
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
MOV
MOV
MOV
CX,LINEAR_SEL
DS,CX
ES,CX
; do board specific initialization
;
;
; ......
;
; See Figure 9-5
; copy EPROM GDT to ram at:
;
MOV
ADD
MOV
RAM_START + size (STARTUP_DATA)
EAX,RAM_START
EAX,OFFSET (end_data)
EBX,RAM_START
Vol. 3 9-27
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
MOV
ADD
MOV
MOV
ECX, CS_BASE
ECX, OFFSET (GDT_EPROM)
ESI, [ECX].table_linear
EDI,EAX
MOVZX ECX, [ECX].table_lim
MOV
INC
MOV
MOV
APP_GDT_ram[EBX].table_lim,CX
ECX
EDX,EAX
APP_GDT_ram[EBX].table_linear,EAX
EAX,ECX
BYTE PTR ES:[EDI],BYTE PTR DS:[ESI]
ADD
REP MOVS
; fixup GDT base in descriptor
MOV
MOV
ROR
MOV
MOV
ECX,EDX
[EDX].bas_0_15+GDT_DESC_OFF,CX
ECX,16
[EDX].bas_16_23+GDT_DESC_OFF,CL
[EDX].bas_24_31+GDT_DESC_OFF,CH
; copy EPROM IDT to ram at:
; RAM_START+size(STARTUP_DATA)+SIZE (EPROM GDT)
MOV
ADD
MOV
MOV
MOVZX ECX, [ECX].table_lim
MOV
INC
MOV
MOV
ECX, CS_BASE
ECX, OFFSET (IDT_EPROM)
ESI, [ECX].table_linear
EDI,EAX
APP_IDT_ram[EBX].table_lim,CX
ECX
APP_IDT_ram[EBX].table_linear,EAX
EBX,EAX
ADD
EAX,ECX
REP MOVS
BYTE PTR ES:[EDI],BYTE PTR DS:[ESI]
; fixup IDT pointer in GDT
[EDX].bas_0_15+IDT_DESC_OFF,BX
EBX,16
[EDX].bas_16_23+IDT_DESC_OFF,BL
[EDX].bas_24_31+IDT_DESC_OFF,BH
MOV
ROR
MOV
MOV
; load GDTR and IDTR
EBX,RAM_START
MOV
DB
APP_GDT_ram[EBX]
DB 66H
APP_IDT_ram[EBX]
66H
; execute a 32 bit LGDT
LGDT
LIDT
; execute a 32 bit LIDT
9-28 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
; move the TSS
EDI,EAX
EBX,TSS_INDEX*SIZE(DESC)
ECX,GDT_DESC_OFF ;build linear address for TSS
GS,CX
DH,GS:[EBX].bas_24_31
DL,GS:[EBX].bas_16_23
EDX,16
DX,GS:[EBX].bas_0_15
ESI,EDX
ECX,EBX
ECX
EDX,EAX
MOV
MOV
MOV
MOV
MOV
MOV
ROL
MOV
MOV
LSL
INC
MOV
ADD
EAX,ECX
REP MOVS
BYTE PTR ES:[EDI],BYTE PTR DS:[ESI]
; fixup TSS pointer
GS:[EBX].bas_0_15,DX
EDX,16
GS:[EBX].bas_24_31,DH
GS:[EBX].bas_16_23,DL
EDX,16
MOV
ROL
MOV
MOV
ROL
;save start of free ram at linear location RAMSTART
MOV free_mem_linear_base+RAM_START,EAX
;code to move the LDT could be added, and should resemble
;that used to move the TSS
; load task register
BX ; No task switch, only descriptor loading
; See Figure 9-6
; load minimal set of registers necessary to simulate task
; switch
LTR
MOV
MOV
MOV
MOV
PUSH
PUSH
AX,[EDX].SS_reg
EDI,[EDX].ESP_reg
SS,AX
ESP,EDI
DWORD PTR [EDX].EFLAGS_reg
DWORD PTR [EDX].CS_reg
; start loading registers
; stack now valid
Vol. 3 9-29
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
289
290
291
292
293
294
295
296
297
PUSH
MOV
MOV
MOV
MOV
DWORD PTR [EDX].EIP_reg
AX,[EDX].DS_reg
BX,[EDX].ES_reg
DS,AX
ES,BX
; DS and ES no longer linear memory
; simulate far jump to initial task
IRETD
298 STARTUP_CODE ENDS
*** WARNING #377 IN 298, (PASS 2) SEGMENT CONTAINS PRIVILEGED
INSTRUCTION(S)
299
300 END STARTUP, DS:STARTUP_DATA, SS:STARTUP_DATA
301
302
ASSEMBLY COMPLETE, 1 WARNING, NO ERRORS.
9-30 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
FFFF FFFFH
FFFF 0000H
START: [CS.BASE+EIP]
• Jump near start
• Construct TEMP_GDT
• LGDT
• Move to protected mode
DS, ES = GDT[1]
4 GB
Base
Limit
GDT_SCRATCH
TEMP_GDT
GDT [1]
GDT [0]
Base=0, Limit=4G
0
Figure 9-4. Constructing Temporary GDT and Switching to Protected Mode (Lines
162-172 of List File)
Vol. 3 9-31
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
FFFF FFFFH
TSS
IDT
GDT
• Move the GDT, IDT, TSS
from ROM to RAM
• Fix Aliases
• LTR
TSS RAM
IDT RAM
GDT RAM
RAM_START
0
Figure 9-5. Moving the GDT, IDT, and TSS from ROM to RAM (Lines 196-261 of List
File)
9-32 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
•
•
EIP
EFLAGS
SS = TSS.SS
ESP = TSS.ESP
PUSH TSS.EFLAG
PUSH TSS.CS
PUSH TSS.EIP
ES = TSS.ES
DS = TSS.DS
IRET
•
•
•
ESP
•
ES
CS
SS
DS
GDT
TSS RAM
IDT RAM
GDT RAM
IDT Alias
GDT Alias
0
RAM_START
Figure 9-6. Task Switching (Lines 282-296 of List File)
9.10.3
MAIN.ASM Source Code
The file MAIN.ASM shown in Example 9-2 defines the data and stack segments for
this application and can be substituted with the main module task written in a high-
level language that is invoked by the IRET instruction executed by STARTUP.ASM.
Example 9-2. MAIN.ASM
NAME
data
main_module
SEGMENT RW
dw 1000 dup(?)
DATA
ENDS
stack stackseg 800
Vol. 3 9-33
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
CODE SEGMENT ER use32 PUBLIC
main_start:
nop
nop
nop
CODE ENDS
END main_start, ds:data, ss:stack
9.10.4
Supporting Files
The batch file shown in Example 9-3 can be used to assemble the source code files
STARTUP.ASM and MAIN.ASM and build the final application.
Example 9-3. Batch File to Assemble and Build the Application
ASM386 STARTUP.ASM
ASM386 MAIN.ASM
BLD386 STARTUP.OBJ, MAIN.OBJ buildfile(EPROM.BLD) bootstrap(STARTUP)
Bootload
BLD386 performs several operations in this example:
It allocates physical memory location to segments and tables.
It generates tables using the build file and the input files.
It links object files and resolves references.
It generates a boot-loadable file to be programmed into the EPROM.
Example 9-4 shows the build file used as an input to BLD386 to perform the above
functions.
Example 9-4. Build File
INIT_BLD_EXAMPLE;
SEGMENT
*SEGMENTS(DPL = 0)
, startup.startup_code(BASE = 0FFFF0000H)
;
TASK
BOOT_TASK(OBJECT = startup, INITIAL,DPL = 0,
NOT INTENABLED)
,
PROTECTED_MODE_TASK(OBJECT = main_module,DPL = 0,
NOT INTENABLED)
;
9-34 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
TABLE
GDT (
LOCATION = GDT_EPROM
, ENTRY = (
10: PROTECTED_MODE_TASK
,
,
,
,
,
startup.startup_code
startup.startup_data
main_module.data
main_module.code
main_module.stack
)
),
IDT (
LOCATION = IDT_EPROM
);
MEMORY
(
RESERVE = (0..3FFFH
-- Area for the GDT, IDT, TSS copied from ROM
60000H..0FFFEFFFFH)
,
-- Eprom size 64K
, RANGE = (RAM_AREA = RAM (4000H..05FFFFH))
);
END
Table 9-5 shows the relationship of each build item with an ASM source file.
Table 9-5. Relationship Between BLD Item and ASM Source File
Item
ASM386 and
Startup.A58
BLD386 Controls
and BLD file
Effect
Bootstrap
public startup
startup:
bootstrap
Near jump at 0FFFFFFF0H
to start.
start(startup)
GDT location
public GDT_EPROM
TABLE
The location of the GDT
will be programmed into
the GDT_EPROM location.
GDT_EPROM TABLE_REG <> GDT(location = GDT_EPROM)
IDT location
public IDT_EPROM
TABLE
The location of the IDT
will be programmed into
the IDT_EPROM location.
IDT_EPROM TABLE_REG <>
IDT(location = IDT_EPROM
Vol. 3 9-35
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
Table 9-5. Relationship Between BLD Item and ASM Source File (Contd.)
Item
ASM386 and
Startup.A58
BLD386 Controls
and BLD file
Effect
RAM start
RAM_START equ 400H
memory (reserve = (0..3FFFH))
RAM_START is used as
the ram destination for
moving the tables. It must
be excluded from the
application's segment
area.
Location of the TSS_INDEX EQU 10
application TSS
TABLE GDT(
ENTRY = (10:
PROTECTED_MODE_
TASK))
Put the descriptor of the
application TSS in GDT
entry 10.
in the GDT
EPROM size
and location
size and location of the
initialization code
SEGMENT startup.code (base =
0FFFF0000H) ...memory
(RANGE(
Initialization code size
must be less than 64K
and resides at upper most
64K of the 4-GByte
memory space.
ROM_AREA = ROM(x..y))
9.11
MICROCODE UPDATE FACILITIES
The Pentium 4, Intel Xeon, and P6 family processors have the capability to correct
errata by loading an Intel-supplied data block into the processor. The data block is
called a microcode update. This section describes the mechanisms the BIOS needs to
provide in order to use this feature during system initialization. It also describes a
specification that permits the incorporation of future updates into a system BIOS.
Intel considers the release of a microcode update for a silicon revision to be the
equivalent of a processor stepping and completes a full-stepping level validation for
releases of microcode updates.
A microcode update is used to correct errata in the processor. The BIOS, which has
an update loader, is responsible for loading the update on processors during system
initialization (Figure 9-7). There are two steps to this process: the first is to incorpo-
rate the necessary update data blocks into the BIOS; the second is to load update
data blocks into the processor.
9-36 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
Update
Loader
Update
Blocks
New Update
CPU
BIOS
Figure 9-7. Applying Microcode Updates
9.11.1
Microcode Update
A microcode update consists of an Intel-supplied binary that contains a descriptive
update is tailored for a specific list of processor signatures. A mismatch of the
processor’s signature with the signature contained in the update will result in a
failure to load. A processor signature includes the extended family, extended model,
type, family, model, and stepping of the processor (starting with processor family
0fH, model 03H, a given microcode update may be associated with one of multiple
processor signatures; see Section 9.11.2 for detail).
Microcode updates are composed of a multi-byte header, followed by encrypted data
and then by an optional extended signature table. Table 9-6 provides a definition of
the fields; Table 9-7 shows the format of an update.
The header is 48 bytes. The first 4 bytes of the header contain the header version.
The update header and its reserved fields are interpreted by software based upon the
header version. An encoding scheme guards against tampering and provides a
means for determining the authenticity of any given update. For microcode updates
with a data size field equal to 00000000H, the size of the microcode update is 2048
bytes. The first 48 bytes contain the microcode update header. The remaining 2000
bytes contain encrypted data.
For microcode updates with a data size not equal to 00000000H, the total size field
specifies the size of the microcode update. The first 48 bytes contain the microcode
update header. The second part of the microcode update is the encrypted data. The
data size field of the microcode update header specifies the encrypted data size, its
value must be a multiple of the size of DWORD. The total size field of the microcode
update header specifies the encrypted data size plus the header size; its value must
be in multiples of 1024 bytes (1 KBytes). The optional extended signature table if
implemented follows the encrypted data, and its size is calculated by (Total Size –
(Data Size + 48)).
Vol. 3 9-37
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
NOTE
The optional extended signature table is supported starting with
processor family 0FH, model 03H.
.
Table 9-6. Microcode Update Field Definitions
Field Name
Offset
(bytes)
Length
(bytes)
Description
Header Version
Update Revision
0
4
4
4
Version number of the update header.
Unique version number for the update, the basis for the
update signature provided by the processor to indicate
the current update functioning within the processor.
Used by the BIOS to authenticate the update and verify
that the processor loads successfully. The value in this
field cannot be used for processor stepping identification
alone. This is a signed 32-bit number.
Date
8
4
4
Date of the update creation in binary format: mmddyyyy
(e.g. 07/18/98 is 07181998H).
Processor
Signature
12
Extended family, extended model, type, family, model,
and stepping of processor that requires this particular
update revision (e.g., 00000650H). Each microcode
update is designed specifically for a given extended
family, extended model, type, family, model, and stepping
of the processor.
The BIOS uses the processor signature field in
conjunction with the CPUID instruction to determine
whether or not an update is appropriate to load on a
processor. The information encoded within this field
exactly corresponds to the bit representations returned
by the CPUID instruction.
Checksum
16
4
Checksum of Update Data and Header. Used to verify the
integrity of the update header and data. Checksum is
correct when the summation of all the DWORDs (including
the extended Processor Signature Table) that comprise
the microcode update result in 00000000H.
Loader Revision
Processor Flags
20
24
4
4
Version number of the loader program needed to
correctly load this update. The initial version is
00000001H.
Platform type information is encoded in the lower 8 bits
of this 4-byte field. Each bit represents a particular
platform type for a given CPUID. The BIOS uses the
processor flags field in conjunction with the platform Id
bits in MSR (17H) to determine whether or not an update
is appropriate to load on a processor. Multiple bits may be
set representing support for multiple platform IDs.
Data Size
Total Size
28
32
4
4
Specifies the size of the encrypted data in bytes, and
must be a multiple of DWORDs. If this value is
00000000H, then the microcode update encrypted data
is 2000 bytes (or 500 DWORDs).
Specifies the total size of the microcode update in bytes.
It is the summation of the header size, the encrypted
data size and the size of the optional extended signature
table. This value is always a multiple of 1024.
9-38 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
Table 9-6. Microcode Update Field Definitions (Contd.)
Field Name
Offset
(bytes)
Length
(bytes)
Description
Reserved
36
48
12
Reserved fields for future expansion
Update Data
Data Size or Update data
2000
ExtendedSignature Data Size +
4
Specifies the number of extended signature structures
Count
48
(Processor Signature[n], processor flags[n] and
checksum[n]) that exist in this microcode update.
Extended
Checksum
Data Size +
52
4
Checksum of update extended processor signature table.
Used to verify the integrity of the extended processor
signature table. Checksum is correct when the
summation of the DWORDs that comprise the extended
processor signature table results in 00000000H.
Reserved
Data Size +
56
12
4
Reserved fields
Processor
Signature[n]
Data Size +
68 + (n * 12)
Extended family, extended model, type, family, model,
and stepping of processor that requires this particular
update revision (e.g., 00000650H). Each microcode
update is designed specifically for a given extended
family, extended model, type, family, model, and stepping
of the processor.
The BIOS uses the processor signature field in
conjunction with the CPUID instruction to determine
whether or not an update is appropriate to load on a
processor. The information encoded within this field
exactly corresponds to the bit representations returned
by the CPUID instruction.
Processor Flags[n]
Checksum[n]
Data Size +
4
4
Platform type information is encoded in the lower 8 bits
of this 4-byte field. Each bit represents a particular
platform type for a given CPUID. The BIOS uses the
processor flags field in conjunction with the platform Id
bits in MSR (17H) to determine whether or not an update
is appropriate to load on a processor. Multiple bits may be
set representing support for multiple platform IDs.
72 + (n * 12)
Data Size +
Used by utility software to decompose a microcode
update into multiple microcode updates where each of
the new updates is constructed without the optional
Extended Processor Signature Table.
76 + (n * 12)
To calculate the Checksum, substitute the Primary
Processor Signature entry and the Processor Flags entry
with the corresponding Extended Patch entry. Delete the
Extended Processor Signature Table entries. The
Checksum is correct when the summation of all DWORDs
that comprise the created Extended Processor Patch
results in 00000000H.
Vol. 3 9-39
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
Table 9-7. Microcode Update Format
16
31
24
8
0
Bytes
Header Version
Update Revision
Month: 8
0
4
8
Day: 8
Year: 16
Processor Signature (CPUID)
12
Checksum
16
20
24
Loader Revision
Processor Flags
Reserved (24 bits)
Data Size
28
32
36
48
Total Size
Reserved (12 Bytes)
Update Data (Data Size bytes, or 2000 Bytes if Data Size = 00000000H)
Extended Signature Count ‘n’
Data Size
+ 48
Extended Processor Signature Table Checksum
Reserved (12 Bytes)
Data Size
+ 52
Data Size
+ 56
Processor Signature[n]
Processor Flags[n]
Checksum[n]
Data Size
+ 68 +
(n * 12)
Data Size
+ 72 +
(n * 12)
Data Size
+ 76 +
(n * 12)
9-40 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
9.11.2
Optional Extended Signature Table
The extended signature table is a structure that may be appended to the end of the
(optional case). The extended signature table will always be present when the
encrypted data supports multiple processor steppings and/or models (required
case).
ture, which contains the extended signature count, the extended processor signature
table checksum, and 12 reserved bytes (Table 9-8). Following the extended signa-
ture header structure, the extended signature table contains 0-to-n extended
processor signature structures.
Each processor signature structure consist of the processor signature, processor
flags, and a checksum (Table 9-9).
The extended signature count in the extended signature header structure indicates
the number of processor signature structures that exist in the extended signature
table.
The extended processor signature table checksum is a checksum of all DWORDs that
comprise the extended signature table. That includes the extended signature count,
extended processor signature table checksum, 12 reserved bytes and the n
processor signature structures. A valid extended signature table exists when the
result of a DWORD checksum is 00000000H.
Table 9-8. Extended Processor Signature Table Header Structure
Extended Signature Count ‘n’
Extended Processor Signature Table Checksum
Reserved (12 Bytes)
Data Size + 48
Data Size + 52
Data Size + 56
Table 9-9. Processor Signature Structure
Processor Signature[n]
Processor Flags[n]
Checksum[n]
Data Size + 68 + (n * 12)
Data Size + 72 + (n * 12)
Data Size + 76 + (n * 12)
9.11.3
Processor Identification
Each microcode update is designed to for a specific processor or set of processors. To
determine the correct microcode update to load, software must ensure that one of
the processor signatures embedded in the microcode update matches the 32-bit
processor signature returned by the CPUID instruction when executed by the target
processor with EAX = 1. Attempting to load a microcode update that does not match
Vol. 3 9-41
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
a processor signature embedded in the microcode update with the processor signa-
ture returned by CPUID will cause the BIOS to reject the update.
Example 9-5 shows how to check for a valid processor signature match between the
processor and microcode update.
Example 9-5. Pseudo Code to Validate the Processor Signature
ProcessorSignature ← CPUID(1):EAX
If (Update.HeaderVersion == 00000001h)
{
// first check the ProcessorSignature field
If (ProcessorSignature == Update.ProcessorSignature)
Success
// if extended signature is present
Else If (Update.TotalSize > (Update.DataSize + 48))
{
//
// Assume the Data Size has been used to calculate the
// location of Update.ProcessorSignature[0].
//
For (N ← 0; ((N < Update.ExtendedSignatureCount) AND
(ProcessorSignature != Update.ProcessorSignature[N])); N++);
// if the loops ended when the iteration count is
// less than the number of processor signatures in
// the table, we have a match
If (N < Update.ExtendedSignatureCount)
Success
Else
Fail
}
Else
Fail
Else
Fail
9.11.4
Platform Identification
In addition to verifying the processor signature, the intended processor platform type
must be determined to properly target the microcode update. The intended
processor platform type is determined by reading the IA32_PLATFORM_ID register,
(MSR 17H). This 64-bit register must be read using the RDMSR instruction.
9-42 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
The three platform ID bits, when read as a binary coded decimal (BCD) number, indi-
cate the bit position in the microcode update header’s processor flags field associated
with the installed processor. The processor flags in the 48-byte header and the
processor flags field associated with the extended processor signature structures
may have multiple bits set. Each set bit represents a different platform ID that the
update supports.
Register Name:
MSR Address:
Access:
IA32_PLATFORM_ID
017H
Read Only
IA32_PLATFORM_ID is a 64-bit register accessed only when referenced as a Qword through a
RDMSR instruction.
Table 9-10. Processor Flags
Bit
Descriptions
63:53
52:50
Reserved
Platform Id Bits (RO). The field gives information concerning the intended platform for
the processor. See also Table 9-7.
52 51 50
0
0
0
0
1
1
1
1
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
Processor Flag 0
Processor Flag 1
Processor Flag 2
Processor Flag 3
Processor Flag 4
Processor Flag 5
Processor Flag 6
Processor Flag 7
49:0
Reserved
To validate the platform information, software may implement an algorithm similar to
the algorithms in Example 9-6.
Example 9-6. Pseudo Code Example of Processor Flags Test
Flag ← 1 << IA32_PLATFORM_ID[52:50]
If (Update.HeaderVersion == 00000001h)
{
If (Update.ProcessorFlags & Flag)
{
Load Update
Vol. 3 9-43
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
}
Else
{
//
// Assume the Data Size has been used to calculate the
// location of Update.ProcessorSignature[N] and a match
// on Update.ProcessorSignature[N] has already succeeded
//
If (Update.ProcessorFlags[n] & Flag)
{
Load Update
}
}
}
9.11.5
Microcode Update Checksum
Each microcode update contains a DWORD checksum located in the update header. It
is software’s responsibility to ensure that a microcode update is not corrupt. To check
for a corrupt microcode update, software must perform a unsigned DWORD (32-bit)
checksum of the microcode update. Even though some fields are signed, the
checksum procedure treats all DWORDs as unsigned. Microcode updates with a
header version equal to 00000001H must sum all DWORDs that comprise the micro-
code update. A valid checksum check will yield a value of 00000000H. Any other
value indicates the microcode update is corrupt and should not be loaded.
The checksum algorithm shown by the pseudo code in Example 9-7 treats the micro-
code update as an array of unsigned DWORDs. If the data size DWORD field at byte
offset 32 equals 00000000H, the size of the encrypted data is 2000 bytes, resulting
in 500 DWORDs. Otherwise the microcode update size in DWORDs = (Total Size / 4),
where the total size is a multiple of 1024 bytes (1 KBytes).
Example 9-7. Pseudo Code Example of Checksum Test
N ← 512
If (Update.DataSize != 00000000H)
N ← Update.TotalSize / 4
ChkSum ← 0
For (I ← 0; I < N; I++)
{
ChkSum ← ChkSum + MicrocodeUpdate[I]
}
9-44 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
If (ChkSum == 00000000H)
Success
Else
Fail
9.11.6
Microcode Update Loader
This section describes an update loader used to load an update into a Pentium 4, Intel
to ensure proper loading. The update loader described contains the minimal instruc-
tions needed to load an update. The specific instruction sequence that is required to
load an update is dependent upon the loader revision field contained within the
update header. This revision is expected to change infrequently (potentially, only
when new processor models are introduced).
Example 9-8 below represents the update loader with a loader revision of
00000001H. Note that the microcode update must be aligned on a 16-byte boundary
and the size of the microcode update must be 1-KByte granular.
Example 9-8. Assembly Code Example of Simple Microcode Update Loader
mov ecx,79h
xor eax,eax
xor ebx,ebx
mov ax,cs
; MSR to read in ECX
; clear EAX
; clear EBX
; Segment of microcode update
shl eax,4
mov bx,offset Update
add eax,ebx
add eax,48d
xor edx,edx
WRMSR
; Linear Address of Update in EAX
; Offset of the Update Data within the Update
; Zero in EDX
; microcode update trigger
The loader shown in Example 9-8 assumes that update is the address of a microcode
update (header and data) embedded within the code segment of the BIOS. It also
assumes that the processor is operating in real mode. The data may reside anywhere
in memory, aligned on a 16-byte boundary, that is accessible by the processor within
its current operating mode.
Before the BIOS executes the microcode update trigger (WRMSR) instruction, the
following must be true:
• In 64-bit mode, EAX contains the lower 32-bits of the microcode update linear
address. In protected mode, EAX contains the full 32-bit linear address of the
microcode update.
• In 64-bit mode, EDX contains the upper 32-bits of the microcode update linear
address. In protected mode, EDX equals zero.
Vol. 3 9-45
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
• ECX contains 79H (address of IA32_BIOS_UPDT_TRIG).
Other requirements are:
• If the update is loaded while the processor is in real mode, then the update data
may not cross a segment boundary.
• If the update is loaded while the processor is in real mode, then the update data
may not exceed a segment limit.
• If paging is enabled, pages that are currently present must map the update data.
• The microcode update data requires a 16-byte boundary alignment.
9.11.6.1 Hard Resets in Update Loading
The effects of a loaded update are cleared from the processor upon a hard reset.
Therefore, each time a hard reset is asserted during the BIOS POST, the update must
be reloaded on all processors that observed the reset. The effects of a loaded update
are, however, maintained across a processor INIT. There are no side effects caused
by loading an update into a processor multiple times.
9.11.6.2 Update in a Multiprocessor System
A multiprocessor (MP) system requires loading each processor with update data
appropriate for its CPUID and platform ID bits. The BIOS is responsible for ensuring
that this requirement is met and that the loader is located in a module executed by
all processors in the system. If a system design permits multiple steppings of
Pentium 4, Intel Xeon, and P6 family processors to exist concurrently; then the BIOS
must verify individual processors against the update header information to ensure
appropriate loading. Given these considerations, it is most practical to load the
update during MP initialization.
9.11.6.3 Update in a System Supporting Intel Hyper-Threading Technology
Intel Hyper-Threading Technology has implications on the loading of the microcode
update. The update must be loaded for each core in a physical processor. Thus, for a
processor supporting Intel Hyper-Threading Technology, only one logical processor
per core is required to load the microcode update. Each individual logical processor
can independently load the update. However, MP initialization must provide some
mechanism (e.g. a software semaphore) to force serialization of microcode update
loads and to prevent simultaneous load attempts to the same core.
9.11.6.4 Update in a System Supporting Dual-Core Technology
Dual-core technology has implications on the loading of the microcode update. The
microcode update facility is not shared between processor cores in the same physical
package. The update must be loaded for each core in a physical processor.
9-46 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
If processor core supports Intel Hyper-Threading Technology, the guideline described
in Section 9.11.6.3 also applies.
9.11.6.5 Update Loader Enhancements
The update loader presented in Section 9.11.6, “Microcode Update Loader,” is a
minimal implementation that can be enhanced to provide additional functionality.
Potential enhancements are described below:
• BIOS can incorporate multiple updates to support multiple steppings of the
Pentium 4, Intel Xeon, and P6 family processors. This feature provides for
operating in a mixed stepping environment on an MP system and enables a user
to upgrade to a later version of the processor. In this case, modify the loader to
check the CPUID and platform ID bits of the processor that it is running on
updates is only limited by available BIOS space.
• A loader can load the update and test the processor to determine if the update
was loaded correctly. See Section 9.11.7, “Update Signature and Verification.”
• A loader can verify the integrity of the update data by performing a checksum on
the double words of the update summing to zero. See Section 9.11.5, “Microcode
Update Checksum.”
• A loader can provide power-on messages indicating successful loading of an
update.
9.11.7
Update Signature and Verification
The Pentium 4, Intel Xeon, and P6 family processors provide capabilities to verify the
authenticity of a particular update and to identify the current update revision. This
section describes the model-specific extensions of processors that support this
feature. The update verification method below assumes that the BIOS will only verify
an update that is more recent than the revision currently loaded in the processor.
CPUID returns a value in a model specific register in addition to its usual register
return values. The semantics of CPUID cause it to deposit an update ID value in the
64-bit model-specific register at address 08BH (IA32_BIOS_SIGN_ID). If no update
is present in the processor, the value in the MSR remains unmodified. The BIOS must
pre-load a zero into the MSR before executing CPUID. If a read of the MSR at 8BH still
returns zero after executing CPUID, this indicates that no update is present.
The update ID value returned in the EDX register after RDMSR executes indicates the
revision of the update loaded in the processor. This value, in combination with the
CPUID value returned in the EAX register, uniquely identifies a particular update. The
signature ID can be directly compared with the update revision field in a microcode
update header for verification of a correct load. No consecutive updates released for
a given stepping of a processor may share the same signature. The processor signa-
ture returned by CPUID differentiates updates for different steppings.
Vol. 3 9-47
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
9.11.7.1 Determining the Signature
An update that is successfully loaded into the processor provides a signature that
matches the update revision of the currently functioning revision. This signature is
available any time after the actual update has been loaded. Requesting the signature
does not have a negative impact upon a loaded update.
The procedure for determining this signature shown in Example 9-9.
Example 9-9. Assembly Code to Retrieve the Update Revision
MOV ECX, 08BH
XOR EAX, EAX
XOR EDX, EDX
WRMSR
;IA32_BIOS_SIGN_ID
;clear EAX
;clear EDX
;Load 0 to MSR at 8BH
MOV EAX, 1
cpuid
MOV ECX, 08BH
rdmsr
;IA32_BIOS_SIGN_ID
;Read Model Specific Register
If there is an update active in the processor, its revision is returned in the EDX
register after the RDMSR instruction executes.
IA32_BIOS_SIGN_ID
MSR Address:
Default Value:
Access:
Microcode Update Signature Register
08BH Accessed as a Qword
XXXX XXXX XXXX XXXXh
Read/Write
The IA32_BIOS_SIGN_ID register is used to report the microcode update signature
when CPUID executes. The signature is returned in the upper DWORD (Table 9-11).
Table 9-11. Microcode Update Signature
Bit
Description
63:32
Microcode update signature. This field contains the signature of the currently loaded
microcode update when read following the execution of the CPUID instruction, function
1. It is required that this register field be pre-loaded with zero prior to executing the
CPUID, function 1. If the field remains equal to zero, then there is no microcode update
loaded. Another non-zero value will be the signature.
31:0
Reserved.
9.11.7.2 Authenticating the Update
An update may be authenticated by the BIOS using the signature primitive,
described above, and the algorithm in Example 9-10.
9-48 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
Example 9-10. Pseudo Code to Authenticate the Update
Z ← Obtain Update Revision from the Update Header to be authenticated;
X ← Obtain Current Update Signature from MSR 8BH;
If (Z > X)
{
Load Update that is to be authenticated;
Y ← Obtain New Signature from MSR 8BH;
If (Z == Y)
Success
Else
Fail
}
Else
Fail
Example 9-10 requires that the BIOS only authenticate updates that contain a
numerically larger revision than the currently loaded revision, where Current Signa-
ture (X) < New Update Revision (Z). A processor with no loaded update is considered
to have a revision equal to zero.
This authentication procedure relies upon the decoding provided by the processor to
verify an update from a potentially hostile source. As an example, this mechanism in
conjunction with other safeguards provides security for dynamically incorporating
field updates into the BIOS.
9.11.8
Pentium 4, Intel Xeon, and P6 Family Processor
Microcode Update Specifications
This section describes the interface that an application can use to dynamically inte-
grate processor-specific updates into the system BIOS. In this discussion, the appli-
cation is referred to as the calling program or caller.
The real mode INT15 call specification described here is an Intel extension to an OEM
BIOS. This extension allows an application to read and modify the contents of the
microcode update data in NVRAM. The update loader, which is part of the system
BIOS, cannot be updated by the interface. All of the functions defined in the specifi-
cation must be implemented for a system to be considered compliant with the speci-
fication. The INT15 functions are accessible only from real mode.
9.11.8.1 Responsibilities of the BIOS
If a BIOS passes the presence test (INT 15H, AX = 0D042H, BL = 0H), it must imple-
ment all of the sub-functions defined in the INT 15H, AX = 0D042H specification.
Vol. 3 9-49
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
There are no optional functions. BIOS must load the appropriate update for each
processor during system initialization.
A Header Version of an update block containing the value 0FFFFFFFFH indicates that
the update block is unused and available for storing a new update.
The BIOS is responsible for providing a region of non-volatile storage (NVRAM) for
each potential processor stepping within a system. This storage unit consists of one
or more update blocks. An update block is a contiguous 2048-byte block of memory.
The BIOS for a single processor system need only provide update blocks to store one
microcode update. If the BIOS for a multiple processor system is intended to support
mixed processor steppings, then the BIOS needs to provide enough update blocks to
store each unique microcode update or for each processor socket on the OEM’s
system board.
The BIOS is responsible for managing the NVRAM update blocks. This includes
garbage collection, such as removing microcode updates that exist in NVRAM for
which a corresponding processor does not exist in the system. This specification only
provides the mechanism for ensuring security, the uniqueness of an entry, and that
stale entries are not loaded. The actual update block management is implementation
specific on a per-BIOS basis.
As an example, the BIOS may use update blocks sequentially in ascending order with
CPU signatures sorted versus the first available block. In addition, garbage collection
may be implemented as a setup option to clear all NVRAM slots or as BIOS code that
searches and eliminates unused entries during boot.
NOTES
For IA-32 processors starting with family 0FH and model 03H and
Intel 64 processors, the microcode update may be as large as 16
KBytes. Thus, BIOS must allocate 8 update blocks for each microcode
update. In a MP system, a common microcode update may be
sufficient for each socket in the system.
microcode update is 2 KBytes. An MP-capable BIOS that supports
multiple steppings must allocate a block for each socket in the system.
A single-processor BIOS that supports variable-sized microcode
update and fixed-sized microcode update must allocate one 16-KByte
region and a second region of at least 2 KBytes.
The following algorithm (Example 9-11) describes the steps performed during BIOS
initialization used to load the updates into the processor(s). The algorithm assumes:
• The BIOS ensures that no update contained within NVRAM has a header version
or loader version that does not match one currently supported by the BIOS.
• The update contains a correct checksum.
• The BIOS ensures that (at most) one update exists for each processor stepping.
• Older update revisions are not allowed to overwrite more recent ones.
9-50 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
These requirements are checked by the BIOS during the execution of the write
update function of this interface. The BIOS sequentially scans through all of the
update blocks in NVRAM starting with index 0. The BIOS scans until it finds an update
where the processor fields in the header match the processor signature (extended
family, extended model, type, family, model, and stepping) as well as the platform
bits of the current processor.
Example 9-11. Pseudo Code, Checks Required Prior to Loading an Update
For each processor in the system
{
Determine the Processor Signature via CPUID function 1;
Determine the Platform Bits ← 1 << IA32_PLATFORM_ID[52:50];
For (I ← UpdateBlock 0, I < NumOfBlocks; I++)
{
If (Update.Header_Version == 0x00000001)
{
If ((Update.ProcessorSignature == Processor Signature) &&
(Update.ProcessorFlags & Platform Bits))
{
Load Update.UpdateData into the Processor;
Verify update was correctly loaded into the processor
Go on to next processor
Break;
}
Else If (Update.TotalSize > (Update.DataSize + 48))
{
N ← 0
While (N < Update.ExtendedSignatureCount)
{
If ((Update.ProcessorSignature[N] ==
Processor Signature) &&
(Update.ProcessorFlags[N] & Platform Bits))
{
Load Update.UpdateData into the Processor;
Verify update correctly loaded into the processor
Go on to next processor
Break;
}
N ← N + 1
}
I ← I + (Update.TotalSize / 2048)
If ((Update.TotalSize MOD 2048) == 0)
I ← I + 1
}
}
Vol. 3 9-51
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
}
}
NOTES
The platform Id bits in IA32_PLATFORM_ID are encoded as a three-
bit binary coded decimal field. The platform bits in the microcode
update header are individually bit encoded. The algorithm must do a
translation from one format to the other prior to doing a check.
When performing the INT 15H, 0D042H functions, the BIOS must assume that the
caller has no knowledge of platform specific requirements. It is the responsibility of
BIOS calls to manage all chipset and platform specific prerequisites for managing the
NVRAM device. When writing the update data using the Write Update sub-function,
the BIOS must maintain implementation specific data requirements (such as the
update of NVRAM checksum). The BIOS should also attempt to verify the success of
write operations on the storage device used to record the update.
9.11.8.2 Responsibilities of the Calling Program
This section of the document lists the responsibilities of a calling program using the
interface specifications to load microcode update(s) into BIOS NVRAM.
• The calling program should call the INT 15H, 0D042H functions from a pure real
mode program and should be executing on a system that is running in pure real
mode.
• The caller should issue the presence test function (sub function 0) and verify the
signature and return codes of that function.
• It is important that the calling program provides the required scratch RAM buffers
for the BIOS and the proper stack size as specified in the interface definition.
• The calling program should read any update data that already exists in the BIOS
in order to make decisions about the appropriateness of loading the update. The
BIOS must refuse to overwrite a newer update with an older version. The update
header contains information about version and processor specifics for the calling
program to make an intelligent decision about loading.
• There can be no ambiguous updates. The BIOS must refuse to allow multiple
updates for processors that don’t exist on the system.
• The calling application should implement a verify function that is run after the
update write function successfully completes. This function reads back the
update and verifies that the BIOS returned an image identical to the one that was
written.
Example 9-12 represents a calling program.
9-52 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
Example 9-12. INT 15 DO42 Calling Program Pseudo-code
//
// We must be in real mode
//
If the system is not in Real mode exit
//
// Detect presence of Genuine Intel processor(s) that can be updated
// using(CPUID)
//
If no Intel processors exist that can be updated exit
//
// Detect the presence of the Intel microcode update extensions
//
If the BIOS fails the PresenceTestexit
//
// If the APIC is enabled, see if any other processors are out there
//
Read IA32_APICBASE
If APIC enabled
{
Send Broadcast Message to all processors except self via APIC
Have all processors execute CPUID, record the Processor Signature
(i.e.,Extended Family, Extended Model, Type, Family, Model,
Stepping)
Have all processors read IA32_PLATFORM_ID[52:50], record Platform
Id Bits
If current processor cannot be updated
exit
}
//
// Determine the number of unique update blocks needed for this system
//
NumBlocks = 0
For each processor
{
If ((this is a unique processor stepping) AND
(we have a unique update in the database for this processor))
{
Checksum the update from the database;
If Checksum fails
exit
NumBlocks ← NumBlocks + size of microcode update / 2048
}
}
//
Vol. 3 9-53
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
// Do we have enough update slots for all CPUs?
//
If there are more blocks required to support the unique processor
steppings than update blocks provided by the BIOS exit
//
// Do we need any update blocks at all? If not, we are done
//
If (NumBlocks == 0)
exit
//
// Record updates for processors in NVRAM.
//
For (I=0; I<NumBlocks; I++)
{
//
// Load each Update
//
Issue the WriteUpdate function
If (STORAGE_FULL) returned
{
Display Error -- BIOS is not managing NVRAM appropriately
exit
}
If (INVALID_REVISION) returned
{
Display Message: More recent update already loaded in NVRAM for
this stepping
continue
}
If any other error returned
{
Display Diagnostic
exit
}
//
// Verify the update was loaded correctly
//
Issue the ReadUpdate function
If an error occurred
{
Display Diagnostic
exit
9-54 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
}
//
// Compare the Update read to that written
//
If (Update read != Update written)
{
Display Diagnostic
exit
}
I ← I + (size of microcode update / 2048)
}
//
//
Issue the Update Control function with Task = Enable.
9.11.8.3 Microcode Update Functions
Table 9-12 defines current Pentium 4, Intel Xeon, and P6 family processor microcode
update functions.
Table 9-12. Microcode Update Functions
Microcode Update
Function
Function Description
Number
Required/Optional
Required
Presence test
00H
Returns information about the
supported functions.
Write update data
01H
Writes one of the update data areas
(slots).
Required
Update control
02H
03H
Globally controls the loading of updates. Required
Read update data
Reads one of the update data areas
(slots).
Required
9.11.8.4 INT 15H-based Interface
Intel recommends that a BIOS interface be provided that allows additional microcode
updates to be added to system flash. The INT15H interface is the Intel-defined
method for doing this.
The program that calls this interface is responsible for providing three 64-kilobyte
RAM areas for BIOS use during calls to the read and write functions. These RAM
scratch pads can be used by the BIOS for any purpose, but only for the duration of
the function call. The calling routine places real mode segments pointing to the RAM
blocks in the CX, DX and SI registers. Calls to functions in this interface must be
made with a minimum of 32 kilobytes of stack available to the BIOS.
Vol. 3 9-55
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
In general, each function returns with CF cleared and AH contains the returned
status. The general return codes and other constant definitions are listed in Section
9.11.8.9, “Return Codes.”
The OEM error field (AL) is provided for the OEM to return additional error informa-
tion specific to the platform. If the BIOS provides no additional information about the
error, OEM error must be set to SUCCESS. The OEM error field is undefined if AH
contains either SUCCESS (00H) or NOT_IMPLEMENTED (86H). In all other cases, it
The following sections describe functions provided by the INT15H-based interface.
9.11.8.5 Function 00H—Presence Test
This function verifies that the BIOS has implemented required microcode update
functions. Table 9-13 lists the parameters and return codes for the function.
Table 9-13. Parameters for the Presence Test
Input
AX
Function Code
Sub-function
0D042H
BL
00H - Presence test
Output
CF
Carry Flag
Carry Set - Failure - AH contains status
Carry Clear - All return values valid
AH
AL
Return Code
OEM Error
Additional OEM information.
EBX
ECX
EDX
SI
Signature Part 1
Signature Part 2
Loader Version
Update Count
'INTE' - Part one of the signature
'LPEP'- Part two of the signature
Version number of the microcode update loader
Number of 2048 update blocks in NVRAM the BIOS
allocated to storing microcode updates
Return Codes (see Table 9-18 for code definitions
SUCCESS
The function completed successfully.
The function is not implemented.
NOT_IMPLEMENTED
In order to assure that the BIOS function is present, the caller must verify the carry
flag, the return code, and the 64-bit signature. The update count reflects the number
of 2048-byte blocks available for storage within one non-volatile RAM.
The loader version number refers to the revision of the update loader program that is
included in the system BIOS image.
9-56 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
9.11.8.6 Function 01H—Write Microcode Update Data
This function integrates a new microcode update into the BIOS storage device. Table
9-14 lists the parameters and return codes for the function.
Table 9-14. Parameters for the Write Update Data Function
Input
AX
Function Code
Sub-function
0D042H
BL
01H - Write update
ES:DI
Update Address
Real Mode pointer to the Intel Update structure. This
buffer is 2048 bytes in length if the processor supports
only fixed-size microcode update or...
Real Mode pointer to the Intel Update structure. This
buffer is 64 KBytes in length if the processor supports a
variable-size microcode update.
CX
Scratch Pad1
Scratch Pad2
Scratch Pad3
Stack pointer
Real mode segment address of 64 KBytes of RAM block
Real mode segment address of 64 KBytes of RAM block
Real mode segment address of 64 KBytes of RAM block
32 KBytes of stack minimum
DX
SI
SS:SP
Output
CF
Carry Flag
Carry Set - Failure - AH Contains status
Carry Clear - All return values valid
AH
AL
Return Code
OEM Error
Status of the call
Additional OEM information
Return Codes (see Table 9-18 for code definitions
SUCCESS
The function completed successfully.
NOT_IMPLEMENTED
WRITE_FAILURE
The function is not implemented.
A failure occurred because of the inability to write the
storage device.
ERASE_FAILURE
READ_FAILURE
STORAGE_FULL
A failure occurred because of the inability to erase the
storage device.
A failure occurred because of the inability to read the
storage device.
The BIOS non-volatile storage area is unable to
accommodate the update because all available update
blocks are filled with updates that are needed for
processors in the system.
Vol. 3 9-57
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
Table 9-14. Parameters for the Write Update Data Function (Contd.)
Input
CPU_NOT_PRESENT
The processor stepping does not currently exist in the
system.
INVALID_HEADER
The update header contains a header or loader version
that is not recognized by the BIOS.
INVALID_HEADER_CS
SECURITY_FAILURE
INVALID_REVISION
The update does not checksum correctly.
The processor rejected the update.
The same or more recent revision of the update exists in
the storage device.
Description
The BIOS is responsible for selecting an appropriate update block in the non-volatile
storage for storing the new update. This BIOS is also responsible for ensuring the
integrity of the information provided by the caller, including authenticating the
proposed update before incorporating it into storage.
Before writing the update block into NVRAM, the BIOS should ensure that the update
structure meets the following criteria in the following order:
1. The update header version should be equal to an update header version
recognized by the BIOS.
2. The update loader version in the update header should be equal to the update
loader version contained within the BIOS image.
3. The update block must checksum. This checksum is computed as a 32-bit
summation of all double words in the structure, including the header, data, and
processor signature table.
The BIOS selects update block(s) in non-volatile storage for storing the candidate
update. The BIOS can select any available update block as long as it guarantees that
only a single update exists for any given processor stepping in non-volatile storage.
If the update block selected already contains an update, the following additional
criteria apply to overwrite it:
• The processor signature in the proposed update must be equal to the processor
signature in the header of the current update in NVRAM (Processor Signature +
platform ID bits).
• The update revision in the proposed update should be greater than the update
revision in the header of the current update in NVRAM.
If no unused update blocks are available and the above criteria are not met, the BIOS
can overwrite update block(s) for a processor stepping that is no longer present in
the system. This can be done by scanning the update blocks and comparing the
processor steppings, identified in the MP Specification table, to the processor step-
pings that currently exist in the system.
9-58 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
Finally, before storing the proposed update in NVRAM, the BIOS must verify the
authenticity of the update via the mechanism described in Section 9.11.6, “Micro-
code Update Loader.” This includes loading the update into the current processor,
executing the CPUID instruction, reading MSR 08Bh, and comparing a calculated
value with the update revision in the proposed update header for equality.
When performing the write update function, the BIOS must record the entire update,
applicable). When writing an update, the original contents may be overwritten,
assuming the above criteria have been met. It is the responsibility of the BIOS to
ensure that more recent updates are not overwritten through the use of this BIOS
call, and that only a single update exists within the NVRAM for any processor step-
ping and platform ID.
Figure 9-8 and Figure 9-9 show the process the BIOS follows to choose an update
block and ensure the integrity of the data when it stores the new microcode update.
Vol. 3 9-59
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
Write Microcode Update
Return
CPU_NOT_PRESENT
Does Update Match A
CPU in The System
No
Yes
Return
INVALID_HEADER
Valid Update
Header Version?
No
No
No
Yes
Return
INVALID_HEADER
Loader Revision Match
BIOS’s Loader?
Yes
Return
INVALID_HEADER_CS
Does Update
ChecksumCorrectly?
1
Figure 9-8. Microcode Update Write Operation Flow [1]
9-60 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
1
Replacement
Update Matching CPU
Already In NVRAM?
Space Available in
No
No
policy implemented?
NVRAM?
Yes
No
Yes
Yes
Update Revision Newer
Than NVRAM Update?
Return
INVALID_REVISION
Return
STORAGE_FULL
No
Yes
Update Pass
Return
Authenticity Test?
SECURITY_FAILURE
Yes
Update NMRAM Record
Return
SUCCESS
Figure 9-9. Microcode Update Write Operation Flow [2]
Vol. 3 9-61
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
9.11.8.7 Function 02H—Microcode Update Control
This function enables loading of binary updates into the processor. Table 9-15 lists
the parameters and return codes for the function.
Table 9-15. Parameters for the Control Update Sub-function
Input
AX
Function Code
Sub-function
Task
0D042H
BL
02H - Control update
BH
See the description below.
CX
Scratch Pad1
Scratch Pad2
Scratch Pad3
Stack pointer
Real mode segment of 64 KBytes of RAM block
Real mode segment of 64 KBytes of RAM block
Real mode segment of 64 KBytes of RAM block
32 kilobytes of stack minimum
DX
SI
SS:SP
Output
CF
Carry Flag
Carry Set - Failure - AH contains status
Carry Clear - All return values valid.
AH
AL
BL
Return Code
OEM Error
Status of the call
Additional OEM Information.
Either enable or disable indicator
Update Status
Return Codes (see Table 9-18 for code definitions)
SUCCESS
Function completed successfully.
READ_FAILURE
A failure occurred because of the inability to read the
storage device.
This control is provided on a global basis for all updates and processors. The caller
can determine the current status of update loading (enabled or disabled) without
changing the state. The function does not allow the caller to disable loading of binary
updates, as this poses a security risk.
The caller specifies the requested operation by placing one of the values from Table
9-16 in the BH register. After successfully completing this function, the BL register
contains either the enable or the disable designator. Note that if the function fails, the
update status return value is undefined.
9-62 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
Table 9-16. Mnemonic Values
Mnemonic
Enable
Value
Meaning
1
2
Enable the Update loading at initialization time.
Query
Determine the current state of the update control without
changing its status.
The READ_FAILURE error code returned by this function has meaning only if the
control function is implemented in the BIOS NVRAM. The state of this feature
failure errors cannot occur.
9.11.8.8 Function 03H—Read Microcode Update Data
This function reads a currently installed microcode update from the BIOS storage into
a caller-provided RAM buffer. Table 9-17 lists the parameters and return codes.
Table 9-17. Parameters for the Read Microcode Update Data Function
Input
AX
Function Code
Sub-function
Buffer Address
0D042H
BL
03H - Read Update
ES:DI
Real Mode pointer to the Intel Update
structure that will be written with the
binary data
ECX
ECX
DX
Scratch Pad1
Scratch Pad2
Scratch Pad3
Real Mode Segment address of 64
KBytes of RAM Block (lower 16 bits)
Real Mode Segment address of 64
KBytes of RAM Block (upper 16 bits)
Real Mode Segment address of 64
KBytes of RAM Block
SS:SP
SI
Stack pointer
32 KBytes of Stack Minimum
Update Number
This is the index number of the update
block to be read. This value is zero based
and must be less than the update count
returned from the presence test
function.
Output
CF
Carry Flag
Carry Set - Failure - AH contains Status
Status of the Call
Carry Clear - All return
values are valid.
AH
Return Code
Vol. 3 9-63
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
Table 9-17. Parameters for the Read Microcode Update Data Function (Contd.)
AL
OEM Error
Additional OEM Information
Return Codes (see Table 9-18 for code definitions)
SUCCESS
The function completed successfully.
READ_FAILURE
There was a failure because of the
inability to read the storage device.
UPDATE_NUM_INVALID
NOT_EMPTY
Update number exceeds the maximum
number of update blocks implemented
by the BIOS.
The specified update block is a
subsequent block in use to store a valid
microcode update that spans multiple
blocks.
The specified block is not a header block
and is not empty.
The read function enables the caller to read any microcode update data that already
exists in a BIOS and make decisions about the addition of new updates. As a result
of a successful call, the BIOS copies the microcode update into the location pointed
to by ES:DI, with the contents of all Update block(s) that are used to store the spec-
ified microcode update.
If the specified block is not a header block, but does contain valid data from a micro-
code update that spans multiple update blocks, then the BIOS must return Failure
with the NOT_EMPTY error code in AH.
An update block is considered unused and available for storing a new update if its
Header Version contains the value 0FFFFFFFFH after return from this function call.
The actual implementation of NVRAM storage management is not specified here and
empty block by the BIOS may be zero, rather than 0FFFFFFFFH. The BIOS is respon-
sible for translating this information into the header provided by this function.
9.11.8.9 Return Codes
After the call has been made, the return codes listed in Table 9-18 are available in the
AH register.
9-64 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
Table 9-18. Return Code Definitions
Return Code
Value
00H
86H
90H
Description
SUCCESS
The function completed successfully.
The function is not implemented.
NOT_IMPLEMENTED
ERASE_FAILURE
A failure because of the inability to erase the storage
device.
WRITE_FAILURE
READ_FAILURE
STORAGE_FULL
91H
92H
93H
A failure because of the inability to write the storage
device.
A failure because of the inability to read the storage
device.
The BIOS non-volatile storage area is unable to
accommodate the update because all available update
blocks are filled with updates that are needed for
processors in the system.
CPU_NOT_PRESENT
INVALID_HEADER
94H
95H
The processor stepping does not currently exist in the
system.
The update header contains a header or loader version
that is not recognized by the BIOS.
INVALID_HEADER_CS
SECURITY_FAILURE
INVALID_REVISION
96H
97H
98H
The update does not checksum correctly.
The update was rejected by the processor.
The same or more recent revision of the update exists
in the storage device.
UPDATE_NUM_INVALID
NOT_EMPTY
99H
9AH
The update number exceeds the maximum number of
update blocks implemented by the BIOS.
The specified update block is a subsequent block in use
to store a valid microcode update that spans multiple
blocks.
The specified block is not a header block and is not
empty.
Vol. 3 9-65
Download from Www.Somanuals.com. All Manuals Search And Download.
PROCESSOR MANAGEMENT AND INITIALIZATION
9-66 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
CHAPTER 10
The Advanced Programmable Interrupt Controller (APIC), referred to in the following
sections as the local APIC, was introduced into the IA-32 processors with the Pentium
processor (see Section 19.27, “Advanced Programmable Interrupt Controller
(APIC)”) and is included in the P6 family, Pentium 4, Intel Xeon processors, and other
more recent Intel 64 and IA-32 processor families (see Section 10.4.2, “Presence of
the Local APIC”). The local APIC performs two primary functions for the processor:
• It receives interrupts from the processor’s interrupt pins, from internal sources
and from an external I/O APIC (or other external interrupt controller). It sends
these to the processor core for handling.
• In multiple processor (MP) systems, it sends and receives interprocessor
interrupt (IPI) messages to and from other logical processors on the system bus.
IPI messages can be used to distribute interrupts among the processors in the
system or to execute system wide functions (such as, booting up processors or
distributing work among a group of processors).
The external I/O APIC is part of Intel’s system chip set. Its primary function is to
receive external interrupt events from the system and its associated I/O devices and
relay them to the local APIC as interrupt messages. In MP systems, the I/O APIC also
provides a mechanism for distributing external interrupts to the local APICs of
This chapter provides a description of the local APIC and its programming interface.
It also provides an overview of the interface between the local APIC and the I/O
APIC. Contact Intel for detailed information about the I/O APIC.
When a local APIC has sent an interrupt to its processor core for handling, the
processor uses the interrupt and exception handling mechanism described in Chapter
Overview,” for an introduction to interrupt and exception handling.
10.1
LOCAL AND I/O APIC OVERVIEW
Each local APIC consists of a set of APIC registers (see Table 10-1) and associated
hardware that control the delivery of interrupts to the processor core and the gener-
ation of IPI messages. The APIC registers are memory mapped and can be read and
written to using the MOV instruction.
Local APICs can receive interrupts from the following sources:
• Locally connected I/O devices — These interrupts originate as an edge or
level asserted by an I/O device that is connected directly to the processor’s local
Vol. 3 10-1
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
interrupt pins (LINT0 and LINT1). The I/O devices may also be connected to an
8259-type interrupt controller that is in turn connected to the processor through
one of the local interrupt pins.
• Externally connected I/O devices — These interrupts originate as an edge or
level asserted by an I/O device that is connected to the interrupt input pins of an
I/O APIC. Interrupts are sent as I/O interrupt messages from the I/O APIC to one
or more of the processors in the system.
the IPI mechanism to interrupt another processor or group of processors on the
system bus. IPIs are used for software self-interrupts, interrupt forwarding, or
preemptive scheduling.
• APIC timer generated interrupts — The local APIC timer can be programmed
to send a local interrupt to its associated processor when a programmed count is
reached (see Section 10.6.4, “APIC Timer”).
Intel Xeon processors provide the ability to send an interrupt to its associated
processor when a performance-monitoring counter overflows (see Section
30.8.5.8, “Generating an Interrupt on Overflow”).
ability to send an interrupt to themselves when the internal thermal sensor has
been tripped (see Section 14.5.2, “Thermal Monitor”).
• APIC internal error interrupts — When an error condition is recognized within
the local APIC (such as an attempt to access an unimplemented register), the
APIC can be programmed to send an interrupt to its associated processor (see
Section 10.6.3, “Error Handling”).
Of these interrupt sources: the processor’s LINT0 and LINT1 pins, the APIC timer, the
performance-monitoring counters, the thermal sensor, and the internal APIC error
detector are referred to as local interrupt sources. Upon receiving a signal from a
local interrupt source, the local APIC delivers the interrupt to the processor core
using an interrupt delivery protocol that has been set up through a group of APIC
registers called the local vector table or LVT (see Section 10.6.1, “Local Vector
Table”). A separate entry is provided in the local vector table for each local interrupt
source, which allows a specific interrupt delivery protocol to be set up for each
source. For example, if the LINT1 pin is going to be used as an NMI pin, the LINT1
entry in the local vector table can be set up to deliver an interrupt with vector number
2 (NMI interrupt) to the processor core.
The local APIC handles interrupts from the other two interrupt sources (externally
connected I/O devices and IPIs) through its IPI message handling facilities.
A processor can generate IPIs by programming the interrupt command register (ICR)
in its local APIC (see Section 10.7.1, “Interrupt Command Register (ICR)”). The act
of writing to the ICR causes an IPI message to be generated and issued on the
system bus (for Pentium 4 and Intel Xeon processors) or on the APIC bus (for
Pentium and P6 family processors). See Section 10.2, “System Bus Vs. APIC Bus.”
10-2 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
IPIs can be sent to other processors in the system or to the originating processor
(self-interrupts). When the target processor receives an IPI message, its local APIC
handles the message automatically (using information included in the message such
as vector number and trigger mode). See Section 10.7, “Issuing Interprocessor
Interrupts,” for a detailed explanation of the local APIC’s IPI message delivery and
acceptance mechanism.
The local APIC can also receive interrupts from externally connected devices through
the I/O APIC (see Figure 10-1). The I/O APIC is responsible for receiving interrupts
generated by system hardware and I/O devices and forwarding them to the local
APIC as interrupt messages.
Pentium 4 and
Pentium and P6
Intel Xeon Processors
Family Processors
Processor Core
Local APIC
Processor Core
Local APIC
Interrupt
Messages
Local
Interrupts
Local
Interrupts
Interrupt
Messages
Interrupt
Messages
System Bus
3-Wire APIC Bus
Bridge
External
Interrupts
I/O APIC
PCI
System Chip Set
External
Interrupts
I/O APIC
System Chip Set
Individual pins on the I/O APIC can be programmed to generate a specific interrupt
communicate with a standard 8259A-style external interrupt controller. Note that the
local APIC can be disabled (see Section 10.4.3, “Enabling or Disabling the Local
APIC”). This allows an associated processor core to receive interrupts directly from
an 8259A interrupt controller.
Both the local APIC and the I/O APIC are designed to operate in MP systems (see
Figures 10-2 and 10-3). Each local APIC handles interrupts from the I/O APIC, IPIs
from processors on the system bus, and self-generated interrupts. Interrupts can
Vol. 3 10-3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
also be delivered to the individual processors through the local interrupt pins;
however, this mechanism is commonly not used in MP systems.
Processor #1
CPU
Processor #2
CPU
Processor #3
CPU
Processor #3
CPU
Local APIC
Local APIC
Local APIC
Local APIC
Interrupt
Messages
Interrupt
Messages
Interrupt
Messages
Interrupt
Messages
IPIs
IPIs
IPIs
IPIs
Processor System Bus
Interrupt
Messages
Bridge
PCI
External
Interrupts
I/O APIC
System Chip Set
Figure 10-2. Local APICs and I/O APIC When Intel Xeon Processors Are Used in
Multiple-Processor Systems
Processor #1
CPU
Processor #2
CPU
Processor #3
CPU
Processor #4
CPU
Local APIC
Local APIC
Local APIC
Local APIC
Interrupt
Messages
Interrupt
Messages
IPIs
IPIs
Interrupt
Messages
IPIs
Interrupt
Messages
IPIs
3-wire APIC Bus
Interrupt
Messages
External
Interrupts
I/O APIC
System Chip Set
Figure 10-3. Local APICs and I/O APIC When P6 Family Processors Are Used in
Multiple-Processor Systems
10-4 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
The IPI mechanism is typically used in MP systems to send fixed interrupts (inter-
rupts for a specific vector number) and special-purpose interrupts to processors on
the system bus. For example, a local APIC can use an IPI to forward a fixed interrupt
to another processor for servicing. Special-purpose IPIs (including NMI, INIT, SMI
The following sections focus on the local APIC and its implementation in the
Pentium 4, Intel Xeon, and P6 family processors. In these sections, the terms “local
APIC” and “I/O APIC” refer to local and I/O APICs used with the P6 family processors
and to local and I/O xAPICs used with the Pentium 4 and Intel Xeon processors (see
®
Section 10.3, “the Intel 82489DX External APIC, The APIC, the xAPIC, AND THE
X2APIC”).
10.2
SYSTEM BUS VS. APIC BUS
through the 3-wire inter-APIC bus (see Figure 10-3). Local APICs also use the APIC
bus to send and receive IPIs. The APIC bus and its messages are invisible to software
and are not classed as architectural.
Beginning with the Pentium 4 and Intel Xeon processors, the I/O APIC and local
APICs (using the xAPIC architecture) communicate through the system bus (see
Figure 10-2). The I/O APIC sends interrupt requests to the processors on the system
bus through bridge hardware that is part of the Intel chip set. The bridge hardware
generates the interrupt messages that go to the local APICs. IPIs between local
APICs are transmitted directly on the system bus.
10.3
THE INTEL® 82489DX EXTERNAL APIC,
THE APIC, THE XAPIC, AND THE X2APIC
The local APIC in the P6 family and Pentium processors is an architectural subset of
®
the Intel 82489DX external APIC. See Section 19.27.1, “Software Visible Differ-
The APIC architecture used in the Pentium 4 and Intel Xeon processors (called the
xAPIC architecture) is an extension of the APIC architecture found in the P6 family
processors. The primary difference between the APIC and xAPIC architectures is that
with the xAPIC architecture, the local APICs and the I/O APIC communicate through
the system bus. With the APIC architecture, they communication through the APIC
bus (see Section 10.2, “System Bus Vs. APIC Bus”). Also, some APIC architectural
features have been extended and/or modified in the xAPIC architecture.
The x2APIC architecture is extended from the xAPIC architecture. Extensions to the
xAPIC architecture are intended primarily to increase processor addressability. The
x2APIC architecture provides backward compatibility to the xAPIC architecture and
Vol. 3 10-5
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
forward extendability for future Intel platform innovations. These extensions and
modifications are noted in the following sections.
10.4
LOCAL APIC
The following sections describe the architecture of the local APIC and how to detect
are given in Section 10.6.1, “Local Vector Table,” and Section 10.7.1, “Interrupt
Command Register (ICR).”
10.4.1
The Local APIC Block Diagram
Figure 10-4 gives a functional block diagram for the local APIC. Software interacts
with the local APIC by reading and writing its registers. APIC registers are memory-
mapped to a 4-KByte region of the processor’s physical address space with an initial
starting address of FEE00000H. For correct APIC operation, this address space must
be mapped to an area of memory that has been designated as strong uncacheable
In MP system configurations, the APIC registers for Intel 64 or IA-32 processors on
the system bus are initially mapped to the same 4-KByte region of the physical
APIC to its own 4-KByte region. Section 10.4.5, “Relocating the Local APIC Regis-
ters,” describes how to relocate the base address for APIC registers.
On processors supporting x2APIC architecture (indicated by CPUID.01H:ECX[21]
=1), the local APIC supports operation in the xAPIC mode (as described in Section
10.4. Additionally, software can enable the local APIC to operate in x2APIC mode for
extended processor addressability (see Section 10.5).
NOTE
For P6 family, Pentium 4, and Intel Xeon processors, the APIC
handles all memory accesses to addresses within the 4-KByte APIC
register space internally and no external bus cycles are produced. For
the Pentium processors with an on-chip APIC, bus cycles are
produced for accesses to the APIC register space. Thus, for software
intended to run on Pentium processors, system software should
explicitly not map the APIC register space to regular system memory.
Doing so can result in an invalid opcode exception (#UD) being
generated or unpredictable execution.
10-6 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
DATA/ADDR
EOI Register
Version Register
Timer
Task Priority Register
Current Count
Register
Processor Priority
Initial Count
Register
Register
From
CPU
Core
INTA
Divide Configuration
Register
INTR
To
Prioritizer
EXTINT CPU
Local Vector Table
Timer
Core
In-Service Register (ISR)
Interrupt Request Register (IRR)
Trigger Mode Register (TMR)
Local
Interrupts 0,1
LINT0/1
Perf. Mon.
(Internal
Interrupt)
Performance
Monitoring Counters1
Thermal
Sensor
(Internal
Interrupt)
Thermal Sensor2
Vec[3:0]
& TMR Bit
Register
Select
Error
Arb. ID
Register
Vector
Decode
4
Error Status
Register
Local
Interrupts
Acceptance
Logic
Dest. Mode
& Vector
To
Protocol
Translation Logic
APIC ID
Register
CPU
INIT
NMI
SMI
Core
Logical Destination
Register
Interrupt Command
Register (ICR)
Destination Format
Register
Processor System Bus3
Spurious Vector
Register
1. Introduced in P6 family processors.
2. Introduced in the Pentium 4 and Intel Xeon processors.
3. Three-wire APIC bus in P6 family and Pentium processors.
4. Not implemented in Pentium 4 and Intel Xeon processors.
Figure 10-4. Local APIC Structure
Vol. 3 10-7
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
Table 10-1 shows how the APIC registers are mapped into the 4-KByte APIC register
space. Registers are 32 bits, 64 bits, or 256 bits in width; all are aligned on 128-bit
boundaries. All 32-bit registers should be accessed using 128-bit aligned 32-bit loads
or stores. Some processors may support loads and stores of less than 32 bits to some
of the APIC registers. This is model specific behavior and is not guaranteed to work
touches bytes 4 through 15 of an APIC register may cause undefined behavior and
or unexpected exceptions, including machine checks, and may vary between imple-
mentations. Wider registers (64-bit or 256-bit) must be accessed using multiple 32-
bit loads or stores, with all accesses being 128-bit aligned.
The local APIC registers listed in Table 10-1 are not MSRs. The only MSR associated
with the programming of the local APIC is the IA32_APIC_BASE MSR (see Section
10.4.3, “Enabling or Disabling the Local APIC”).
NOTE
In processors based on Intel Microarchitecture (Nehalem) the Local
APIC ID Register is no longer Read/Write; it is Read Only.
Table 10-1 Local APIC Register Address Map
Address
Register Name
Software
Read/Write
FEE0 0000H
FEE0 0010H
FEE0 0020H
FEE0 0030H
FEE0 0040H
FEE0 0050H
FEE0 0060H
FEE0 0070H
FEE0 0080H
FEE0 0090H
FEE0 00A0H
FEE0 00B0H
FEE0 00C0H
FEE0 00D0H
FEE0 00E0H
Reserved
Reserved
Local APIC ID Register
Local APIC Version Register
Reserved
Read/Write.
Read Only.
Reserved
Reserved
Reserved
Task Priority Register (TPR)
Read/Write.
Read Only.
Read Only.
Write Only.
Read Only
Read/Write.
1
Arbitration Priority Register (APR)
Processor Priority Register (PPR)
EOI Register
1
Remote Read Register (RRD)
Logical Destination Register
Destination Format Register
Bits 0-27 Read only;
bits 28-31
Read/Write.
10-8 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
Table 10-1 Local APIC Register Address Map (Contd.)
Address
Register Name
Software
Read/Write
FEE0 00F0H
Spurious Interrupt Vector Register
Bits 0-8 Read/Write;
bits 9-31 Read Only.
FEE0 0100H
FEE0 0110H
FEE0 0120H
FEE0 0130H
FEE0 0140H
FEE0 0150H
FEE0 0160H
FEE0 0170H
FEE0 0180H
FEE0 0190H
FEE0 01A0H
FEE0 01B0H
FEE0 01C0H
FEE0 01D0H
FEE0 01E0H
FEE0 01F0H
FEE0 0200H
FEE0 0210H
FEE0 0220H
FEE0 0230H
FEE0 0240H
FEE0 0250H
FEE0 0260H
FEE0 0270H
FEE0 0280H
In-Service Register (ISR); bits 0:31
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
In-Service Register (ISR); bits 32:63
In-Service Register (ISR); bits 64:95
In-Service Register (ISR); bits 96:127
In-Service Register (ISR); bits 128:159
In-Service Register (ISR); bits 160:191
In-Service Register (ISR); bits 192:223
In-Service Register (ISR); bits 224:255
Trigger Mode Register (TMR); bits 0:31
Trigger Mode Register (TMR); bits 32:63
Trigger Mode Register (TMR); bits 64:95
Trigger Mode Register (TMR); bits 96:127
Trigger Mode Register (TMR); bits 128:159
Trigger Mode Register (TMR); bits 160:191
Trigger Mode Register (TMR); bits 192:223
Trigger Mode Register (TMR); bits 224:255
Interrupt Request Register (IRR); bits 0:31
Interrupt Request Register (IRR); bits32:63
Interrupt Request Register (IRR); bits 64:95
Interrupt Request Register (IRR); bits 96:127
Interrupt Request Register (IRR); bits 128:159 Read Only.
Interrupt Request Register (IRR); bits 160:191 Read Only.
Interrupt Request Register (IRR); bits 192:223 Read Only.
Interrupt Request Register (IRR); bits 224:255 Read Only.
Error Status Register
Reserved
Read Only.
FEE0 0290H through
FEE0 02E0H
FEE0 02F0H
FEE0 0300H
LVT CMCI Registers
Read/Write.
Read/Write.
Interrupt Command Register (ICR); bits 0-31
Vol. 3 10-9
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
Table 10-1 Local APIC Register Address Map (Contd.)
Register Name Software
Address
Read/Write
FEE0 0310H
FEE0 0320H
FEE0 0330H
FEE0 0340H
Interrupt Command Register (ICR); bits 32-63 Read/Write.
LVT Timer Register
Read/Write.
Read/Write.
Read/Write.
2
LVT Thermal Sensor Register
LVT Performance Monitoring Counters
3
Register
FEE0 0350H
FEE0 0360H
FEE0 0370H
FEE0 0380H
FEE0 0390H
LVT LINT0 Register
Read/Write.
Read/Write.
Read/Write.
Read/Write.
Read Only.
LVT LINT1 Register
LVT Error Register
Initial Count Register (for Timer)
Current Count Register (for Timer)
Reserved
FEE0 03A0H through
FEE0 03D0H
FEE0 03E0H
FEE0 03F0H
NOTES:
Divide Configuration Register (for Timer)
Reserved
Read/Write.
1. Not supported in the Pentium 4 and Intel Xeon processors. The Illegal Register Access bit (7) of
the ESR will not be set when writing to these registers.
2. Introduced in the Pentium 4 and Intel Xeon processors. This APIC register and its associated
function are implementation dependent and may not be present in future IA-32 or Intel 64 pro-
cessors.
3. Introduced in the Pentium Pro processor. This APIC register and its associated function are
implementation dependent and may not be present in future IA-32 or Intel 64 processors.
10.4.2
Presence of the Local APIC
Beginning with the P6 family processors, the presence or absence of an on-chip local
APIC can be detected using the CPUID instruction. When the CPUID instruction is
executed with a source operand of 1 in the EAX register, bit 9 of the CPUID feature
flags returned in the EDX register indicates the presence (set) or absence (clear) of a
local APIC.
10.4.3
Enabling or Disabling the Local APIC
The local APIC can be enabled or disabled in either of two ways:
10-10 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
1. Using the APIC global enable/disable flag in the IA32_APIC_BASE MSR (MSR
address 1BH; see Figure 10-5):
— When IA32_APIC_BASE[11] is 0, the processor is functionally equivalent to
an IA-32 processor without an on-chip APIC. The CPUID feature flag for the
APIC (see Section 10.4.2, “Presence of the Local APIC”) is also set to 0.
— When IA32_APIC_BASE[11] is set to 0, processor APICs based on the 3-wire
APIC bus cannot be generally re-enabled until a system hardware reset. The
3-wire bus loses track of arbitration that would be necessary for complete re-
enabling. Certain APIC functionality can be enabled (for example:
performance and thermal monitoring interrupt generation).
IA32_APIC_BASE[11]. A hardware reset is not required to re-start APIC
IA32_APIC_BASE[11] is cleared.
— When IA32_APIC_BASE[11] is set to 0, prior initialization to the APIC may be
“Local APIC State After Power-Up or Reset.”
register (see Figure 10-31):
— If IA32_APIC_BASE[11] is 1, software can temporarily disable a local APIC at
any time by clearing the APIC software enable/disable flag in the spurious-
interrupt vector register (see Figure 10-31). The state of the local APIC when
in this software-disabled state is described in Section 10.4.7.2, “Local APIC
State After It Has Been Software Disabled.”
— When the local APIC is in the software-disabled state, it can be re-enabled at
any time by setting the APIC software enable/disable flag to 1.
For the Pentium processor, the APICEN pin (which is shared with the PICD1 pin) is
used during power-up or RESET to disable the local APIC.
Note that each entry in the LVT has a mask bit that can be used to inhibit interrupts
from being delivered to the processor from selected local interrupt sources (the
thermal sensor, and/or the internal APIC error detector).
10.4.4
Local APIC Status and Location
The status and location of the local APIC are contained in the IA32_APIC_BASE MSR
(see Figure 10-5). MSR bit functions are described below:
• BSP flag, bit 8 ⎯ Indicates if the processor is the bootstrap processor (BSP).
See Section 8.4, “Multiple-Processor (MP) Initialization.” Following a power-up or
RESET, this flag is set to 1 for the processor selected as the BSP and set to 0 for
the remaining processors (APs).
Vol. 3 10-11
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
• APIC Global Enable flag, bit 11 ⎯ Enables or disables the local APIC (see
Section 10.4.3, “Enabling or Disabling the Local APIC”). This flag is available in
the Pentium 4, Intel Xeon, and P6 family processors. It is not guaranteed to be
available or available at the same location in future Intel 64 or IA-32 processors.
• APIC Base field, bits 12 through 35 ⎯ Specifies the base address of the APIC
registers. This 24-bit value is extended by 12 bits at the low end to form the base
address. This automatically aligns the address on a 4-KByte boundary. Following
a power-up or RESET, the field is set to FEE0 0000H.
1
• Bits 0 through 7, bits 9 and 10, and bits MAXPHYADDR through 63 in the
IA32_APIC_BASE MSR are reserved.
63
MAXPHYADDR
12 11109 8 7
0
Reserved
APIC Base
APIC Base—Base physical address
APIC global enable/disable
BSP—Processor is BSP
Reserved
Figure 10-5. IA32_APIC_BASE MSR (APIC_BASE_MSR in P6 Family)
10.4.5
Relocating the Local APIC Registers
The Pentium 4, Intel Xeon, and P6 family processors permit the starting address of
the APIC registers to be relocated from FEE00000H to another physical address by
modifying the value in the 24-bit base address field of the IA32_APIC_BASE MSR.
This extension of the APIC architecture is provided to help resolve conflicts with
memory maps of existing systems and to allow individual processors in an MP system
to map their APIC registers to different locations in physical memory.
10.4.6
Local APIC ID
At power up, system hardware assigns a unique APIC ID to each local APIC on the
system bus (for Pentium 4 and Intel Xeon processors) or on the APIC bus (for P6
family and Pentium processors). The hardware assigned APIC ID is based on system
topology and includes encoding for socket position and cluster information (see
Figure 8-2).
In MP systems, the local APIC ID is also used as a processor ID by the BIOS and the
operating system. Some processors permit software to modify the APIC ID. However,
the ability of software to modify the APIC ID is processor model specific. Because of
1. The MAXPHYADDR is 36 bits for processors that do not support CPUID leaf 80000008H, or indi-
cated by CPUID.80000008H:EAX[bits 7:0] for processors that support CPUID leaf 80000008H.
10-12 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
this, operating system software should avoid writing to the local APIC ID register. The
value returned by bits 31-24 of the EBX register (when the CPUID instruction is
executed with a source operand value of 1 in the EAX register) is always the Initial
changed the value in the Local APIC ID register.
The processor receives the hardware assigned APIC ID (or Initial APIC ID) by
sampling pins A11# and A12# and pins BR0# through BR3# (for the Pentium 4, Intel
Xeon, and P6 family processors) and pins BE0# through BE3# (for the Pentium
processor). The APIC ID latched from these pins is stored in the APIC ID field of the
local APIC ID register (see Figure 10-6), and is used as the Initial APIC ID for the
processor.
Address: 0FEE0 0020H
Value after reset: 0000 0000H
P6 family and Pentium processors
31 27
24
0
APIC ID
Reserved
Pentium 4 processors, Xeon processors, and later processors
31
24
0
APIC ID
Reserved
MSR Address: 802H
x2APIC Mode
31
0
x2APIC ID
Figure 10-6. Local APIC ID Register
For the P6 family and Pentium processors, the local APIC ID field in the local APIC ID
register is 4 bits. Encodings 0H through EH can be used to uniquely identify 15
different processors connected to the APIC bus. For the Pentium 4 and Intel Xeon
processors, the xAPIC specification extends the local APIC ID field to 8 bits. These
can be used to identify up to 255 processors in the system.
10.4.7
Local APIC State
The following sections describe the state of the local APIC and its registers following
a power-up or RESET, after the local APIC has been software disabled, following an
INIT reset, and following an INIT-deassert message.
Vol. 3 10-13
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
x2APIC will introduce 32-bit ID; see Section 10.5.
10.4.7.1 Local APIC State After Power-Up or Reset
Following a power-up or RESET of the processor, the state of local APIC and its regis-
ters are as follows:
• The following registers are reset to all 0s:
•
•
•
IRR, ISR, TMR, ICR, LDR, and TPR
Timer initial count and timer current count registers
Divide configuration register
• The DFR register is reset to all 1s.
• The LVT register is reset to 0s except for the mask bits; these are set to 1s.
• The local APIC version register is not affected.
• The local APIC ID register is set to a unique APIC ID. (Pentium and P6 family
processors only). The Arb ID register is set to the value in the APIC ID register.
• The spurious-interrupt vector register is initialized to 000000FFH. By setting bit 8
to 0, software disables the local APIC.
• If the processor is the only processor in the system or it is the BSP in an MP
system (see Section 8.4.1, “BSP and AP Processors”); the local APIC will respond
normally to INIT and NMI messages, to INIT# signals and to STPCLK# signals. If
the processor is in an MP system and has been designated as an AP; the local
APIC will respond the same as for the BSP. In addition, it will respond to SIPI
messages. For P6 family processors only, an AP will not respond to a STPCLK#
signal.
10.4.7.2 Local APIC State After It Has Been Software Disabled
When the APIC software enable/disable flag in the spurious interrupt vector register
has been explicitly cleared (as opposed to being cleared during a power up or
RESET), the local APIC is temporarily disabled (see Section 10.4.3, “Enabling or
Disabling the Local APIC”). The operation and response of a local APIC while in this
software-disabled state is as follows:
• The local APIC will respond normally to INIT, NMI, SMI, and SIPI messages.
• Pending interrupts in the IRR and ISR registers are held and require masking or
handling by the CPU.
• The local APIC can still issue IPIs. It is software’s responsibility to avoid issuing
IPIs through the IPI mechanism and the ICR register if sending interrupts
through this mechanism is not desired.
• The reception or transmission of any IPIs that are in progress when the local APIC
is disabled are completed before the local APIC enters the software-disabled
state.
10-14 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
• The mask bits for all the LVT entries are set. Attempts to reset these bits will be
ignored.
• (For Pentium and P6 family processors) The local APIC continues to listen to all
bus messages in order to keep its arbitration ID synchronized with the rest of the
system.
10.4.7.3 Local APIC State After an INIT Reset (“Wait-for-SIPI” State)
An INIT reset of the processor can be initiated in either of two ways:
• By asserting the processor’s INIT# pin.
Upon receiving an INIT through either of these mechanisms, the processor responds
by beginning the initialization process of the processor core and the local APIC. The
state of the local APIC following an INIT reset is the same as it is after a power-up or
hardware RESET, except that the APIC ID and arbitration ID registers are not
affected. This state is also referred to at the “wait-for-SIPI” state (see also: Section
8.4.2, “MP Initialization Protocol Requirements and Restrictions”).
10.4.7.4 Local APIC State After It Receives an INIT-Deassert IPI
Only the Pentium and P6 family processors support the INIT-deassert IPI. An INIT-
tion ID register with the value in the APIC ID register.
10.4.8
Local APIC Version Register
The local APIC contains a hardwired version register. Software can use this register to
identify the APIC version (see Figure 10-7). In addition, the register specifies the
number of entries in the local vector table (LVT) for a specific implementation.
The fields in the local APIC version register are as follows:
Version
The version numbers of the local APIC:
1XH
Local APIC. For Pentium 4 and Intel Xeon
processors, 14H is returned.
82489DX external APIC.
Reserved.
0XH
20H - FFH
Max LVT Entry
Shows the number of LVT entries minus 1. For the Pentium 4 and
Intel Xeon processors (which have 6 LVT entries), the value
returned in the Max LVT field is 5; for the P6 family processors
(which have 5 LVT entries), the value returned is 4; for the
Pentium processor (which has 4 LVT entries), the value returned
is 3.
Vol. 3 10-15
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
31
24 23
16 15
8 7
0
Max. LVT
Entry
Reserved
Reserved
Version
Value after reset: 000N 00VVH
V = Version, N = # of LVT entries minus 1
Address: FEE0 0030H
Figure 10-7. Local APIC Version Register
10.5
EXTENDED XAPIC (X2APIC)
The x2APIC architecture extends the xAPIC architecture (described in Section 9.4) in
a backward compatible manner and provides forward extendability for future Intel
platform innovations. Specifically, x2APIC
• Retains all key elements of compatibility to the xAPIC architecture:
— delivery modes,
— interrupt and processor priorities,
— interrupt sources,
— interrupt destination types;
• Provides extensions to scale processor addressability for both the logical and
physical destination modes;
• Adds new features to enhance performance of interrupt delivery;
• Reduces complexity of logical destination mode interrupt delivery on link based
platform architectures.
• Uses MSR programming interface to access APIC registers in x2APIC mode
instead of memory-mapped interfaces. Memory-mapped interface is supported
when operating in xAPIC mode.
10.5.1
DETECTING AND ENABLING x2APIC
Processor support for x2APIC mode can be detected by executing CPUID with EAX=1
and then checking ECX, bit 21 ECX. If CPUID.(EAX=1):ECX.21 is set , the processor
supports the x2APIC capability and can be placed into the x2APIC mode.
• System software can place the local APIC in the x2APIC mode by setting the
x2APIC mode enable bit (bit 10) in the IA32_APIC_BASE MSR at MSR address
01BH. The layout for the IA32_APIC_BASE MSR is shown in Figure 10-8.
10-16 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
63
36 35
12 11109 8 7
0
Reserved
APIC Base
APIC Base—Base physical address
EN—xAPIC global enable/disable
EXTD—Enable x2APIC mode
BSP—Processor is BSP
Reserved
Figure 10-8. IA32_APIC_BASE MSR Supporting x2APIC
Table 10-2, “x2APIC operating mode configurations” describe the possible combina-
tions of the enable bit (EN - bit 11) and the extended mode bit (EXTD - bit 10) in the
IA32_APIC_BASE MSR.
Table 10-2. x2APIC Operating Mode Configurations
xAPIC global enable
x2APIC enable
(IA32_APIC_BASE[11]) (IA32_APIC_BASE[10]) Description
0
0
1
1
0
1
0
1
local APIC is disabled
Invalid
local APIC is enabled in xAPIC mode
local APIC is enabled in x2APIC mode
Once the local APIC has been switched to x2APIC mode (EN = 1, EXTD = 1),
switching back to xAPIC mode would require system software to disable the local
APIC unit. Specifically, attempting to write a value to the IA32_APIC_BASE MSR that
has (EN= 1, EXTD = 0) when the local APIC is enabled and in x2APIC mode will raise
a GP exception. Once bit 10 in IA32_APIC_BASE MSR is set, the only way to leave
x2APIC mode using IA32_APIC_BASE would require a WRMSR to set both bit 11 and
bit 10 to zero. Section 10.5.6, “x2APIC State Transitions” provides a detailed state
diagram for the state transitions allowed for the local APIC.
10.5.1.1 Instructions to Access APIC Registers
In x2APIC mode, system software uses RDMSR and WRMSR to access the APIC regis-
ters. The MSR addresses for accessing the x2APIC registers are architecturally
defined and specified in Section 10.5.1.2, “APIC Register Address Space”. Executing
the RDMSR instruction with APIC register address specified in ECX returns the
content of bits 0 through 31 of the APIC registers in EAX. Bits 32 through 63 are
returned in register EDX - these bits are reserved if the APIC register being read is a
Vol. 3 10-17
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
32-bit register. Similarly executing the WRMSR instruction with the APIC register
address in ECX, writes bits 0 to 31 of register EAX to bits 0 to 31 of the specified APIC
register. If the register is a 64-bit register then bits 0 to 31 of register EDX are written
to bits 32 to 63 of the APIC register. The Interrupt Command Register is the only APIC
register that is implemented as a 64-bit MSR. The semantics of handling reserved
bits are defined in Section 10.5.1.3, “Reserved Bit Checking”.
10.5.1.2 APIC Register Address Space
The MSR address range between 0000_0800H through 0000_0BFFH is architectur-
ally reserved and dedicated for accessing APIC registers in x2APIC mode. Table 10-3
provides the detailed list of the APIC registers in xAPIC mode and x2APIC mode. The
MSR address offset specified in the table is relative to the base MSR address of 800H.
The MMIO offset specified in the table is relative to the default base address of
FEE00000H.
There is a one-to-one mapping between the legacy xAPIC register MMIO offset and
the MSR address offset with the following exceptions:
• The Interrupt Command Register (ICR): The two 32-bit ICR registers in xAPIC
mode are merged into a single 64-bit MSR in x2APIC mode.
• The Destination Format Register (DFR) is not supported in x2APIC mode.
• The SELF IPI register is available only if x2APIC mode is enabled.
The MSR address space is compressed to allow for future growth. Every 32 bit
register on a 128- bit boundary in the legacy MMIO space is mapped to a single MSR
in the local x2APIC MSR address space. The upper 32-bits of all x2APIC MSRs (except
for the ICR) are reserved.
Table 10-3. Local APIC Register Address Map Supported by x2APIC
MSR Offset
MMIO Offset (x2APIC
(xAPIC mode) mode)
R/W
Register Name
Semantics Comments
0000H-
0010H
000H-001H
Reserved
0020H
002H
003H
Local APIC ID Register
Read only
Read only.
See Section 10.5.6.1
for initial values.
0030H
Local APIC Version
Register
Same version between
extended and legacy
modes. Bit 24 is available
only to an x2APIC unit (in
xAPIC mode and x2APIC
modes, See Section
2.5.1).
0040H-
0070H
004H-007H
Reserved
10-18 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
Table 10-3. Local APIC Register Address Map Supported by x2APIC (Contd.)
MSR Offset
MMIO Offset (x2APIC
(xAPIC mode) mode)
R/W
Register Name
Semantics Comments
0080H
008H
Task Priority Register
(TPR)
Read/Write. Bits 7:0 are RW. Bits 31:8
are Reserved.
0090H
00A0H
009H
00AH
Reserved
Processor Priority
Register (PPR)
Read only.
00B0H
00BH
EOI Register
Write only. 0 is the only valid value
to write. GP fault on non-
zero write
00C0H
00D0H
00CH
00DH
Reserved
Logical Destination
Register
Read only.
Read/Write in xAPIC
mode)
1
00E0H
00F0H
0100H
00EH
00FH
010H
Reserved
GP fault on Read Write in
x2APIC mode.
Spurious Interrupt
Vector Register
Read/Write. Bits 0-8, 12 Read/Write;
other bits reserved.
In-Service Register
(ISR); bits 0:31
Read Only.
0110H
0120H
0130H
0140H
0150H
0160H
0170H
0180H
011H
012H
013H
014H
015H
016H
017H
018H
ISR bits 32:63
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
ISR bits 64:95
ISR bits 96:127
ISR bits 128:159
ISR bits 160:191
ISR bits 192:223
ISR bits 224:255
Trigger Mode Register
(TMR); bits 0:31
0190H
01A0H
01B0H
01C0H
01D0H
01E0H
019H
01AH
01BH
01CH
01DH
01EH
TMR bits 32:63
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
TMR bits 64:95
TMR bits 96:127
TMR bits 128:159
TMR bits 160:191
TMR bits 192:223
Vol. 3 10-19
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
Table 10-3. Local APIC Register Address Map Supported by x2APIC (Contd.)
MSR Offset
MMIO Offset (x2APIC
(xAPIC mode) mode)
R/W
Register Name
Semantics Comments
01F0H
0200H
01FH
020H
TMR bits 224:255
Read Only.
Read Only.
Interrupt Request
Register (IRR); bits 0:31
0210H
0220H
0230H
0240H
0250H
0260H
0270H
0280H
021H
022H
023H
024H
025H
026H
027H
028H
IRR bits32:63
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
Read Only.
IRR bits 64:95
IRR bits 96:127
IRR bits 128:159
IRR bits 160:191
IRR bits 192:223
IRR bits 224:255
Error Status Register
Read/Write. GP fault on non-zero
writes
0290H-
02E0H
029H-02EH
02FH
Reserved
02F0H
LVT CMCI Register
Read/Write.
Read/Write.
3
0300H-
030H
Interrupt Command
Register (ICR); bits 0-63
2
0310H
0320H
0330H
032H
033H
LVT Timer Register
Read/Write.
Read/Write.
LVT Thermal Sensor
Register
0340H
034H
LVT Performance
Monitoring Register
Read/Write.
0350H
0360H
0370H
0380H
035H
036H
037H
038H
LVT LINT0 Register
LVT LINT1 Register
LVT Error Register
Read/Write.
Read/Write.
Read/Write.
Read/Write.
Initial Count Register
(for Timer)
0390H
039H
Current Count Register Read Only.
(for Timer)
03A0H-
03D0H
03AH-03DH
Reserved
10-20 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
Table 10-3. Local APIC Register Address Map Supported by x2APIC (Contd.)
MSR Offset
MMIO Offset (x2APIC
(xAPIC mode) mode)
R/W
Register Name
Semantics Comments
03E0H
03EH
Divide Configuration
Register (for Timer)
Read/Write.
4
Not supported 03FH
040H-3FFH
SELF IPI
Write only
Only in x2APIC mode
Reserved
NOTES:
1. Destination format register (DFR) is supported in xAPIC mode at
MMIO offset 00E0H.
2. APIC register at MMIO offset 0310H is accessible in xAPIC mode
only
3. MSR 831H (offset 31H) is reserved; read/write operations will result
in a GP fault.
4. SELF IPI register is supported only if x2APIC mode is enabled.
10.5.1.3 Reserved Bit Checking
Section 10.5.1.2 and Table 10-3 specifies the reserved bit definitions for the APIC
registers in x2APIC mode. Non-zero writes (by WRMSR instruction) to reserved bits
to these registers will raise a general protection fault exception while reads return
zeros (RsvdZ semantics).
In x2APIC mode, the local APIC ID register is increased to 32 bits wide. This enables
2^32 -1 processors to be addressable in physical destination mode. This 32-bit value
is referred to as “x2APIC ID”. A processor implementation may choose to support less
than 32 bits in its hardware. System software should be agnostic to the actual
number of bits that are implemented. All non-implemented bits will return zeros on
reads by software.
The APIC ID value of FFFF_FFFFH and the highest value corresponding to the imple-
mented bit-width of the local APIC ID register in the system are reserved and cannot
be assigned to any logical processor.
In x2APIC mode, the local APIC ID register is a read-only register to system software
and will be initialized by hardware. It is accessed via the RDMSR instruction reading
the MSR at address 0802H.
Each logical processor in the system (including clusters with a communication fabric)
must be configured with an unique x2APIC ID to avoid collisions of x2APIC IDs. On
DP and high-end MP processors targeted to specific market segments and depending
on the system configuration, it is possible that logical processors in different and "un-
connected" clusters power up initialized with overlapping x2APIC IDs. In these
configurations, a model-specific means may be provided in those product segments
Vol. 3 10-21
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
to enable BIOS and/or platform firmware to re-configure the x2APIC IDs in some
clusters to provide for unique and non-overlapping system wide IDs before config-
uring the disconnected components into a single system.
10.5.2
x2APIC Register Availability
The local APIC registers can be accessed via the MSR interface only when the local
APIC has been switched to the x2APIC mode as described in Section 10.5.1.
Accessing any APIC register in the MSR address range 0800H through 0BFFH via
RDMSR or WRMSR when the local APIC is not in x2APIC mode will cause the instruc-
tions to raise a GP fault. In x2APIC mode, the memory mapped interface is not avail-
able and any access to the MMIO interface will behave similar to that of a legacy
xAPIC in globally disabled state. Table 10-4 provides the interactions between the
legacy & extended modes and the legacy and register interfaces.
Table 10-4. MSR/MMIO Interface of a Local x2APIC in Different Modes of Operation
MMIO Interface
MSR Interface
GP Fault
xAPIC mode
Available
x2APIC mode
Behavior identical to xAPIC in globally
disabled state
Available
10.5.3
MSR Access in x2APIC Mode
To allow for efficient access to the APIC registers in x2APIC mode, the serializing
semantics of WRMSR are relaxed when writing to the APIC registers. Thus, system
software should not use “WRMSR to APIC registers in x2APIC mode” as a serializing
instruction. Read and write accesses to the APIC registers will occur in program
order. A WRMSR to an APIC register may complete before all preceding stores are
globally visible; software can prevent this by inserting a serializing instruction or
MFENCE before the WRMSR.
The RDMSR instruction is not serializing and this behavior is unchanged when
reading APIC registers in x2APIC mode. System software accessing the APIC regis-
ters using the RDMSR instruction should not expect a serializing behavior. (Note: The
MMIO-based xAPIC interface is mapped by system software as an un-cached region.
Consequently, read/writes to the xAPIC-MMIO interface have serializing semantics in
the xAPIC mode.)
10.5.4
VM-exit Controls for MSRs and x2APIC Registers
The VMX architecture allows a VMM to specify lists of MSRs to be loaded or stored on
VMX transitions using the VMX-transition MSR areas (see VM-exit MSR-store address
10-22 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
field, VM-exit MSR-load address filed, and VM-entry MSR-load address field in Intel®
64 and IA-32 Architectures Software Developer’s Manual, Volume 3B).
The X2APIC MSRs cannot to be loaded and stored on VMX transitions. A VMX transi-
tion fails if the VMM has specified that the transition should access any MSRs in the
address range from 0000_0800H to 0000_08FFH (the range used for accessing the
X2APIC registers). Specifically, processing of an 128-bit entry in any of the VMX-
transition MSR areas fails if bits 31:0 of that entry (represented as ENTRY_LOW_DW)
satisfies the expression: “ENTRY_LOW_DW & FFFFF800H = 00000800H”. Such a
failure causes an associated VM entry to fail (by reloading host state) and causes an
associated VM exit to lead to VMX abort.
10.5.5
Directed EOI with x2APIC Mode
Directed EOI capability is intended to enable system software to perform directed
EOIs to specific IOxAPICs in the system. System software desiring to perform a
directed EOI would do the following:
• inhibit the broadcast of EOI message by setting bit 12 of the Spurious Interrupt
Vector Register, and
• following the EOI to the local x2APIC unit for a level triggered interrupt, perform
a directed EOI to the IOxAPIC generating the interrupt by writing to its EOI
register.
Supporting directed EOI capability would require system software to retain a
mapping associating level triggered interrupts with IOxAPICs in the system.
Bit 12 of the Spurious Interrupt Vector Register (SVR) in the local x2APIC unit
controls the generation of the EOI broadcast if the Directed EOI capability is
supported. This bit is reserved to 0 if the processor doesn't support Directed EOI. If
SVR[bit 12] is set, a broadcast EOI is not generated on an EOI cycle even if the asso-
ciated TMR bit is indicating the current interrupt is a level triggered interrupt. Layout
of the Spurious Interrupt Vector Register is shown in Figure 10-9.
Vol. 3 10-23
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
31
12
11
9
8
7
0
Reserved
EOI Broadcast Disable
APIC Software Enable/Disable
0: APIC Disabled
1: APIC Enabled
Spurious Vector
MMIO Address: FEE0 00F0H
MSR Address: 080FH
Figure 10-9. Spurious Interrupt Vector Register (SVR) of x2APIC
The default value for SVR[bit 12] is clear, indicating that an EOI broadcast will be
performed.
The support for Directed EOI capability can be detected by means of bit 24 in the
Local APIC Version Register. This feature is supported in both the xAPIC mode and
x2APIC modes of a local x2APIC unit. Layout of the Local APIC Version register is as
shown in Figure 10-10. The Directed EOI feature is supported if bit 24 is set to 1.
31
8
7
0
25 24
23
15
16
Vector
Max LVT Entry
Reserved
Reserved
Directed EOI Support
MMIO Address: FEE0 0030H
MSR Address: 0803H
Figure 10-10. Local APIC Version Register of x2APIC
10.5.6
x2APIC State Transitions
This section provides a detailed description of the x2APIC states of a local x2APIC
unit, transitions between these states as well as interactions of these states with INIT
and RESET.
10.5.6.1 x2APIC States
The valid states for a local x2APIC unit is listed in Table 10-2:
• APIC disabled: IA32_APIC_BASE[EN]=0 and IA32_APIC_BASE[EXTD]=0
10-24 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
• xAPIC mode: IA32_APIC_BASE[EN]=1 and IA32_APIC_BASE[EXTD]=0
• x2APIC mode: IA32_APIC_BASE[EN]=1 and IA32_APIC_BASE[EXTD]=1
• Invalid: IA32_APIC_BASE[EN]=0 and IA32_APIC_BASE[EXTD]=1
The state corresponding to EXTD=1 and EN=0 is not valid and it is not possible to get
into this state. Values written to the IA32_APIC_BASE_MSR that attempt a transition
from a valid state to this invalid state will cause a GP fault. Figure 10-11 shows the
comprehensive state transition diagram for a local x2APIC unit.
On coming out of RESET, the local APIC unit is enabled and is in the xAPIC mode:
IA32_APIC_BASE[EN]=1 and IA32_APIC_BASE[EXTD]=0. The APIC registers are
initialized as:
• The local APIC ID is initialized by hardware with a 32 bit ID (x2APIC ID). The
lowest 8 bits of the x2APIC ID is the legacy local xAPIC ID, and is stored in the
upper 8 bits of the APIC register for access in xAPIC mode.
• The following APIC registers are reset to all zeros for those fields that are defined
in the xAPIC mode:
— IRR, ISR, TMR, ICR, LDR, TPR, Divide Configuration Register (See Chapter 8
of “Intel® 64 and IA-32 Architectures Software Developer’s Manual“, Vol. 3B
for details of individual APIC registers),
— Timer initial count and timer current count registers,
• The LVT registers are reset to 0s except for the mask bits; these are set to 1s.
• The local APIC version register is not affected.
• The Spurious Interrupt Vector Register is initialized to 000000FFH.
• The DFR (available only in xAPIC mode) is reset to all 1s.
• SELF IPI register is reset to zero.
Vol. 3 10-25
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
Reset
Init
Disabled
EN = 0
Extd = 0
Init
Illegal
Transition
Extd = 1
EN =1
EN = 0
Illegal
Transition
EN = 0
Extd = 0
xAPIC Mode
Invalid
State
EN=1, Extd=0
Extd = 1
EN = 0
Illegal
Transition
Extd = 0
Extd = 1
Illegal
Transition
EN = 0
Reset
Extended
Mode
EN=1, Extd=1
Reset
Init
Figure 10-11. Local x2APIC State Transitions with IA32_APIC_BASE, INIT, and RESET
x2APIC After RESET
The valid transitions from the xAPIC mode state are:
• to the x2APIC mode by setting EXT to 1 (resulting EN=1, EXTD= 1). The physical
x2APIC ID (see Figure 10-6) is preserved across this transition and the logical
x2APIC ID (see Figure 10-21) is initialized by hardware during this transition as
documented in Section 10.7.2.4. The state of the extended fields in other APIC
registers, which was not initialized at RESET, is not architecturally defined across
this transition and system software should explicitly initialize those program-
mable APIC registers.
EXTD= 0. The state of the local APIC ID register is preserved (the 8-bit xAPIC ID is in
the upper 8 bits of the APIC ID register). All the other APIC registers are initialized as
a result of INIT.
A RESET in this state places the APIC in the state with EN= 1, EXTD= 0. The state of
the local APIC ID register is initialized as described in Section 10.5.6.1. All the other
APIC registers are initialized described in Section 10.5.6.1.
10-26 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
x2APIC Transitions From x2APIC Mode
From the x2APIC mode, the only valid x2APIC transition using IA32_APIC_BASE is to
the state where the x2APIC is disabled by setting EN to 0 and EXTD to 0. The x2APIC
ID (32 bits) and the legacy local xAPIC ID (8 bits) are preserved across this transi-
tion. A transition from the x2APIC mode to xAPIC mode is not valid and the corre-
sponding WRMSR to the IA32_APIC_BASE MSR will raise a GP fault.
A RESET in this state places the x2APIC in xAPIC mode. All APIC registers (including
the local APIC ID register) are initialized as described in Section 10.5.6.1.
An INIT in this state keeps the x2APIC in the x2APIC mode. The state of the local
APIC ID register is preserved (all 32 bits). However, all the other APIC registers are
initialized as a result of the INIT transition.
x2APIC Transitions From Disabled Mode
From the disabled state, the only valid x2APIC transition using IA32_APIC_BASE is to
the xAPIC mode (EN= 1, EXTD = 0). Thus the only means to transition from x2APIC
mode to xAPIC mode is a two-step process:
• first transition from x2APIC mode to local APIC disabled mode (EN= 0, EXTD =
0),
• followed by another transition from disabled mode to xAPIC mode (EN= 1,
EXTD= 0).
Consequently, all the APIC register states in the x2APIC, except for the x2APIC ID
(32 bits), are not preserved across mode transitions.
A RESET in the disabled state places the x2APIC in the xAPIC mode. All APIC registers
(including the local APIC ID register) are initialized as described in Section 10.5.6.1.
An INIT in the disabled state keeps the x2APIC in the disabled state.
State Changes From xAPIC Mode to x2APIC Mode
After APIC register states have been initialized by software in xAPIC mode, a transi-
tion from xAPIC mode to x2APIC mode does not affect most of the APIC register
states, except the following:
• The Logical Destination Register is not preserved.
• Any APIC ID value written to the memory-mapped local APIC ID register is not
preserved.
• The high half of the Interrupt Command Register is not preserved.
10.5.7
System Software Transitions
This section describes implications for the x2APIC across system state transitions -
specifically initialization and booting.
Vol. 3 10-27
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
Support for the x2APIC architecture can be implemented in the local APIC unit. All
existing PCI/MSI capable devices and IOxAPIC unit should work with the x2APIC
extensions defined in this document. The x2APIC architecture also provides flexibility
to cope with the underlying fabrics that connect the PCI devices, IOxAPICs and Local
APIC units.
The extensions provided in this specification translate into modifications to:
• the local APIC unit,
• the underlying fabrics connecting Message Signaled Interrupts (MSI) capable PCI
devices to local xAPICs,
• the underlying fabrics connecting the IOxAPICs to the local APIC units.
However no modifications are required to PCI or PCIe devices that support direct
interrupt delivery to the processors via Message Signaled Interrupts. Similarly no
modifications are required to the IOxAPIC. The routing of interrupts from these
devices in x2APIC mode leverages the interrupt remapping architecture specified in
®
the Intel Virtualization Technology for Directed I/O, Rev 1.2 specification.
Modifications to ACPI interfaces to support x2APIC are described in Appendix A,
®
“ACPI Extensions for x2APIC Support”, of the Intel 64 Architecture x2APIC Specifi-
cation.
The default will be for the BIOS to pass the control to the OS with the local x2APICs
in xAPIC mode if all x2APIC IDs reported by CPUID.0BH:EDX are less than 255, and
in x2APIC mode if there are any logical processor reporting its x2APIC ID at 255 or
greater.
10.5.8
CPUID Extensions And Topology Enumeration
For Intel 64 and IA-32 processors that support x2APIC, a value of 1 reported by
CPUID.01H:ECX[21] indicates that the processor supports x2APIC and the extended
topology enumeration leaf (CPUID.0BH).
The extended topology enumeration leaf can be accessed by executing CPUID with
EAX = 0BH. Processors that do not support x2APIC may support CPUID leaf 0BH.
Software can detect the availability of the extended topology enumeration leaf (0BH)
by performing two steps:
• Check maximum input value for basic CPUID information by executing CPUID
with EAX= 0. If CPUID.0H:EAX is greater than or equal or 11 (0BH), then proceed
to next step
• Check CPUID.EAX=0BH, ECX=0H:EBX is non-zero.
If both of the above conditions are true, extended topology enumeration leaf is avail-
able. The presence of CPUID leaf 0BH in a processor does not guarantee support for
x2APIC. If CPUID.EAX=0BH, ECX=0H:EBX returns zero and maximum input value for
basic CPUID information is greater than 0BH, then CPUID.0BH leaf is not supported
on that processor.
10-28 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
The extended topology enumeration leaf is intended to assist software with enumer-
ating processor topology on systems that requires 32-bit x2APIC IDs to address indi-
vidual logical processors. For example, a system with greater than 256 logical
processors or greater than 64 processor cores will require the OS to use 32-bit
x2APIC IDs. Details of CPUID leaf 0BH can be found in the reference pages of CPUID
in Chapter 3 of Intel® 64 and IA-32 Architectures Software Developer’s Manual,
Volume 2A.
Processor topology enumeration algorithm for processors supporting the extended
topology enumeration leaf of CPUID and processors that do not support CPUID leaf
0BH are treated in Section 8.9.4, “Algorithm for Three-Level Mappings of APIC_ID”.
10.5.8.1 Consistency of APIC IDs and CPUID
The consistency of physical x2APIC ID in MSR 802H in x2APIC mode and the 32-bit
value returned in CPUID.0BH:EDX is facilitated by processor hardware.
CPUID.0BH:EDX will report the full 32 bit ID, in xAPIC and x2APIC mode. This allows
BIOS to determine if a system has processors with IDs exceeding the 8-bit initial
APIC ID limit (CPUID.01H:EBX[31:24]). Initial APIC ID (CPUID.01H:EBX[31:24]) is
always equal to CPUID.0BH:EDX[7:0].
If the values of CPUID.0BH:EDX reported by all logical processors in a system are
less than 255, BIOS can transfer control to OS in xAPIC mode.
If the values of CPUID.0BH:EDX reported by some logical processors in a system are
greater or equal than 255, BIOS must support two options to hand off to OS:
• If BIOS enables logical processors with x2APIC IDs greater than 255, then it
should enable X2APIC in Boot Strap Processor (BSP) and all Application
Processors (AP) before passing control to the OS. Application requiring processor
topology information must use OS provided services based on x2APIC IDs or
CPUID.0BH leaf.
• If a BIOS transfers control to OS in xAPIC mode, then the BIOS must ensure that
only logical processors with CPUID.0BH.EDX value less than 255 are enabled.
BIOS initialization on all logical processors with CPUID.0B.EDX values greater
than or equal to 255 must (a) disable APIC and execute CLI in each logical
processor, and (b) leave these logical processor in the lowest power state so that
these processors do not respond to INIT IPI during OS boot. The BSP and all the
enabled logical processor operate in xAPIC mode after BIOS passed control to
OS. Application requiring processor topology information can use OS provided
legacy services based on 8-bit initial APIC IDs or legacy topology information
from CPUID.01H and CPUID 04H leaves. Even if the BIOS passes control in xAPIC
mode, an OS can switch the processors to x2APIC mode later. BIOS SMM handler
should always read the APIC_BASE_MSR, determine the APIC mode and use the
corresponding access method.
Vol. 3 10-29
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
10.6
HANDLING LOCAL INTERRUPTS
The following sections describe facilities that are provided in the local APIC for
handling local interrupts. These include: the processor’s LINT0 and LINT1 pins, the
APIC timer, the performance-monitoring counters, the thermal sensor, and the
internal APIC error detector. Local interrupt handling facilities include: the LVT, the
error status register (ESR), the divide configuration register (DCR), and the initial
count and current count registers.
10.6.1
Local Vector Table
The local vector table (LVT) allows software to specify the manner in which the local
interrupts are delivered to the processor core. It consists of the following 32-bit APIC
• LVT Timer Register (FEE0 0320H) — Specifies interrupt delivery when the
APIC timer signals an interrupt (see Section 10.6.4, “APIC Timer”).
• LVT Thermal Monitor Register (FEE0 0330H) — Specifies interrupt delivery
when the thermal sensor generates an interrupt (see Section 14.5.2, “Thermal
Monitor”). This LVT entry is implementation specific, not architectural. If imple-
mented, it will always be at base address FEE0 0330H.
• LVT Performance Counter Register (FEE0 0340H) — Specifies interrupt
delivery when a performance counter generates an interrupt on overflow (see
Section 30.8.5.8, “Generating an Interrupt on Overflow”). This LVT entry is
implementation specific, not architectural. If implemented, it is not guaranteed
to be at base address FEE0 0340H.
• LVT LINT0 Register (FEE0 0350H) — Specifies interrupt delivery when an
interrupt is signaled at the LINT0 pin.
• LVT LINT1 Register (FEE0 0360H) — Specifies interrupt delivery when an
• LVT Error Register (FEE0 0370H) — Specifies interrupt delivery when the
APIC detects an internal error (see Section 10.6.3, “Error Handling”).
• CMCI LVT Register (FEE0 02F0H) — Specifies interrupt delivery when an
overflow condition of corrected machine check error count reaching a threshold
“CMCI Local APIC Interface”).
The LVT performance counter register and its associated interrupt were introduced in
the P6 processors and are also present in the Pentium 4 and Intel Xeon processors.
The LVT thermal monitor register and its associated interrupt were introduced in the
Pentium 4 and Intel Xeon processors.
As shown in Figures 10-12, some of these fields and flags are not available (and
reserved) for some entries.
10-30 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
31
18 17 16 15
13 12 11
8 7
0
Timer
Vector
Address: FEE0 0320H
Value after Reset: 0001 0000H
Timer Mode
0: One-shot
1: Periodic
Delivery Status
0: Idle
1: Send Pending
Mask†
0: Not Masked
1: Masked
Interrupt Input
Pin Polarity
Delivery Mode
000: Fixed
010: SMI
100: NMI
111: ExtlNT
101: INIT
Remote
IRR
All other combinations
are Reserved
Trigger Mode
0: Edge
1: Level
31
17
11 10
8
7
0
LINT0
Vector
Vector
Vector
Vector
LINT1
Error
Performance
Mon. Counters
Thermal
Sensor
Vector
16 15 14 13 12
Address: FEE0 0350H
Address: FEE0 0360H
Reserved
Address: FEE0 0370H
Address: FEE0 0340H
Address: FEE0 0330H
† (Pentium 4 and Intel Xeon processors.) When a
performance monitoring counters interrupt is generated,
the mask bit for its associated LVT entry is set.
Value After Reset: 0001 0000H
Figure 10-12. Local Vector Table (LVT)
Vol. 3 10-31
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
The setup information that can be specified in the registers of the LVT table is as
follows:
Vector
Interrupt vector number.
Delivery Mode
Specifies the type of interrupt to be sent to the processor. Some
delivery modes will only operate as intended when used in
conjunction with a specific trigger mode. The allowable delivery
modes are as follows:
000 (Fixed) Delivers the interrupt specified in the vector
field.
010 (SMI)
Delivers an SMI interrupt to the processor
core through the processor’s local SMI signal
path. When using this delivery mode, the
vector field should be set to 00H for future
compatibility.
100 (NMI)
101 (INIT)
Delivers an NMI interrupt to the processor.
The vector information is ignored.
Delivers an INIT request to the processor
core, which causes the processor to perform
an INIT. When using this delivery mode, the
vector field should be set to 00H for future
compatibility.
111 (ExtINT) Causes the processor to respond to the in-
terrupt as if the interrupt originated in an
externally connected (8259A-compatible)
interrupt controller. A special INTA bus cycle
corresponding to ExtINT, is routed to the ex-
ternal controller. The external controller is
expected to supply the vector information.
The APIC architecture supports only one Ex-
tINT source in a system, usually contained in
the compatibility bridge.
Delivery Status (Read Only)
Indicates the interrupt delivery status, as follows:
0 (Idle)
There is currently no activity for this inter-
rupt source, or the previous interrupt from
core and accepted.
1 (Send Pending)
Indicates that an interrupt from this source
has been delivered to the processor core,
but has not yet been accepted (see Section
10.6.5, “Local Interrupt Acceptance”).
10-32 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
Interrupt Input Pin Polarity
Specifies the polarity of the corresponding interrupt pin: (0)
active high or (1) active low.
Remote IRR Flag (Read Only)
For fixed mode, level-triggered interrupts; this flag is set when
the local APIC accepts the interrupt for servicing and is reset
when an EOI command is received from the processor. The
meaning of this flag is undefined for edge-triggered interrupts
and other delivery modes.
Trigger Mode
Selects the trigger mode for the local LINT0 and LINT1 pins: (0)
edge sensitive and (1) level sensitive. This flag is only used
when the delivery mode is Fixed. When the delivery mode is
NMI, SMI, or INIT, the trigger mode is always edge sensitive.
When the delivery mode is ExtINT, the trigger mode is always
level sensitive. The timer and error interrupts are always treated
as edge sensitive.
If the local APIC is not used in conjunction with an I/O APIC and
fixed delivery mode is selected; the Pentium 4, Intel Xeon, and
P6 family processors will always use level-sensitive triggering,
regardless if edge-sensitive triggering is selected.
Mask
Interrupt mask: (0) enables reception of the interrupt and (1)
inhibits reception of the interrupt. When the local APIC handles
a performance-monitoring counters interrupt, it automatically
sets the mask flag in the corresponding LVT entry. This flag will
remain set until software clears it.
Timer Mode
Selects the timer mode: (0) one-shot and (1) periodic (see
Section 10.6.4, “APIC Timer”).
10.6.2
The Intel 64 and IA-32 architectures define 256 vector numbers, ranging from 0
APICs support 240 of these vectors (in the range of 16 to 255) as valid interrupts.
When an interrupt vector in the range of 0 to 15 is sent or received through the local
APIC, the APIC indicates an illegal vector in its Error Status Register (see Section
10.6.3, “Error Handling”). The Intel 64 and IA-32 architectures reserve vectors 16
through 31 for predefined interrupts, exceptions, and Intel-reserved encodings (see
Table 6-1). However, the local APIC does not treat vectors in this range as illegal.
When an illegal vector value (0 to 15) is written to an LVT entry and the delivery
mode is Fixed (bits 8-11equal 0), the APIC may signal an illegal vector error, without
regard to whether the mask bit is set or whether an interrupt is actually seen on the
input.
Vol. 3 10-33
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
10.6.3
Error Handling
The local APIC provides an error status register (ESR) that it uses to record errors
that it detects when handling interrupts (see Figure 10-13). An APIC error interrupt
is generated when the local APIC sets one of the error bits in the ESR. The LVT error
register allows selection of the interrupt vector to be delivered to the processor core
when APIC error is detected. The LVT error register also provides a means of masking
an APIC error interrupt.
8 7 6 5 4 3 2
1
0
31
Reserved
Illegal Register Address1
Received Illegal Vector
Send Illegal Vector
Reserved
Receive Accept Error2
Send Accept Error2
Receive Checksum Error2
Send Checksum Error2
Address: FEE0 0280H
Value after reset: 0H
NOTES:
1. Used in Intel Core, Pentium 4, Intel Xeon, and P6 family
processors; reserved in the Pentium processor.
2. Only used in the P6 family and Pentium processors;
reserved in Intel Core, Pentium 4 and Intel Xeon processors.
Figure 10-13. Error Status Register (ESR)
The ESR is a write/read register. A write (of any value) to the ESR must be done just
prior to reading the ESR to update the register. This initial write causes the ESR
contents to be updated with the latest error status. Back-to-back writes clear the ESR
register. After an error bit is set in the register, it remains set until the register is
cleared.
The functions of the ESR are listed in Table 10-5.
10-34 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
Table 10-5. ESR Flags
FLAG
Function
Send Checksum Error
(P6 family and Pentium processors only) Set when the local APIC
detects a checksum error for a message that it sent on the APIC bus.
Receive Checksum Error
Send Accept Error
(P6 family and Pentium processors only) Set when the local APIC
detects a checksum error for a message that it received on the APIC
bus.
(P6 family and Pentium processors only) Set when the local APIC
detects that a message it sent was not accepted by any APIC on the
APIC bus.
Receive Accept Error
(P6 family and Pentium processors only) Set when the local APIC
detects that the message it received was not accepted by any APIC
on the APIC bus, including itself.
Send Illegal Vector
Set when the local APIC detects an illegal vector in the message that
it is sending.
Receive Illegal Vector
received, including an illegal vector code in the local vector table
interrupts or in a self-interrupt.
Illegal Reg. Address
(Intel Core, Intel Atom, Pentium 4, Intel Xeon, and P6 family
processors only) Set when the processor is trying to access a
register in the processor's local APIC register address space that is
reserved (see Table 10-1). Addresses in one of the 0x10 byte
regions marked reserved are illegal register addresses.
The Local APIC Register Map is the address range of the APIC
register base address (specified in the IA32_APIC_BASE MSR) plus
4 KBytes.
10.6.3.1 x2APIC Differences in Error Handling
RDMSR and WRMSR operations to reserved addresses in the x2APIC mode will raise
a GP fault. Additionally reserved bit violations cause GP faults as detailed in Section
10.5.1.3. Beyond illegal register access and reserved bit violations, other APIC errors
are logged in Error Status Register. Writes of a non-zero value to the Error Status
Register in x2APIC mode will raise a GP fault.
Write to the ICR (in xAPIC and x2APIC modes) or to SELF IPI register (x2APIC mode
only) with an illegal vector (vector <= 0FH) will set the "Send Illegal Vector" bit.
On receiving an IPI with an illegal vector (vector <= 0FH), the "Receive Illegal
Vector" bit will be set. On receiving an interrupt with illegal vector in the range 0H -
0FH, the interrupt will not be delivered to the processor nor will an IRR bit be set in
that range. Only the ESR "Receive Illegal Vector" bit will be set.
Vol. 3 10-35
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
If the ICR is programmed with lowest priority delivery mode then the "Re-directible
IPI" bit will be set in x2APIC modes (same as legacy xAPIC behavior) and the inter-
rupt will not be processed.
Write to the ICR with both lowest priority delivery mode and illegal vector, will set the
"re-directible IPI" error bit. The interrupt will not be processed and hence the "Send
Illegal Vector" error bit will not be set.
8 7 6 5 4 3 2
1
0
31
Reserved
Illegal Register Address
Received Illegal Vector
Send Illegal Vector
Redirectible IPI
Reserved
MSR Address: 828H
Figure 10-14. Error Status Register (ESR) in x2APIC Mode
10.6.4
APIC Timer
The local APIC unit contains a 32-bit programmable timer that is available to soft-
ware to time events or operations. This timer is set up by programming four regis-
ters: the divide configuration register (see Figure 10-15), the initial-count and
current-count registers (see Figure 10-16), and the LVT timer register (see
Figure 10-12).
If CPUID.06H:EAX.ARAT[bit 2] = 1, the processor’s APIC timer runs at a constant
rate regardless of P-state transitions and it continues to run at the same rate in deep
C-states.
If CPUID.06H:EAX.ARAT[bit 2] = 0 or if CPUID 06H is not supported, the APIC timer
may temporarily stop while the processor is in deep C-states or during transitions
caused by Enhanced Intel SpeedStep® Technology.
10-36 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
4
3 2
0
1
0
31
Reserved
Address: FEE0 03E0H
Value after reset: 0H
Divide Value (bits 0, 1 and 3)
000: Divide by 2
001: Divide by 4
010: Divide by 8
011: Divide by 16
100: Divide by 32
101: Divide by 64
110: Divide by 128
111: Divide by 1
Figure 10-15. Divide Configuration Register
31
0
Initial Count
Current Count
Address: Initial Count
FEE0 0380H
Current Count FEE0 0390H
Value after reset: 0H
Figure 10-16. Initial Count and Current Count Registers
The time base for the timer is derived from the processor’s bus clock, divided by the
value specified in the divide configuration register.
The timer can be configured through the timer LVT entry for one-shot or periodic
operation. In one-shot mode, the timer is started by programming its initial-count
register. The initial count value is then copied into the current-count register and
count-down begins. After the timer reaches zero, an timer interrupt is generated and
the timer remains at its 0 value until reprogrammed.
In periodic mode, the current-count register is automatically reloaded from the
initial-count register when the count reaches 0 and a timer interrupt is generated,
and the count-down is repeated. If during the count-down process the initial-count
register is set, counting will restart, using the new initial-count value. The initial-
count register is a read-write register; the current-count register is read only.
The LVT timer register determines the vector number that is delivered to the
processor with the timer interrupt that is generated when the timer count reaches
zero. The mask flag in the LVT timer register can be used to mask the timer interrupt.
Vol. 3 10-37
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
10.6.5
Local Interrupt Acceptance
When a local interrupt is sent to the processor core, it is subject to the acceptance
criteria specified in the interrupt acceptance flow chart in Figure 10-25. If the inter-
rupt is accepted, it is logged into the IRR register and handled by the processor
according to its priority (see Section 10.9.4, “Interrupt Acceptance for Fixed Inter-
rupts”). If the interrupt is not accepted, it is sent back to the local APIC and retried.
10.7
ISSUING INTERPROCESSOR INTERRUPTS
The following sections describe the local APIC facilities that are provided for issuing
interprocessor interrupts (IPIs) from software. The primary local APIC facility for
issuing IPIs is the interrupt command register (ICR). The ICR can be used for the
following functions:
• To send an interrupt to another processor.
• To allow a processor to forward an interrupt that it received but did not service to
another processor for servicing.
• To direct the processor to interrupt itself (perform a self interrupt).
• To deliver special IPIs, such as the start-up IPI (SIPI) message, to other
processors.
Interrupts generated with this facility are delivered to the other processors in the
system through the system bus (for Pentium 4 and Intel Xeon processors) or the
APIC bus (for P6 family and Pentium processors). The ability for a processor to send
system software.
10.7.1
Interrupt Command Register (ICR)
The interrupt command register (ICR) is a 64-bit local APIC register (see
Figure 10-17) that allows software running on the processor to specify and send
interprocessor interrupts (IPIs) to other processors in the system.
To send an IPI, software must set up the ICR to indicate the type of IPI message to
be sent and the destination processor or processors. (All fields of the ICR are read-
write by software with the exception of the delivery status field, which is read-only.)
The act of writing to the low doubleword of the ICR causes the IPI to be sent.
10-38 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
63
55
56
32
Destination Field
Reserved
31
20 1918 1716 1514 1312 1110
8 7
0
Vector
Reserved
Destination Shorthand
00: No Shorthand
01: Self
Delivery Mode
000: Fixed
001: Lowest Priority1
010: SMI
10: All Including Self
11: All Excluding Self
011: Reserved
100: NMI
101: INIT
110: Start Up
111: Reserved
Reserved
Destination Mode
0: Physical
1: Logical
Delivery Status
0: Idle
1: Send Pending
Level
0 = De-assert
1 = Assert
Address: FEE0 0300H (0 - 31)
FEE0 0310H (32 - 63)
Value after Reset: 0H
Trigger Mode
0: Edge
1: Level
NOTE:
1. The ability of a processor to send Lowest Priority IPI is model specific.
Figure 10-17. Interrupt Command Register (ICR)
The ICR consists of the following fields.
Vector
The vector number of the interrupt being sent.
Delivery Mode
Specifies the type of IPI to be sent. This field is also know as the
IPI message type field.
000 (Fixed) Delivers the interrupt specified in the vector
field to the target processor or processors.
001 (Lowest Priority)
Same as fixed mode, except that the inter-
rupt is delivered to the processor executing
at the lowest priority among the set of pro-
cessors specified in the destination field. The
Vol. 3 10-39
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
ability for a processor to send a lowest prior-
ity IPI is model specific and should be avoid-
ed by BIOS and operating system software.
010 (SMI)
Delivers an SMI interrupt to the target pro-
cessor or processors. The vector field must
be programmed to 00H for future compati-
bility.
011 (Reserved)
100 (NMI)
Delivers an NMI interrupt to the target pro-
cessor or processors. The vector information
is ignored.
101 (INIT)
Delivers an INIT request to the target pro-
cessor or processors, which causes them to
perform an INIT. As a result of this IPI mes-
sage, all the target processors perform an
INIT. The vector field must be programmed
to 00H for future compatibility.
(Not supported in the Pentium 4 and Intel
Xeon processors.) Sends a synchronization
message to all the local APICs in the system
to set their arbitration IDs (stored in their
Arb ID registers) to the values of their APIC
IDs (see Section 10.8, “System and APIC
Bus Arbitration”). For this delivery mode,
the level flag must be set to 0 and trigger
mode flag to 1. This IPI is sent to all proces-
sors, regardless of the value in the destina-
tion field or the destination shorthand field;
however, software should specify the “all in-
110 (Start-Up)
Sends a special “start-up” IPI (called a SIPI)
to the target processor or processors. The
vector typically points to a start-up routine
that is part of the BIOS boot-strap code (see
Section 8.4, “Multiple-Processor (MP) Initial-
ization”). IPIs sent with this delivery mode
are not automatically retried if the source
APIC is unable to deliver it. It is up to the
software to determine if the SIPI was not
successfully delivered and to reissue the
SIPI if necessary.
10-40 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
Destination Mode Selects either physical (0) or logical (1) destination mode (see
Section 10.7.2, “Determining IPI Destination”).
Delivery Status (Read Only)
Indicates the IPI delivery status, as follows:
0 (Idle)
There is currently no IPI activity for this local
APIC, or the previous IPI sent from this local
APIC was delivered and accepted by the tar-
get processor or processors.
1 (Send Pending)
Indicates that the last IPI sent from this lo-
cal APIC has not yet been accepted by the
target processor or processors.
Level
For the INIT level de-assert delivery mode this flag must be set
to 0; for all other delivery modes it must be set to 1. (This flag
has no meaning in Pentium 4 and Intel Xeon processors, and will
always be issued as a 1.)
Trigger Mode
Selects the trigger mode when using the INIT level de-assert
delivery mode: edge (0) or level (1). It is ignored for all other
delivery modes. (This flag has no meaning in Pentium 4 and
Intel Xeon processors, and will always be issued as a 0.)
Destination Shorthand
Indicates whether a shorthand notation is used to specify the
destination of the interrupt and, if so, which shorthand is used.
Destination shorthands are used in place of the 8-bit destination
field, and can be sent by software using a single write to the low
doubleword of the ICR. Shorthands are defined for the following
cases: software self interrupt, IPIs to all processors in the
system including the sender, IPIs to all processors in the system
excluding the sender.
00: (No Shorthand)
The destination is specified in the destination
field.
01: (Self)
The issuing APIC is the one and only destina-
tion of the IPI. This destination shorthand al-
lows software to interrupt the processor on
which it is executing. An APIC implementa-
tion is free to deliver the self-interrupt mes-
sage internally or to issue the message to
the bus and “snoop” it as with any other IPI
message.
10: (All Including Self)
The IPI is sent to all processors in the system
including the processor sending the IPI. The
APIC will broadcast an IPI message with the
Vol. 3 10-41
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
destination field set to FH for Pentium and P6
family processors and to FFH for Pentium 4
and Intel Xeon processors.
11: (All Excluding Self)
The IPI is sent to all processors in a system
with the exception of the processor sending
the IPI. The APIC broadcasts a message with
the physical destination mode and destina-
tion field set to 0xFH for Pentium and P6
family processors and to 0xFFH for Pentium
4 and Intel Xeon processors. Support for this
destination shorthand in conjunction with
the lowest-priority delivery mode is model
specific. For Pentium 4 and Intel Xeon pro-
cessors, when this shorthand is used togeth-
er with lowest priority delivery mode, the IPI
may be redirected back to the issuing pro-
cessor.
Destination
Specifies the target processor or processors. This field is only
used when the destination shorthand field is set to 00B. If the
destination mode is set to physical, then bits 56 through 59
contain the APIC ID of the target processor for Pentium and P6
the target processor the for Pentium 4 and Intel Xeon proces-
sors. If the destination mode is set to logical, the interpretation
of the 8-bit destination field depends on the settings of the DFR
and LDR registers of the local APICs in all the processors in the
system (see Section 10.7.2, “Determining IPI Destination”).
Not all combinations of options for the ICR are valid. Table 10-6 shows the valid
combinations for the fields in the ICR for the Pentium 4 and Intel Xeon processors;
Table 10-7 shows the valid combinations for the fields in the ICR for the P6 family
processors. Also note that the lower half of the ICR may not be preserved over tran-
sitions to the deepest C-States.
Table 10-6 Valid Combinations for the Pentium 4 and Intel Xeon Processors’
Local xAPIC Interrupt Command Register
Destination
Shorthand
Valid/
Trigger
Destination
Mode
Invalid Mode
Delivery Mode
1
No Shorthand
No Shorthand
Self
Valid
Edge
Level
Edge
Level
All Modes
Physical or Logical
Physical or Logical
2
2
Invalid
Valid
All Modes
Fixed
3
X
Self
Invalid
Fixed
X
10-42 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
Table 10-6 Valid Combinations for the Pentium 4 and Intel Xeon Processors’
Local xAPIC Interrupt Command Register (Contd.)
Destination
Shorthand
Valid/
Trigger
Destination
Mode
Invalid Mode
Delivery Mode
Self
Invalid
X
Lowest Priority, NMI, INIT, SMI, Start-
Up
X
All Including Self Valid
All Including Self Invalid
All Including Self Invalid
Edge
Level
X
Fixed
Fixed
X
X
X
2
Lowest Priority, NMI, INIT, SMI, Start-
Up
,
1 4
All Excluding
Self
Valid
Edge
Level
Fixed, Lowest Priority , NMI, INIT,
SMI, Start-Up
X
X
2
4
All Excluding
Self
Invalid
FIxed, Lowest Priority , NMI, INIT,
SMI, Start-Up
NOTES:
1. The ability of a processor to send a lowest priority IPI is model specific.
2. For these interrupts, if the trigger mode bit is 1 (Level), the local xAPIC will override the bit set-
ting and issue the interrupt as an edge triggered interrupt.
3. X means the setting is ignored.
4. When using the “lowest priority” delivery mode and the “all excluding self” destination, the IPI
can be redirected back to the issuing APIC, which is essentially the same as the “all including
self” destination mode.
Table 10-7 Valid Combinations for the P6 Family Processors’
Local APIC Interrupt Command Register
Destination
Shorthand
Valid/
Trigger
Mode
Invalid
Delivery Mode
Destination Mode
Physical or Logical
Physical or Logical
Physical or Logical
1
No Shorthand
No Shorthand
No Shorthand
Self
Valid
Edge
Level
Level
Edge
Level
X
All Modes
2
1
Valid
Fixed, Lowest Priority , NMI
3
Valid
INIT
4
Valid
1
Fixed
Fixed
X
Self
X
X
5
Self
Invalid
Lowest Priority, NMI, INIT,
SMI, Start-Up
All including Self
All including Self
All including Self
Valid
Edge
Level
X
Fixed
Fixed
X
X
X
2
Valid
5
Invalid
Lowest Priority, NMI, INIT,
SMI, Start-Up
Vol. 3 10-43
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
Table 10-7 Valid Combinations for the P6 Family Processors’
Local APIC Interrupt Command Register (Contd.)
Destination
Shorthand
Valid/
Invalid
Trigger
Mode
Delivery Mode
Destination Mode
1
All excluding Self Valid
All excluding Self Valid
Edge
Level
Level
Level
Level
All Modes
X
X
X
X
X
2
1
Fixed, Lowest Priority , NMI
SMI, Start-Up
5
5
All excluding Self Invalid
3
All excluding Self Valid
INIT
X
Invalid
SMI, Start-Up
NOTES:
1. The ability of a processor to send a lowest priority IPI is model specific.
2. Treated as edge triggered if level bit is set to 1, otherwise ignored.
3. Treated as edge triggered when Level bit is set to 1; treated as “INIT Level Deassert” message
when level bit is set to 0 (deassert). Only INIT level deassert messages are allowed to have the
level bit set to 0. For all other messages the level bit must be set to 1.
5. The behavior of the APIC is undefined.
10.7.1.1 ICR Operation in x2APIC Mode
In x2APIC mode, the layout of the Interrupt Command Register is shown in Figure
10-17. The lower 32 bits of ICR in x2APIC mode is identical to the lower half of the
10-44 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
ICR in xAPIC mode, except the Delivery Status bit is removed since it is not needed
in x2APIC mode. The destination ID field is expanded to 32 bits in x2APIC mode.
63
32
Destination Field
31
20 1918 1716 1514 1312 1110
8 7
0
Vector
Reserved
Destination Shorthand
00: No Shorthand
01: Self
Delivery Mode
000: Fixed
001: Reserved
010: SMI
10: All Including Self
11: All Excluding Self
011: Reserved
100: NMI
101: INIT
110: Start Up
111: Reserved
Reserved
Destination Mode
0: Physical
1: Logical
Level
0 = De-assert
1 = Assert
Address: 830H (63 - 0)
Value after Reset: 0H
Trigger Mode
0: Edge
1: Level
Figure 10-18. Interrupt Command Register (ICR) in x2APIC Mode
To send an IPI, software must set up the ICR or SELF IPI register to indicate the type
of IPI message to be sent and the destination processor or processors.
A single MSR write to the Interrupt Command Register is required for dispatching an
interrupt in x2APIC mode. With the removal of the Delivery Status bit, system soft-
ware no longer has a reason to read the ICR. It remains readable only to aid in
debugging.
A destination ID value of FFFF_FFFFH is used for broadcast of interrupts in both
logical destination and physical destination modes.
Vol. 3 10-45
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
10.7.2
Determining IPI Destination
The destination of an IPI can be one, all, or a subset (group) of the processors on the
system bus. The sender of the IPI specifies the destination of an IPI with the
following APIC registers and fields within the registers:
• ICR Register — The following fields in the ICR register are used to specify the
destination of an IPI:
— Destination Mode — Selects one of two destination modes (physical or
logical).
— Destination Field — In physical destination mode, used to specify the APIC
ID of the destination processor; in logical destination mode, used to specify a
message destination address (MDA) that can be used to select specific
processors in clusters.
— Destination Shorthand — A quick method of specifying all processors, all
excluding self, or self as the destination.
— Delivery mode, Lowest Priority — Architecturally specifies that a lowest-
priority arbitration mechanism be used to select a destination processor from
a specified group of processors. The ability of a processor to send a lowest
priority IPI is model specific and should be avoided by BIOS and operating
system software.
• Local destination register (LDR) — Used in conjunction with the logical
destination mode and MDAs to select the destination processors.
• Destination format register (DFR) — Used in conjunction with the logical
destination mode and MDAs to select the destination processors.
How the ICR, LDR, and DFR are used to select an IPI destination depends on the
mode. These destination modes are described in the following sections.
10.7.2.1 Physical Destination Mode
In physical destination mode, the destination processor is specified by its local APIC
ID (see Section 10.4.6, “Local APIC ID”). For Pentium 4 and Intel Xeon processors,
either a single destination (local APIC IDs 00H through FEH) or a broadcast to all
APICs (the APIC ID is FFH) may be specified in physical destination mode.
A broadcast IPI (bits 28-31 of the MDA are 1's) or I/O subsystem initiated interrupt
with lowest priority delivery mode is not supported in physical destination mode and
must not be configured by software. Also, for any non-broadcast IPI or I/O
subsystem initiated interrupt with lowest priority delivery mode, software must
ensure that APICs defined in the interrupt address are present and enabled to receive
interrupts.
For the P6 family and Pentium processors, a single destination is specified in physical
destination mode with a local APIC ID of 0H through 0EH, allowing up to 15 local
10-46 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
APICs to be addressed on the APIC bus. A broadcast to all local APICs is specified with
0FH.
NOTE
The number of local APICs that can be addressed on the system bus
may be restricted by hardware.
10.7.2.2 Logical Destination Mode
In logical destination mode, IPI destination is specified using an 8-bit message desti-
nation address (MDA), which is entered in the destination field of the ICR. Upon
compares the MDA in the message with the values in its LDR and DFR to determine if
it should accept and handle the IPI. For both configurations of logical destination
mode, when combined with lowest priority delivery mode, software is responsible for
ensuring that all of the local APICs included in or addressed by the IPI or I/O
subsystem interrupt are present and enabled to receive the interrupt.
Figure 10-19 shows the layout of the logical destination register (LDR). The 8-bit
logical APIC ID field in this register is used to create an identifier that can be
compared with the MDA.
NOTE
The logical APIC ID should not be confused with the local APIC ID that
is contained in the local APIC ID register.
31
24 23
0
Logical APIC ID
Reserved
Address: 0FEE0 00D0H
Value after reset: 0000 0000H
Figure 10-19. Logical Destination Register (LDR)
Figure 10-20 shows the layout of the destination format register (DFR). The 4-bit
model field in this register selects one of two models (flat or cluster) that can be used
to interpret the MDA when using logical destination mode.
Vol. 3 10-47
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
31
28
0
Model
Reserved (All 1s)
Flat model: 1111B
Cluster model: 0000B
Address: 0FEE0 00E0H
Value after reset: FFFF FFFFH
Figure 10-20. Destination Format Register (DFR)
The interpretation of MDA for the two models is described in the following para-
graphs.
1. Flat Model — This model is selected by programming DFR bits 28 through 31 to
1111. Here, a unique logical APIC ID can be established for up to 8 local APICs by
setting a different bit in the logical APIC ID field of the LDR for each local APIC. A
group of local APICs can then be selected by setting one or more bits in the MDA.
Each local APIC performs a bit-wise AND of the MDA and its logical APIC ID. If a
true condition is detected, the local APIC accepts the IPI message. A broadcast to
all APICs is achieved by setting the MDA to 1s.
2. Cluster Model — This model is selected by programming DFR bits 28 through 31
to 0000. This model supports two basic destination schemes: flat cluster and
hierarchical cluster.
The flat cluster destination model is only supported for P6 family and Pentium
processors. Using this model, all APICs are assumed to be connected through the
APIC bus. Bits 28 through 31 of the MDA contains the encoded address of the
destination cluster and bits 24 through 27 identify up to four local APICs within
the cluster (each bit is assigned to one local APIC in the cluster, as in the flat
connection model). To identify one or more local APICs, bits 28 through 31 of the
MDA are compared with bits 28 through 31 of the LDR to determine if a local APIC
is part of the cluster. Bits 24 through 27 of the MDA are compared with Bits 24
through 27 of the LDR to identify a local APICs within the cluster.
Sets of processors within a cluster can be specified by writing the target cluster
address in bits 28 through 31 of the MDA and setting selected bits in bits 24
through 27 of the MDA, corresponding to the chosen members of the cluster. In
this mode, 15 clusters (with cluster addresses of 0 through 14) each having 4
local APICs can be specified in the message. For the P6 and Pentium processor’s
local APICs, however, the APIC arbitration ID supports only 15 APIC agents.
Therefore, the total number of processors and their local APICs supported in
this mode is limited to 15. Broadcast to all local APICs is achieved by setting all
destination bits to one. This guarantees a match on all clusters and selects all
APICs in each cluster. A broadcast IPI or I/O subsystem broadcast interrupt with
10-48 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
lowest priority delivery mode is not supported in cluster mode and must not be
configured by software.
The hierarchical cluster destination model can be used with Pentium 4, Intel
Xeon, P6 family, or Pentium processors. With this model, a hierarchical network
can be created by connecting different flat clusters via independent system or
APIC buses. This scheme requires a cluster manager within each cluster, which is
responsible for handling message passing between system or APIC buses. One
cluster contains up to 4 agents. Thus 15 cluster managers, each with 4 agents,
can form a network of up to 60 APIC agents. Note that hierarchical APIC networks
requires a special cluster manager device, which is not part of the local or the I/O
APIC units.
NOTES
All processors that have their APIC software enabled (using the
spurious vector enable/disable bit) must have their DFRs (Desti-
nation Format Registers) programmed identically.
The default mode for DFR is flat mode. If you are using cluster mode,
DFRs must be programmed before the APIC is software enabled.
Since some chipsets do not accurately track a system view of the
logical mode, program DFRs as soon as possible after starting the
processor.
10.7.2.3 Logical Destination Mode in x2APIC Mode
In x2APIC mode, the Logical Destination Register (LDR) is increased to 32 bits wide.
It is a read-only register to system software. This 32-bit value is referred to as
“logical x2APIC ID”. System software accesses this register via the RDMSR instruc-
tion reading the MSR at address 80DH. Figure 10-21 provides the layout of the
Logical Destination Register in x2APIC mode.
MSR Address: 80DH
31
0
Logical x2APIC ID
Figure 10-21. Logical Destination Register in x2APIC Mode
In the xAPIC mode, the Destination Format Register (DFR) through MMIO interface
determines the choice of a flat logical mode or a clustered logical mode. Flat logical
Vol. 3 10-49
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
mode is not supported in the x2APIC mode. Hence the Destination Format Register
(DFR) is eliminated in x2APIC mode.
The 32-bit logical x2APIC ID field of LDR is partitioned into two sub-fields:
• Cluster ID (LDR[31:16]): is the address of the destination cluster
• Logical ID (LDR[15:0]): defines a logical ID of the individual local x2APIC within
the cluster specified by LDR[31:16].
This layout enables 2^16-1 clusters each with up to 16 unique logical IDs - effec-
tively providing an addressability of ((2^20) - 16) processors in logical destination
mode.
It is likely that processor implementations may choose to support less than 16 bits of
the cluster ID or less than 16-bits of the Logical ID in the Logical Destination Register.
However system software should be agnostic to the number of bits implemented in
the cluster ID and logical ID sub-fields. The x2APIC hardware initialization will ensure
that the appropriately initialized logical x2APIC IDs are available to system software
and reads of non-implemented bits return zero. This is a read-only register that soft-
ware must read to determine the logical x2APIC ID of the processor. Specifically,
software can apply a 16-bit mask to the lowest 16 bits of the logical x2APIC ID to
identify the logical address of a processor within a cluster without needing to know
the number of implemented bits in cluster ID and Logical ID sub-fields. Similarly,
the Logical X2APIC ID (31:0) of processors that have matching Cluster ID(31:16).
To enable cluster ID assignment in a fashion that matches the system topology char-
acteristics and to enable efficient routing of logical mode lowest priority device inter-
rupts in link based platform interconnects, the LDR are initialized by hardware based
on the value of x2APIC ID upon x2APIC state transitions. Details of this initialization
are provided in Section 10.5.6.
10.7.2.4 Deriving Logical x2APIC ID from the Local x2APIC ID
In x2APIC mode, the 32-bit logical x2APIC ID, which can be read from LDR, is derived
from the 32-bit local x2APIC ID. Specifically, the 16-bit logical ID sub-field is derived
by shifting 1 by the lowest 4 bits of the x2APIC ID, i.e. Logical ID = 1 << x2APIC
ID[3:0]. The rest of the bits of the x2APIC ID then form the cluster ID portion of the
logical x2APIC ID:
Logical x2APIC ID = [(x2APIC ID[31:4] << 16) | (1 << x2APIC ID[3:0])]
The use of lowest 4 bits in x2APIC ID implies that at least 16 APIC IDs are reserved
for logical processors within a socket in multi-socket configurations. If more than 16
APIC IDS are reserved for logical processors in a socket/package then multiple
cluster IDs can exist within the package.
The LDR initialization occurs whenever the x2APIC mode is enabled. This is described
in Section 10.5.6.
10-50 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
10.7.2.5 Broadcast/Self Delivery Mode
The destination shorthand field of the ICR allows the delivery mode to be by-passed
in favor of broadcasting the IPI to all the processors on the system bus and/or back
to itself (see Section 10.7.1, “Interrupt Command Register (ICR)”). Three destina-
tion shorthands are supported: self, all excluding self, and all including self. The
destination mode is ignored when a destination shorthand is used.
10.7.2.6 Lowest Priority Delivery Mode
With lowest priority delivery mode, the ICR is programmed to send an IPI to several
processors on the system bus, using the logical or shorthand destination mechanism
for selecting the processor. The selected processors then arbitrate with one another
over the system bus or the APIC bus, with the lowest-priority processor accepting the
IPI.
For systems based on the Intel Xeon processor, the chipset bus controller accepts
messages from the I/O APIC agents in the system and directs interrupts to the
processors on the system bus. When using the lowest priority delivery mode, the
chipset chooses a target processor to receive the interrupt out of the set of possible
informs the chipset of the current task priority for each logical processor in the
system. The chipset saves this information and uses it to choose the lowest priority
processor when an interrupt is received.
For systems based on P6 family processors, the processor priority used in lowest-
priority arbitration is contained in the arbitration priority register (APR) in each local
APIC. Figure 10-22 shows the layout of the APR.
31
8 7
4 3
0
Reserved
Arbitration Priority
Arbitration Priority Sub-Class
Address: FEE0 0090H
Value after reset: 0H
Figure 10-22. Arbitration Priority Register (APR)
The APR value is computed as follows:
IF (TPR[7:4] ≥ IRRV[7:4]) AND (TPR[7:4] > ISRV[7:4])
THEN
APR[7:0] ← TPR[7:0]
ELSE
APR[7:4] ← max(TPR[7:4] AND ISRV[7:4], IRRV[7:4])
APR[3:0] ← 0.
Vol. 3 10-51
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
Here, the TPR value is the task priority value in the TPR (see Figure 10-26), the IRRV
value is the vector number for the highest priority bit that is set in the IRR (see
Figure 10-28) or 00H (if no IRR bit is set), and the ISRV value is the vector number
for the highest priority bit that is set in the ISR (see Figure 10-28). Following arbitra-
tion among the destination processors, the processor with the lowest value in its APR
handles the IPI and the other processors ignore it.
(P6 family and Pentium processors.) For these processors, if a focus processor
exists, it may accept the interrupt, regardless of its priority. A processor is said to be
the focus of an interrupt if it is currently servicing that interrupt or if it has a pending
request for that interrupt. For Intel Xeon processors, the concept of a focus processor
is not supported.
In operating systems that use the lowest priority delivery mode but do not update
the TPR, the TPR information saved in the chipset will potentially cause the interrupt
to be always delivered to the same processor from the logical set. This behavior is
functionally backward compatible with the P6 family processor but may result in
unexpected performance implications.
10.7.3
IPI Delivery and Acceptance
When the low double-word of the ICR is written to, the local APIC creates an IPI
message from the information contained in the ICR and sends the message out on
the system bus (Pentium 4 and Intel Xeon processors) or the APIC bus (P6 family and
Pentium processors). The manner in which these IPIs are handled after being issues
in described in Section 10.9, “Handling Interrupts.”
10.7.4
SELF IPI Register
SELF IPIs are used extensively by some system software. The x2APIC architecture
introduces a new register interface. This new register is dedicated to the purpose of
sending self-IPIs with the intent of enabling a highly optimized path for sending self-
IPIs.
Figure 10-23 provides the layout of the SELF IPI register. System software only spec-
ifies the vector associated with the interrupt to be sent. The semantics of sending a
self-IPI via the SELF IPI register are identical to sending a self targeted edge trig-
gered fixed interrupt with the specified vector. Specifically the semantics are identical
to the following settings for an inter-processor interrupt sent via the ICR - Destina-
tion Shorthand (ICR[19:18] = 01 (Self)), Trigger Mode (ICR[15] = 0 (Edge)),
Delivery Mode (ICR[10:8] = 000 (Fixed)), Vector (ICR[7:0] = Vector).
10-52 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
MSR Address: 083FH
31
8 7
0
Vector
Reserved
Figure 10-23. SELF IPI register
The SELF IPI register is a write-only register. A RDMSR instruction with address of the
SELF IPI register will raise a GP fault.
The handling and prioritization of a self-IPI sent via the SELF IPI register is architec-
turally identical to that for an IPI sent via the ICR from a legacy xAPIC unit. Specifi-
cally the state of the interrupt would be tracked via the Interrupt Request Register
(IRR) and In Service Register (ISR) and Trigger Mode Register (TMR) as if it were
received from the system bus. Also sending the IPI via the Self Interrupt Register
ensures that interrupt is delivered to the processor core. Specifically completion of
the WRMSR instruction to the SELF IPI register implies that the interrupt has been
logged into the IRR. As expected for edge triggered interrupts, depending on the
processor priority and readiness to accept interrupts, it is possible that interrupts
sent via the SELF IPI register or via the ICR with identical vectors can be combined.
10.8
SYSTEM AND APIC BUS ARBITRATION
When several local APICs and the I/O APIC are sending IPI and interrupt messages
on the system bus (or APIC bus), the order in which the messages are sent and
handled is determined through bus arbitration.
For the Pentium 4 and Intel Xeon processors, the local and I/O APICs use the arbitra-
tion mechanism defined for the system bus to determine the order in which IPIs are
handled. This mechanism is non-architectural and cannot be controlled by software.
For the P6 family and Pentium processors, the local and I/O APICs use an APIC-based
arbitration mechanism to determine the order in which IPIs are handled. Here, each
local APIC is given an arbitration priority of from 0 to 15, which the I/O APIC uses
during arbitration to determine which local APIC should be given access to the APIC
bus. The local APIC with the highest arbitration priority always wins bus access. Upon
completion of an arbitration round, the winning local APIC lowers its arbitration
priority to 0 and the losing local APICs each raise theirs by 1.
The current arbitration priority for a local APIC is stored in a 4-bit, software-trans-
parent arbitration ID (Arb ID) register. During reset, this register is initialized to the
APIC ID number (stored in the local APIC ID register). The INIT level-deassert IPI,
which is issued with and ICR command, can be used to resynchronize the arbitration
Vol. 3 10-53
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
priorities of the local APICs by resetting Arb ID register of each agent to its current
APIC ID value. (The Pentium 4 and Intel Xeon processors do not implement the Arb
ID register.)
Section 10.11, “APIC Bus Message Passing Mechanism and Protocol (P6 Family,
Pentium Processors),” describes the APIC bus arbitration protocols and bus message
formats, while Section 10.7.1, “Interrupt Command Register (ICR),” describes the
INIT level de-assert IPI message.
Note that except for the SIPI IPI (see Section 10.7.1, “Interrupt Command Register
(ICR)”), all bus messages that fail to be delivered to their specified destination or
destinations are automatically retried. Software should avoid situations in which IPIs
are sent to disabled or nonexistent local APICs, causing the messages to be resent
repeatedly.
10.9
HANDLING INTERRUPTS
When a local APIC receives an interrupt from a local source, an interrupt message
from an I/O APIC, or and IPI, the manner in which it handles the message depends
on processor implementation, as described in the following sections.
10.9.1
Interrupt Handling with the Pentium 4 and Intel Xeon
Processors
With the Pentium 4 and Intel Xeon processors, the local APIC handles the local inter-
rupts, interrupt messages, and IPIs it receives as follows:
1. It determines if it is the specified destination or not (see Figure 10-24). If it is the
specified destination, it accepts the message; if it is not, it discards the message.
Wait to Receive
Bus Message
Belong to
Destination?
Yes
No
Discard
Accept
Message
Message
Figure 10-24. Interrupt Acceptance Flow Chart for the Local APIC (Pentium 4 and
Intel Xeon Processors)
2. If the local APIC determines that it is the designated destination for the interrupt
and if the interrupt request is an NMI, SMI, INIT, ExtINT, or SIPI, the interrupt is
sent directly to the processor core for handling.
10-54 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
but the interrupt request is not one of the interrupts given in step 2, the local
APIC sets the appropriate bit in the IRR.
4. When interrupts are pending in the IRR and ISR register, the local APIC
dispatches them to the processor one at a time, based on their priority and the
current task and processor priorities in the TPR and PPR (see Section 10.9.3.1,
“Task and Processor Priorities”).
5. When a fixed interrupt has been dispatched to the processor core for handling,
the completion of the handler routine is indicated with an instruction in the
instruction handler code that writes to the end-of-interrupt (EOI) register in the
local APIC (see Section 10.9.5, “Signaling Interrupt Servicing Completion”). The
act of writing to the EOI register causes the local APIC to delete the interrupt
from its ISR queue and (for level-triggered interrupts) send a message on the
bus indicating that the interrupt handling has been completed. (A write to the EOI
register must not be included in the handler routine for an NMI, SMI, INIT,
ExtINT, or SIPI.)
10.9.2
Interrupt Handling with the P6 Family and Pentium
Processors
With the P6 family and Pentium processors, the local APIC handles the local inter-
rupts, interrupt messages, and IPIs it receives as follows (see Figure 10-25).
Vol. 3 10-55
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
Wait to Receive
Bus Message
Belong
No
Discard
Message
to
Destination?
Yes
Is it
NMI/SMI/INIT
/ExtINT?
Yes
Accept
Message
No
Lowes
Priority
Fixed
Delivery
P6 Family
Processor Specific
No
Set Status
to Retry
Yes
Am I
Focus?
Accept
Message
Is Interrupt Slot
Available?
Yes
No
No
Yes
Yes
Discard
Message
Is Status a
Retry?
Other
Focus?
No
No
Accept
Message
Is Interrupt
Slot Avail-
able?
Yes
No
Set Status
to Retry
Arbitrate
Yes
Accept
Message
Am I Winner?
Pentium Processors)
1. (IPIs only) It examines the IPI message to determines if it is the specified
destination for the IPI as described in Section 10.7.2, “Determining IPI Desti-
nation.” If it is the specified destination, it continues its acceptance procedure; if
it is not the destination, it discards the IPI message. When the message specifies
lowest-priority delivery mode, the local APIC will arbitrate with the other
processors that were designated on recipients of the IPI message (see Section
10.7.2.6, “Lowest Priority Delivery Mode”).
2. If the local APIC determines that it is the designated destination for the interrupt
and if the interrupt request is an NMI, SMI, INIT, ExtINT, or INIT-deassert
10-56 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
interrupt is sent directly to the processor core for handling.
3. If the local APIC determines that it is the designated destination for the interrupt
but the interrupt request is not one of the interrupts given in step 2, the local
APIC looks for an open slot in one of its two pending interrupt queues contained
slot. If a slot is not available, it rejects the interrupt request and sends it back to
the sender with a retry message.
4. When interrupts are pending in the IRR and ISR register, the local APIC
dispatches them to the processor one at a time, based on their priority and the
current task and processor priorities in the TPR and PPR (see Section 10.9.3.1,
“Task and Processor Priorities”).
5. When a fixed interrupt has been dispatched to the processor core for handling,
the completion of the handler routine is indicated with an instruction in the
instruction handler code that writes to the end-of-interrupt (EOI) register in the
local APIC (see Section 10.9.5, “Signaling Interrupt Servicing Completion”). The
act of writing to the EOI register causes the local APIC to delete the interrupt
from its queue and (for level-triggered interrupts) send a message on the bus
indicating that the interrupt handling has been completed. (A write to the EOI
register must not be included in the handler routine for an NMI, SMI, INIT,
ExtINT, or SIPI.)
The following sections describe the acceptance of interrupts and their handling by the
local APIC and processor in greater detail.
10.9.3
Interrupt, Task, and Processor Priority
For interrupts that are delivered to the processor through the local APIC, each inter-
rupt has an implied priority based on its vector number. The local APIC uses this
priority to determine when to service the interrupt relative to the other activities of
the processor, including the servicing of other interrupts.
For interrupt vectors in the range of 16 to 255, the interrupt priority is determined
using the following relationship:
priority = vector / 16
Here the quotient is rounded down to the nearest integer value to determine the
priority, with 1 being the lowest priority and 15 is the highest. Because vectors 0
through 31 are reserved for dedicated uses by the Intel 64 and IA-32 architectures,
the priorities of user defined interrupts range from 2 to 15.
Each interrupt priority level (sometimes interpreted by software as an interrupt
priority class) encompasses 16 vectors. Prioritizing interrupts within a priority level is
determined by the vector number. The higher the vector number, the higher the
priority within that priority level. In determining the priority of a vector and ranking
Vol. 3 10-57
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
of vectors within a priority group, the vector number is often divided into two parts,
with the high 4 bits of the vector indicating its priority and the low 4 bit indicating its
ranking within the priority group.
10.9.3.1 Task and Processor Priorities
The local APIC also defines a task priority and a processor priority that it uses in
determining the order in which interrupts should be handled. The task priority is a
software selected value between 0 and 15 (see Figure 10-26) that is written into the
task priority register (TPR). The TPR is a read/write register.
31
8 7
4 3
0
Reserved
Task Priority
Task Priority Sub-Class
Address: FEE0 0080H
Value after reset: 0H
Figure 10-26. Task Priority Register (TPR)
NOTE
In this discussion, the term “task” refers to a software defined task,
process, thread, program, or routine that is dispatched to run on the
processor by the operating system. It does not refer to an IA-32
architecture defined task as described in Chapter 7, “Task
Management.”
The task priority allows software to set a priority threshold for interrupting the
processor. The processor will service only those interrupts that have a priority higher
than that specified in the TPR. If software sets the task priority in the TPR to 0, the
processor will handle all interrupts; it is it set to 15, all interrupts are inhibited from
being handled, except those delivered with the NMI, SMI, INIT, ExtINT, INIT-deas-
temporarily block specific interrupts (generally low priority interrupts) from
disturbing high-priority work that the processor is doing.
Note that the task priority is also used to determine the arbitration priority of the
local processor (see Section 10.7.2.6, “Lowest Priority Delivery Mode”).
The processor priority is set by the processor, also to value between 0 and 15 (see
Figure 10-27) that is written into the processor priority register (PPR). The PPR is a
read only register. The processor priority represents the current priority at which the
processor is executing. It is used to determine whether a pending interrupt can be
dispensed to the processor.
10-58 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
31
8 7
4 3
0
Reserved
Processor Priority
Processor Priority Sub-Class
Address: FEE0 00A0H
Value after reset: 0H
Figure 10-27. Processor Priority Register (PPR)
Its value in the PPR is computed as follows:
IF TPR[7:4] ≥ ISRV[7:4]
THEN
PPR[7:0] ← TPR[7:0]
ELSE
PPR[7:4] ← ISRV[7:4]
PPR[3:0] ← 0
Here, the ISRV value is the vector number of the highest priority ISR bit that is set,
or 00H if no ISR bit is set. Essentially, the processor priority is set to either to the
highest priority pending interrupt in the ISR or to the current task priority, whichever
is higher.
10.9.4
Interrupt Acceptance for Fixed Interrupts
The local APIC queues the fixed interrupts that it accepts in one of two interrupt
pending registers: the interrupt request register (IRR) or in-service register (ISR).
These two 256-bit read-only registers are shown in Figure 10-28. The 256 bits in
these registers represent the 256 possible vectors; vectors 0 through 15 are
reserved by the APIC (see also: Section 10.6.2, “Valid Interrupt Vectors”).
NOTE
All interrupts with an NMI, SMI, INIT, ExtINT, start-up, or INIT-
deassert delivery mode bypass the IRR and ISR registers and are
sent directly to the processor core for servicing.
Vol. 3 10-59
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
255
16 15
0
Reserved
Reserved
Reserved
IRR
ISR
TMR
Addresses: IRR FEE0 0200H - FEE0 0270H
ISR FEE0 0100H - FEE0 0170H
TMR FEE0 0180H - FEE0 01F0H
Value after reset: 0H
Figure 10-28. IRR, ISR and TMR Registers
The IRR contains the active interrupt requests that have been accepted, but not yet
dispatched to the processor for servicing. When the local APIC accepts an interrupt,
it sets the bit in the IRR that corresponds the vector of the accepted interrupt. When
the processor core is ready to handle the next interrupt, the local APIC clears the
the highest priority bit set in the ISR is then dispatched to the processor core for
servicing.
While the processor is servicing the highest priority interrupt, the local APIC can send
additional fixed interrupts by setting bits in the IRR. When the interrupt service
routine issues a write to the EOI register (see Section 10.9.5, “Signaling Interrupt
Servicing Completion”), the local APIC responds by clearing the highest priority ISR
bit that is set. It then repeats the process of clearing the highest priority bit in the IRR
and setting the corresponding bit in the ISR. The processor core then begins
executing the service routing for the highest priority bit set in the ISR.
If more than one interrupt is generated with the same vector number, the local APIC
can set the bit for the vector both in the IRR and the ISR. This means that for the
Pentium 4 and Intel Xeon processors, the IRR and ISR can queue two interrupts for
each interrupt vector: one in the IRR and one in the ISR. Any additional interrupts
issued for the same interrupt vector are collapsed into the single bit in the IRR.
For the P6 family and Pentium processors, the IRR and ISR registers can queue no
more than two interrupts per priority level, and will reject other interrupts that are
received within the same priority level.
If the local APIC receives an interrupt with a priority higher than that of the interrupt
currently in serviced, and interrupts are enabled in the processor core, the local APIC
waiting for a write to the EOI register). The currently executing interrupt handler is
then interrupted so the higher-priority interrupt can be handled. When the handling
of the higher-priority interrupt has been completed, the servicing of the interrupted
interrupt is resumed.
The trigger mode register (TMR) indicates the trigger mode of the interrupt (see
Figure 10-28). Upon acceptance of an interrupt into the IRR, the corresponding TMR
10-60 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
bit is cleared for edge-triggered interrupts and set for level-triggered interrupts. If a
TMR bit is set when an EOI cycle for its corresponding interrupt vector is generated,
an EOI message is sent to all I/O APICs.
10.9.5
Signaling Interrupt Servicing Completion
For all interrupts except those delivered with the NMI, SMI, INIT, ExtINT, the start-
up, or INIT-Deassert delivery mode, the interrupt handler must include a write to the
end-of-interrupt (EOI) register (see Figure 10-29). This write must occur at the end
of the handler routine, sometime before the IRET instruction. This action indicates
that the servicing of the current interrupt is complete and the local APIC can issue the
next interrupt from the ISR.
31
0
Address: 0FEE0 00B0H
Value after reset: 0H
Figure 10-29. EOI Register
Upon receiving and EOI, the APIC clears the highest priority bit in the ISR and
dispatches the next highest priority interrupt to the processor. If the terminated
interrupt was a level-triggered interrupt, the local APIC also sends an end-of-inter-
rupt message to all I/O APICs.
10.9.5.1 Signaling Interrupt Servicing Completion in x2APIC Mode
In the x2APIC mode, the write of a zero value to EOI register is enforced. Writes of a
non-zero value to the EOI register in x2APIC mode will raise a GP fault. System soft-
ware continues to have to perform the EOI write to indicate interrupt service comple-
tion. But in x2APIC mode, the EOI write is with a value of zero.
10.9.6
Task Priority in IA-32e Mode
In IA-32e mode, operating systems can manage the 16 priority classes of external
interrupts (see Section 10.9.3, “Interrupt, Task, and Processor Priority”) explicitly
using the task priority register (TPR). Operating systems can use the TPR to tempo-
rarily block specific (low-priority) interrupts from interrupting a high-priority task.
This is done by loading TPR with a value corresponding to the highest-priority inter-
rupt that is to be blocked. For example:
Vol. 3 10-61
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
8 or less while allowing all interrupts with a priority of nine or more to be
recognized.
• Loading the TPR with zero enables all external interrupts.
• Loading the TPR with 0F (01111B) disables all external interrupts.
The TPR (shown in Figure 10-26) is cleared to 0 on reset. In 64-bit mode, software
can read and write the TPR using an alternate interface, MOV CR8 instruction. The
new priority level is established when the MOV CR8 instruction completes execution.
Software does not need to force serialization after loading the TPR using MOV CR8.
Use of the MOV CRn instruction requires a privilege level of 0. Programs running at
privilege level greater than 0 cannot read or write the TPR. An attempt to do so
results in a general-protection exception, #GP(0). The TPR is abstracted from the
the processor. The IC can be an external device, such as an APIC or 8259. Typically,
the IC provides a priority mechanism similar or identical to the TPR. The IC, however,
is considered implementation-dependent with the under-lying priority mechanisms
subject to change. CR8, by contrast, is part of the Intel 64 architecture. Software can
depend on this definition remaining unchanged.
Figure 10-30 shows the layout of CR8; only the low four bits are used. The remaining
60 bits are reserved and must be written with zeros. Failure to do this results in a
general-protection exception, #GP(0).
63
0
4 3
Reserved
Value after reset: 0H
Figure 10-30. CR8 Register
10.9.6.1 Interaction of Task Priorities between CR8 and APIC
The first implementation of Intel 64 architecture includes a local advanced program-
mable interrupt controller (APIC) that is similar to the APIC used with previous IA-32
processors. Some aspects of the local APIC affect the operation of the architecturally
defined task priority register and the programming interface using CR8.
Notable CR8 and APIC interactions are:
• The processor powers up with the local APIC enabled.
• The APIC must be enabled for CR8 to function as the TPR. Writes to CR8 are
reflected into the APIC Task Priority Register.
• APIC.TPR[bits 7:4] = CR8[bits 3:0], APIC.TPR[bits 3:0] = 0. A read of CR8
returns a 64-bit value which is the value of TPR[bits 7:4], zero extended to 64
bits.
10-62 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
There are no ordering mechanisms between direct updates of the APIC.TPR and CR8.
Operating software should implement either direct APIC TPR updates or CR8 style
TPR updates but not mix them. Software can use a serializing instruction (for
example, CPUID) to serialize updates between MOV CR8 and stores to the APIC.
10.10 SPURIOUS INTERRUPT
A special situation may occur when a processor raises its task priority to be greater
than or equal to the level of the interrupt for which the processor INTR signal is
currently being asserted. If at the time the INTA cycle is issued, the interrupt that
was to be dispensed has become masked (programmed by software), the local APIC
will deliver a spurious-interrupt vector. Dispensing the spurious-interrupt vector does
not affect the ISR, so the handler for this vector should return without an EOI.
The vector number for the spurious-interrupt vector is specified in the spurious-inter-
rupt vector register (see Figure 10-31). The functions of the fields in this register are
as follows:
Spurious Vector Determines the vector number to be delivered to the processor
when the local APIC generates a spurious vector.
(Pentium 4 and Intel Xeon processors.) Bits 0 through 7 of the
this field are programmable by software.
(P6 family and Pentium processors). Bits 4 through 7 of the this
field are programmable by software, and bits 0 through 3 are
hardwired to logical ones. Software writes to bits 0 through 3
have no effect.
APIC Software
Enable/Disable
Allows software to temporarily enable (1) or disable (0) the local
APIC (see Section 10.4.3, “Enabling or Disabling the Local
APIC”).
Focus Processor Determines if focus processor checking is enabled (0) or
disabled (1)
Checking
when using the lowest-priority delivery mode. In Pentium 4 and
Intel Xeon processors, this bit is reserved and should be cleared
to 0.
NOTE
Do not program an LVT or IOAPIC RTE with a spurious vector even if
you set the mask bit. A spurious vector ISR does not do an EOI. If for
some reason an interrupt is generated by an LVT or RTE entry, the bit
in the in-service register will be left set for the spurious vector. This
will mask all interrupts at the same or lower priority
Vol. 3 10-63
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
31
10
9
8
7
0
Reserved
Focus Processor Checking1
0: Enabled
1: Disabled
APIC Software Enable/Disable
0: APIC Disabled
1: APIC Enabled
Spurious Vector2
Address: FEE0 00F0H
Value after reset: 0000 00FFH
1. Not supported in Pentium 4 and Intel Xeon processors.
2. For the P6 family and Pentium processors, bits 0 through 3
of the spurious vector are hardwired to 1.
Figure 10-31. Spurious-Interrupt Vector Register (SVR)
10.11 APIC BUS MESSAGE PASSING MECHANISM AND
PROTOCOL (P6 FAMILY, PENTIUM PROCESSORS)
The Pentium 4 and Intel Xeon processors pass messages among the local and I/O
APICs on the system bus, using the system bus message passing mechanism and
protocol.
The P6 family and Pentium processors, pass messages among the local and I/O
APICs on the serial APIC bus, as follows. Because only one message can be sent at a
time on the APIC bus, the I/O APIC and local APICs employ a “rotating priority” arbi-
tration protocol to gain permission to send a message on the APIC bus. One or more
APICs may start sending their messages simultaneously. At the beginning of every
message, each APIC presents the type of the message it is sending and its current
arbitration priority on the APIC bus. This information is used for arbitration. After
each arbitration cycle (within an arbitration round), only the potential winners keep
driving the bus. By the time all arbitration cycles are completed, there will be only
one APIC left driving the bus. Once a winner is selected, it is granted exclusive use of
the bus, and will continue driving the bus to send its actual message.
After each successfully transmitted message, all APICs increase their arbitration
priority by 1. The previous winner (that is, the one that has just successfully trans-
mitted its message) assumes a priority of 0 (lowest). An agent whose arbitration
priority was 15 (highest) during arbitration, but did not send a message, adopts the
previous winner’s arbitration priority, increments by 1.
Note that the arbitration protocol described above is slightly different if one of the
APICs issues a special End-Of-Interrupt (EOI). This high-priority message is granted
10-64 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
the bus regardless of its sender’s arbitration priority, unless more than one APIC
issues an EOI message simultaneously. In the latter case, the APICs sending the EOI
messages arbitrate using their arbitration priorities.
If the APICs are set up to use “lowest priority” arbitration (see Section 10.7.2.6,
“Lowest Priority Delivery Mode”) and multiple APICs are currently executing at the
lowest priority (the value in the APR register), the arbitration priorities (unique
values in the Arb ID register) are used to break ties. All 8 bits of the APR are used for
the lowest priority arbitration.
10.11.1 Bus Message Formats
See Appendix F, “APIC Bus Message Formats,” for a description of bus message
formats used to transmit messages on the serial APIC bus.
10.12 MESSAGE SIGNALLED INTERRUPTS
The PCI Local Bus Specification, Rev 2.2 (www.pcisig.com) introduces the concept of
message signalled interrupts. Intel processors and chipsets with this capability
currently include the Pentium 4 and Intel Xeon processors. As the specification indi-
cates:
“Message signalled interrupts (MSI) is an optional feature that
enables PCI devices to request service by writing a system-specified
message to a system-specified address (PCI DWORD memory write
transaction). The transaction address specifies the message
destination while the transaction data specifies the message. System
software is expected to initialize the message destination and
message during device configuration, allocating one or more non-
shared messages to each MSI capable function.”
identify and configure MSI capable PCI devices. Among other fields, this structure
contains a Message Data Register and a Message Address Register. To request
service, the PCI device function writes the contents of the Message Data Register to
the address contained in the Message Address Register (and the Message Upper
Address register for 64-bit message addresses).
Section 10.12.1 and Section 10.12.2 provide layout details for the Message Address
Register and the Message Data Register. The operation issued by the device is a PCI
write command to the Message Address Register with the Message Data Register
contents. The operation follows semantic rules as defined for PCI write operations
and is a DWORD operation.
Vol. 3 10-65
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
10.12.1 Message Address Register Format
The format of the Message Address Register (lower 32-bits) is shown in
Figure 10-32.
31
20 19
12 11
4
3
2
1
0
0FEEH
Destination ID Reserved
RH
DM
XX
Figure 10-32. Layout of the MSI Message Address Register
Fields in the Message Address Register are as follows:
1. Bits 31-20 — These bits contain a fixed value for interrupt messages (0FEEH).
This value locates interrupts at the 1-MByte area with a base address of 4G –
18M. All accesses to this region are directed as interrupt messages. Care must to
be taken to ensure that no other device claims the region as I/O space.
2. Destination ID — This field contains an 8-bit destination ID. It identifies the
message’s target processor(s). The destination ID corresponds to bits 63:56 of
the I/O APIC Redirection Table Entry if the IOAPIC is used to dispatch the
interrupt to the processor(s).
3. Redirection hint indication (RH) — This bit indicates whether the message
should be directed to the processor with the lowest interrupt priority among
processors that can receive the interrupt.
•
When RH is 0, the interrupt is directed to the processor listed in the
Destination ID field.
•
When RH is 1 and the physical destination mode is used, the Destination
ID field must not be set to 0xFF; it must point to a processor that is
present and enabled to receive the interrupt.
•
•
When RH is 1 and the logical destination mode is active in a system using
a flat addressing model, the Destination ID field must be set so that bits
set to 1 identify processors that are present and enabled to receive the
interrupt.
If RH is set to 1 and the logical destination mode is active in a system
using cluster addressing model, then Destination ID field must not be set
to 0xFF; the processors identified with this field must be present and
enabled to receive the interrupt.
4. Destination mode (DM) — This bit indicates whether the Destination ID field
should be interpreted as logical or physical APIC ID for delivery of the lowest
priority interrupt. If RH is 1 and DM is 0, the Destination ID field is in physical
10-66 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
destination mode and only the processor in the system that has the matching
APIC ID is considered for delivery of that interrupt (this means no re-direction).
If RH is 1 and DM is 1, the Destination ID Field is interpreted as in logical
part of the logical group of processors based on the processor’s logical APIC ID
and the Destination ID field in the message. The logical group of processors
consists of those identified by matching the 8-bit Destination ID with the logical
destination identified by the Destination Format Register and the Logical
Destination Register in each local APIC. The details are similar to those described
ignored and the message is sent ahead independent of whether the physical or
logical destination mode is used.
10.12.2 Message Data Register Format
The layout of the Message Data Register is shown in Figure 10-33.
63
32
0
Reserved
31
16
15
14
13
11 10
8
7
Reserved
Reserved
Vector
Trigger Mode
0 - Edge
1 - Level
Delivery Mode
000 - Fixed
001 - Lowest Priority
010 - SMI
011 - Reserved
100 - NMI
101 - INIT
110 - Reserved
111 - ExtINT
Level for Trigger Mode = 0
X - Don’t care
Level for Trigger Mode = 1
0 - Deassert
1 - Assert
Figure 10-33. Layout of the MSI Message Data Register
Vol. 3 10-67
Download from Www.Somanuals.com. All Manuals Search And Download.
ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC)
Reserved fields are not assumed to be any value. Software must preserve their
contents on writes. Other fields in the Message Data Register are described below.
1. Vector — This 8-bit field contains the interrupt vector associated with the
message. Values range from 010H to 0FEH. Software must guarantee that the
field is not programmed with vector 00H to 0FH.
2. Delivery Mode — This 3-bit field specifies how the interrupt receipt is handled.
Delivery Modes operate only in conjunction with specified Trigger Modes. Correct
Trigger Modes must be guaranteed by software. Restrictions are indicated below:
a. 000B (Fixed Mode) — Deliver the signal to all the agents listed in the
destination. The Trigger Mode for fixed delivery mode can be edge or level.
b. 001B (Lowest Priority) — Deliver the signal to the agent that is executing
at the lowest priority of all agents listed in the destination field. The trigger
mode can be edge or level.
c. 010B (System Management Interrupt or SMI) — The delivery mode is
edge only. For systems that rely on SMI semantics, the vector field is ignored
but must be programmed to all zeroes for future compatibility.
d. 100B (NMI) — Deliver the signal to all the agents listed in the destination
field. The vector information is ignored. NMI is an edge triggered interrupt
regardless of the Trigger Mode Setting.
e. 101B (INIT) — Deliver this signal to all the agents listed in the destination
field. The vector information is ignored. INIT is an edge triggered interrupt
regardless of the Trigger Mode Setting.
f. 111B (ExtINT) — Deliver the signal to the INTR signal of all agents in the
destination field (as an interrupt that originated from an 8259A compatible
interrupt controller). The vector is supplied by the INTA cycle issued by the
activation of the ExtINT. ExtINT is an edge triggered interrupt.
3. Level — Edge triggered interrupt messages are always interpreted as assert
messages. For edge triggered interrupts this field is not used. For level triggered
interrupts, this bit reflects the state of the interrupt input.
4. Trigger Mode — This field indicates the signal type that will trigger a message.
a. 0 — Indicates edge sensitive.
b. 1 — Indicates level sensitive.
10-68 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
CHAPTER 11
MEMORY CACHE CONTROL
This chapter describes the memory cache and cache control mechanisms, the TLBs,
and the store buffer in Intel 64 and IA-32 processors. It also describes the memory
type range registers (MTRRs) introduced in the P6 family processors and how they
are used to control caching of physical memory locations.
11.1
INTERNAL CACHES, TLBS, AND BUFFERS
The Intel 64 and IA-32 architectures support cache, translation look aside buffers
(TLBs), and a store buffer for temporary on-chip (and external) storage of instruc-
tions and data. (Figure 11-1 shows the arrangement of caches, TLBs, and the store
buffer for the Pentium 4 and Intel Xeon processors.) Table 11-1 shows the character-
istics of these caches and buffers for the Pentium 4, Intel Xeon, P6 family, and
Pentium processors. The sizes and characteristics of these units are machine
specific and may change in future versions of the processor. The CPUID
instruction returns the sizes and characteristics of the caches and buffers for the
processor on which the instruction is executed. See “CPUID—CPU Identification” in
Chapter 3, “Instruction Set Reference, A-M,” of the Intel® 64 and IA-32 Architec-
tures Software Developer’s Manual, Volume 2A.
Physical
Memory
System Bus
(External)
Data Cache
Unit (L1)
L2 Cache
L3 Cache†
Instruction
TLBs
Bus Interface Unit
Data TLBs
Instruction Decoder
Trace Cache
Store Buffer
† Intel Xeon processors only
Figure 11-1. Cache Structure of the Pentium 4 and Intel Xeon Processors
Vol. 3 11-1
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
Instruction
Cache
Instruction Decoder and front end
ITLB
Chipset
Out-of-Order Engine
Data TLB
QPI
STLB
IMC
Data Cache
Unit (L1)
L2 Cache
L3 Cache
Figure 11-2. Cache Structure of the Intel Core i7 Processors
Figure 11-2 shows the cache arrangement of Intel Core i7 processor.
Table 11-1. Characteristics of the Caches, TLBs, Store Buffer, and
Write Combining Buffer in Intel 64 and IA-32 Processors
Cache or Buffer
Characteristics
1
Trace Cache
•
•
Pentium 4 and Intel Xeon processors (Based on Intel NetBurst
microarchitecture): 12 Kμops, 8-way set associative.
Intel Core i7, Intel Core 2 Duo, Intel Atom, Intel Core Duo, Intel Core Solo,
Pentium M processor: not implemented.
•
•
P6 family and Pentium processors: not implemented.
L1 Instruction Cache
Pentium 4 and Intel Xeon processors (Based on Intel NetBurst
microarchitecture): not implemented.
Intel Core i7 processor: 32-KByte, 4-way set associative.
Intel Core 2 Duo, Intel Atom, Intel Core Duo, Intel Core Solo, Pentium M
processor: 32-KByte, 8-way set associative.
P6 family and Pentium processors: 8- or 16-KByte, 4-way set associative,
32-byte cache line size; 2-way set associative for earlier Pentium
processors.
•
•
•
11-2 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
Table 11-1. Characteristics of the Caches, TLBs, Store Buffer, and
Write Combining Buffer in Intel 64 and IA-32 Processors (Contd.)
Cache or Buffer Characteristics
L1 Data Cache
•
Pentium 4 and Intel Xeon processors (Based on Intel NetBurst
microarchitecture): 8-KByte, 4-way set associative, 64-byte cache line
size.
Pentium 4 and Intel Xeon processors (Based on Intel NetBurst
microarchitecture): 16-KByte, 8-way set associative, 64-byte cache line
size.
Intel Atom processors: 24-KByte, 6-way set associative, 64-byte cache
line size.
•
•
•
Intel Core i7, Intel Core 2 Duo, Intel Core Duo, Intel Core Solo, Pentium M
and Intel Xeon processors: 32-KByte, 8-way set associative, 64-byte
cache line size.
•
•
P6 family processors: 16-KByte, 4-way set associative, 32-byte cache
line size; 8-KBytes, 2-way set associative for earlier P6 family
processors.
Pentium processors: 16-KByte, 4-way set associative, 32-byte cache line
size; 8-KByte, 2-way set associative for earlier Pentium processors.
L2 Unified Cache
•
•
•
•
•
•
•
•
•
Intel Core 2 Duo and Intel Xeon processors: up to 4-MByte (or 4MBx2 in
quadcore processors), 16-way set associative, 64-byte cache line size.
Intel Core 2 Duo and Intel Xeon processors: up to 6-MByte (or 6MBx2 in
quadcore processors), 24-way set associative, 64-byte cache line size.
Intel Core i7 processor: 256KBbyte, 8-way set associative, 64-byte cache
line size.
Intel Atom processors: 512-KByte, 8-way set associative, 64-byte cache
line size.
Intel Core Duo, Intel Core Solo processors: 2-MByte, 8-way set
associative, 64-byte cache line size
Pentium 4 and Intel Xeon processors: 256, 512, 1024, or 2048-KByte, 8-
way set associative, 64-byte cache line size, 128-byte sector size.
Pentium M processor: 1 or 2-MByte, 8-way set associative, 64-byte
cache line size.
P6 family processors: 128-KByte, 256-KByte, 512-KByte, 1-MByte, or 2-
MByte, 4-way set associative, 32-byte cache line size.
Pentium processor (external optional): System specific, typically 256- or
512-KByte, 4-way set associative, 32-byte cache line size.
L3 Unified Cache
•
•
Intel Xeon processors: 512-KByte, 1-MByte, 2-MByte, or 4-MByte, 8-way
set associative, 64-byte cache line size, 128-byte sector size.
Intel Core i7 processor: Up to 8MByte, 16-way set associative, 64-byte
cache line size.
Vol. 3 11-3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
Table 11-1. Characteristics of the Caches, TLBs, Store Buffer, and
Write Combining Buffer in Intel 64 and IA-32 Processors (Contd.)
Cache or Buffer Characteristics
Instruction TLB
(4-KByte Pages)
•
Pentium 4 and Intel Xeon processors (Based on Intel NetBurst
microarchitecture): 128 entries, 4-way set associative.
Intel Atom processors: 32-entries, fully associative.
Intel Core i7 processor: 64-entries per thread (128-entries per core), 4-
way set associative.
•
•
•
Intel Core 2 Duo, Intel Core Duo, Intel Core Solo processors, Pentium M
processor: 128 entries, 4-way set associative.
•
•
P6 family processors: 32 entries, 4-way set associative.
Pentium processor: 32 entries, 4-way set associative; fully set
associative for Pentium processors with MMX technology.
Data TLB (4-KByte
Pages)
•
•
Intel Core i7 processor, DTLB0: 64-entries, 4-way set associative.
Intel Core 2 Duo processors: DTLB0, 16 entries, DTLB1, 256 entries, 4
ways.
Intel Atom processors: 16-entry-per-thread micro-TLB, fully associative;
64-entry DTLB, 4-way set associative; 16-entry PDE cache, fully
associative.
Pentium 4 and Intel Xeon processors (Based on Intel NetBurst
microarchitecture): 64 entry, fully set associative, shared with large page
DTLB.
Intel Core Duo, Intel Core Solo processors, Pentium M processor: 128
entries, 4-way set associative.
Pentium and P6 family processors: 64 entries, 4-way set associative;
fully set, associative for Pentium processors with MMX technology.
•
•
•
•
Instruction TLB
(Large Pages)
•
•
•
•
Intel Core i7 processor: 7-entries per thread, fully associative.
Intel Core 2 Duo processors: 4 entries, 4 ways.
Pentium 4 and Intel Xeon processors: large pages are fragmented.
Intel Core Duo, Intel Core Solo, Pentium M processor: 2 entries, fully
associative.
•
•
P6 family processors: 2 entries, fully associative.
Pentium processor: Uses same TLB as used for 4-KByte pages.
Data TLB (Large
Pages)
•
•
Intel Core i7 processor, DTLB0: 32-entries, 4-way set associative.
Intel Core 2 Duo processors: DTLB0, 16 entries, DTLB1, 32 entries, 4
ways.
Intel Atom processors: 8 entries, 4-way set associative.
Pentium 4 and Intel Xeon processors: 64 entries, fully set associative;
shared with small page data TLBs.
Intel Core Duo, Intel Core Solo, Pentium M processor: 8 entries, fully
associative.
•
•
•
•
•
P6 family processors: 8 entries, 4-way set associative.
Pentium processor: 8 entries, 4-way set associative; uses same TLB as
used for 4-KByte pages in Pentium processors with MMX technology.
Second-levelUnified
TLB (4-KByte
Pages)
•
Intel Core i7 processor, STLB: 512-entries, 4-way set associative.
11-4 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
Table 11-1. Characteristics of the Caches, TLBs, Store Buffer, and
Write Combining Buffer in Intel 64 and IA-32 Processors (Contd.)
Cache or Buffer Characteristics
Store Buffer
•
•
•
•
•
•
•
Intel Core i7 processors: 32entries.
Intel Core 2 Duo processors: 20 entries.
Intel Atom processors: 8 entries, used for both WC and store buffers.
Pentium 4 and Intel Xeon processors: 24 entries.
Pentium M processor: 16 entries.
P6 family processors: 12 entries.
Pentium processor: 2 buffers, 1 entry each (Pentium processors with
MMX technology have 4 buffers for 4 entries).
Write Combining
(WC) Buffer
•
•
•
•
•
Intel Core 2 Duo processors: 8 entries.
Intel Atom processors: 8 entries, used for both WC and store buffers.
Pentium 4 and Intel Xeon processors: 6 or 8 entries.
Intel Core Duo, Intel Core Solo, Pentium M processors: 6 entries.
P6 family processors: 4 entries.
NOTES:
1 Introduced to the IA-32 architecture in the Pentium 4 and Intel Xeon processors.
Intel 64 and IA-32 processors may implement four types of caches: the trace cache,
the level 1 (L1) cache, the level 2 (L2) cache, and the level 3 (L3) cache. See
Figure 11-1. Cache availability is described below:
• Intel Core i7 processor Family — The L1 cache is divided into two sections:
one section is dedicated to caching instructions (pre-decoded instructions) and
the other caches data. The L2 cache is a unified data and instruction cache. Each
processor core has its own L1 and L2. The L3 cache is an inclusive, unified data
and instruction cache, shared by all processor cores inside a physical package. No
trace cache is implemented.
• Intel Core 2 processor and Intel Xeon processor Family based on Intel
Core microarchitecture — The L1 cache is divided into two sections: one
section is dedicated to caching instructions (pre-decoded instructions) and the
other caches data. The L2 cache is a unified data and instruction cache located on
the processor chip; it is shared between two processor cores in a dual-core
processor implementation. Quad-core processors have two L2, each shared by
two processor cores. No trace cache is implemented.
• Intel Atom processor — The L1 cache is divided into two sections: one section
is dedicated to caching instructions (pre-decoded instructions) and the other
caches data. The L2 cache is a unified data and instruction cache is located on the
processor chip. No trace cache is implemented.
• Intel Core Solo and Intel Core Duo processors — The L1 cache is divided into
two sections: one section is dedicated to caching instructions (pre-decoded
instructions) and the other caches data. The L2 cache is a unified data and
instruction cache located on the processor chip. It is shared between two
processor cores in a dual-core processor implementation. No trace cache is
implemented.
Vol. 3 11-5
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
• Pentium 4 and Intel Xeon processors Based on Intel NetBurst microar-
chitecture — The trace cache caches decoded instructions (μops) from the
instruction decoder and the L1 cache contains data. The L2 and L3 caches are
unified data and instruction caches located on the processor chip. Dualcore
processors have two L2, one in each processor core. Note that the L3 cache is
only implemented on some Intel Xeon processors.
• P6 family processors — The L1 cache is divided into two sections: one
dedicated to caching instructions (pre-decoded instructions) and the other to
caching data. The L2 cache is a unified data and instruction cache located on the
processor chip. P6 family processors do not implement a trace cache.
• Pentium processors — The L1 cache has the same structure as on P6 family
processors. There is no trace cache. The L2 cache is a unified data and instruction
cache external to the processor chip on earlier Pentium processors and
implemented on the processor chip in later Pentium processors. For Pentium
processors where the L2 cache is external to the processor, access to the cache is
through the system bus.
For Intel Core i7 processors and processors based on Intel Core, Intel Atom, and Intel
NetBurst microarchitectures, Intel Core Duo, Intel Core Solo and Pentium M proces-
sors, the cache lines for the L1 and L2 caches (and L3 caches if supported) are 64
bytes wide. The processor always reads a cache line from system memory beginning
on a 64-byte boundary. (A 64-byte aligned cache line begins at an address with its 6
least-significant bits clear.) A cache line can be filled from memory with a 8-transfer
burst transaction. The caches do not support partially-filled cache lines, so caching
even a single doubleword requires caching an entire line.
The L1 and L2 cache lines in the P6 family and Pentium processors are 32 bytes wide,
with cache line reads from system memory beginning on a 32-byte boundary (5
least-significant bits of a memory address clear.) A cache line can be filled from
memory with a 4-transfer burst transaction. Partially-filled cache lines are not
supported.
The trace cache in processors based on Intel NetBurst microarchitecture is available
in all execution modes: protected mode, system management mode (SMM), and
real-address mode. The L1,L2, and L3 caches are also available in all execution
modes; however, use of them must be handled carefully in SMM (see Section 26.4.2,
“SMRAM Caching”).
The TLBs store the most recently used page-directory and page-table entries. They
speed up memory accesses when paging is enabled by reducing the number of
memory. The TLBs are divided into four groups: instruction TLBs for 4-KByte pages,
data TLBs for 4-KByte pages; instruction TLBs for large pages (2-MByte or 4-MByte
pages), and data TLBs for large pages. The TLBs are normally active only in protected
mode with paging enabled. When paging is disabled or the processor is in real-
address mode, the TLBs maintain their contents until explicitly or implicitly flushed
(see Section 11.9, “Invalidating the Translation Lookaside Buffers (TLBs)”).
11-6 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
Processors based on Intel Core microarchitectures implement one level of instruction
TLB and two levels of data TLB. Intel Core i7 processor provides a second-level
unified TLB.
The store buffer is associated with the processors instruction execution units. It
allows writes to system memory and/or the internal caches to be saved and in some
cases combined to optimize the processor’s bus accesses. The store buffer is always
enabled in all execution modes.
The processor’s caches are for the most part transparent to software. When enabled,
instructions and data flow through these caches without the need for explicit soft-
ware control. However, knowledge of the behavior of these caches may be useful in
optimizing software performance. For example, knowledge of cache dimensions and
replacement algorithms gives an indication of how large of a data structure can be
operated on at once without causing cache thrashing.
In multiprocessor systems, maintenance of cache consistency may, in rare circum-
stances, require intervention by system software. For these rare cases, the processor
provides privileged cache control instructions for use in flushing caches and forcing
memory ordering.
The Pentium III, Pentium 4, and Intel Xeon processors introduced several instructions
that software can use to improve the performance of the L1, L2, and L3 caches,
including the PREFETCHh and CLFLUSH instructions and the non-temporal move
instructions (MOVNTI, MOVNTQ, MOVNTDQ, MOVNTPS, and MOVNTPD). The use of
these instructions are discussed in Section 11.5.5, “Cache Management Instruc-
tions.”
11.2
CACHING TERMINOLOGY
IA-32 processors (beginning with the Pentium processor) and Intel 64 processors use
the MESI (modified, exclusive, shared, invalid) cache protocol to maintain consis-
tency with internal caches and caches in other processors (see Section 11.4, “Cache
Control Protocol”).
When the processor recognizes that an operand being read from memory is cache-
able, the processor reads an entire cache line into the appropriate cache (L1, L2, L3,
or all). This operation is called a cache line fill. If the memory location containing
that operand is still cached the next time the processor attempts to access the
operand, the processor can read the operand from the cache instead of going back to
memory. This operation is called a cache hit.
When the processor attempts to write an operand to a cacheable area of memory, it
first checks if a cache line for that memory location exists in the cache. If a valid
cache line does exist, the processor (depending on the write policy currently in force)
can write the operand into the cache instead of writing it out to system memory. This
operation is called a write hit. If a write misses the cache (that is, a valid cache line
is not present for area of memory being written to), the processor performs a cache
line fill, write allocation. Then it writes the operand into the cache line and
Vol. 3 11-7
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
(depending on the write policy currently in force) can also write it out to memory. If
the operand is to be written out to memory, it is written first into the store buffer, and
then written from the store buffer to memory when the system bus is available.
(Note that for the Pentium processor, write misses do not result in a cache line fill;
they always result in a write to memory. For this processor, only read misses result in
cache line fills.)
When operating in an MP system, IA-32 processors (beginning with the Intel486
processor) and Intel 64 processors have the ability to snoop other processor’s
accesses to system memory and to their internal caches. They use this snooping
ability to keep their internal caches consistent both with system memory and with
the caches in other processors on the bus. For example, in the Pentium and P6 family
processors, if through snooping one processor detects that another processor
intends to write to a memory location that it currently has cached in shared state,
the snooping processor will invalidate its cache line forcing it to perform a cache line
fill the next time it accesses the same memory location.
Beginning with the P6 family processors, if a processor detects (through snooping)
that another processor is trying to access a memory location that it has modified in
its cache, but has not yet written back to system memory, the snooping processor
will signal the other processor (by means of the HITM# signal) that the cache line is
held in modified state and will preform an implicit write-back of the modified data.
The implicit write-back is transferred directly to the initial requesting processor and
snooped by the memory controller to assure that system memory has been updated.
Here, the processor with the valid data may pass the data to the other processors
without actually writing it to system memory; however, it is the responsibility of the
memory controller to snoop this operation and update memory.
11.3
METHODS OF CACHING AVAILABLE
The processor allows any area of system memory to be cached in the L1, L2, and L3
caches. In individual pages or regions of system memory, it allows the type of
caching (also called memory type) to be specified (see Section 11.5). Memory types
currently defined for the Intel 64 and IA-32 architectures are (see Table 11-2):
• Strong Uncacheable (UC) —System memory locations are not cached. All
reads and writes appear on the system bus and are executed in program order
without reordering. No speculative memory accesses, page-table walks, or
prefetches of speculated branch targets are made. This type of cache-control is
useful for memory-mapped I/O devices. When used with normal RAM, it greatly
reduces processor performance.
NOTE
The behavior of FP and SSE/SSE2 operations on operands in UC
memory is implementation dependent. In some implementations,
accesses to UC memory may occur more than once. To ensure
predictable behavior, use loads and stores of general purpose
11-8 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
registers to access UC memory that may have read or write side
effects.
Table 11-2. Memory Types and Their Properties
Memory Type and
Mnemonic
Cacheable
Writeback Allows
Cacheable Speculative
Reads
Memory Ordering Model
Strong Uncacheable
(UC)
No
No
No
No
Strong Ordering
Uncacheable (UC-)
No
No
Strong Ordering. Can only be
selected through the PAT. Can
be overridden by WC in MTRRs.
Write Combining (WC) No
No
Yes
Weak Ordering. Available by
programming MTRRs or by
selecting it through the PAT.
Write Through (WT)
Write Back (WB)
Yes
Yes
No
Yes
No
Yes
Yes
Yes
Speculative Processor Ordering.
Speculative Processor Ordering.
Write Protected (WP) Yes for
reads; no for
writes
Speculative Processor Ordering.
Available by programming
MTRRs.
• Uncacheable (UC-) — Has same characteristics as the strong uncacheable (UC)
memory type, except that this memory type can be overridden by programming
the MTRRs for the WC memory type. This memory type is available in processor
families starting from the Pentium III processors and can only be selected through
the PAT.
• Write Combining (WC) — System memory locations are not cached (as with
uncacheable memory) and coherency is not enforced by the processor’s bus
coherency protocol. Speculative reads are allowed. Writes may be delayed and
If the WC buffer is partially filled, the writes may be delayed until the next
occurrence of a serializing event; such as, an SFENCE or MFENCE instruction,
CPUID execution, a read or write to uncached memory, an interrupt occurrence,
or a LOCK instruction execution. This type of cache-control is appropriate for
video frame buffers, where the order of writes is unimportant as long as the
writes update memory so they can be seen on the graphics display. See Section
11.3.1, “Buffering of Write Combining Memory Locations,” for more information
about caching the WC memory type. This memory type is available in the
Pentium Pro and Pentium II processors by programming the MTRRs; or in
processor families starting from the Pentium III processors by programming the
MTRRs or by selecting it through the PAT.
• Write-through (WT) — Writes and reads to and from system memory are
cached. Reads come from cache lines on cache hits; read misses cause cache
fills. Speculative reads are allowed. All writes are written to a cache line (when
Vol. 3 11-9
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
possible) and through to system memory. When writing through to memory,
invalid cache lines are never filled, and valid cache lines are either filled or inval-
idated. Write combining is allowed. This type of cache-control is appropriate for
frame buffers or when there are devices on the system bus that access system
memory, but do not perform snooping of memory accesses. It enforces
coherency between caches in the processors and system memory.
• Write-back (WB) — Writes and reads to and from system memory are cached.
Reads come from cache lines on cache hits; read misses cause cache fills.
Speculative reads are allowed. Write misses cause cache line fills (in processor
families starting with the P6 family processors), and writes are performed
entirely in the cache, when possible. Write combining is allowed. The write-back
memory type reduces bus traffic by eliminating many unnecessary writes to
system memory. Writes to a cache line are not immediately forwarded to system
memory; instead, they are accumulated in the cache. The modified cache lines
are written to system memory later, when a write-back operation is performed.
Write-back operations are triggered when cache lines need to be deallocated,
such as when new cache lines are being allocated in a cache that is already full.
They also are triggered by the mechanisms used to maintain cache consistency.
This type of cache-control provides the best performance, but it requires that all
devices that access system memory on the system bus be able to snoop memory
misses cause cache fills. Writes are propagated to the system bus and cause
corresponding cache lines on all processors on the bus to be invalidated.
Speculative reads are allowed. This memory type is available in processor
families starting from the P6 family processors by programming the MTRRs (see
Table 11-6).
Table 11-3 shows which of these caching methods are available in the Pentium, P6
Family, Pentium 4, and Intel Xeon processors.
Table 11-3. Methods of Caching Available in Intel Core 2 Duo, Intel Atom, Intel Core
Duo, Pentium M, Pentium 4, Intel Xeon, P6 Family, and Pentium Processors
Memory Type
Intel Core 2 Duo, Intel Atom, Intel
Core Duo, Pentium M, Pentium 4
and Intel Xeon Processors
P6 Family
Processors
Pentium
Processor
Strong Uncacheable (UC) Yes
Yes
Yes*
Yes
Yes
Yes
Yes
Yes
No
Uncacheable (UC-)
Write Combining (WC)
Write Through (WT)
Write Back (WB)
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
No
Write Protected (WP)
11-10 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
Table 11-3. Methods of Caching Available in Intel Core 2 Duo, Intel Atom, Intel Core
Duo, Pentium M, Pentium 4, Intel Xeon, P6 Family, and Pentium Processors (Contd.)
Memory Type
Intel Core 2 Duo, Intel Atom, Intel
Core Duo, Pentium M, Pentium 4
and Intel Xeon Processors
P6 Family
Processors
Pentium
Processor
NOTE:
* Introduced in the Pentium III processor; not available in the Pentium Pro or Pentium II processors
11.3.1
Buffering of Write Combining Memory Locations
Writes to the WC memory type are not cached in the typical sense of the word
cached. They are retained in an internal write combining buffer (WC buffer) that is
separate from the internal L1, L2, and L3 caches and the store buffer. The WC buffer
is not snooped and thus does not provide data coherency. Buffering of writes to WC
memory is done to allow software a small window of time to supply more modified
data to the WC buffer while remaining as non-intrusive to software as possible. The
buffering of writes to WC memory also causes data to be collapsed; that is, multiple
writes to the same memory location will leave the last data written in the location and
the other writes will be lost.
The size and structure of the WC buffer is not architecturally defined. For the Intel
Core 2 Duo, Intel Atom, Intel Core Duo, Pentium M, Pentium 4 and Intel Xeon proces-
sors; the WC buffer is made up of several 64-byte WC buffers. For the P6 family
processors, the WC buffer is made up of several 32-byte WC buffers.
When software begins writing to WC memory, the processor begins filling the WC
buffers one at a time. When one or more WC buffers has been filled, the processor
has the option of evicting the buffers to system memory. The protocol for evicting the
WC buffers is implementation dependent and should not be relied on by software for
system memory coherency. When using the WC memory type, software must be
sensitive to the fact that the writing of data to system memory is being delayed and
must deliberately empty the WC buffers when system memory coherency is
required.
Once the processor has started to evict data from the WC buffer into system
memory, it will make a bus-transaction style decision based on how much of the
buffer contains valid data. If the buffer is full (for example, all bytes are valid), the
processor will execute a burst-write transaction on the bus. This results in all 32
bytes (P6 family processors) or 64 bytes (Pentium 4 and more recent processor)
being transmitted on the data bus in a single burst transaction. If one or more of the
WC buffer’s bytes are invalid (for example, have not been written by software), the
processor will transmit the data to memory using “partial write” transactions (one
chunk at a time, where a “chunk” is 8 bytes).
This will result in a maximum of 4 partial write transactions (for P6 family processors)
or 8 partial write transactions (for the Pentium 4 and more recent processors) for one
WC buffer of data sent to memory.
Vol. 3 11-11
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
The WC memory type is weakly ordered by definition. Once the eviction of a WC
buffer has started, the data is subject to the weak ordering semantics of its defini-
tion. Ordering is not maintained between the successive allocation/deallocation of
WC buffers (for example, writes to WC buffer 1 followed by writes to WC buffer 2 may
appear as buffer 2 followed by buffer 1 on the system bus). When a WC buffer is
evicted to memory as partial writes there is no guaranteed ordering between succes-
sive partial writes (for example, a partial write for chunk 2 may appear on the bus
before the partial write for chunk 1 or vice versa).
The only elements of WC propagation to the system bus that are guaranteed are
those provided by transaction atomicity. For example, with a P6 family processor, a
completely full WC buffer will always be propagated as a single 32-bit burst transac-
tion using any chunk order. In a WC buffer eviction where data will be evicted as
partials, all data contained in the same chunk (0 mod 8 aligned) will be propagated
simultaneously. Likewise, for more recent processors starting with those based on
Intel NetBurst microarchitectures, a full WC buffer will always be propagated as a
single burst transactions, using any chunk order within a transaction. For partial
buffer propagations, all data contained in the same chunk will be propagated simul-
taneously.
11.3.2
Choosing a Memory Type
The simplest system memory model does not use memory-mapped I/O with read or
write side effects, does not include a frame buffer, and uses the write-back memory
type for all memory. An I/O agent can perform direct memory access (DMA) to write-
back memory and the cache protocol maintains cache coherency.
A system can use strong uncacheable memory for other memory-mapped I/O, and
should always use strong uncacheable memory for memory-mapped I/O with read
side effects.
Dual-ported memory can be considered a write side effect, making relatively prompt
writes desirable, because those writes cannot be observed at the other port until they
reach the memory agent. A system can use strong uncacheable, uncacheable, write-
through, or write-combining memory for frame buffers or dual-ported memory that
contains pixel values displayed on a screen. Frame buffer memory is typically large (a
few megabytes) and is usually written more than it is read by the processor. Using
strong uncacheable memory for a frame buffer generates very large amounts of bus
traffic, because operations on the entire buffer are implemented using partial writes
rather than line writes. Using write-through memory for a frame buffer can displace
almost all other useful cached lines in the processor's L2 and L3 caches and L1 data
cache. Therefore, systems should use write-combining memory for frame buffers
whenever possible.
Software can use page-level cache control, to assign appropriate effective memory
types when software will not access data structures in ways that benefit from write-
back caching. For example, software may read a large data structure once and not
access the structure again until the structure is rewritten by another agent. Such a
11-12 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
large data structure should be marked as uncacheable, or reading it will evict cached
lines that the processor will be referencing again.
A similar example would be a write-only data structure that is written to (to export
the data to another agent), but never read by software. Such a structure can be
marked as uncacheable, because software never reads the values that it writes
(though as uncacheable memory, it will be written using partial writes, while as
write-back memory, it will be written using line writes, which may not occur until the
other agent reads the structure and triggers implicit write-backs).
provided that give software greater control over the caching, prefetching, and the
write-back characteristics of data. These instructions allow software to use weakly
ordered or processor ordered memory types to improve processor performance, but
when necessary to force strong ordering on memory reads and/or writes. They also
allow software greater control over the caching of data. For a description of these
instructions and there intended use, see Section 11.5.5, “Cache Management
Instructions.”
11.3.3
Code Fetches in Uncacheable Memory
Programs may execute code from uncacheable (UC) memory, but the implications
are different from accessing data in UC memory. When doing code fetches, the
processor never transitions from cacheable code to UC code speculatively. It also
never speculatively fetches branch targets that result in UC code.
The processor may fetch the same UC cache line multiple times in order to decode an
instruction once. It may decode consecutive UC instructions in a cacheline without
fetching between each instruction. It may also fetch additional cachelines from the
same or a consecutive 4-KByte page in order to decode one non-speculative UC
instruction (this can be true even when the instruction is contained fully in one line).
Because of the above and because cacheline sizes may change in future processors,
software should avoid placing memory-mapped I/O with read side effects in the
same page or in a subsequent page used to execute UC code.
11.4
CACHE CONTROL PROTOCOL
Intel 64 and IA-32 architectures.
In the L1 data cache and in the L2/L3 unified caches, the MESI (modified, exclusive,
shared, invalid) cache protocol maintains consistency with caches of other proces-
sors. The L1 data cache and the L2/L3 unified caches have two MESI status flags per
cache line. Each line can be marked as being in one of the states defined in Table
11-4. In general, the operation of the MESI protocol is transparent to programs.
Vol. 3 11-13
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
Table 11-4. MESI Cache Line States
Cache Line State
M (Modified)
E (Exclusive)
S (Shared)
Yes
I (Invalid)
No
This cache line is valid? Yes
Yes
Valid
No
The memory copy is…
Out of date
Valid
—
Copies exist in caches No
of other processors?
Maybe
Maybe
A write to this line …
Does not go to Does not go to Causes the
the system bus. the system bus. processor to gain
Goes directly to
the system bus.
exclusive
ownership of the
The L1 instruction cache in P6 family processors implements only the “SI” part of the
MESI protocol, because the instruction cache is not writable. The instruction cache
monitors changes in the data cache to maintain consistency between the caches
when instructions are modified. See Section 11.6, “Self-Modifying Code,” for more
information on the implications of caching instructions.
11.5
CACHE CONTROL
The Intel 64 and IA-32 architectures provide a variety of mechanisms for controlling
the caching of data and instructions and for controlling the ordering of reads and
writes between the processor, the caches, and memory. These mechanisms can be
divided into two groups:
• Cache control registers and bits — The Intel 64 and IA-32 architectures
define several dedicated registers and various bits within control registers and
page- and directory-table entries that control the caching system memory
locations in the L1, L2, and L3 caches. These mechanisms control the caching of
virtual memory pages and of regions of physical memory.
• Cache control and memory ordering instructions — The Intel 64 and IA-32
architectures provide several instructions that control the caching of data, the
ordering of memory reads and writes, and the prefetching of data. These instruc-
tions allow software to control the caching of specific data structures, to control
memory coherency for specific locations in memory, and to force strong memory
ordering at specific locations in a program.
The following sections describe these two groups of cache control mechanisms.
11-14 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
11.5.1
Cache Control Registers and Bits
the matter of memory address space, these work the same in Intel 64 processors.
The Intel 64 and IA-32 architectures provide the following cache-control registers
and bits for use in enabling or restricting caching to various pages or regions in
memory:
• CD flag, bit 30 of control register CR0 — Controls caching of system memory
locations (see Section 2.5, “Control Registers”). If the CD flag is clear, caching is
enabled for the whole of system memory, but may be restricted for individual
pages or regions of memory by other cache-control mechanisms. When the CD
flag is set, caching is restricted in the processor’s caches (cache hierarchy) for
the P6 and more recent processor families and prevented for the Pentium
processor (see note below). With the CD flag set, however, the caches will still
respond to snoop traffic. Caches should be explicitly flushed to insure memory
control register CR0 should be cleared. Table 11-5 shows the interaction of the
CD and NW flags.
The effect of setting the CD flag is somewhat different for processor families
memory coherency after the CD flag is set, the caches should be explicitly
flushed (see Section 11.5.3, “Preventing Caching”). Setting the CD flag for the
P6 and more recent processor families modify cache line fill and update
behaviour. Also, setting the CD flag on these processors do not force strict
ordering of memory accesses unless the MTRRs are disabled and/or all memory
is referenced as uncached (see Section 8.2.5, “Strengthening or Weakening the
Memory-Ordering Model”).
Vol. 3 11-15
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
CR4
P
G
E
Enables global pages
designated with G flag
Physical Memory
CR3
CR0
FFFFFFFFH2
PAT4
P P
C W
D T
Control caching of
page directory
PAT controls caching
of virtual memory
pages
Page-Directory or
Page-Table Entry
4
P P
C W
D T
P
A
T
C N
D W
1
G
CD and NW Flags
control overall caching
of system memory
PCD and PWT flags
control page-level
caching
MTRRs3
MTRRs control caching
of selected regions of
physical memory
G flag controls page-
level flushing of TLBs
0
Store Buffer
TLBs
1. G flag only available in P6 and later processor families
2. The maximum physical address size is reported by CPUID leaf
function 80000008H. The maximum physical address size of
FFFFFFFFFH applies only If 36-bit physical addressing is used.
3. MTRRs available only in P6 and later processor families;
similar control available in Pentium processor with the KEN#
and WB/WT# pins.
4. PAT available only in Pentium III and later processor families.
5. L3 in processors based on Intel NetBurst microarchitecture can
be disabled using IA32_MISC_ENABLES MSR.
Figure 11-3. Cache-Control Registers and Bits Available in Intel 64 and IA-32
Processors
11-16 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
Table 11-5. Cache Operating Modes
Caching and Read/Write Policy
1
CD NW
L1
L2/L3
0
0
Normal Cache Mode. Highest performance cache operation.
Yes
Yes
Yes
Yes
Yes
Yes
•
•
•
Read hits access the cache; read misses may cause replacement.
Write hits update the cache.
Only writes to shared lines and write misses update system
memory.
•
•
Write misses cause cache line fills.
Yes
Yes
Yes
Write hits can change shared lines to modified under control of
the MTRRs and with associated read invalidation cycle.
(Pentium processor only.) Write misses do not cause cache line
fills.
•
•
Yes
Yes
(Pentium processor only.) Write hits can change shared lines to
exclusive under control of WB/WT#.
Invalidation is allowed.
•
•
Yes
Yes
Yes
Yes
External snoop traffic is supported.
0
1
1
0
Invalid setting.
NA
NA
Generates a general-protection exception (#GP) with an error code
of 0.
3
No-fill Cache Mode. Memory coherency is maintained.
Yes
Yes
Yes
Yes
•
(Pentium 4 and later processor families.) State of processor after
a power up or reset.
Read hits access the cache; read misses do not cause
replacement (see Pentium 4 and Intel Xeon processors reference
below).
Write hits update the cache.
Only writes to shared lines and write misses update system
memory.
•
•
•
Yes
Yes
Yes
Yes
•
•
Write misses access memory.
Yes
Yes
Yes
Yes
Write hits can change shared lines to exclusive under control of
the MTRRs and with associated read invalidation cycle.
(Pentium processor only.) Write hits can change shared lines to
exclusive under control of the WB/WT#.
•
•
Yes
Yes
1
0
(P6 and later processor families only.) Strict memory ordering is
not enforced unless the MTRRs are disabled and/or all memory is
referenced as uncached (see Section 7.2.4., “Strengthening or
Weakening the Memory Ordering Model”).
Invalidation is allowed.
External snoop traffic is supported.
Yes
•
•
Yes
Yes
Yes
Yes
Vol. 3 11-17
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
Table 11-5. Cache Operating Modes
1
CD NW
Caching and Read/Write Policy
L1
L2/L3
2, 3
1
1
Memory coherency is not maintained.
Yes
Yes
Yes
Yes
Yes
Yes
•
•
•
(P6 family and Pentium processors.) State of the processor after
a power up or reset.
Read hits access the cache; read misses do not cause
replacement.
Write hits update the cache and change exclusive lines to
modified.
•
•
•
Shared lines remain shared after write hit.
Write misses access memory.
Yes
Yes
Yes
Yes
Yes
Yes
Invalidation is inhibited when snooping; but is allowed with INVD
and WBINVD instructions.
•
External snoop traffic is supported.
No
Yes
NOTES:
1. The L2/L3 column in this table is definitive for the Pentium 4, Intel Xeon, and P6 family proces-
sors. It is intended to represent what could be implemented in a system based on a Pentium pro-
cessor with an external, platform specific, write-back L2 cache.
2. The Pentium 4 and more recent processor families do not support this mode; setting the CD and
NW bits to 1 selects the no-fill cache mode.
3. Not supported In Intel Atom processors. If CD = 1 in an Intel Atom processor, caching is disabled.
• NW flag, bit 29 of control register CR0 — Controls the write policy for system
memory locations (see Section 2.5, “Control Registers”). If the NW and CD flags
are clear, write-back is enabled for the whole of system memory, but may be
restricted for individual pages or regions of memory by other cache-control
mechanisms. Table 11-5 shows how the other combinations of CD and NW flags
affects caching.
NOTES
For the Pentium 4 and Intel Xeon processors, the NW flag is a don’t
care flag; that is, when the CD flag is set, the processor uses the no-
fill cache mode, regardless of the setting of the NW flag.
For Intel Atom processors, the NW flag is a don’t care flag; that is,
when the CD flag is set, the processor disables caching, regardless of
the setting of the NW flag.
For the Pentium processor, when the L1 cache is disabled (the CD and
NW flags in control register CR0 are set), external snoops are
accepted in DP (dual-processor) systems and inhibited in unipro-
cessor systems.
When snoops are inhibited, address parity is not checked and
APCHK# is not asserted for a corrupt address; however, when snoops
are accepted, address parity is checked and APCHK# is asserted for
11-18 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
corrupt addresses.
for individual page tables and pages, respectively (see Section 4.9, “Paging and
Memory Typing”). This flag only has effect when paging is enabled and the CD
flag in control register CR0 is clear. The PCD flag enables caching of the page
table or page when clear and prevents caching when set.
policy for individual page tables and pages, respectively (see Section 4.9, “Paging
and Memory Typing”). This flag only has effect when paging is enabled and the
NW flag in control register CR0 is clear. The PWT flag enables write-back caching
of the page table or page when clear and write-through caching when set.
• PCD and PWT flags in control register CR3 — Control the global caching and
write policy for the page directory (see Section 2.5, “Control Registers”). The PCD
flag enables caching of the page directory when clear and prevents caching when
set. The PWT flag enables write-back caching of the page directory when clear
write policy for individual page tables. These flags only have effect when paging
is enabled and the CD flag in control register CR0 is clear.
to the IA-32 architecture in the P6 family processors) — Controls the
flushing of TLB entries for individual pages. See Section 4.10, “Caching
Information,” for more information about this flag.
• Memory type range registers (MTRRs) (introduced in P6 family
Available,” can be selected. See Section 11.11, “Memory Type Range Registers
(MTRRs),” for a detailed description of the MTRRs.
• Page Attribute Table (PAT) MSR (introduced in the Pentium III processor)
— Extends the memory typing capabilities of the processor to permit memory
types to be assigned on a page-by-page basis (see Section 11.12, “Page Attribute
Table (PAT)”).
• Third-Level Cache Disable flag, bit 6 of the IA32_MISC_ENABLES MSR
(Available only in processors based on Intel NetBurst microarchitecture)
— Allows the L3 cache to be disabled and enabled, independently of the L1 and
L2 caches.
• KEN# and WB/WT# pins (Pentium processor) — Allow external hardware to
control the caching method used for specific areas of memory. They perform
similar (but not identical) functions to the MTRRs in the P6 family processors.
• PCD and PWT pins (Pentium processor) — These pins (which are associated
with the PCD and PWT flags in control register CR3 and in the page-directory and
Vol. 3 11-19
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
page-table entries) permit caching in an external L2 cache to be controlled on a
page-by-page basis, consistent with the control exercised on the L1 cache of
these processors. The P6 and more recent processor families do not provide
these pins because the L2 cache in internal to the chip package.
11.5.2
Precedence of Cache Controls
The cache control flags and MTRRs operate hierarchically for restricting caching. That
is, if the CD flag is set, caching is prevented globally (see Table 11-5). If the CD flag
is clear, the page-level cache control flags and/or the MTRRs can be used to restrict
caching. If there is an overlap of page-level and MTRR caching controls, the mecha-
nism that prevents caching has precedence. For example, if an MTRR makes a region
of system memory uncacheable, a page-level caching control cannot be used to
enable caching for a page in that region. The converse is also true; that is, if a page-
level caching control designates a page as uncacheable, an MTRR cannot be used to
make the page cacheable.
In cases where there is a overlap in the assignment of the write-back and write-
through caching policies to a page and a region of memory, the write-through policy
takes precedence. The write-combining policy (which can only be assigned through
an MTRR or the PAT) takes precedence over either write-through or write-back.
The selection of memory types at the page level varies depending on whether PAT is
being used to select memory types for pages, as described in the following sections.
On processors based on Intel NetBurst microarchitecture, the third-level cache can
be disabled by bit 6 of the IA32_MISC_ENABLE MSR. Using IA32_MISC_ENALBES[bit
6] takes precedence over the CD flag, MTRRs, and PAT for the L3 cache in those
processors. That is, when the third-level cache disable flag is set (cache disabled),
the other cache controls have no affect on the L3 cache; when the flag is clear
(enabled), the cache controls have the same affect on the L3 cache as they have on
the L1 and L2 caches.
IA32_MISC_ENALBES[bit 6] is not supported in Intel Core i7 processors, nor proces-
sors based on Intel Core, and Intel Atom microarchitectures.
11.5.2.1 Selecting Memory Types for Pentium Pro and Pentium II
Processors
The Pentium Pro and Pentium II processors do not support the PAT. Here, the effec-
tive memory type for a page is selected with the MTRRs and the PCD and PWT bits in
the page-table or page-directory entry for the page. Table 11-6 describes the
mapping of MTRR memory types and page-level caching attributes to effective
memory types, when normal caching is in effect (the CD and NW flags in control
register CR0 are clear). Combinations that appear in gray are implementation-
defined for the Pentium Pro and Pentium II processors. System designers are encour-
aged to avoid these implementation-defined combinations.
11-20 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
Table 11-6. Effective Page-Level Memory Type for Pentium Pro and
Pentium II Processors
PCD Value PWT Value
1
MTRR Memory Type
Effective Memory Type
UC
X
0
0
1
1
0
1
0
0
1
1
0
0
1
X
0
1
0
1
X
X
0
1
0
1
0
1
X
UC
WC
WC
WC
UC
WC
WT
WP
WT
UC
WP
WP
WC
UC
WB
WB
WT
UC
NOTE:
1. These effective memory types also apply to the Pentium 4, Intel Xeon, and Pentium III proces-
sors when the PAT bit is not used (set to 0) in page-table and page-directory entries.
When normal caching is in effect, the effective memory type shown in Table 11-6 is
determined using the following rules:
1. If the PCD and PWT attributes for the page are both 0, then the effective
memory type is identical to the MTRR-defined memory type.
2. If the PCD flag is set, then the effective memory type is UC.
3. If the PCD flag is clear and the PWT flag is set, the effective memory type is WT
for the WB memory type and the MTRR-defined memory type for all other
memory types.
4. Setting the PCD and PWT flags to opposite values is considered model-specific for
the WP and WC memory types and architecturally-defined for the WB, WT, and
UC memory types.
Vol. 3 11-21
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
The Intel Core 2 Duo, Intel Atom, Intel Core Duo, Intel Core Solo, Pentium M,
Pentium 4, Intel Xeon, and Pentium III processors use the PAT to select effective
page-level memory types. Here, a memory type for a page is selected by the MTRRs
and the value in a PAT entry that is selected with the PAT, PCD and PWT bits in a
page-table or page-directory entry (see Section 11.12.3, “Selecting a Memory Type
from the PAT”). Table 11-7 describes the mapping of MTRR memory types and PAT
entry types to effective memory types, when normal caching is in effect (the CD and
NW flags in control register CR0 are clear).
Table 11-7. Effective Page-Level Memory Types for Pentium III and More Recent
Processor Families
MTRR Memory Type
PAT Entry Value
Effective Memory Type
1
UC
UC
UC-
WC
WT
WB
WP
UC
UC
1
UC
WC
1
UC
1
UC
1
UC
2
WC
WT
UC
UC-
WC
WT
WB
WP
UC
WC
WC
2,3
UC
WC
2,3
UC
2
UC
2
UC-
WC
WT
WB
WP
UC
WC
WT
WT
3
WP
11-22 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
Table 11-7. Effective Page-Level Memory Types for Pentium III and More Recent
Processor Families (Contd.)
MTRR Memory Type
PAT Entry Value
Effective Memory Type
2
WB
UC
UC-
WC
WT
WB
WP
UC
UC
2
UC
WC
WT
WB
WP
2
WP
UC
3
UC-
WC
WT
WB
WP
WC
WC
3
WT
WP
WP
NOTES:
1. The UC attribute comes from the MTRRs and the processors are not required to snoop their
caches since the data could never have been cached. This attribute is preferred for performance
reasons.
2. The UC attribute came from the page-table or page-directory entry and processors are required
to check their caches because the data may be cached due to page aliasing, which is not recom-
mended.
3. These combinations were specified as “undefined” in previous editions of the Intel® 64 and IA-32
Architectures Software Developer’s Manual. However, all processors that support both the PAT
and the MTRRs determine the effective page-level memory types for these combinations as
given.
11.5.2.3 Writing Values Across Pages with Different Memory Types
If two adjoining pages in memory have different memory types, and a word or longer
operand is written to a memory location that crosses the page boundary between
those two pages, the operand might be written to memory twice. This action does not
present a problem for writes to actual memory; however, if a device is mapped the
memory space assigned to the pages, the device might malfunction.
Vol. 3 11-23
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
11.5.3
Preventing Caching
To disable the L1, L2, and L3 caches after they have been enabled and have received
cache fills, perform the following steps:
1. Enter the no-fill cache mode. (Set the CD flag in control register CR0 to 1 and
the NW flag to 0.
2. Flush all caches using the WBINVD instruction.
3. Disable the MTRRs and set the default memory type to uncached or set all MTRRs
for the uncached memory type (see the discussion of the discussion of the TYPE
field and the E flag in Section 11.11.2.1, “IA32_MTRR_DEF_TYPE MSR”).
The caches must be flushed (step 2) after the CD flag is set to insure system memory
coherency. If the caches are not flushed, cache hits on reads will still occur and data
will be read from valid cache lines.
The intent of the three separate steps listed above address three distinct require-
ments: (i) discontinue new data replacing existing data in the cache (ii) ensure data
already in the cache are evicted to memory, (iii) ensure subsequent memory refer-
ences observe UC memory type semantics. Different processor implementation of
caching control hardware may allow some variation of software implementation of
these three requirements. See note below.
NOTES
Setting the CD flag in control register CR0 modifies the processor’s
caching behaviour as indicated in Table 11-5, but setting the CD flag
alone may not be sufficient across all processor families to force the
effective memory type for all physical memory to be UC nor does it
force strict memory ordering, due to hardware implementation
variations across different processor families. To force the UC
memory type and strict memory ordering on all of physical memory,
it is sufficient to either program the MTRRs for all physical memory to
be UC memory type or disable all MTRRs.
For the Pentium 4 and Intel Xeon processors, after the sequence of
steps given above has been executed, the cache lines containing the
code between the end of the WBINVD instruction and before the
MTRRS have actually been disabled may be retained in the cache
hierarchy. Here, to remove code from the cache completely, a second
WBINVD instruction must be executed after the MTRRs have been
disabled.
For Intel Atom processors, setting the CD flag forces all physical
memory to observe UC semantics (without requiring memory type of
physical memory to be set explicitly). Consequently, software does
not need to issue a second WBINVD as some other processor
generations might require.
11-24 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
11.5.4
On processors based on Intel NetBurst microarchitecture, the third-level cache can
be disabled by bit 6 of the IA32_MISC_ENABLE MSR. The third-level cache disable
flag (bit 6 of the IA32_MISC_ENABLE MSR) allows the L3 cache to be disabled and
enabled, independently of the L1 and L2 caches. Prior to using this control to disable
or enable the L3 cache, software should disable and flush all the processor caches, as
described earlier in Section 11.5.3, “Preventing Caching,” to prevent of loss of infor-
mation stored in the L3 cache. After the L3 cache has been disabled or enabled,
caching for the whole processor can be restored.
Newer Intel 64 processor with L3 do not support IA32_MISC_ENABLES[bit 6], the
procedure described in Section 11.5.3, “Preventing Caching,” apply to the entire
cache hierarchy.
11.5.5
Cache Management Instructions
The Intel 64 and IA-32 architectures provide several instructions for managing the
L1, L2, and L3 caches. The INVD, WBINVD, and WBINVD instructions are system
instructions that operate on the L1, L2, and L3 caches as a whole. The PREFETCHh
and CLFLUSH instructions and the non-temporal move instructions (MOVNTI,
MOVNTQ, MOVNTDQ, MOVNTPS, and MOVNTPD), which were introduced in
SSE/SSE2 extensions, offer more granular control over caching.
The INVD and WBINVD instructions are used to invalidate the contents of the L1, L2,
and L3 caches. The INVD instruction invalidates all internal cache entries, then
generates a special-function bus cycle that indicates that external caches also should
be invalidated. The INVD instruction should be used with care. It does not force a
write-back of modified cache lines; therefore, data stored in the caches and not
written back to system memory will be lost. Unless there is a specific requirement or
benefit to invalidating the caches without writing back the modified lines (such as,
during testing or fault recovery where cache coherency with main memory is not a
concern), software should use the WBINVD instruction.
The WBINVD instruction first writes back any modified lines in all the internal caches,
then invalidates the contents of both the L1, L2, and L3 caches. It ensures that cache
coherency with main memory is maintained regardless of the write policy in effect
(that is, write-through or write-back). Following this operation, the WBINVD instruc-
tion generates one (P6 family processors) or two (Pentium and Intel486 processors)
special-function bus cycles to indicate to external cache controllers that write-back of
time or cycles for WBINVD to complete will vary due to the size of different cache
hierarchies and other factors. As a consequence, the use of the WBINVD instruction
can have an impact on interrupt/event response time.
The PREFETCHh instructions allow a program to suggest to the processor that a
cache line from a specified location in system memory be prefetched into the cache
hierarchy (see Section 11.8, “Explicit Caching”).
Vol. 3 11-25
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
The CLFLUSH instruction allow selected cache lines to be flushed from memory. This
instruction give a program the ability to explicitly free up cache space, when it is
known that cached section of system memory will not be accessed in the near future.
The non-temporal move instructions (MOVNTI, MOVNTQ, MOVNTDQ, MOVNTPS, and
MOVNTPD) allow data to be moved from the processor’s registers directly into
system memory without being also written into the L1, L2, and/or L3 caches. These
instructions can be used to prevent cache pollution when operating on data that is
going to be modified only once before being stored back into system memory. These
instructions operate on data in the general-purpose, MMX, and XMM registers.
11.5.6
L1 Data Cache Context Mode
L1 data cache context mode is a feature of processors based on the Intel NetBurst
microarchitecture that support Intel Hyper-Threading Technology. When
CPUID.1:ECX[bit 10] = 1, the processor supports setting L1 data cache context
mode using the L1 data cache context mode flag ( IA32_MISC_ENABLE[bit 24] ).
Selectable modes are adaptive mode (default) and shared mode.
The BIOS is responsible for configuring the L1 data cache context mode.
11.5.6.1 Adaptive Mode
Adaptive mode facilitates L1 data cache sharing between logical processors. When
running in adaptive mode, the L1 data cache is shared across logical processors in
the same core if:
• CR3 control registers for logical processors sharing the cache are identical.
• The same paging mode is used by logical processors sharing the cache.
In this situation, the entire L1 data cache is available to each logical processor
(instead of being competitively shared).
If CR3 values are different for the logical processors sharing an L1 data cache or the
logical processors use different paging modes, processors compete for cache
resources. This reduces the effective size of the cache for each logical processor.
Aliasing of the cache is not allowed (which prevents data thrashing).
11.5.6.2 Shared Mode
In shared mode, the L1 data cache is competitively shared between logical proces-
sors. This is true even if the logical processors use identical CR3 registers and paging
modes.
In shared mode, linear addresses in the L1 data cache can be aliased, meaning that
one linear address in the cache can point to different physical locations. The mecha-
nism for resolving aliasing can lead to thrashing. For this reason,
IA32_MISC_ENABLE[bit 24] = 0 is the preferred configuration for processors based
11-26 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
on the Intel NetBurst microarchitecture that support Intel Hyper-Threading Tech-
nology.
11.6
SELF-MODIFYING CODE
A write to a memory location in a code segment that is currently cached in the
processor causes the associated cache line (or lines) to be invalidated. This check is
based on the physical address of the instruction. In addition, the P6 family and
Pentium processors check whether a write to a code segment may modify an instruc-
tion that has been prefetched for execution. If the write affects a prefetched instruc-
tion, the prefetch queue is invalidated. This latter check is based on the linear
address of the instruction. For the Pentium 4 and Intel Xeon processors, a write or a
snoop of an instruction in a code segment, where the target instruction is already
decoded and resident in the trace cache, invalidates the entire trace cache. The latter
behavior means that programs that self-modify code can cause severe degradation
of performance when run on the Pentium 4 and Intel Xeon processors.
among IA-32 processors. Applications that include self-modifying code use the same
linear address for modifying and fetching the instruction. Systems software, such as
a debugger, that might possibly modify an instruction using a different linear address
than that used to fetch the instruction, will execute a serializing operation, such as a
CPUID instruction, before the modified instruction is executed, which will automati-
cally resynchronize the instruction cache and prefetch queue. (See Section 8.1.3,
“Handling Self- and Cross-Modifying Code,” for more information about the use of
self-modifying code.)
For Intel486 processors, a write to an instruction in the cache will modify it in both
the cache and memory, but if the instruction was prefetched before the write, the old
version of the instruction could be the one executed. To prevent the old instruction
from being executed, flush the instruction prefetch unit by coding a jump instruction
immediately after any write that modifies an instruction.
11.7
IMPLICIT CACHING (PENTIUM 4, INTEL XEON,
AND P6 FAMILY PROCESSORS)
Implicit caching occurs when a memory element is made potentially cacheable,
although the element may never have been accessed in the normal von Neumann
sequence. Implicit caching occurs on the P6 and more recent processor families due
to aggressive prefetching, branch prediction, and TLB miss handling. Implicit caching
is an extension of the behavior of existing Intel386, Intel486, and Pentium processor
systems, since software running on these processor families also has not been able
to deterministically predict the behavior of instruction prefetch.
Vol. 3 11-27
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
To avoid problems related to implicit caching, the operating system must explicitly
invalidate the cache when changes are made to cacheable data that the cache coher-
ency mechanism does not automatically handle. This includes writes to dual-ported
or physically aliased memory boards that are not detected by the snooping mecha-
nisms of the processor, and changes to page- table entries in memory.
The code in Example 11-1 shows the effect of implicit caching on page-table entries.
The linear address F000H points to physical location B000H (the page-table entry for
F000H contains the value B000H), and the page-table entry for linear address F000
is PTE_F000.
Example 11-1. Effect of Implicit Caching on Page-Table Entries
mov EAX, CR3; Invalidate the TLB
mov CR3, EAX; by copying CR3 to itself
mov PTE_F000, A000H; Change F000H to point to A000H
mov EBX, [F000H];
Because of speculative execution in the P6 and more recent processor families, the
last MOV instruction performed would place the value at physical location B000H into
EBX, rather than the value at the new physical address A000H. This situation is
remedied by placing a TLB invalidation between the load and the store.
11.8
EXPLICIT CACHING
The Pentium III processor introduced four new instructions, the PREFETCHh instruc-
tions, that provide software with explicit control over the caching of data. These
instructions provide “hints” to the processor that the data requested by a PREFETCHh
instruction should be read into cache hierarchy now or as soon as possible, in antici-
pation of its use. The instructions provide different variations of the hint that allow
selection of the cache level into which data will be read.
The PREFETCHh instructions can help reduce the long latency typically associated
with reading data from memory and thus help prevent processor “stalls.” However,
these instructions should be used judiciously. Overuse can lead to resource conflicts
and hence reduce the performance of an application. Also, these instructions should
only be used to prefetch data from memory; they should not be used to prefetch
instructions. For more detailed information on the proper use of the prefetch instruc-
tion, refer to Chapter 7, “Optimizing Cache Usage,” in the Intel® 64 and IA-32 Archi-
tectures Optimization Reference Manual.
11-28 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
11.9
The processor updates its address translation caches (TLBs) transparently to soft-
ware. Several mechanisms are available, however, that allow software and hardware
to invalidate the TLBs either explicitly or as a side effect of another operation. Most
details are given in Section 4.10.3, “Invalidation of TLBs and Paging-Structure
Caches.” In addition, the following operations invalidate all TLB entries, irrespective
of the setting of the G flag:
• Asserting or de-asserting the FLUSH# pin.
WRMSR instruction).
• Writing to control register CR0 to modify the PG or PE flag.
• (Pentium 4, Intel Xeon, and later processors only.) Writing to control register CR4
to modify the PSE, PGE, or PAE flag.
See Section 4.10, “Caching Translation Information,” for additional information about
the TLBs.
11.10 STORE BUFFER
Intel 64 and IA-32 processors temporarily store each write (store) to memory in a
store buffer. The store buffer improves processor performance by allowing the
processor to continue executing instructions without having to wait until a write to
memory and/or to a cache is complete. It also allows writes to be delayed for more
efficient use of memory-access bus cycles.
In general, the existence of the store buffer is transparent to software, even in
systems that use multiple processors. The processor ensures that write operations
are always carried out in program order. It also insures that the contents of the store
buffer are always drained to memory in the following situations:
• When an exception or interrupt is generated.
• (P6 and more recent processor families only) When a serializing instruction is
executed.
• When an I/O instruction is executed.
• When a LOCK operation is performed.
• (P6 and more recent processor families only) When a BINIT operation is
performed.
• (Pentium III, and more recent processor families only) When using an SFENCE
instruction to order stores.
• (Pentium 4 and more recent processor families only) When using an MFENCE
instruction to order stores.
Vol. 3 11-29
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
The discussion of write ordering in Section 8.2, “Memory Ordering,” gives a detailed
description of the operation of the store buffer.
11.11 MEMORY TYPE RANGE REGISTERS (MTRRS)
The following section pertains only to the P6 and more recent processor families.
The memory type range registers (MTRRs) provide a mechanism for associating the
memory types (see Section 11.3, “Methods of Caching Available”) with physical-
address ranges in system memory. They allow the processor to optimize operations
nating the memory control pins used for this function on earlier IA-32 processors and
the external logic needed to drive them.
The MTRR mechanism allows up to 96 memory ranges to be defined in physical
memory, and it defines a set of model-specific registers (MSRs) for specifying the
type of memory that is contained in each range. Table 11-8 shows the memory types
that can be specified and their properties; Figure 11-4 shows the mapping of physical
memory with MTRRs. See Section 11.3, “Methods of Caching Available,” for a more
detailed description of each memory type.
Following a hardware reset, the P6 and more recent processor families disable all the
fixed and variable MTRRs, which in effect makes all of physical memory uncacheable.
Initialization software should then set the MTRRs to a specific, system-defined
memory map. Typically, the BIOS (basic input/output system) software configures
the MTRRs. The operating system or executive is then free to modify the memory
map using the normal page-level cacheability attributes.
In a multiprocessor system using a processor in the P6 family or a more recent
family, each processor MUST use the identical MTRR memory map so that software
will have a consistent view of memory.
NOTE
In multiple processor systems, the operating system must maintain
MTRR consistency between all the processors in the system (that is,
all processors must use the same MTRR values). The P6 and more
recent processor families provide no hardware support for
maintaining this consistency.
Table 11-8. Memory Types That Can Be Encoded in MTRRs
Memory Type and Mnemonic
Uncacheable (UC)
Encoding in MTRR
00H
01H
02H
Write Combining (WC)
Reserved*
11-30 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
Table 11-8. Memory Types That Can Be Encoded in MTRRs (Contd.)
Reserved*
03H
04H
Write-through (WT)
Write-protected (WP)
Writeback (WB)
Reserved*
05H
06H
7H through FFH
NOTE:
*
Use of these encodings results in a general-protection exception (#GP).
Physical Memory
FFFFFFFFH
Address ranges not
mapped by an MTRR
are set to a default type
Variable ranges
(from 4 KBytes to
maximum size of
physical memory)
100000H
FFFFFH
64 fixed ranges
(4 KBytes each)
16 fixed ranges
(16 KBytes each)
256 KBytes
256 KBytes
C0000H
BFFFFH
80000H
7FFFFH
8 fixed ranges
(64-KBytes each)
512 KBytes
0
Figure 11-4. Mapping Physical Memory With MTRRs
Vol. 3 11-31
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
11.11.1 MTRR Feature Identification
The availability of the MTRR feature is model-specific. Software can determine if
MTRRs are supported on a processor by executing the CPUID instruction and reading
the state of the MTRR flag (bit 12) in the feature information register (EDX).
If the MTRR flag is set (indicating that the processor implements MTRRs), additional
information about MTRRs can be obtained from the 64-bit IA32_MTRRCAP MSR
(named MTRRcap MSR for the P6 family processors). The IA32_MTRRCAP MSR is a
read-only MSR that can be read with the RDMSR instruction. Figure 11-5 shows the
contents of the IA32_MTRRCAP MSR. The functions of the flags and field in this
register are as follows:
• VCNT (variable range registers count) field, bits 0 through 7 — Indicates
the number of variable ranges implemented on the processor.
• FIX (fixed range registers supported) flag, bit 8 — Fixed range MTRRs
(IA32_MTRR_FIX64K_00000 through IA32_MTRR_FIX4K_0F8000) are
supported when set; no fixed range registers are supported when clear.
• WC (write combining) flag, bit 10 — The write-combining (WC) memory type
is supported when set; the WC type is not supported when clear.
• SMRR (System-Management Range Register) flag, bit 11 — The system-
management range register (SMRR) interface is supported when bit 11 is set; the
SMRR interface is not supported when clear.
Bit 9 and bits 11 through 63 in the IA32_MTRRCAP MSR are reserved. If software
attempts to write to the IA32_MTRRCAP MSR, a general-protection exception (#GP)
is generated.
Software must read IA32_MTRRCAP VCNT field to determine the number of variable
MTRRs and query other feature bits in IA32_MTRRCAP to determine additional capa-
bilities that are supported in a processor. For example, some processors may report
a value of ‘8’ in the VCNT field, other processors may report VCNT with different
values.
63
1110 9 8 7
0
F
W
I
Reserved
VCNT
C
X
SMRR — SMRR interface supported
WC — Write-combining memory type supported
FIX — Fixed range registers supported
VCNT — Number of variable range registers
Reserved
Figure 11-5. IA32_MTRRCAP Register
11-32 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
11.11.2 Setting Memory Ranges with MTRRs
The memory ranges and the types of memory specified in each range are set by three
groups of registers: the IA32_MTRR_DEF_TYPE MSR, the fixed-range MTRRs, and
the variable range MTRRs. These registers can be read and written to using the
RDMSR and WRMSR instructions, respectively. The IA32_MTRRCAP MSR indicates
the availability of these registers on the processor (see Section 11.11.1, “MTRR
Feature Identification”).
11.11.2.1 IA32_MTRR_DEF_TYPE MSR
processors) sets the default properties of the regions of physical memory that are not
encompassed by MTRRs. The functions of the flags and field in this register are as
follows:
• Type field, bits 0 through 7 — Indicates the default memory type used for
those physical memory address ranges that do not have a memory type specified
for them by an MTRR (see Table 11-8 for the encoding of this field). The legal
values for this field are 0, 1, 4, 5, and 6. All other values result in a general-
protection exception (#GP) being generated.
Intel recommends the use of the UC (uncached) memory type for all physical
memory addresses where memory does not exist. To assign the UC type to
nonexistent memory locations, it can either be specified as the default type in the
Type field or be explicitly assigned with the fixed and variable MTRRs.
63
12 1110 9 8 7
0
F
E
Reserved
Type
E
E — MTRR enable/disable
FE — Fixed-range MTRRs enable/disable
Type — Default memory type
Reserved
Figure 11-6. IA32_MTRR_DEF_TYPE MSR
• FE (fixed MTRRs enabled) flag, bit 10 — Fixed-range MTRRs are enabled
when set; fixed-range MTRRs are disabled when clear. When the fixed-range
MTRRs are enabled, they take priority over the variable-range MTRRs when
overlaps in ranges occur. If the fixed-range MTRRs are disabled, the variable-
range MTRRs can still be used and can map the range ordinarily covered by the
fixed-range MTRRs.
• E (MTRRs enabled) flag, bit 11 — MTRRs are enabled when set; all MTRRs are
disabled when clear, and the UC memory type is applied to all of physical
Vol. 3 11-33
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
memory. When this flag is set, the FE flag can disable the fixed-range MTRRs;
when the flag is clear, the FE flag has no affect. When the E flag is set, the type
specified in the default memory type field is used for areas of memory not
already mapped by either a fixed or variable MTRR.
Bits 8 and 9, and bits 12 through 63, in the IA32_MTRR_DEF_TYPE MSR are
reserved; the processor generates a general-protection exception (#GP) if software
attempts to write nonzero values to them.
11.11.2.2 Fixed Range MTRRs
The fixed memory ranges are mapped with 11 fixed-range registers of 64 bits each.
Each of these registers is divided into 8-bit fields that are used to specify the memory
type for each of the sub-ranges the register controls:
• Register IA32_MTRR_FIX64K_00000 — Maps the 512-KByte address range
from 0H to 7FFFFH. This range is divided into eight 64-KByte sub-ranges.
• Registers IA32_MTRR_FIX16K_80000 and IA32_MTRR_FIX16K_A0000
• Registers IA32_MTRR_FIX4K_C0000 through
IA32_MTRR_FIX4K_F8000 — Maps eight 32-KByte address ranges from
C0000H to FFFFFH. This range is divided into sixty-four 4-KByte sub-ranges, 8
ranges per register.
Table 11-9 shows the relationship between the fixed physical-address ranges and the
corresponding fields of the fixed-range MTRRs; Table 11-8 shows memory type
encoding for MTRRs.
For the P6 family processors, the prefix for the fixed range MTRRs is MTRRfix.
11.11.2.3 Variable Range MTRRs
The Pentium 4, Intel Xeon, and P6 family processors permit software to specify the
memory type for eight variable-size address ranges, using a pair of MTRRs for each
range. The first entry in each pair (IA32_MTRR_PHYSBASEn) defines the base
address and memory type for the range; the second entry
(IA32_MTRR_PHYSMASKn) contains a mask used to determine the address range.
The “n” suffix indicates register pairs 0 through 7.
For P6 family processors, the prefixes for these variable range MTRRs are MTRRphys-
Base and MTRRphysMask.
11-34 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
Table 11-9. Address Mapping for Fixed-Range MTRRs
Address Range (hexadecimal)
MTRR
63 56
55 48
47 40
39 32
31 24 23 16 15
8
7
0
70000-
7FFFF
60000-
6FFFF
50000-
5FFFF
40000-
4FFFF
30000-
3FFFF
20000-
2FFFF
10000-
1FFFF
00000-
0FFFF
IA32_MTRR_
FIX64K_00000
9C000
9FFFF
98000-
98FFF
94000-
97FFF
90000-
93FFF
8C000-
8FFFF
88000-
8BFFF
84000-
87FFF
80000-
83FFF
IA32_MTRR_
FIX16K_80000
BC000
BFFFF
B8000-
BBFFF
B4000-
B7FFF
B0000-
B3FFF
AC000-
AFFFF
A8000-
ABFFF
A4000-
A7FFF
A0000-
A3FFF
IA32_MTRR_
FIX16K_A0000
C7000
C7FFF
C6000-
C6FFF
C5000-
C5FFF
C4000-
C4FFF
C3000-
C3FFF
C2000-
C2FFF
C1000-
C1FFF
C0000-
C0FFF
IA32_MTRR_
FIX4K_C0000
CF000
CFFFF
CE000-
CEFFF
CD000-
CDFFF
CC000-
CCFFF
CB000-
CBFFF
CA000-
CAFFF
C9000-
C9FFF
C8000-
C8FFF
IA32_MTRR_
FIX4K_C8000
D7000
D7FFF
D6000-
D6FFF
D5000-
D5FFF
D4000-
D4FFF
D3000-
D3FFF
D2000-
D2FFF
D1000-
D1FFF
D0000-
D0FFF
IA32_MTRR_
FIX4K_D0000
DF000
DFFFF
DE000-
DEFFF
DD000-
DDFFF
DC000-
DCFFF
DB000-
DBFFF
DA000-
DAFFF
D9000-
D9FFF
D8000-
D8FFF
IA32_MTRR_
FIX4K_D8000
E7000
E7FFF
E6000-
E6FFF
E5000-
E5FFF
E4000-
E4FFF
E3000-
E3FFF
E2000-
E2FFF
E1000-
E1FFF
E0000-
E0FFF
IA32_MTRR_
FIX4K_E0000
EF000
EFFFF
EE000-
EEFFF
ED000-
EC000-
ECFFF
EB000-
EBFFF
EA000-
EAFFF
E9000-
E9FFF
E8000-
E8FFF
IA32_MTRR_
FIX4K_E8000
F7000
F7FFF
F6000-
F6FFF
F5000-
F5FFF
F4000-
F4FFF
F3000-
F3FFF
F2000-
F2FFF
F1000-
F1FFF
F0000-
F0FFF
IA32_MTRR_
FIX4K_F0000
FF000
FFFFF
FE000-
FEFFF
FB000-
FBFFF
FA000-
FAFFF
F9000-
F9FFF
F8000-
F8FFF
IA32_MTRR_
FIX4K_F8000
Figure 11-7 shows flags and fields in these registers. The functions of these flags and
fields are:
• Type field, bits 0 through 7 — Specifies the memory type for the range (see
Table 11-8 for the encoding of this field).
• PhysBase field, bits 12 through (MAXPHYADDR-1) — Specifies the base
address of the address range. This 24-bit value, in the case where MAXPHYADDR
is 36 bits, is extended by 12 bits at the low end to form the base address (this
automatically aligns the address on a 4-KByte boundary).
physical address size is 40 bits). The mask determines the range of the region
being mapped, according to the following relationships:
— Address_Within_Range AND PhysMask = PhysBase AND PhysMask
— This value is extended by 12 bits at the low end to form the mask value. For
more information: see Section 11.11.3, “Example Base and Mask Calcula-
tions.”
Vol. 3 11-35
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
— The width of the PhysMask field depends on the maximum physical address
size supported by the processor.
CPUID.80000008H reports the maximum physical address size supported by
the processor. If CPUID.80000008H is not available, software may assume
that the processor supports a 36-bit physical address size (then PhysMask is
24 bits wide and the upper 28 bits of IA32_MTRR_PHYSMASKn are reserved).
See the Note below.
• V (valid) flag, bit 11 — Enables the register pair when set; disables register
pair when clear.
IA32_MTRR_PHYSBASEn Register
63
MAXPHYADDR 1211
8 7
0
Reserved
PhysBase
Type
PhysBase — Base address of range
Type — Memory type for range
IA32_MTRR_PHYSMASKn Register
63
MAXPHYADDR
121110
0
V
Reserved
PhysMask
Reserved
PhysMask — Sets range mask
V — Valid
Reserved
MAXPHYADDR: The bit position indicated by MAXPHYADDR depends on the maximum
physical address range supported by the processor. It is reported by CPUID leaf
function 80000008H. If CPUID does not support leaf 80000008H, the processor
supports 36-bit physical address size, then bit PhysMask consists of bits 35:12, and
bits 63:36 are reserved.
Figure 11-7. IA32_MTRR_PHYSBASEn and IA32_MTRR_PHYSMASKn Variable-Range
Register Pair
All other bits in the IA32_MTRR_PHYSBASEn and IA32_MTRR_PHYSMASKn registers
are reserved; the processor generates a general-protection exception (#GP) if soft-
ware attempts to write to them.
Some mask values can result in ranges that are not continuous. In such ranges, the
area not mapped by the mask value is set to the default memory type. Intel does not
encourage the use of “discontinuous” ranges because they could require physical
memory to be present throughout the entire 4-GByte physical memory map. If
memory is not provided, the behaviour is undefined.
11-36 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
NOTE
BIOS provides by using the ACPI/INT15 e820 interface mechanism.
This information then can be used to determine how MTRRs are
initialized (for example: allowing the BIOS to define valid memory
ranges and the maximum memory range supported by the platform,
including the processor).
See Section 11.11.4.1, “MTRR Precedences,” for information on overlapping variable
MTRR ranges.
11.11.2.4 System-Management Range Register Interface
If IA32_MTRRCAP[bit 11] is set, the processor supports the SMRR interface to
restrict access to a specified memory address range used by system-management
mode (SMM) software (see Section 26.4.2.1). If the SMRR interface is supported,
SMM software is strongly encouraged to use it to protect the SMI code and data
stored by SMI handler in the SMRAM region.
The system-management range registers consist of a pair of MSRs (see Figure 11-8).
The IA32_SMRR_PHYSBASE MSR defines the base address for the SMRAM memory
range and the memory type used to access it in SMM. The IA32_SMRR_PHYSMASK
MSR contains a valid bit and a mask that determines the SMRAM address range
protected by the SMRR interface. These MSRs may be written only in SMM; an
1
attempt to write them outside of SMM causes a general-protection exception.
Figure 11-8 shows flags and fields in these registers. The functions of these flags and
fields are the following:
• Type field, bits 0 through 7 — Specifies the memory type for the range (see
Table 11-8 for the encoding of this field).
• PhysBase field, bits 12 through 31 — Specifies the base address of the
address range. The address must be less than 4 GBytes and is automatically
aligned on a 4-KByte boundary.
• PhysMask field, bits 12 through 31 — Specifies a mask that determines the
range of the region being mapped, according to the following relationships:
— Address_Within_Range AND PhysMask = PhysBase AND PhysMask
— This value is extended by 12 bits at the low end to form the mask value. For
more information: see Section 11.11.3, “Example Base and Mask Calcula-
tions.”
• V (valid) flag, bit 11 — Enables the register pair when set; disables register
pair when clear.
1. For some processor models, these MSRs can be accessed by RDMSR and WRMSR only if the
SMRR interface has been enabled in the IA32_FEATURE_CONTROL MSR. See Appendix B.
Vol. 3 11-37
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
Before attempting to access these SMRR registers, software must test bit 11 in the
IA32_MTRRCAP register. If SMRR is not supported, reads from or writes to registers
cause general-protection exceptions.
When the valid flag in the IA32_SMRR_PHYSMASK MSR is 1, accesses to the specified
address range are treated as follows:
• If the logical processor is in SMM, accesses uses the memory type in the
IA32_SMRR_PHYSBASE MSR.
• If the logical processor is not in SMM, write accesses are ignored and read
accesses return a fixed value for each byte. The uncacheable memory type (UC)
is used in this case.
The above items apply even if the address range specified overlaps with a range
specified by the MTRRs.
IA32_SMRR_PHYSBASE Register
63
31
1211
8 7
0
Reserved
PhysBase
Type
PhysBase — Base address of range
Type — Memory type for range
IA32_SMRR_PHYSMASK Register
63
31
121110
0
V
Reserved
PhysMask
Reserved
PhysMask — Sets range mask
V — Valid
Reserved
Figure 11-8. IA32_SMRR_PHYSBASE and IA32_SMRR_PHYSMASK SMRR Pair
11.11.3 Example Base and Mask Calculations
The examples in this section apply to processors that support a maximum physical
address size of 36 bits. The base and mask values entered in variable-range MTRR
pairs are 24-bit values that the processor extends to 36-bits.
For example, to enter a base address of 2 MBytes (200000H) in the
IA32_MTRR_PHYSBASE3 register, the 12 least-significant bits are truncated and the
value 000200H is entered in the PhysBase field. The same operation must be
performed on mask values. For example, to map the address range from 200000H to
11-38 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
3FFFFFH (2 MBytes to 4 MBytes), a mask value of FFFE00000H is required. Again, the
12 least-significant bits of this mask value are truncated, so that the value entered in
the PhysMask field of IA32_MTRR_PHYSMASK3 is FFFE00H. This mask is chosen so
that when any address in the 200000H to 3FFFFFH range is AND’d with the mask
value, it will return the same value as when the base address is AND’d with the mask
value (which is 200000H).
To map the address range from 400000H to 7FFFFFH (4 MBytes to 8 MBytes), a base
value of 000400H is entered in the PhysBase field and a mask value of FFFC00H is
entered in the PhysMask field.
Example 11-2. Setting-Up Memory for a System
Here is an example of setting up the MTRRs for an system. Assume that the system
has the following characteristics:
• 96 MBytes of system memory is mapped as write-back memory (WB) for highest
system performance.
• A custom 4-MByte I/O card is mapped to uncached memory (UC) at a base
address of 64 MBytes. This restriction forces the 96 MBytes of system memory to
be addressed from 0 to 64 MBytes and from 68 MBytes to 100 MBytes, leaving a
4-MByte hole for the I/O card.
• An 8-MByte graphics card is mapped to write-combining memory (WC) beginning
at address A0000000H.
• The BIOS area from 15 MBytes to 16 MBytes is mapped to UC memory.
The following settings for the MTRRs will yield the proper mapping of the physical
address space for this system configuration.
IA32_MTRR_PHYSBASE0 = 0000 0000 0000 0006H
IA32_MTRR_PHYSMASK0 = 0000 000F FC00 0800H
Caches 0-64 MByte as WB cache type.
IA32_MTRR_PHYSBASE1 = 0000 0000 0400 0006H
IA32_MTRR_PHYSMASK1 = 0000 000F FE00 0800H
Caches 64-96 MByte as WB cache type.
IA32_MTRR_PHYSBASE2 = 0000 0000 0600 0006H
IA32_MTRR_PHYSMASK2 = 0000 000F FFC0 0800H
Caches 96-100 MByte as WB cache type.
IA32_MTRR_PHYSBASE3 = 0000 0000 0400 0000H
IA32_MTRR_PHYSMASK3 = 0000 000F FFC0 0800H
Caches 64-68 MByte as UC cache type.
IA32_MTRR_PHYSBASE4 = 0000 0000 00F0 0000H
IA32_MTRR_PHYSMASK4 = 0000 000F FFF0 0800H
Caches 15-16 MByte as UC cache type.
Vol. 3 11-39
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
IA32_MTRR_PHYSBASE5 = 0000 0000 A000 0001H
IA32_MTRR_PHYSMASK5 = 0000 000F FF80 0800H
Caches A0000000-A0800000 as WC type.
This MTRR setup uses the ability to overlap any two memory ranges (as long as the
ranges are mapped to WB and UC memory types) to minimize the number of MTRR
registers that are required to configure the memory environment. This setup also
fulfills the requirement that two register pairs are left for operating system usage.
11.11.3.1 Base and Mask Calculations for Greater-Than 36-bit Physical
Address Support
For Intel 64 and IA-32 processors that support greater than 36 bits of physical
physical address. See the example.
Example 11-3. Setting-Up Memory for a System with a 40-Bit Address Size
If a processor supports 40-bits of physical address size, then the PhysMask field (in
IA32_MTRR_PHYSMASKn registers) is 28 bits instead of 24 bits. For this situation,
Example 11-2 should be modified as follows:
IA32_MTRR_PHYSBASE0 = 0000 0000 0000 0006H
IA32_MTRR_PHYSMASK0 = 0000 00FF FC00 0800H
Caches 0-64 MByte as WB cache type.
IA32_MTRR_PHYSBASE1 = 0000 0000 0400 0006H
IA32_MTRR_PHYSMASK1 = 0000 00FF FE00 0800H
Caches 64-96 MByte as WB cache type.
IA32_MTRR_PHYSBASE2 = 0000 0000 0600 0006H
IA32_MTRR_PHYSMASK2 = 0000 00FF FFC0 0800H
Caches 96-100 MByte as WB cache type.
IA32_MTRR_PHYSBASE3 = 0000 0000 0400 0000H
IA32_MTRR_PHYSMASK3 = 0000 00FF FFC0 0800H
Caches 64-68 MByte as UC cache type.
IA32_MTRR_PHYSBASE4 = 0000 0000 00F0 0000H
IA32_MTRR_PHYSMASK4 = 0000 00FF FFF0 0800H
Caches 15-16 MByte as UC cache type.
IA32_MTRR_PHYSBASE5 = 0000 0000 A000 0001H
IA32_MTRR_PHYSMASK5 = 0000 00FF FF80 0800H
Caches A0000000-A0800000 as WC type.
11-40 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
11.11.4 Range Size and Alignment Requirement
A range that is to be mapped to a variable-range MTRR must meet the following
“power of 2” size and alignment rules:
1. The minimum range size is 4 KBytes and the base address of the range must be
on at least a 4-KByte boundary.
n
2. For ranges greater than 4 KBytes, each range must be of length 2 and its base
n
address must be aligned on a 2 boundary, where n is a value equal to or greater
than 12. The base-address alignment value cannot be less than its length. For
example, an 8-KByte range cannot be aligned on a 4-KByte boundary. It must be
aligned on at least an 8-KByte boundary.
11.11.4.1 MTRR Precedences
If the MTRRs are not enabled (by setting the E flag in the IA32_MTRR_DEF_TYPE
MSR), then all memory accesses are of the UC memory type. If the MTRRs are
enabled, then the memory type used for a memory access is determined as follows:
1. If the physical address falls within the first 1 MByte of physical memory and
fixed MTRRs are enabled, the processor uses the memory type stored for the
appropriate fixed-range MTRR.
2. Otherwise, the processor attempts to match the physical address with a memory
type set by the variable-range MTRRs:
— If one variable memory range matches, the processor uses the memory type
stored in the IA32_MTRR_PHYSBASEn register for that range.
— If two or more variable memory ranges match and the memory types are
identical, then that memory type is used.
— If two or more variable memory ranges match and one of the memory types
is UC, the UC memory type used.
— If two or more variable memory ranges match and the memory types are WT
and WB, the WT memory type is used.
— For overlaps not defined by the above rules, processor behavior is undefined.
3. If no fixed or variable memory range matches, the processor uses the default
memory type.
11.11.5 MTRR Initialization
On a hardware reset, the P6 and more recent processors clear the valid flags in vari-
able-range MTRRs and clear the E flag in the IA32_MTRR_DEF_TYPE MSR to disable
all MTRRs. All other bits in the MTRRs are undefined.
Prior to initializing the MTRRs, software (normally the system BIOS) must initialize all
fixed-range and variable-range MTRR register fields to 0. Software can then initialize
Vol. 3 11-41
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
the MTRRs according to known types of memory, including memory on devices that it
auto-configures. Initialization is expected to occur prior to booting the operating
system.
See Section 11.11.8, “MTRR Considerations in MP Systems,” for information on
initializing MTRRs in MP (multiple-processor) systems.
11.11.6 Remapping Memory Types
A system designer may re-map memory types to tune performance or because a
future processor may not implement all memory types supported by the Pentium 4,
Intel Xeon, and P6 family processors. The following rules support coherent memory-
type re-mappings:
1. A memory type should not be mapped into another memory type that has a
weaker memory ordering model. For example, the uncacheable type cannot be
mapped into any other type, and the write-back, write-through, and write-
protected types cannot be mapped into the weakly ordered write-combining
type.
2. A memory type that does not delay writes should not be mapped into a memory
type that does delay writes, because applications of such a memory type may
rely on its write-through behavior. Accordingly, the write-back type cannot be
mapped into the write-through type.
3. A memory type that views write data as not necessarily stored and read back by
a subsequent read, such as the write-protected type, can only be mapped to
another type with the same behaviour (and there are no others for the
Pentium 4, Intel Xeon, and P6 family processors) or to the uncacheable type.
In many specific cases, a system designer can have additional information about how
a memory type is used, allowing additional mappings. For example, write-through
memory with no associated write side effects can be mapped into write-back
memory.
11.11.7 MTRR Maintenance Programming Interface
The operating system maintains the MTRRs after booting and sets up or changes the
memory types for memory-mapped devices. The operating system should provide a
driver and application programming interface (API) to access and set the MTRRs. The
function calls MemTypeGet() and MemTypeSet() define this interface.
11.11.7.1 MemTypeGet() Function
The MemTypeGet() function returns the memory type of the physical memory range
specified by the parameters base and size. The base address is the starting physical
address and the size is the number of bytes for the memory range. The function
11-42 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
automatically aligns the base address and size to 4-KByte boundaries. Pseudocode
for the MemTypeGet() function is given in Example 11-4.
Example 11-4. MemTypeGet() Pseudocode
#define MIXED_TYPES -1 /* 0 < MIXED_TYPES || MIXED_TYPES > 256 */
IF CPU_FEATURES.MTRR /* processor supports MTRRs */
THEN
Align BASE and SIZE to 4-KByte boundary;
IF (BASE + SIZE) wrap 4-GByte address space
THEN return INVALID;
FI;
IF MTRRdefType.E = 0
THEN return UC;
FI;
FirstType ¨ Get4KMemType (BASE);
/* Obtains memory type for first 4-KByte range. */
/* See Get4KMemType (4KByteRange) in Example 11-5. */
FOR each additional 4-KByte range specified in SIZE
NextType ¨ Get4KMemType (4KByteRange);
IF NextType ¼ FirstType
THEN return MixedTypes;
FI;
ROF;
return FirstType;
ELSE return UNSUPPORTED;
FI;
If the processor does not support MTRRs, the function returns UNSUPPORTED. If the
MTRRs are not enabled, then the UC memory type is returned. If more than one
memory type corresponds to the specified range, a status of MIXED_TYPES is
returned. Otherwise, the memory type defined for the range (UC, WC, WT, WB, or
WP) is returned.
The pseudocode for the Get4KMemType() function in Example 11-5 obtains the
memory type for a single 4-KByte range at a given physical address. The sample
code determines whether an PHY_ADDRESS falls within a fixed range by comparing
the address with the known fixed ranges: 0 to 7FFFFH (64-KByte regions), 80000H to
BFFFFH (16-KByte regions), and C0000H to FFFFFH (4-KByte regions). If an address
falls within one of these ranges, the appropriate bits within one of its MTRRs deter-
mine the memory type.
Vol. 3 11-43
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
Example 11-5. Get4KMemType() Pseudocode
IF IA32_MTRRCAP.FIX AND MTRRdefType.FE /* fixed registers enabled */
THEN IF PHY_ADDRESS is within a fixed range
return IA32_MTRR_FIX.Type;
FI;
FOR each variable-range MTRR in IA32_MTRRCAP.VCNT
IF IA32_MTRR_PHYSMASK.V = 0
THEN continue;
FI;
IF (PHY_ADDRESS AND IA32_MTRR_PHYSMASK.Mask) =
(IA32_MTRR_PHYSBASE.Base
AND IA32_MTRR_PHYSMASK.Mask)
THEN
return IA32_MTRR_PHYSBASE.Type;
FI;
ROF;
return MTRRdefType.Type;
11.11.7.2 MemTypeSet() Function
The MemTypeSet() function in Example 11-6 sets a MTRR for the physical memory
range specified by the parameters base and size to the type specified by type. The
base address and size are multiples of 4 KBytes and the size is not 0.
Example 11-6. MemTypeSet Pseudocode
IF CPU_FEATURES.MTRR (* processor supports MTRRs *)
THEN
IF BASE and SIZE are not 4-KByte aligned or size is 0
THEN return INVALID;
FI;
IF (BASE + SIZE) wrap 4-GByte address space
THEN return INVALID;
FI;
IF TYPE is invalid for Pentium 4, Intel Xeon, and P6 family
processors
THEN return UNSUPPORTED;
FI;
IF TYPE is WC and not supported
THEN return UNSUPPORTED;
FI;
IF IA32_MTRRCAP.FIX is set AND range can be mapped using a
fixed-range MTRR
11-44 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
THEN
pre_mtrr_change();
update affected MTRR;
post_mtrr_change();
FI;
ELSE (* try to map using a variable MTRR pair *)
IF IA32_MTRRCAP.VCNT = 0
THEN return UNSUPPORTED;
FI;
IF conflicts with current variable ranges
THEN return RANGE_OVERLAP;
FI;
IF no MTRRs available
THEN return VAR_NOT_AVAILABLE;
FI;
IF BASE and SIZE do not meet the power of 2 requirements for
variable MTRRs
THEN return INVALID_VAR_REQUEST;
FI;
pre_mtrr_change();
Update affected MTRRs;
post_mtrr_change();
FI;
pre_mtrr_change()
BEGIN
disable interrupts;
Save current value of CR4;
disable and flush caches;
flush TLBs;
disable MTRRs;
IF multiprocessing
THEN maintain consistency through IPIs;
FI;
END
post_mtrr_change()
BEGIN
flush caches and TLBs;
enable MTRRs;
enable caches;
restore value of CR4;
enable interrupts;
Vol. 3 11-45
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
END
The physical address to variable range mapping algorithm in the MemTypeSet func-
tion detects conflicts with current variable range registers by cycling through them
and determining whether the physical address in question matches any of the current
ranges. During this scan, the algorithm can detect whether any current variable
ranges overlap and can be concatenated into a single range.
The pre_mtrr_change() function disables interrupts prior to changing the MTRRs, to
avoid executing code with a partially valid MTRR setup. The algorithm disables
caching by setting the CD flag and clearing the NW flag in control register CR0. The
caches are invalidated using the WBINVD instruction. The algorithm flushes all TLB
entries either by clearing the page-global enable (PGE) flag in control register CR4 (if
PGE was already set) or by updating control register CR3 (if PGE was already clear).
Finally, it disables MTRRs by clearing the E flag in the IA32_MTRR_DEF_TYPE MSR.
After the memory type is updated, the post_mtrr_change() function re-enables the
MTRRs and again invalidates the caches and TLBs. This second invalidation is
required because of the processor's aggressive prefetch of both instructions and
data. The algorithm restores interrupts and re-enables caching by setting the CD
flag.
An operating system can batch multiple MTRR updates so that only a single pair of
cache invalidations occur.
11.11.8 MTRR Considerations in MP Systems
In MP (multiple-processor) systems, the operating systems must maintain MTRR
consistency between all the processors in the system. The Pentium 4, Intel Xeon, and
P6 family processors provide no hardware support to maintain this consistency. In
general, all processors must have the same MTRR values.
This requirement implies that when the operating system initializes an MP system, it
must load the MTRRs of the boot processor while the E flag in register MTRRdefType
is 0. The operating system then directs other processors to load their MTRRs with the
same memory map. After all the processors have loaded their MTRRs, the operating
system signals them to enable their MTRRs. Barrier synchronization is used to
prevent further memory accesses until all processors indicate that the MTRRs are
enabled. This synchronization is likely to be a shoot-down style algorithm, with
shared variables and interprocessor interrupts.
Any change to the value of the MTRRs in an MP system requires the operating system
to repeat the loading and enabling process to maintain consistency, using the
following procedure:
1. Broadcast to all processors to execute the following code sequence.
2. Disable interrupts.
3. Wait for all processors to reach this point.
11-46 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
4. Enter the no-fill cache mode. (Set the CD flag in control register CR0 to 1 and the
NW flag to 0.)
5. Flush all caches using the WBINVD instructions. Note on a processor that
supports self-snooping, CPUID feature flag bit 27, this step is unnecessary.
6. If the PGE flag is set in control register CR4, flush all TLBs by clearing that flag.
7. If the PGE flag is clear in control register CR4, flush all TLBs by executing a MOV
from control register CR3 to another register and then a MOV from that register
back to CR3.
8. Disable all range registers (by clearing the E flag in register MTRRdefType). If
only variable ranges are being modified, software may clear the valid bits for the
affected register pairs instead.
9. Update the MTRRs.
10. Enable all range registers (by setting the E flag in register MTRRdefType). If only
variable-range registers were modified and their individual valid bits were
cleared, then set the valid bits for the affected ranges instead.
11. Flush all caches and all TLBs a second time. (The TLB flush is required for
Pentium 4, Intel Xeon, and P6 family processors. Executing the WBINVD
instruction is not needed when using Pentium 4, Intel Xeon, and P6 family
processors, but it may be needed in future systems.)
12. Enter the normal cache mode to re-enable caching. (Set the CD and NW flags in
control register CR0 to 0.)
13. Set PGE flag in control register CR4, if cleared in Step 6 (above).
14. Wait for all processors to reach this point.
15. Enable interrupts.
11.11.9 Large Page Size Considerations
The MTRRs provide memory typing for a limited number of regions that have a
4 KByte granularity (the same granularity as 4-KByte pages). The memory type for a
given page is cached in the processor’s TLBs. When using large pages (2 or
4 MBytes), a single page-table entry covers multiple 4-KByte granules, each with a
single memory type. Because the memory type for a large page is cached in the TLB,
the processor can behave in an undefined manner if a large page is mapped to a
region of memory that MTRRs have mapped with multiple memory types.
Undefined behavior can be avoided by insuring that all MTRR memory-type ranges
within a large page are of the same type. If a large page maps to a region of memory
containing different MTRR-defined memory types, the PCD and PWT flags in the
page-table entry should be set for the most conservative memory type for that
range. For example, a large page used for memory mapped I/O and regular memory
is mapped as UC memory. Alternatively, the operating system can map the region
using multiple 4-KByte pages each with its own memory type.
Vol. 3 11-47
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
The requirement that all 4-KByte ranges in a large page are of the same memory
type implies that large pages with different memory types may suffer a performance
penalty, since they must be marked with the lowest common denominator memory
type.
The Pentium 4, Intel Xeon, and P6 family processors provide special support for the
physical memory range from 0 to 4 MBytes, which is potentially mapped by both the
fixed and variable MTRRs. This support is invoked when a Pentium 4, Intel Xeon, or
P6 family processor detects a large page overlapping the first 1 MByte of this
memory range with a memory type that conflicts with the fixed MTRRs. Here, the
processor maps the memory range as multiple 4-KByte pages within the TLB. This
operation insures correct behavior at the cost of performance. To avoid this perfor-
mance penalty, operating-system software should reserve the large page option for
regions of memory at addresses greater than or equal to 4 MBytes.
11.12 PAGE ATTRIBUTE TABLE (PAT)
The Page Attribute Table (PAT) extends the IA-32 architecture’s page-table format to
allow memory types to be assigned to regions of physical memory based on linear
address mappings. The PAT is a companion feature to the MTRRs; that is, the MTRRs
allow mapping of memory types to regions of the physical address space, where the
PAT allows mapping of memory types to pages within the linear address space. The
MTRRs are useful for statically describing memory types for physical ranges, and are
typically set up by the system BIOS. The PAT extends the functions of the PCD and
PWT bits in page tables to allow all five of the memory types that can be assigned
with the MTRRs (plus one additional memory type) to also be assigned dynamically
to pages of the linear address space.
The PAT was introduced to IA-32 architecture on the Pentium III processor. It is also
available in the Pentium 4 and Intel Xeon processors.
11.12.1 Detecting Support for the PAT Feature
An operating system or executive can detect the availability of the PAT by executing
the CPUID instruction with a value of 1 in the EAX register. Support for the PAT is indi-
cated by the PAT flag (bit 16 of the values returned to EDX register). If the PAT is
supported, the operating system or executive can use the IA32_PAT MSR to program
the PAT. When memory types have been assigned to entries in the PAT, software can
then use of the PAT-index bit (PAT) in the page-table and page-directory entries
along with the PCD and PWT bits to assign memory types from the PAT to individual
pages.
Note that there is no separate flag or control bit in any of the control registers that
enables the PAT. The PAT is always enabled on all processors that support it, and the
table lookup always occurs whenever paging is enabled, in all paging modes.
11-48 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
11.12.2 IA32_PAT MSR
The IA32_PAT MSR is located at MSR address 277H (see to Appendix B, “Model-
Specific Registers (MSRs),” and this address will remain at the same address on
of the 64-bit IA32_PAT MSR.
The IA32_PAT MSR contains eight page attribute fields: PA0 through PA7. The three
low-order bits of each field are used to specify a memory type. The five high-order
bits of each field are reserved, and must be set to all 0s. Each of the eight page
attribute fields can contain any of the memory type encodings specified in Table
11-10.
31
27
26
24
56
23
19
18
16
48
15
11
10
8
7
3
2
0
Reserved PA3
Reserved PA2
Reserved PA1
Reserved PA0
63
59
58
55
51
50
47
43
42
40
39
35
34
32
Reserved PA7
Reserved PA6
Reserved PA5
Reserved PA4
Figure 11-9. IA32_PAT MSR
Note that for the P6 family processors, the IA32_PAT MSR is named the PAT MSR.
Table 11-10. Memory Types That Can Be Encoded With PAT
Encoding
00H
Mnemonic
Uncacheable (UC)
Write Combining (WC)
Reserved*
01H
02H
03H
Reserved*
04H
Write Through (WT)
Write Protected (WP)
Write Back (WB)
Uncached (UC-)
Reserved*
05H
06H
07H
08H - FFH
NOTE:
* Using these encodings will result in a general-protection exception (#GP).
Vol. 3 11-49
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
11.12.3 Selecting a Memory Type from the PAT
To select a memory type for a page from the PAT, a 3-bit index made up of the PAT,
PCD, and PWT bits must be encoded in the page-table or page-directory entry for the
page. Table 11-11 shows the possible encodings of the PAT, PCD, and PWT bits and
that point to 4-KByte pages and bit 12 in paging-structure entries that point to larger
pages. The PCD and PWT bits are bits 4 and 3, respectively, in paging-structure
entries that point to pages of any size.
The PAT entry selected for a page is used in conjunction with the MTRR setting for the
region of physical memory in which the page is mapped to determine the effective
memory type for the page, as shown in Table 11-7.
Table 11-11. Selection of PAT Entries with PAT, PCD, and PWT Flags
PAT
PCD
PWT
PAT Entry
0
0
0
PAT0
0
0
1
PAT1
0
1
0
PAT2
0
1
1
PAT3
1
0
0
PAT4
1
0
1
PAT5
1
1
0
PAT6
1
1
1
PAT7
11.12.4 Programming the PAT
Table 11-12 shows the default setting for each PAT entry following a power up or
reset of the processor. The setting remain unchanged following a soft reset (INIT
reset).
Table 11-12. Memory Type Setting of PAT Entries Following a Power-up or Reset
PAT Entry
PAT0
Memory Type Following Power-up or Reset
WB
WT
UC-
UC
PAT1
PAT2
PAT3
PAT4
WB
WT
UC-
UC
PAT5
PAT6
PAT7
11-50 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
The values in all the entries of the PAT can be changed by writing to the IA32_PAT
MSR using the WRMSR instruction. The IA32_PAT MSR is read and write accessible
(use of the RDMSR and WRMSR instructions, respectively) to software operating at a
Attempting to write an undefined memory type encoding into the PAT causes a
general-protection (#GP) exception to be generated.
The operating system is responsible for insuring that changes to a PAT entry occur in
a manner that maintains the consistency of the processor caches and translation
lookaside buffers (TLB). This is accomplished by following the procedure as specified
in Section 11.11.8, “MTRR Considerations in MP Systems,” for changing the value of
an MTRR in a multiple processor system. It requires a specific sequence of operations
that includes flushing the processors caches and TLBs.
The PAT allows any memory type to be specified in the page tables, and therefore it
is possible to have a single physical page mapped to two or more different linear
addresses, each with different memory types. Intel does not support this practice
because it may lead to undefined operations that can result in a system failure. In
particular, a WC page must never be aliased to a cacheable page because WC writes
may not check the processor caches.
When remapping a page that was previously mapped as a cacheable memory type to
a WC page, an operating system can avoid this type of aliasing by doing the
following:
1. Remove the previous mapping to a cacheable memory type in the page tables;
that is, make them not present.
2. Flush the TLBs of processors that may have used the mapping, even specula-
tively.
3. Create a new mapping to the same physical address with a new memory type, for
instance, WC.
4. Flush the caches on all processors that may have used the mapping previously.
Note on processors that support self-snooping, CPUID feature flag bit 27, this
step is unnecessary.
Operating systems that use a page directory as a page table (to map large pages)
and enable page size extensions must carefully scrutinize the use of the PAT index bit
for the 4-KByte page-table entries. The PAT index bit for a page-table entry (bit 7)
corresponds to the page size bit in a page-directory entry. Therefore, the operating
a page table that is also used as a page directory. If the operating system attempts
to use PAT entries PA4 through PA7 when using this memory as a page table, it effec-
tively sets the PS bit for the access to this memory as a page directory.
For compatibility with earlier IA-32 processors that do not support the PAT, care
should be taken in selecting the encodings for entries in the PAT (see Section
11.12.5, “PAT Compatibility with Earlier IA-32 Processors”).
Vol. 3 11-51
Download from Www.Somanuals.com. All Manuals Search And Download.
MEMORY CACHE CONTROL
11.12.5 PAT Compatibility with Earlier IA-32 Processors
For IA-32 processors that support the PAT, the IA32_PAT MSR is always active. That
is, the PCD and PWT bits in page-table entries and in page-directory entries (that
point to pages) are always select a memory type for a page indirectly by selecting an
entry in the PAT. They never select the memory type for a page directly as they do in
earlier IA-32 processors that do not implement the PAT (see Table 11-6).
To allow compatibility for code written to run on earlier IA-32 processor that do not
bility to earlier processors. This compatibility is provided through the ordering of the
PAT, PCD, and PWT bits in the 3-bit PAT entry index. For processors that do not imple-
ment the PAT, the PAT index bit (bit 7 in the page-table entries and bit 12 in the page-
directory entries) is reserved and set to 0. With the PAT bit reserved, only the first
four entries of the PAT can be selected with the PCD and PWT bits. At power-up or
reset (see Table 11-12), these first four entries are encoded to select the same
memory types as the PCD and PWT bits would normally select directly in an IA-32
processor that does not implement the PAT. So, if encodings of the first four entries
in the PAT are left unchanged following a power-up or reset, code written to run on
earlier IA-32 processors that do not implement the PAT will run correctly on IA-32
processors that do implement the PAT.
11-52 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
CHAPTER 12
INTEL® MMX™ TECHNOLOGY SYSTEM
PROGRAMMING
®
This chapter describes those features of the Intel MMX™ technology that must be
considered when designing or enhancing an operating system to support MMX tech-
nology. It covers MMX instruction set emulation, the MMX state, aliasing of MMX
registers, saving MMX state, task and context switching considerations, exception
handling, and debugging.
12.1
The IA-32 or Intel 64 architecture does not support emulation of the MMX instruc-
tions, as it does for x87 FPU instructions. The EM flag in control register CR0
(provided to invoke emulation of x87 FPU instructions) cannot be used for MMX
instruction emulation. If an MMX instruction is executed when the EM flag is set, an
invalid opcode exception (UD#) is generated. Table 12-1 shows the interaction of the
EM, MP, and TS flags in control register CR0 when executing MMX instructions.
Table 12-1. Action Taken By MMX Instructions
for Different Combinations of EM, MP and TS
CR0 Flags
EM
0
MP*
1
TS
0
Action
Execute.
0
1
1
#NM exception.
#UD exception.
#UD exception.
1
1
0
1
1
1
NOTE:
* For processors that support the MMX instructions, the MP flag should be set.
12.2
THE MMX STATE AND MMX REGISTER ALIASING
The MMX state consists of eight 64-bit registers (MM0 through MM7). These registers
are aliased to the low 64-bits (bits 0 through 63) of floating-point registers R0
through R7 (see Figure 12-1). Note that the MMX registers are mapped to the phys-
ical locations of the floating-point registers (R0 through R7), not to the relative loca-
tions of the registers in the floating-point register stack (ST0 through ST7). As a
Vol. 3 12-1
Download from Www.Somanuals.com. All Manuals Search And Download.
®
™
INTEL MMX TECHNOLOGY SYSTEM PROGRAMMING
result, the MMX register mapping is fixed and is not affected by value in the Top Of
Stack (TOS) field in the floating-point status word (bits 11 through 13).
x87 FPU Tag
Register
Floating-Point Registers
0
79
64 63
00
R7
R6
R5
R4
R3
R2
R1
R0
00
00
00
00
00
00
00
x87 FPU Status Register
11
13
000
TOS
MMX Registers
0
63
MM7
MM6
MM5
MM4
MM3
MM2
MM1
MM0
TOS = 0
Figure 12-1. Mapping of MMX Registers to Floating-Point Registers
When a value is written into an MMX register using an MMX instruction, the value also
appears in the corresponding floating-point register in bits 0 through 63. Likewise,
when a floating-point value written into a floating-point register by a x87 FPU, the
low 64 bits of that value also appears in a the corresponding MMX register.
The execution of MMX instructions have several side effects on the x87 FPU state
status word. These side effects are as follows:
• When an MMX instruction writes a value into an MMX register, at the same time,
bits 64 through 79 of the corresponding floating-point register are set to all 1s.
• When an MMX instruction (other than the EMMS instruction) is executed, each of
the tag fields in the x87 FPU tag word is set to 00B (valid). (See also Section
12.2.1, “Effect of MMX, x87 FPU, FXSAVE, and FXRSTOR Instructions on the x87
FPU Tag Word.”)
12-2 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
®
™
INTEL MMX TECHNOLOGY SYSTEM PROGRAMMING
• When the EMMS instruction is executed, each tag field in the x87 FPU tag word is
set to 11B (empty).
• Each time an MMX instruction is executed, the TOS value is set to 000B.
Execution of MMX instructions does not affect the other bits in the x87 FPU status
word (bits 0 through 10 and bits 14 and 15) or the contents of the other x87 FPU
registers that comprise the x87 FPU state (the x87 FPU control word, instruction
pointer, data pointer, or opcode registers).
Table 12-2 summarizes the effects of the MMX instructions on the x87 FPU state.
Table 12-2. Effects of MMX Instructions on x87 FPU State
MMX
Instruction
Type
x87 FPU Tag TOS Field of Other x87
Bits 64
Bits 0
Word
x87 FPU
Status
Word
FPU Registers Through 79 of Through 63 of
x87 FPU Data x87 FPU Data
Registers
Registers
Read from
MMX register to 00B (Valid)
All tags set
000B
Unchanged
Unchanged
Unchanged
Unchanged
Unchanged
Write to MMX All tags set
register
EMMS
000B
Set to all 1s
Unchanged
Overwritten
with MMX data
to 00B (Valid)
All fields set 000B
Unchanged
(Empty)
12.2.1
Effect of MMX, x87 FPU, FXSAVE, and FXRSTOR
Instructions on the x87 FPU Tag Word
and FXRSTOR instructions on the tags in the x87 FPU tag word and the corresponding
tags in an image of the tag word stored in memory.
The values in the fields of the x87 FPU tag word do not affect the contents of the MMX
registers or the execution of MMX instructions. However, the MMX instructions do
modify the contents of the x87 FPU tag word, as is described in Section 12.2, “The
MMX State and MMX Register Aliasing.” These modifications may affect the operation
of the x87 FPU when executing x87 FPU instructions, if the x87 FPU state is not
initialized or restored prior to beginning x87 FPU instruction execution.
Note that the FSAVE, FXSAVE, and FSTENV instructions (which save x87 FPU state
information) read the x87 FPU tag register and contents of each of the floating-point
registers, determine the actual tag values for each register (empty, nonzero, zero, or
special), and store the updated tag word in memory. After executing these instruc-
tions, all the tags in the x87 FPU tag word are set to empty (11B). Likewise, the
EMMS instruction clears MMX state from the MMX/floating-point registers by setting
all the tags in the x87 FPU tag word to 11B.
Vol. 3 12-3
Download from Www.Somanuals.com. All Manuals Search And Download.
®
™
INTEL MMX TECHNOLOGY SYSTEM PROGRAMMING
Table 12-3. Effect of the MMX, x87 FPU, and FXSAVE/FXRSTOR Instructions on the
x87 FPU Tag Word
Instruction
Type
Instruction
x87 FPU Tag Word
Image of x87 FPU Tag Word
Stored in Memory
MMX
MMX
All (except EMMS) All tags are set to 00B (valid). Not affected.
EMMS
All tags are set to 11B
(empty).
Not affected.
x87 FPU
All (except FSAVE, Tag for modified floating-
FSTENV, FRSTOR, point register is set to 00B or
Not affected.
FLDENV)
x87 FPU and FSAVE, FSTENV,
FXSAVE FXSAVE
11B.
Tags and register values are
read and interpreted; then all actual values in the floating-
tags are set to 11B.
Tags are set according to the
point registers; that is, empty
registers are marked 11B and
valid registers are marked
00B (nonzero), 01B (zero), or
10B (special).
x87 FPU and FRSTOR, FLDENV, All tags marked 11B in
Tags are read and
interpreted, but not modified.
FXRSTOR
FXRSTOR
memory are set to 11B; all
other tags are set according
to the value in the
corresponding floating-point
register: 00B (nonzero), 01B
(zero), or 10B (special).
12.3
SAVING AND RESTORING THE MMX STATE AND
REGISTERS
Because the MMX registers are aliased to the x87 FPU data registers, the MMX state
can be saved to memory and restored from memory as follows:
• Execute an FSAVE, FNSAVE, or FXSAVE instruction to save the MMX state to
registers.)
• Execute an FRSTOR or FXRSTOR instruction to restore the MMX state from
memory. (The FXRSTOR instruction also restores the state of the XMM and
MXCSR registers.)
The save and restore methods described above are required for operating systems
(see Section 12.4, “Saving MMX State on Task or Context Switches”). Applications
can in some cases save and restore only the MMX registers in the following way:
12-4 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
®
™
INTEL MMX TECHNOLOGY SYSTEM PROGRAMMING
• Execute eight MOVQ instructions to save the contents of the MMX0 through
MMX7 registers to memory. An EMMS instruction may then (optionally) be
executed to clear the MMX state in the x87 FPU.
• Execute eight MOVQ instructions to read the saved contents of MMX registers
from memory into the MMX0 through MMX7 registers.
NOTE
The IA-32 architecture does not support scanning the x87 FPU tag
word and then only saving valid entries.
12.4
SAVING MMX STATE ON TASK OR CONTEXT
When switching from one task or context to another, it is often necessary to save the
MMX state. As a general rule, if the existing task switching code for an operating
system includes facilities for saving the state of the x87 FPU, these facilities can also
be relied upon to save the MMX state, without rewriting the task switch code. This
Section 12.2, “The MMX State and MMX Register Aliasing”).
With the introduction of the FXSAVE and FXRSTOR instructions and of
SSE/SSE2/SSE3/SSSE3 extensions, it is possible (and more efficient) to create state
saving facilities in the operating system or executive that save the x87
FPU/MMX/SSE/SSE2/SSE3/SSSE3 state in one operation. Section 13.5, “Designing
OS Facilities for AUTOMATICALLY Saving x87 FPU, MMX, and
SSE/SSE2/SSE3/SSSE3/SSE4 state on Task or Context Switches,” describes how to
design such facilities. The techniques describes in this section can be adapted to
saving only the MMX and x87 FPU state if needed.
12.5
EXCEPTIONS THAT CAN OCCUR WHEN EXECUTING
MMX INSTRUCTIONS
MMX instructions do not generate x87 FPU floating-point exceptions, nor do they
affect the processor’s status flags in the EFLAGS register or the x87 FPU status word.
The following exceptions can be generated during the execution of an MMX instruc-
tion:
• Exceptions during memory accesses:
— Stack-segment fault (#SS).
— General protection (#GP).
— Page fault (#PF).
— Alignment check (#AC), if alignment checking is enabled.
Vol. 3 12-5
Download from Www.Somanuals.com. All Manuals Search And Download.
®
™
MMX instruction is executed (see Section 12.1, “Emulation of the MMX
— Device not available (#NM), if an MMX instruction is executed when the TS
flag in control register CR0 is set. (See Section 13.5.1, “Using the TS Flag to
Control the Saving of the x87 FPU, MMX, SSE, SSE2, SSE3 SSSE3 and SSE4
State.”)
• Floating-point error (#MF). (See Section 12.5.1, “Effect of MMX Instructions on
Pending x87 Floating-Point Exceptions.”)
• Other exceptions can occur indirectly due to the faulty execution of the exception
handlers for the above exceptions.
12.5.1
Effect of MMX Instructions on Pending x87 Floating-Point
Exceptions
MMX instruction, the processor generates a x87 FPU floating-point error (#MF) prior
to executing the MMX instruction, to allow the pending exception to be handled by
the x87 FPU floating-point error exception handler. While this exception handler is
executing, the x87 FPU state is maintained and is visible to the handler. Upon
returning from the exception handler, the MMX instruction is executed, which will
alter the x87 FPU state, as described in Section 12.2, “The MMX State and MMX
Register Aliasing.”
12.6
DEBUGGING MMX CODE
The debug facilities operate in the same manner when executing MMX instructions as
when executing other IA-32 or Intel 64 architecture instructions.
To correctly interpret the contents of the MMX or x87 FPU registers from the
FSAVE/FNSAVE or FXSAVE image in memory, a debugger needs to take account of
the relationship between the x87 FPU register’s logical locations relative to TOS and
the MMX register’s physical locations.
In the x87 FPU context, STn refers to an x87 FPU register at location n relative to the
locations of the x87 FPU registers (R0 through R7). The MMX registers always refer
to the physical locations of the registers (with MM0 through MM7 being mapped to R0
through R7). Figure 12-2 shows this relationship. Here, the inner circle refers to the
physical location of the x87 FPU and MMX registers. The outer circle refers to the x87
FPU registers’s relative location to the current TOS.
When the TOS equals 0 (case A in Figure 12-2), ST0 points to the physical location
R0 on the floating-point stack. MM0 maps to ST0, MM1 maps to ST1, and so on.
12-6 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
®
™
INTEL MMX TECHNOLOGY SYSTEM PROGRAMMING
When the TOS equals 2 (case B in Figure 12-2), ST0 points to the physical location
R2. MM0 maps to ST6, MM1 maps to ST7, MM2 maps to ST0, and so on.
x87 FPU “push”
x87 FPU “push”
x87 FPU “pop”
ST1
ST6
ST0
ST7
MM1
ST7
MM0
(R0)
MM0
(R0)
MM7
MM6
MM5
MM7
MM1
MM2
TOS
MM2
(R2)
TOS
ST0
MM6
MM5
ST2
(R2)
MM3
MM3
x87 FPU “pop”
MM4
MM4
ST1
Case B: TOS=2
Case A: TOS=0
Outer circle = x87 FPU data register’s logical location relative to TOS
Inner circle = x87 FPU tags = MMX register’s location = FP registers’s physical location
Figure 12-2. Mapping of MMX Registers to x87 FPU Data Register Stack
Vol. 3 12-7
Download from Www.Somanuals.com. All Manuals Search And Download.
®
™
INTEL MMX TECHNOLOGY SYSTEM PROGRAMMING
12-8 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
CHAPTER 13
SYSTEM PROGRAMMING FOR INSTRUCTION SET
EXTENSIONS AND PROCESSOR EXTENDED STATES
operating on the processor state extension known as the SSE state (XMM registers,
MXCSR) and for processor extended states. Instruction set extensions operating on
the SSE state include the streaming SIMD extensions (SSE), streaming SIMD exten-
sions 2 (SSE2), streaming SIMD extensions 3 (SSE3), Supplemental SSE3 (SSSE3),
and SSE4.
Sections 13.1 through 13.5 cover system programming requirements to enable
SSE/SSE2/SSE3/SSSE3/SSE4 extensions, providing operating system or executive
support for the SSE/SSE2/SSE3/SSSE3/SSE4 extensions, SIMD floating-point
exceptions, exception handling, and task (context) switching.
Operating system support for SSE state, once implemented using FXSAVE/FXRSTOR,
provides a limited degree of forward support for subsequent instruction set exten-
sions operating on the same known set of processor state. Processor extended states
refer to an extension in Intel 64 architecture that will allow system executives to
implement support for multiple processor state extensions that may be introduced
over time without requiring the system executive to be modified each time a new
processor state extension is introduced.
Managing processor extended states requires the following aspects:
• using instructions like XSAVE, XRSTOR, to save/restore state information to a
memory region consistent with the processor state extensions supported in
hardware,
supported by the processor,
• using XSETBV instruction to enable individual processor state extensions,
• maintaining various system programming resources.
System programming for managing processor extended states is described in the
sections starting 13.6.
13.1
PROVIDING OPERATING SYSTEM SUPPORT FOR
SSE/SSE2/SSE3/SSSE3/SSE4 EXTENSIONS
To use SSE/SSE2/SSE3/SSSE3/SSE4 extensions, the operating system or executive
must provide support for initializing the processor to use these extensions, for
handling the FXSAVE and FXRSTOR state saving instructions, and for handling SIMD
floating-point exceptions. The following sections provide system programming
Vol. 3 13-1
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND PROCESSOR
guidelines for this support. Because SSE/SSE2/SSE3/SSSE3/SSE4 extensions share
the same state, experience the same sets of non-numerical and numerical exception
behavior, these guidelines that apply to SSE also apply to other sets of SIMD exten-
sions that operate on the same processor state and subject to the same sets of of
non-numerical and numerical exception behavior.
Chapter 11, “Programming with Streaming SIMD Extensions 2 (SSE2),” and Chapter
12, “Programming with SSE3, SSSE3 and SSE4,” in the Intel® 64 and IA-32 Archi-
tectures Software Developer’s Manual, Volume 1, discuss support for
SSE/SSE2/SSE3/SSSE3/SSE4 from an applications point of view program.
13.1.1
Adding Support to an Operating System for
SSE/SSE2/SSE3/SSSE3/SSE4 Extensions
The following guidelines describe functions that an operating system or executive
must perform to support SSE/SSE2/SSE3/SSSE3/SSE4 extensions:
1. Check that the processor supports the SSE/SSE2/SSE3/SSSE3/SSE4 extensions.
2. Check that the processor supports the FXSAVE and FXRSTOR instructions.
3. Provide an initialization for the SSE, SSE2 SSE3, SSSE3 and SSE4 states.
4. Provide support for the FXSAVE and FXRSTOR instructions.
5. Provide support (if necessary) in non-numeric exception handlers for exceptions
generated by the SSE, SSE2, SSE3 and SSE4 instructions.
6. Provide an exception handler for the SIMD floating-point exception (#XM).
The following sections describe how to implement each of these guidelines.
13.1.2
Checking for SSE/SSE2/SSE3/SSSE3/SSE4 Extension
Support
If the processor attempts to execute an unsupported SSE/SSE2/SSE3/SSSE3/SSE4
instruction, the processor generates an invalid-opcode exception (#UD).
Before an operating system or executive attempts to use
SSE/SSE2/SSE3/SSSE3/SSE4 extensions, it should check that support is present.
Make sure:
• CPUID.1:EDX.SSE[bit 25] = 1
• CPUID.1:EDX.SSE2[bit 26] = 1
• CPUID.1:ECX.SSE3[bit 0] = 1
• CPUID.1:ECX.SSSE3[bit 9] = 1
• CPUID.1:ECX.SSE4_1[bit 19] = 1
• CPUID.1:ECX.SSE4_2[bit 20] = 1
13-2 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND
To use POPCNT instruction, software must check CPUID.1:ECX.POPCNT[bit 23] = 1
13.1.3
Checking for Support for the FXSAVE and FXRSTOR
Instructions
A separate check must be made to insure that the processor supports FXSAVE and
FXRSTOR. Make sure:
• CPUID.1:EDX.FXSR[bit 24] = 1
13.1.4
Initialization of the SSE/SSE2/SSE3/SSSE3/SSE4 Extensions
The operating system or executive should carry out the following steps to set up
provides facilities for saving and restoring SSE/SSE2/SSE3/SSSE3/SSE4 states
using FXSAVE and FXRSTOR instructions. These instructions are commonly used
to save the SSE/SSE2/SSE3/SSSE3/SSE4 state during task switches and when
invoking the SIMD floating-point exception (#XM) handler (see Section 13.4,
“Saving the SSE/SSE2/SSE3/SSSE3/SSE4 State on Task or Context Switches,”
and Section 13.1.6, “Providing an Handler for the SIMD Floating-Point Exception
(#XM),” respectively).
If the processor does not support the FXSAVE and FXRSTOR instructions,
attempting to set the OSFXSR flag will cause an exception (#GP) to be
generated.
2. Set CR4.OSXMMEXCPT[bit 10] = 1. Setting this flag assumes that the operating
system provides an SIMD floating-point exception (#XM) handler (see Section
13.1.6, “Providing an Handler for the SIMD Floating-Point Exception (#XM)”).
NOTE
set by the operating system. The processor has no other way of
detecting operating-system support for the FXSAVE and FXRSTOR
required when executing SSE/SSE2/SSE3/SSSE3/SSE4 instructions (see Section
2.5, “Control Registers”).
4. Set CR0.MP[bit 1] = 1. This setting is the required setting for Intel 64 and IA-32
processors that support the SSE/SSE2/SSE3/SSSE3/SSE4 extensions (see
Section 9.2.1, “Configuring the x87 FPU Environment”).
Table 13-1 and Table 13-2 show the actions of the processor when an
SSE/SSE2/SSE3/SSSE3/SSE4 instruction is executed, depending on the:
Vol. 3 13-3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND PROCESSOR
• OSFXSR and OSXMMEXCPT flags in control register CR4
• SSE/SSE2/SSE3/SSSE3/SSE4 feature flags returned by CPUID
• EM, MP, and TS flags in control register CR0
Table 13-1. Action Taken for Combinations of OSFXSR, OSXMMEXCPT, SSE, SSE2,
SSE3, EM, MP, and TS1
CR4
CPUID
CR0 Flags
4
OSFXSR OSXMMEXCPT
SSE,
EM MP
TS
Action
SSE2,
2
SSE3
3
SSE4_1
5
0
1
1
1
X
X
0
1
1
X
X
1
0
1
1
1
1
X
X
X
0
#UD exception.
#UD exception.
#UD exception.
X
X
0
Execute instruction; #UD exception
if unmasked SIMD floating-point
exception is detected.
1
1
X
1
1
0
0
1
1
0
1
Execute instruction; #XM exception
if unmasked SIMD floating-point
exception is detected.
1
#NM exception.
NOTES:
1. For execution of any SSE/SSE2/SSE3 instruction except the PAUSE, PREFETCHh, SFENCE,
LFENCE, MFENCE, MOVNTI, and CLFLUSH instructions.
2. Exception conditions due to CR4.OSFXSR or CR4.OSXMMEXCPT do not apply to FISTTP.
3. Only applies to DPPS, DPPD, ROUNDPS, ROUNDPD, ROUNDSS, ROUNDSD.
4. For processors that support the MMX instructions, the MP flag should be set.
5. X — Don’t care.
13-4 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND
Table 13-2. Action Taken for Combinations of OSFXSR, SSSE3, SSE4, EM, and TS
CR4
CPUID
CR0 Flags
EM TS
OSFXSR
SSSE3
SSE4_1*
SSE4_2**
Action
0
1
1
1
X***
0
X
X
1
0
X
X
X
1
#UD exception.
#UD exception.
#UD exception.
#NM exception.
1
1
NOTES:
* Applies to SSE4_1 instructions except DPPS, DPPD, ROUNDPS, ROUNDPD, ROUNDSS, ROUNDSD.
** Applies to SSE4_2 instructions except CRC32 and POPCNT.
***X — Don’t care.
The SIMD floating-point exception mask bits (bits 7 through 12), the flush-to-zero
flag (bit 15), the denormals-are-zero flag (bit 6), and the rounding control field (bits
13 and 14) in the MXCSR register should be left in their default values of 0. This
permits the application to determine how these features are to be used.
13.1.5
Providing Non-Numeric Exception Handlers for Exceptions
Generated by the SSE/SSE2/SSE3/SSSE3/SSE4 Instructions
SSE/SSE2/SSE3/SSSE3/SSE4 instructions can generate the same type of memory
access exceptions (such as, page fault, segment not present, and limit violations)
and other non-numeric exceptions as other Intel 64 and IA-32 architecture instruc-
tions generate.
Ordinarily, existing exception handlers can handle these and other non-numeric
exceptions without code modification. However, depending on the mechanisms used
in existing exception handlers, some modifications might need to be made.
The SSE/SSE2/SSE3/SSSE3/SSE4 extensions can generate the non-numeric excep-
tions listed below:
• Memory Access Exceptions:
— Invalid opcode (#UD).
— Stack-segment fault (#SS).
— General protection (#GP). Executing most SSE/SSE2/SSE3 instructions with
an unaligned 128-bit memory reference generates a general-protection
exception. (The MOVUPS and MOVUPD instructions allow unaligned a loads or
stores of 128-bit memory locations, without generating a general-protection
exception.) A 128-bit reference within the stack segment that is not aligned
Vol. 3 13-5
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND PROCESSOR
to a 16-byte boundary will also generate a general-protection exception,
instead a stack-segment fault exception (#SS).
— Page fault (#PF).
— Alignment check (#AC). When enabled, this type of alignment check
operates on operands that are less than 128-bits in size: 16-bit, 32-bit, and
64-bit. To enable the generation of alignment check exceptions, do the
following:
•
•
•
Set the AM flag (bit 18 of control register CR0)
Set the AC flag (bit 18 of the EFLAGS register)
CPL must be 3
If alignment check exceptions are enabled, 16-bit, 32-bit, and 64-bit
misalignment will be detected for the MOVUPD and MOVUPS instructions;
detection of 128-bit misalignment is not guaranteed and may vary with
implementation.
• System Exceptions:
— Invalid-opcode exception (#UD). This exception is generated when executing
SSE/SSE2/SSE3/SSSE3 instructions under the following conditions:
•
•
•
•
SSE/SSE2/SSE3/SSSE3/SSE4_1/SSE4_2 feature flags returned by
CPUID are set to 0. This condition does not affect the CLFLUSH
instruction, nor POPCNT.
The CLFSH feature flag returned by the CPUID instruction is set to 0. This
exception condition only pertains to the execution of the CLFLUSH
instruction.
The POPCNT feature flag returned by the CPUID instruction is set to 0.
This exception condition only pertains to the execution of the POPCNT
instruction.
The EM flag (bit 2) in control register CR0 is set to 1, regardless of the
value of TS flag (bit 3) of CR0. This condition does not affect the PAUSE,
PREFETCHh, MOVNTI, SFENCE, LFENCE, MFENSE, CLFLUSH, CRC32 and
POPCNT instructions.
•
•
The OSFXSR flag (bit 9) in control register CR4 is set to 0. This condition
MASKMOVQ, MOVNTQ, MOVNTI, PAUSE, PREFETCHh, SFENCE, LFENCE,
MFENCE, CLFLUSH, CRC32 and POPCNT instructions.
Executing a instruction that causes a SIMD floating-point exception when
the OSXMMEXCPT flag (bit 10) in control register CR4 is set to 0. See
Section 13.5.1, “Using the TS Flag to Control the Saving of the x87 FPU,
MMX, SSE, SSE2, SSE3 SSSE3 and SSE4 State.”
13-6 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND
— Device not available (#NM). This exception is generated by executing a
SSE/SSE2/SSE3/SSSE3/SSE4 instruction when the TS flag (bit 3) of CR0 is
set to 1.
Other exceptions can occur indirectly due to faulty execution of the above
exceptions.
13.1.6
Providing an Handler for the SIMD Floating-Point Exception
(#XM)
SSE/SSE2/SSE3/SSSE3/SSE4 instructions do not generate numeric exceptions on
packed integer operations. They can generate the following numeric (SIMD floating-
point) exceptions on packed and scalar single-precision and double-precision
floating-point operations.
• Invalid operation (#I)
• Divide-by-zero (#Z)
• Denormal operand (#D)
• Numeric overflow (#O)
• Numeric underflow (#U)
• Inexact result (Precision) (#P)
These SIMD floating-point exceptions (with the exception of the denormal operand
exception) are defined in the IEEE Standard 754 for Binary Floating-Point Arithmetic
tions (#MF) to be generated for x87 FPU instructions.
Each of these exceptions can be masked, in which case the processor returns a
reasonable result to the destination operand without invoking an exception handler.
However, if any of these exceptions are left unmasked, detection of the exception
condition results in a SIMD floating-point exception (#XM) being generated. See
Chapter 6, “Interrupt 19—SIMD Floating-Point Exception (#XM).”
To handle unmasked SIMD floating-point exceptions, the operating system or execu-
tive must provide an exception handler. The section titled “SSE and SSE2 SIMD
Floating-Point Exceptions” in Chapter 11, “Programming with Streaming SIMD
Extensions 2 (SSE2),” of the Intel® 64 and IA-32 Architectures Software Developer’s
Manual, Volume 1, describes the SIMD floating-point exception classes and gives
suggestions for writing an exception handler to handle them.
To indicate that the operating system provides a handler for SIMD floating-point
exceptions (#XM), the OSXMMEXCPT flag (bit 10) must be set in control register
CR0.
Vol. 3 13-7
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND PROCESSOR
13.1.6.1 Numeric Error flag and IGNNE#
SSE/SSE2/SSE3/SSE4 extensions ignore the NE flag in control register CR0 (that is,
treats it as if it were always set) and the IGNNE# pin. When an unmasked SIMD
floating-point exception is detected, it is always reported by generating a SIMD
floating-point exception (#XM).
13.2
EMULATION OF SSE/SSE2/SSE3/SSSE3/SSE4
EXTENSIONS
The Intel 64 and IA-32 architecture does not support emulation of the
SSE/SSE2/SSE3/SSSE3/SSE4 instructions, as they do for x87 FPU instructions.
The EM flag in control register CR0 (provided to invoke emulation of x87 FPU instruc-
tions) cannot be used to invoke emulation of SSE/SSE2/SSE3/SSSE3/SSE4 instruc-
tions. If an SSE/SSE2/SSE3/SSSE3/SSE4 instruction is executed when CR0.EM = 1,
an invalid opcode exception (#UD) is generated. See Table 13-1.
13.3
SAVING AND RESTORING THE
SSE/SSE2/SSE3/SSSE3/SSE4 STATE
The SSE/SSE2/SSE3/SSSE3/SSE4 state consists of the state of the XMM and MXCSR
registers. The recommended method for saving and restoring this state follows:
to memory.
• Execute an FXRSTOR instruction to restore the state of the XMM and MXCSR
registers from the image saved in memory by the FXSAVE instruction.
This save and restore method is required for all operating systems. See Section 13.5,
“Designing OS Facilities for AUTOMATICALLY Saving x87 FPU, MMX, and
SSE/SSE2/SSE3/SSSE3/SSE4 state on Task or Context Switches.”
In some cases, applications can only save the XMM and MXCSR registers in the
following way:
• Execute MOVDQ instructions to save the contents of each XMM registers to
memory.
• Execute a STMXCSR instruction to save the state of the MXCSR register to
memory.
In some cases, applications can only restore the XMM and MXCSR registers in the
following way:
• Execute MOVDQ instructions to read the saved contents of each XMM registers
from memory to XMM registers.
13-8 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND
• Execute a LDMXCSR instruction to restore the state of the MXCSR register from
memory.
13.4
TASK OR CONTEXT SWITCHES
simple method for saving and restoring this state. See Section 13.3, “Saving and
Restoring the SSE/SSE2/SSE3/SSSE3/SSE4 State.” These instructions offer the
added benefit of saving x87 FPU and MMX state as well.
Guidelines for writing such procedures are in Section 13.5, “Designing OS Facilities
for AUTOMATICALLY Saving x87 FPU, MMX, and SSE/SSE2/SSE3/SSSE3/SSE4 state
on Task or Context Switches.”
13.5
DESIGNING OS FACILITIES FOR AUTOMATICALLY
SAVING X87 FPU, MMX, AND
SSE/SSE2/SSE3/SSSE3/SSE4 STATE ON TASK OR
CONTEXT SWITCHES
The x87 FPU/MMX/SSE/SSE2/SSE3/SSSE3/SSE4 state consist of the state of the x87
FPU, MMX, XMM, and MXCSR registers. The FXSAVE and FXRSTOR instructions
provide a fast method for saving ad restoring this state. If task or context switching
facilities are already implemented in an operating system or executive and they use
FSAVE/FNSAVE and FRSTOR to save the x87 FPU and MMX state, these facilities can
be extended to save and restore SSE/SSE2/SSE3/SSSE3/SSE4 state by substituting
FXSAVE/FXRSTOR for FSAVE/FNSAVE and FRSTOR.
Where task or content switching facilities must be written from scratch, several
approaches can be taken for using the FXSAVE and FXRSTOR instructions to save and
restore x87 FPU/MMX/SSE/SSE2/SSE3/SSSE3/SSE4 state:
• The operating system can require applications that are intended be run as tasks
take responsibility for saving the state of the x87 FPU, MMX, XMM, and MXCSR
registers prior to a task suspension during a task switch and for restoring the
registers when the task is resumed. This approach is appropriate for cooperative
multitasking operating systems, where the application has control over (or is able
to determine) when a task switch is about to occur and can save state prior to the
task switch.
• The operating system can take the responsibility for automatically saving the x87
FPU, MMX, XMM, and MXCSR registers as part of the task switch process (using
an FXSAVE instruction) and automatically restoring the state of the registers
Vol. 3 13-9
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND PROCESSOR
when a suspended task is resumed (using an FXRSTOR instruction). Here, the
x87 FPU/MMX/SSE/SSE2/SSE3/SSE4 state must be saved as part of the task
state. This approach is appropriate for preemptive multitasking operating
systems, where the application cannot know when it is going to be preempted
and cannot prepare in advance for task switching. Here, the operating system is
responsible for saving and restoring the task and the x87
FPU/MMX/SSE/SSE2/SSE3 state when necessary.
saving of the MMX and x87 FPU state until an x87 FPU, MMX, or
SSE/SSE2/SSE3/SSSE3/SSE4 instruction is actually executed by the new task.
Using this approach, the x87 FPU/MMX/SSE/SSE2/SSE3/SSSE3/SSE4 state is
saved only if an x87 FPU/MMX/SSE/SSE2/SSE3/SSSE3/SSE4 instruction needs
to be executed in the new task. (See Section 13.5.1, “Using the TS Flag to
Control the Saving of the x87 FPU, MMX, SSE, SSE2, SSE3 SSSE3 and SSE4
State,” for more information.)
13.5.1
Using the TS Flag to Control the Saving of the
x87 FPU, MMX, SSE, SSE2, SSE3 SSSE3 and SSE4 State
Saving the x87 FPU/MMX/SSE/SSE2/SSE3/SSSE3/SSE4 state using FXSAVE requires
processor overhead. If the new task does not access x87 FPU, MMX, XMM, and
MXCSR registers, avoid overhead by not automatically saving the state on a task
switch.
The TS flag in control register CR0 is provided to allow the operating system to delay
saving the x87 FPU/MMX/SSE/SSE2/SSE3/SSSE3/SSE4 state until an instruction
that actually accesses this state is encountered in a new task. When the TS flag is
set, the processor monitors the instruction stream for an x87
FPU/MMX/SSE/SSE2/SSE3/SSSE3/SSE4 instruction. When the processor detects
one of these instructions, it raises a device-not-available exception (#NM) prior to
executing the instruction. The device-not-available exception handler can then be
used to save the x87 FPU/MMX/SSE/SSE2/SSE3/SSSE3/SSE4 state for the previous
task (using an FXSAVE instruction) and load the x87
FPU/MMX/SSE/SSE2/SSE3/SSSE3/SSE4 state for the current task (using an
FXRSTOR instruction). If the task never encounters an x87
FPU/MMX/SSE/SSE2/SSE3/SSSE3/SSE4 instruction, the device-not-available excep-
tion will not be raised and a task state will not be saved unnecessarily.
NOTE
The CRC32 and POPCNT instructions do not operate on the x87
FPU/MMX/SSE/SSE2/SSE3/SSSE3/SSE4 state. They operate on the
general-purpose registers and are not involved in the OS’s lazy
FXSAVE/FXRSTOR technique.
13-10 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND
register CR0) or implicitly (using the IA-32 architecture’s native task switching mech-
anism). When the native task switching mechanism is used, the processor automati-
cally sets the TS flag on a task switch. After the device-not-available handler has
saved the x87 FPU/MMX/SSE/SSE2/SSE3/SSSE3/SSE4 state, it should execute the
CLTS instruction to clear the TS flag.
Figure 13-1 gives an example of an operating system that implements x87
FPU/MMX/SSE/SSE2/SSE3/SSSE3/SSE4 state saving using the TS flag. In this
example, task A is the currently running task and task B is the new task. The oper-
ating system maintains a save area for the x87
FPU/MMX/SSE/SSE2/SSE3/SSSE3/SSE4 state for each task and defines a variable
(x87_MMX_SSE_SSE2_SSE3_StateOwner) that indicates the task that “owns” the
state. In this example, task A is the current owner.
On a task switch, the operating system task switching code must execute the
following pseudo-code to set the TS flag according to the current owner of the x87
FPU/MMX/SSE/SSE2/SSE3/SSSE3/SSE4 state. If the new task (task B in this
example) is not the current owner of this state, the TS flag is set to 1; otherwise, it is
set to 0.
Task A
Task B
Owner of x87 FPU,
MMX, XMM,
Application
MXCSR State
Operating System
Task B
Task A
CR0.TS=1 and x87 FPU
MMX, SSEx
Instruction is encountered
x87 FPU/MMX/
XMM/MXCSR
State Save Area
x87 FPU/MMX/
XMM/MXCSR
State Save Area
Operating System
Task Switching Code
Loads Task B
x87 FPU/MMX/
XMM/MXCSR State
Saves Task A
x87 FPU/MMX/
XMM/MXCSR State
Device-Not-Available
Exception Handler
Figure 13-1. Example of Saving the x87 FPU, MMX, SSE, SSE2, SSE3, and SSSE3
State During an Operating-System Controlled Task Switch
IF Task_Being_Switched_To ≠ x87FPU_MMX_XMM_MXCSR_StateOwner
THEN
CR0.TS ← 1;
ELSE
CR0.TS ← 0;
FI;
Vol. 3 13-11
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND PROCESSOR
If a new task attempts to access an x87 FPU, MMX, XMM, or MXCSR register while the
TS flag is set to 1, a device-not-available exception (#NM) is generated. The device-
not-available exception handler executes the following pseudo-code.
FXSAVE “To x87FPU/MMX/XMM/MXCSR State Save Area for Current
x87FPU_MMX_XMM_MXCSR_StateOwner”;
FXRSTOR “x87FPU/MMX/XMM/MXCSR State From Current Task’s
x87FPU/MMX/XMM/MXCSR State Save Area”;
x87FPU_MMX_XMM_MXCSR_StateOwner ← Current_Task;
CR0.TS ← 0;
This exception handler code performs the following tasks:
• Saves the x87 FPU, MMX, XMM, or MXCSR registers in the state save area for the
current owner of the x87 FPU/MMX/XMM/MXCSR state.
• Restores the x87 FPU, MMX, XMM, or MXCSR registers from the new task’s save
area for the x87 FPU/MMX/XMM/MXCSR state.
• Updates the current x87 FPU/MMX/XMM/MXCSR state owner to be the current
task.
• Clears the TS flag.
13.6
XSAVE/XRSTOR AND PROCESSOR EXTENDED STATE
MANAGEMENT
The features associated with managing processor extended states include
• An extensible data layout for existing and future processor state extensions. The
layout of the XSAVE/XRSTOR area extends from the 512-byte FXSAVE/FXRSTOR
layout to provide compatibility and migration path from managing the legacy
FXSAVE/FXRSTOR area. Specifically, the XSAVE/XRSTOR area layout consists of:
— The FXSAVE/FXRSTOR area (512 bytes, the layout is identical to the
FXSAVE/FXRSTOR area),
— The XSAVE header area (64 bytes),
— A finite set of save areas, each corresponding to a processor extended state
(see Intel® 64 and IA-32 Architectures Software Developer’s Manual,
Volume 2B, XSAVE instruction). The number of save areas, the offset and the
size of each save area is enumerated by CPUID leaf function 0DH.
• CPUID Enhancement: CPUID instruction provides information on
— CPUID.01H.ECX.XSAVE[bit 26]. A feature flag indicating the processor’s
support of XSAVE/XRSTOR architecture extensions
— CPUID.01H.ECX.OSXSAVE[bit 27]. A feature flag indicating whether OS has
enabled extensible state management and communicating that the OS
supports processor extended state management.
13-12 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND
— CPUID leaf function 0DH enumerates the list of processor states (including
legacy x87 FPU, SSE states and processor extended states), the offset and
size of individual save area for each processor extended state.
• Control register enhancement and dedicated register for enabling each processor
extended state: CR4. OSXSAVE[bit 18] and the XFEATURE_ENABLED_MASK
register (XCR0) are described in Chapter 2, “System Architecture Overview”.
XCR0 can be read at all privilege levels but written only at ring 0.
• Instructions to manage the XFEATURE_ENABLED_MASK register (XCR0) and the
XSAVE/XRSTOR area (see Intel® 64 and IA-32 Architectures Software
Developer’s Manual, Volume 2B):
— XGETBV: reads XCR0.
— XSETBV: writes to XCR0, ring 0 only.
— XRSTOR: restores from memory the processor states specified by a bit vector
mask specified in EDX:EAX.
— XSAVE: saves the current processor states to memory according to a bit
vector mask in EDX:EAX.
13.6.1
XSAVE Header
The header section includes a “XSTATE_BV“ bit vector field. If the value of a bit in
HEADER.XSTATE_BV is 1, it indicates that the corresponding processor extended
state was written to the respective save area in memory by the XSAVE instruction.
If software modifies the save area image of a particular processor state component
directly, it is responsible to update the corresponding bit in HEADER.XSTATE_BV to 1.
by XRSTOR.
The order of bit vectors in XSTATE_BV matches those of the
XFEATURE_ENABLED_MASK register (XCR0). Although XCR0 has only two bits
initially defined for state management, the general relationship between the value of
XSTATE_BV and the corresponding processor state in the XSAVE/XRSTOR layout is
depicted in Figure 13-2.
Vol. 3 13-13
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND PROCESSOR
XState_BV
Bit Position
63
4
1
3
0
2
1
0
..................................
1
1
1
FXSAVE
X87 FPU State
SSE State
FXRSTOR
Save Area
XState_BV, ..
Updated
Header
Ext_SaveArea2
Not updated
Ext_SaveArea3
Ext_SaveArea4
Updated
.........................
of Processor State Extensions
The XSAVE header is 64 bytes in length and must be aligned on 64 byte boundary.
Therefore, the XSAVE/XRSTOR region must be aligned on 64-byte boundary. The
format of the header is as follows (see Table 13-3):
Table 13-3. XSAVE Header Format
15:8
Reserved (Must be zero)
Reserved
7:0
XSTATE_BV
Byte Offset
0
Reserved (Must be zero)
Reserved
16
32
48
Reserved
Reserved
Reserved
The value of each bit in HEADER.XSTATE_BV may affect the action performed by
XRSTOR, depending on the logical value of the respective bits in the
XFEATURE_ENABLED_MASK register (XCR0), the restore bit mask (EDX:EAX input to
XRSTOR), and HEADER.XSTATE_BV. When an XRSTOR instruction is executed with a
restore bit mask selecting the i’th bit vector (and the corresponding XCR0 bit is
13-14 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND
enabled), a value of "1" in the corresponding bit of HEADER.XSTATE_BV causes the
processor state to be updated with contents of the save area read from the memory
image. A value of "0" in HEADER.XSTATE_BV causes the processor state to be initial-
ized by hardware supplied values instead of from memory (See the operation detail
of XRSTOR in Intel® 64 and IA-32 Architectures Software Developer’s Manual,
Volume 2B).
The save area image corresponding to a bit with "0" value in HEADER.XSTATE_BV
may or may not contain the correct state information. XRSTOR will ensure the
register state for a component is properly initialized regardless of the value of the
save area when the component header bit is zero.
13.7
INTEROPERABILITY OF XSAVE/XRSTOR AND
FXSAVE/FXRSTOR
FXSAVE instruction writes x87 FPU and SSE state information to a 512-byte FXSAVE,
FXRSTOR save area. FXRSTOR restores the processor’s x87 FPU and SSE states from
FXSAVE/FXRSTOR save area image. XSAVE/XRSTOR instructions support x87 FPU
and SSE states using the same layout as the FXSAVE/FXRSTOR area to provide
interoperability of FXSAVE versus XSAVE, and FXRSTOR versus XRSTOR.
XSAVE/XRSTOR provides the additional flexibility for system software to manage SSE
state independent of x87 FPU states. Thus system software that had been using
FXSAVE/FXRSTOR to manage x87 FPU and SSE states can transition to
XSAVE/XRSTOR to manage x87 FPU, SSE and other processor extended states in a
systematic and forward-looking manner.
It is also possible for system software to adopt an alternate approach of using
FXSAVE/FXRSTOR for x87 and SSE state management, and implementing forward
processor extended state management using XSAVE/XRSTOR. In this case, system
software must specify the bit vector mask in EDX:EAX appropriately when executing
XSAVE/XRSTOR instructions.
For instance, when using the XSAVE instruction, the OS can supply a bit vector in
equal to 0. Then, the XSAVE instruction will not write the processor’s x87 FPU and
SSE state into memory. Similarly for the XRSTOR instruction a bit vector mask in
EDX:EAX with the least two significant bit equal to 0 will cause the XRSTOR instruc-
tion to not restore nor initialize the processor’s x87 FPU and SSE state.
The processor’s action as a result of executing XRSTOR, on the x87 FPU state,
XFEATURE_ENABLED_MASK register are presumed to be 1). The x87 FPU or XMM
registers may be initialized by the processor (See XRSTOR operation in Intel® 64 and
IA-32 Architectures Software Developer’s Manual, Volume 2B). When the MXCSR
register is updated from memory, reserved bit checking is enforced. The
saving/restoring of MXCSR is bound to the SSE state, independent of the x87 FPU
state. The action of XSAVE is listed in Table 13-5.
Vol. 3 13-15
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND PROCESSOR
Table 13-4. XRSTOR Action on MXCSR, x87 FPU, XMM Register
EDX:EAX
XSTATE_BV
Bit 1 Bit 0
MXCSR
XMM Registers x87 FPU State
Bit 1
Bit 0
0
0
0
0
1
1
1
1
1
1
X
X
X
0
1
0
0
1
1
X
0
1
X
X
0
1
0
1
None
None
None
None
None
None
None
Init by processor
Load
1
1
0
Load/Check Init by processor
Load/Check Load
None
0
None
1
Load/Check Init by processor Init by processor
1
Load/Check Init by processor
Load
Init by processor
Load
1
Load/Check
Load/Check
Load
Load
1
Table 13-5. XSAVE Action on MXCSR, x87 FPU, XMM Register
1
EDX:EAX
XCR0
MXCSR
XMM Registers x87 FPU State
Bit 1
Bit 0
Bit 1
Bit 0
0
0
1
1
1
1
0
1
0
0
1
1
X
X
0
1
0
1
1
1
1
1
1
1
None
None
None
Store
None
Store
None
None
None
Store
None
Store
None
Store
None
None
Store
Store
NOTES:
1. XCR0 is the XFEATURE_ENABLED_MASK register. Note that attempts to set XCR0[0] to 0 cause
#GP.
XSAVE, XRSTOR instructions operating on FP or SSE state will cause a #NM Device
Not Available) exception, if CR0.TS is set. Using this feature, system software can
implement the “lazy restore” technique of managing x87 FPU/SSE state using either
FXSAVE/FXRSTOR or XSAVE/XRSTOR. It can be accomplished even with the inter-
mixing of FXSAVE and XSAVE instructions.
13-16 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND
13.8
DETECTION, ENUMERATION, ENABLING PROCESSOR
EXTENDED STATE SUPPORT
An OS can determine if the XSAVE/XRSTOR/XGETBV/XSETBV instructions and the
XFEATURE_ENABLED_MASK register (XCR0) are available in the processor by
checking the value of CPUID.1.ECX.XSAVE to be 1. The OS must set CR4.OSXSAVE to
1 to enable the new instructions. The OS uses XSETBV to enable the processor state
component (setting the corresponding bit in XCR0 to 1) that it will manage using
XSAVE/XRSTOR. Bit 0 of XCR0 must be set to 1. The value of CR4.OSXSAVE is
reflected in CPUID.01H:ECX.OSXSAVE (bit 27) to communicate the setting to non-
privileged software.
Check
CPUID.1H:ECX.XSAVE?
HW support XSAVE, XRSTOR, XSETBV, XFEM
Enumerate
Extended state features
Buffer size requirement
Clear buffer to 0
XSETBV enabled
Set valid bits in
XCR0 via XSETBV
Set CR4.OSXSAVE
to 1
Figure 13-3. OS Enabling of Processor Extended State Support
The bits that must be enabled in the XFEATURE_ENABLED_MASK register (XCR0)
and the size of the memory region needed to save processor extended state informa-
tion must be enumerated by CPUID leaf 0DH with ECX = 0 as input. However, the
recommended usage by system software to use XSAVE/XRSTOR is to:
• Allocate a memory buffer according to the size reported by CPUID.(EAX=0DH,
ECX=0H):ECX. The value reported by CPUID.(EAX=0DH, ECX=0H):ECX always
includes the size of the header. Clear the entire buffer prior to being used by
XSAVE.
• Provide EDX:EAX with all bits set to 1 for XSAVE and XRSTOR instructions.
An alternative approach is to read the master bit vector mask EDX:EAX reported by
CPUID.(EAX=0D, ECX=0H). This mask may be used as input to the XSAVE/XRSTOR
Vol. 3 13-17
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND PROCESSOR
instructions, and provides a more constrained list of features than using all 1's in the
save mask.
The advantage of using a mask value of all-bits-set-to-1 for XSAVE/XRSTOR is that it
can simplify system software’s support for processor extended state management,
when multiple generations of hardware may support different number of processor
extended states as reported by CPUID. However, there may be additional implemen-
tation requirement of software modification that may arise due to a particular system
software or specific details introduced by a new processor extended state.
13.8.1
Application Programming Model and Processor Extended
States
New instruction set extensions may be introduced over time and operating on a
processor extended state that must be enabled in the XFEATURE_ENABLED_MASK
register (XCR0). The general application programming model for using such instruc-
tion set extensions are:
• Check if OS has enabled processor extended state management. If
CPUID.01H:ECX.OSXSAVE is 1, the OS has enabled the
XSAVE/XRSTOR/XSETBV/XGETBV instructions and the
XFEATURE_ENABLED_MASK register, and it has indicated support for the
processor extended state management.
Applications do not need to check the value of CPUID.01H:ECX.XSAVE because
“CPUID.01H:ECX.OSXSAVE = 1” implies OS has successfully verified
CPUID.01H:ECX.XSAVE = 1. CPUID.01H:ECX.OSXSAVE reflects the value of
CR4.OSXSAVE, and this bit cannot be set to 1 unless CPUID.01H:ECX.XSAVE = 1.
• Check whether the processor extended state component associated with a given
instruction set extension is enabled by the OS. The bits of EDX:EAX returned by
XGETBV as 1 indicate which processor extended state components have been
enabled by OS. Note, the CR4.OSFXSR is not used by OS to enable instruction
extensions requiring processor extended state support.
• Check the target instruction set extension is supported in the processor. Each
new instruction set extension is expected to provide a feature flag in CPUID when
it is introduced.
13-18 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND
Check feature flag
CPUID.1H:ECX.OXSAVE = 1?
OS provides processor
extended state management
Yes
Implied HW support for
XSAVE, XRSTOR, XGETBV, XCR0
Check enabled state in
XCR0 via XGETBV
Check feature flag
for Instruction set
State
enabled
ok to use
Instructions
Figure 13-4. Application Detection of New Instruction Extensions and Processor
Extended State
If all three requirements are met, applications can use the target new instruction set
instruction operating on a processor extended state corresponding to bit offset
higher than 1 in the XFEATURE_ENABLED_MASK register (XCR0) will cause a #UD
exception.
Newer instruction extensions operating on SSE state, but not on any processor
extended states corresponding bits in XCR0 with an offset higher than 1, follow the
programming model described by Section 13.1 through Section 13.5. XCR0 is not
required to enable OS support for SSE state management, but CR4.OSFXSR is
required.
Vol. 3 13-19
Download from Www.Somanuals.com. All Manuals Search And Download.
SYSTEM PROGRAMMING FOR INSTRUCTION SET EXTENSIONS AND PROCESSOR
13-20 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
CHAPTER 14
POWER AND THERMAL MANAGEMENT
This chapter describes facilities of Intel 64 and IA-32 architecture used for power
management and thermal monitoring.
14.1
ENHANCED INTEL SPEEDSTEP® TECHNOLOGY
®
Enhanced Intel SpeedStep Technology was introduced in the Pentium M processor;
it is available in Pentium 4, Intel Xeon, Intel Core™ Solo, Intel Core™ Duo, Intel
Atom™ and Intel Core™2 Duo processors. The technology manages processor
®
®
®
®
power consumption using performance state transitions. These states are defined as
discrete operating points associated with different frequencies.
Enhanced Intel SpeedStep Technology differs from previous generations of Intel
SpeedStep Technology in two ways:
• Centralization of the control mechanism and software interface in the processor
by using model-specific registers.
• Reduced hardware overhead; this permits more frequent performance state
transitions.
Previous generations of the Intel SpeedStep Technology require processors to be a
deep sleep state, holding off bus master transfers for the duration of a performance
state transition. Performance state transitions under the Enhanced Intel SpeedStep
Technology are discrete transitions to a new target frequency.
Support is indicated by CPUID, using ECX feature bit 07. Enhanced Intel SpeedStep
Technology is enabled by setting IA32_MISC_ENABLE MSR, bit 16. On reset, bit 16 of
IA32_MISC_ENABLE MSR is cleared.
14.1.1
Software Interface For Initiating Performance State
Transitions
State transitions are initiated by writing a 16-bit value to the IA32_PERF_CTL
register, see Figure 14-2. If a transition is already in progress, transition to a new
value will subsequently take effect.
Reads of IA32_PERF_CTL determine the last targeted operating point. The current
operating point can be read from IA32_PERF_STATUS. IA32_PERF_STATUS is
updated dynamically.
The 16-bit encoding that defines valid operating points is model-specific. Applications
and performance tools are not expected to use either IA32_PERF_CTL or
IA32_PERF_STATUS and should treat both as reserved. Performance monitoring
Vol. 3 14-1
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
tools can access model-specific events and report the occurrences of state
transitions.
14.2
P-STATE HARDWARE COORDINATION
The Advanced Configuration and Power Interface (ACPI) defines performance states
(P-state) that are used facilitate system software’s ability to manage processor
power consumption. Different P-state correspond to different performance levels
that are applied while the processor is actively executing instructions. Enhanced Intel
SpeedStep Technology supports P-state by providing software interfaces that control
the operating frequency and voltage of a processor.
With multiple processor cores residing in the same physical package, hardware
dependencies may exist for a subset of logical processors on a platform. These
dependencies may impose requirements that impact coordination of P-state transi-
tions. As a result, multi-core processors may require an OS to provide additional soft-
ware support for coordinating P-state transitions for those subsets of logical
processors.
A BIOS (following ACPI 3.0 specification) can choose to expose P-state as dependent
and hardware-coordinated to OS power management (OSPM) policy. To support
OSPMs, multi-core processors must have additional built-in support for P-state hard-
ware coordination and feedback.
Intel 64 and IA-32 processors with dependent P-state amongst a subset of logical
processors permit hardware coordination of P-state and provide a hardware-coordi-
nation feedback mechanism using IA32_MPERF MSR and IA32_APERF MSR. See
Figure 14-1 for an overview of the two 64-bit MSRs and the bullets below for a
detailed description:
63
0
63
0
IA32_MPERF (Addr: E7H)
IA32_APERF (Addr: E8H)
Figure 14-1. IA32_MPERF MSR and IA32_APERF MSR for P-state Coordination
• Use CPUID to check the P-State hardware coordination feedback capability bit.
CPUID.06H.ECX[Bit 0] = 1 indicates IA32_MPERF MSR and IA32_APERF MSR are
present.
• IA32_MPERF MSR (0xE7) increments in proportion to a fixed frequency, which is
configured when the processor is booted.
14-2 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
• IA32_APERF MSR (0xE8) increments in proportion to actual performance, while
accounting for hardware coordination of P-state and TM1/TM2; or software
initiated throttling.
• The MSRs are per logical processor; they measure performance only when the
targeted processor is in the C0 state.
should not attach meaning to the content of the individual of IA32_APERF or
IA32_MPERF MSRs.
• When either MSR overflows, both MSRs are reset to zero and continue to
increment.
• Both MSRs are full 64-bits counters. Each MSR can be written to independently.
However, software should follow the guidelines illustrated in Example 14-1.
If P-states are exposed by the BIOS as hardware coordinated, software is expected
to confirm processor support for P-state hardware coordination feedback and use the
away the current MSR values (for determination of the delta of the counter ratio at a
later time) or reset both MSRs (execute WRMSR with 0 to these MSRs individually) at
the start of the time window used for making the P-state decision. When not reset-
ting the values, overflow of the MSRs can be detected by checking whether the new
values read are less than the previously saved values.
Example 14-1 demonstrates steps for using the hardware feedback mechanism
provided by IA32_APERF MSR and IA32_MPERF MSR to determine a target P-state.
Example 14-1. Determine Target P-state From Hardware Coordinated Feedback
DWORD PercentBusy; // Percentage of processor time not idle.
// Measure “PercentBusy“ during previous sampling window.
// Typically, “PercentBusy“ is measure over a time scale suitable for
// power management decisions
//
// RDMSR of MCNT and ACNT should be performed without delay.
// Software needs to exercise care to avoid delays between
// the two RDMSRs (for example, interrupts).
MCNT = RDMSR(IA32_MPERF);
ACNT = RDMSR(IA32_APERF);
// PercentPerformance indicates the percentage of the processor
// that is in use. The calculation is based on the PercentBusy,
// that is the percentage of processor time not idle and the P-state
// hardware coordinated feedback using the ACNT/MCNT ratio.
// Note that both values need to be calculated over the same
// time window.
PercentPerformance = PercentBusy * (ACNT/MCNT);
Vol. 3 14-3
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
// This example does not cover the additional logic or algorithms
// necessary to coordinate multiple logical processors to a target P-state.
TargetPstate = FindPstate(PercentPerformance);
if (TargetPstate != currentPstate) {
SetPState(TargetPstate);
}
// WRMSR of MCNT and ACNT should be performed without delay.
// Software needs to exercise care to avoid delays between
// the two WRMSRs (for example, interrupts).
WRMSR(IA32_MPERF, 0);
WRMSR(IA32_APERF, 0);
14.3
SYSTEM SOFTWARE CONSIDERATIONS AND
OPPORTUNISTIC PROCESSOR PERFORMANCE
OPERATION
An Intel 64 processor may support a form of processor operation that takes advan-
tage of design headroom to opportunistically increase performance. In Intel Core i7
processors, Intel Turbo Boost Technology can convert thermal headroom into higher
performance across multi-threaded and single-threaded workloads. In Intel Core 2
processors, Intel Dynamic Acceleration can convert thermal headroom into higher
performance if only one thread is active.
14.3.1
Intel Dynamic Acceleration
Intel Core 2 Duo processor T 7700 introduces Intel Dynamic Acceleration (IDA). IDA
takes advantage of thermal design headroom and opportunistically allows a single
core to operate at a higher performance level when the operating system requests
increased performance.
14.3.2
System Software Interfaces for Opportunistic Processor
Performance Operation
Opportunistic processor operation, applicable to Intel Dynamic Acceleration and Intel
Turbo Boost Technology, has the following characteristics:
• A transition from a normal state of operation (e.g. IDA/Turbo mode disengaged)
to a target state is not guaranteed, but may occur opportunistically after the
14-4 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
corresponding enable mechanism is activated, the headroom is available and
certain criteria are met.
• The opportunistic processor performance operation is generally transparent to
most application software.
• System software (BIOS and Operating system) must be aware of hardware
support for opportunistic processor performance operation and may need to
temporarily disengage opportunistic processor performance operation when it
requires more predictable processor operation.
• When opportunistic processor performance operation is engaged, the OS should
use hardware coordination feedback mechanisms to prevent un-intended policy
effects if it is activated during inappropriate situations.
14.3.2.1 Discover Hardware Support and Enabling of Opportunistic
Processor Operation
If an Intel 64 processor has hardware support for opportunistic processor perfor-
mance operation, the power-on default state of IA32_MISC_ENABLES[38] indicates
the presence of such hardware support. For Intel 64 processors that support oppor-
tunistic processor performance operation, the default value is 1, indicating its pres-
ence. For processors that do not support opportunistic processor performance
IA32_MISC_ENABLES[38] allows BIOS to detect the presence of hardware support of
opportunistic processor performance operation.
IA32_MISC_ENABLES[38] is shared across all logical processors in a physical
package. It is written by BIOS during platform initiation to enable/disable opportu-
nistic processor operation in conjunction of OS power management capabilities, see
Section 14.3.2.2. BIOS can set IA32_MISC_ENABLES[38] with 1 to disable opportu-
nistic processor performance operation; it must clear the default value of
IA32_MISC_ENABLES[38] to 0 to enable opportunistic processor performance oper-
ation. OS and applications must use CPUID leaf 06H if it needs to detect processors
that has opportunistic processor operation enabled.
When CPUID is executed with EAX = 06H on input, Bit 1 of EAX in Leaf 06H (i.e.
CPUID.06H:EAX[1]) indicates opportunistic processor performance operation, such
as IDA, has been enabled by BIOS.
Opportunistic processor performance operation can be disabled by setting bit 38 of
IA32_MISC_ENABLES. This mechanism is intended for BIOS only. If
IA32_MISC_ENABLES[38] is set, CPUID.06H:EAX[1] will return 0.
14.3.2.2 OS Control of Opportunistic Processor Performance Operation
There may be phases of software execution in which system software cannot tolerate
the non-deterministic aspects of opportunistic processor performance operation. For
example, when calibrating a real-time workload to make a CPU reservation request
Vol. 3 14-5
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
to the OS, it may be undesirable to allow the possibility of the processor delivering
increased performance that cannot be sustained after the calibration phase.
System software can temporarily disengage opportunistic processor performance
operation by setting bit 32 of the IA32_PERF_CTL MSR (0199H), using a read-
modify-write sequence on the MSR. The opportunistic processor performance opera-
modify-write sequence. The DISENAGE bit in IA32_PERF_CTL is not reflected in bit
32 of the IA32_PERF_STATUS MSR (0198H), and it is not shared between logical
processors in a physical package. In order for OS to engage IDA/Turbo mode, the
BIOS must
• enable opportunistic processor performance operation, as described in Section
14.3.2.1,
• expose the operating points associated with IDA/Turbo mode to the OS.
33
16 15
32 31
63
0
Reserved
IDA/Turbo DISENGAGE
EIST Transition Target
Figure 14-2. IA32_PERF_CTL Register
Intel Dynamic Acceleration (IDA) and Intel Turbo Boost Technology can provide
opportunistic performance greater than the performance level corresponding to the
maximum qualified frequency of the processor (see CPUID’s brand string informa-
tion). System software can use a pair of MSRs to observe performance feedback.
Software must query for the presence of IA32_APERF and IA32_MPERF (see Section
14.2). The ratio between IA32_APERF and IA32_MPERF is architecturally defined and
a value greater than unity indicates performance increase occurred during the obser-
vation period due to IDA. Without incorporating such performance feedback, the
target P-state evaluation algorithm can result in a non-optimal P-state target.
There are other scenarios under which OS power management may want to disable
IDA, some of these are listed below:
• When engaging ACPI defined passive thermal management, it may be more
effective to disable IDA for the duration of passive thermal management.
• When the user has indicated a policy preference of power savings over perfor-
mance, OS power management may want to disable IDA while that policy is in
effect.
14-6 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
14.3.2.4 Application Awareness of Opportunistic Processor Operation
(Optional)
There may be situations that an end user or application software wishes to be aware
of turbo mode activity. It is possible for an application-level utility to periodically
check the occurrences of opportunistic processor operation. The basic elements of an
algorithm is described below, using the characteristics of Intel Turbo Boost Tech-
nology as example.
Using an OS-provided timer service, application software can periodically calculate
the ratio between unhalted-core-clockticks (UCC) relative to the unhalted-reference-
clockticks (URC) on each logical processor to determine if that logical processor had
been requested by OS to run at some frequency higher than the invariant TSC
frequency, or the OS has determined system-level demand has reduced sufficiently
state.
If an application software have access to information of the base operating ratio
between the invariant TSC frequency and the base clock (133.33 MHz), it can convert
the sampled ratio into a dynamic frequency estimate for each prior sampling period.
The base operating ratio can be read from MSR_PLATFORM_INFO[15:8].
The periodic sampling technique is depicted in Figure 14-3 and described below:
n-1
n
n+1
n+2
n+3
Sample period
Unhalted core clockticks
FixedCtr1
UCCn+1, 0
URCn+1, 0
UCCn+3, 0
URCn+3, 0
UCCn, 0
URCn, 0
UCCn+2, 0
URCn+2, 0
FixedCtr2
LP 0
LP 1
LP 2
LP 0
LP 1
LP 2
LP 0
LP 1
LP 0
LP 1
LP 2
Unhalted reference
clockticks
LP 2
.....
.....
.....
.....
Logical Processor i Turbo Activity Ratio = (UCCn+1, i - UCCn, i) / (URCn+1, i - URCn, i
)
Figure 14-3. Periodic Query of Activity Ratio of Opportunistic Processor Operation
• The sampling period chosen by the application (to program an OS timer service)
should be sufficiently large to avoid excessive polling overhead to other applica-
tions or tasks managed by the OS.
Vol. 3 14-7
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
• When the OS timer service transfers control, the application can use RDPMC
(with ECX = 4000_0001H) to read IA32_PERF_FIXED_CTR1 (MSR address 30AH)
to record the unhalted core clocktick (UCC) value; followed by RDPMC
(ECX=4000_0002H) to read IA32_PERF_FIXED_CTR2 (MSR address 30BH) to
record the unhalted reference clocktick (URC) value. This pair of values is needed
for each logical processor for each sampling period.
• The application can calculate the Turbo activity ratio based on the difference of
UCC between each sample period, over the difference of URC difference. The
effective frequency of each sample period of the logical processor, i, can be
estimated by:
(UCC
- UCC
)/(URC
- URC
)* Base_operating_ratio* 133.33MHz
n, i
n+1, i
n, i
n+1, i
It is possible that the OS had requested a lower-performance P-state during a
sampling period. Thus the ratio (UCC - UCC )/(URC - URC ) can reflect
n+1, i
n, i
n+1, i
n, i
the average of Turbo activity (driving the ratio above unity) and some lower P-state
transitions (causing the ratio to be < 1).
It is also possible that the OS might requested C-state transitions when the demand
is low. The above ratio generally does not account for cycles any logical processor
was idle. On Intel Core i7 processors, an application can make use of the time stamp
counter (IA-32_TSC) running at a constant frequency (i.e. Base_operating_ratio*
133.33MHz) during C-states. Thus software can calculate ratios that can indicate
fractions of sample period spent in the C0 state, using the unhalted reference clock-
ticks and the invariant TSC. Note the estimate of fraction spent in C0 may be affected
by SMM handler if the system software makes use of the “FREEZE_WHILE_SMM_EN“
capability to freeze performance counter values while the SMM handler is servicing
an SMI (see Chapter 20, “Introduction to Virtual-Machine Extensions”).
14.3.3
Intel Turbo Boost Technology
Intel Turbo Boost Technology is supported in Intel Core i7 processors and Intel Xeon
processors based on Intel microarchitecture (Nehalem). It uses the same principle of
leveraging thermal headroom to dynamically increase processor performance for
single-threaded and multi-threaded/multi-tasking environment. The programming
interface described in Section 14.3.2 also applies to Intel Turbo Boost Technology.
14.3.4
Performance and Energy Bias Hint support
Intel 64 processors may support additional software hint to guide the hardware
heuristic of power management features to favor increasing dynamic performance or
conserve energy consumption.
Software can detect processor's capability to support performance-energy bias pref-
erence hint by examining bit 3 of ECX in CPUID leaf 6. The processor supports this
capability if CPUID.06H:ECX.SETBH[bit 3] is set and it also implies the presence of a
new architectural MSR called IA32_ENERGY_PERF_BIAS (1B0H).
14-8 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
Software can program the lowest four bits of IA32_ENERGY_PERF_BIAS MSR with a
value from 0 - 15. The values represent a sliding scale, where a value of 0 (the
default reset value) corresponds to a hint preference for highest performance and a
value of 15 corresponds to the maximum energy savings. A value of 7 roughly trans-
lates into a hint to balance performance with energy consumption
4
3
63
0
Reserved
Energy Policy Preference Hint
Figure 14-4. IA32_ENERGY_PERF_BIAS Register
The layout of IA32_ENERGY_PERF_BIAS is shown in Figure 14-4. The scope of
IA32_ENERGY_PERF_BIAS is per logical processor, which means that each of the
logical processors in the package can be programmed with a different value. This
may be especially important in virtualization scenarios, where the performance /
energy requirements of one logical processor may differ from the other. Conflicting
"hints" from various logical processors at higher hierarchy level will be resolved in
favor of performance over energy savings.
Software can use whatever criteria it sees fit to program the MSR with the appro-
priate value. However, the value only serves as a hint to the hardware and the actual
impact on performance and energy savings is model specific.
14.4
MWAIT EXTENSIONS FOR ADVANCED POWER
MANAGEMENT
1
IA-32 processors may support a number of C-states that reduce power consumption
for inactive states. Intel Core Solo and Intel Core Duo processors support both
deeper C-state and MWAIT extensions that can be used by OS to implement power
management policy.
Software should use CPUID to discover if a target processor supports the enumera-
tion of MWAIT extensions. If CPUID.05H.ECX[Bit 0] = 1, the target processor
supports MWAIT extensions and their enumeration (see Chapter 3, “Instruction Set
1. The processor-specific C-states defined in MWAIT extensions can map to ACPI defined C-state
types (C0, C1, C2, C3). The mapping relationship depends on the definition of a C-state by proces-
sor implementation and is exposed to OSPM by the BIOS using the ACPI defined _CST table.
Vol. 3 14-9
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
Reference, A-M,” of Intel® 64 and IA-32 Architectures Software Developer’s Manual,
Volume 2A).
If CPUID.05H.ECX[Bit 1] = 1, the target processor supports using interrupts as
break-events for MWAIT, even when interrupts are disabled. Use this feature to
measure C-state residency as follows:
• Software can write to bit 0 in the MWAIT Extensions register (ECX) when issuing
an MWAIT to enter into a processor-specific C-state or sub C-state.
• When a processor comes out of an inactive C-state or sub C-state, software can
read a timestamp before an interrupt service routine (ISR) is potentially
executed.
CPUID.05H.EDX allows software to enumerate processor-specific C-states and sub
C-states available for use with MWAIT extensions. IA-32 processors may support
more than one C-state of a given C-state type. These are called sub C-states. Numer-
ically higher C-state have higher power savings and latency (upon entering and
exiting) than lower-numbered C-state.
At CPL = 0, system software can specify desired C-state and sub C-state by using the
MWAIT hints register (EAX). Processors will not go to C-state and sub C-state deeper
than what is specified by the hint register. If CPL > 0 and if MONITOR/MWAIT is
supported at CPL > 0, the processor will only enter C1-state (regardless of the
C-state request in the hints register).
Executing MWAIT generates an exception on processors operating at a privilege level
where MONITOR/MWAIT are not supported.
NOTE
If MWAIT is used to enter a C-state (including sub C-state) that is
numerically higher than C1, a store to the address range armed by
MONITOR instruction will cause the processor to exit MWAIT if the
store was originated by other processor agents. A store from non-
processor agent may not cause the processor to exit MWAIT.
14.5
THERMAL MONITORING AND PROTECTION
The IA-32 architecture provides the following mechanisms for monitoring tempera-
ture and controlling thermal power:
1. The catastrophic shutdown detector forces processor execution to stop if the
processor’s core temperature rises above a preset limit.
2. Automatic and adaptive thermal monitoring mechanisms force the
processor to reduce it’s power consumption in order to operate within predeter-
mined temperature limits.
3. The software controlled clock modulation mechanism permits operating
systems to implement power management policies that reduce power
14-10 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
consumption; this is in addition to the reduction offered by automatic thermal
monitoring mechanisms.
4. On-die digital thermal sensor and interrupt mechanisms permit the OS to
manage thermal conditions natively without relying on BIOS or other system
board components.
The first mechanism is not visible to software. The other three mechanisms are
visible to software using processor feature information returned by executing CPUID
with EAX = 1.
The second mechanism includes:
• Automatic thermal monitoring provides two modes of operation. One mode
modulates the clock duty cycle; the second mode changes the processor’s
frequency. Both modes are used to control the core temperature of the processor.
• Adaptive thermal monitoring can provide flexible thermal management on
processors made of multiple cores.
The third mechanism modulates the clock duty cycle of the processor. As shown in
Figure 14-5, the phrase ‘duty cycle’ does not refer to the actual duty cycle of the
clock signal. Instead it refers to the time period during which the clock signal is
allowed to drive the processor chip. By using the stop clock mechanism to control
how often the processor is clocked, processor power consumption can be modulated.
Clock Applied to Processor
Stop-Clock Duty Cycle
25% Duty Cycle (example only)
Figure 14-5. Processor Modulation Through Stop-Clock Mechanism
For previous automatic thermal monitoring mechanisms, software controlled mecha-
nisms that changed processor operating parameters to impact changes in thermal
conditions. Software did not have native access to the native thermal condition of the
processor; nor could software alter the trigger condition that initiated software
program control.
The fourth mechanism (listed above) provides access to an on-die digital thermal
sensor using a model-specific register and uses an interrupt mechanism to alert soft-
ware to initiate digital thermal monitoring.
Vol. 3 14-11
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
14.5.1
Catastrophic Shutdown Detector
P6 family processors introduced a thermal sensor that acts as a catastrophic shut-
down detector. This catastrophic shutdown detector was also implemented in
Pentium 4, Intel Xeon and Pentium M processors. It is always enabled. When
processor core temperature reaches a factory preset level, the sensor trips and
processor execution is halted until after the next reset cycle.
14.5.2
Thermal Monitor
Pentium 4, Intel Xeon and Pentium M processors introduced a second temperature
sensor that is factory-calibrated to trip when the processor’s core temperature
crosses a level corresponding to the recommended thermal design envelop. The trip-
temperature of the second sensor is calibrated below the temperature assigned to
the catastrophic shutdown detector.
14.5.2.1 Thermal Monitor 1
The Pentium 4 processor uses the second temperature sensor in conjunction with a
mechanism called Thermal Monitor 1 (TM1) to control the core temperature of the
processor. TM1 controls the processor’s temperature by modulating the duty cycle of
the processor clock. Modulation of duty cycles is processor model specific. Note that
the processors STPCLK# pin is not used here; the stop-clock circuitry is controlled
internally.
Support for TM1 is indicated by CPUID.1:EDX.TM[bit 29] = 1.
TM1 is enabled by setting the thermal-monitor enable flag (bit 3) in
IA32_MISC_ENABLE [see Appendix B, “Model-Specific Registers (MSRs)”]. Following
a power-up or reset, the flag is cleared, disabling TM1. BIOS is required to enable
only one automatic thermal monitoring modes. Operating systems and applications
must not disable the operation of these mechanisms.
14.5.2.2 Thermal Monitor 2
An additional automatic thermal protection mechanism, called Thermal Monitor 2
(TM2), was introduced in the Intel Pentium M processor and also incorporated in
newer models of the Pentium 4 processor family. Intel Core Duo and Solo processors,
and Intel Core 2 Duo processor family all support TM1 and TM2. TM2 controls the
core temperature of the processor by reducing the operating frequency and voltage
of the processor and offers a higher performance level for a given level of power
reduction than TM1.
TM2 is triggered by the same temperature sensor as TM1. The mechanism to enable
TM2 may be implemented differently across various IA-32 processor families with
different CPUID signatures in the family encoding value, but will be uniform within an
IA-32 processor family.
14-12 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
Support for TM2 is indicated by CPUID.1:ECX.TM2[bit 8] = 1.
14.5.2.3 Two Methods for Enabling TM2
On processors with CPUID family/model/stepping signature encoded as 0x69n or
0x6Dn (early Pentium M processors), TM2 is enabled if the TM_SELECT flag (bit 16)
of the MSR_THERM2_CTL register is set to 1 (Figure 14-6) and bit 3 of the
IA32_MISC_ENABLE register is set to 1.
Following a power-up or reset, the TM_SELECT flag may be cleared. BIOS is required
to enable either TM1 or TM2. Operating systems and applications must not disable
mechanisms that enable TM1 or TM2. If bit 3 of the IA32_MISC_ENABLE register is
set and TM_SELECT flag of the MSR_THERM2_CTL register is cleared, TM1 is
enabled.
0
31
16
Reserved
Reserved
TM_SELECT
Figure 14-6. MSR_THERM2_CTL Register On Processors with CPUID
Family/Model/Stepping Signature Encoded as 0x69n or 0x6Dn
On processors introduced after the Pentium 4 processor (this includes most Pentium
M processors), the method used to enable TM2 is different. TM2 is enable by setting
bit 13 of IA32_MISC_ENABLE register to 1. This applies to Intel Core Duo, Core Solo,
and Intel Core 2 processor family.
The target operating frequency and voltage for the TM2 transition after TM2 is trig-
gered is specified by the value written to MSR_THERM2_CTL, bits 15:0 (Figure 14-7).
Following a power-up or reset, BIOS is required to enable at least one of these two
thermal monitoring mechanisms. If both TM1 and TM2 are supported, BIOS may
choose to enable TM2 instead of TM1. Operating systems and applications must not
disable the mechanisms that enable TM1or TM2; and they must not alter the value in
bits 15:0 of the MSR_THERM2_CTL register.
Vol. 3 14-13
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
15
63
0
Reserved
TM2 Transition Target
Figure 14-7. MSR_THERM2_CTL Register for Supporting TM2
14.5.2.4 Performance State Transitions and Thermal Monitoring
If the thermal control circuitry (TCC) for thermal monitor (TM1/TM2) is active, writes
to the IA32_PERF_CTL will effect a new target operating point as follows:
• If TM1 is enabled and the TCC is engaged, the performance state transition can
commence before the TCC is disengaged.
• If TM2 is enabled and the TCC is engaged, the performance state transition
specified by a write to the IA32_PERF_CTL will commence after the TCC has
disengaged.
14.5.2.5 Thermal Status Information
The status of the temperature sensor that triggers the thermal monitor (TM1/TM2) is
indicated through the thermal status flag and thermal status log flag in the
IA32_THERM_STATUS MSR (see Figure 14-8).
The functions of these flags are:
• Thermal Status flag, bit 0 — When set, indicates that the processor core
temperature is currently at the trip temperature of the thermal monitor and that
the processor power consumption is being reduced via either TM1 or TM2,
depending on which is enabled. When clear, the flag indicates that the core
temperature is below the thermal monitor trip temperature. This flag is read only.
• Thermal Status Log flag, bit 1 — When set, indicates that the thermal sensor
has tripped since the last power-up or reset or since the last time that software
cleared this flag. This flag is a sticky bit; once set it remains set until cleared by
software or until a power-up or reset of the processor. The default state is clear.
14-14 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
63
2 1 0
Reserved
Thermal Status Log
Thermal Status
Figure 14-8. IA32_THERM_STATUS MSR
After the second temperature sensor has been tripped, the thermal monitor
(TM1/TM2) will remain engaged for a minimum time period (on the order of 1 ms).
The thermal monitor will remain engaged until the processor core temperature drops
below the preset trip temperature of the temperature sensor, taking hysteresis into
account.
While the processor is in a stop-clock state, interrupts will be blocked from inter-
but does not cause interrupts to be lost. Outstanding interrupts remain pending until
clock modulation is complete.
The thermal monitor can be programmed to generate an interrupt to the processor
when the thermal sensor is tripped. The delivery mode, mask and vector for this
interrupt can be programmed through the thermal entry in the local APIC’s LVT (see
Section 10.6.1, “Local Vector Table”). The low-temperature interrupt enable and
high-temperature interrupt enable flags in the IA32_THERM_INTERRUPT MSR (see
Figure 14-9) control when the interrupt is generated; that is, on a transition from a
temperature below the trip point to above and/or vice-versa.
63
2 1 0
Reserved
Low-Temperature Interrupt Enable
High-Temperature Interrupt Enable
Figure 14-9. IA32_THERM_INTERRUPT MSR
• High-Temperature Interrupt Enable flag, bit 0 — Enables an interrupt to be
generated on the transition from a low-temperature to a high-temperature when
set; disables the interrupt when clear.(R/W).
• Low-Temperature Interrupt Enable flag, bit 1 — Enables an interrupt to be
generated on the transition from a high-temperature to a low-temperature when
set; disables the interrupt when clear.
The thermal monitor interrupt can be masked by the thermal LVT entry. After a
power-up or reset, the low-temperature interrupt enable and high-temperature
Vol. 3 14-15
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
interrupt enable flags in the IA32_THERM_INTERRUPT MSR are cleared (interrupts
are disabled) and the thermal LVT entry is set to mask interrupts. This interrupt
should be handled either by the operating system or system management mode
(SMM) code.
Note that the operation of the thermal monitoring mechanism has no effect upon the
clock rate of the processor's internal high-resolution timer (time stamp counter).
14.5.2.6 Adaptive Thermal Monitor
The Intel Core 2 Duo processor family supports enhanced thermal management
mechanism, referred to as Adaptive Thermal Monitor (Adaptive TM).
Unlike TM2, Adaptive TM is not limited to one TM2 transition target. During a thermal
trip event, Adaptive TM (if enabled) selects an optimal target operating point based
on whether or not the current operating point has effectively cooled the processor.
and TM2 feature flags and enable all available thermal control mechanisms (including
Adaptive TM) at platform initiation.
Adaptive TM is available only to a subset of processors that support TM2.
In each chip-multiprocessing (CMP) silicon die, each core has a unique thermal
sensor that triggers independently. These thermal sensor can trigger TM1 or TM2
transitions in the same manner as described in Section 14.5.2.1 and Section
14.5.2.2. The trip point of the thermal sensor is not programmable by software since
it is set during the fabrication of the processor.
Each thermal sensor in a processor core may be triggered independently to engage
thermal management features. In Adaptive TM, both cores will transition to a lower
frequency and/or lower voltage level if one sensor is triggered.
Triggering of this sensor is visible to software via the thermal interrupt LVT entry in
the local APIC of a given core.
14.5.3
Software Controlled Clock Modulation
Pentium 4, Intel Xeon and Pentium M processors also support software-controlled
clock modulation. This provides a means for operating systems to implement a power
management policy to reduce the power consumption of the processor. Here, the
stop-clock duty cycle is controlled by software through the
IA32_CLOCK_MODULATION MSR (see Figure 14-10).
14-16 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
63
5 4 3 1 0
Reserved
On-Demand Clock Modulation Enable
On-Demand Clock Modulation Duty Cycle
Reserved
Figure 14-10. IA32_CLOCK_MODULATION MSR
The IA32_CLOCK_MODULATION MSR contains the following flag and field used to
enable software-controlled clock modulation and to select the clock modulation duty
cycle:
• On-Demand Clock Modulation Enable, bit 4 — Enables on-demand software
controlled clock modulation when set; disables software-controlled clock
modulation when clear.
• On-Demand Clock Modulation Duty Cycle, bits 1 through 3 — Selects the
on-demand clock modulation duty cycle (see Table 14-1). This field is only active
when the on-demand clock modulation enable flag is set.
Note that the on-demand clock modulation mechanism (like the thermal monitor)
controls the processor’s stop-clock circuitry internally to modulate the clock signal.
The STPCLK# pin is not used in this mechanism.
Table 14-1. On-Demand Clock Modulation Duty Cycle Field Encoding
Duty Cycle Field Encoding
Duty Cycle
000B
001B
010B
011B
100B
101B
110B
111B
Reserved
12.5% (Default)
25.0%
37.5%
50.0%
63.5%
75%
87.5%
The on-demand clock modulation mechanism can be used to control processor power
consumption. Power management software can write to the
IA32_CLOCK_MODULATION MSR to enable clock modulation and to select a modula-
tion duty cycle. If on-demand clock modulation and TM1 are both enabled and the
thermal status of the processor is hot (bit 0 of the IA32_THERM_STATUS MSR is set),
Vol. 3 14-17
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
clock modulation at the duty cycle specified by TM1 takes precedence, regardless of
the setting of the on-demand clock modulation duty cycle.
For Hyper-Threading Technology enabled processors, the
IA32_CLOCK_MODULATION register is duplicated for each logical processor. In order
for the On-demand clock modulation feature to work properly, the feature must be
enabled on all the logical processors within a physical processor. If the programmed
duty cycle is not identical for all the logical processors, the processor clock will modu-
late to the highest duty cycle programmed.
For the P6 family processors, on-demand clock modulation was implemented
through the chipset, which controlled clock modulation through the processor’s
STPCLK# pin.
14.5.4
Detection of Thermal Monitor and Software Controlled
Clock Modulation Facilities
The ACPI flag (bit 22) of the CPUID feature flags indicates the presence of the
IA32_THERM_STATUS, IA32_THERM_INTERRUPT, IA32_CLOCK_MODULATION
MSRs, and the xAPIC thermal LVT entry.
The TM1 flag (bit 29) of the CPUID feature flags indicates the presence of the auto-
matic thermal monitoring facilities that modulate clock duty cycles.
14.5.5
On Die Digital Thermal Sensors
On die digital thermal sensor can be read using an MSR (no I/O interface). In Intel
Core Duo processors, each core has a unique digital sensor whose temperature is
accessible using an MSR. The digital thermal sensor is the preferred method for
reading the die temperature because (a) it is located closer to the hottest portions of
the die, (b) it enables software to accurately track the die temperature and the
potential activation of thermal throttling.
14.5.5.1 Digital Thermal Sensor Enumeration
The processor supports a digital thermal sensor if CPUID.06H.EAX[0] = 1. If the
processor supports digital thermal sensor, EBX[bits 3:0] determine the number of
thermal thresholds that are available for use.
Software sets thermal thresholds by using the IA32_THERM_INTERRUPT MSR. Soft-
ware reads output of the digital thermal sensor using the IA32_THERM_STATUS
MSR.
14-18 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
14.5.5.2 Reading the Digital Sensor
Unlike traditional analog thermal devices, the output of the digital thermal sensor is
a temperature relative to the maximum supported operating temperature of the
processor.
Temperature measurements returned by digital thermal sensors are always at or
below TCC activation temperature. Critical temperature conditions are detected
using the “Critical Temperature Status” bit. When this bit is set, the processor is
operating at a critical temperature and immediate shutdown of the system should
occur. Once the “Critical Temperature Status” bit is set, reliable operation is not guar-
anteed.
See Figure 14-11 for the layout of IA32_THERM_STATUS MSR. Bit fields include:
• Thermal Status (bit 0, RO) — This bit indicates whether the digital thermal
sensor high-temperature output signal (PROCHOT#) is currently active. Bit 0 = 1
indicates the feature is active. This bit may not be written by software; it reflects
the state of the digital thermal sensor.
• Thermal Status Log (bit 1, R/WC0) — This is a sticky bit that indicates the
history of the thermal sensor high temperature output signal (PROCHOT#).
Bit 1 = 1 if PROCHOT# has been asserted since a previous RESET or the last time
software cleared the bit. Software may clear this bit by writing a zero.
• PROCHOT# or FORCEPR# Event (bit 2, RO) — Indicates whether PROCHOT#
or FORCEPR# is being asserted by another agent on the platform.
63
32 31
27
23 22 16 15
10
9
8
7
6
5
4
3
2
1
0
Reserved
Reading Valid
Resolution in Deg. Celsius
Digital Readout
Thermal Threshold #2 Log
Thermal Threshold #2 Status
Thermal Threshold #1 Log
Thermal Threshold #1 Status
Critical Temperature Log
Critical Temperature Status
PROCHOT# or FORCEPR# Log
PROCHOT# or FORCEPR# Event
Thermal Status Log
Thermal Status
Figure 14-11. IA32_THERM_STATUS Register
Vol. 3 14-19
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
• PROCHOT# or FORCEPR# Log (bit 3, R/WC0) — Sticky bit that indicates
whether PROCHOT# or FORCEPR# has been asserted by another agent on the
platform since the last clearing of this bit or a reset. If bit 3 = 1, PROCHOT# or
FORCEPR# has been externally asserted. Software may clear this bit by writing a
zero. External PROCHOT# assertions are only acknowledged if the Bidirectional
Prochot feature is enabled.
• Critical Temperature Status (bit 4, RO) — Indicates whether the critical
temperature detector output signal is currently active. If bit 4 = 1, the critical
temperature detector output signal is currently active.
• Critical Temperature Log (bit 5, R/WC0) — Sticky bit that indicates whether
the critical temperature detector output signal has been asserted since the last
clearing of this bit or reset. If bit 5 = 1, the output signal has been asserted.
Software may clear this bit by writing a zero.
• Thermal Threshold #1 Status (bit 6, RO) — Indicates whether the actual
temperature is currently higher than or equal to the value set in Thermal
Threshold #1. If bit 6 = 0, the actual temperature is lower. If bit 6 = 1, the
actual temperature is greater than or equal to TT#1. Quantitative information of
actual temperature can be inferred from Digital Readout, bits 22:16.
• Thermal Threshold #1 Log (bit 7, R/WC0) — Sticky bit that indicates
whether the Thermal Threshold #1 has been reached since the last clearing of
this bit or a reset. If bit 7 = 1, the Threshold #1 has been reached. Software may
clear this bit by writing a zero.
• Thermal Threshold #2 Status (bit 8, RO) — Indicates whether actual
temperature is currently higher than or equal to the value set in Thermal
Threshold #2. If bit 8 = 0, the actual temperature is lower. If bit 8 = 1, the
actual temperature is greater than or equal to TT#2. Quantitative information of
actual temperature can be inferred from Digital Readout, bits 22:16.
• Thermal Threshold #2 Log (bit 9, R/WC0) — Sticky bit that indicates
whether the Thermal Threshold #2 has been reached since the last clearing of
this bit or a reset. If bit 9 = 1, the Thermal Threshold #2 has been reached.
Software may clear this bit by writing a zero.
• Digital Readout (bits 22:16, RO) — Digital temperature reading in 1 degree
Celsius relative to the TCC activation temperature.
0: TCC Activation temperature,
1: (TCC Activation - 1) , etc. See the processor’s data sheet for details regarding
TCC activation.
A lower reading in the Digital Readout field (bits 22:16) indicates a higher actual
temperature.
• Resolution in Degrees Celsius (bits 30:27, RO) — Specifies the resolution
(or tolerance) of the digital thermal sensor. The value is in degrees Celsius. It is
recommended that new threshold values be offset from the current temperature
by at least the resolution + 1 in order to avoid hysteresis of interrupt generation.
14-20 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
• Reading Valid (bit 31, RO) — Indicates if the digital readout in bits 22:16 is
valid. The readout is valid if bit 31 = 1.
Changes to temperature can be detected using two thresholds (see Figure 14-12);
one is set above and the other below the current temperature. These thresholds have
the capability of generating interrupts using the core's local APIC which software
must then service. Note that the local APIC entries used by these thresholds are also
®
used by the Intel Thermal Monitor; it is up to software to determine the source of a
specific interrupt.
63
24 23
22
16
15
14
8
5
4
3
2
1
0
Reserved
Threshold #2 Interrupt Enable
Threshold #2 Value
Threshold #1 Interrupt Enable
Threshold #1 Value
Overheat Interrupt Enable
FORCPR# Interrupt Enable
PROCHOT# Interrupt Enable
Low Temp. Interrupt Enable
High Temp. Interrupt Enable
Figure 14-12. IA32_THERM_INTERRUPT Register
See Figure 14-12 for the layout of IA32_THERM_INTERRUPT MSR. Bit fields include:
• High-Temperature Interrupt Enable (bit 0, R/W) — This bit allows the BIOS
to enable the generation of an interrupt on the transition from low-temperature
to a high-temperature threshold. Bit 0 = 0 (default) disables interrupts;
bit 0 = 1 enables interrupts.
• Low-Temperature Interrupt Enable (bit 1, R/W) — This bit allows the BIOS
to enable the generation of an interrupt on the transition from high-temperature
to a low-temperature (TCC de-activation). Bit 1 = 0 (default) disables interrupts;
bit 1 = 1 enables interrupts.
• PROCHOT# Interrupt Enable (bit 2, R/W) — This bit allows the BIOS or OS
to enable the generation of an interrupt when PROCHOT# has been asserted by
another agent on the platform and the Bidirectional Prochot feature is enabled.
Bit 2 = 0 disables the interrupt; bit 2 = 1 enables the interrupt.
• FORCEPR# Interrupt Enable (bit 3, R/W) — This bit allows the BIOS or OS to
enable the generation of an interrupt when FORCEPR# has been asserted by
another agent on the platform. Bit 3 = 0 disables the interrupt; bit 3 = 1 enables
the interrupt.
Vol. 3 14-21
Download from Www.Somanuals.com. All Manuals Search And Download.
POWER AND THERMAL MANAGEMENT
• Critical Temperature Interrupt Enable (bit 4, R/W) — Enables the
generation of an interrupt when the Critical Temperature Detector has detected a
critical thermal condition. The recommended response to this condition is a
system shutdown. Bit 4 = 0 disables the interrupt; bit 4 = 1 enables the
interrupt.
• Threshold #1 Value (bits 14:8, R/W) — A temperature threshold, encoded
relative to the TCC Activation temperature (using the same format as the Digital
Readout). This threshold is compared against the Digital Readout and is used to
generate the Thermal Threshold #1 Status and Log bits as well as the Threshold
#1 thermal interrupt delivery.
• Threshold #1 Interrupt Enable (bit 15, R/W) — Enables the generation of
an interrupt when the actual temperature crosses the Threshold #1 setting in any
direction. Bit 15 = 0 enables the interrupt; bit 15 = 1 disables the interrupt.
• Threshold #2 Value (bits 22:16, R/W) —A temperature threshold, encoded
relative to the TCC Activation temperature (using the same format as the Digital
Readout). This threshold is compared against the Digital Readout and is used to
generate the Thermal Threshold #2 Status and Log bits as well as the Threshold
#2 thermal interrupt delivery.
• Threshold #2 Interrupt Enable (bit 23, R/W) — Enables the generation of
an interrupt when the actual temperature crosses the Threshold #2 setting in any
direction. Bit 23 = 0 enables the interrupt; bit 23 = 1 disables the interrupt.
14-22 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
CHAPTER 15
This chapter describes the machine-check architecture and machine-check
exception mechanism found in the Pentium 4, Intel Xeon, and P6 family
processors. See Chapter 6, “Interrupt 18—Machine-Check Exception
(#MC),” for more information on machine-check exceptions. A brief descrip-
tion of the Pentium processor’s machine check capability is also given.
Additionally, a signaling mechanism for software to respond to hardware
corrected machine check error is covered.
15.1
MACHINE-CHECK ARCHITECTURE
The Pentium 4, Intel Xeon, and P6 family processors implement a machine-
check architecture that provides a mechanism for detecting and reporting
hardware (machine) errors, such as: system bus errors, ECC errors, parity
errors, cache errors, and TLB errors. It consists of a set of model-specific
registers (MSRs) that are used to set up machine checking and additional
banks of MSRs used for recording errors that are detected.
The processor signals the detection of an uncorrected machine-check error
by generating a machine-check exception (#MC), which is an abort class
exception. The implementation of the machine-check architecture does not
ordinarily permit the processor to be restarted reliably after generating a
machine-check exception. However, the machine-check-exception handler
can collect information about the machine-check error from the machine-
check MSRs.
Starting with 45nm Intel 64 processor with CPUID signature
DisplayFamily_DisplayModel encoding of 06H_1AH (see CPUID instruction in
Chapter 3, “Instruction Set Reference, A-M” in the Intel® 64 and IA-32
Architectures Software Developer’s Manual, Volume 2A), the processor can
corrected machine-check error interrupt (CMCI). See Section 15.5 for detail.
Intel 64 processors supporting machine-check architecture and CMCI may
also support an additional enhancement, namely, support for software
recovery from certain uncorrected recoverable machine check errors. See
Section 15.6 for detail.
Vol. 3 15-1
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
15.2
COMPATIBILITY WITH PENTIUM PROCESSOR
The Pentium 4, Intel Xeon, and P6 family processors support and extend the
machine-check exception mechanism introduced in the Pentium processor.
The Pentium processor reports the following machine-check errors:
• data parity errors during read cycles
• unsuccessful completion of a bus cycle
The above errors are reported using the P5_MC_TYPE and P5_MC_ADDR
MSRs (implementation specific for the Pentium processor). Use the RDMSR
instruction to read these MSRs. See Appendix B, “Model-Specific Registers
When an error is detected, it is recorded in P5_MC_TYPE and P5_MC_ADDR;
the processor then generates a machine-check exception (#MC).
See Section 15.3.3, “Mapping of the PentiumProcessor Machine-Check
Errors to the Machine-Check Architecture,” and Section 15.10.2, “Pentium
Processor Machine-Check Exception Handling,” for information on compati-
bility between machine-check code written to run on the Pentium processors
and code written to run on P6 family processors.
15.3
MACHINE-CHECK MSRS
Machine check MSRs in the Pentium 4, Intel Xeon, and P6 family processors
consist of a set of global control and status registers and several error-
reporting register banks. See Figure 15-1.
15-2 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
Error-Reporting Bank Registers
(One Set for Each Hardware Unit)
Global Control MSRs
IA32_MCG_CAP MSR
63
63
63
0
0
0
63
63
63
63
63
0
0
0
0
0
IA32_MCi_CTL MSR
IA32_MCG_STATUS MSR
IA32_MCG_CTL MSR
IA32_MCi_STATUS MSR
IA32_MCi_ADDR MSR
IA32_MCi_MISC MSR
IA32_MCi_CTL2 MSR
Figure 15-1. Machine-Check MSRs
Each error-reporting bank is associated with a specific hardware unit (or
group of hardware units) in the processor. Use RDMSR and WRMSR to read
and to write these registers.
15.3.1
Machine-Check Global Control MSRs
The machine-check global control MSRs include the IA32_MCG_CAP,
IA32_MCG_STATUS, and IA32_MCG_CTL. See Appendix B, “Model-Specific
Registers (MSRs),” for the addresses of these registers.
15.3.1.1 IA32_MCG_CAP MSR
The IA32_MCG_CAP MSR is a read-only register that provides information
about the machine-check architecture of the processor. Figure 15-2 shows
the structure of the register in Pentium 4, Intel Xeon, and P6 family proces-
sors.
Vol. 3 15-3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
63
0
25 24 23
1615
12 11 10 9
8
7
Count
Reserved
MCG_SER_P[24]
MCG_EXT_CNT[23:16]
MCG_TES_P[11]
MCG_CMCI_P[10]
MCG_EXT_P[9]
MCG_CTL_P[8]
Figure 15-2. IA32_MCG_CAP Register
Where:
• Count field, bits 7:0 — Indicates the number of hardware unit error-reporting
banks available in a particular processor implementation.
• MCG_CTL_P (control MSR present) flag, bit 8 — Indicates that the processor
implements the IA32_MCG_CTL MSR when set; this register is absent when clear.
• MCG_EXT_P (extended MSRs present) flag, bit 9 — Indicates that the
processor implements the extended machine-check state registers found starting
at MSR address 180H; these registers are absent when clear.
• MCG_CMCI_P (Corrected MC error counting/signaling extension
present) flag, bit 10 — Indicates (when set) that extended state and
associated MSRs necessary to support the reporting of an interrupt on a
corrected MC error event and/or count threshold of corrected MC errors, is
present. When this bit is set, it does not imply this feature is supported across all
banks. Software should check the availability of the necessary logic on a bank by
bank basis when using this signaling capability (i.e. bit 30 settable in individual
IA32_MCi_CTL2 register).
• MCG_TES_P (threshold-based error status present) flag, bit 11 —
Indicates (when set) that bits 56:53 of the IA32_MCi_STATUS MSR are part of
the architectural space. Bits 56:55 are reserved, and bits 54:53 are used to
report threshold-based error status. Note that when MCG_TES_P is not set, bits
56:53 of the IA32_MCi_STATUS MSR are model-specific.
• MCG_EXT_CNT, bits 23:16 — Indicates the number of extended machine-
check state registers present. This field is meaningful only when the MCG_EXT_P
flag is set.
• MCG_SER_P (software error recovery support present) flag, bit 24—
Indicates (when set) that the processor supports software error recovery (see
15-4 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
Section 15.6), and IA32_MCi_STATUS MSR bits 56:55 are used to report the
signaling of uncorrected recoverable errors and whether software must take
recovery actions for uncorrected errors. Note that when MCG_TES_P is not set,
bits 56:53 of the IA32_MCi_STATUS MSR are model-specific. If MCG_TES_P is set
but MCG_SER_P is not set, bits 56:55 are reserved.
The effect of writing to the IA32_MCG_CAP MSR is undefined.
15.3.1.2 IA32_MCG_STATUS MSR
The IA32_MCG_STATUS MSR describes the current state of the processor
after a machine-check exception has occurred (see Figure 15-3).
63
3 2
1
0
M
R
I
P
V
E
I
P
V
C
I
Reserved
P
MCIP—Machine check in progress flag
EIPV—Error IP valid flag
RIPV—Restart IP valid flag
Figure 15-3. IA32_MCG_STATUS Register
Where:
• RIPV (restart IP valid) flag, bit 0 — Indicates (when set) that program
execution can be restarted reliably at the instruction pointed to by the instruction
pointer pushed on the stack when the machine-check exception is generated.
When clear, the program cannot be reliably restarted at the pushed instruction
pointer.
• EIPV (error IP valid) flag, bit 1 — Indicates (when set) that the instruction
pointed to by the instruction pointer pushed onto the stack when the machine-
• MCIP (machine check in progress) flag, bit 2 — Indicates (when set) that a
machine-check exception was generated. Software can set or clear this flag. The
occurrence of a second Machine-Check Event while MCIP is set will cause the
processor to enter a shutdown state. For information on processor behavior in
the shutdown state, please refer to the description in Chapter 6, “Interrupt and
Exception Handling”: “Interrupt 8—Double Fault Exception (#DF)”.
Bits 63:03 in IA32_MCG_STATUS are reserved.
Vol. 3 15-5
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
15.3.1.3 IA32_MCG_CTL MSR
The IA32_MCG_CTL MSR is present if the capability flag MCG_CTL_P is set in
the IA32_MCG_CAP MSR.
IA32_MCG_CTL controls the reporting of machine-check exceptions. If
present, writing 1s to this register enables machine-check features and
writing all 0s disables machine-check features. All other values are unde-
fined and/or implementation specific.
15.3.2
Error-Reporting Register Banks
Each error-reporting register bank can contain the IA32_MCi_CTL,
IA32_MCi_STATUS, IA32_MCi_ADDR, and IA32_MCi_MISC MSRs. The
number of reporting banks is indicated by bits [7:0] of IA32_MCG_CAP MSR
(address 0179H). The first error-reporting register (IA32_MC0_CTL) always
starts at address 400H.
See Appendix B, “Model-Specific Registers (MSRs),” for addresses of the
error-reporting registers in the Pentium 4 and Intel Xeon processors; and for
addresses of the error-reporting registers P6 family processors.
The IA32_MCi_CTL MSR controls error reporting for errors produced by a
particular hardware unit (or group of hardware units). Each of the 64 flags
(EEj) represents a potential error. Setting an EEj flag enables reporting of
the associated error and clearing it disables reporting of the error. The
processor does not write changes to bits that are not implemented.
Figure 15-4 shows the bit fields of IA32_MCi_CTL.
63 62 61
3 2
1
0
E
E
6
2
E
E
6
1
E
E
0
2
E
E
0
0
E
E
6
3
E
E
0
1
. . . . .
EEj—Error reporting enable flag
(where j is 00 through 63)
Figure 15-4. IA32_MCi_CTL Register
NOTE
For P6 family processors, processors based on Intel Core microarchi-
tecture (excluding processors with DisplayFamily_DisplayModel
15-6 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
encoding of 06H_1AH and onward): the operating system or
executive software must not modify the contents of the
IA32_MC0_CTL MSR. This MSR is internally aliased to the
EBL_CR_POWERON MSR and controls platform-specific error
handling features. System specific firmware (the BIOS) is responsible
for the appropriate initialization of the IA32_MC0_CTL MSR. P6 family
processors only allow the writing of all 1s or all 0s to the
IA32_MCi_CTL MSR.
15.3.2.2 IA32_MCi_STATUS MSRS
check error if its VAL (valid) flag is set (see Figure 15-5). Software is respon-
sible for clearing IA32_MCi_STATUS MSRs by explicitly writing 0s to them;
writing 1s to them causes a general-protection exception.
NOTE
Figure 15-5 depicts the IA32_MCi_STATUS MSR when
IA32_MCG_CAP[10] = 1. When IA32_MCG_CAP[24] = 0 and
IA32_MCG_CAP[11] = 1, bits 56:55 is reserved and bits 54:53 for
threshold-based error reporting. When IA32_MCG_CAP[11] = 0, bits
56:53 are part of the “Other Information” field. The use of bits 54:53
for threshold-based error reporting began with Intel Core Duo
processors, and is currently used for cache memory. See Section
15.4, “Enhanced Cache Error reporting,” for more information. When
IA32_MCG_CAP[10] = 0, bits 52:38 are part of the “Other Infor-
mation” field. The use of bits 52:38 for corrected MC error count is
Vol. 3 15-7
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
introduced with Intel 64 processor having CPUID
DisplayFamily_DisplayModel encoding of 06H_1AH.
63626160 5958 5756555453 52
38 37
Other
32 31
MSCOD Model
Specific Error Code
16 15
0
O
V
E
R
V
A
L
U
C
E
N
S
P
C
C
A
R
Corrected Error
Count
MCA Error Code
Info
Threshold-based error status (54:53)*
AR — Recovery action required for UCR error (55)**
S — Signaling an uncorrected recoverable (UCR) error (56)**
PCC — Processor context corrupted (57)
ADDRV — MCi_ADDR register valid (58)
MISCV — MCi_MISC register valid (59)
EN — Error reporting enabled (60)
UC — Uncorrected error (61)
OVER — Error overflow (62)
VAL — MCi_STATUS register valid (63)
* When IA32_MCG_CAP[11] (MCG_TES_P) is not set, these bits are model-specific
(part of “Other Information”).
** When IA32_MCG_CAP[11] or IA32_MCG_CAP[24] are not set, these bits are reserved, or
model-specific (part of “Other Information”).
Figure 15-5. IA32_MCi_STATUS Register
Where:
• MCA (machine-check architecture) error code field, bits 15:0 — Specifies
the machine-check architecture-defined error code for the machine-check error
condition detected. The machine-check architecture-defined error codes are
guaranteed to be the same for all IA-32 processors that implement the machine-
check architecture. See Section 15.9, “Interpreting the MCA Error Codes,” and
Appendix E, “Interpreting Machine-Check Error Codes”, for information on
machine-check error codes.
• Model-specific error code field, bits 31:16 — Specifies the model-specific
error code that uniquely identifies the machine-check error condition detected.
The model-specific error codes may differ among IA-32 processors for the same
machine-check error condition. See Appendix E, “Interpreting Machine-Check
Error Codes”for information on model-specific error codes.
• Reserved, Error Status, and Other Information fields, bits 56:32 —
•
Bits 37:32 always contain “Other Information” that is implementation-
specific and is not part of the machine-check architecture. Software that
is intended to be portable among IA-32 processors should not rely on
these values.
15-8 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
•
•
If IA32_MCG_CAP[10] is 0, bits 52:38 also contain “Other Information”
(in the same sense as bits 37:32).
If IA32_MCG_CAP[10] is 1, bits 52:38 are architectural (not model-
specific). In this case, bits 52:38 reports the value of a 15 bit counter that
increments each time a corrected error is observed by the MCA recording
bank. This count value will continue to increment until cleared by
software. The most significant bit, 52, is a sticky count overflow bit.
•
•
If IA32_MCG_CAP[11] is 0, bits 56:53 also contain “Other Information”
(in the same sense).
specific). In this case, bits 56:53 have the following functionality:
•
•
•
If IA32_MCG_CAP[24] is 0, bits 56:55 are reserved.
If IA32_MCG_CAP[24] is 1, bits 56:55 are defined as follows:
MC bank. See Section 15.6.2 for additional detail.
•
AR (Action Required) flag, bit 55 - Indicates (when set) that MCA
error code specific recovery action must be performed by system
software at the time this error was signaled. See Section 15.6.2 for
additional detail.
•
•
If the UC bit (Figure 15-5) is 1, bits 54:53 are undefined.
If the UC bit (Figure 15-5) is 0, bits 54:53 indicate the status of the
hardware structure that reported the threshold-based error. See
Table 15-1.
Table 15-1. Bits 54:53 in IA32_MCi_STATUS MSRs
when IA32_MCG_CAP[11] = 1 and UC = 0
Bits 54:53 Meaning
00
event.
01
Green - Status tracking is provided for the structure posting the event; the current
status is green (below threshold). For more information, see Section 15.4, “Enhanced
Cache Error reporting”.
10
11
Yellow - Status tracking is provided for the structure posting the event; the current
status is yellow (above threshold). For more information, see Section 15.4, “Enhanced
Cache Error reporting”.
Reserved
• PCC (processor context corrupt) flag, bit 57 — Indicates (when set) that the
state of the processor might have been corrupted by the error condition detected
and that reliable restarting of the processor may not be possible. When clear, this
Vol. 3 15-9
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
flag indicates that the error did not affect the processor’s state. Software
restarting might be possible.
• ADDRV (IA32_MCi_ADDR register valid) flag, bit 58 — Indicates (when set)
that the IA32_MCi_ADDR register contains the address where the error occurred
(see Section 15.3.2.3, “IA32_MCi_ADDR MSRs”). When clear, this flag indicates
that the IA32_MCi_ADDR register is either not implemented or does not contain
the address where the error occurred. Do not read these registers if they are not
implemented in the processor.
• MISCV (IA32_MCi_MISC register valid) flag, bit 59 — Indicates (when set)
that the IA32_MCi_MISC register contains additional information regarding the
error. When clear, this flag indicates that the IA32_MCi_MISC register is either
not implemented or does not contain additional information regarding the error.
Do not read these registers if they are not implemented in the processor.
• EN (error enabled) flag, bit 60 — Indicates (when set) that the error was
enabled by the associated EEj bit of the IA32_MCi_CTL register.
• UC (error uncorrected) flag, bit 61 — Indicates (when set) that the processor
did not or was not able to correct the error condition. When clear, this flag
indicates that the processor was able to correct the error condition.
• OVER (machine check overflow) flag, bit 62 — Indicates (when set) that a
machine-check error occurred while the results of a previous error were still in
the error-reporting register bank (that is, the VAL bit was already set in the
IA32_MCi_STATUS register). The processor sets the OVER flag and software is
responsible for clearing it. In general, enabled errors are written over disabled
errors, and uncorrected errors are written over corrected errors. Uncorrected
errors are not written over previous valid uncorrected errors. For more infor-
mation, see Section 15.3.2.2.1, “Overwrite Rules for Machine Check Overflow”.
• VAL (IA32_MCi_STATUS register valid) flag, bit 63 — Indicates (when set)
that the information within the IA32_MCi_STATUS register is valid. When this flag
is set, the processor follows the rules given for the OVER flag in the
IA32_MCi_STATUS register when overwriting previously valid entries. The
processor sets the VAL flag and software is responsible for clearing it.
15.3.2.2.1 Overwrite Rules for Machine Check Overflow
Table 15-2 shows the overwrite rules for how to treat a second event if the
cache has already posted an event to the MC bank – that is, what to do if the
valid bit for an MC bank already is set to 1. When more than one structure
posts events in a given bank, these rules specify whether a new event will
overwrite a previous posting or not. These rules define a priority for uncor-
rected (highest priority), yellow, and green/unmonitored (lowest priority)
status.
15-10 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
In Table 15-2, the values in the two left-most columns are
IA32_MCi_STATUS[54:53].
Table 15-2. Overwrite Rules for Enabled Errors
First Event
00/green
00/green
yellow
Second Event
00/green
yellow
UC bit
Color
MCA Info
second
0
0
0
0
1
1
00/green
yellow
second error
first error
either
00/green
yellow
yellow
yellow
yellow
00/green/yellow UC
UC 00/green/yellow
undefined
undefined
second
first
If a second event overwrites a previously posted event, the information (as
guarded by individual valid bits) in the MCi bank is entirely from the second
event. Similarly, if a first event is retained, all of the information previously
posted for that event is retained. In either case, the OVER bit
(MCi_Status[62]) will be set to indicate an overflow.
After software polls a posting and clears the register, the valid bit is no
longer set and therefore the meaning of the rest of the bits, including the
yellow/green/00 status field in bits 54:53, is undefined. The yellow/green
indication will only be posted for events associated with monitored struc-
tures – otherwise the unmonitored (00) code will be posted in
MCi_Status[54:53].
15.3.2.3 IA32_MCi_ADDR MSRs
The IA32_MCi_ADDR MSR contains the address of the code or data memory
location that produced the machine-check error if the ADDRV flag in the
IA32_MCi_STATUS register is set (see Section 15-6, “IA32_MCi_ADDR
MSR”). The IA32_MCi_ADDR register is either not implemented or contains
no address if the ADDRV flag in the IA32_MCi_STATUS register is clear.
cause a general protection exception.
The address returned is an offset into a segment, linear address, or physical
address. This depends on the error encountered. When these registers are
implemented, these registers can be cleared by explicitly writing 0s to these
registers. Writing 1s to these registers will cause a general-protection
exception. See Figure 15-6.
Vol. 3 15-11
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
Processor Without Support For Intel 64 Architecture
63
0
36 35
Address
Reserved
Processor With Support for Intel 64 Architecture
63
0
*
Address
* Useful bits in this field depend on the address methodology in use when the
the register state is saved.
Figure 15-6. IA32_MCi_ADDR MSR
15.3.2.4 IA32_MCi_MISC MSRs
The IA32_MCi_MISC MSR contains additional information describing the
machine-check error if the MISCV flag in the IA32_MCi_STATUS register is
set. The IA32_MCi_MISC_MSR is either not implemented or does not contain
additional information if the MISCV flag in the IA32_MCi_STATUS register is
clear.
When not implemented in the processor, all reads and writes to this MSR will
cause a general protection exception. When implemented in a processor,
is not implemented in any of the error-reporting register banks for the P6
family processors.
If both MISCV and IA32_MCG_CAP[24] are set, the IA32_MCi_MISC_MSR is
defined according to Figure 15-7 to support software recovery of uncor-
rected errors (see Section 15.6):
15-12 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
9 8 6 5
0
63
Model Specific Information
Address Mode
Recoverable Address LSB
Figure 15-7. UCR Support in IA32_MCi_MISC Register
Indicates the position of the least significant bit (LSB) of the recoverable error
address. For example, if the processor logs bits [43:9] of the address, the LSB
sub-field in IA32_MCi_MISC is 01001b (9 decimal). For this example, bits [8:0]
of the recoverable error address in IA32_MCi_ADDR should be ignored.
• Address Mode (bits 8:6): Address mode for the address logged in
IA32_MCi_ADDR. The supported address modes are given in Table 15-3.
Table 15-3. Address Mode in IA32_MCi_MISC[8:6]
IA32_MCi_MISC[8:6] Encoding
Definition
000
001
Segment Offset
Linear Address
Physical Address
Memory Address
Reserved
010
011
100 to 110
111
Generic
• Model Specific Information (bits 63:9): Not architecturally defined.
15.3.2.5 IA32_MCi_CTL2 MSRs
The IA32_MCi_CTL2 MSR provides the programming interface to use
corrected MC error signaling capability that is indicated by
IA32_MCG_CAP[10] = 1. Software must check for the presence of
IA32_MCi_CTL2 on a per-bank basis.
Vol. 3 15-13
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
When IA32_MCG_CAP[10] = 1, the IA32_MCi_CTL2 MSR for each bank
exists, i.e. reads and writes to these MSR are supported. However, signaling
interface for corrected MC errors may not be supported in all banks.
The layout of IA32_MCi_CTL2 is shown in Figure 15-8:
63
31 30 29
0
15 14
Reserved
Reserved
CMCI_EN—Enable/disable CMCI
Corrected error count threshold
• Corrected error count threshold, bits 14:0 — Software must initialize this
field. The value is compared with the corrected error count field in
IA32_MCi_STATUS, bits 38 through 52. An overflow event is signaled to the CMCI
LVT entry (see Table 10-1) in the APIC when the count value equals the threshold
value. The new LVT entry in the APIC is at 02F0H offset from the APIC_BASE. If
CMCI interface is not supported for a particular bank (but IA32_MCG_CAP[10]
= 1), this field will always read 0.
• CMCI_EN-Corrected error interrupt enable/disable/indicator, bits 30 —
Software sets this bit to enable the generation of corrected machine-check error
interrupt (CMCI). If CMCI interface is not supported for a particular bank (but
IA32_MCG_CAP[10] = 1), this bit is writeable but will always return 0 for that
bank. This bit also indicates CMCI is supported or not supported in the corre-
sponding bank. See Section 15.5 for details of software detection of CMCI facility.
Some microarchitectural sub-systems that are the source of corrected MC
errors may be shared by more than one logical processors. Consequently,
the facilities for reporting MC errors and controlling mechanisms may be
shared by more than one logical processors. For example, the
IA32_MCi_CTL2 MSR is shared between logical processors sharing a
processor core. Software is responsible to program IA32_MCi_CTL2 MSR in
a consistent manner with CMCI delivery and usage.
After processor reset, IA32_MCi_CTL2 MSRs are zero’ed.
15-14 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
The Pentium 4 and Intel Xeon processors implement a variable number of
extended machine-check state MSRs. The MCG_EXT_P flag in the
IA32_MCG_CAP MSR indicates the presence of these extended registers,
and the MCG_EXT_CNT field indicates the number of these registers actually
implemented. See Section 15.3.1.1, “IA32_MCG_CAP MSR.” Also see Table
15-4.
Table 15-4. Extended Machine Check State MSRs
in Processors Without Support for Intel 64 Architecture
MSR
Address
Description
IA32_MCG_EAX
180H
Contains state of the EAX register at the time of the machine-
check error.
IA32_MCG_EBX
IA32_MCG_ECX
IA32_MCG_EDX
IA32_MCG_ESI
IA32_MCG_EDI
IA32_MCG_EBP
IA32_MCG_ESP
IA32_MCG_EFLAGS
IA32_MCG_EIP
IA32_MCG_MISC
181H
182H
183H
184H
185H
186H
187H
188H
189H
18AH
Contains state of the EBX register at the time of the machine-
check error.
Contains state of the ECX register at the time of the machine-
check error.
Contains state of the EDX register at the time of the machine-
check error.
Contains state of the ESI register at the time of the machine-
check error.
Contains state of the EDI register at the time of the machine-
check error.
Contains state of the EBP register at the time of the machine-
check error.
Contains state of the ESP register at the time of the machine-
check error.
Contains state of the EFLAGS register at the time of the
machine-check error.
Contains state of the EIP register at the time of the machine-
check error.
When set, indicates that a page assist or page fault occurred
during DS normal operation.
In processors with support for Intel 64 architecture, 64-bit machine check
state MSRs are aliased to the legacy MSRs. In addition, there may be regis-
ters beyond IA32_MCG_MISC. These may include up to five reserved MSRs
(IA32_MCG_RESERVED[1:5]) and save-state MSRs for registers introduced
in 64-bit mode. See Table 15-5.
Vol. 3 15-15
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
Table 15-5. Extended Machine Check State MSRs
In Processors With Support For Intel 64 Architecture
MSR
Address
Description
IA32_MCG_RAX
180H
Contains state of the RAX register at the time of the machine-
check error.
IA32_MCG_RBX
IA32_MCG_RCX
IA32_MCG_RDX
IA32_MCG_RSI
IA32_MCG_RDI
IA32_MCG_RBP
IA32_MCG_RSP
181H
182H
183H
184H
185H
186H
187H
Contains state of the RBX register at the time of the machine-
check error.
Contains state of the RCX register at the time of the machine-
check error.
Contains state of the RDX register at the time of the machine-
check error.
Contains state of the RSI register at the time of the machine-
check error.
Contains state of the RDI register at the time of the machine-
check error.
Contains state of the RBP register at the time of the machine-
check error.
Contains state of the RSP register at the time of the machine-
check error.
IA32_MCG_RFLAGS 188H
Contains state of the RFLAGS register at the time of the
machine-check error.
IA32_MCG_RIP
IA32_MCG_MISC
189H
18AH
Contains state of the RIP register at the time of the machine-
check error.
When set, indicates that a page assist or page fault occurred
during DS normal operation.
IA32_MCG_
RSERVED[1:5]
18BH-
18FH
These registers, if present, are reserved.
IA32_MCG_R8
IA32_MCG_R9
IA32_MCG_R10
IA32_MCG_R11
IA32_MCG_R12
IA32_MCG_R13
190H
191H
192H
193H
194H
195H
Contains state of the R8 register at the time of the machine-
check error.
Contains state of the R9 register at the time of the machine-
check error.
Contains state of the R10 register at the time of the machine-
check error.
Contains state of the R11 register at the time of the machine-
check error.
Contains state of the R12 register at the time of the machine-
check error.
Contains state of the R13 register at the time of the machine-
check error.
15-16 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
Table 15-5. Extended Machine Check State MSRs
In Processors With Support For Intel 64 Architecture (Contd.)
MSR
Address
Description
IA32_MCG_R14
196H
Contains state of the R14 register at the time of the machine-
check error.
IA32_MCG_R15
197H
Contains state of the R15 register at the time of the machine-
check error.
When a machine-check error is detected on a Pentium 4 or Intel Xeon
processor, the processor saves the state of the general-purpose registers,
the R/EFLAGS register, and the R/EIP in these extended machine-check
state MSRs. This information can be used by a debugger to analyze the error.
These registers are read/write to zero registers. This means software can
read them; but if software writes to them, only all zeros is allowed. If soft-
ware attempts to write a non-zero value into one of these registers, a
general-protection (#GP) exception is generated. These registers are
cleared on a hardware reset (power-up or RESET), but maintain their
contents following a soft reset (INIT reset).
15.3.3
Mapping of the Pentium Processor Machine-Check Errors
to the Machine-Check Architecture
The Pentium processor reports machine-check errors using two registers:
P5_MC_TYPE and P5_MC_ADDR. The Pentium 4, Intel Xeon, and P6 family
processors map these registers to the IA32_MCi_STATUS and
IA32_MCi_ADDR in the error-reporting register bank. This bank reports on
the same type of external bus errors reported in P5_MC_TYPE and
P5_MC_ADDR.
The information in these registers can then be accessed in two ways:
• By reading the IA32_MCi_STATUS and IA32_MCi_ADDR registers as part of a
general machine-check exception handler written for Pentium 4 and P6 family
processors.
• By reading the P5_MC_TYPE and P5_MC_ADDR registers using the RDMSR
instruction.
The second capability permits a machine-check exception handler written to
run on a Pentium processor to be run on a Pentium 4, Intel Xeon, or P6
family processor. There is a limitation in that information returned by the
Pentium 4, Intel Xeon, and P6 family processors is encoded differently than
information returned by the Pentium processor. To run a Pentium processor
machine-check exception handler on a Pentium 4, Intel Xeon, or P6 family
Vol. 3 15-17
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
processor; the handler must be written to interpret P5_MC_TYPE encodings
correctly.
15.4
ENHANCED CACHE ERROR REPORTING
Starting with Intel Core Duo processors, cache error reporting was
enhanced. In earlier Intel processors, cache status was based on the
number of correction events that occurred in a cache. In the new paradigm,
called “threshold-based error status”, cache status is based on the number
of lines (ECC blocks) in a cache that incur repeated corrections. The
threshold is chosen by Intel, based on various factors. If a processor
supports threshold-based error status, it sets IA32_MCG_CAP[11]
(MCG_TES_P) to 1; if not, to 0.
A processor that supports enhanced cache error reporting contains hard-
ware that tracks the operating status of certain caches and provides an indi-
cator of their “health”. The hardware reports a “green” status when the
number of lines that incur repeated corrections is at or below a pre-defined
threshold, and a “yellow” status when the number of affected lines exceeds
the threshold. Yellow status means that the cache reporting the event is
operating correctly, but you should schedule the system for servicing within
a few weeks.
Intel recommends that you rely on this mechanism for structures supported
by threshold-base error reporting.
The CPU/system/platform response to a yellow event should be less severe
than its response to an uncorrected error. An uncorrected error means that
a serious error has actually occurred, whereas the yellow condition is a
warning that the number of affected lines has exceeded the threshold but is
not, in itself, a serious event: the error was corrected and system state was
not compromised.
The green/yellow status indicator is not a foolproof early warning for an
uncorrected error resulting from the failure of two bits in the same ECC
block. Such a failure can occur and cause an uncorrected error before the
yellow threshold is reached. However, the chance of an uncorrected error
increases as the number of affected lines increases.
15.5
CORRECTED MACHINE CHECK ERROR INTERRUPT
Corrected machine-check error interrupt (CMCI) is an architectural
enhancement to the machine-check architecture. It provides capabilities
15-18 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
beyond those of threshold-based error reporting (Section 15.4). With
threshold-based error reporting, software is limited to use periodic polling to
query the status of hardware corrected MC errors. CMCI provides a signaling
mechanism to deliver a local interrupt based on threshold values that soft-
ware can program using the IA32_MCi_CTL2 MSRs.
CMCI is disabled by default. System software is required to enable CMCI for
each IA32_MCi bank that support the reporting of hardware corrected errors
if IA32_MCG_CAP[10] = 1.
System software use IA32_MCi_CTL2 MSR to enable/disable the CMCI capa-
bility for each bank and program threshold values into IA32_MCi_CTL2 MSR.
CMCI is not affected by the CR4.MCE bit, and it is not affected by the
IA32_MCi_CTL MSR’s.
To detect the existence of thresholding for a given bank, software writes only
bits 14:0 with the threshold value. If the bits persist, then thresholding is
available (and CMCI is available). If the bits are all 0's, then no thresholding
exists. To detect that CMCI signaling exists, software writes a 1 to bit 30 of
the MCi_CTL2 register. Upon subsequent read, If Bit 30 = 0, no CMCI is
available for this bank. If Bit 30 = 1, then CMCI is available and enabled.
15.5.1
CMCI Local APIC Interface
The interaction of CMCI is depicted in Figure 15-9.
Software write 1 to enable
0
63
31 30 29
14
MCi_CTL2
Error threshold
Count overflow threshold -> CMCI LVT in local APIC
?=
APIC_BASE + 2F0H
53 52
38 37
0
MCi_STATUS
Error count
Figure 15-9. CMCI Behavior
Vol. 3 15-19
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
CMCI interrupt delivery is configured by writing to the LVT CMCI register
entry in the local APIC register space at default address of APIC_BASE +
2F0H. A CMCI interrupt can be delivered to more than one logical processors
if multiple logical processors are affected by the associated MC errors. For
example, if a corrected bit error in a cache shared by two logical processors
caused a CMCI, the interrupt will be delivered to both logical processors
sharing that microarchitectural sub-system. Similarly, package level errors
may cause CMCI to be delivered to all logical processors within the package.
However, system level errors will not be handled by CMCI.
The format of the LVT CMCI register is shown in Figure 15-10. The LVT entry
allows the 4 delivery modes, an 8 bit interrupt vector, and masking.
13
12
11 10
31
17 16 15
8
7
0
Reserved
Reserved
APIC_BASE + 2F0H
MASK
Delivery Status
Delivery Mode
Vector Number
Figure 15-10. Local APIC CMCI LVT Register
• Vector, bits 7:0 — The interrupt vector number. The local APIC allows values
16-255 in this register. Values of 0 through 15 will result in an illegal vector to be
logged in the APIC error status register.
• Delivery mode, bits 10:8 — The following delivery modes are supported:
— 000B: Fixed delivery. Delivers an interrupt to the vector specified in bits 7:0.
— 010B: SMI Delivers an SMI interrupt to the processor core through the
processor's local SMI signal path. When using this delivery mode, the vector
field should be set to 00H for future compatibility.
— 100B: NMI Delivers an NMI interrupt to the logical processor affected by the
error. The vector information is ignored.
— Entry 101B (INIT) and entry 111B (ExtINT) are not supported by CMCI LVT.
— Other bit patterns are reserved.
15-20 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
• Delivery status, bits 12 — It is a read-only bit that, when set, indicates that an
interrupt from this source has been delivered to the processor core, but has not
yet been accepted.
• Mask, bits 16 — When set, inhibits reception of the interrupt. (Unlike the
PerfMon LVT entry, this bit is not set when an interrupt is received. When clear,
CMCI is not masked. The mask bit is set by default.
• Bits 31:17, 15:13 and 11 are reserved.
15.5.2
System Software Recommendation for Managing CMCI and
Machine Check Resources
System software must enable and manage CMCI, set up interrupt handlers
to service CMCI interrupts delivered to affected logical processors, program
CMCI LVT entry, and query machine check banks that are shared by more
than one logical processors.
This section describes techniques system software can implement to
manage CMCI initialization, service CMCI interrupts in a efficient manner to
minimize contentions to access shared MSR resources.
15.5.2.1 CMCI Initialization
Although a CMCI interrupt may be delivered to more than one logical proces-
sors depending on the nature of the corrected MC error, only one instance of
the interrupt service routine needs to perform the necessary service and
make queries to the machine-check banks. The following steps describes a
technique that limits the amount of work the system has to do in response to
a CMCI.
• To provide maximum flexibility, system software should define per-thread data
structure for each logical processor to allow equal-opportunity and efficient
response to interrupt delivery. Specifically, the per-thread data structure should
include a set of per-bank fields to track which machine check bank it needs to
access in response to a delivered CMCI interrupt. The number of banks that
needs to be tracked is determined by IA32_MCG_CAP[7:0].
• Initialization of per-thread data structure. The initialization of per-thread data
structure must be done serially on each logical processor in the system. The
sequencing order to start the per-thread initialization between different logical
processor is arbitrary. But it must observe the following specific detail to satisfy
the shared nature of specific MSR resources:
a. Each thread initializes its data structure to indicate that it does not own any
MC bank registers.
Vol. 3 15-21
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
b. Each thread examines IA32_MCi_CTL2[30] indicator for each bank to
determine if another thread has already claimed ownership of that bank.
•
If IA32_MCi_CTL2[30] had been set by another thread. This thread can
not own bank i and should proceed to step b. and examine the next
machine check bank until all of the machine check banks are exhausted.
•
If IA32_MCi_CTL2[30] = 0, proceed to step c.
c. Check whether writing a 1 into IA32_MCi_CTL2[30] can return with 1 on a
subsequent read to determine this bank can support CMCI.
•
can not own bank i and should proceed to step b. and examine the next
machine check bank until all of the machine check banks are exhausted.
•
If IA32_MCi_CTL2[30] = 1, modify the per-thread data structure to
indicate this thread claims ownership to the MC bank; proceed to initialize
the error threshold count (bits 15:0) of that bank as described in Chapter
the next machine check bank until all of the machine check banks are
exhausted.
• After the thread has examined all of the machine check banks, it sees if it owns
any MC banks to service CMCI. If any bank has been claimed by this thread:
— Ensure that the CMCI interrupt handler has been set up as described in
Chapter 15, “CMCI Interrupt Handler”.
— Initialize the CMCI LVT entry, as described in Chapter 15, “CMCI Local APIC
Interface”.
— Log and clear all of IA32_MCi_Status registers for the banks that this thread
owns. This will allow new errors to be logged.
15.5.2.2 CMCI Threshold Management
The Corrected MC error threshold field, IA32_MCi_CTL2[15:0], is architec-
turally defined. Specifically, all these bits are writable by software, but
different processor implementations may choose to implement less than 15
bits as threshold for the overflow comparison with
IA32_MCi_STATUS[52:38]. The following describes techniques that soft-
ware can manage CMCI threshold to be compatible with changes in imple-
mentation characteristics:
• Software can set the initial threshold value to 1 by writing 1 to
IA32_MCi_CTL2[15:0]. This will cause overflow condition on every
corrected MC error and generates a CMCI interrupt.
• To increase the threshold and reduce the frequency of CMCI servicing:
a. Find the maximum threshold value a given processor implementation
supports. The steps are:
15-22 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
•
•
Write 7FFFH to IA32_MCi_CTL2[15:0],
Read back IA32_MCi_CTL2[15:0], the lower 15 bits (14:0) is the
maximum threshold supported by the processor.
b. Increase the threshold to a value below the maximum value discovered using
step a.
15.5.2.3 CMCI Interrupt Handler
The following describes techniques system software may consider to imple-
ment a CMCI service routine:
• The service routine examines its private per-thread data structure to check which
set of MC banks it has ownership. If the thread does not have ownership of a
given MC bank, proceed to the next MC bank. Ownership is determined at initial-
ization time which is described in Section [Cross Reference to 14.5.2.1].
• If the thread had claimed ownership to an MC bank,
— Check for valid MC errors by testing IA32_MCi_STATUS.VALID[63],
•
•
Log MC errors,
Clear the MSRs of this MC bank.
— If no valid error, proceed to next MC bank.
• When all MC banks have been processed, exit service routine and return to
original program execution.
This technique will allow each logical processors to handle corrected MC
errors independently and requires no synchronization to access shared MSR
resources.
15.6
RECOVERY OF UNCORRECTED RECOVERABLE (UCR)
ERRORS
Recovery of uncorrected recoverable machine check errors is an enhance-
ment in machine-check architecture. The first processor that supports this
feature is 45nm Intel 64 processor with CPUID signature
DisplayFamily_DisplayModel encoding of 06H_2EH. This allow system soft-
ware to perform recovery action on certain class of uncorrected errors and
continue execution.
Vol. 3 15-23
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
15.6.1
Detection of Software Error Recovery Support
Software must use bit 24 of IA32_MCG_CAP (MCG_SER_P) to detect the
presence of software error recovery support (see Figure 15-2). When
IA32_MCG_CAP[24] is set, this indicates that the processor supports soft-
ware error recovery. When this bit is clear, this indicates that there is no
support for error recovery from the processor and the primary responsibility
of the machine check handler is logging the machine check error information
and shutting down the system.
The new class of architectural MCA errors from which system software can
attempt recovery is called Uncorrected Recoverable (UCR) Errors. UCR
errors are uncorrected errors that have been detected and signaled but have
not corrupted the processor context. For certain UCR errors, this means that
once system software has performed a certain recovery action, it is possible
to continue execution on this processor. UCR error reporting provides an
error containment mechanism for data poisoning. The machine check
handler will use the error log information from the error reporting registers
to analyze and implement specific error recovery actions for UCR errors.
15.6.2
UCR Error Reporting and Logging
IA32_MCi_STATUS MSR is used for reporting UCR errors and existing
corrected or uncorrected errors. The definitions of IA32_MCi_STATUS,
including bit fields to identify UCR errors, is shown in Figure 15-5. UCR
errors can be signaled through either the corrected machine check interrupt
(CMCI) or machine check exception (MCE) path depending on the type of the
UCR error.
When IA32_MCG_CAP[24] is set, a UCR error is indicated by the following
bit settings in the IA32_MCi_STATUS register:
• Valid (bit 63) = 1
• UC (bit 61) = 1
• PCC (bit 57) = 0
Additional information from the IA32_MCi_MISC and the IA32_MCi_ADDR
MCA error code field of the IA32_MCi_STATUS register indicates the type of
UCR error. System software can interpret the MCA error code field to analyze
and identify the necessary recovery action for the given UCR error.
In addition, the IA32_MCi_STATUS register bit fields, bits 56:55, are defined
(see Figure 15-5) to provide additional information to help system software
to properly identify the necessary recovery action for the UCR error:
15-24 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
• S (Signaling) flag, bit 56 - Indicates (when set) that a machine check exception
was generated for the UCR error reported in this MC bank and system software
needs to check the AR flag and the MCA error code fields in the
IA32_MCi_STATUS register to identify the necessary recovery action for this
error. When the S flag in the IA32_MCi_STATUS register is clear, this UCR error
was not signaled via a machine check exception and instead was reported as a
corrected machine check (CMC). System software is not required to take any
recovery action when the S flag in the IA32_MCi_STATUS register is clear.
• AR (Action Required) flag, bit 55 - Indicates (when set) that MCA error code
specific recovery action must be performed by system software at the time this
error was signaled. This recovery action must be completed successfully before
any additional work is scheduled for this processor When the RIPV flag in the
IA32_MCG_STATUS is clear, an alternative execution stream needs to be
provided; when the MCA error code specific recovery specific recovery action
cannot be successfully completed, system software must shut down the system.
When the AR flag in the IA32_MCi_STATUS register is clear, system software may
still take MCA error code specific recovery action but this is optional; system
software can safely resume program execution at the instruction pointer saved
on the stack from the machine check exception when the RIPV flag in the
IA32_MCG_STATUS register is set.
Both the S and the AR flags in the IA32_MCi_STATUS register are defined to
be sticky bits, which mean that once set, the processor does not clear them.
Only software and good power-on reset can clear the S and the AR-flags.
Both the S and the AR flags are only set when the processor reports the UCR
errors (MCG_CAP[24] is set).
15.6.3
UCR Error Classification
With the S and AR flag encoding in the IA32_MCi_STATUS register, UCR
errors can be classified as:
• Uncorrected no action required (UCNA) - is a UCR error that is not signaled via a
machine check exception and, instead, is reported to system software as a
corrected machine check error. UCNA errors indicate that some data in the
system is corrupted, but the data has not been consumed and the processor
state is valid and you may continue execution on this processor. UCNA errors
require no action from system software to continue execution. A UNCA error is
indicated with UC=1, PCC=0, S=0 and AR=0 in the IA32_MCi_STATUS register.
• software recoverable action optional (SRAO) - a UCR error is signaled via a
machine check exception and a system software recovery action is optional and
not required to continue execution from this machine check exception. SRAO
errors indicate that some data in the system is corrupt, but the data has not been
consumed and the processor state is valid. SRAO errors provide the additional
error information for system software to perform a recovery action. An SRAO
error is indicated with UC=1, PCC=0, S=1, EN=1 and AR=0 in the
Vol. 3 15-25
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
IA32_MCi_STATUS register. Recovery actions for SRAO errors are MCA error code
specific. The MISCV and the ADDRV flags in the IA32_MCi_STATUS register are
set when the additional error information is available from the IA32_MCi_MISC
and the IA32_MCi_ADDR registers. System software needs to inspect the MCA
error code fields in the IA32_MCi_STATUS register to identify the specific
recovery action for a given SRAO error. If MISCV and ADDRV are not set, it is
recommended that no system software error recovery be performed however,
you can resume execution.
• software recoverable action required (SRAR) - a UCR error that requires system
software to take a recovery action on this processor before scheduling another
stream of execution on this processor. SRAR errors indicate that the error was
detected and raised at the point of the consumption in the execution flow. An
SRAR error is indicated with UC=1, PCC=0, S=1, EN=1 and AR=1 in the
IA32_MCi_STATUS register. Recovery actions are MCA error code specific. The
MISCV and the ADDRV flags in the IA32_MCi_STATUS register are set when the
additional error information is available from the IA32_MCi_MISC and the
IA32_MCi_ADDR registers. System software needs to inspect the MCA error code
fields in the IA32_MCi_STATUS register to identify the specific recovery action for
a given SRAR error. If MISCV and ADDRV are not set, it is recommended that
system software shutdown the system.
Table 15-6 summarizes UCR, corrected, and uncorrected errors.
Table 15-6. MC Error Classifications
1
Type of Error
UC PCC
S
AR Signaling Software Action
Example
Uncorrected Error
(UC)
1
1
x
x
MCE
Reset the system
SRAR
1
0
1
1
1
MCE
For known MCACOD, take Cache to
specific recovery action; processor load
error
For unknown MCACOD,
must bugcheck
SRAO
1
0
0
MCE
For known MCACOD, take Patrol scrub and
specific recovery action; explicit writeback
poison errors
For unknown MCACOD,
OK to keep the system
running
UCNA
1
0
0
0
0
x
0
x
CMC
CMC
Log the error and Ok to Poison detection
keep the system running error
Corrected Error (CE)
Log the error and no
corrective action
required
ECC in caches and
memory
NOTES:
1. VAL=1, EN=1 for UC=1 errors; OVER=0 for UC=1 and PCC=0 errors SRAR, SRAO and UCNA errors
are supported by the processor only when IA32_MCG_CAP[24] (MCG_SER_P) is set.
15-26 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
15.6.4
UCR Error Overwrite Rules
In general, the overwrite rules are as follows:
• UCR errors will overwrite corrected errors.
• Uncorrected (PCC=1) errors overwrite UCR (PCC=0) errors.
• UCR errors are not written over previous UCR errors.
• Corrected errors do not write over previous UCR errors.
Regardless of whether the 1st error is retained or the 2nd error is over-
written over the 1st error, the OVER flag in the IA32_MCi_STATUS register
will be set to indicate an overflow condition. As the S flag and AR flag in the
IA32_MCi_STATUS register are defined to be sticky flags, a second event
cannot clear these 2 flags once set, however the MC bank information may
be filled in for the 2nd error. The table below shows the overwrite rules and
along with the resulting bit setting of the UC, PCC, and AR flags in the
IA32_MCi_STATUS register. As UCNA and SRA0 errors do not require
recovery action from system software to continue program execution, a
system reset by system software is not required unless the AR flag or PCC
flag is set for the UCR overflow case (OVER=1, VAL=1, UC=1, PCC=0).
Table 15-7 lists overwrite rules for uncorrected errors, corrected errors, and
UCR errors.
Table 15-7. Overwrite Rules for UC, CE, and UCR Errors
First Event Second Event UC PCC S
AR
MCABank Reset System
CE
UCR
1
0
0 if UCNA,
else 1
1 if SRAR,
else 0
second
yes, if AR=1
UCR
CE
1
0
0 if UCNA,
else 1
1 if SRAR,
else 0
first
yes, if AR=1
UCNA
UCNA
UCNA
SRAO
SRAO
SRAO
SRAR
SRAR
SRAR
UCR
UCNA
SRAO
SRAR
UCNA
SRAO
SRAR
UCNA
SRAO
SRAR
UC
1
1
1
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
1
1
0
0
first
first
first
first
first
first
first
first
first
second
first
no
1
0
no
1
1
yes
no
1
0
1
0
no
1
1
yes
yes
yes
yes
yes
yes
1
1
1
1
1
1
undefined
undefined
undefined
undefined
UC
UCR
Vol. 3 15-27
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
15.7
MACHINE-CHECK AVAILABILITY
The machine-check architecture and machine-check exception (#MC) are
model-specific features. Software can execute the CPUID instruction to
determine whether a processor implements these features. Following the
execution of the CPUID instruction, the settings of the MCA flag (bit 14) and
MCE flag (bit 7) in EDX indicate whether the processor implements the
machine-check architecture and machine-check exception.
15.8
MACHINE-CHECK INITIALIZATION
To use the processors machine-check architecture, software must initialize
the processor to activate the machine-check exception and the error-
reporting mechanism.
Example 15-1 gives pseudocode for performing this initialization. This
pseudocode checks for the existence of the machine-check architecture and
exception; it then enables machine-check exception and the error-reporting
register banks. The pseudocode shown is compatible with the Pentium 4,
Intel Xeon, P6 family, and Pentium processors.
Following power up or power cycling, IA32_MCi_STATUS registers are not
guaranteed to have valid data until after they are initially cleared to zero by
software (as shown in the initialization pseudocode in Example 15-1). In
addition, when using P6 family processors, software must set MCi_STATUS
registers to zero when doing a soft-reset.
Example 15-1. Machine-Check Initialization Pseudocode
Check CPUID Feature Flags for MCE and MCA support
IF CPU supports MCE
THEN
IF CPU supports MCA
THEN
IF (IA32_MCG_CAP.MCG_CTL_P = 1)
(* IA32_MCG_CTL register is present *)
THEN
IA32_MCG_CTL ← FFFFFFFFFFFFFFFFH;
15-28 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
(* enables all MCA features *)
FI
(* Determine number of error-reporting banks supported *)
COUNT← IA32_MCG_CAP.Count;
MAX_BANK_NUMBER ← COUNT - 1;
IF (Processor Family is 6H and Processor EXTMODEL:MODEL is less than 1AH)
THEN
(* Enable logging of all errors except for MC0_CTL register *)
FOR error-reporting banks (1 through MAX_BANK_NUMBER)
DO
IA32_MCi_CTL ← 0FFFFFFFFFFFFFFFFH;
OD
ELSE
(* Enable logging of all errors including MC0_CTL register *)
FOR error-reporting banks (0 through MAX_BANK_NUMBER)
DO
IA32_MCi_CTL ← 0FFFFFFFFFFFFFFFFH;
OD
FI
(* BIOS clears all errors only on power-on reset *)
IF (BIOS detects Power-on reset)
THEN
FOR error-reporting banks (0 through MAX_BANK_NUMBER)
DO
IA32_MCi_STATUS ← 0;
OD
ELSE
FOR error-reporting banks (0 through MAX_BANK_NUMBER)
DO
(Optional for BIOS and OS) Log valid errors
(OS only) IA32_MCi_STATUS ← 0;
OD
FI
FI
Setup the Machine Check Exception (#MC) handler for vector 18 in IDT
Set the MCE bit (bit 6) in CR4 register to enable Machine-Check Exceptions
FI
15.9
INTERPRETING THE MCA ERROR CODES
When the processor detects a machine-check error condition, it writes a 16-
bit error code to the MCA error code field of one of the IA32_MCi_STATUS
registers and sets the VAL (valid) flag in that register. The processor may
Vol. 3 15-29
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
also write a 16-bit model-specific error code in the IA32_MCi_STATUS
register depending on the implementation of the machine-check architec-
ture of the processor.
The MCA error codes are architecturally defined for Intel 64 and IA-32
processors. To determine the cause of a machine-check exception, the
machine-check exception handler must read the VAL flag for each
IA32_MCi_STATUS register. If the flag is set, the machine check-exception
handler must then read the MCA error code field of the register. It is the
encoding of the MCA error code field [15:0] that determines the type of error
being reported and not the register bank reporting it.
error codes.
15.9.1
Simple Error Codes
Table 15-8 shows the simple error codes. These unique codes indicate global
error information.
Table 15-8. IA32_MCi_Status [15:0] Simple Error Code Encoding
Error Code
Binary Encoding
Meaning
No Error
0000 0000 0000 0000
No error has been reported to this bank of
error-reporting registers.
Unclassified
0000 0000 0000 0001
This error has not been classified into the
MCA error classes.
Microcode ROM Parity 0000 0000 0000 0010
Error
Parity error in internal microcode ROM
External Error
0000 0000 0000 0011
The BINIT# from another processor caused
this processor to enter machine check.
1
FRC Error
0000 0000 0000 0100
FRC (functional redundancy check)
master/slave error
Internal Parity Error
Internal Timer Error
Internal Unclassified
NOTES:
0000 0000 0000 0101
0000 0100 0000 0000
0000 01xx xxxx xxxx
Internal parity error.
Internal timer error.
2
Internal unclassified errors.
1. BINIT# assertion will cause a machine check exception if the processor (or any processor on the
same external bus) has BINIT# observation enabled during power-on configuration (hardware
strapping) and if machine check exceptions are enabled (by setting CR4.MCE = 1).
2. At least one X must equal one. Internal unclassified errors have not been classified.
15-30 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
15.9.2
Compound Error Codes
Compound error codes describe errors related to the TLBs, memory, caches,
bus and interconnect logic, and internal timer. A set of sub-fields is common
to all of compound errors. These sub-fields describe the type of access, level
in the cache hierarchy, and type of request. Table 15-9 shows the general
form of the compound error codes.
Table 15-9. IA32_MCi_Status [15:0] Compound Error Code Encoding
Type
Form
Interpretation
Generic Cache Hierarchy
TLB Errors
000F 0000 0000 11LL Generic cache hierarchy error
000F 0000 0001 TTLL {TT}TLB{LL}_ERR
Memory Controller Errors
Cache Hierarchy Errors
000F 0000 1MMM CCCC {MMM}_CHANNEL{CCCC}_ERR
000F 0001 RRRR TTLL {TT}CACHE{LL}_{RRRR}_ERR
Bus and Interconnect Errors 000F 1PPT RRRR IILL
BUS{LL}_{PP}_{RRRR}_{II}_{T}_ERR
The “Interpretation” column in the table indicates the name of a compound
ICACHEL1_RD_ERR is constructed from the form:
{TT}CACHE{LL}_{RRRR}_ERR,
where {TT} is replaced by I, {LL} is replaced by L1, and {RRRR} is replaced by RD.
For more information on the “Form” and “Interpretation” columns, see
Sections Section 15.9.2.1, “Correction Report Filtering (F) Bit” through
Section 15.9.2.5, “Bus and Interconnect Errors”.
15.9.2.1 Correction Report Filtering (F) Bit
Starting with Intel Core Duo processors, bit 12 in the “Form” column in Table
15-9 is used to indicate that a particular posting to a log may be the last
posting for corrections in that line/entry, at least for some time:
meaning).
• 1 in bit 12 indicates “corrected” filtering (filtering is activated for the line/entry in
the posting). Filtering means that some or all of the subsequent corrections to
this entry (in this structure) will not be posted. The enhanced error reporting
introduced with the Intel Core Duo processors is based on tracking the lines
affected by repeated corrections (see Section 15.4, “Enhanced Cache Error
reporting”). This capability is indicated by IA32_MCG_CAP[11]. Only the first few
correction events for a line are posted; subsequent redundant correction events
to the same line are not posted. Uncorrected events are always posted.
Vol. 3 15-31
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
The behavior of error filtering after crossing the yellow threshold is model-
specific.
15.9.2.2 Transaction Type (TT) Sub-Field
The 2-bit TT sub-field (Table 15-10) indicates the type of transaction (data,
instruction, or generic). The sub-field applies to the TLB, cache, and inter-
connect error conditions. Note that interconnect error conditions are prima-
rily associated with P6 family and Pentium processors, which utilize an
external APIC bus separate from the system bus. The generic type is
reported when the processor cannot determine the transaction type.
Table 15-10. Encoding for TT (Transaction Type) Sub-Field
Transaction Type
Mnemonic
Binary Encoding
Instruction
I
00
01
10
Data
D
G
Generic
15.9.2.3 Level (LL) Sub-Field
The 2-bit LL sub-field (see Table 15-11) indicates the level in the memory
hierarchy where the error occurred (level 0, level 1, level 2, or generic). The
LL sub-field also applies to the TLB, cache, and interconnect error condi-
tions. The Pentium 4, Intel Xeon, and P6 family processors support two
levels in the cache hierarchy and one level in the TLBs. Again, the generic
type is reported when the processor cannot determine the hierarchy level.
Table 15-11. Level Encoding for LL (Memory Hierarchy Level) Sub-Field
Hierarchy Level
Mnemonic
Binary Encoding
Level 0
Level 1
Level 2
Generic
L0
L1
L2
LG
00
01
10
11
15.9.2.4 Request (RRRR) Sub-Field
The 4-bit RRRR sub-field (see Table 15-12) indicates the type of action asso-
ciated with the error. Actions include read and write operations, prefetches,
cache evictions, and snoops. Generic error is returned when the type of error
cannot be determined. Generic read and generic write are returned when
the processor cannot determine the type of instruction or data request that
15-32 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
caused the error. Eviction and snoop requests apply only to the caches. All of
the other requests apply to TLBs, caches and interconnects.
Table 15-12. Encoding of Request (RRRR) Sub-Field
Request Type
Generic Error
Generic Read
Generic Write
Data Read
Mnemonic
Binary Encoding
ERR
0000
RD
0001
WR
0010
DRD
0011
Data Write
Instruction Fetch
Prefetch
DWR
0100
IRD
0101
PREFETCH
EVICT
SNOOP
0110
Eviction
0111
Snoop
1000
15.9.2.5 Bus and Interconnect Errors
The bus and interconnect errors are defined with the 2-bit PP (participation),
1-bit T (time-out), and 2-bit II (memory or I/O) sub-fields, in addition to the
LL and RRRR sub-fields (see Table 15-13). The bus error conditions are
implementation dependent and related to the type of bus implemented by
the processor. Likewise, the interconnect error conditions are predicated on
a specific implementation-dependent interconnect model that describes the
connections between the different levels of the storage hierarchy. The type
of bus is implementation dependent, and as such is not specified in this
document. A bus or interconnect transaction consists of a request involving
an address and a response.
Table 15-13. Encodings of PP, T, and II Sub-Fields
Sub-Field
Transaction
Mnemonic
Binary Encoding
PP (Participation)
Local processor* originated request
SRC
00
01
10
Local processor* responded to request RES
Local processor* observed error as
third party
OBS
Generic
11
1
T (Time-out)
Request timed out
Request did not time out
Memory Access
Reserved
TIMEOUT
NOTIMEOUT
M
0
II (Memory or I/O)
00
01
Vol. 3 15-33
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
Table 15-13. Encodings of PP, T, and II Sub-Fields (Contd.)
I/O
IO
10
Other transaction
11
NOTE:
* Local processor differentiates the processor reporting the error from other system components
(including the APIC, other processors, etc.).
15.9.2.6 Memory Controller Errors
The memory controller errors are defined with the 3-bit MMM (memory
transaction type), and 4-bit CCCC (channel) sub-fields. The encodings for
MMM and CCCC are defined in Table 15-14.
Table 15-14. Encodings of MMM and CCCC Sub-Fields
Sub-Field
Transaction
Mnemonic
GEN
RD
Binary Encoding
MMM
Generic undefined request
Memory read error
Memory write error
Address/Command Error
Memory Scrubbing Error
Reserved
000
001
WR
010
AC
011
MS
100
101-111
0000-1110
1111
CCCC
Channel number
CHN
Channel not specified
15.9.3
Architecturally Defined UCR Errors
Software recoverable compound error code are defined in this section.
15.9.3.1 Architecturally Defined SRAO Errors
The following two SRAO errors are architecturally defined.
• UCR Errors detected by memory controller scrubbing; and
• UCR Errors detected during L3 cache (L3) explicit writebacks.
The MCA error code encodings for these two architecturally-defined UCR
errors corresponds to sub-classes of compound MCA error codes (see Table
15-34 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
15-9). Their values and compound encoding format are given in Table
15-15.
Table 15-15. MCA Compound Error Code Encoding for SRAO Errors
1
Type
MCACOD Value MCA Error Code Encoding
Memory Scrubbing
0xC0 - 0xCF
0000_0000_1100_CCCC
000F 0000 1MMM CCCC (Memory Controller Error), where
Memory subfield MMM = 100B (memory scrubbing)
Channel subfield CCCC = channel # or generic
0000_0001_0111_1010
L3 Explicit Writeback 0x17A
000F 0001 RRRR TTLL (Cache Hierarchy Error) where
Request subfields RRRR = 0111B (Eviction)
Transaction Type subfields TT = 10B (Generic)
Level subfields LL = 10B
NOTES:
1. Note that for both of these errors the correction report filtering (F) bit (bit 12) of the MCA error is
0, indicating "normal" filtering.
Table 15-16 lists values of relevant bit fields of IA32_MCi_STATUS for archi-
tecturally defined SRAO errors.
Table 15-16. IA32_MCi_STATUS Values for SRAO Errors
SRAO Error
Valid OVER UC EN MISCV ADDRV PCC
S
1
1
AR MCACOD
Memory Scrubbing
1
0
0
1
1
1
1
1
1
1
1
0
0
0
0
0xC0-0xCF
0x17A
L3 Explicit Writeback 1
For both the memory scrubbing and L3 explicit writeback errors, the ADDRV
and MISCV flags in the IA32_MCi_STATUS register are set to indicate that
the offending physical address information is available from the
IA32_MCi_MISC and the IA32_MCi_ADDR registers. For the memory scrub-
bing and L3 explicit writeback errors, the address mode in the
IA32_MCi_MISC register should be set as physical address mode (010b) and
the address LSB information in the IA32_MCi_MISC register should indicate
the lowest valid address bit in the address information provided from the
An MCE signal is broadcast to all logical processors on the system on which
the UCR errors are supported. MCi_STATUS banks can be shared by logical
processors within a core or within the same package. So several logical
processors may find an SRAO error in the shared IA32_MCi_STATUS bank
but other processors do not find it in any of the IA32_MCi_STATUS banks.
Table 15-17 shows the RIPV and EIPV flag indication in the
Vol. 3 15-35
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
IA32_MCG_STATUS register for the memory scrubbing and L3 explicit write-
back errors on both the reporting and non-reporting logical processors.
Table 15-17. IA32_MCG_STATUS Flag Indication for SRAO Errors
SRAO Type
Reporting Logical Processors
Non-reporting Logical Processors
RIPV
EIPV
RIPV
EIPV
0
Memory Scrubbing
1
1
0
0
1
1
L3 Explicit Writeback
0
15.9.3.2 Architecturally Defined SRAR Errors
The following two SRAR errors are architecturally defined.
• UCR Errors detected on instruction fetch.
The MCA error code encodings for these two architecturally-defined UCR
errors corresponds to sub-classes of compound MCA error codes (see Table
15-9). Their values and compound encoding format are given in Table
15-18.
Table 15-18. MCA Compound Error Code Encoding for SRAR Errors
1
Type
MCACOD Value MCA Error Code Encoding
Data Load
0x134
0000_0001_0011_0100
000F 0001 RRRR TTLL (Cache Hierarchy Error), where
Request subfield RRRR = 0011B (Data Load)
Transaction Type subfield TT= 01B (Data)
Level subfield LL = 00B (Level 0)
Instruction Fetch
0x150
0000_0001_0101_0000
000F 0001 RRRR TTLL (Cache Hierarchy Error), where
Request subfield RRRR = 0101B (Instruction Fetch)
Transaction Type subfield TT= 00B (Instruction)
Level subfield LL = 00B (Level 0)
NOTES:
1. Note that for both of these errors the correction report filtering (F) bit (bit 12) of the MCA error is
0, indicating "normal" filtering.
15-36 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
Table 15-19 lists values of relevant bit fields of IA32_MCi_STATUS for archi-
tecturally defined SRAR errors.
Table 15-19. IA32_MCi_STATUS Values for SRAR Errors
SRAR Error
Valid OVER UC EN MISCV ADDRV PCC
S
1
1
AR MCACOD
Data Load
1
1
0
0
1
1
1
1
1
1
1
1
0
0
1
1
0x134
0x150
Instruction Fetch
For both the data load and instruction fetch errors, the ADDRV and MISCV
flags in the IA32_MCi_STATUS register are set to indicate that the offending
physical address information is available from the IA32_MCi_MISC and the
IA32_MCi_ADDR registers. For the memory scrubbing and L3 explicit write-
back errors, the address mode in the IA32_MCi_MISC register should be set
as physical address mode (010b) and the address LSB information in the
IA32_MCi_MISC register should indicate the lowest valid address bit in the
address information provided from the IA32_MCi_ADDR register.
An MCE signal is broadcast to all logical processors on the system on which
the UCR errors are supported. The IA32_MCG_STATUS MSR allows system
software to distinguish the affected logical processor of an SRAR error
amongst logical processors that observed SRAR via a shared MCi_STATUS
bank.
Table 15-20 shows the RIPV and EIPV flag indication in the
IA32_MCG_STATUS register for the data load and instruction fetch errors on
both the reporting and non-reporting logical processors.
Table 15-20. IA32_MCG_STATUS Flag Indication for SRAR Errors
SRAR Type
Affected Logical Processors
Non-Affected Logical Processors
RIPV
EIPV
1
RIPV
EIPV
Data Load
0
0
1
1
0
0
instruction Fetch
0
The affected logical processor is the one that has detected and raised an
SRAR error at the point of the consumption in the execution flow. The
affected logical processor should find the Data Load or the Instruction Fetch
error information in the IA32_MCi_STATUS register that is reporting the
SRAR error.
For Data Load recoverable errors, the affected logical processor should find
that the IA32_MCG_STATUS.RIPV flag is cleared and the
IA32_MCG_STATUS.EIPV flag is set indicating that the error is detected at
the instruction pointer saved on the stack for this machine check exception
and restarting execution with the interrupted context is not possible.
Vol. 3 15-37
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
For Instruction Fetch recoverable error, the affected logical processor should
find that the RIPV flag and the EIPV Flag in the IA32_MCG_STATUS register
are cleared, indicating that the error is detected at the instruction pointer
saved on the stack may not be associated with this error and restarting the
execution with the interrupted context is not possible.
The logical processors that observed but not affected by an SRAR error
should find that the RIPV flag in the IA32_MCG_STATUS register is set and
the EIPV flag in the IA32_MCG_STATUS register is cleared, indicating that it
is safe to restart the execution at the instruction saved on the stack for the
machine check exception on these processors after the recovery action is
successfully taken by system software.
For the Data-Load and the Instruction-Fetch recoverable errors, system
software may take the following recovery actions for the affected logical
processor:
• The current executing thread cannot be continued. You must terminate the
interrupted stream of execution and provide a new stream of execution on return
from the machine check handler for the affected logical processor
In addition to taking the recovery action described above, system software
may also need to disable the use of the affected page from the program. This
recovery action by system software may prevent the occurrence of future
consumption errors from that affected page.
15.9.4
Multiple MCA Errors
When multiple MCA errors are detected within a certain detection window,
the processor may aggregate the reporting of these errors together as a
single event, i.e. a single machine exception condition. If this occurs,
system software may find multiple MCA errors logged in different MC banks
on one logical processor or find multiple MCA errors logged across different
processors for a single machine check broadcast event. In order to handle
multiple UCR errors reported from a single machine check event and
possibly recover from multiple errors, system software may consider the
following:
• Whether it can recover from multiple errors is determined by the most severe
error reported on the system. If the most severe error is found to be an unrecov-
erable error (VAL=1, UC=1, PCC=1 and EN=1) after system software examines
the MC banks of all processors to which the MCA signal is broadcast, recovery
from the multiple errors is not possible and system software needs to reset the
system.
15-38 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
• When multiple recoverable errors are reported and no other fatal condition (e.g..
overflowed condition for SRAR error) is found for the reported recoverable errors,
it is possible for system software to recover from the multiple recoverable errors
by taking necessary recovery action for each individual recoverable error.
However, system software can no longer expect one to one relationship with the
error information recorded in the IA32_MCi_STATUS register and the states of
the RIPV and EIPV flags in the IA32_MCG_STATUS register as the states of the
RIPV and the EIPV flags in the IA32_MCG_STATUS register may indicate the
information for the most severe error recorded on the processor. System
software is required to use the RIPV flag indication in the IA32_MCG_STATUS
register to make a final decision of recoverability of the errors and find the
restart-ability requirement after examining each IA32_MCi_STATUS register
error information in the MC banks.
15.9.5
Machine-Check Error Codes Interpretation
Appendix E, “Interpreting Machine-Check Error Codes,” provides information
on interpreting the MCA error code, model-specific error code, and other
information error code fields. For P6 family processors, information has been
included on decoding external bus errors. For Pentium 4 and Intel Xeon
processors; information is included on external bus, internal timer and cache
hierarchy errors.
15.10 GUIDELINES FOR WRITING MACHINE-CHECK
SOFTWARE
The machine-check architecture and error logging can be used in three
different ways:
• To detect machine errors during normal instruction execution, using the
machine-check exception (#MC).
• To periodically check and log machine errors.
• To examine recoverable UCR errors, determine software recoverability and
perform recovery actions via a machine-check exception handler or a corrected
machine-check interrupt handler.
To use the machine-check exception, the operating system or executive
software must provide a machine-check exception handler. This handler may
need to be designed specifically for each family of processors.
A special program or utility is required to log machine errors.
Vol. 3 15-39
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
Guidelines for writing a machine-check exception handler or a machine-
error logging utility are given in the following sections.
15.10.1 Machine-Check Exception Handler
The machine-check exception (#MC) corresponds to vector 18. To service
machine-check exceptions, a trap gate must be added to the IDT. The
pointer in the trap gate must point to a machine-check exception handler.
Two approaches can be taken to designing the exception handler:
1. The handler can merely log all the machine status and error information, then call
a debugger or shut down the system.
2. The handler can analyze the reported error information and, in some cases,
attempt to correct the error and restart the processor.
For Pentium 4, Intel Xeon, P6 family, and Pentium processors; virtually all
machine-check conditions cannot be corrected (they result in abort-type
exceptions). The logging of status and error information is therefore a base-
line implementation requirement.
When recovery from a machine-check error may be possible, consider the
following when writing a machine-check exception handler:
• To determine the nature of the error, the handler must read each of the error-
reporting register banks. The count field in the IA32_MCG_CAP register gives
• The VAL (valid) flag in each IA32_MCi_STATUS register indicates whether the
error information in the register is valid. If this flag is clear, the registers in that
bank do not contain valid error information and do not need to be checked.
• To write a portable exception handler, only the MCA error code field in the
IA32_MCi_STATUS register should be checked. See Section 15.9, “Interpreting
the MCA Error Codes,” for information that can be used to write an algorithm to
interpret this field.
• The RIPV, PCC, and OVER flags in each IA32_MCi_STATUS register indicate
whether recovery from the error is possible. If PCC or OVER are set, recovery is
not possible. If RIPV is not set, program execution can not be restarted reliably.
When recovery is not possible, the handler typically records the error information
and signals an abort to the operating system.
• Correctable errors are corrected automatically by the processor. The UC flag in
each IA32_MCi_STATUS register indicates whether the processor automatically
corrected an error.
• The RIPV flag in the IA32_MCG_STATUS register indicates whether the program
can be restarted at the instruction indicated by the instruction pointer (the
address of the instruction pushed on the stack when the exception was
15-40 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
generated). If this flag is clear, the processor may still be able to be restarted (for
debugging purposes) but not without loss of program continuity.
• For unrecoverable errors, the EIPV flag in the IA32_MCG_STATUS register
indicates whether the instruction indicated by the instruction pointer pushed on
the stack (when the exception was generated) is related to the error. If the flag is
clear, the pushed instruction may not be related to the error.
• The MCIP flag in the IA32_MCG_STATUS register indicates whether a machine-
check exception was generated. Before returning from the machine-check
exception handler, software should clear this flag so that it can be used reliably by
an error logging utility. The MCIP flag also detects recursion. The machine-check
architecture does not support recursion. When the processor detects machine-
check recursion, it enters the shutdown state.
Example 15-2 gives typical steps carried out by a machine-check exception
handler.
Example 15-2. Machine-Check Exception Handler Pseudocode
IF CPU supports MCE
THEN
IF CPU supports MCA
THEN
call errorlogging routine; (* returns restartability *)
FI;
ELSE (* Pentium(R) processor compatible *)
READ P5_MC_ADDR
READ P5_MC_TYPE;
report RESTARTABILITY to console;
FI;
IF error is not restartable
THEN
report RESTARTABILITY to console;
abort system;
FI;
CLEAR MCIP flag in IA32_MCG_STATUS;
15.10.2 Pentium Processor Machine-Check Exception Handling
Machine-check exception handler on P6 family and later processor families,
should follow the guidelines described in Section 15.10.1 and Example 15-2
that check the processor’s support of MCA.
NOTE
On processors that support MCA (CPUID.1.EDX.MCA == 1) reading
the P5_MC_TYPE and P5_MC_ADDR registers may produce invalid
data.
Vol. 3 15-41
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
When machine-check exceptions are enabled for the Pentium processor
(MCE flag is set in control register CR4), the machine-check exception
handler uses the RDMSR instruction to read the error type from the
P5_MC_TYPE register and the machine check address from the
P5_MC_ADDR register. The handler then normally reports these register
values to the system console before aborting execution (see Example 15-2).
15.10.3 Logging Correctable Machine-Check Errors
The error handling routine for servicing the machine-check exceptions is
responsible for logging uncorrected errors.
If a machine-check error is correctable, the processor does not generate a
machine-check exception for it. To detect correctable machine-check errors,
a utility program must be written that reads each of the machine-check
error-reporting register banks and logs the results in an accounting file or
data structure. This utility can be implemented in either of the following
ways.
• A system daemon that polls the register banks on an infrequent basis, such as
hourly or daily.
• A user-initiated application that polls the register banks and records the
exceptions. Here, the actual polling service is provided by an operating-system
driver or through the system call interface.
• An interrupt service routine servicing CMCI can read the MC banks and log the
error.
Example 15-3 gives pseudocode for an error logging utility.
Example 15-3. Machine-Check Error Logging Pseudocode
Assume that execution is restartable;
IF the processor supports MCA
THEN
FOR each bank of machine-check registers
DO
READ IA32_MCi_STATUS;
IF VAL flag in IA32_MCi_STATUS = 1
THEN
IF ADDRV flag in IA32_MCi_STATUS = 1
THEN READ IA32_MCi_ADDR;
FI;
IF MISCV flag in IA32_MCi_STATUS = 1
THEN READ IA32_MCi_MISC;
FI;
IF MCIP flag in IA32_MCG_STATUS = 1
(* Machine-check exception is in progress *)
15-42 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
AND PCC flag in IA32_MCi_STATUS = 1
OR RIPV flag in IA32_MCG_STATUS = 0
(* execution is not restartable *)
THEN
RESTARTABILITY = FALSE;
return RESTARTABILITY to calling procedure;
FI;
Save time-stamp counter and processor ID;
Set IA32_MCi_STATUS to all 0s;
Execute serializing instruction (i.e., CPUID);
FI;
OD;
FI;
If the processor supports the machine-check architecture, the utility reads
through the banks of error-reporting registers looking for valid register
entries. It then saves the values of the IA32_MCi_STATUS, IA32_MCi_ADDR,
IA32_MCi_MISC and IA32_MCG_STATUS registers for each bank that is
valid. The routine minimizes processing time by recording the raw data into
a system data structure or file, reducing the overhead associated with
polling. User utilities analyze the collected data in an off-line environment.
When the MCIP flag is set in the IA32_MCG_STATUS register, a machine-
check exception is in progress and the machine-check exception handler has
called the exception logging routine.
Once the logging process has been completed the exception-handling
routine must determine whether execution can be restarted, which is usually
possible when damage has not occurred (The PCC flag is clear, in the
IA32_MCi_STATUS register) and when the processor can guarantee that
execution is restartable (the RIPV flag is set in the IA32_MCG_STATUS
register). If execution cannot be restarted, the system is not recoverable
and the exception-handling routine should signal the console appropriately
before returning the error status to the Operating System kernel for subse-
quent shutdown.
The machine-check architecture allows buffering of exceptions from a given
error-reporting bank although the Pentium 4, Intel Xeon, and P6 family
processors do not implement this feature. The error logging routine should
provide compatibility with future processors by reading each hardware
error-reporting bank's IA32_MCi_STATUS register and then writing 0s to
clear the OVER and VAL flags in this register. The error logging utility should
re-read the IA32_MCi_STATUS register for the bank ensuring that the valid
bit is clear. The processor will write the next error into the register bank and
set the VAL flags.
Additional information that should be stored by the exception-logging
routine includes the processor’s time-stamp counter value, which provides a
Vol. 3 15-43
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
mechanism to indicate the frequency of exceptions. A multiprocessing oper-
ating system stores the identity of the processor node incurring the excep-
tion using a unique identifier, such as the processor’s APIC ID (see Section
10.9, “Handling Interrupts”).
The basic algorithm given in Example 15-3 can be modified to provide more
robust recovery techniques. For example, software has the flexibility to
attempt recovery using information unavailable to the hardware. Specifi-
cally, the machine-check exception handler can, after logging carefully
analyze the error-reporting registers when the error-logging routine reports
an error that does not allow execution to be restarted. These recovery tech-
niques can use external bus related model-specific information provided
with the error report to localize the source of the error within the system and
determine the appropriate recovery strategy.
15.10.4 Machine-Check Software Handler Guidelines for Error
Recovery
15.10.4.1 Machine-Check Exception Handler for Error Recovery
When writing a machine-check exception (MCE) handler to support software
recovery from Uncorrected Recoverable (UCR) errors, consider the
following:
• When IA32_MCG_CAP [24] is zero, there are no recoverable errors supported
and all machine-check are fatal exceptions. The logging of status and error
information is therefore a baseline implementation requirement.
• When IA32_MCG_CAP [24] is 1, certain uncorrected errors called uncorrected
recoverable (UCR) errors may be software recoverable. The handler can analyze
the reported error information, and in some cases attempt to recover from the
uncorrected error and continue execution.
• For processors with DisplayFamily_DisplayModel encoding of 06H_EH and above,
a MCA signal is broadcast to all logical processors in the system. Due to the
potentially shared machine check MSR resources among the logical processors
on the same package/core, the MCE handler may be required to synchronize with
the other processors that received a machine check error and serialize access to
the machine check registers when analyzing, logging and clearing the
information in the machine check registers.
• The VAL (valid) flag in each IA32_MCi_STATUS register indicates whether the
error information in the register is valid. If this flag is clear, the registers in that
bank do not contain valid error information and should not be checked.
• The MCE handler is primarily responsible for processing uncorrected errors. The
UC flag in each IA32_MCi_Status register indicates whether the reported error
15-44 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
was corrected (UC=0) or uncorrected (UC=1). The MCE handler can optionally
log and clear the corrected errors in the MC banks if it can implement software
algorithm to avoid the undesired race conditions with the CMCI or CMC polling
handler.
• For uncorrectable errors, the EIPV flag in the IA32_MCG_STATUS register
indicates (when set) that the instruction pointed to by the instruction pointer
pushed onto the stack when the machine-check exception is generated is directly
associated with the error. When this flag is cleared, the instruction pointed to
may not be associated with the error.
• The MCIP flag in the IA32_MCG_STATUS register indicates whether a machine-
check exception was generated. When a machine check exception is generated,
it is expected that the MCIP flag in the IA32_MCG_STATUS register is set to 1. If
it is not set, this machine check was generated by either an INT 18 instruction or
some piece of hardware signaling an interrupt with vector 18.
When IA32_MCG_CAP [24] is 1, the following rules can apply when writing a
machine check exception (MCE) handler to support software recovery:
• The PCC flag in each IA32_MCi_STATUS register indicates whether recovery from
the error is possible for uncorrected errors (UC=1). If the PCC flag is set for
uncorrected errors (UC=1), recovery is not possible. When recovery is not
possible, the MCE handler typically records the error information and signals the
operating system to reset the system.
• The RIPV flag in the IA32_MCG_STATUS register indicates whether restarting the
program execution from the instruction pointer saved on the stack for the
machine check exception is possible. When the RIPV is set, program execution
can be restarted reliably when recovery is possible. If the RIPV flag is not set,
program execution cannot be restarted reliably. In this case the recovery
algorithm may involve terminating the current program execution and resuming
an alternate thread of execution upon return from the machine check handler
when recovery is possible. When recovery is not possible, the MCE handler
signals the operating system to reset the system.
• When the EN flag is zero but the VAL and UC flags are one in the
IA32_MCi_STATUS register, the reported uncorrected error in this bank is not
enabled. As uncorrected errors with the EN flag = 0 are not the source of
machine check exceptions, the MCE handler should log and clear non-enabled
errors when the S bit is set and should continue searching for enabled errors from
the other IA32_MCi_STATUS registers. Note that when IA32_MCG_CAP [24] is 0,
any uncorrected error condition (VAL =1 and UC=1) including the one with the
EN flag cleared are fatal and the handler must signal the operating system to
reset the system. For the errors that do not generate machine check exceptions,
the EN flag has no meaning. See Appendix A: Table A-4 to find the errors that do
not generate machine check exceptions.
• When the VAL flag is one, the UC flag is one, the EN flag is one and the PCC flag
is zero in the IA32_MCi_STATUS register, the error in this bank is an uncorrected
recoverable (UCR) error. The MCE handler needs to examine the S flag and the
Vol. 3 15-45
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
AR flag to find the type of the UCR error for software recovery and determine if
software error recovery is possible.
• When both the S and the AR flags are clear in the IA32_MCi_STATUS register for
the UCR error (VAL=1, UC=1, EN=x and PCC=0), the error in this bank is an
uncorrected no-action required error (UCNA). UCNA errors are uncorrected but
do not require any OS recovery action to continue execution. These errors
indicate that some data in the system is corrupt, but that data has not been
consumed and may not be consumed. If that data is consumed a non-UNCA
machine check exception will be generated. UCNA errors are signaled in the same
way as corrected machine check errors and the CMCI and CMC polling handler is
primarily responsible for handling UCNA errors. Like corrected errors, the MCA
handler can optionally log and clear UCNA errors as long as it can avoid the
undesired race condition with the CMCI or CMC polling handler. As UCNA errors
are not the source of machine check exceptions, the MCA handler should
continue searching for uncorrected or software recoverable errors in all other MC
banks.
• When the S flag in the IA32_MCi_STATUS register is set for the UCR error
((VAL=1, UC=1, EN=1 and PCC=0), the error in this bank is software recoverable
and it was signaled through a machine-check exception. The AR flag in the
IA32_MCi_STATUS register further clarifies the type of the software recoverable
errors.
• When the AR flag in the IA32_MCi_STATUS register is clear for the software
recoverable error (VAL=1, UC=1, EN=1, PCC=0 and S=1), the error in this bank
is a software recoverable action optional (SRAO) error. The MCE handler and the
operating system can analyze the IA32_MCi_STATUS [15:0] to implement MCA
error code specific optional recovery action, but this recovery action is optional.
System software can resume the program execution from the instruction pointer
saved on the stack for the machine check exception when the RIPV flag in the
IA32_MCG_STATUS register is set.
• When the OVER flag in the IA32_MCi_STATUS register is set for the SRAO error
(VAL=1, UC=1, EN=1, PCC=0, S=1 and AR=0), the MCE handler cannot take
recovery action as the information of the SRAO error in the IA32_MCi_STATUS
register was potentially lost due to the overflow condition. Since the recovery
action for SRAO errors is optional, restarting the program execution from the
instruction pointer saved on the stack for the machine check exception is still
possible for the overflowed SRAO error if the RIPV flag in the IA32_MCG_STATUS
is set.
• When the AR flag in the IA32_MCi_STATUS register is set for the software
recoverable error (VAL=1, UC=1, EN=1, PCC=0 and S=1), the error in this bank
is a software recoverable action required (SRAR) error. The MCE handler and the
operating system must take recovery action in order to continue execution after
the machine-check exception. The MCA handler and the operating system need
to analyze the IA32_MCi_STATUS [15:0] to determine the MCA error code
specific recovery action. If no recovery action can be performed, the operating
system must reset the system.
15-46 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
• When the OVER flag in the IA32_MCi_STATUS register is set for the SRAR error
(VAL=1, UC=1, EN=1, PCC=0, S=1 and AR=1), the MCE handler cannot take
recovery action as the information of the SRAR error in the IA32_MCi_STATUS
register was potentially lost due to the overflow condition. Since the recovery
action for SRAR errors must be taken, the MCE handler must signal the operating
system to reset the system.
• When the MCE handler cannot find any uncorrected (VAL=1, UC=1 and EN=1) or
any software recoverable errors (VAL=1, UC=1, EN=1, PCC=0 and S=1) in any
of the IA32_MCi banks of the processors, this is an unexpected condition for the
MCE handler and the handler should signal the operating system to reset the
system.
• Before returning from the machine-check exception handler, software must clear
the MCIP flag in the IA32_MCG_STATUS register. The MCIP flag is used to detect
recursion. The machine-check architecture does not support recursion. When the
processor receives a machine check when MCIP is set, it automatically enters the
shutdown state.
Example 15-4 gives pseudocode for an MC exception handler that supports
recovery of UCR.
Example 15-4. Machine-Check Error Handler Pseudocode Supporting UCR
MACHINE CHECK HANDLER: (* Called from INT 18 handler *)
NOERROR = TRUE;
ProcessorCount = 0;
IF CPU supports MCA
THEN
RESTARTABILITY = TRUE;
IF (Processor Family = 6 AND Display_Model >= 0EH) OR (Processor Family > 6)
THEN
MCA_BROADCAST = TRUE;
Acquire SpinLock;
ProcessorCount++; (* Allowing one logical processor at a time to examine machine check
registers *)
CALL MCA ERROR PROCESSING; (* returns RESTARTABILITY and NOERROR *)
ELSE
MCA_BROADCAST = FALSE;
(* Implement a rendezvous mechanism with the other processors if necessary *)
CALL MCA ERROR PROCESSING;
FI;
ELSE (* Pentium(R) processor compatible *)
READ P5_MC_ADDR
READ P5_MC_TYPE;
RESTARTABILITY = FALSE;
FI;
IF NOERROR = TRUE
THEN
IF NOT (MCG_RIPV = 1 AND MCG_EIPV = 0)
THEN
Vol. 3 15-47
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
RESTARTABILITY = FALSE;
FI
FI;
IF RESTARTABILITY = FALSE
THEN
Report RESTARTABILITY to console;
Reset system;
FI;
IF MCA_BROADCAST = TRUE
THEN
IF ProcessorCount = MAX_PROCESSORS
AND NOERROR = TRUE
THEN
Report RESTARTABILITY to console;
Reset system;
FI;
Release SpinLock;
Wait till ProcessorCount = MAX_PROCESSRS on system;
(* implement a timeout and abort function if necessary *)
FI;
CLEAR MCIP flag in IA32_MCG_STATUS;
RESUME Execution;
(* End of MACHINE CHECK HANDLER*)
MCA ERROR PROCESSING: (* MCA Error Processing Routine called from MCA Handler *)
IF MCIP flag in IA32_MCG_STATUS = 0
THEN (* MCIP=0 upon MCA is unexpected *)
RESTARTABILITY = FALSE;
FI;
FOR each bank of machine-check registers
DO
CLEAR_MC_BANK = FALSE;
READ IA32_MCi_STATUS;
IF VAL Flag in IA32_MCi_STATUS = 1
THEN
IF UC Flag in IA32_MCi_STATUS = 1
THEN
IF Bit 24 in IA32_MCG_CAP = 0
THEN (* the processor does not support software error recovery *)
RESTARTABILITY = FALSE;
NOERROR = FALSE;
GOTO LOG MCA REGISTER;
FI;
(* the processor supports software error recovery *)
IF EN Flag in IA32_MCi_STATUS = 0 AND OVER Flag in IA32_MCi_STATUS=0
THEN (* It is a spurious MCA Log. Log and clear the register *)
CLEAR_MC_BANK = TRUE;
GOTO LOG MCA REGISTER;
FI;
15-48 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
IF PCC Flag in IA32_MCi_STATUS = 1
THEN (* processor context might have been corrupted *)
RESTARTABILITY = FALSE;
ELSE (* It is a uncorrected recoverable (UCR) error *)
IF S Flag in IA32_MCi_STATUS = 0
THEN
IF AR Flag in IA32_MCi_STATUS = 0
THEN (* It is a uncorrected no action required (UCNA) error *)
GOTO CONTINUE; (* let CMCI and CMC polling handler to process *)
ELSE
FESTARTABILITY = FALSE; (* S=0, AR=1 is illegal *)
FI
FI;
IF RESTARTABILITY = FALSE
THEN (* no need to take recovery action if RESTARTABILITY is already false *)
NOERROR = FALSE;
GOTO LOG MCA REGISTER;
FI;
(* S in IA32_MCi_STATUS = 1 *)
IF AR Flag in IA32_MCi_STATUS = 1
THEN (* It is a software recoverable and action required (SRAR) error *)
IF OVER Flag in IA32_MCi_STATUS = 1
THEN
RESTARTABILITY = FALSE;
NOERROR = FALSE;
GOTO LOG MCA REGISTER;
FI
IF MCACOD Value in IA32_MCi_STATUS is recognized
AND Current Processor is an Affected Processor
THEN
Implement MCACOD specific recovery action;
CLEAR_MC_BANK = TURE;
ELSE
RESTARTABILITY = FALSE;
FI;
ELSE (* It is a software recoverable and action optional (SRAO) error *)
IF OVER Flag in IA32_MCi_STATUS = 0 AND
MCACOD in IA32_MCi_STATUS is recognized
THEN
Implement MCACOD specific recovery action;
FI;
CLEAR_MC_BANK = TRUE;
FI; AR
FI; PCC
NOERROR = FALSE;
GOTO LOG MCA REGISTER;
ELSE (* It is a corrected error; continue to the next IA32_MCi_STATUS *)
GOTO CONTINUE;
FI; UC
FI; VAL
LOG MCA REGISTER:
SAVE IA32_MCi_STATUS;
Vol. 3 15-49
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
If MISCV in IA32_MCi_STATUS
THEN
SAVE IA32_MCi_MISC;
FI;
IF ADDRV in IA32_MCi_STATUS
THEN
SAVE IA32_MCi_ADDR;
FI;
IF CLEAR_MC_BANK = TRUE
THEN
SET all 0 to IA32_MCi_STATUS;
If MISCV in IA32_MCi_STATUS
THEN
SET all 0 to IA32_MCi_MISC;
FI;
IF ADDRV in IA32_MCi_STATUS
THEN
SET all 0 to IA32_MCi_ADDR;
FI;
FI;
CONTINUE:
OD;
( *END FOR *)
RETURN;
(* End of MCA ERROR PROCESSING*)
15.10.4.2 Corrected Machine-Check Handler for Error Recovery
When writing a corrected machine check handler, which is invoked as a
result of CMCI or called from an OS CMC Polling dispatcher, consider the
following:
• The VAL (valid) flag in each IA32_MCi_STATUS register indicates whether the
error information in the register is valid. If this flag is clear, the registers in that
bank does not contain valid error information and does not need to be checked.
• The CMCI or CMC polling handler is responsible for logging and clearing corrected
errors. The UC flag in each IA32_MCi_Status register indicates whether the
reported error was corrected (UC=0) or not (UC=1).
• When IA32_MCG_CAP [24] is one, the CMC handler is also responsible for
logging and clearing uncorrected no-action required (UCNA) errors. When the
UC flag is one but the PCC, S, and AR flags are zero in the IA32_MCi_STATUS
register, the reported error in this bank is an uncorrected no-action required
(UCNA) error.
• In addition to corrected errors and UCNA errors, the CMC handler optionally logs
uncorrected (UC=1 and PCC=1), software recoverable machine check errors
(UC=1, PCC=0 and S=1), but should avoid clearing those errors from the MC
banks. Clearing these errors may result in accidentally removing these errors
15-50 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
before these errors are actually handled and processed by the MCE handler for
attempted software error recovery.
Example 15-5 gives pseudocode for a CMCI handler with UCR support.
Example 15-5. Corrected Error Handler Pseudocode with UCR Support
Corrected Error HANDLER: (* Called from CMCI handler or OS CMC Polling Dispatcher*)
IF CPU supports MCA
THEN
FOR each bank of machine-check registers
DO
READ IA32_MCi_STATUS;
IF VAL flag in IA32_MCi_STATUS = 1
THEN
IF UC Flag in IA32_MCi_STATUS = 0 (* It is a corrected error *)
THEN
GOTO LOG CMC ERROR;
ELSE
IF Bit 24 in IA32_MCG_CAP = 0
THEN
GOTO CONTINUE;
FI;
IF S Flag in IA32_MCi_STATUS = 0 AND AR Flag in IA32_MCi_STATUS = 0
THEN (* It is a uncorrected no action required error *)
GOTO LOG CMC ERROR
FI
IF EN Flag in IA32_MCi_STATUS = 0
THEN (* It is a spurious MCA error *)
GOTO LOG MCM ERROR
FI;
FI;
FI;
GOTO CONTINUE;
LOG CMC ERROR:
SAVE IA32_MCi_STATUS;
If MISCV Flag in IA32_MCi_STATUS
THEN
SAVE IA32_MCi_MISC;
SET all 0 to IA32_MCi_MISC;
FI;
IF ADDRV Flag in IA32_MCi_STATUS
THEN
SAVE IA32_MCi_ADDR;
SET all 0 to IA32_MCi_ADDR
FI;
SET all 0 to IA32_MCi_STATUS;
CONTINUE:
OD;
( *END FOR *)
FI;
Vol. 3 15-51
Download from Www.Somanuals.com. All Manuals Search And Download.
MACHINE-CHECK ARCHITECTURE
15-52 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
CHAPTER 16
DEBUGGING, PROFILING BRANCHES AND TIME-
STAMP COUNTER
Intel 64 and IA-32 architectures provide debug facilities for use in debugging code
and monitoring performance. These facilities are valuable for debugging application
software, system software, and multitasking operating systems. Debug support is
accessed using debug registers (DB0 through DB7) and model-specific registers
(MSRs):
• Debug registers hold the addresses of memory and I/O locations called break-
points. Breakpoints are user-selected locations in a program, a data-storage area
in memory, or specific I/O ports. They are set where a programmer or system
designer wishes to halt execution of a program and examine the state of the
processor by invoking debugger software. A debug exception (#DB) is generated
when a memory or I/O access is made to a breakpoint address.
• MSRs monitor branches, interrupts, and exceptions; they record addresses of the
last branch, interrupt or exception taken and the last branch taken before an
interrupt or exception.
16.1
OVERVIEW OF DEBUG SUPPORT FACILITIES
The following processor facilities support debugging and performance monitoring:
• Debug exception (#DB) — Transfers program control to a debug procedure or
task when a debug event occurs.
• Breakpoint exception (#BP) — See breakpoint instruction (INT 3) below.
• Breakpoint-address registers (DR0 through DR3) — Specifies the
addresses of up to 4 breakpoints.
• Debug status register (DR6) — Reports the conditions that were in effect
when a debug or breakpoint exception was generated.
• Debug control register (DR7) — Specifies the forms of memory or I/O access
that cause breakpoints to be generated.
• T (trap) flag, TSS — Generates a debug exception (#DB) when an attempt is
made to switch to a task with the T flag set in its TSS.
• RF (resume) flag, EFLAGS register — Suppresses multiple exceptions to the
same instruction.
• TF (trap) flag, EFLAGS register — Generates a debug exception (#DB) after
every execution of an instruction.
• Breakpoint instruction (INT 3) — Generates a breakpoint exception (#BP)
that transfers program control to the debugger procedure or task. This
Vol. 3 16-1
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
instruction is an alternative way to set code breakpoints. It is especially useful
when more than four breakpoints are desired, or when breakpoints are being
placed in the source code.
• Last branch recording facilities — Store branch records in the last branch
record (LBR) stack MSRs for the most recent taken branches, interrupts, and/or
exceptions in MSRs. A branch record consist of a branch-from and a branch-to
instruction address. Send branch records out on the system bus as branch trace
messages (BTMs).
These facilities allow a debugger to be called as a separate task or as a procedure in
the context of the current program or task. The following conditions can be used to
invoke the debugger:
• Task switch to a specific task.
• Execution of the breakpoint instruction.
• Execution of any instruction.
• Execution of an instruction at a specified address.
• Read or write to a specified memory address/range.
• Write to a specified memory address/range.
• Input from a specified I/O address/range.
• Output to a specified I/O address/range.
• Attempt to change the contents of a debug register.
16.2
DEBUG REGISTERS
Eight debug registers (see Figure 16-1) control the debug operation of the processor.
These registers can be written to and read using the move to/from debug register
form of the MOV instruction. A debug register may be the source or destination
operand for one of these instructions.
Debug registers are privileged resources; a MOV instruction that accesses these
registers can only be executed in real-address mode, in SMM or in protected mode at
a CPL of 0. An attempt to read or write the debug registers from any other privilege
level generates a general-protection exception (#GP).
The primary function of the debug registers is to set up and monitor from 1 to 4
breakpoints, numbered 0 though 3. For each breakpoint, the following information
can be specified:
• The linear address where the breakpoint is to occur.
• The length of the breakpoint location (1, 2, or 4 bytes).
• The operation that must be performed at the address for a debug exception to be
generated.
• Whether the breakpoint is enabled.
16-2 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
• Whether the breakpoint condition was present when the debug exception was
generated.
31
23
21 20 19 18 1716
15 14 13 12 11
30 29 28 27 26 25 24
22
6
5
4
3
2
1
0
10 9
8
7
LEN R/W LEN
R/W LEN R/W LEN R/W
G
D
0 0 1 G L G L G L G L G L
E E 3 3 2 2 1 1 0 0
0 0
DR7
DR6
DR5
DR4
DR3
DR2
3
3
2
2
1
1
0
0
31
16
15 14 13 12 11
10 9
8
7
6
5
4
3
2
1
0
B B B
T S D
B B B B
3 2 1 0
Reserved (set to 1)
0 1 1 1 1 1 1 1 1 1
31
31
31
31
31
31
0
0
0
0
0
0
Breakpoint 3 Linear Address
Breakpoint 2 Linear Address
Breakpoint 1 Linear Address
Breakpoint 0 Linear Address
DR1
DR0
Reserved
Figure 16-1. Debug Registers
The following paragraphs describe the functions of flags and fields in the debug
registers.
Vol. 3 16-3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
16.2.1
Debug Address Registers (DR0-DR3)
Each of the debug-address registers (DR0 through DR3) holds the 32-bit linear
address of a breakpoint (see Figure 16-1). Breakpoint comparisons are made before
physical address translation occurs. The contents of debug register DR7 further spec-
ifies breakpoint conditions.
16.2.2
Debug Registers DR4 and DR5
Debug registers DR4 and DR5 are reserved when debug extensions are enabled
(when the DE flag in control register CR4 is set) and attempts to reference the DR4
and DR5 registers cause invalid-opcode exceptions (#UD). When debug extensions
are not enabled (when the DE flag is clear), these registers are aliased to debug
registers DR6 and DR7.
16.2.3
Debug Status Register (DR6)
The debug status register (DR6) reports debug conditions that were sampled at the
time the last debug exception was generated (see Figure 16-1). Updates to this
register only occur when an exception is generated. The flags in this register show
the following information:
• B0 through B3 (breakpoint condition detected) flags (bits 0 through 3)
— Indicates (when set) that its associated breakpoint condition was met when a
debug exception was generated. These flags are set if the condition described for
each breakpoint by the LENn, and R/Wn flags in debug control register DR7 is
true. They may or may not be set if the breakpoint is not enabled by the Ln or the
Gn flags in register DR7. Therefore on a #DB, a debug handler should check only
those B0-B3 bits which correspond to an enabled breakpoint.
• BD (debug register access detected) flag (bit 13) — Indicates that the next
instruction in the instruction stream accesses one of the debug registers (DR0
through DR7). This flag is enabled when the GD (general detect) flag in debug
control register DR7 is set. See Section 16.2.4, “Debug Control Register (DR7),”
for further explanation of the purpose of this flag.
was triggered by the single-step execution mode (enabled with the TF flag in the
EFLAGS register). The single-step mode is the highest-priority debug exception.
When the BS flag is set, any of the other debug status bits also may be set.
• BT (task switch) flag (bit 15) — Indicates (when set) that the debug
exception resulted from a task switch where the T flag (debug trap flag) in the
TSS of the target task was set. See Section 7.2.1, “Task-State Segment (TSS),”
for the format of a TSS. There is no flag in debug control register DR7 to enable
or disable this exception; the T flag of the TSS is the only enabling flag.
Certain debug exceptions may clear bits 0-3. The remaining contents of the DR6
register are never cleared by the processor. To avoid confusion in identifying debug
16-4 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
exceptions, debug handlers should clear the register before returning to the inter-
rupted task.
16.2.4
Debug Control Register (DR7)
The debug control register (DR7) enables or disables breakpoints and sets break-
point conditions (see Figure 16-1). The flags and fields in this register control the
following things:
• L0 through L3 (local breakpoint enable) flags (bits 0, 2, 4, and 6) —
Enables (when set) the breakpoint condition for the associated breakpoint for the
current task. When a breakpoint condition is detected and its associated Ln flag
is set, a debug exception is generated. The processor automatically clears these
flags on every task switch to avoid unwanted breakpoint conditions in the new
task.
• G0 through G3 (global breakpoint enable) flags (bits 1, 3, 5, and 7) —
Enables (when set) the breakpoint condition for the associated breakpoint for all
tasks. When a breakpoint condition is detected and its associated Gn flag is set,
a debug exception is generated. The processor does not clear these flags on a
task switch, allowing a breakpoint to be enabled for all tasks.
• LE and GE (local and global exact breakpoint enable) flags (bits 8, 9) —
This feature is not supported in the P6 family processors, later IA-32 processors,
and Intel 64 processors. When set, these flags cause the processor to detect the
exact instruction that caused a data breakpoint condition. For backward and
forward compatibility with other Intel processors, we recommend that the LE and
GE flags be set to 1 if exact breakpoints are required.
• GD (general detect enable) flag (bit 13) — Enables (when set) debug-
register protection, which causes a debug exception to be generated prior to any
MOV instruction that accesses a debug register. When such a condition is
detected, the BD flag in debug status register DR6 is set prior to generating the
exception. This condition is provided to support in-circuit emulators.
When the emulator needs to access the debug registers, emulator software can
set the GD flag to prevent interference from the program currently executing on
the processor.
The processor clears the GD flag upon entering to the debug exception handler,
to allow the handler access to the debug registers.
• R/W0 through R/W3 (read/write) fields (bits 16, 17, 20, 21, 24, 25, 28,
and 29) — Specifies the breakpoint condition for the corresponding breakpoint.
The DE (debug extensions) flag in control register CR4 determines how the bits in
the R/Wn fields are interpreted. When the DE flag is set, the processor interprets
bits as follows:
00 — Break on instruction execution only.
01 — Break on data writes only.
Vol. 3 16-5
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
10 — Break on I/O reads or writes.
11 — Break on data reads or writes but not instruction fetches.
When the DE flag is clear, the processor interprets the R/Wn bits the same as for
the Intel386™ and Intel486™ processors, which is as follows:
00 — Break on instruction execution only.
01 — Break on data writes only.
10 — Undefined.
11 — Break on data reads or writes but not instruction fetches.
• LEN0 through LEN3 (Length) fields (bits 18, 19, 22, 23, 26, 27, 30, and
31) — Specify the size of the memory location at the address specified in the
corresponding breakpoint address register (DR0 through DR3). These fields are
interpreted as follows:
01 — 2-byte length.
10 — Undefined (or 8 byte length, see note below).
11 — 4-byte length.
If the corresponding RWn field in register DR7 is 00 (instruction execution), then the
LENn field should also be 00. The effect of using other lengths is undefined. See
Section 16.2.5, “Breakpoint Field Recognition,” below.
NOTES
®
®
®
For Pentium 4 and Intel Xeon processors with a CPUID signature
corresponding to family 15 (model 3, 4, and 6), break point
conditions permit specifying 8-byte length on data read/write with an
of encoding 10B in the LENn field.
Encoding 10B is also supported in processors based on Intel Core
microarchitecture or enhanced Intel Core microarchitecture, the
respective CPUID signatures corresponding to family 6, model 15,
and family 6, display_model value 23. The Encoding 10B is supported
in processors based on Intel Atom microarchitecture, with CPUID
signature of family 6, display_model value 28. The encoding 10B is
undefined for other processors.
16.2.5
Breakpoint Field Recognition
Breakpoint address registers (debug registers DR0 through DR3) and the LENn fields
for each breakpoint define a range of sequential byte addresses for a data or I/O
breakpoint. The LENn fields permit specification of a 1-, 2-, 4-, or 8-byte range,
beginning at the linear address specified in the corresponding debug register (DRn).
Two-byte ranges must be aligned on word boundaries; 4-byte ranges must be
aligned on doubleword boundaries. I/O addresses are zero-extended (from 16 to 32
bits, for comparison with the breakpoint address in the selected debug register).
These requirements are enforced by the processor; it uses LENn field bits to mask
16-6 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
the lower address bits in the debug registers. Unaligned data or I/O breakpoint
addresses do not yield valid results.
A data breakpoint for reading or writing data is triggered if any of the bytes partici-
pating in an access is within the range defined by a breakpoint address register and
its LENn field. Table 16-1 provides an example setup of debug registers and data
accesses that would subsequently trap or not trap on the breakpoints.
A data breakpoint for an unaligned operand can be constructed using two break-
points, where each breakpoint is byte-aligned and the two breakpoints together
cover the operand. The breakpoints generate exceptions only for the operand, not for
neighboring bytes.
Instruction breakpoint addresses must have a length specification of 1 byte (the
LENn field is set to 00). Code breakpoints for other operand sizes are undefined. The
processor recognizes an instruction breakpoint address only when it points to the
first byte of an instruction. If the instruction has prefixes, the breakpoint address
must point to the first prefix.
Table 16-1. Breakpoint Examples
Debug Register Setup
Debug Register
R/Wn
R/W0 = 11 (Read/Write) A0001H
R/W1 = 01 (Write) A0002H
R/W2 = 11 (Read/Write) B0002H
Breakpoint Address
LENn
DR0
DR1
DR2
DR3
LEN0 = 00 (1 byte)
LEN1 = 00 (1 byte)
LEN2 = 01) (2 bytes)
LEN3 = 11 (4 bytes)
R/W3 = 01 (Write)
C0000H
Data Accesses
Operation
Address
Access Length
(In Bytes)
Data operations that trap
- Read or write
- Read or write
- Write
A0001H
A0001H
A0002H
A0002H
B0001H
B0002H
B0002H
C0000H
C0001H
C0003H
1
2
1
2
4
1
2
4
2
1
- Write
- Read or write
- Read or write
- Read or write
- Write
- Write
- Write
Vol. 3 16-7
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
Table 16-1. Breakpoint Examples (Contd.)
Debug Register Setup
Debug Register
R/Wn
Breakpoint Address
LENn
Data operations that do not trap
- Read or write
- Read
- Read or write
- Read or write
- Read
A0000H
A0002H
A0003H
B0000H
C0000H
C0004H
1
1
4
2
2
4
- Read or write
16.2.6
Debug Registers and Intel® 64 Processors
For Intel 64 architecture processors, debug registers DR0–DR7 are 64 bits. In 16-bit
fill the upper 32 bits with zeros. Reads from a debug register return the lower 32 bits.
In 64-bit mode, MOV DRn instructions read or write all 64 bits. Operand-size prefixes
are ignored.
In 64-bit mode, the upper 32 bits of DR6 and DR7 are reserved and must be written
with zeros. Writing 1 to any of the upper 32 bits results in a #GP(0) exception (see
Figure 16-2). All 64 bits of DR0–DR3 are writable by software. However, MOV DRn
instructions do not check that addresses written to DR0–DR3 are in the linear-
address limits of the processor implementation (address matching is supported only
on valid addresses generated by the processor implementation). Break point condi-
tions for 8-byte memory read/writes are supported in all modes.
16-8 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
63
31
32
DR7
30 29 28 27 26 25 24
23
22
21 20 19 18 1716
15 14 13 12 11
7
6
5
4
3
2
1
0
10 9
8
LEN R/W LEN
R/W LEN R/W LEN R/W 0 0 G
G L G L G L G L G L
E E 3 3 2 2 1 1 0 0
0 0 1
DR7
3
3
2
2
1
1
0
0
D
63
31
32
DR6
DR6
16
15 14 13 12 11
10 9
8
7
6
5
4
3
2
1
0
Reserved (set to 1)
0 1 1 1 1 1 1 1 1 1
B B B
T S D
B B B B
3 2 1 0
Reserved
Figure 16-2. DR6/DR7 Layout on Processors Supporting Intel 64 Technology
16.3
DEBUG EXCEPTIONS
The Intel 64 and IA-32 architectures dedicate two interrupt vectors to handling
debug exceptions: vector 1 (debug exception, #DB) and vector 3 (breakpoint excep-
tion, #BP). The following sections describe how these exceptions are generated and
typical exception handler operations.
16.3.1
Debug Exception (#DB)—Interrupt Vector 1
The debug-exception handler is usually a debugger program or part of a larger soft-
tions. The debugger checks flags in the DR6 and DR7 registers to determine which
condition caused the exception and which other conditions might apply. Table 16-2
shows the states of these flags following the generation of each kind of breakpoint
condition.
Instruction-breakpoint and general-detect condition (see Section 16.3.1.3, “General-
Detect Exception Condition”) result in faults; other debug-exception conditions result
in traps. The debug exception may report one or both at one time. The following
sections describe each class of debug exception.
Vol. 3 16-9
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
See also: Chapter 6, “Interrupt 1—Debug Exception (#DB),” in the Intel® 64 and
IA-32 Architectures Software Developer’s Manual, Volume 3A.
Table 16-2. Debug Exception Conditions
Debug or Breakpoint Condition
DR6 Flags
Tested
DR7 Flags
Tested
Exception Class
Single-step trap
BS = 1
Trap
Instruction breakpoint, at addresses
defined by DRn and LENn
Bn = 1 and
(Gn or Ln = 1)
R/Wn = 0
R/Wn = 1
R/Wn = 2
R/Wn = 3
Fault
Data write breakpoint, at addresses
defined by DRn and LENn
Bn = 1 and
(Gn or Ln = 1)
Trap
Trap
Trap
I/O read or write breakpoint, at
addresses defined by DRn and LENn
Bn = 1 and
(Gn or Ln = 1)
Data read or write (but not instruction Bn = 1 and
fetches), at addresses defined by DRn (Gn or Ln = 1)
and LENn
General detect fault, resulting from an BD = 1
attempt to modify debug registers
(usually in conjunction with in-circuit
emulation)
Fault
Trap
Task switch
BT = 1
16.3.1.1 Instruction-Breakpoint Exception Condition
The processor reports an instruction breakpoint when it attempts to execute an
instruction at an address specified in a breakpoint-address register (DB0 through
DR3) that has been set up to detect instruction execution (R/W flag is set to 0). Upon
reporting the instruction breakpoint, the processor generates a fault-class, debug
exception (#DB) before it executes the target instruction for the breakpoint.
Instruction breakpoints are the highest priority debug exceptions. They are serviced
before any other exceptions detected during the decoding or execution of an instruc-
tion. However, if a code instruction breakpoint is placed on an instruction located
immediately after a POP SS/MOV SS instruction, the breakpoint may not be trig-
gered. In most situations, POP SS/MOV SS will inhibit such interrupts (see
“MOV—Move” and “POP—Pop a Value from the Stack” in Chapters 3 and 4 of the
Intel® 64 and IA-32 Architectures Software Developer’s Manual, Volumes
2A & 2B).
Because the debug exception for an instruction breakpoint is generated before the
instruction is executed, if the instruction breakpoint is not removed by the exception
handler; the processor will detect the instruction breakpoint again when the instruc-
tion is restarted and generate another debug exception. To prevent looping on an
instruction breakpoint, the Intel 64 and IA-32 architectures provide the RF flag
16-10 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
(resume flag) in the EFLAGS register (see Section 2.3, “System Flags and Fields in
the EFLAGS Register,” in the Intel® 64 and IA-32 Architectures Software Developer’s
Manual, Volume 3A). When the RF flag is set, the processor ignores instruction
breakpoints.
All Intel 64 and IA-32 processors manage the RF flag as follows. The RF Flag is
cleared at the start of the instruction after the check for code breakpoint, CS limit
violation and FP exceptions. Task Switches and IRETD/IRETQ instructions transfer
the RF image from the TSS/stack to the EFLAGS register.
When calling an event handler, Intel 64 and IA-32 processors establish the value of
the RF flag in the EFLAGS image pushed on the stack:
• For any fault-class exception except a debug exception generated in response to
an instruction breakpoint, the value pushed for RF is 1.
• For any interrupt arriving after any iteration of a repeated string instruction but
the last iteration, the value pushed for RF is 1.
• For any trap-class exception generated by any iteration of a repeated string
instruction but the last iteration, the value pushed for RF is 1.
• For other cases, the value pushed for RF is the value that was in EFLAG.RF at the
time the event handler was called. This includes:
— Debug exceptions generated in response to instruction breakpoints
— Hardware-generated interrupts arriving between instructions (including
those arriving after the last iteration of a repeated string instruction)
— Trap-class exceptions generated after an instruction completes (including
those generated after the last iteration of a repeated string instruction)
— Software-generated interrupts (RF is pushed as 0, since it was cleared at the
start of the software interrupt)
As noted above, the processor does not set the RF flag prior to calling the debug
exception handler for debug exceptions resulting from instruction breakpoints. The
debug exception handler can prevent recurrence of the instruction breakpoint by
setting the RF flag in the EFLAGS image on the stack. If the RF flag in the EFLAGS
image is set when the processor returns from the exception handler, it is copied into
the RF flag in the EFLAGS register by IRETD/IRETQ or a task switch that causes the
return. The processor then ignores instruction breakpoints for the duration of the
next instruction. (Note that the POPF, POPFD, and IRET instructions do not transfer
the RF image into the EFLAGS register.) Setting the RF flag does not prevent other
types of debug-exception conditions (such as, I/O or data breakpoints) from being
detected, nor does it prevent non-debug exceptions from being generated.
For the Pentium processor, when an instruction breakpoint coincides with another
fault-type exception (such as a page fault), the processor may generate one spurious
debug exception after the second exception has been handled, even though the
debug exception handler set the RF flag in the EFLAGS image. To prevent a spurious
exception with Pentium processors, all fault-class exception handlers should set the
RF flag in the EFLAGS image.
Vol. 3 16-11
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
16.3.1.2 Data Memory and I/O Breakpoint Exception Conditions
Data memory and I/O breakpoints are reported when the processor attempts to
access a memory or I/O address specified in a breakpoint-address register (DB0
through DR3) that has been set up to detect data or I/O accesses (R/W flag is set to
1, 2, or 3). The processor generates the exception after it executes the instruction
that made the access, so these breakpoint condition causes a trap-class exception to
be generated.
Because data breakpoints are traps, the original data is overwritten before the trap
exception is generated. If a debugger needs to save the contents of a write break-
point location, it should save the original contents before setting the breakpoint. The
handler can report the saved value after the breakpoint is triggered. The address in
the debug registers can be used to locate the new value stored by the instruction that
triggered the breakpoint.
Intel486 and later processors ignore the GE and LE flags in DR7. In Intel386 proces-
sors, exact data breakpoint matching does not occur unless it is enabled by setting
the LE and/or the GE flags.
P6 family processors are unable to report data breakpoints exactly for the REP MOVS
and REP STOS instructions until the completion of the iteration after the iteration in
which the breakpoint occurred.
For repeated INS and OUTS instructions that generate an I/O-breakpoint debug
exception, the processor generates the exception after the completion of the first
iteration. Repeated INS and OUTS instructions generate a memory-breakpoint debug
exception after the iteration in which the memory address breakpoint location is
accessed.
16.3.1.3 General-Detect Exception Condition
When the GD flag in DR7 is set, the general-detect debug exception occurs when a
program attempts to access any of the debug registers (DR0 through DR7) at the
same time they are being used by another application, such as an emulator or
debugger. This protection feature guarantees full control over the debug registers
when required. The debug exception handler can detect this condition by checking
the state of the BD flag in the DR6 register. The processor generates the exception
before it executes the MOV instruction that accesses a debug register, which causes
a fault-class exception to be generated.
16.3.1.4 Single-Step Exception Condition
The processor generates a single-step debug exception if (while an instruction is
being executed) it detects that the TF flag in the EFLAGS register is set. The excep-
tion is a trap-class exception, because the exception is generated after the instruc-
tion is executed. The processor will not generate this exception after the instruction
that sets the TF flag. For example, if the POPF instruction is used to set the TF flag, a
16-12 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
single-step trap does not occur until after the instruction that follows the POPF
instruction.
The processor clears the TF flag before calling the exception handler. If the TF flag
was set in a TSS at the time of a task switch, the exception occurs after the first
instruction is executed in the new task.
The TF flag normally is not cleared by privilege changes inside a task. The INT n and
INTO instructions, however, do clear this flag. Therefore, software debuggers that
single-step code must recognize and emulate INT n or INTO instructions rather than
executing them directly. To maintain protection, the operating system should check
the CPL after any single-step trap to see if single stepping should continue at the
current privilege level.
The interrupt priorities guarantee that, if an external interrupt occurs, single step-
ping stops. When both an external interrupt and a single-step interrupt occur
together, the single-step interrupt is processed first. This operation clears the TF flag.
After saving the return address or switching tasks, the external interrupt input is
examined before the first instruction of the single-step handler executes. If the
external interrupt is still pending, then it is serviced. The external interrupt handler
does not run in single-step mode. To single step an interrupt handler, single step an
INT n instruction that calls the interrupt handler.
16.3.1.5 Task-Switch Exception Condition
The processor generates a debug exception after a task switch if the T flag of the new
task's TSS is set. This exception is generated after program control has passed to the
new task, and prior to the execution of the first instruction of that task. The exception
handler can detect this condition by examining the BT flag of the DR6 register.
not be set. Failure to observe this rule will put the processor in a loop.
16.3.2
Breakpoint Exception (#BP)—Interrupt Vector 3
The breakpoint exception (interrupt 3) is caused by execution of an INT 3 instruction.
See Chapter 6, “Interrupt 3—Breakpoint Exception (#BP).” Debuggers use break
exceptions in the same way that they use the breakpoint registers; that is, as a
mechanism for suspending program execution to examine registers and memory
locations. With earlier IA-32 processors, breakpoint exceptions are used extensively
for setting instruction breakpoints.
With the Intel386 and later IA-32 processors, it is more convenient to set break-
points with the breakpoint-address registers (DR0 through DR3). However, the
breakpoint exception still is useful for breakpointing debuggers, because a break-
point exception can call a separate exception handler. The breakpoint exception is
also useful when it is necessary to set more breakpoints than there are debug regis-
ters or when breakpoints are being placed in the source code of a program under
development.
Vol. 3 16-13
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
16.4
LAST BRANCH, INTERRUPT, AND EXCEPTION
RECORDING OVERVIEW
P6 family processors introduced the ability to set breakpoints on taken branches,
interrupts, and exceptions, and to single-step from one branch to the next. This
®
®
®
®
®
Intel Atom™ processors to allow logging of branch trace messages in a branch trace
store (BTS) buffer in memory.
See the following sections for processor specific implementation of last branch, inter-
rupt and exception recording:
®
— Section 16.5, “Last Branch, Interrupt, and Exception Recording (Intel
™
®
— Section 16.6, “Last Branch, Interrupt, and Exception Recording (Intel
™
— Section 16.7, “Last Branch, Interrupt, and Exception Recording (Processors
®
™
— Section 16.8, “Last Branch, Interrupt, and Exception Recording (Intel Core
®
™
— Section 16.9, “Last Branch, Interrupt, and Exception Recording (Pentium M
Processors)”
— Section 16.10, “Last Branch, Interrupt, and Exception Recording (P6 Family
Processors)”
The following subsections of Section 16.4 describe common features of profiling
branches. These features are generally enabled using the IA32_DEBUGCTL MSR
(older processor may have implemented a subset or model-specific features, see
definitions of MSR_DEBUGCTLA, MSR_DEBUGCTLB, MSR_DEBUGCTL).
IA32_DEBUGCTL MSR
The IA32_DEBUGCTL MSR provides bit field controls to enable debug trace inter-
rupts, debug trace stores, trace messages enable, single stepping on branches, last
branch record recording, and to control freezing of LBR stack or performance
counters on a PMI request. IA32_DEBUGCTL MSR is located at register address
01D9H.
See Figure 16-3 for the MSR layout and the bullets below for a description of the
flags:
• LBR (last branch/interrupt/exception) flag (bit 0) — When set, the
processor records a running trace of the most recent branches, interrupts, and/or
exceptions taken by the processor (prior to a debug exception being generated)
16-14 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
16.5.1, “LBR Stack”.
•
BTF (single-step on branches) flag (bit 1) — When set, the processor treats
the TF flag in the EFLAGS register as a “single-step on branches” flag rather than
a “single-step on instructions” flag. This mechanism allows single-stepping the
“Single-Stepping on Branches, Exceptions, and Interrupts,” for more information
about the BTF flag.
• TR (trace message enable) flag (bit 6) — When set, branch trace messages
it sends the branch record out on the system bus as a branch trace message
(BTM). See Section 16.4.4, “Branch Trace Messages,” for more information about
the TR flag.
• BTS (branch trace store) flag (bit 7) — When set, the flag enables BTS
facilities to log BTMs to a memory-resident BTS buffer that is part of the DS save
area. See Section 16.4.9, “BTS and DS Save Area.”
• BTINT (branch trace interrupt) flag (bit 8) — When set, the BTS facilities
generate an interrupt when the BTS buffer is full. When clear, BTMs are logged to
the BTS buffer in a circular fashion. See Section 16.4.5, “Branch Trace Store (BTS),”
for a description of this mechanism.
31
14 12
0
11 10 9 8 7 6 5 4 3 2 1
Reserved
FREEZE_WHILE_SMM_EN
FREEZE_PERFMON_ON_PMI
FREEZE_LBRS_ON_PMI
BTS_OFF_USR — BTS off in user code
BTS_OFF_OS — BTS off in OS
BTINT — Branch trace interrupt
BTS — Branch trace store
TR — Trace messages enable
Reserved
BTF — Single-step on branches
LBR — Last branch/interrupt/exception
Figure 16-3. IA32_DEBUGCTL MSR for Processors based
on Intel Core microarchitecture
• BTS_OFF_OS (branch trace off in privileged code) flag (bit 9) — When set,
BTS or BTM is skipped if CPL is 0. See Section 16.7.2.
• BTS_OFF_USR (branch trace off in user code) flag (bit 10) — When set,
BTS or BTM is skipped if CPL is greater than 0. See Section 16.7.2.
Vol. 3 16-15
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
• FREEZE_LBRS_ON_PMI flag (bit 11) — When set, the LBR stack is frozen on a
hardware PMI request (e.g. when a counter overflows and is configured to trigger
PMI).
• FREEZE_PERFMON_ON_PMI flag (bit 12) — When set, a PMI request clears
each of the “ENABLE” field of MSR_PERF_GLOBAL_CTRL MSR (see Figure 30-3) to
disable all the counters.
• FREEZE_WHILE_SMM_EN (bit 14) — If this bit is set, upon the delivery of an
SMI, the processor will clear all the enable bits of IA32_PERF_GLOBAL_CTRL,
save a copy of the content of IA32_DEBUGCTL and disable LBR, BTF, TR, and BTS
fields of IA32_DEBUGCTL before transferring control to the SMI handler. Subse-
quently, the enable bits of IA32_PERF_GLOBAL_CTRL will be set to 1, the saved
copy of IA32_DEBUGCTL prior to SMI delivery will be restored, after the SMI
handler issues RSM to complete its service. Note that system software must
check IA32_DEBUGCTL. to determine if the processor supports the
FREEZE_WHILE_SMM_EN control bit. FREEZE_WHILE_SMM_EN is supported if
IA32_PERF_CAPABILITIES.FREEZE_WHILE_SMM[Bit 12] is reporting 1. See
Section 30.11 for details of detecting the presence of IA32_PERF_CAPABILITIES
MSR.
16.4.2
Monitoring Branches, Exceptions, and Interrupts
When the LBR flag (bit 0) in the IA32_DEBUGCTL MSR is set, the processor automat-
ically begins recording branch records for taken branches, interrupts, and exceptions
(except for debug exceptions) in the LBR stack MSRs.
When the processor generates a a debug exception (#DB), it automatically clears the
LBR flag before executing the exception handler. This action does not clear previously
stored LBR stack MSRs. The branch record for the last four taken branches, interrupts
and/or exceptions are retained for analysis.
A debugger can use the linear addresses in the LBR stack to re-set breakpoints in the
breakpoint address registers (DR0 through DR3). This allows a backward trace from
the manifestation of a particular bug toward its source.
If the LBR flag is cleared and TR flag in the IA32_DEBUGCTL MSR remains set, the
processor will continue to update LBR stack MSRs. This is because BTM information
must be generated from entries in the LBR stack. A #DB does not automatically clear
the TR flag.
16.4.3
Single-Stepping on Branches, Exceptions, and Interrupts
When software sets both the BTF flag (bit 1) in the IA32_DEBUGCTL MSR and the TF
flag in the EFLAGS register, the processor generates a single-step debug exception
the next time it takes a branch, services an interrupt, or generates an exception. This
mechanism allows the debugger to single-step on control transfers caused by
branches, interrupts, and exceptions. This “control-flow single stepping” helps isolate
16-16 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
a bug to a particular block of code before instruction single-stepping further narrows
the search. If the BTF flag is set when the processor generates a debug exception,
the processor clears the BTF flag along with the TF flag. The debugger must reset the
BTF and TF flags before resuming program execution to continue control-flow single
stepping.
16.4.4
Branch Trace Messages
Setting the TR flag (bit 6) in the IA32_DEBUGCTL MSR enables branch trace
interrupt, it sends a branch record out on the system bus as a BTM. A debugging
device that is monitoring the system bus can read these messages and synchronize
operations with taken branch, interrupt, and exception events.
When interrupts or exceptions occur in conjunction with a taken branch, additional
BTMs are sent out on the bus, as described in Section 16.4.2, “Monitoring Branches,
Exceptions, and Interrupts.”
Unlike the P6 family and Core family processors, the Pentium 4, Atom, and Intel Xeon
processors can collect branch records in the LBR stack MSRs while at the same time
sending/storing BTMs when both the TR and LBR flags are set in the IA32_DEBUGCTL
MSR (in the case of Pentium 4, processor, MSR_DEBUGCTLA).
16.4.5
Branch Trace Store (BTS)
A trace of taken branches, interrupts, and exceptions is useful for debugging code by
providing a method of determining the decision path taken to reach a particular code
location. The LBR flag (bit 0) of IA32_DEBUGCTL provides a mechanism for capturing
records of taken branches, interrupts, and exceptions and saving them in the last
branch record (LBR) stack MSRs, setting the TR flag for sending them out onto the
system bus as BTMs. The branch trace store (BTS) mechanism provides the addi-
tional capability of saving the branch records in a memory-resident BTS buffer, which
is part of the DS save area. The BTS buffer can be configured to be circular so that
the most recent branch records are always available or it can be configured to
generate an interrupt when the buffer is nearly full so that all the branch records can
be saved. The BTINT flag (bit 8) can be used to enable the generation of interrupt
when the BTS buffer is full. See Section 16.4.9.2, “Setting Up the DS Save Area.” for
additional details.
Setting this flag (BTS) alone can greatly reduce the performance of the processor.
CPL-qualified branch trace storing mechanism can help mitigate the performance
impact of sending/logging branch trace messages.
Vol. 3 16-17
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
16.4.6
CPL-Qualified Branch Trace Mechanism
CPL-qualified branch trace mechanism is available to a subset of Intel 64 and IA-32
processors that support the branch trace storing mechanism. The processor supports
the CPL-qualified branch trace mechanism if CPUID.01H:ECX[bit 4] = 1.
The CPL-qualified branch trace mechanism is described in Section 16.4.9.4. System
software can selectively specify CPL qualification to not send/store Branch Trace
Messages associated with a specified privilege level. Two bit fields, BTS_OFF_USR
(bit 10) and BTS_OFF_OS (bit 9), are provided in the debug control register to
specify the CPL of BTMs that will not be logged in the BTS buffer or sent on the bus.
16.4.7
Freezing LBR and Performance Counters on PMI
Many issues may generate a performance monitoring interrupt (PMI); a PMI service
handler will need to determine cause to handle the situation. Two capabilities that
allow a PMI service routine to improve branch tracing and performance monitoring
are:
• Freezing LBRs on PMI (bit 11)— The processor freezes LBRs on a PMI request
by clearing the LBR bit (bit 0) in IA32_DEBUGCTL. Software must then re-enable
IA32_DEBUGCTL.[0] to continue monitoring branches. When using this feature,
software should be careful about writes to IA32_DEBUGCTL to avoid re-enabling
LBRs by accident if they were just disabled.
• Freezing PMCs on PMI (bit 12) — The processor freezes the performance
counters on a PMI request by clearing the MSR_PERF_GLOBAL_CTRL MSR (see
Figure 30-3). The PMCs affected include both general-purpose counters and
fixed-function counters (see Section 30.4.1, “Fixed-function Performance
Counters”). Software must re-enable counts by writing 1s to the corresponding
enable bits in MSR_PERF_GLOBAL_CTRL before leaving a PMI service routine to
continue counter operation.
Freezing LBRs and PMCs on PMIs occur when:
• A performance counter had an overflow and was programmed to signal a PMI in
case of an overflow.
— For the general-purpose counters; this is done by setting bit 20 of the
IA32_PERFEVTSELx register.
— For the fixed-function counters; this is done by setting the 3rd bit in the
corresponding 4-bit control field of the MSR_PERF_FIXED_CTR_CTRL register
(see Figure 30-1) or IA32_FIXED_CTR_CTRL MSR (see Figure 30-2).
• The PEBS buffer is almost full and reaches the interrupt threshold.
• The BTS buffer is almost full and reaches the interrupt threshold.
16-18 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
16.4.8
LBR Stack
The last branch record stack and top-of-stack (TOS) pointer MSRs are supported
across Intel 64 and IA-32 processor families. However, the number of MSRs in the
LBR stack and the valid range of TOS pointer value can vary between different
processor families. Table 16-3 lists the LBR stack size and TOS pointer range for
several processor families according to the CPUID signatures of Display-
Family/DisplayModel encoding (see CPUID instruction in Chapter 3 of Intel® 64 and
IA-32 Architectures Software Developer’s Manual, Volume 2A).
Table 16-3. LBR Stack Size and TOS Pointer Range
DisplayFamily_DisplayModel Size of LBR Stack
Range of TOS Pointer
06_1AH, 06_1EH, 06_1FH, 16
06_2EH
0 to 15
06_17H, 06_1DH
06_0FH
4
4
8
0 to 3
0 to 3
0 to 7
06_1CH
The last branch recording mechanism tracks not only branch instructions (like JMP,
the instruction pointer (like external interrupts, traps and faults). The branch
recording mechanisms generally employs a set of MSRs, referred to as last branch
record (LRB) stack. The size and exact locations of the LRB stack are generally
model-specific.
• Last Branch Record (LBR) Stack — The LBR consists of N pairs of MSRs (N is
listed in the LBR stack size column of Table 16-3) that store source and
destination address of recent branches (see Figure 16-3):
— MSR_LASTBRANCH_0_FROM_IP (address is model specific) through the next
consecutive (N-1) MSR address store source addresses
consecutive (N-1) MSR address store destination addresses.
• Last Branch Record Top-of-Stack (TOS) Pointer — The lowest significant M
bits of the TOS Pointer MSR (MSR_LASTBRANCH_TOS, address is model specific)
contains an M-bit pointer to the MSR in the LBR stack that contains the most
recent branch, interrupt, or exception recorded. The valid range of the M-bit POS
pointer is given in Table 16-3.
Vol. 3 16-19
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
16.4.8.1 LBR Stack and Intel® 64 Processors
LBR MSRs are 64-bits. If IA-32e mode is disabled, only the lower 32-bits of the
address is recorded. If IA-32e mode is enabled, the processor writes 64-bit values
into the MSR.
In 64-bit mode, last branch records store 64-bit addresses; in compatibility mode,
the upper 32-bits of last branch records are cleared.
MSR_LASTBRANCH_0_FROM_IP through MSR_LASTBRANCH_(N-1)_FROM_IP
63
0
Source Address
MSR_LASTBRANCH_0_TO_IP through MSR_LASTBRANCH_(N-1)_TO_IP
0
63
Destination Address
Figure 16-4. 64-bit Address Layout of LBR MSR
Software should query an architectural MSR IA32_PERF_CAPABILITIES[5:0]
about the format of the address that is stored in the LBR stack. Four formats are
defined by the following encoding:
— 000000B (32-bit record format) — Stores 32-bit offset in current CS of
respective source/destination,
— 000001B (64-bit LIP record format) — Stores 64-bit linear address of
respective source/destination,
— 000010B (64-bit EIP record format) — Stores 64-bit offset (effective
address) of respective source/destination.
— 000011B (64-bit EIP record format) and Flags — Stores 64-bit offset
(effective address) of respective source/destination. LBR flags are supported
in the upper bits of ‘FROM’ register in the LBR stack. See LBR stack details
below for flag support and definition.
Processor’s support for the architectural MSR IA32_PERF_CAPABILITIES is
provided by CPUID.01H:ECX[PERF_CAPAB_MSR] (bit 15).
16.4.8.2 LBR Stack and IA-32 Processors
The LBR MSRs in IA-32 processors introduced prior to Intel 64 architecture store the
32-bit “To Linear Address” and “From Linear Address“ using the high and low half of
each 64-bit MSR.
16-20 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
16.4.8.3 Last Exception Records and Intel 64 Architecture
Intel 64 and IA-32 processors also provide MSRs that store the branch record for the
last branch taken prior to an exception or an interrupt. The location of the last excep-
tion record (LER) MSRs are model specific. The MSRs that store last exception
records are 64-bits. If IA-32e mode is disabled, only the lower 32-bits of the address
is recorded. If IA-32e mode is enabled, the processor writes 64-bit values into the
MSR. In 64-bit mode, last exception records store 64-bit addresses; in compatibility
mode, the upper 32-bits of last exception records are cleared.
16.4.9
BTS and DS Save Area
The Debug store (DS) feature flag (bit 21), returned by CPUID.1:EDX[21] Indicates
that the processor provides the debug store (DS) mechanism. This mechanism
allows BTMs to be stored in a memory-resident BTS buffer. See Section 16.4.5,
“Branch Trace Store (BTS).” Precise event-based sampling (PEBS, see Section
30.4.4, “Precise Event Based Sampling (PEBS),”) also uses the DS save area
provided by debug store mechanism. When CPUID.1:EDX[21] is set, the following
BTS facilities are available:
• The BTS_UNAVAILABLE flag in the IA32_MISC_ENABLE MSR indicates (when
clear) the availability of the BTS facilities, including the ability to set the BTS and
BTINT bits in the MSR_DEBUGCTLA MSR.
• The IA32_DS_AREA MSR can be programmed to point to the DS save area.
The debug store (DS) save area is a software-designated area of memory that is
used to collect the following two types of information:
• Branch records — When the BTS flag in the IA32_DEBUGCTL MSR is set, a
branch record is stored in the BTS buffer in the DS save area whenever a taken
branch, interrupt, or exception is detected.
• PEBS records — When a performance counter is configured for PEBS, a PEBS
record is stored in the PEBS buffer in the DS save area after the counter overflow
occurs. This record contains the architectural state of the processor (state of the
8 general purpose registers, EIP register, and EFLAGS register) at the next
occurrence of the PEBS event that caused the counter to overflow. When the
state information has been logged, the counter is automatically reset to a
preselected value, and event counting begins again. This feature is available only
for a subset of the performance events on processors that support PEBS.
NOTES
DS save area and recording mechanism is not available in the SMM.
The feature is disabled on transition to the SMM mode. Similarly DS
recording is disabled on the generation of a machine check exception
Vol. 3 16-21
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
and is cleared on processor RESET and INIT. DS recording is available
in real address mode.
The BTS and PEBS facilities may not be available on all processors.
The availability of these facilities is indicated by the
BTS_UNAVAILABLE and PEBS_UNAVAILABLE flags, respectively, in
the IA32_MISC_ENABLE MSR (see Appendix B).
The DS save area is divided into three parts (see Figure 16-5): buffer management
area, branch trace store (BTS) buffer, and PEBS buffer. The buffer management area
is used to define the location and size of the BTS and PEBS buffers. The processor
then uses the buffer management area to keep track of the branch and/or PEBS
records in their respective buffers and to record the performance counter reset value.
The linear address of the first byte of the DS buffer management area is specified
with the IA32_DS_AREA MSR.
The fields in the buffer management area are as follows:
• BTS buffer base — Linear address of the first byte of the BTS buffer. This
address should point to a natural doubleword boundary.
• BTS index — Linear address of the first byte of the next BTS record to be written
to. Initially, this address should be the same as the address in the BTS buffer
base field.
• BTS absolute maximum — Linear address of the next byte past the end of the
BTS buffer. This address should be a multiple of the BTS record size (12 bytes)
plus 1.
• BTS interrupt threshold — Linear address of the BTS record on which an
interrupt is to be generated. This address must point to an offset from the BTS
buffer base that is a multiple of the BTS record size. Also, it must be several
records short of the BTS absolute maximum address to allow a pending interrupt
to be handled prior to processor writing the BTS absolute maximum record.
• PEBS buffer base — Linear address of the first byte of the PEBS buffer. This
address should point to a natural doubleword boundary.
• PEBS index — Linear address of the first byte of the next PEBS record to be
written to. Initially, this address should be the same as the address in the PEBS
buffer base field.
16-22 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
IA32_DS_AREA MSR
DS Buffer Management Area
BTS Buffer
BTS Buffer Base
BTS Index
0H
Branch Record 0
4H
8H
BTS Absolute
Maximum
BTS Interrupt
Threshold
Branch Record 1
CH
PEBS Buffer Base
PEBS Index
10H
14H
18H
1CH
PEBS Absolute
Maximum
PEBS Interrupt
Threshold
Branch Record n
20H
24H
30H
PEBS
Counter Reset
PEBS Buffer
Reserved
PEBS Record 0
PEBS Record 1
PEBS Record n
Figure 16-5. DS Save Area
• PEBS absolute maximum — Linear address of the next byte past the end of the
PEBS buffer. This address should be a multiple of the PEBS record size (40 bytes)
plus 1.
• PEBS interrupt threshold — Linear address of the PEBS record on which an
interrupt is to be generated. This address must point to an offset from the PEBS
buffer base that is a multiple of the PEBS record size. Also, it must be several
records short of the PEBS absolute maximum address to allow a pending
interrupt to be handled prior to processor writing the PEBS absolute maximum
record.
Vol. 3 16-23
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
• PEBS counter reset value — A 40-bit value that the counter is to be reset to
after state information has collected following counter overflow. This value allows
state information to be collected after a preset number of events have been
counted.
Figures 16-6 shows the structure of a 12-byte branch record in the BTS buffer. The
fields in each record are as follows:
• Last branch from — Linear address of the instruction from which the branch,
interrupt, or exception was taken.
• Last branch to — Linear address of the branch target or the first instruction in
the interrupt or exception service routine.
• Branch predicted — Bit 4 of field indicates whether the branch that was taken
was predicted (set) or not predicted (clear).
31
4
0
Last Branch From
Last Branch To
0H
4H
8H
Branch Predicted
Figure 16-6. 32-bit Branch Trace Record Format
Figures 16-7 shows the structure of the 40-byte PEBS records. Nominally the register
values are those at the beginning of the instruction that caused the event. However,
there are cases where the registers may be logged in a partially modified state. The
linear IP field shows the value in the EIP register translated from an offset into the
current code segment to a linear address.
16-24 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
31
0
EFLAGS
Linear IP
EAX
0H
4H
8H
EBX
CH
ECX
10H
14H
18H
1CH
EDX
ESI
EDI
EBP
20H
24H
ESP
Figure 16-7. PEBS Record Format
16.4.9.1 DS Save Area and IA-32e Mode Operation
When IA-32e mode is active (IA32_EFER.LMA = 1), the structure of the DS save area
is shown in Figure 16-8. The organization of each field in IA-32e mode operation is
similar to that of non-IA-32e mode operation. However, each field now stores a
64-bit address. The IA32_DS_AREA MSR holds the 64-bit linear address of the first
byte of the DS buffer management area.
Vol. 3 16-25
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
IA32_DS_AREA MSR
DS Buffer Management Area
BTS Buffer
BTS Buffer Base
BTS Index
0H
Branch Record 0
8H
BTS Absolute
Maximum
BTS Interrupt
Threshold
10H
Branch Record 1
18H
PEBS Buffer Base
PEBS Index
20H
28H
30H
38H
PEBS Absolute
Maximum
PEBS Interrupt
Threshold
Branch Record n
40H
48H
50H
PEBS
Counter Reset
PEBS Buffer
Reserved
PEBS Record 0
PEBS Record 1
PEBS Record n
Figure 16-8. IA-32e Mode DS Save Area
When IA-32e mode is active, the structure of a branch trace record is similar to that
shown in Figure 16-6, but each field is 8 bytes in length. This makes each BTS record
24 bytes (see Figure 16-9). The structure of a PEBS record is similar to that shown in
Figure 16-7, but each field is 8 bytes in length and architectural states include
register R8 through R15. This makes the size of a PEBS record in 64-bit mode 144
bytes (see Figure 16-10).
16-26 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
63
4
0
Last Branch From
Last Branch To
0H
8H
10H
Branch Predicted
Figure 16-9. 64-bit Branch Trace Record Format
63
0
RFLAGS
0H
8H
RIP
RAX
RBX
RCX
RDX
RSI
RDI
RBP
RSP
R8
10H
18H
20H
28H
30H
38H
40H
48H
50H
...
...
R15
88H
Figure 16-10. 64-bit PEBS Record Format
Fields in the buffer management area of a DS save area are described in Section
16.4.9.
The format of a branch trace record and a PEBS record are the same as the 64-bit
record formats shown in Figures 16-9 and Figures 16-10, with the exception that the
branch predicted bit is not supported by Intel Core microarchitecture or Intel Atom
microarchitecture. The 64-bit record formats for BTS and PEBS apply to DS save area
for all operating modes.
Vol. 3 16-27
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
The procedures used to program IA32_DEBUG_CTRL MSR to set up a BTS buffer or a
CPL-qualified BTS are described in Section 16.4.9.3 and Section 16.4.9.4.
Required elements for writing a DS interrupt service routine are largely the same on
processors that support using DS Save area for BTS or PEBS records. However, on
®
processors based on Intel NetBurst microarchitecture, re-enabling counting
requires writing to CCCRs. But a DS interrupt service routine on processors based on
Intel Core or Intel Atom microarchitecture should:
• Re-enable the enable bits in IA32_PERF_GLOBAL_CTRL MSR if it is servicing an
overflow PMI due to PEBS.
• Clear overflow indications by writing to IA32_PERF_GLOBAL_OVF_CTRL when a
counting configuration is changed. This includes bit 62 (ClrOvfBuffer) and the
overflow indication of counters used in either PEBS or general-purpose counting
(specifically: bits 0 or 1; see Figures 30-3).
16.4.9.2 Setting Up the DS Save Area
memory as described in the following procedure (See Section 30.4.4.1, “Setting up
the PEBS Buffer,” for instructions for setting up a PEBS buffer, respectively, in the DS
save area):
16.4.9, “BTS and DS Save Area,” and Section 16.4.9.1, “DS Save Area and IA-
32e Mode Operation”). Also see the additional notes in this section.
2. Write the base linear address of the DS buffer management area into the
IA32_DS_AREA MSR.
3. Set up the performance counter entry in the xAPIC LVT for fixed delivery and
edge sensitive. See Section 10.6.1, “Local Vector Table.”
4. Establish an interrupt handler in the IDT for the vector associated with the
performance counter entry in the xAPIC LVT.
5. Write an interrupt service routine to handle the interrupt. See Section 16.4.9.5,
“Writing the DS Interrupt Service Routine.”
The following restrictions should be applied to the DS save area.
• The three DS save area sections should be allocated from a non-paged pool, and
marked accessed and dirty. It is the responsibility of the operating system to
keep the pages that contain the buffer present and to mark them accessed and
dirty. The implication is that the operating system cannot do “lazy” page-table
entry propagation for these pages.
• The DS save area can be larger than a page, but the pages must be mapped to
contiguous linear addresses. The buffer may share a page, so it need not be
aligned on a 4-KByte boundary. For performance reasons, the base of the buffer
must be aligned on a doubleword boundary and should be aligned on a cache line
boundary.
16-28 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
• It is recommended that the buffer size for the BTS buffer and the PEBS buffer be
an integer multiple of the corresponding record sizes.
• The precise event records buffer should be large enough to hold the number of
precise event records that can occur while waiting for the interrupt to be
serviced.
• The DS save area should be in kernel space. It must not be on the same page as
code, to avoid triggering self-modifying code actions.
• There are no memory type restrictions on the buffers, although it is
recommended that the buffers be designated as WB memory type for
performance considerations.
• Either the system must be prevented from entering A20M mode while DS save
area is active, or bit 20 of all addresses within buffer bounds must be 0.
• Pages that contain buffers must be mapped to the same physical addresses for all
processes, such that any change to control register CR3 will not change the DS
addresses.
• The DS save area is expected to used only on systems with an enabled APIC. The
gate instead of the trap gate.
16.4.9.3 Setting Up the BTS Buffer
Three flags in the MSR_DEBUGCTLA MSR (see Table 16-4), IA32_DEBUGCTL (see
Figure 16-3), or MSR_DEBUGCTLB (see Figure 16-16) control the generation of
branch records and storing of them in the BTS buffer; these are TR, BTS, and BTINT.
The TR flag enables the generation of BTMs. The BTS flag determines whether the
BTMs are sent out on the system bus (clear) or stored in the BTS buffer (set). BTMs
cannot be simultaneously sent to the system bus and logged in the BTS buffer. The
BTINT flag enables the generation of an interrupt when the BTS buffer is full. When
this flag is clear, the BTS buffer is a circular buffer.
Table 16-4. IA32_DEBUGCTL Flag Encodings
TR
0
BTS
X
BTINT
Description
X
X
0
1
Branch trace messages (BTMs) off
Generate BTMs
1
0
1
1
Store BTMs in the BTS buffer, used here as a circular buffer
1
1
Store BTMs in the BTS buffer, and generate an interrupt when
the buffer is nearly full
The following procedure describes how to set up a DS Save area to collect branch
records in the BTS buffer:
1. Place values in the BTS buffer base, BTS index, BTS absolute maximum, and BTS
interrupt threshold fields of the DS buffer management area to set up the BTS
buffer in memory.
Vol. 3 16-29
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
2. Set the TR and BTS flags in the IA32_DEBUGCTL for Intel Core Solo and Intel
Core Duo processors or later processors (or MSR_DEBUGCTLA MSR for
processors based on Intel NetBurst Microarchitecture; or MSR_DEBUGCTLB for
Pentium M processors).
3. Clear the BTINT flag in the corresponding IA32_DEBUGCTL (or MSR_DEBUGCTLA
MSR; or MSR_DEBUGCTLB) if a circular BTS buffer is desired.
NOTES
If the buffer size is set to less than the minimum allowable value (i.e.
BTS absolute maximum < 1 + size of BTS record), the results of BTS
is undefined.
In order to prevent generating an interrupt, when working with
circular BTS buffer, SW need to set BTS interrupt threshold to a value
greater than BTS absolute maximum (fields of the DS buffer
management area). It's not enough to clear the BTINT flag itself only.
16.4.9.4 Setting Up CPL-Qualified BTS
If the processor supports CPL-qualified last branch recording mechanism, the gener-
ation of branch records and storing of them in the BTS buffer are determined by: TR,
BTS, BTS_OFF_OS, BTS_OFF_USR, and BTINT. The encoding of these five bits are
shown in Table 16-5.
Table 16-5. CPL-Qualified Branch Trace Store Encodings
TR
BTS
BTS_OFF_OS BTS_OFF_USR
BTINT
Description
0
X
X
X
0
1
0
1
0
X
X
0
0
1
1
0
X
Branch trace messages (BTMs)
off
1
1
1
1
1
1
0
1
1
1
1
1
X
0
0
0
X
1
Generates BTMs but do not
store BTMs
Store all BTMs in the BTS buffer,
used here as a circular buffer
Store BTMs with CPL > 0 in the
BTS buffer
Store BTMs with CPL = 0 in the
BTS buffer
Generate BTMs but do not store
BTMs
Store all BTMs in the BTS buffer;
generate an interrupt when the
buffer is nearly full
16-30 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
Table 16-5. CPL-Qualified Branch Trace Store Encodings (Contd.)
TR
BTS
BTS_OFF_OS BTS_OFF_USR
BTINT
Description
1
1
1
0
1
Store BTMs with CPL > 0 in the
BTS buffer; generate an
interrupt when the buffer is
nearly full
1
1
0
1
1
Store BTMs with CPL = 0 in the
BTS buffer; generate an
interrupt when the buffer is
nearly full
16.4.9.5 Writing the DS Interrupt Service Routine
The BTS, non-precise event-based sampling, and PEBS facilities share the same
interrupt vector and interrupt service routine (called the debug store interrupt
service routine or DS ISR). To handle BTS, non-precise event-based sampling, and
PEBS interrupts: separate handler routines must be included in the DS ISR. Use the
following guidelines when writing a DS ISR to handle BTS, non-precise event-based
sampling, and/or PEBS interrupts.
• The DS interrupt service routine (ISR) must be part of a kernel driver and operate
at a current privilege level of 0 to secure the buffer storage area.
• Because the BTS, non-precise event-based sampling, and PEBS facilities share
the same interrupt vector, the DS ISR must check for all the possible causes of
interrupts from these facilities and pass control on to the appropriate handler.
BTS and PEBS buffer overflow would be the sources of the interrupt if the buffer
index matches/exceeds the interrupt threshold specified. Detection of non-
precise event-based sampling as the source of the interrupt is accomplished by
checking for counter overflow.
• There must be separate save areas, buffers, and state for each processor in an
MP system.
• Upon entering the ISR, branch trace messages and PEBS should be disabled to
prevent race conditions during access to the DS save area. This is done by
clearing TR flag in the IA32_DEBUGCTL (or MSR_DEBUGCTLA MSR) and by
clearing the precise event enable flag in the MSR_PEBS_ENABLE MSR. These
settings should be restored to their original values when exiting the ISR.
• The processor will not disable the DS save area when the buffer is full and the
circular mode has not been selected. The current DS setting must be retained
and restored by the ISR on exit.
• After reading the data in the appropriate buffer, up to but not including the
current index into the buffer, the ISR must reset the buffer index to the beginning
of the buffer. Otherwise, everything up to the index will look like new entries upon
the next invocation of the ISR.
Vol. 3 16-31
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
• The ISR must clear the mask bit in the performance counter LVT entry.
• The ISR must re-enable the counters to count via
IA32_PERF_GLOBAL_CTRL/IA32_PERF_GLOBAL_OVF_CTRL if it is servicing an
overflow PMI due to PEBS (or via CCCR's ENABLE bit on processor based on Intel
NetBurst microarchitecture).
• The Pentium 4 Processor and Intel Xeon Processor mask PMIs upon receiving an
interrupt. Clear this condition before leaving the interrupt handler.
16.5
LAST BRANCH, INTERRUPT, AND EXCEPTION
RECORDING (INTEL® CORE™2 DUO AND INTEL®
ATOM™ PROCESSOR FAMILY)
The Intel Core 2 Duo processor family and Intel Xeon processors based on Intel Core
microarchitecture or enhanced Intel Core microarchitecture provide last branch
Intel Atom processor family. These capabilities are similar to those found in Pentium
4 processors, including support for the following facilities:
provide bit fields for software to configure mechanisms related to debug trace,
branch recording, branch trace store, and performance counter operations. See
• Last branch record (LBR) stack — There are a collection of MSR pairs that
store the source and destination addresses related to recently executed
branches. See Section 16.5.1.
LBR stack on a PMI request is available.
FREEZE_LBRS_ON_PMI flag is set.
• Branch trace messages — See Section 16.4.4.
• Last exception records — See Section 16.7.3.
• FREEZE_LBRS_ON_PMI flag (bit 11) — see Section 16.4.7.
• FREEZE_PERFMON_ON_PMI flag (bit 12) — see Section 16.4.7.
• FREEZE_WHILE_SMM_EN (bit 14) — FREEZE_WHILE_SMM_EN is supported
if IA32_PERF_CAPABILITIES.FREEZE_WHILE_SMM[Bit 12] is reporting 1. See
Section 16.4.1.
16-32 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
16.5.1
LBR Stack
The last branch record stack and top-of-stack (TOS) pointer MSRs are supported
across Intel Core 2, Intel Xeon and Intel Atom processor families. Four pair of MSRs
are supported in the LBR stack
• Last Branch Record (LBR) Stack
— MSR_LASTBRANCH_0_FROM_IP (address 40H) through
MSR_LASTBRANCH_3_FROM_IP (address 43H) store source addresses
— MSR_LASTBRANCH_0_TO_IP (address 60H) through
MSR_LASTBRANCH_3_To_IP (address 63H) store destination addresses.
• Last Branch Record Top-of-Stack (TOS) Pointer — The lowest significant 2
bits of the TOS Pointer MSR (MSR_LASTBRANCH_TOS, address 1C9H) contains a
pointer to the MSR in the LBR stack that contains the most recent branch,
interrupt, or exception recorded.
For compatibility, the MSR_LER_TO_LIP and the MSR_LER_FROM_LIP MSRs) dupli-
cate functions of the LastExceptionToIP and LastExceptionFromIP MSRs found in P6
family processors.
16.6
LAST BRANCH, INTERRUPT, AND EXCEPTION
RECORDING (INTEL® CORE™I7 PROCESSOR FAMILY)
The Intel Core i7 processor family and Intel Xeon processors based on Intel microar-
capabilities are similar to those found in Intel Core 2 processors and adds additional
capabilities:
provides bit fields for software to configure mechanisms related to debug trace,
Section 16.4.1 for a description of the flags. See Figure 16-11 for the MSR layout.
• Last branch record (LBR) stack — There are 16 MSR pairs that store the
source and destination addresses related to recently executed branches. See
Section 16.6.1.
• Monitoring and single-stepping of branches, exceptions, and interrupts
LBR stack on a PMI request is available.
• Branch trace messages — The IA32_DEBUGCTL MSR provides bit fields for
software to enable each logical processor to generate branch trace messages.
See Section 16.4.4. However, not all BTM messages are observable using the
®
Intel QPI link.
• Last exception records — See Section 16.7.3.
Vol. 3 16-33
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
• Branch trace store and CPL-qualified BTS — See Section 16.4.6 and Section
16.4.5.
• FREEZE_LBRS_ON_PMI flag (bit 11) — see Section 16.4.7.
• FREEZE_PERFMON_ON_PMI flag (bit 12) — see Section 16.4.7.
if IA32_PERF_CAPABILITIES.FREEZE_WHILE_SMM[Bit 12] is reporting 1. See
Section 16.4.1.
Processors based on Intel microarchitecture (Nehalem) provide additional capabili-
ties:
• Independent control of uncore PMI — The IA32_DEBUGCTL MSR provides a
bit field (see Figure 16-11) for software to enable each logical processor to
receive an uncore counter overflow interrupt.
• LBR filtering — Processors based on Intel microarchitecture (Nehalem) support
filtering of LBR based on combination of CPL and branch type conditions. When
LBR filtering is enabled, the LBR stack only captures the subset of branches that
are specified by MSR_LBR_SELECT.
31
14 1312
0
11 10 9 8 7 6 5 4 3 2 1
Reserved
FREEZE_WHILE_SMM_EN
UNCORE_PMI_EN
FREEZE_PERFMON_ON_PMI
FREEZE_LBRS_ON_PMI
BTS_OFF_USR — BTS off in user code
BTS_OFF_OS — BTS off in OS
BTINT — Branch trace interrupt
BTS — Branch trace store
TR — Trace messages enable
Reserved
BTF — Single-step on branches
LBR — Last branch/interrupt/exception
Figure 16-11. IA32_DEBUGCTL MSR for Processors based
16.6.1
LBR Stack
Processors based on Intel microarchitecture (Nehalem) provide 16 pairs of MSR to
record last branch record information. The layout of each MSR pair is shown in
Table 16-6 and Table 16-7.
16-34 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
Table 16-6. IA32_LASTBRACH_x_FROM_IP
Bit Field
Data
Bit Offset Access
Description
47:0
R/O
The linear address of the branch instruction itself,
This is the “branch from“ address
SIGN_EXt
MISPRED
62:48
63
R/0
R/O
Signed extension of bit 47 of this register
When set, indicates the branch was predicted;
otherwise, the branch was mispredicted.
Table 16-7. IA32_LASTBRACH_x_TO_IP
Bit Field
Data
Bit Offset Access
Description
47:0
R/O
The linear address of the target of the branch
instruction itself, This is the “branch to“ address
SIGN_EXt
63:48
R/0
Signed extension of bit 47 of this register
Processors based on Intel microarchitecture (Nehalem) have an LBR MSR Stack as
shown in Table 16-8.
Table 16-8. LBR Stack Size and TOS Pointer Range
DisplayFamily_DisplayModel Size of LBR Stack
Range of TOS Pointer
06_1AH
16
0 to 15
16.6.2
Filtering of Last Branch Records
MSR_LBR_SELECT is cleared to zero at RESET, and LBR filtering is disabled, i.e. all
branches will be captured. MSR_LBR_SELECT provides bit fields to specify the condi-
tions of subsets of branches that will not be captured in the LBR. The layout of
MSR_LBR_SELECT is shown in Table 16-9.
Table 16-9. MSR_LBR_SELECT
Bit Field
Bit Offset Access
Description
CPL_EQ_0
CPL_NEQ_0
0
1
R/W
R/W
When set, do not capture branches occurring in ring 0
When set, do not capture branches occurring in ring
>0
Vol. 3 16-35
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
Table 16-9. MSR_LBR_SELECT (Contd.)
Bit Field
Bit Offset Access
Description
JCC
2
R/W
R/W
R/W
R/W
R/W
R/W
R/W
When set, do not capture conditional branches
When set, do not capture near relative calls
When set, do not capture near indirect calls
When set, do not capture near returns
When set, do not capture near indirect jumps
When set, do not capture near relative jumps
When set, do not capture far branches
Must be zero
NEAR_REL_CALL
NEAR_IND_CALL
NEAR_RET
3
4
5
NEAR_IND_JMP
NEAR_REL_JMP
FAR_BRANCH
Reserved
6
7
8
63:9
16.7
LAST BRANCH, INTERRUPT, AND EXCEPTION
RECORDING (PROCESSORS BASED ON INTEL
NETBURST® MICROARCHITECTURE)
Pentium 4 and Intel Xeon processors based on Intel NetBurst microarchitecture
provide the following methods for recording taken branches, interrupts and excep-
tions:
• Store branch records in the last branch record (LBR) stack MSRs for the most
recent taken branches, interrupts, and/or exceptions in MSRs. A branch record
consist of a branch-from and a branch-to instruction address.
• Send the branch records out on the system bus as branch trace messages
(BTMs).
• Log BTMs in a memory-resident branch trace store (BTS) buffer.
To support these functions, the processor provides the following MSRs and related
facilities:
• MSR_DEBUGCTLA MSR — Enables last branch, interrupt, and exception
recording; single-stepping on taken branches; branch trace messages (BTMs);
and branch trace store (BTS). This register is named DebugCtlMSR in the P6
family processors.
• Debug store (DS) feature flag (CPUID.1:EDX.DS[bit 21]) — Indicates that
the processor provides the debug store (DS) mechanism, which allows BTMs to
be stored in a memory-resident BTS buffer.
• CPL-qualified debug store (DS) feature flag (CPUID.1:ECX.DS-CPL[bit
4]) — Indicates that the processor provides a CPL-qualified debug store (DS)
mechanism, which allows software to selectively skip sending and storing BTMs,
according to specified current privilege level settings, into a memory-resident
BTS buffer.
16-36 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
• IA32_MISC_ENABLE MSR — Indicates that the processor provides the BTS
facilities.
• Last branch record (LBR) stack — The LBR stack is a circular stack that
consists of four MSRs (MSR_LASTBRANCH_0 through MSR_LASTBRANCH_3) for
the Pentium 4 and Intel Xeon processor family [CPUID family 0FH, models 0H-
02H]. The LBR stack consists of 16 MSR pairs (MSR_LASTBRANCH_0_FROM_LIP
through MSR_LASTBRANCH_15_FROM_LIP and MSR_LASTBRANCH_0_TO_LIP
through MSR_LASTBRANCH_15_TO_LIP) for the Pentium 4 and Intel Xeon
processor family [CPUID family 0FH, model 03H].
• Last branch record top-of-stack (TOS) pointer — The TOS Pointer MSR
contains a 2-bit pointer (0-3) to the MSR in the LBR stack that contains the most
recent branch, interrupt, or exception recorded for the Pentium 4 and Intel Xeon
processor family [CPUID family 0FH, models 0H-02H]. This pointer becomes a
4-bit pointer (0-15) for the Pentium 4 and Intel Xeon processor family [CPUID
family 0FH, model 03H]. See also: Table 16-10, Figure 16-12, and Section
16.7.2, “LBR Stack for Processors Based on Intel NetBurst Microarchitecture.”
• Last exception record — See Section 16.7.3, “Last Exception Records.”
16.7.1
MSR_DEBUGCTLA MSR
The MSR_DEBUGCTLA MSR enables and disables the various last branch recording
mechanisms described in the previous section. This register can be written to using
the WRMSR instruction, when operating at privilege level 0 or when in real-address
mode. A protected-mode operating system procedure is required to provide user
access to this register. Figure 16-12 shows the flags in the MSR_DEBUGCTLA MSR.
The functions of these flags are as follows:
processor records a running trace of the most recent branches, interrupts, and/or
exceptions taken by the processor (prior to a debug exception being generated)
in the last branch record (LBR) stack. Each branch, interrupt, or exception is
recorded as a 64-bit branch record. The processor clears this flag whenever a
debug exception is generated (for example, when an instruction or data
Processors Based on Intel NetBurst Microarchitecture.”
• BTF (single-step on branches) flag (bit 1) — When set, the processor treats
the TF flag in the EFLAGS register as a “single-step on branches” flag rather than
processor on taken branches, interrupts, and exceptions. See Section 16.4.3,
“Single-Stepping on Branches, Exceptions, and Interrupts.”
• TR (trace message enable) flag (bit 2) — When set, branch trace messages
are enabled. Thereafter, when the processor detects a taken branch, interrupt, or
exception, it sends the branch record out on the system bus as a branch trace
message (BTM). See Section 16.4.4, “Branch Trace Messages.”
Vol. 3 16-37
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
31
7
6
5 4 3 2 1 0
Reserved
BTS_OFF_USR — Disable storing non-CPL_0 BTS
BTS_OFF_OS — Disable storing CPL_0 BTS
BTINT — Branch trace interrupt
BTS — Branch trace store
TR — Trace messages enable
BTF — Single-step on branches
LBR — Last branch/interrupt/exception
Figure 16-12. MSR_DEBUGCTLA MSR for Pentium 4 and Intel Xeon Processors
• BTS (branch trace store) flag (bit 3) — When set, enables the BTS facilities to
log BTMs to a memory-resident BTS buffer that is part of the DS save area. See
generate an interrupt when the BTS buffer is full. When clear, BTMs are logged to
the BTS buffer in a circular fashion. See Section 16.4.5, “Branch Trace Store (BTS).”
resident BTS buffer. See Section 16.7.2, “LBR Stack for Processors Based on Intel
NetBurst Microarchitecture.”
• BTS_OFF_USR (disable ring 0 branch trace store) flag (bit 6) — When set,
enables the BTS facilities to skip sending/logging non-CPL_0 BTMs to the
memory-resident BTS buffer. See Section 16.7.2, “LBR Stack for Processors
Based on Intel NetBurst Microarchitecture.”
The initial implementation of BTS_OFF_USR and BTS_OFF_OS in
MSR_DEBUGCTLA is shown in Figure 16-12. The BTS_OFF_USR and
BTS_OFF_OS fields may be implemented on other model-specific
debug control register at different locations.
See Appendix B, “Model-Specific Registers (MSRs),” for a detailed description of each
of the last branch recording MSRs.
16.7.2
LBR Stack for Processors Based on Intel NetBurst
Microarchitecture
The LBR stack is made up of LBR MSRs that are treated by the processor as a circular
stack. The TOS pointer (MSR_LASTBRANCH_TOS MSR) points to the LBR MSR (or
16-38 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
LBR MSR pair) that contains the most recent (last) branch record placed on the stack.
Prior to placing a new branch record on the stack, the TOS is incremented by 1. When
the TOS pointer reaches it maximum value, it wraps around to 0. See Table 16-10
and Figure 16-12.
®
Table 16-10. LBR MSR Stack Size and TOS Pointer Range for the Pentium 4 and the
®
®
Intel Xeon Processor Family
DisplayFamily_DisplayModel Size of LBR Stack Range of TOS Pointer
Family 0FH, Models 0H-02H;
MSRs at locations 1DBH-
1DEH.
4
0 to 3
Family 0FH, Models; MSRs at
locations 680H-68FH.
16
0 to 15
0 to 15
at locations 6C0H-6CFH.
The registers in the LBR MSR stack and the MSR_LASTBRANCH_TOS MSR are read-
only and can be read using the RDMSR instruction.
Figure 16-13 shows the layout of a branch record in an LBR MSR (or MSR pair). Each
branch record consists of two linear addresses, which represent the “from” and “to”
instruction pointers for a branch, interrupt, or exception. The contents of the from
and to addresses differ, depending on the source of the branch:
• Taken branch — If the record is for a taken branch, the “from” address is the
address of the branch instruction and the “to” address is the target instruction of
the branch.
• Interrupt — If the record is for an interrupt, the “from” address the return
instruction pointer (RIP) saved for the interrupt and the “to” address is the
address of the first instruction in the interrupt handler routine. The RIP is the
linear address of the next instruction to be executed upon returning from the
interrupt handler.
• Exception — If the record is for an exception, the “from” address is the linear
address of the instruction that caused the exception to be generated and the “to”
address is the address of the first instruction in the exception handler routine.
Vol. 3 16-39
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
CPUID Family 0FH, Models 0H-02H
MSR_LASTBRANCH_0 through MSR_LASTBRANCH_3
0
63
32 - 31
From Linear Address
To Linear Address
CPUID Family 0FH, Model 03H-04H
MSR_LASTBRANCH_0_FROM_LIP through MSR_LASTBRANCH_15_FROM_LIP
0
63
32 - 31
Reserved
From Linear Address
MSR_LASTBRANCH_0_TO_LIP through MSR_LASTBRANCH_15_TO_LIP
0
63
32 - 31
Reserved
To Linear Address
Figure 16-13. LBR MSR Branch Record Layout for the Pentium 4
and Intel Xeon Processor Family
Additional information is saved if an exception or interrupt occurs in conjunction with
a branch instruction. If a branch instruction generates a trap type exception, two
branch records are stored in the LBR stack: a branch record for the branch instruction
followed by a branch record for the exception.
If a branch instruction is immediately followed by an interrupt, a branch record is
stored in the LBR stack for the branch instruction followed by a record for the
interrupt.
16.7.3
Last Exception Records
®
®
®
The Pentium 4, Intel Xeon, Pentium M, Intel Core™ Solo, Intel Core™ Duo, Intel
®
®
Core™2 Duo, Intel Core™ i7 and Intel Atom™ processors provide two MSRs (the
MSR_LER_TO_LIP and the MSR_LER_FROM_LIP MSRs) that duplicate the functions
of the LastExceptionToIP and LastExceptionFromIP MSRs found in the P6 family
processors. The MSR_LER_TO_LIP and MSR_LER_FROM_LIP MSRs contain a branch
record for the last branch that the processor took prior to an exception or interrupt
being generated.
16-40 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
16.8
LAST BRANCH, INTERRUPT, AND EXCEPTION
RECORDING (INTEL® CORE™ SOLO AND INTEL®
CORE™DUO PROCESSORS)
Intel Core Solo and Intel Core Duo processors provide last branch interrupt and
exception recording. This capability is almost identical to that found in Pentium 4 and
Intel Xeon processors. There are differences in the stack and in some MSR names
and locations.
Note the following:
• IA32_DEBUGCTL MSR — Enables debug trace interrupt, debug trace store,
trace messages enable, performance monitoring breakpoint flags, single
stepping on branches, and last branch. IA32_DEBUGCTL MSR is located at
register address 01D9H.
See Figure 16-14 for the layout and the entries below for a description of the
flags:
— LBR (last branch/interrupt/exception) flag (bit 0) — When set, the
processor records a running trace of the most recent branches, interrupts,
and/or exceptions taken by the processor (prior to a debug exception being
generated) in the last branch record (LBR) stack. For more information, see
— BTF (single-step on branches) flag (bit 1) — When set, the processor
treats the TF flag in the EFLAGS register as a “single-step on branches” flag
rather than a “single-step on instructions” flag. This mechanism allows
single-stepping the processor on taken branches, interrupts, and exceptions.
See Section 16.4.3, “Single-Stepping on Branches, Exceptions, and Inter-
rupts,” for more information about the BTF flag.
— TR (trace message enable) flag (bit 6) — When set, branch trace
messages are enabled. When the processor detects a taken branch,
interrupt, or exception; it sends the branch record out on the system bus as
a branch trace message (BTM). See Section 16.4.4, “Branch Trace Messages,”
for more information about the TR flag.
facilities to log BTMs to a memory-resident BTS buffer that is part of the DS
save area. See Section 16.4.9, “BTS and DS Save Area.”
— BTINT (branch trace interrupt) flag (bits 8) — When set, the BTS
facilities generate an interrupt when the BTS buffer is full. When clear, BTMs are
logged to the BTS buffer in a circular fashion. See Section 16.4.5, “Branch Trace
Store (BTS),” for a description of this mechanism.
Vol. 3 16-41
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
31
8 7 6 5 4 3 2 1
0
Reserved
BTINT — Branch trace interrupt
BTS — Branch trace store
TR — Trace messages enable
Reserved
BTF — Single-step on branches
LBR — Last branch/interrupt/exception
Figure 16-14. IA32_DEBUGCTL MSR for Intel Core Solo
• Debug store (DS) feature flag (bit 21), returned by the CPUID
instruction — Indicates that the processor provides the debug store (DS)
mechanism, which allows BTMs to be stored in a memory-resident BTS buffer.
See Section 16.4.5, “Branch Trace Store (BTS).”
• Last Branch Record (LBR) Stack — The LBR stack consists of 8 MSRs
(MSR_LASTBRANCH_0 through MSR_LASTBRANCH_7); bits 31-0 hold the ‘from’
address, bits 63-32 hold the ‘to’ address (MSR addresses start at 40H). See
Figure 16-15.
• Last Branch Record Top-of-Stack (TOS) Pointer — The TOS Pointer MSR
contains a 3-bit pointer (bits 2-0) to the MSR in the LBR stack that contains the
most recent branch, interrupt, or exception recorded. For Intel Core Solo and
For compatibility, the Intel Core Solo and Intel Core Duo processors provide two 32-
bit MSRs (the MSR_LER_TO_LIP and the MSR_LER_FROM_LIP MSRs) that duplicate
functions of the LastExceptionToIP and LastExceptionFromIP MSRs found in P6 family
processors.
For details, see Section 16.7, “Last Branch, Interrupt, and Exception Recording
®
(Processors based on Intel NetBurst Microarchitecture),” and Appendix B.6, “MSRs
®
™
®
™
In Intel Core Solo and Intel Core Duo Processors.”
16-42 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
MSR_LASTBRANCH_0 through MSR_LASTBRANCH_7
0
63
32 - 31
To Linear Address
From Linear Address
Figure 16-15. LBR Branch Record Layout for the Intel Core Solo
and Intel Core Duo Processor
16.9
LAST BRANCH, INTERRUPT, AND EXCEPTION
RECORDING (PENTIUM M PROCESSORS)
Like the Pentium 4 and Intel Xeon processor family, Pentium M processors provide
cally to that found in Pentium 4 and Intel Xeon processors. There are differences in
the shape of the stack and in some MSR names and locations. Note the following:
• MSR_DEBUGCTLB MSR — Enables debug trace interrupt, debug trace store,
trace messages enable, performance monitoring breakpoint flags, single
stepping on branches, and last branch. For Pentium M processors, this MSR is
located at register address 01D9H. See Figure 16-16 and the entries below for a
description of the flags.
— LBR (last branch/interrupt/exception) flag (bit 0) — When set, the
processor records a running trace of the most recent branches, interrupts,
and/or exceptions taken by the processor (prior to a debug exception being
generated) in the last branch record (LBR) stack. For more information, see
the “Last Branch Record (LBR) Stack” bullet below.
— BTF (single-step on branches) flag (bit 1) — When set, the processor
treats the TF flag in the EFLAGS register as a “single-step on branches” flag
rather than a “single-step on instructions” flag. This mechanism allows
single-stepping the processor on taken branches, interrupts, and exceptions.
See Section 16.4.3, “Single-Stepping on Branches, Exceptions, and Inter-
rupts,” for more information about the BTF flag.
— PBi (performance monitoring/breakpoint pins) flags (bits 5-2) —
When these flags are set, the performance monitoring/breakpoint pins on the
processor (BP0#, BP1#, BP2#, and BP3#) report breakpoint matches in the
corresponding breakpoint-address registers (DR0 through DR3). The
processor asserts then deasserts the corresponding BPi# pin when a
breakpoint match occurs. When a PBi flag is clear, the performance
monitoring/breakpoint pins report performance events. Processor execution
is not affected by reporting performance events.
Vol. 3 16-43
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
— TR (trace message enable) flag (bit 6) — When set, branch trace
messages are enabled. When the processor detects a taken branch,
interrupt, or exception, it sends the branch record out on the system bus as a
branch trace message (BTM). See Section 16.4.4, “Branch Trace Messages,”
for more information about the TR flag.
facilities to log BTMs to a memory-resident BTS buffer that is part of the DS
save area. See Section 16.4.9, “BTS and DS Save Area.”
— BTINT (branch trace interrupt) flag (bits 8) — When set, the BTS
facilities generate an interrupt when the BTS buffer is full. When clear, BTMs are
logged to the BTS buffer in a circular fashion. See Section 16.4.5, “Branch Trace
Store (BTS),” for a description of this mechanism.
31
8 7 6 5 4 3 2 1
0
Reserved
BTINT — Branch trace interrupt
BTS — Branch trace store
TR — Trace messages enable
PB3/2/1/0 — Performance monitoring breakpoint flags
BTF — Single-step on branches
LBR — Last branch/interrupt/exception
• Debug store (DS) feature flag (bit 21), returned by the CPUID
mechanism, which allows BTMs to be stored in a memory-resident BTS buffer.
See Section 16.4.5, “Branch Trace Store (BTS).”
• Last Branch Record (LBR) Stack — The LBR stack consists of 8 MSRs
(MSR_LASTBRANCH_0 through MSR_LASTBRANCH_7); bits 31-0 hold the ‘from’
address, bits 63-32 hold the ‘to’ address. For Pentium M Processors, these pairs
are located at register addresses 040H-047H. See Figure 16-17.
• Last Branch Record Top-of-Stack (TOS) Pointer — The TOS Pointer MSR
contains a 3-bit pointer (bits 2-0) to the MSR in the LBR stack that contains the
most recent branch, interrupt, or exception recorded. For Pentium M Processors,
this MSR is located at register address 01C9H.
16-44 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
MSR_LASTBRANCH_0
through MSR_LASTBRANCH_7
0
63
32 - 31
To Linear Address
From Linear Address
Figure 16-17. LBR Branch Record Layout for the Pentium M Processor
For more detail on these capabilities, see Section 16.7.3, “Last Exception Records,”
and Appendix B.7, “MSRs In the Pentium M Processor.”
16.10 LAST BRANCH, INTERRUPT, AND EXCEPTION
RECORDING (P6 FAMILY PROCESSORS)
The P6 family processors provide five MSRs for recording the last branch, interrupt,
or exception taken by the processor: DEBUGCTLMSR, LastBranchToIP, LastBranch-
FromIP, LastExceptionToIP, and LastExceptionFromIP. These registers can be used to
collect last branch records, to set breakpoints on branches, interrupts, and excep-
tions, and to single-step from one branch to the next.
See Appendix B, “Model-Specific Registers (MSRs),” for a detailed description of each
of the last branch recording MSRs.
The version of the DEBUGCTLMSR register found in the P6 family processors enables
last branch, interrupt, and exception recording; taken branch breakpoints; the
breakpoint reporting pins; and trace messages. This register can be written to using
the WRMSR instruction, when operating at privilege level 0 or when in real-address
mode. A protected-mode operating system procedure is required to provide user
access to this register. Figure 16-18 shows the flags in the DEBUGCTLMSR register
for the P6 family processors. The functions of these flags are as follows:
• LBR (last branch/interrupt/exception) flag (bit 0) — When set, the
processor records the source and target addresses (in the LastBranchToIP,
LastBranchFromIP, LastExceptionToIP, and LastExceptionFromIP MSRs) for the
last branch and the last exception or interrupt taken by the processor prior to a
debug exception being generated. The processor clears this flag whenever a
debug exception, such as an instruction or data breakpoint or single-step trap
occurs.
Vol. 3 16-45
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
31
7 6 5 4 3 2 1 0
P
B
3
P
B
2
P
B
1
P
B
0
B
T
F
L
B
R
T
R
Reserved
TR — Trace messages enable
PBi — Performance monitoring/breakpoint pins
BTF — Single-step on branches
LBR — Last branch/interrupt/exception
Figure 16-18. DEBUGCTLMSR Register (P6 Family Processors)
• BTF (single-step on branches) flag (bit 1) — When set, the processor treats
the TF flag in the EFLAGS register as a “single-step on branches” flag. See
Section 16.4.3, “Single-Stepping on Branches, Exceptions, and Interrupts.”
• PBi (performance monitoring/breakpoint pins) flags (bits 2 through 5)
— When these flags are set, the performance monitoring/breakpoint pins on the
processor (BP0#, BP1#, BP2#, and BP3#) report breakpoint matches in the
corresponding breakpoint-address registers (DR0 through DR3). The processor
asserts then deasserts the corresponding BPi# pin when a breakpoint match
occurs. When a PBi flag is clear, the performance monitoring/breakpoint pins
report performance events. Processor execution is not affected by reporting
performance events.
• TR (trace message enable) flag (bit 6) — When set, trace messages are
enabled as described in Section 16.4.4, “Branch Trace Messages.” Setting this
flag greatly reduces the performance of the processor. When trace messages are
enabled, the values stored in the LastBranchToIP, LastBranchFromIP, LastExcep-
tionToIP, and LastExceptionFromIP MSRs are undefined.
16.10.2 Last Branch and Last Exception MSRs
The LastBranchToIP and LastBranchFromIP MSRs are 32-bit registers for recording
the instruction pointers for the last branch, interrupt, or exception that the processor
took prior to a debug exception being generated. When a branch occurs, the
processor loads the address of the branch instruction into the LastBranchFromIP MSR
and loads the target address for the branch into the LastBranchToIP MSR.
When an interrupt or exception occurs (other than a debug exception), the address
of the instruction that was interrupted by the exception or interrupt is loaded into the
LastBranchFromIP MSR and the address of the exception or interrupt handler that is
called is loaded into the LastBranchToIP MSR.
The LastExceptionToIP and LastExceptionFromIP MSRs (also 32-bit registers) record
the instruction pointers for the last branch that the processor took prior to an excep-
16-46 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
tion or interrupt being generated. When an exception or interrupt occurs, the
contents of the LastBranchToIP and LastBranchFromIP MSRs are copied into these
registers before the to and from addresses of the exception or interrupt are recorded
in the LastBranchToIP and LastBranchFromIP MSRs.
These registers can be read using the RDMSR instruction.
Note that the values stored in the LastBranchToIP, LastBranchFromIP, LastException-
ToIP, and LastExceptionFromIP MSRs are offsets into the current code segment, as
opposed to linear addresses, which are saved in last branch records for the Pentium
4 and Intel Xeon processors.
16.10.3 Monitoring Branches, Exceptions, and Interrupts
When the LBR flag in the DEBUGCTLMSR register is set, the processor automatically
begins recording branches that it takes, exceptions that are generated (except for
debug exceptions), and interrupts that are serviced. Each time a branch, exception,
or interrupt occurs, the processor records the to and from instruction pointers in the
LastBranchToIP and LastBranchFromIP MSRs. In addition, for interrupts and excep-
tions, the processor copies the contents of the LastBranchToIP and LastBranch-
FromIP MSRs into the LastExceptionToIP and LastExceptionFromIP MSRs prior to
recording the to and from addresses of the interrupt or exception.
When the processor generates a debug exception (#DB), it automatically clears the
LBR flag before executing the exception handler, but does not touch the last branch
and last exception MSRs. The addresses for the last branch, interrupt, or exception
taken are thus retained in the LastBranchToIP and LastBranchFromIP MSRs and the
addresses of the last branch prior to an interrupt or exception are retained in the
LastExceptionToIP, and LastExceptionFromIP MSRs.
The debugger can use the last branch, interrupt, and/or exception addresses in
combination with code-segment selectors retrieved from the stack to reset break-
points in the breakpoint-address registers (DR0 through DR3), allowing a backward
trace from the manifestation of a particular bug toward its source. Because the
instruction pointers recorded in the LastBranchToIP, LastBranchFromIP, LastExcepti-
onToIP, and LastExceptionFromIP MSRs are offsets into a code segment, software
must determine the segment base address of the code segment associated with the
control transfer to calculate the linear address to be placed in the breakpoint-address
registers. The segment base address can be determined by reading the segment
selector for the code segment from the stack and using it to locate the segment
descriptor for the segment in the GDT or LDT. The segment base address can then be
read from the segment descriptor.
Before resuming program execution from a debug-exception handler, the handler
must set the LBR flag again to re-enable last branch and last exception/interrupt
recording.
Vol. 3 16-47
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
16.11 TIME-STAMP COUNTER
The Intel 64 and IA-32 architectures (beginning with the Pentium processor) define a
time-stamp counter mechanism that can be used to monitor and identify the relative
time occurrence of processor events. The counter’s architecture includes the
following components:
• TSC flag — A feature bit that indicates the availability of the time-stamp counter.
The counter is available in an if the function CPUID.1:EDX.TSC[bit 4] = 1.
• IA32_TIME_STAMP_COUNTER MSR (called TSC MSR in P6 family and
Pentium processors) — The MSR used as the counter.
• RDTSC instruction — An instruction used to read the time-stamp counter.
• TSD flag — A control register flag is used to enable or disable the time-stamp
counter (enabled if CR4.TSD[bit 2] = 1).
The time-stamp counter (as implemented in the P6 family, Pentium, Pentium M,
Pentium 4, Intel Xeon, Intel Core Solo and Intel Core Duo processors and later
processors) is a 64-bit counter that is set to 0 following a RESET of the processor.
Following a RESET, the counter increments even when the processor is halted by the
HLT instruction or the external STPCLK# pin. Note that the assertion of the external
DPSLP# pin may cause the time-stamp counter to stop.
Processor families increment the time-stamp counter differently:
• For Pentium M processors (family [06H], models [09H, 0DH]); for Pentium 4
processors, Intel Xeon processors (family [0FH], models [00H, 01H, or 02H]);
and for P6 family processors: the time-stamp counter increments with every
internal processor clock cycle.
The internal processor clock cycle is determined by the current core-clock to bus-
clock ratio. Intel® SpeedStep® technology transitions may also impact the
processor clock.
• For Pentium 4 processors, Intel Xeon processors (family [0FH], models [03H and
higher]); for Intel Core Solo and Intel Core Duo processors (family [06H], model
[0EH]); for the Intel Xeon processor 5100 series and Intel Core 2 Duo processors
(family [06H], model [0FH]); for Intel Core 2 and Intel Xeon processors (family
[06H], display_model [17H]); for Intel Atom processors (family [06H],
display_model [1CH]): the time-stamp counter increments at a constant rate.
That rate may be set by the maximum core-clock to bus-clock ratio of the
processor or may be set by the maximum resolved frequency at which the
processor is booted. The maximum resolved frequency may differ from the
maximum qualified frequency of the processor, see Section 30.10.5 for more
detail.
The specific processor configuration determines the behavior. Constant TSC
behavior ensures that the duration of each clock tick is uniform and supports the
use of the TSC as a wall clock timer even if the processor core changes frequency.
This is the architectural behavior moving forward.
16-48 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
NOTE
To determine average processor clock frequency, Intel recommends
the use of EMON logic to count processor core clocks over the period
of time for which the average is required. See Section 30.10,
“Counting Clocks,” and Appendix A, “Performance-
Monitoring Events,” for more information.
The RDTSC instruction reads the time-stamp counter and is guaranteed to return a
monotonically increasing unique value whenever executed, except for a 64-bit
counter wraparound. Intel guarantees that the time-stamp counter will not wrap-
around within 10 years after being reset. The period for counter wrap is longer for
Pentium 4, Intel Xeon, P6 family, and Pentium processors.
Normally, the RDTSC instruction can be executed by programs and procedures
running at any privilege level and in virtual-8086 mode. The TSD flag allows use of
this instruction to be restricted to programs and procedures running at privilege level
0. A secure operating system would set the TSD flag during system initialization to
disable user access to the time-stamp counter. An operating system that disables
user access to the time-stamp counter should emulate the instruction through a
user-accessible programming interface.
The RDTSC instruction is not serializing or ordered with other instructions. It does not
necessarily wait until all previous instructions have been executed before reading the
counter. Similarly, subsequent instructions may begin execution before the RDTSC
instruction operation is performed.
The RDMSR and WRMSR instructions read and write the time-stamp counter, treating
the time-stamp counter as an ordinary MSR (address 10H). In the Pentium 4, Intel
Xeon, and P6 family processors, all 64-bits of the time-stamp counter are read using
RDMSR (just as with RDTSC). When WRMSR is used to write the time-stamp counter
on processors before family [0FH], models [03H, 04H]: only the low-order 32-bits of
the time-stamp counter can be written (the high-order 32 bits are cleared to 0). For
family [0FH], models [03H, 04H, 06H]; for family [06H]], model [0EH, 0FH]; for
family [06H]], display_model [17H, 1AH, 1CH, 1DH]: all 64 bits are writable.
16.11.1 Invariant TSC
The time stamp counter in newer processors may support an enhancement, referred
to as invariant TSC. Processor’s support for invariant TSC is indicated by
CPUID.80000007H:EDX[8].
The invariant TSC will run at a constant rate in all ACPI P-, C-. and T-states. This is
the architectural behavior moving forward. On processors with invariant TSC
support, the OS may use the TSC for wall clock timer services (instead of ACPI or
HPET timers). TSC reads are much more efficient and do not incur the overhead
associated with a ring transition or access to a platform resource.
Vol. 3 16-49
Download from Www.Somanuals.com. All Manuals Search And Download.
DEBUGGING, PROFILING BRANCHES AND TIME-STAMP COUNTER
16.11.2 IA32_TSC_AUX Register and RDTSCP Support
Processor based on Intel microarchitecture (Nehalem) provides an auxiliary TSC
register, IA32_TSC_AUX that is designed to be used in conjunction with IA32_TSC.
IA32_TSC_AUX provides a 32-bit field that is initialized by privileged software with a
signature value (for example, a logical processor ID).
The primary usage of IA32_TSC_AUX in conjunction with IA32_TSC is to allow soft-
ware to read the 64-bit time stamp in IA32_TSC and signature value in
IA32_TSC_AUX with the instruction RDTSCP in an atomic operation. RDTSCP returns
the 64-bit time stamp in EDX:EAX and the 32-bit TSC_AUX signature value in ECX.
The atomicity of RDTSCP ensures that no context switch can occur between the reads
of the TSC and TSC_AUX values.
Support for RDTSCP is indicated by CPUID.80000001H:EDX[27]. As with RDTSC
instruction, non-ring 0 access is controlled by CR4.TSD (Time Stamp Disable flag).
User mode software can use RDTSCP to detect if CPU migration has occurred
between successive reads of the TSC. It can also be used to adjust for per-CPU differ-
ences in TSC values in a NUMA system.
16-50 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
CHAPTER 17
8086 EMULATION
execute new or legacy programs that are assembled and/or compiled to run on an
Intel 8086 processor:
• Real-address mode.
• Virtual-8086 mode.
Figure 2-3 shows the relationship of these operating modes to protected mode and
system management mode (SMM).
When the processor is powered up or reset, it is placed in the real-address mode.
This operating mode almost exactly duplicates the execution environment of the
Intel 8086 processor, with some extensions. Virtually any program assembled and/or
compiled to run on an Intel 8086 processor will run on an IA-32 processor in this
mode.
When running in protected mode, the processor can be switched to virtual-8086
mode to run 8086 programs. This mode also duplicates the execution environment of
the Intel 8086 processor, with extensions. In virtual-8086 mode, an 8086 program
runs as a separate protected-mode task. Legacy 8086 programs are thus able to run
under an operating system (such as Microsoft Windows*) that takes advantage of
protected mode and to use protected-mode facilities, such as the protected-mode
interrupt- and exception-handling facilities. Protected-mode multitasking permits
multiple virtual-8086 mode tasks (with each task running a separate 8086 program)
to be run on the processor along with other non-virtual-8086 mode tasks.
This section describes both the basic real-address mode execution environment and
the virtual-8086-mode execution environment, available on the IA-32 processors
beginning with the Intel386 processor.
17.1
REAL-ADDRESS MODE
The IA-32 architecture’s real-address mode runs programs written for the Intel 8086,
Intel 8088, Intel 80186, and Intel 80188 processors, or for the real-address mode of
the Intel 286, Intel386, Intel486, Pentium, P6 family, Pentium 4, and Intel Xeon
processors.
The execution environment of the processor in real-address mode is designed to
duplicate the execution environment of the Intel 8086 processor. To an 8086
program, a processor operating in real-address mode behaves like a high-speed
8086 processor. The principal features of this architecture are defined in Chapter 3,
“Basic Execution Environment”, of the Intel® 64 and IA-32 Architectures Software
Developer’s Manual, Volume 1.
Vol. 3 17-1
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
The following is a summary of the core features of the real-address mode execution
environment as would be seen by a program written for the 8086:
• The processor supports a nominal 1-MByte physical address space (see Section
address space is divided into segments, each of which can be up to 64 KBytes in
length. The base of a segment is specified with a 16-bit segment selector, which
is zero extended to form a 20-bit offset from address 0 in the address space. An
operand within a segment is addressed with a 16-bit offset from the base of the
segment. A physical address is thus formed by adding the offset to the 20-bit
segment base (see Section 17.1.1, “Address Translation in Real-Address Mode”).
• All operands in “native 8086 code” are 8-bit or 16-bit values. (Operand size
override prefixes can be used to access 32-bit operands.)
• Eight 16-bit general-purpose registers are provided: AX, BX, CX, DX, SP, BP, SI,
and DI. The extended 32 bit registers (EAX, EBX, ECX, EDX, ESP, EBP, ESI, and
EDI) are accessible to programs that explicitly perform a size override operation.
• Four segment registers are provided: CS, DS, SS, and ES. (The FS and GS
registers are accessible to programs that explicitly access them.) The CS register
contains the segment selector for the code segment; the DS and ES registers
contain segment selectors for data segments; and the SS register contains the
segment selector for the stack segment.
register. Note this register is a 32-bit register and unintentional address wrapping
may occur.
• The 16-bit FLAGS register contains status and control flags. (This register is
mapped to the 16 least significant bits of the 32-bit EFLAGS register.)
• All of the Intel 8086 instructions are supported (see Section 17.1.3, “Instructions
Supported in Real-Address Mode”).
• A single, 16-bit-wide stack is provided for handling procedure calls and
invocations of interrupt and exception handlers. This stack is contained in the
stack segment identified with the SS register. The SP (stack pointer) register
contains an offset into the stack segment. The stack grows down (toward lower
segment offsets) from the stack pointer. The BP (base pointer) register also
contains an offset into the stack segment that can be used as a pointer to a
parameter list. When a CALL instruction is executed, the processor pushes the
current instruction pointer (the 16 least-significant bits of the EIP register and,
on far calls, the current value of the CS register) onto the stack. On a return,
initiated with a RET instruction, the processor pops the saved instruction pointer
from the stack into the EIP register (and CS register on far returns). When an
implicit call to an interrupt or exception handler is executed, the processor
pushes the EIP, CS, and EFLAGS (low-order 16-bits only) registers onto the
stack. On a return from an interrupt or exception handler, initiated with an IRET
instruction, the processor pops the saved instruction pointer and EFLAGS image
from the stack into the EIP, CS, and EFLAGS registers.
17-2 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
provided for handling interrupts and exceptions (see Figure 17-2). The interrupt
table (which has 4-byte entries) takes the place of the interrupt descriptor table
(IDT, with 8-byte entries) used when handling protected-mode interrupts and
exceptions. Interrupt and exception vector numbers provide an index to entries
in the interrupt table. Each entry provides a pointer (called a “vector”) to an
interrupt- or exception-handling procedure. See Section 17.1.4, “Interrupt and
Exception Handling”, for more details. It is possible for software to relocate the
IDT by means of the LIDT instruction on IA-32 processors beginning with the
Intel386 processor.
• The x87 FPU is active and available to execute x87 FPU instructions in real-
address mode. Programs written to run on the Intel 8087 and Intel 287 math
coprocessors can be run in real-address mode without modification.
The following extensions to the Intel 8086 execution environment are available in the
Intel 8086 processors is required, these features should not be used in new programs
written to run in real-address mode.
• Two additional segment registers (FS and GS) are available.
• Many of the integer and system instructions that have been added to later IA-32
processors can be executed in real-address mode (see Section 17.1.3, “Instruc-
tions Supported in Real-Address Mode”).
• The 32-bit operand prefix can be used in real-address mode programs to execute
the 32-bit forms of instructions. This prefix also allows real-address mode
programs to use the processor’s 32-bit general-purpose registers.
• The 32-bit address prefix can be used in real-address mode programs, allowing
32-bit offsets.
The following sections describe address formation, registers, available instructions,
and interrupt and exception handling in real-address mode. For information on I/O in
real-address mode, see Chapter 13, “Input/Output”, of the Intel® 64 and IA-32
Architectures Software Developer’s Manual, Volume 1.
17.1.1
Address Translation in Real-Address Mode
In real-address mode, the processor does not interpret segment selectors as indexes
into a descriptor table; instead, it uses them directly to form linear addresses as the
8086 processor does. It shifts the segment selector left by 4 bits to form a 20-bit
base address (see Figure 17-1). The offset into a segment is added to the base
address to create a linear address that maps directly to the physical address space.
When using 8086-style address translation, it is possible to specify addresses larger
than 1 MByte. For example, with a segment selector value of FFFFH and an offset of
FFFFH, the linear (and physical) address would be 10FFEFH (1 megabyte plus 64
KBytes). The 8086 processor, which can form addresses only up to 20 bits long, trun-
cates the high-order bit, thereby “wrapping” this address to FFEFH. When operating
Vol. 3 17-3
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
in real-address mode, however, the processor does not truncate such an address and
uses it as a physical address. (Note, however, that for IA-32 processors beginning
with the Intel486 processor, the A20M# signal can be used in real-address mode to
mask address line A20, thereby mimicking the 20-bit wrap-around behavior of the
8086 processor.) Care should be take to ensure that A20M# based address wrapping
is handled correctly in multiprocessor based system.
19
4
3
0
0
Base
16-bit Segment Selector
0 0 0 0
+
Offset
=
19
16 15
0 0 0 0
16-bit Effective Address
19
0
Linear
Address
20-bit Linear Address
Figure 17-1. Real-Address Mode Address Translation
The IA-32 processors beginning with the Intel386 processor can generate 32-bit
offsets using an address override prefix; however, in real-address mode, the value of
a 32-bit offset may not exceed FFFFH without causing an exception.
For full compatibility with Intel 286 real-address mode, pseudo-protection faults
(interrupt 12 or 13) occur if a 32-bit offset is generated outside the range 0 through
FFFFH.
17.1.2
Registers Supported in Real-Address Mode
The register set available in real-address mode includes all the registers defined for
the 8086 processor plus the new registers introduced in later IA-32 processors, such
as the FS and GS segment registers, the debug registers, the control registers, and
the floating-point unit registers. The 32-bit operand prefix allows a real-address
mode program to use the 32-bit general-purpose registers (EAX, EBX, ECX, EDX,
ESP, EBP, ESI, and EDI).
17.1.3
Instructions Supported in Real-Address Mode
The following instructions make up the core instruction set for the 8086 processor. If
backwards compatibility to the Intel 286 and Intel 8086 processors is required, only
these instructions should be used in a new program written to run in real-address
mode.
17-4 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
• Move (MOV) instructions that move operands between general-purpose
registers, segment registers, and between memory and general-purpose
registers.
• The exchange (XCHG) instruction.
• Load segment register instructions LDS and LES.
• Arithmetic instructions ADD, ADC, SUB, SBB, MUL, IMUL, DIV, IDIV, INC, DEC,
CMP, and NEG.
• Logical instructions AND, OR, XOR, and NOT.
• Decimal instructions DAA, DAS, AAA, AAS, AAM, and AAD.
• Stack instructions PUSH and POP (to general-purpose registers and segment
registers).
• Type conversion instructions CWD, CDQ, CBW, and CWDE.
• Shift and rotate instructions SAL, SHL, SHR, SAR, ROL, ROR, RCL, and RCR.
• TEST instruction.
• Control instructions JMP, Jcc, CALL, RET, LOOP, LOOPE, and LOOPNE.
• Interrupt instructions INT n, INTO, and IRET.
• EFLAGS control instructions STC, CLC, CMC, CLD, STD, LAHF, SAHF, PUSHF, and
POPF.
• I/O instructions IN, INS, OUT, and OUTS.
• Load effective address (LEA) instruction, and translate (XLATB) instruction.
• LOCK prefix.
• Repeat prefixes REP, REPE, REPZ, REPNE, and REPNZ.
• Processor halt (HLT) instruction.
• No operation (NOP) instruction.
The following instructions, added to later IA-32 processors (some in the Intel 286
processor and the remainder in the Intel386 processor), can be executed in real-
address mode, if backwards compatibility to the Intel 8086 processor is not required.
• Move (MOV) instructions that operate on the control and debug registers.
• Load segment register instructions LSS, LFS, and LGS.
• Generalized multiply instructions and multiply immediate data.
• Shift and rotate by immediate counts.
• Stack instructions PUSHA, PUSHAD, POPA and POPAD, and PUSH immediate
data.
• Move with sign extension instructions MOVSX and MOVZX.
• Long-displacement Jcc instructions.
• Exchange instructions CMPXCHG, CMPXCHG8B, and XADD.
• String instructions MOVS, CMPS, SCAS, LODS, and STOS.
Vol. 3 17-5
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
• Bit test and bit scan instructions BT, BTS, BTR, BTC, BSF, and BSR; the byte-set-
on condition instruction SETcc; and the byte swap (BSWAP) instruction.
• Double shift instructions SHLD and SHRD.
• EFLAGS control instructions PUSHF and POPF.
• ENTER and LEAVE control instructions.
• BOUND instruction.
• CPU identification (CPUID) instruction.
• System instructions CLTS, INVD, WINVD, INVLPG, LGDT, SGDT, LIDT, SIDT,
LMSW, SMSW, RDMSR, WRMSR, RDTSC, and RDPMC.
Execution of any of the other IA-32 architecture instructions (not given in the
previous two lists) in real-address mode result in an invalid-opcode exception (#UD)
being generated.
17.1.4
Interrupt and Exception Handling
When operating in real-address mode, software must provide interrupt and excep-
tion-handling facilities that are separate from those provided in protected mode.
Even during the early stages of processor initialization when the processor is still in
real-address mode, elementary real-address mode interrupt and exception-handling
facilities must be provided to insure reliable operation of the processor, or the initial-
ization code must insure that no interrupts or exceptions will occur.
The IA-32 processors handle interrupts and exceptions in real-address mode similar
to the way they handle them in protected mode. When a processor receives an inter-
rupt or generates an exception, it uses the vector number of the interrupt or excep-
tion as an index into the interrupt table. (In protected mode, the interrupt table is
called the interrupt descriptor table (IDT), but in real-address mode, the table is
usually called the interrupt vector table, or simply the interrupt table.) The entry
in the interrupt vector table provides a pointer to an interrupt- or exception-handler
procedure. (The pointer consists of a segment selector for a code segment and a 16-
bit offset into the segment.) The processor performs the following actions to make an
implicit call to the selected handler:
1. Pushes the current values of the CS and EIP registers onto the stack. (Only the 16
least-significant bits of the EIP register are pushed.)
2. Pushes the low-order 16 bits of the EFLAGS register onto the stack.
3. Clears the IF flag in the EFLAGS register to disable interrupts.
4. Clears the TF, RC, and AC flags, in the EFLAGS register.
5. Transfers program control to the location specified in the interrupt vector table.
An IRET instruction at the end of the handler procedure reverses these steps to
return program control to the interrupted program. Exceptions do not return error
codes in real-address mode.
17-6 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
The interrupt vector table is an array of 4-byte entries (see Figure 17-2). Each entry
consists of a far pointer to a handler procedure, made up of a segment selector and
an offset. The processor scales the interrupt or exception vector by 4 to obtain an
offset into the interrupt table. Following reset, the base of the interrupt vector table
is located at physical address 0 and its limit is set to 3FFH. In the Intel 8086
processor, the base address and limit of the interrupt vector table cannot be
changed. In the later IA-32 processors, the base address and limit of the interrupt
vector table are contained in the IDTR register and can be changed using the LIDT
instruction.
(For backward compatibility to Intel 8086 processors, the default base address and
limit of the interrupt vector table should not be changed.)
Up to Entry 255
Entry 3
12
Entry 2
8
Entry 1
4
Segment Selector
Offset
2
0
Interrupt Vector 0*
0
15
* Interrupt vector number 0 selects entry 0
IDTR
(called “interrupt vector 0”) in the interrupt
points to the start of the interrupt handler
for interrupt 0.
Figure 17-2. Interrupt Vector Table in Real-Address Mode
Table 17-1 shows the interrupt and exception vectors that can be generated in real-
address mode and virtual-8086 mode, and in the Intel 8086 processor. See Chapter
6, “Interrupt and Exception Handling”, for a description of the exception conditions.
Vol. 3 17-7
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
Virtual-8086 mode is actually a special type of a task that runs in protected mode.
When the operating-system or executive switches to a virtual-8086-mode task, the
processor emulates an Intel 8086 processor. The execution environment of the
processor while in the 8086-emulation state is the same as is described in Section
17.1, “Real-Address Mode” for real-address mode, including the extensions. The
major difference between the two modes is that in virtual-8086 mode the 8086
emulator uses some protected-mode services (such as the protected-mode interrupt
and exception-handling and paging facilities).
As in real-address mode, any new or legacy program that has been assembled
and/or compiled to run on an Intel 8086 processor will run in a virtual-8086-mode
task. And several 8086 programs can be run as virtual-8086-mode tasks concur-
rently with normal protected-mode tasks, using the processor’s multitasking
facilities.
Table 17-1. Real-Address Mode Exceptions and Interrupts
Vector Description
No.
Real-Address
Mode
Virtual-8086
Mode
Intel 8086
Processor
0
1
Divide Error (#DE)
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Debug Exception (#DB)
NMI Interrupt
Yes
No
2
Yes
Yes
3
Breakpoint (#BP)
Overflow (#OF)
Yes
Yes
4
Yes
Yes
5
BOUND Range Exceeded (#BR) Yes
Yes
Reserved
Reserved
Reserved
Reserved
Reserved
Reserved
Reserved
Reserved
Reserved
Reserved
Reserved
Reserved
Reserved
Reserved
6
Invalid Opcode (#UD)
Device Not Available (#NM)
Double Fault (#DF)
Yes
Yes
7
Yes
Yes
8
Yes
Yes
9
(Intel reserved. Do not use.)
Invalid TSS (#TS)
Reserved
Reserved
Reserved
Yes
Reserved
Yes
10
11
12
13
14
15
16
17
18
Segment Not Present (#NP)
Stack Fault (#SS)
Yes
Yes
General Protection (#GP)*
Page Fault (#PF)
Yes
Yes
Reserved
Reserved
Yes
Yes
(Intel reserved. Do not use.)
Floating-Point Error (#MF)
Alignment Check (#AC)
Machine Check (#MC)
Reserved
Yes
Reserved
Yes
Yes
Yes
17-8 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
Table 17-1. Real-Address Mode Exceptions and Interrupts (Contd.)
Vector Description
No.
Real-Address
Mode
Virtual-8086
Mode
Intel 8086
Processor
19-31 (Intel reserved. Do not use.)
Reserved
Yes
Reserved
Yes
Reserved
Yes
32-
User Defined Interrupts
255
NOTE:
* In the real-address mode, vector 13 is the segment overrun exception. In protected and vir-
tual-8086 modes, this exception covers all general-protection error conditions, including traps
to the virtual-8086 monitor from virtual-8086 mode.
17.2.1
Enabling Virtual-8086 Mode
The processor runs in virtual-8086 mode when the VM (virtual machine) flag in the
EFLAGS register is set. This flag can only be set when the processor switches to a
new protected-mode task or resumes virtual-8086 mode via an IRET instruction.
System software cannot change the state of the VM flag directly in the EFLAGS
register (for example, by using the POPFD instruction). Instead it changes the flag in
the image of the EFLAGS register stored in the TSS or on the stack following a call to
an interrupt- or exception-handler procedure. For example, software sets the VM flag
in the EFLAGS image in the TSS when first creating a virtual-8086 task.
The processor tests the VM flag under three general conditions:
• When loading segment registers, to determine whether to use 8086-style
address translation.
• When decoding instructions, to determine which instructions are not supported in
virtual-8086 mode and which instructions are sensitive to IOPL.
• When checking privileged instructions, on page accesses, or when performing
other permission checks. (Virtual-8086 mode always executes at CPL 3.)
17.2.2
Structure of a Virtual-8086 Task
A virtual-8086-mode task consists of the following items:
• A 32-bit TSS for the task.
• The 8086 program.
• A virtual-8086 monitor.
• 8086 operating-system services.
The TSS of the new task must be a 32-bit TSS, not a 16-bit TSS, because the 16-bit
TSS does not load the most-significant word of the EFLAGS register, which contains
the VM flag. All TSS’s, stacks, data, and code used to handle exceptions when in
virtual-8086 mode must also be 32-bit segments.
Vol. 3 17-9
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
The processor enters virtual-8086 mode to run the 8086 program and returns to
protected mode to run the virtual-8086 monitor.
The virtual-8086 monitor is a 32-bit protected-mode code module that runs at a CPL
of 0. The monitor consists of initialization, interrupt- and exception-handling, and I/O
emulation procedures that emulate a personal computer or other 8086-based plat-
form. Typically, the monitor is either part of or closely associated with the protected-
mode general-protection (#GP) exception handler, which also runs at a CPL of 0. As
with any protected-mode code module, code-segment descriptors for the virtual-
8086 monitor must exist in the GDT or in the task’s LDT. The virtual-8086 monitor
also may need data-segment descriptors so it can examine the IDT or other parts of
the 8086 program in the first 1 MByte of the address space. The linear addresses
above 10FFEFH are available for the monitor, the operating system, and other system
software.
The 8086 operating-system services consists of a kernel and/or operating-system
procedures that the 8086 program makes calls to. These services can be imple-
mented in either of the following two ways:
• They can be included in the 8086 program. This approach is desirable for either
of the following reasons:
— The 8086 program code modifies the 8086 operating-system services.
— There is not sufficient development time to merge the 8086 operating-
system services into main operating system or executive.
• They can be implemented or emulated in the virtual-8086 monitor. This approach
is desirable for any of the following reasons:
— The 8086 operating-system procedures can be more easily coordinated
among several virtual-8086 tasks.
— Memory can be saved by not duplicating 8086 operating-system procedure
code for several virtual-8086 tasks.
— The 8086 operating-system procedures can be easily emulated by calls to the
main operating system or executive.
The approach chosen for implementing the 8086 operating-system services may
result in different virtual-8086-mode tasks using different 8086 operating-system
services.
17.2.3
Paging of Virtual-8086 Tasks
Even though a program running in virtual-8086 mode can use only 20-bit linear
addresses, the processor converts these addresses into 32-bit linear addresses
before mapping them to the physical address space. If paging is being used, the
8086 address space for a program running in virtual-8086 mode can be paged and
located in a set of pages in physical address space. If paging is used, it is transparent
to the program running in virtual-8086 mode just as it is for any task running on the
processor.
17-10 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
Paging is not necessary for a single virtual-8086-mode task, but paging is useful or
necessary in the following situations:
MByte of the linear address space for each virtual-8086-mode task to be mapped
to a different physical address location.
• When emulating the 8086 address-wraparound that occurs at 1 MByte. When
using 8086-style address translation, it is possible to specify addresses larger
than 1 MByte. These addresses automatically wraparound in the Intel 8086
processor (see Section 17.1.1, “Address Translation in Real-Address Mode”). If
any 8086 programs depend on address wraparound, the same effect can be
achieved in a virtual-8086-mode task by mapping the linear addresses between
100000H and 110000H and linear addresses between 0 and 10000H to the same
physical addresses.
• When sharing the 8086 operating-system services or ROM code that is common
to several 8086 programs running as different 8086-mode tasks.
• When redirecting or trapping references to memory-mapped I/O devices.
17.2.4
Protection within a Virtual-8086 Task
Protection is not enforced between the segments of an 8086 program. Either of the
following techniques can be used to protect the system software running in a virtual-
8086-mode task from the 8086 program:
• Reserve the first 1 MByte plus 64 KBytes of each task’s linear address space for
the 8086 program. An 8086 processor task cannot generate addresses outside
this range.
• Use the U/S flag of page-table entries to protect the virtual-8086 monitor and
other system software in the virtual-8086 mode task space. When the processor
only user privileges. If the pages of the virtual-8086 monitor have supervisor
privilege, they cannot be accessed by the 8086 program.
17.2.5
Entering Virtual-8086 Mode
Figure 17-3 summarizes the methods of entering and leaving virtual-8086 mode.
The processor switches to virtual-8086 mode in either of the following situations:
• Task switch when the VM flag is set to 1 in the EFLAGS register image stored in
the TSS for the task. Here the task switch can be initiated in either of two ways:
— A CALL or JMP instruction.
— An IRET instruction, where the NT flag in the EFLAGS image is set to 1.
• Return from a protected-mode interrupt or exception handler when the VM flag is
set to 1 in the EFLAGS register image on the stack.
Vol. 3 17-11
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
When a task switch is used to enter virtual-8086 mode, the TSS for the virtual-8086-
mode task must be a 32-bit TSS. (If the new TSS is a 16-bit TSS, the upper word of
the EFLAGS register is not in the TSS, causing the processor to clear the VM flag
loading the segment registers from their images in the new TSS. The new setting of
the VM flag determines whether the processor interprets the contents of the segment
registers as 8086-style segment selectors or protected-mode segment selectors.
When the VM flag is set, the segment registers are loaded from the TSS, using 8086-
style address translation to form base addresses.
See Section 17.3, “Interrupt and Exception Handling in Virtual-8086 Mode”, for infor-
mation on entering virtual-8086 mode on a return from an interrupt or exception
handler.
17-12 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
Real Mode
Code
Real-Address
Mode
PE=0 or
RESET
PE=1
Task Switch
VM=0
CALL
RET
Protected-
Protected
Mode
Mode Interrupt
Virtual-8086
Monitor
Protected-
Mode Tasks
and Exception
Handlers
Task Switch1
VM = 0
VM = 1
Interrupt or
Exception2
Virtual-8086
Mode
Virtual-8086
Mode Tasks
(8086
#GP Exception3
RESET
IRET4
IRET5
Programs)
Redirect Interrupt to 8086 Program
Interrupt or Exception Handler6
NOTES:
1. Task switch carried out in either of two ways:
- CALL or JMP where the VM flag in the EFLAGS image is 1.
- IRET where VM is 1 and NT is 1.
2. Hardware interrupt or exception; software interrupt (INT n) when IOPL is 3.
3. General-protection exception caused by software interrupt (INT n), IRET,
POPF, PUSHF, IN, or OUT when IOPL is less than 3.
4. Normal return from protected-mode interrupt or exception handler.
5. A return from the 8086 monitor to redirect an interrupt or exception back
to an interrupt or exception handler in the 8086 program running in virtual-
8086 mode.
6. Internal redirection of a software interrupt (INT n) when VME is 1,
IOPL is <3, and the redirection bit is 1.
Figure 17-3. Entering and Leaving Virtual-8086 Mode
Vol. 3 17-13
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
17.2.6
Leaving Virtual-8086 Mode
The processor can leave the virtual-8086 mode only through an interrupt or excep-
tion. The following are situations where an interrupt or exception will lead to the
processor leaving virtual-8086 mode (see Figure 17-3):
• The processor services a hardware interrupt generated to signal the suspension
of execution of the virtual-8086 application. This hardware interrupt may be
generated by a timer or other external mechanism. Upon receiving the hardware
interrupt, the processor enters protected mode and switches to a protected-
mode (or another virtual-8086 mode) task either through a task gate in the
protected-mode IDT or through a trap or interrupt gate that points to a handler
that initiates a task switch. A task switch from a virtual-8086 task to another task
loads the EFLAGS register from the TSS of the new task. The value of the VM flag
in the new EFLAGS determines if the new task executes in virtual-8086 mode or
not.
• The processor services an exception caused by code executing the virtual-8086
task or services a hardware interrupt that “belongs to” the virtual-8086 task.
Here, the processor enters protected mode and services the exception or
hardware interrupt through the protected-mode IDT (normally through an
interrupt or trap gate) and the protected-mode exception- and interrupt-
handlers. The processor may handle the exception or interrupt within the context
of the virtual 8086 task and return to virtual-8086 mode on a return from the
exception or interrupt in the context of another task.
• The processor services a software interrupt generated by code executing in the
virtual-8086 task (such as a software interrupt to call a MS-DOS* operating
system routine). The processor provides several methods of handling these
software interrupts, which are discussed in detail in Section 17.3.3, “Class
3—Software Interrupt Handling in Virtual-8086 Mode”. Most of them involve the
processor entering protected mode, often by means of a general-protection
(#GP) exception. In protected mode, the processor can send the interrupt to the
virtual-8086 monitor for handling and/or redirect the interrupt back to the
application program running in virtual-8086 mode task for handling.
IA-32 processors that incorporate the virtual mode extension (enabled with the
VME flag in control register CR4) are capable of redirecting software-generated
interrupts back to the program’s interrupt handlers without leaving virtual-8086
mode. See Section 17.3.3.4, “Method 5: Software Interrupt Handling”, for more
information on this mechanism.
• A hardware reset initiated by asserting the RESET or INIT pin is a special kind of
interrupt. When a RESET or INIT is signaled while the processor is in virtual-8086
mode, the processor leaves virtual-8086 mode and enters real-address mode.
• Execution of the HLT instruction in virtual-8086 mode will cause a general-
protection (GP#) fault, which the protected-mode handler generally sends to the
virtual-8086 monitor. The virtual-8086 monitor then determines the correct
17-14 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
execution sequence after verifying that it was entered as a result of a HLT
execution.
See Section 17.3, “Interrupt and Exception Handling in Virtual-8086 Mode”, for infor-
mation on leaving virtual-8086 mode to handle an interrupt or exception generated
in virtual-8086 mode.
17.2.7
Sensitive Instructions
When an IA-32 processor is running in virtual-8086 mode, the CLI, STI, PUSHF, POPF,
INT n, and IRET instructions are sensitive to IOPL. The IN, INS, OUT, and OUTS
instructions, which are sensitive to IOPL in protected mode, are not sensitive in
virtual-8086 mode.
The CPL is always 3 while running in virtual-8086 mode; if the IOPL is less than 3, an
attempt to use the IOPL-sensitive instructions listed above triggers a general-protec-
tion exception (#GP). These instructions are sensitive to IOPL to give the virtual-
8086 monitor a chance to emulate the facilities they affect.
17.2.8
Virtual-8086 Mode I/O
Many 8086 programs written for non-multitasking systems directly access I/O ports.
This practice may cause problems in a multitasking environment. If more than one
program accesses the same port, they may interfere with each other. Most multi-
tasking systems require application programs to access I/O ports through the oper-
ating system. This results in simplified, centralized control.
The processor provides I/O protection for creating I/O that is compatible with the
environment and transparent to 8086 programs. Designers may take any of several
possible approaches to protecting I/O ports:
• Protect the I/O address space and generate exceptions for all attempts to
perform I/O directly.
• Let the 8086 program perform I/O directly.
• Generate exceptions on attempts to access specific I/O ports.
• Generate exceptions on attempts to access specific memory-mapped I/O ports.
The method of controlling access to I/O ports depends upon whether they are
I/O-port mapped or memory mapped.
17.2.8.1 I/O-Port-Mapped I/O
The I/O permission bit map in the TSS can be used to generate exceptions on
attempts to access specific I/O port addresses. The I/O permission bit map of each
virtual-8086-mode task determines which I/O addresses generate exceptions for
that task. Because each task may have a different I/O permission bit map, the
addresses that generate exceptions for one task may be different from the addresses
Vol. 3 17-15
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
for another task. This differs from protected mode in which, if the CPL is less than or
equal to the IOPL, I/O access is allowed without checking the I/O permission bit map.
See Chapter 13, “Input/Output”, in the Intel® 64 and IA-32 Architectures Software
Developer’s Manual, Volume 1, for more information about the I/O permission bit
map.
17.2.8.2 Memory-Mapped I/O
In systems which use memory-mapped I/O, the paging facilities of the processor can
be used to generate exceptions for attempts to access I/O ports. The virtual-8086
monitor may use paging to control memory-mapped I/O in these ways:
• Map part of the linear address space of each task that needs to perform I/O to the
physical address space where I/O ports are placed. By putting the I/O ports at
different addresses (in different pages), the paging mechanism can enforce
isolation between tasks.
• Map part of the linear address space to pages that are not-present. This
generates an exception whenever a task attempts to perform I/O to those pages.
System software then can interpret the I/O operation being attempted.
Software emulation of the I/O space may require too much operating system inter-
vention under some conditions. In these cases, it may be possible to generate an
exception for only the first attempt to access I/O. The system software then may
determine whether a program can be given exclusive control of I/O temporarily, the
protection of the I/O space may be lifted, and the program allowed to run at full
speed.
17.2.8.3 Special I/O Buffers
Buffers of intelligent controllers (for example, a bit-mapped frame buffer) also can be
emulated using page mapping. The linear space for the buffer can be mapped to a
different physical space for each virtual-8086-mode task. The virtual-8086 monitor
then can control which virtual buffer to copy onto the real buffer in the physical
address space.
17.3
INTERRUPT AND EXCEPTION HANDLING
IN VIRTUAL-8086 MODE
When the processor receives an interrupt or detects an exception condition while in
virtual-8086 mode, it invokes an interrupt or exception handler, just as it does in
protected or real-address mode. The interrupt or exception handler that is invoked
and the mechanism used to invoke it depends on the class of interrupt or exception
that has been detected or generated and the state of various system flags and fields.
17-16 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
In virtual-8086 mode, the interrupts and exceptions are divided into three classes for
the purposes of handling:
• Class 1 — All processor-generated exceptions and all hardware interrupts,
including the NMI interrupt and the hardware interrupts sent to the processor’s
external interrupt delivery pins. All class 1 exceptions and interrupts are handled
by the protected-mode exception and interrupt handlers.
• Class 2 — Special case for maskable hardware interrupts (Section 6.3.2,
“Maskable Hardware Interrupts”) when the virtual mode extensions are enabled.
1
the INT n instruction .
The method the processor uses to handle class 2 and 3 interrupts depends on the
setting of the following flags and fields:
• IOPL field (bits 12 and 13 in the EFLAGS register) — Controls how class 3
software interrupts are handled when the processor is in virtual-8086 mode (see
Section 2.3, “System Flags and Fields in the EFLAGS Register”). This field also
VME flag is set. The VIF and VIP flags are provided to assist in the handling of
class 2 maskable hardware interrupts.
• VME flag (bit 0 in control register CR4) — Enables the virtual mode extension
for the processor when set (see Section 2.5, “Control Registers”).
• Software interrupt redirection bit map (32 bytes in the TSS, see
Figure 17-5) — Contains 256 flags that indicates how class 3 software
interrupts should be handled when they occur in virtual-8086 mode. A software
currently running 8086 program or to the protected-mode interrupt and
exception handlers.
• The virtual interrupt flag (VIF) and virtual interrupt pending flag (VIP)
in the EFLAGS register — Provides virtual interrupt support for the handling
of class 2 maskable hardware interrupts (see Section 17.3.2, “Class 2—Maskable
Hardware Interrupt Handling in Virtual-8086 Mode Using the Virtual Interrupt
Mechanism”).
NOTE
The VME flag, software interrupt redirection bit map, and VIF and VIP
flags are only available in IA-32 processors that support the virtual
mode extensions. These extensions were introduced in the IA-32
architecture with the Pentium processor.
The following sections describe the actions that processor takes and the possible
actions of interrupt and exception handlers for the two classes of interrupts described
1. The INT 3 instruction is a special case (see the description of the INT n instruction in Chapter 3,
“Instruction Set Reference, A-M”, of the Intel® 64 and IA-32 Architectures Software Developer’s
Manual, Volume 2A).
Vol. 3 17-17
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
in the previous paragraphs. These sections describe three possible types of interrupt
and exception handlers:
• Protected-mode interrupt and exceptions handlers — These are the
standard handlers that the processor calls through the protected-mode IDT.
• Virtual-8086 monitor interrupt and exception handlers — These handlers
are resident in the virtual-8086 monitor, and they are commonly accessed
through a general-protection exception (#GP, interrupt 13) that is directed to the
protected-mode general-protection exception handler.
• 8086 program interrupt and exception handlers — These handlers are part
of the 8086 program that is running in virtual-8086 mode.
The following sections describe how these handlers are used, depending on the
selected class and method of interrupt and exception handling.
17.3.1
Class 1—Hardware Interrupt and Exception Handling in
Virtual-8086 Mode
In virtual-8086 mode, the Pentium, P6 family, Pentium 4, and Intel Xeon processors
or exception handler that the interrupt or exception vector points to in the IDT. Here,
the IDT entry must contain either a 32-bit trap or interrupt gate or a task gate. The
following sections describe various ways that a virtual-8086 mode interrupt or excep-
tion can be handled after the protected-mode handler has been invoked.
See Section 17.3.2, “Class 2—Maskable Hardware Interrupt Handling in Virtual-8086
Mode Using the Virtual Interrupt Mechanism”, for a description of the virtual interrupt
mechanism that is available for handling maskable hardware interrupts while in
virtual-8086 mode. When this mechanism is either not available or not enabled,
maskable hardware interrupts are handled in the same manner as exceptions, as
described in the following sections.
17.3.1.1 Handling an Interrupt or Exception Through a Protected-Mode
Trap or Interrupt Gate
IDT, the gate must in turn point to a nonconforming, privilege-level 0, code segment.
When accessing this code segment, processor performs the following steps.
1. Switches to 32-bit protected mode and privilege level 0.
2. Saves the state of the processor on the privilege-level 0 stack. The states of the
EIP, CS, EFLAGS, ESP, SS, ES, DS, FS, and GS registers are saved (see
Figure 17-4).
3. Clears the segment registers. Saving the DS, ES, FS, and GS registers on the
stack and then clearing the registers lets the interrupt or exception handler safely
17-18 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
save and restore these registers regardless of the type segment selectors they
contain (protected-mode or 8086-style). The interrupt and exception handlers,
which may be called in the context of either a protected-mode task or a virtual-
8086-mode task, can use the same code sequences for saving and restoring the
registers for any task. Clearing these registers before execution of the IRET
instruction does not cause a trap in the interrupt handler. Interrupt procedures
that expect values in the segment registers or that return values in the segment
registers must use the register images saved on the stack for privilege level 0.
4. Clears VM, NT, RF and TF flags (in the EFLAGS register). If the gate is an interrupt
gate, clears the IF flag.
5. Begins executing the selected interrupt or exception handler.
If the trap or interrupt gate references a procedure in a conforming segment or in a
segment at a privilege level other than 0, the processor generates a general-protec-
tion exception (#GP). Here, the error code is the segment selector of the code
segment to which a call was attempted.
Without Error Code
With Error Code
ESP from
TSS
ESP from
TSS
Unused
Old GS
Unused
Old GS
Old FS
Old FS
Old DS
Old DS
Old ES
Old ES
Old SS
Old SS
Old ESP
Old EFLAGS
Old CS
Old ESP
Old EFLAGS
Old CS
Old EIP
New ESP
Old EIP
New ESP
Error Code
Figure 17-4. Privilege Level 0 Stack After Interrupt or
Exception in Virtual-8086 Mode
Vol. 3 17-19
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
Interrupt and exception handlers can examine the VM flag on the stack to determine
if the interrupted procedure was running in virtual-8086 mode. If so, the interrupt or
exception can be handled in one of three ways:
• The protected-mode interrupt or exception handler that was called can handle
the interrupt or exception.
• The protected-mode interrupt or exception handler can call the virtual-8086
monitor to handle the interrupt or exception.
• The virtual-8086 monitor (if called) can in turn pass control back to the 8086
program’s interrupt and exception handler.
If the interrupt or exception is handled with a protected-mode handler, the handler
can return to the interrupted program in virtual-8086 mode by executing an IRET
instruction. This instruction loads the EFLAGS and segment registers from the
images saved in the privilege level 0 stack (see Figure 17-4). A set VM flag in the
EFLAGS image causes the processor to switch back to virtual-8086 mode. The CPL at
the time the IRET instruction is executed must be 0, otherwise the processor does
not change the state of the VM flag.
The virtual-8086 monitor runs at privilege level 0, like the protected-mode interrupt
and exception handlers. It is commonly closely tied to the protected-mode general-
protection exception (#GP, vector 13) handler. If the protected-mode interrupt or
exception handler calls the virtual-8086 monitor to handle the interrupt or exception,
the return from the virtual-8086 monitor to the interrupted virtual-8086 mode
protected-mode handler and an IRET instruction to return to the interrupted
program.
The virtual-8086 monitor has the option of directing the interrupt and exception back
to an interrupt or exception handler that is part of the interrupted 8086 program, as
described in Section 17.3.1.2, “Handling an Interrupt or Exception With an 8086
Program Interrupt or Exception Handler”.
17.3.1.2 Handling an Interrupt or Exception With an 8086 Program
Interrupt or Exception Handler
Because it was designed to run on an 8086 processor, an 8086 program running in a
virtual-8086-mode task contains an 8086-style interrupt vector table, which starts at
linear address 0. If the virtual-8086 monitor correctly directs an interrupt or excep-
tion vector back to the virtual-8086-mode task it came from, the handlers in the
8086 program can handle the interrupt or exception. The virtual-8086 monitor must
carry out the following steps to send an interrupt or exception back to the 8086
program:
1. Use the 8086 interrupt vector to locate the appropriate handler procedure in the
8086 program interrupt table.
17-20 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
2. Store the EFLAGS (low-order 16 bits only), CS and EIP values of the 8086
program on the privilege-level 3 stack. This is the stack that the virtual-8086-
mode task is using. (The 8086 handler may use or modify this information.)
3. Change the return link on the privilege-level 0 stack to point to the privilege-level
3 handler procedure.
4. Execute an IRET instruction to pass control to the 8086 program handler.
5. When the IRET instruction from the privilege-level 3 handler triggers a general-
protection exception (#GP) and thus effectively again calls the virtual-8086
monitor, restore the return link on the privilege-level 0 stack to point to the
original, interrupted, privilege-level 3 procedure.
6. Copy the low order 16 bits of the EFLAGS image from the privilege-level 3 stack
to the privilege-level 0 stack (because some 8086 handlers modify these flags to
return information to the code that caused the interrupt).
7. Execute an IRET instruction to pass control back to the interrupted 8086
program.
Note that if an operating system intends to support all 8086 MS-DOS-based
programs, it is necessary to use the actual 8086 interrupt and exception handlers
supplied with the program. The reason for this is that some programs modify their
own interrupt vector table to substitute (or hook in series) their own specialized
interrupt and exception handlers.
17.3.1.3 Handling an Interrupt or Exception Through a Task Gate
When an interrupt or exception vector points to a task gate in the IDT, the processor
performs a task switch to the selected interrupt- or exception-handling task. The
following actions are carried out as part of this task switch:
1. The EFLAGS register with the VM flag set is saved in the current TSS.
2. The link field in the TSS of the called task is loaded with the segment selector of
the TSS for the interrupted virtual-8086-mode task.
3. The EFLAGS register is loaded from the image in the new TSS, which clears the
VM flag and causes the processor to switch to protected mode.
4. The NT flag in the EFLAGS register is set.
5. The processor begins executing the selected interrupt- or exception-handler
task.
When an IRET instruction is executed in the handler task and the NT flag in the
EFLAGS register is set, the processors switches from a protected-mode interrupt- or
exception-handler task back to a virtual-8086-mode task. Here, the EFLAGS and
segment registers are loaded from images saved in the TSS for the virtual-8086-
mode task. If the VM flag is set in the EFLAGS image, the processor switches back to
virtual-8086 mode on the task switch. The CPL at the time the IRET instruction is
Vol. 3 17-21
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
executed must be 0, otherwise the processor does not change the state of the VM
flag.
17.3.2
Class 2—Maskable Hardware Interrupt Handling in
Virtual-8086 Mode Using the Virtual Interrupt Mechanism
Maskable hardware interrupts are those interrupts that are delivered through the
INTR# pin or through an interrupt request to the local APIC (see Section 6.3.2,
“Maskable Hardware Interrupts”). These interrupts can be inhibited (masked) from
interrupting an executing program or task by clearing the IF flag in the EFLAGS
register.
When the VME flag in control register CR4 is set and the IOPL field in the EFLAGS
register is less than 3, two additional flags are activated in the EFLAGS register:
• VIF (virtual interrupt) flag, bit 19 of the EFLAGS register.
• VIP (virtual interrupt pending) flag, bit 20 of the EFLAGS register.
These flags provide the virtual-8086 monitor with more efficient control over
handling maskable hardware interrupts that occur during virtual-8086 mode tasks.
They also reduce interrupt-handling overhead, by eliminating the need for all IF
related operations (such as PUSHF, POPF, CLI, and STI instructions) to trap to the
virtual-8086 monitor. The purpose and use of these flags are as follows.
NOTE
The VIF and VIP flags are only available in IA-32 processors that
support the virtual mode extensions. These extensions were
introduced in the IA-32 architecture with the Pentium processor.
When this mechanism is either not available or not enabled,
maskable hardware interrupts are handled as class 1 interrupts.
Here, if VIF and VIP flags are needed, the virtual-8086 monitor can
implement them in software.
Existing 8086 programs commonly set and clear the IF flag in the EFLAGS register to
enable and disable maskable hardware interrupts, respectively; for example, to
disable interrupts while handling another interrupt or an exception. This practice
works well in single task environments, but can cause problems in multitasking and
multiple-processor environments, where it is often desirable to prevent an applica-
tion program from having direct control over the handling of hardware interrupts.
When using earlier IA-32 processors, this problem was often solved by creating a
virtual IF flag in software. The IA-32 processors (beginning with the Pentium
processor) provide hardware support for this virtual IF flag through the VIF and VIP
flags.
The VIF flag is a virtualized version of the IF flag, which an application program
running from within a virtual-8086 task can used to control the handling of maskable
hardware interrupts. When the VIF flag is enabled, the CLI and STI instructions
operate on the VIF flag instead of the IF flag. When an 8086 program executes the
17-22 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
CLI instruction, the processor clears the VIF flag to request that the virtual-8086
monitor inhibit maskable hardware interrupts from interrupting program execution;
when it executes the STI instruction, the processor sets the VIF flag requesting that
the virtual-8086 monitor enable maskable hardware interrupts for the 8086
program. But actually the IF flag, managed by the operating system, always controls
whether maskable hardware interrupts are enabled. Also, if under these circum-
stances an 8086 program tries to read or change the IF flag using the PUSHF or POPF
instructions, the processor will change the VIF flag instead, leaving IF unchanged.
The VIP flag provides software a means of recording the existence of a deferred (or
pending) maskable hardware interrupt. This flag is read by the processor but never
explicitly written by the processor; it can only be written by software.
If the IF flag is set and the VIF and VIP flags are enabled, and the processor receives
operations:
1. The processor invokes the protected-mode interrupt handler for the interrupt
received, as described in the following steps. These steps are almost identical to
those described for method 1 interrupt and exception handling in Section
17.3.1.1, “Handling an Interrupt or Exception Through a Protected-Mode Trap or
Interrupt Gate”:
a. Switches to 32-bit protected mode and privilege level 0.
b. Saves the state of the processor on the privilege-level 0 stack. The states of
the EIP, CS, EFLAGS, ESP, SS, ES, DS, FS, and GS registers are saved (see
Figure 17-4).
c. Clears the segment registers.
d. Clears the VM flag in the EFLAGS register.
e. Begins executing the selected protected-mode interrupt handler.
2. The recommended action of the protected-mode interrupt handler is to read the
VM flag from the EFLAGS image on the stack. If this flag is set, the handler makes
a call to the virtual-8086 monitor.
3. The virtual-8086 monitor should read the VIF flag in the EFLAGS register.
— If the VIF flag is clear, the virtual-8086 monitor sets the VIP flag in the
EFLAGS image on the stack to indicate that there is a deferred interrupt
pending and returns to the protected-mode handler.
— If the VIF flag is set, the virtual-8086 monitor can handle the interrupt if it
“belongs” to the 8086 program running in the interrupted virtual-8086 task;
otherwise, it can call the protected-mode interrupt handler to handle the
interrupt.
4. The protected-mode handler executes a return to the program executing in
virtual-8086 mode.
Vol. 3 17-23
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
5. Upon returning to virtual-8086 mode, the processor continues execution of the
8086 program.
When the 8086 program is ready to receive maskable hardware interrupts, it
executes the STI instruction to set the VIF flag (enabling maskable hardware
interrupts). Prior to setting the VIF flag, the processor automatically checks the VIP
flag and does one of the following, depending on the state of the flag:
• If the VIP flag is clear (indicating no pending interrupts), the processor sets the
VIF flag.
• If the VIP flag is set (indicating a pending interrupt), the processor generates a
general-protection exception (#GP).
The recommended action of the protected-mode general-protection exception
handler is to then call the virtual-8086 monitor and let it handle the pending inter-
rupt. After handling the pending interrupt, the typical action of the virtual-8086
monitor is to clear the VIP flag and set the VIF flag in the EFLAGS image on the stack,
and then execute a return to the virtual-8086 mode. The next time the processor
receives a maskable hardware interrupt, it will then handle it as described in steps 1
through 5 earlier in this section.
If the processor finds that both the VIF and VIP flags are set at the beginning of an
instruction, it generates a general-protection exception. This action allows the
virtual-8086 monitor to handle the pending interrupt for the virtual-8086 mode task
for which the VIF flag is enabled. Note that this situation can only occur immediately
following execution of a POPF or IRET instruction or upon entering a virtual-8086
mode task through a task switch.
during transitions between real-address and protected modes.
NOTE
The virtual interrupt mechanism described in this section is also
available for use in protected mode, see Section 17.4, “Protected-
Mode Virtual Interrupts”.
17.3.3
Class 3—Software Interrupt Handling in Virtual-8086 Mode
When the processor receives a software interrupt (an interrupt generated with the
INT n instruction) while in virtual-8086 mode, it can use any of six different methods
to handle the interrupt. The method selected depends on the settings of the VME flag
in control register CR4, the IOPL field in the EFLAGS register, and the software inter-
rupt redirection bit map in the TSS. Table 17-2 lists the six methods of handling soft-
ware interrupts in virtual-8086 mode and the respective settings of the VME flag,
IOPL field, and the bits in the interrupt redirection bit map for each method. The table
also summarizes the various actions the processor takes for each method.
The VME flag enables the virtual mode extensions for the Pentium and later IA-32
processors. When this flag is clear, the processor responds to interrupts and excep-
17-24 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
tions in virtual-8086 mode in the same manner as an Intel386 or Intel486 processor
does. When this flag is set, the virtual mode extension provides the following
enhancements to virtual-8086 mode:
• Speeds up the handling of software-generated interrupts in virtual-8086 mode by
allowing the processor to bypass the virtual-8086 monitor and redirect software
8086 program.
• Supports virtual interrupts for software written to run on the 8086 processor.
The IOPL value interacts with the VME flag and the bits in the interrupt redirection bit
map to determine how specific software interrupts should be handled.
The software interrupt redirection bit map (see Figure 17-5) is a 32-byte field in the
TSS. This map is located directly below the I/O permission bit map in the TSS. Each
bit in the interrupt redirection bit map is mapped to an interrupt vector. Bit 0 in the
interrupt redirection bit map (which maps to vector zero in the interrupt table) is
located at the I/O base map address in the TSS minus 32 bytes. When a bit in this bit
map is set, it indicates that the associated software interrupt (interrupt generated
with an INT n instruction) should be handled through the protected-mode IDT and
interrupt and exception handlers. When a bit in this bit map is clear, the processor
redirects the associated software interrupt back to the interrupt table in the 8086
program (located at linear address 0 in the program’s address space).
NOTE
The software interrupt redirection bit map does not affect hardware
generated interrupts and exceptions. Hardware generated interrupts
and exceptions are always handled by the protected-mode interrupt
and exception handlers.
Vol. 3 17-25
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
Table 17-2. Software Interrupt Handling Methods While in Virtual-8086 Mode
Bit in
Redir.
Method VME IOPL Bitmap*
Processor Action
1
0
3
X
Interrupt directed to a protected-mode interrupt handler:
•
•
•
Switches to privilege-level 0 stack
Pushes GS, FS, DS and ES onto privilege-level 0 stack
Pushes SS, ESP, EFLAGS, CS and EIP of interrupted task onto
privilege-level 0 stack
•
•
•
•
Clears VM, RF, NT, and TF flags
If serviced through interrupt gate, clears IF flag
Clears GS, FS, DS and ES to 0
Sets CS and EIP from interrupt gate
2
3
0
1
< 3
< 3
X
1
Interrupt directed to protected-mode general-protection
exception (#GP) handler.
Interrupt directed to a protected-mode general-protection
exception (#GP) handler; VIF and VIP flag support for handling
class 2 maskable hardware interrupts.
4
5
1
1
3
3
1
0
Interrupt directed to protected-mode interrupt handler: (see
method 1 processor action).
Interrupt redirected to 8086 program interrupt handler:
•
•
•
•
•
Pushes EFLAGS
Pushes CS and EIP (lower 16 bits only)
Clears IF flag
Clears TF flag
Loads CS and EIP (lower 16 bits only) from selected entry in
the interrupt vector table of the current virtual-8086 task
6
1
< 3
0
Interrupt redirected to 8086 program interrupt handler; VIF and
VIP flag support for handling class 2 maskable hardware
interrupts:
•
•
•
•
•
Pushes EFLAGS with IOPL set to 3 and VIF copied to IF
Pushes CS and EIP (lower 16 bits only)
Clears the VIF flag
Clears TF flag
Loads CS and EIP (lower 16 bits only) from selected entry in
the interrupt vector table of the current virtual-8086 task
NOTE:
* When set to 0, software interrupt is redirected back to the 8086 program interrupt handler;
when set to 1, interrupt is directed to protected-mode handler.
17-26 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
Task-State Segment (TSS)
I/O Permission Bit Map
31
1
24 23
1
0
Last byte of
bit
1
1
1
1
1
1
map must be
Software Interrupt Redirection Bit Map (32 Bytes)
64H
I/O Map Base
I/O map
base must
not exceed
DFFFH.
0
Figure 17-5. Software Interrupt Redirection Bit Map in TSS
Redirecting software interrupts back to the 8086 program potentially speeds up
interrupt handling because a switch back and forth between virtual-8086 mode and
protected mode is not required. This latter interrupt-handling technique is particu-
larly useful for 8086 operating systems (such as MS-DOS) that use the INT n instruc-
tion to call operating system procedures.
The CPUID instruction can be used to verify that the virtual mode extension is imple-
“Instruction Set Reference, A-M”, of the Intel® 64 and IA-32 Architectures Software
Developer’s Manual, Volume 2A).
The following sections describe the six methods (or mechanisms) for handling soft-
ware interrupts in virtual-8086 mode. See Section 17.3.2, “Class 2—Maskable Hard-
ware Interrupt Handling in Virtual-8086 Mode Using the Virtual Interrupt
Mechanism”, for a description of the use of the VIF and VIP flags in the EFLAGS
register for handling maskable hardware interrupts.
17.3.3.1 Method 1: Software Interrupt Handling
When the VME flag in control register CR4 is clear and the IOPL field is 3, a Pentium
or later IA-32 processor handles software interrupts in the same manner as they are
handled by an Intel386 or Intel486 processor. It executes an implicit call to the inter-
Vol. 3 17-27
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
rupt handler in the protected-mode IDT pointed to by the interrupt vector. See
Section 17.3.1, “Class 1—Hardware Interrupt and Exception Handling in Virtual-8086
Mode”, for a complete description of this mechanism and its possible uses.
17.3.3.2 Methods 2 and 3: Software Interrupt Handling
When a software interrupt occurs in virtual-8086 mode and the method 2 or 3 condi-
tions are present, the processor generates a general-protection exception (#GP).
Here the IOPL value is used to bypass the protected-mode interrupt handlers and
cause any software interrupt that occurs in virtual-8086 mode to be treated as a
protected-mode general-protection exception (#GP). The general-protection excep-
tion handler calls the virtual-8086 monitor, which can then emulate an 8086-
program interrupt handler or pass control back to the 8086 program’s handler, as
described in Section 17.3.1.2, “Handling an Interrupt or Exception With an 8086
Program Interrupt or Exception Handler”.
Method 3 is enabled when the VME flag is set to 1, the IOPL value is less than 3, and
the corresponding bit for the software interrupt in the software interrupt redirection
bit map is set to 1. Here, the processor performs the same operation as it does for
method 2 software interrupt handling. If the corresponding bit for the software inter-
rupt in the software interrupt redirection bit map is set to 0, the interrupt is handled
using method 6 (see Section 17.3.3.5, “Method 6: Software Interrupt Handling”).
17.3.3.3 Method 4: Software Interrupt Handling
Method 4 handling is enabled when the VME flag is set to 1, the IOPL value is 3, and
the bit for the interrupt vector in the redirection bit map is set to 1. Method 4 soft-
ware interrupt handling allows method 1 style handling when the virtual mode exten-
sion is enabled; that is, the interrupt is directed to a protected-mode handler (see
Section 17.3.3.1, “Method 1: Software Interrupt Handling”).
17.3.3.4 Method 5: Software Interrupt Handling
Method 5 software interrupt handling provides a streamlined method of redirecting
software interrupts (invoked with the INT n instruction) that occur in virtual 8086
mode back to the 8086 program’s interrupt vector table and its interrupt handlers.
Method 5 handling is enabled when the VME flag is set to 1, the IOPL value is 3, and
the bit for the interrupt vector in the redirection bit map is set to 0. The processor
performs the following actions to make an implicit call to the selected 8086 program
interrupt handler:
1. Pushes the low-order 16 bits of the EFLAGS register onto the stack.
2. Pushes the current values of the CS and EIP registers onto the current stack.
(Only the 16 least-significant bits of the EIP register are pushed and no stack
switch occurs.)
17-28 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
3. Clears the IF flag in the EFLAGS register to disable interrupts.
4. Clears the TF flag, in the EFLAGS register.
5. Locates the 8086 program interrupt vector table at linear address 0 for the 8086-
mode task.
6. Loads the CS and EIP registers with values from the interrupt vector table entry
pointed to by the interrupt vector number. Only the 16 low-order bits of the EIP
are loaded and the 16 high-order bits are set to 0. The interrupt vector table is
assumed to be at linear address 0 of the current virtual-8086 task.
7. Begins executing the selected interrupt handler.
An IRET instruction at the end of the handler procedure reverses these steps to
return program control to the interrupted 8086 program.
Note that with method 5 handling, a mode switch from virtual-8086 mode to
protected mode does not occur. The processor remains in virtual-8086 mode
throughout the interrupt-handling operation.
takes when handling software interrupts in real-address mode. The benefit of using
method 5 handling to access the 8086 program handlers is that it avoids the over-
head of methods 2 and 3 handling, which requires first going to the virtual-8086
monitor, then to the 8086 program handler, then back again to the virtual-8086
monitor, before returning to the interrupted 8086 program (see Section 17.3.1.2,
“Handling an Interrupt or Exception With an 8086 Program Interrupt or Exception
Handler”).
NOTE
Methods 1 and 4 handling can handle a software interrupt in a virtual-
8086 task with a regular protected-mode handler, but this approach
requires all virtual-8086 tasks to use the same software interrupt
handlers, which generally does not give sufficient latitude to the
programs running in the virtual-8086 tasks, particularly MS-DOS
programs.
17.3.3.5 Method 6: Software Interrupt Handling
Method 6 handling is enabled when the VME flag is set to 1, the IOPL value is less
than 3, and the bit for the interrupt or exception vector in the redirection bit map is
same manner as was described for method 5 handling (see Section 17.3.3.4,
“Method 5: Software Interrupt Handling”).
Method 6 differs from method 5 in that with the IOPL value set to less than 3, the VIF
and VIP flags in the EFLAGS register are enabled, providing virtual interrupt support
for handling class 2 maskable hardware interrupts (see Section 17.3.2, “Class
2—Maskable Hardware Interrupt Handling in Virtual-8086 Mode Using the Virtual
Interrupt Mechanism”). These flags provide the virtual-8086 monitor with an effi-
Vol. 3 17-29
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
cient means of handling maskable hardware interrupts that occur during a virtual-
8086 mode task. Also, because the IOPL value is less than 3 and the VIF flag is
enabled, the information pushed on the stack by the processor when invoking the
interrupt handler is slightly different between methods 5 and 6 (see Table 17-2).
17.4
PROTECTED-MODE VIRTUAL INTERRUPTS
The IA-32 processors (beginning with the Pentium processor) also support the VIF
and VIP flags in the EFLAGS register in protected mode by setting the PVI (protected-
mode virtual interrupt) flag in the CR4 register. Setting the PVI flag allows applica-
When the PVI flag is set to 1, the CPL is 3, and the IOPL is less than 3, the STI and
CLI instructions set and clear the VIF flag in the EFLAGS register, leaving IF unaf-
fected. In this mode of operation, an application running in protected mode and at a
CPL of 3 can inhibit interrupts in the same manner as is described in Section 17.3.2,
“Class 2—Maskable Hardware Interrupt Handling in Virtual-8086 Mode Using the
Virtual Interrupt Mechanism”, for a virtual-8086 mode task. When the application
executes the CLI instruction, the processor clears the VIF flag. If the processor
receives a maskable hardware interrupt, the processor invokes the protected-mode
interrupt handler. This handler checks the state of the VIF flag in the EFLAGS register.
If the VIF flag is clear (indicating that the active task does not want to have interrupts
handled now), the handler sets the VIP flag in the EFLAGS image on the stack and
returns to the privilege-level 3 application, which continues program execution.
When the application executes a STI instruction to set the VIF flag, the processor
automatically invokes the general-protection exception handler, which can then
handle the pending interrupt. After handing the pending interrupt, the handler typi-
cally sets the VIF flag and clears the VIP flag in the EFLAGS image on the stack and
executes a return to the application program. The next time the processor receives a
maskable hardware interrupt, the processor will handle it in the normal manner for
interrupts received while the processor is operating at a CPL of 3.
As with the virtual mode extension (enabled with the VME flag in the CR4 register),
the protected-mode virtual interrupt extension only affects maskable hardware
interrupts (interrupt vectors 32 through 255). NMI interrupts and exceptions are
handled in the normal manner.
When protected-mode virtual interrupts are disabled (that is, when the PVI flag in
control register CR4 is set to 0, the CPL is less than 3, or the IOPL value is 3), then
the CLI and STI instructions execute in a manner compatible with the Intel486
processor. That is, if the CPL is greater (less privileged) than the I/O privilege level
(IOPL), a general-protection exception occurs. If the IOPL value is 3, CLI and STI
clear or set the IF flag, respectively.
PUSHF, POPF, IRET and INT are executed like in the Intel486 processor, regardless of
whether protected-mode virtual interrupts are enabled.
17-30 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
It is only possible to enter virtual-8086 mode through a task switch or the execution
of an IRET instruction, and it is only possible to leave virtual-8086 mode by faulting
to a protected-mode interrupt handler (typically the general-protection exception
handler, which in turn calls the virtual 8086-mode monitor). In both cases, the
EFLAGS register is saved and restored. This is not true, however, in protected mode
when the PVI flag is set and the processor is not in virtual-8086 mode. Here, it is
possible to call a procedure at a different privilege level, in which case the EFLAGS
register is not saved or modified. However, the states of VIF and VIP flags are never
examined by the processor when the CPL is not 3.
Vol. 3 17-31
Download from Www.Somanuals.com. All Manuals Search And Download.
8086 EMULATION
17-32 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
CHAPTER 18
MIXING 16-BIT AND 32-BIT CODE
Program modules written to run on IA-32 processors can be either 16-bit modules or
32-bit modules. Table 18-1 shows the characteristic of 16-bit and 32-bit modules.
Table 18-1. Characteristics of 16-Bit and 32-Bit Program Modules
Characteristic
Segment Size
16-Bit Program Modules
0 to 64 KBytes
32-Bit Program Modules
0 to 4 GBytes
Operand Sizes
8 bits and 16 bits
16 bits
8 bits and 32 bits
32 bits
Pointer Offset Size (Address
Size)
Stack Pointer Size
16 Bits
16 Bits
32 Bits
32 Bits
Control Transfers Allowed to
Code Segments of This Size
The IA-32 processors function most efficiently when executing 32-bit program
modules. They can, however, also execute 16-bit program modules, in any of the
following ways:
• In real-address mode.
• In virtual-8086 mode.
• System management mode (SMM).
• As a protected-mode task, when the code, data, and stack segments for the task
are all configured as a 16-bit segments.
• By integrating 16-bit and 32-bit segments into a single protected-mode task.
Real-address mode, virtual-8086 mode, and SMM are native 16-bit modes. A legacy
program assembled and/or compiled to run on an Intel 8086 or Intel 286 processor
should run in real-address mode or virtual-8086 mode without modification. Sixteen-
bit program modules can also be written to run in real-address mode for handling
system initialization or to run in SMM for handling system management functions.
See Chapter 17, “8086 Emulation,” for detailed information on real-address mode
and virtual-8086 mode; see Chapter 26, “System Management Mode,” for informa-
tion on SMM.
This chapter describes how to integrate 16-bit program modules with 32-bit program
modules when operating in protected mode and how to mix 16-bit and 32-bit code
within 32-bit code segments.
Vol. 3 18-1
Download from Www.Somanuals.com. All Manuals Search And Download.
MIXING 16-BIT AND 32-BIT CODE
18.1
DEFINING 16-BIT AND 32-BIT PROGRAM MODULES
The following IA-32 architecture mechanisms are used to distinguish between and
support 16-bit and 32-bit segments and operations:
• The D (default operand and address size) flag in code-segment descriptors.
• The B (default stack size) flag in stack-segment descriptors.
• 16-bit and 32-bit call gates, interrupt gates, and trap gates.
• Operand-size and address-size instruction prefixes.
• 16-bit and 32-bit general-purpose registers.
The D flag in a code-segment descriptor determines the default operand-size and
address-size for the instructions of a code segment. (In real-address mode and
virtual-8086 mode, which do not use segment descriptors, the default is 16 bits.) A
code segment with its D flag set is a 32-bit segment; a code segment with its D flag
clear is a 16-bit segment.
The B flag in the stack-segment descriptor specifies the size of stack pointer (the
32-bit ESP register or the 16-bit SP register) used by the processor for implicit stack
references. The B flag for all data descriptors also controls upper address range for
expand down segments.
When transferring program control to another code segment through a call gate,
interrupt gate, or trap gate, the operand size used during the transfer is determined
by the type of gate used (16-bit or 32-bit), (not by the D-flag or prefix of the transfer
instruction). The gate type determines how return information is saved on the stack
(or stacks).
For most efficient and trouble-free operation of the processor, 32-bit programs or
tasks should have the D flag in the code-segment descriptor and the B flag in the
stack-segment descriptor set, and 16-bit programs or tasks should have these flags
clear. Program control transfers from 16-bit segments to 32-bit segments (and vice
versa) are handled most efficiently through call, interrupt, or trap gates.
Instruction prefixes can be used to override the default operand size and address size
of a code segment. These prefixes can be used in real-address mode as well as in
protected mode and virtual-8086 mode. An operand-size or address-size prefix only
changes the size for the duration of the instruction.
18.2
MIXING 16-BIT AND 32-BIT OPERATIONS WITHIN A
CODE SEGMENT
The following two instruction prefixes allow mixing of 32-bit and 16-bit operations
within one segment:
• The operand-size prefix (66H)
• The address-size prefix (67H)
18-2 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MIXING 16-BIT AND 32-BIT CODE
These prefixes reverse the default size selected by the D flag in the code-segment
descriptor. For example, the processor can interpret the (MOV mem, reg) instruction
in any of four ways:
• In a 32-bit code segment:
— Moves 32 bits from a 32-bit register to memory using a 32-bit effective
address.
— If preceded by an operand-size prefix, moves 16 bits from a 16-bit register to
memory using a 32-bit effective address.
— If preceded by an address-size prefix, moves 32 bits from a 32-bit register to
memory using a 16-bit effective address.
— If preceded by both an address-size prefix and an operand-size prefix, moves
16 bits from a 16-bit register to memory using a 16-bit effective address.
• In a 16-bit code segment:
— Moves 16 bits from a 16-bit register to memory using a 16-bit effective
address.
— If preceded by an operand-size prefix, moves 32 bits from a 32-bit register to
memory using a 16-bit effective address.
— If preceded by an address-size prefix, moves 16 bits from a 16-bit register to
memory using a 32-bit effective address.
— If preceded by both an address-size prefix and an operand-size prefix, moves
32 bits from a 32-bit register to memory using a 32-bit effective address.
The previous examples show that any instruction can generate any combination of
operand size and address size regardless of whether the instruction is in a 16- or
32-bit segment. The choice of the 16- or 32-bit default for a code segment is
normally based on the following criteria:
• Performance — Always use 32-bit code segments when possible. They run
much faster than 16-bit code segments on P6 family processors, and somewhat
faster on earlier IA-32 processors.
• The operating system the code segment will be running on — If the
operating system is a 16-bit operating system, it may not support 32-bit program
modules.
• Mode of operation — If the code segment is being designed to run in real-
address mode, virtual-8086 mode, or SMM, it must be a 16-bit code segment.
• Backward compatibility to earlier IA-32 processors — If a code segment
must be able to run on an Intel 8086 or Intel 286 processor, it must be a 16-bit
code segment.
Vol. 3 18-3
Download from Www.Somanuals.com. All Manuals Search And Download.
MIXING 16-BIT AND 32-BIT CODE
18.3
SHARING DATA AMONG MIXED-SIZE CODE
SEGMENTS
Data segments can be accessed from both 16-bit and 32-bit code segments. When a
data segment that is larger than 64 KBytes is to be shared among 16- and 32-bit
code segments, the data that is to be accessed from the 16-bit code segments must
be located within the first 64 KBytes of the data segment. The reason for this is that
16-bit pointers by definition can only point to the first 64 KBytes of a segment.
A stack that spans less than 64 KBytes can be shared by both 16- and 32-bit code
segments. This class of stacks includes:
• Stacks in expand-up segments with the G (granularity) and B (big) flags in the
• Stacks in expand-down segments with the G and B flags clear.
• Stacks in expand-up segments with the G flag set and the B flag clear and where
the stack is contained completely within the lower 64 KBytes. (Offsets greater
than FFFFH can be used for data, other than the stack, which is not shared.)
See Section 3.4.5, “Segment Descriptors,” for a description of the G and B flags and
the expand-down stack type.
The B flag cannot, in general, be used to change the size of stack used by a 16-bit
code segment. This flag controls the size of the stack pointer only for implicit stack
references such as those caused by interrupts, exceptions, and the PUSH, POP, CALL,
and RET instructions. It does not control explicit stack references, such as accesses
to parameters or local variables. A 16-bit code segment can use a 32-bit stack only if
the code is modified so that all explicit references to the stack are preceded by the
32-bit address-size prefix, causing those references to use 32-bit addressing and
explicit writes to the stack pointer are preceded by a 32-bit operand-size prefix.
In 32-bit, expand-down segments, all offsets may be greater than 64 KBytes; there-
fore, 16-bit code cannot use this kind of stack segment unless the code segment is
modified to use 32-bit addressing.
18.4
TRANSFERRING CONTROL AMONG MIXED-SIZE CODE
SEGMENTS
There are three ways for a procedure in a 16-bit code segment to safely make a call
to a 32-bit code segment:
• Make the call through a 32-bit call gate.
• Make a 16-bit call to a 32-bit interface procedure. The interface procedure then
makes a 32-bit call to the intended destination.
• Modify the 16-bit procedure, inserting an operand-size prefix before the call, to
change it to a 32-bit call.
18-4 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MIXING 16-BIT AND 32-BIT CODE
Likewise, there are three ways for procedure in a 32-bit code segment to safely make
a call to a 16-bit code segment:
• Make the call through a 16-bit call gate. Here, the EIP value at the CALL
instruction cannot exceed FFFFH.
• Make a 32-bit call to a 16-bit interface procedure. The interface procedure then
makes a 16-bit call to the intended destination.
• Modify the 32-bit procedure, inserting an operand-size prefix before the call,
changing it to a 16-bit call. Be certain that the return offset does not exceed
FFFFH.
These methods of transferring program control overcome the following architectural
limitations imposed on calls between 16-bit and 32-bit code segments:
• Pointers from 16-bit code segments (which by default can only be 16 bits) cannot
be used to address data or code located beyond FFFFH in a 32-bit segment.
• The operand-size attributes for a CALL and its companion RETURN instruction
must be the same to maintain stack coherency. This is also true for implicit calls
to interrupt and exception handlers and their companion IRET instructions.
• A 32-bit parameters (particularly a pointer parameter) greater than FFFFH
cannot be squeezed into a 16-bit parameter location on a stack.
• The size of the stack pointer (SP or ESP) changes when switching between 16-bit
and 32-bit code segments.
These limitations are discussed in greater detail in the following sections.
18.4.1
Code-Segment Pointer Size
For control-transfer instructions that use a pointer to identify the next instruction
(that is, those that do not use gates), the operand-size attribute determines the size
of the offset portion of the pointer. The implications of this rule are as follows:
always possible using a 32-bit operand size, providing the 32-bit pointer does not
exceed FFFFH.
• A JMP, CALL, or RET instruction from a 16-bit segment to a 32-bit segment
cannot address a destination greater than FFFFH, unless the instruction is given
an operand-size prefix.
See Section 18.4.5, “Writing Interface Procedures,” for an interface procedure that
can transfer program control from 16-bit segments to destinations in 32-bit
segments beyond FFFFH.
18.4.2
Stack Management for Control Transfer
Because the stack is managed differently for 16-bit procedure calls than for 32-bit
calls, the operand-size attribute of the RET instruction must match that of the CALL
Vol. 3 18-5
Download from Www.Somanuals.com. All Manuals Search And Download.
MIXING 16-BIT AND 32-BIT CODE
instruction (see Figure 18-1). On a 16-bit call, the processor pushes the contents of
the 16-bit IP register and (for calls between privilege levels) the 16-bit SP register.
The matching RET instruction must also use a 16-bit operand size to pop these 16-bit
values from the stack into the 16-bit registers.
A 32-bit CALL instruction pushes the contents of the 32-bit EIP register and (for
inter-privilege-level calls) the 32-bit ESP register. Here, the matching RET instruction
must use a 32-bit operand size to pop these 32-bit values from the stack into the
32-bit registers. If the two parts of a CALL/RET instruction pair do not have matching
operand sizes, the stack will not be managed correctly and the values of the instruc-
tion pointer and stack pointer will not be restored to correct values.
Without Privilege Transition
After 16-bit Call
After 32-bit Call
31
0
31
0
PARM 2 PARM 1
CS IP
PARM 2
PARM 1
Stack
Growth
SP
CS
EIP
ESP
With Privilege Transition
After 16-bit Call
After 32-bit Call
31
0
31
0
SS
PARM 2 PARM 1
CS IP
SP
SS
ESP
Stack
Growth
SP
PARM 2
PARM 1
CS
EIP
ESP
Undefined
Figure 18-1. Stack after Far 16- and 32-Bit Calls
18-6 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MIXING 16-BIT AND 32-BIT CODE
While executing 32-bit code, if a call is made to a 16-bit code segment which is at the
same or a more privileged level (that is, the DPL of the called code segment is less
than or equal to the CPL of the calling code segment) through a 16-bit call gate, then
the upper 16-bits of the ESP register may be unreliable upon returning to the 32-bit
code segment (that is, after executing a RET in the 16-bit code segment).
When the CALL instruction and its matching RET instruction are in code segments
that have D flags with the same values (that is, both are 32-bit code segments or
both are 16-bit code segments), the default settings may be used. When the CALL
instruction and its matching RET instruction are in segments which have different
D-flag settings, an operand-size prefix must be used.
18.4.2.1 Controlling the Operand-Size Attribute For a Call
Three things can determine the operand-size of a call:
• The D flag in the segment descriptor for the calling code segment.
• An operand-size instruction prefix.
• The type of call gate (16-bit or 32-bit), if a call is made through a call gate.
When a call is made with a pointer (rather than a call gate), the D flag for the calling
code segment determines the operand-size for the CALL instruction. This operand-
size attribute can be overridden by prepending an operand-size prefix to the CALL
instruction. So, for example, if the D flag for a code segment is set for 16 bits and the
operand-size prefix is used with a CALL instruction, the processor will cause the infor-
mation stored on the stack to be stored in 32-bit format. If the call is to a 32-bit code
segment, the instructions in that code segment will be able to read the stack coher-
ently. Also, a RET instruction from the 32-bit code segment without an operand-size
prefix will maintain stack coherency with the 16-bit code segment being returned to.
When a CALL instruction references a call-gate descriptor, the type of call is deter-
mined by the type of call gate (16-bit or 32-bit). The offset to the destination in the
code segment being called is taken from the gate descriptor; therefore, if a 32-bit call
gate is used, a procedure in a 16-bit code segment can call a procedure located more
than 64 KBytes from the base of a 32-bit code segment, because a 32-bit call gate
uses a 32-bit offset.
Note that regardless of the operand size of the call and how it is determined, the size
of the stack pointer used (SP or ESP) is always controlled by the B flag in the stack-
segment descriptor currently in use (that is, when B is clear, SP is used, and when B
is set, ESP is used).
An unmodified 16-bit code segment that has run successfully on an 8086 processor
or in real-mode on a later IA-32 architecture processor will have its D flag clear and
will not use operand-size override prefixes. As a result, all CALL instructions in this
code segment will use the 16-bit operand-size attribute. Procedures in these code
Vol. 3 18-7
Download from Www.Somanuals.com. All Manuals Search And Download.
MIXING 16-BIT AND 32-BIT CODE
segments can be modified to safely call procedures to 32-bit code segments in either
of two ways:
• Relink the CALL instruction to point to 32-bit call gates (see Section 18.4.2.2,
“Passing Parameters With a Gate”).
• Add a 32-bit operand-size prefix to each CALL instruction.
18.4.2.2 Passing Parameters With a Gate
When referencing 32-bit gates with 16-bit procedures, it is important to consider the
number of parameters passed in each procedure call. The count field of the gate
descriptor specifies the size of the parameter string to copy from the current stack to
the stack of a more privileged (numerically lower privilege level) procedure. The
count field of a 16-bit gate specifies the number of 16-bit words to be copied,
whereas the count field of a 32-bit gate specifies the number of 32-bit doublewords
to be copied. The count field for a 32-bit gate must thus be half the size of the
number of words being placed on the stack by a 16-bit procedure. Also, the 16-bit
procedure must use an even number of words as parameters.
18.4.3
Interrupt Control Transfers
A program-control transfer caused by an exception or interrupt is always carried out
through an interrupt or trap gate (located in the IDT). Here, the type of the gate
(16-bit or 32-bit) determines the operand-size attribute used in the implicit call to
the exception or interrupt handler procedure in another code segment.
A 32-bit interrupt or trap gate provides a safe interface to a 32-bit exception or inter-
rupt handler when the exception or interrupt occurs in either a 32-bit or a 16-bit code
segment. It is sometimes impractical, however, to place exception or interrupt
handlers in 16-bit code segments, because only 16-bit return addresses are saved on
the stack. If an exception or interrupt occurs in a 32-bit code segment when the EIP
was greater than FFFFH, the 16-bit handler procedure cannot provide the correct
return address.
18.4.4
Parameter Translation
When segment offsets or pointers (which contain segment offsets) are passed as
parameters between 16-bit and 32-bit procedures, some translation is required. If a
32-bit procedure passes a pointer to data located beyond 64 KBytes to a 16-bit
procedure, the 16-bit procedure cannot use it. Except for this limitation, interface
code can perform any format conversion between 32-bit and 16-bit pointers that
may be needed.
Parameters passed by value between 32-bit and 16-bit code also may require trans-
lation between 32-bit and 16-bit formats. The form of the translation is application-
dependent.
18-8 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
MIXING 16-BIT AND 32-BIT CODE
18.4.5
Writing Interface Procedures
Placing interface code between 32-bit and 16-bit procedures can be the solution to
the following interface problems:
• Allowing procedures in 16-bit code segments to call procedures with offsets
greater than FFFFH in 32-bit code segments.
• Matching operand-size attributes between companion CALL and RET instructions.
• Translating parameters (data), including managing parameter strings with a
variable count or an odd number of 16-bit words.
• The possible invalidation of the upper bits of the ESP register.
The interface procedure is simplified where these rules are followed.
1. The interface procedure must reside in a 32-bit code segment (the D flag for the
code-segment descriptor is set).
2. All procedures that may be called by 16-bit procedures must have offsets not
greater than FFFFH.
3. All return addresses saved by 16-bit procedures must have offsets not greater
than FFFFH.
The interface procedure becomes more complex if any of these rules are violated. For
example, if a 16-bit procedure calls a 32-bit procedure with an entry point beyond
FFFFH, the interface procedure will need to provide the offset to the entry point. The
mapping between 16- and 32-bit addresses is only performed automatically when a
call gate is used, because the gate descriptor for a call gate contains a 32-bit
address. When a call gate is not used, the interface code must provide the 32-bit
address.
The structure of the interface procedure depends on the types of calls it is going to
support, as follows:
• Calls from 16-bit procedures to 32-bit procedures — Calls to the interface
procedure from a 16-bit code segment are made with 16-bit CALL instructions
(by default, because the D flag for the calling code-segment descriptor is clear),
and 16-bit operand-size prefixes are used with RET instructions to return from
the interface procedure to the calling procedure. Calls from the interface
procedure to 32-bit procedures are performed with 32-bit CALL instructions (by
default, because the D flag for the interface procedure’s code segment is set),
and returns from the called procedures to the interface procedure are performed
with 32-bit RET instructions (also by default).
• Calls from 32-bit procedures to 16-bit procedures — Calls to the interface
procedure from a 32-bit code segment are made with 32-bit CALL instructions
(by default), and returns to the calling procedure from the interface procedure
are made with 32-bit RET instructions (also by default). Calls from the interface
procedure to 16-bit procedures require the CALL instructions to have the
operand-size prefixes, and returns from the called procedures to the interface
procedure are performed with 16-bit RET instructions (by default).
Vol. 3 18-9
Download from Www.Somanuals.com. All Manuals Search And Download.
MIXING 16-BIT AND 32-BIT CODE
18-10 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
CHAPTER 19
ARCHITECTURE COMPATIBILITY
Intel 64 and IA-32 processors are binary compatible. Compatibility means that,
within limited constraints, programs that execute on previous generations of proces-
sors will produce identical results when executed on later processors. The compati-
bility constraints and any implementation differences between the Intel 64 and IA-32
processors are described in this chapter.
Each new processor has enhanced the software visible architecture from that found
in earlier Intel 64 and IA-32 processors. Those enhancements have been defined
with consideration for compatibility with previous and future processors. This chapter
also summarizes the compatibility considerations for those extensions.
19.1
PROCESSOR FAMILIES AND CATEGORIES
IA-32 processors are referred to in several different ways in this chapter, depending
on the type of compatibility information being related, as described in the following:
• IA-32 Processors — All the Intel processors based on the Intel IA-32 Archi-
tecture, which include the 8086/88, Intel 286, Intel386, Intel486, Pentium,
Pentium Pro, Pentium II, Pentium III, Pentium 4, and Intel Xeon processors.
• 32-bit Processors — All the IA-32 processors that use a 32-bit architecture,
which include the Intel386, Intel486, Pentium, Pentium Pro, Pentium II,
Pentium III, Pentium 4, and Intel Xeon processors.
• 16-bit Processors — All the IA-32 processors that use a 16-bit architecture,
which include the 8086/88 and Intel 286 processors.
• P6 Family Processors — All the IA-32 processors that are based on the P6
microarchitecture, which include the Pentium Pro, Pentium II, and Pentium III
processors.
• Pentium 4 Processors — A family of IA-32 and Intel 64 processors that are
based on the Intel NetBurst microarchitecture.
• Intel Pentium M Processors — A family of IA-32 processors that are based on
the Intel Pentium M processor microarchitecture.
• Intel Core Duo and Solo Processors — Families of IA-32 processors that are
based on an improved Intel Pentium M processor microarchitecture.
• Intel Xeon Processors — A family of IA-32 and Intel 64 processors that are
based on the Intel NetBurst microarchitecture. This family includes the Intel Xeon
processor and the Intel Xeon processor MP based on the Intel NetBurst microar-
chitecture. Intel Xeon processors 3000, 3100, 3200, 3300, 3200, 5100, 5200,
5300, 5400, 7200, 7300 series are based on Intel Core microarchitectures and
support Intel 64 architecture.
Vol. 3 19-1
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
• Pentium D Processors — A family of dual-core Intel 64 processors that
provides two processor cores in a physical package. Each core is based on the
Intel NetBurst microarchitecture.
• Pentium Processor Extreme Editions — A family of dual-core Intel 64
processors that provides two processor cores in a physical package. Each core is
based on the Intel NetBurst microarchitecture and supports Intel Hyper-
Threading Technology.
• Intel Core 2 Processors — A family of Intel 64 processors that are based on the
Intel Core microarchitecture. Intel Pentium Dual-Core processors are also based
on the Intel Core microarchitecture.
• Intel Atom Processors — A family of IA-32 and Intel 64 processors that are
based on the Intel Atom microarchitecture.
19.2
RESERVED BITS
Throughout this manual, certain bits are marked as reserved in many register and
memory layout descriptions. When bits are marked as undefined or reserved, it is
essential for compatibility with future processors that software treat these bits as
having a future, though unknown effect. Software should follow these guidelines in
dealing with reserved bits:
• Do not depend on the states of any reserved bits when testing the values of
registers or memory locations that contain such bits. Mask out the reserved bits
before testing.
• Do not depend on the states of any reserved bits when storing them to memory
or to a register.
• Do not depend on the ability to retain information written into any reserved bits.
• When loading a register, always load the reserved bits with the values indicated
in the documentation, if any, or reload them with values previously read from the
same register.
Software written for existing IA-32 processor that handles reserved bits correctly will
port to future IA-32 processors without generating protection exceptions.
19.3
ENABLING NEW FUNCTIONS AND MODES
Most of the new control functions defined for the P6 family and Pentium processors
are enabled by new mode flags in the control registers (primarily register CR4). This
register is undefined for IA-32 processors earlier than the Pentium processor.
Attempting to access this register with an Intel486 or earlier IA-32 processor results
in an invalid-opcode exception (#UD). Consequently, programs that execute
correctly on the Intel486 or earlier IA-32 processor cannot erroneously enable these
functions. Attempting to set a reserved bit in register CR4 to a value other than its
19-2 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
original value results in a general-protection exception (#GP). So, programs that
execute on the P6 family and Pentium processors cannot erroneously enable func-
tions that may be implemented in future IA-32 processors.
The P6 family and Pentium processors do not check for attempts to set reserved bits
in model-specific registers; however these bits may be checked on more recent
processors. It is the obligation of the software writer to enforce this discipline. These
reserved bits may be used in future Intel processors.
19.4
DETECTING THE PRESENCE OF NEW FEATURES
THROUGH SOFTWARE
Software can check for the presence of new architectural features and extensions in
either of two ways:
1. Test for the presence of the feature or extension. Software can test for the
presence of new flags in the EFLAGS register and control registers. If these flags
are reserved (meaning not present in the processor executing the test), an
exception is generated. Likewise, software can attempt to execute a new
instruction, which results in an invalid-opcode exception (#UD) being generated
if it is not supported.
2. Execute the CPUID instruction. The CPUID instruction (added to the IA-32 in the
Pentium processor) indicates the presence of new features directly.
See Chapter 14, “Processor Identification and Feature Determination,” in the Intel®
64 and IA-32 Architectures Software Developer’s Manual, Volume 1, for detailed
information on detecting new processor features and extensions.
19.5
INTEL MMX TECHNOLOGY
The Pentium processor with MMX technology introduced the MMX technology and a
set of MMX instructions to the IA-32. The MMX instructions are described in Chapter
9, “Programming with Intel® MMX™ Technology,” in the Intel® 64 and IA-32 Archi-
tectures Software Developer’s Manual, Volume 1, and in the Intel® 64 and IA-32
Architectures Software Developer’s Manual, Volumes 2A & 2B. The MMX technology
and MMX instructions are also included in the Pentium II, Pentium III, Pentium 4, and
Intel Xeon processors.
19.6
STREAMING SIMD EXTENSIONS (SSE)
The Streaming SIMD Extensions (SSE) were introduced in the Pentium III processor.
The SSE extensions consist of a new set of instructions and a new set of registers.
The new registers include the eight 128-bit XMM registers and the 32-bit MXCSR
Vol. 3 19-3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
control and status register. These instructions and registers are designed to allow
SIMD computations to be made on single-precision floating-point numbers. Several
of these new instructions also operate in the MMX registers. SSE instructions and
registers are described in Section 10, “Programming with Streaming SIMD Exten-
sions (SSE),” in the Intel® 64 and IA-32 Architectures Software Developer’s Manual,
Volume 1, and in the Intel® 64 and IA-32 Architectures Software Developer’s
Manual, Volumes 2A & 2B.
19.7
STREAMING SIMD EXTENSIONS 2 (SSE2)
The Streaming SIMD Extensions 2 (SSE2) were introduced in the Pentium 4 and Intel
Xeon processors. They consist of a new set of instructions that operate on the XMM
and MXCSR registers and perform SIMD operations on double-precision floating-
point values and on integer values. Several of these new instructions also operate in
the MMX registers. SSE2 instructions and registers are described in Chapter 11,
“Programming with Streaming SIMD Extensions 2 (SSE2),” in the Intel® 64 and
IA-32 Architectures Software Developer’s Manual, Volume 1, and in the Intel® 64
and IA-32 Architectures Software Developer’s Manual, Volumes 2A & 2B.
19.8
STREAMING SIMD EXTENSIONS 3 (SSE3)
The Streaming SIMD Extensions 3 (SSE3) were introduced in Pentium 4 processors
supporting Intel Hyper-Threading Technology and Intel Xeon processors. SSE3
extensions include 13 instructions. Ten of these 13 instructions support the single
instruction multiple data (SIMD) execution model used with SSE/SSE2 extensions.
One SSE3 instruction accelerates x87 style programming for conversion to integer.
The remaining two instructions (MONITOR and MWAIT) accelerate synchronization
of threads. SSE3 instructions are described in Chapter 12, “Programming with SSE3,
SSSE3 and SSE4,” in the Intel® 64 and IA-32 Architectures Software Developer’s
Manual, Volume 1, and in the Intel® 64 and IA-32 Architectures Software Devel-
oper’s Manual, Volumes 2A & 2B.
19.9
ADDITIONAL STREAMING SIMD EXTENSIONS
The Supplemental Streaming SIMD Extensions 3 (SSSE3) were introduced in the
Intel Core 2 processor and Intel Xeon processor 5100 series. Streaming SIMD Exten-
sions 4 provided 54 new instructions introduced in 45nm Intel Xeon processors and
Intel Core 2 processors. SSSE3, SSE4.1 and SSE4.2 instructions are described in
Chapter 12, “Programming with SSE3, SSSE3 and SSE4,” in the Intel® 64 and IA-32
Architectures Software Developer’s Manual, Volume 1, and in the Intel® 64 and
IA-32 Architectures Software Developer’s Manual, Volumes 2A & 2B.
19-4 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
19.10 INTEL HYPER-THREADING TECHNOLOGY
two separate code streams (called threads) concurrently by using shared resources
in a single processor core or in a physical package.
This feature was introduced in the Intel Xeon processor MP and later steppings of the
Intel Xeon processor, and Pentium 4 processors supporting Intel Hyper-Threading
Technology. The feature is also found in the Pentium processor Extreme Edition. See
®
also: Section 8.7, “Intel Hyper-Threading Technology Architecture.”
Intel Atom processors also support Intel Hyper-Threading Technology.
19.11 MULTI-CORE TECHNOLOGY
The Pentium D processor and Pentium processor Extreme Edition provide two
®
processor cores in each physical processor package. See also: Section 8.5, “Intel
®
Hyper-Threading Technology and Intel Multi-Core Technology,” and Section 8.8,
“Multi-Core Architecture.” Intel Core 2 Duo, Intel Pentium Dual-Core processors,
Intel Xeon processors 3000, 3100, 5100, 5200 series provide two processor cores in
each physical processor package. Intel Core 2 Extreme, Intel Core 2 Quad proces-
sors, Intel Xeon processors 3200, 3300, 5300, 5400, 7300 series provide two
processor cores in each physical processor package.
19.12 SPECIFIC FEATURES OF DUAL-CORE PROCESSOR
Dual-core processors may have some processor-specific features. Use CPUID feature
flags to detect the availability features. Note the following:
• CPUID Brand String — On Pentium processor Extreme Edition, the process will
report the correct brand string only after the correct microcode updates are
loaded.
• Enhanced Intel SpeedStep Technology — This feature is supported in
Pentium D processor but not in Pentium processor Extreme Edition.
19.13 NEW INSTRUCTIONS IN THE PENTIUM AND LATER
IA-32 PROCESSORS
Table 19-1 identifies the instructions introduced into the IA-32 in the Pentium
processor and later IA-32 processors.
Vol. 3 19-5
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
19.13.1 Instructions Added Prior to the Pentium Processor
The following instructions were added in the Intel486 processor:
• BSWAP (byte swap) instruction.
• XADD (exchange and add) instruction.
• CMPXCHG (compare and exchange) instruction.
•
ΙNVD (invalidate cache) instruction.
• WBINVD (write-back and invalidate cache) instruction.
• INVLPG (invalidate TLB entry) instruction.
Table 19-1. New Instruction in the Pentium Processor and
Later IA-32 Processors
Instruction
CPUID Identification Bits
EDX, Bit 15
Introduced In
CMOVcc (conditional move)
Pentium Pro processor
FCMOVcc (floating-point conditional
EDX, Bits 0 and 15
move)
FCOMI (floating-point compare and set
EFLAGS)
EDX, Bits 0 and 15
RDPMC (read performance monitoring
counters)
EAX, Bits 8-11, set to 6H;
see Note 1
UD2 (undefined)
EAX, Bits 8-11, set to 6H
EDX, Bit 8
CMPXCHG8B (compare and exchange 8
bytes)
Pentium processor
CPUID (CPU identification)
RDTSC (read time-stamp counter)
RDMSR (read model-specific register)
WRMSR (write model-specific register)
MMX Instructions
None; see Note 2
EDX, Bit 4
EDX, Bit 5
EDX, Bit 5
EDX, Bit 23
NOTES:
1. The RDPMC instruction was introduced in the P6 family of processors and added to later model
Pentium processors. This instruction is model specific in nature and not architectural.
2. The CPUID instruction is available in all Pentium and P6 family processors and in later models of
the Intel486 processors. The ability to set and clear the ID flag (bit 21) in the EFLAGS register
indicates the availability of the CPUID instruction.
The following instructions were added in the Intel386 processor:
• LSS, LFS, and LGS (load SS, FS, and GS registers).
• Long-displacement conditional jumps.
19-6 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
• Single-bit instructions.
• Bit scan instructions.
• Double-shift instructions.
• Byte set on condition instruction.
• Move with sign/zero extension.
• Generalized multiply instruction.
• MOV to and from control registers.
• MOV to and from test registers (now obsolete).
• MOV to and from debug registers.
• RSM (resume from SMM). This instruction was introduced in the Intel386 SL and
Intel486 SL processors.
The following instructions were added in the Intel 387 math coprocessor:
• FPREM1.
• FUCOM, FUCOMP, and FUCOMPP.
19.14 OBSOLETE INSTRUCTIONS
The MOV to and from test registers instructions were removed from the Pentium
processor and future IA-32 processors. Execution of these instructions generates an
invalid-opcode exception (#UD).
19.15 UNDEFINED OPCODES
All new instructions defined for IA-32 processors use binary encodings that were
reserved on earlier-generation processors. Attempting to execute a reserved opcode
always results in an invalid-opcode (#UD) exception being generated. Consequently,
programs that execute correctly on earlier-generation processors cannot erroneously
execute these instructions and thereby produce unexpected results when executed
on later IA-32 processors.
19.16 NEW FLAGS IN THE EFLAGS REGISTER
The section titled “EFLAGS Register” in Chapter 3, “Basic Execution Environment,” of
the Intel® 64 and IA-32 Architectures Software Developer’s Manual, Volume 1,
shows the configuration of flags in the EFLAGS register for the P6 family processors.
No new flags have been added to this register in the P6 family processors. The flags
added to this register in the Pentium and Intel486 processors are described in the
following sections.
Vol. 3 19-7
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
The following flags were added to the EFLAGS register in the Pentium processor:
• VIF (virtual interrupt flag), bit 19.
• VIP (virtual interrupt pending), bit 20.
• ID (identification flag), bit 21.
The AC flag (bit 18) was added to the EFLAGS register in the Intel486 processor.
19.16.1 Using EFLAGS Flags to Distinguish Between 32-Bit IA-32
Processors
The following bits in the EFLAGS register that can be used to differentiate between
the 32-bit IA-32 processors:
• Bit 18 (the AC flag) can be used to distinguish an Intel386 processor from the P6
family, Pentium, and Intel486 processors. Since it is not implemented on the
Intel386 processor, it will always be clear.
• Bit 21 (the ID flag) indicates whether an application can execute the CPUID
instruction. The ability to set and clear this bit indicates that the processor is a P6
family or Pentium processor. The CPUID instruction can then be used to
determine which processor.
• Bits 19 (the VIF flag) and 20 (the VIP flag) will always be zero on processors that
do not support virtual mode extensions, which includes all 32-bit processors prior
to the Pentium processor.
See Chapter 14, “Processor Identification and Feature Determination,” in the Intel®
64 and IA-32 Architectures Software Developer’s Manual, Volume 1, for more infor-
mation on identifying processors.
19.17 STACK OPERATIONS
This section identifies the differences in stack implementation between the various
IA-32 processors.
19.17.1 PUSH SP
The P6 family, Pentium, Intel486, Intel386, and Intel 286 processors push a different
value on the stack for a PUSH SP instruction than the 8086 processor. The 32-bit
processors push the value of the SP register before it is decremented as part of the
push operation; the 8086 processor pushes the value of the SP register after it is
decremented. If the value pushed is important, replace PUSH SP instructions with the
following three instructions:
PUSH BP
MOV BP, SP
19-8 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
XCHG BP, [BP]
This code functions as the 8086 processor PUSH SP instruction on the P6 family,
Pentium, Intel486, Intel386, and Intel 286 processors.
19.17.2 EFLAGS Pushed on the Stack
The setting of the stored values of bits 12 through 15 (which includes the IOPL field
and the NT flag) in the EFLAGS register by the PUSHF instruction, by interrupts, and
by exceptions is different with the 32-bit IA-32 processors than with the 8086 and
Intel 286 processors. The differences are as follows:
• 8086 processor—bits 12 through 15 are always set.
• Intel 286 processor—bits 12 through 15 are always cleared in real-address mode.
• 32-bit processors in real-address mode—bit 15 (reserved) is always cleared, and
bits 12 through 14 have the last value loaded into them.
19.18 X87 FPU
This section addresses the issues that must be faced when porting floating-point
software designed to run on earlier IA-32 processors and math coprocessors to a
Pentium 4, Intel Xeon, P6 family, or Pentium processor with integrated x87 FPU. To
software, a Pentium 4, Intel Xeon, or P6 family processor looks very much like a
Pentium processor. Floating-point software which runs on a Pentium or Intel486 DX
processor, or on an Intel486 SX processor/Intel 487 SX math coprocessor system or
an Intel386 processor/Intel 387 math coprocessor system, will run with at most
minor modifications on a Pentium 4, Intel Xeon, or P6 family processor. To port code
directly from an Intel 286 processor/Intel 287 math coprocessor system or an
Intel 8086 processor/8087 math coprocessor system to a Pentium 4, Intel Xeon, P6
family, or Pentium processor, certain additional issues must be addressed.
In the following sections, the term “32-bit x87 FPUs” refers to the P6 family, Pentium,
and Intel486 DX processors, and to the Intel 487 SX and Intel 387 math coproces-
sors; the term “16-bit IA-32 math coprocessors” refers to the Intel 287 and 8087
math coprocessors.
19.18.1 Control Register CR0 Flags
The ET, NE, and MP flags in control register CR0 control the interface between the
integer unit of an IA-32 processor and either its internal x87 FPU or an external math
coprocessor. The effect of these flags in the various IA-32 processors are described in
the following paragraphs.
The ET (extension type) flag (bit 4 of the CR0 register) is used in the Intel386
processor to indicate whether the math coprocessor in the system is an Intel 287
Vol. 3 19-9
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
math coprocessor (flag is clear) or an Intel 387 DX math coprocessor (flag is set).
This bit is hardwired to 1 in the P6 family, Pentium, and Intel486 processors.
The NE (Numeric Exception) flag (bit 5 of the CR0 register) is used in the P6 family,
Pentium, and Intel486 processors to determine whether unmasked floating-point
exceptions are reported internally through interrupt vector 16 (flag is set) or exter-
nally through an external interrupt (flag is clear). On a hardware reset, the NE flag is
initialized to 0, so software using the automatic internal error-reporting mechanism
must set this flag to 1. This flag is nonexistent on the Intel386 processor.
As on the Intel 286 and Intel386 processors, the MP (monitor coprocessor) flag (bit 1
of register CR0) determines whether the WAIT/FWAIT instructions or waiting-type
floating-point instructions trap when the context of the x87 FPU is different from that
of the currently-executing task. If the MP and TS flag are set, then a WAIT/FWAIT
instruction and waiting instructions will cause a device-not-available exception
(interrupt vector 7). The MP flag is used on the Intel 286 and Intel386 processors to
support the use of a WAIT/FWAIT instruction to wait on a device other than a math
coprocessor. The device reports its status through the BUSY# pin. Since the P6
family, Pentium, and Intel486 processors do not have such a pin, the MP flag has no
relevant use and should be set to 1 for normal operation.
19.18.2 x87 FPU Status Word
This section identifies differences to the x87 FPU status word for the different IA-32
processors and math coprocessors, the reason for the differences, and their impact
on software.
19.18.2.1 Condition Code Flags (C0 through C3)
The following information pertains to differences in the use of the condition code
flags (C0 through C3) located in bits 8, 9, 10, and 14 of the x87 FPU status word.
After execution of an FINIT instruction or a hardware reset on a 32-bit x87 FPU, the
condition code flags are set to 0. The same operations on a 16-bit IA-32 math copro-
cessor leave these flags intact (they contain their prior value). This difference in
operation has no impact on software and provides a consistent state after reset.
Transcendental instruction results in the core range of the P6 family and Pentium
processors may differ from the Intel486 DX processor and Intel 487 SX math copro-
cessor by 2 to 3 units in the last place (ulps)—(see “Transcendental Instruction Accu-
racy” in Chapter 8, “Programming with the x87 FPU,” of the Intel® 64 and IA-32
Architectures Software Developer’s Manual, Volume 1). As a result, the value saved
in the C1 flag may also differ.
After an incomplete FPREM/FPREM1 instruction, the C0, C1, and C3 flags are set to 0
on the 32-bit x87 FPUs. After the same operation on a 16-bit IA-32 math copro-
cessor, these flags are left intact.
19-10 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
On the 32-bit x87 FPUs, the C2 flag serves as an incomplete flag for the FTAN instruc-
tion. On the 16-bit IA-32 math coprocessors, the C2 flag is undefined for the FPTAN
instruction. This difference has no impact on software, because Intel 287 or 8087
programs do not check C2 after an FPTAN instruction. The use of this flag on later
processors allows fast checking of operand range.
19.18.2.2 Stack Fault Flag
When unmasked stack overflow or underflow occurs on a 32-bit x87 FPU, the IE flag
(bit 0) and the SF flag (bit 6) of the x87 FPU status word are set to indicate a stack
fault and condition code flag C1 is set or cleared to indicate overflow or underflow,
respectively. When unmasked stack overflow or underflow occurs on a 16-bit IA-32
math coprocessor, only the IE flag is set. Bit 6 is reserved on these processors. The
addition of the SF flag on a 32-bit x87 FPU has no impact on software. Existing excep-
tion handlers need not change, but may be upgraded to take advantage of the addi-
tional information.
19.18.3 x87 FPU Control Word
Only affine closure is supported for infinity control on a 32-bit x87 FPU. The infinity
control flag (bit 12 of the x87 FPU control word) remains programmable on these
processors, but has no effect. This change was made to conform to the IEEE Stan-
dard 754 for Binary Floating-Point Arithmetic. On a 16-bit IA-32 math coprocessor,
both affine and projective closures are supported, as determined by the setting of bit
12. After a hardware reset, the default value of bit 12 is projective. Software that
requires projective infinity arithmetic may give different results.
19.18.4 x87 FPU Tag Word
When loading the tag word of a 32-bit x87 FPU, using an FLDENV, FRSTOR, or
FXRSTOR (Pentium III processor only) instruction, the processor examines the
incoming tag and classifies the location only as empty or non-empty. Thus, tag
values of 00, 01, and 10 are interpreted by the processor to indicate a non-empty
location. The tag value of 11 is interpreted by the processor to indicate an empty
location. Subsequent operations on a non-empty register always examine the value
in the register, not the value in its tag. The FSTENV, FSAVE, and FXSAVE (Pentium III
processor only) instructions examine the non-empty registers and put the correct
values in the tags before storing the tag word.
The corresponding tag for a 16-bit IA-32 math coprocessor is checked before each
register access to determine the class of operand in the register; the tag is updated
after every change to a register so that the tag always reflects the most recent status
of the register. Software can load a tag with a value that disagrees with the contents
of a register (for example, the register contains a valid value, but the tag says
special). Here, the 16-bit IA-32 math coprocessors honor the tag and do not examine
the register.
Vol. 3 19-11
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
Software written to run on a 16-bit IA-32 math coprocessor may not operate
correctly on a 16-bit x87 FPU, if it uses the FLDENV, FRSTOR, or FXRSTOR instruc-
tions to change tags to values (other than to empty) that are different from actual
register contents.
The encoding in the tag word for the 32-bit x87 FPUs for unsupported data formats
(including pseudo-zero and unnormal) is special (10B), to comply with IEEE Standard
754. The encoding in the 16-bit IA-32 math coprocessors for pseudo-zero and
unnormal is valid (00B) and the encoding for other unsupported data formats is
special (10B). Code that recognizes the pseudo-zero or unnormal format as valid
must therefore be changed if it is ported to a 32-bit x87 FPU.
19.18.5 Data Types
This section discusses the differences of data types for the various x87 FPUs and
math coprocessors.
19.18.5.1 NaNs
The 32-bit x87 FPUs distinguish between signaling NaNs (SNaNs) and quiet NaNs
(QNaNs). These x87 FPUs only generate QNaNs and normally do not generate an
exception upon encountering a QNaN. An invalid-operation exception (#I) is gener-
ated only upon encountering a SNaN, except for the FCOM, FIST, and FBSTP instruc-
tions, which also generates an invalid-operation exceptions for a QNaNs. This
behavior matches IEEE Standard 754.
The 16-bit IA-32 math coprocessors only generate one kind of NaN (the equivalent of
a QNaN), but the raise an invalid-operation exception upon encountering any kind of
NaN.
When porting software written to run on a 16-bit IA-32 math coprocessor to a 32-bit
x87 FPU, uninitialized memory locations that contain QNaNs should be changed to
SNaNs to cause the x87 FPU or math coprocessor to fault when uninitialized memory
locations are referenced.
19.18.5.2 Pseudo-zero, Pseudo-NaN, Pseudo-infinity, and Unnormal
Formats
The 32-bit x87 FPUs neither generate nor support the pseudo-zero, pseudo-NaN,
pseudo-infinity, and unnormal formats. Whenever they encounter them in an arith-
metic operation, they raise an invalid-operation exception. The 16-bit IA-32 math
coprocessors define and support special handling for these formats. Support for
these formats was dropped to conform with IEEE Standard 754 for Binary Floating-
Point Arithmetic.
This change should not impact software ported from 16-bit IA-32 math coprocessors
to 32-bit x87 FPUs. The 32-bit x87 FPUs do not generate these formats, and there-
fore will not encounter them unless software explicitly loads them in the data regis-
19-12 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
ters. The only affect may be in how software handles the tags in the tag word (see
also: Section 19.18.4, “x87 FPU Tag Word”).
19.18.6 Floating-Point Exceptions
This section identifies the implementation differences in exception handling for
floating-point instructions in the various x87 FPUs and math coprocessors.
19.18.6.1 Denormal Operand Exception (#D)
When the denormal operand exception is masked, the 32-bit x87 FPUs automatically
normalize denormalized numbers when possible; whereas, the 16-bit IA-32 math
coprocessors return a denormal result. A program written to run on a 16-bit IA-32
math coprocessor that uses the denormal exception solely to normalize denormal-
ized operands is redundant when run on the 32-bit x87 FPUs. If such a program is run
on 32-bit x87 FPUs, performance can be improved by masking the denormal excep-
tion. Floating-point programs run faster when the FPU performs normalization of
denormalized operands.
The denormal operand exception is not raised for transcendental instructions and the
FXTRACT instruction on the 16-bit IA-32 math coprocessors. This exception is raised
for these instructions on the 32-bit x87 FPUs. The exception handlers ported to these
latter processors need to be changed only if the handlers gives special treatment to
different opcodes.
19.18.6.2 Numeric Overflow Exception (#O)
On the 32-bit x87 FPUs, when the numeric overflow exception is masked and the
rounding mode is set to chop (toward 0), the result is the largest positive or smallest
negative number. The 16-bit IA-32 math coprocessors do not signal the overflow
exception when the masked response is not ∞; that is, they signal overflow only
when the rounding control is not set to round to 0. If rounding is set to chop (toward
0), the result is positive or negative ∞. Under the most common rounding modes, this
difference has no impact on existing software.
If rounding is toward 0 (chop), a program on a 32-bit x87 FPU produces, under over-
flow conditions, a result that is different in the least significant bit of the significand,
compared to the result on a 16-bit IA-32 math coprocessor. The reason for this differ-
ence is IEEE Standard 754 compatibility.
When the overflow exception is not masked, the precision exception is flagged on the
32-bit x87 FPUs. When the result is stored in the stack, the significand is rounded
according to the precision control (PC) field of the FPU control word or according to
the opcode. On the 16-bit IA-32 math coprocessors, the precision exception is not
flagged and the significand is not rounded. The impact on existing software is that if
the result is stored on the stack, a program running on a 32-bit x87 FPU produces a
different result under overflow conditions than on a 16-bit IA-32 math coprocessor.
Vol. 3 19-13
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
The difference is apparent only to the exception handler. This difference is for IEEE
Standard 754 compatibility.
19.18.6.3 Numeric Underflow Exception (#U)
When the underflow exception is masked on the 32-bit x87 FPUs, the underflow
exception is signaled when both the result is tiny and denormalization results in a
loss of accuracy. When the underflow exception is unmasked and the instruction is
supposed to store the result on the stack, the significand is rounded to the appro-
priate precision (according to the PC flag in the FPU control word, for those instruc-
tions controlled by PC, otherwise to extended precision), after adjusting the
exponent.
When the underflow exception is masked on the 16-bit IA-32 math coprocessors and
rounding is toward 0, the underflow exception flag is raised on a tiny result, regard-
less of loss of accuracy. When the underflow exception is not masked and the desti-
nation is the stack, the significand is not rounded, but instead is left as is.
When the underflow exception is masked, this difference has no impact on existing
software. The underflow exception occurs less often when rounding is toward 0.
When the underflow exception not masked. A program running on a 32-bit x87 FPU
produces a different result during underflow conditions than on a 16-bit IA-32 math
coprocessor if the result is stored on the stack. The difference is only in the least
significant bit of the significand and is apparent only to the exception handler.
19.18.6.4 Exception Precedence
There is no difference in the precedence of the denormal-operand exception on the
32-bit x87 FPUs, whether it be masked or not. When the denormal-operand excep-
tion is not masked on the 16-bit IA-32 math coprocessors, it takes precedence over
all other exceptions. This difference causes no impact on existing software, but some
unneeded normalization of denormalized operands is prevented on the Intel486
processor and Intel 387 math coprocessor.
19.18.6.5 CS and EIP For FPU Exceptions
On the Intel 32-bit x87 FPUs, the values from the CS and EIP registers saved for
floating-point exceptions point to any prefixes that come before the floating-point
instruction. On the 8087 math coprocessor, the saved CS and IP registers points to
the floating-point instruction.
19.18.6.6 FPU Error Signals
The floating-point error signals to the P6 family, Pentium, and Intel486 processors do
not pass through an interrupt controller; an INT# signal from an Intel 387, Intel 287
or 8087 math coprocessors does. If an 8086 processor uses another exception for
19-14 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
the 8087 interrupt, both exception vectors should call the floating-point-error excep-
tion handler. Some instructions in a floating-point-error exception handler may need
to be deleted if they use the interrupt controller. The P6 family, Pentium, and Intel486
processors have signals that, with the addition of external logic, support reporting for
emulation of the interrupt mechanism used in many personal computers.
On the P6 family, Pentium, and Intel486 processors, an undefined floating-point
opcode will cause an invalid-opcode exception (#UD, interrupt vector 6). Undefined
floating-point opcodes, like legal floating-point opcodes, cause a device not available
exception (#NM, interrupt vector 7) when either the TS or EM flag in control register
CR0 is set. The P6 family, Pentium, and Intel486 processors do not check for floating-
point error conditions on encountering an undefined floating-point opcode.
19.18.6.7 Assertion of the FERR# Pin
When using the MS-DOS compatibility mode for handing floating-point exceptions,
the FERR# pin must be connected to an input to an external interrupt controller. An
external interrupt is then generated when the FERR# output drives the input to the
interrupt controller and the interrupt controller in turn drives the INTR pin on the
processor.
For the P6 family and Intel386 processors, an unmasked floating-point exception
always causes the FERR# pin to be asserted upon completion of the instruction that
caused the exception. For the Pentium and Intel486 processors, an unmasked
floating-point exception may cause the FERR# pin to be asserted either at the end of
the instruction causing the exception or immediately before execution of the next
floating-point instruction. (Note that the next floating-point instruction would not be
executed until the pending unmasked exception has been handled.) See Appendix D,
“Guidelines for Writing x87 FPU Extension Handlers,” in the Intel® 64 and IA-32
Architectures Software Developer’s Manual, Volume 1, for a complete description of
the required mechanism for handling floating-point exceptions using the MS-DOS
compatibility mode.
Using FERR# and IGNNE# to handle floating-point exception is deprecated by
modern operating systems; this approach also limits newer processors to operate
with one logical processor active.
19.18.6.8 Invalid Operation Exception On Denormals
An invalid-operation exception is not generated on the 32-bit x87 FPUs upon encoun-
tering a denormal value when executing a FSQRT, FDIV, or FPREM instruction or upon
conversion to BCD or to integer. The operation proceeds by first normalizing the
value. On the 16-bit IA-32 math coprocessors, upon encountering this situation, the
invalid-operation exception is generated. This difference has no impact on existing
software. Software running on the 32-bit x87 FPUs continues to execute in cases
where the 16-bit IA-32 math coprocessors trap. The reason for this change was to
eliminate an exception from being raised.
Vol. 3 19-15
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
19.18.6.9 Alignment Check Exceptions (#AC)
If alignment checking is enabled, a misaligned data operand on the P6 family,
Pentium, and Intel486 processors causes an alignment check exception (#AC) when
a program or procedure is running at privilege-level 3, except for the stack portion of
the FSAVE/FNSAVE, FXSAVE, FRSTOR, and FXRSTOR instructions.
19.18.6.10 Segment Not Present Exception During FLDENV
On the Intel486 processor, when a segment not present exception (#NP) occurs in
the middle of an FLDENV instruction, it can happen that part of the environment is
loaded and part not. In such cases, the FPU control word is left with a value of 007FH.
The P6 family and Pentium processors ensure the internal state is correct at all times
by attempting to read the first and last bytes of the environment before updating the
internal state.
19.18.6.11 Device Not Available Exception (#NM)
The device-not-available exception (#NM, interrupt 7) will occur in the P6 family,
Pentium, and Intel486 processors as described in Section 2.5, “Control Registers,”
Table 2-1, and Chapter 6, “Interrupt 7—Device Not Available Exception (#NM).”
19.18.6.12 Coprocessor Segment Overrun Exception
The coprocessor segment overrun exception (interrupt 9) does not occur in the P6
family, Pentium, and Intel486 processors. In situations where the Intel 387 math
coprocessor would cause an interrupt 9, the P6 family, Pentium, and Intel486 proces-
sors simply abort the instruction. To avoid undetected segment overruns, it is recom-
mended that the floating-point save area be placed in the same page as the TSS. This
placement will prevent the FPU environment from being lost if a page fault occurs
during the execution of an FLDENV, FRSTOR, or FXRSTOR instruction while the oper-
ating system is performing a task switch.
19.18.6.13 General Protection Exception (#GP)
A general-protection exception (#GP, interrupt 13) occurs if the starting address of a
floating-point operand falls outside a segment’s size. An exception handler should be
included to report these programming errors.
19.18.6.14 Floating-Point Error Exception (#MF)
In real mode and protected mode (not including virtual-8086 mode), interrupt vector
16 must point to the floating-point exception handler. In virtual 8086 mode, the
virtual-8086 monitor can be programmed to accommodate a different location of the
interrupt vector for floating-point exceptions.
19-16 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
19.18.7 Changes to Floating-Point Instructions
This section identifies the differences in floating-point instructions for the various
Intel FPU and math coprocessor architectures, the reason for the differences, and
their impact on software.
19.18.7.1 FDIV, FPREM, and FSQRT Instructions
The 32-bit x87 FPUs support operations on denormalized operands and, when
detected, an underflow exception can occur, for compatibility with the IEEE Standard
754. The 16-bit IA-32 math coprocessors do not operate on denormalized operands
or return underflow results. Instead, they generate an invalid-operation exception
when they detect an underflow condition. An existing underflow exception handler
will require change only if it gives different treatment to different opcodes. Also, it is
possible that fewer invalid-operation exceptions will occur.
19.18.7.2 FSCALE Instruction
With the 32-bit x87 FPUs, the range of the scaling operand is not restricted. If (0 < |
ST(1) < 1), the scaling factor is 0; therefore, ST(0) remains unchanged. If the
rounded result is not exact or if there was a loss of accuracy (masked underflow), the
precision exception is signaled. With the 16-bit IA-32 math coprocessors, the range
of the scaling operand is restricted. If (0 < | ST(1) | < 1), the result is undefined and
no exception is signaled. The impact of this difference on exiting software is that
different results are delivered on the 32-bit and 16-bit FPUs and math coprocessors
when (0 < | ST(1) | < 1).
19.18.7.3 FPREM1 Instruction
The 32-bit x87 FPUs compute a partial remainder according to IEEE Standard 754.
This instruction does not exist on the 16-bit IA-32 math coprocessors. The avail-
ability of the FPREM1 instruction has is no impact on existing software.
19.18.7.4 FPREM Instruction
On the 32-bit x87 FPUs, the condition code flags C0, C3, C1 in the status word
correctly reflect the three low-order bits of the quotient following execution of the
FPREM instruction. On the 16-bit IA-32 math coprocessors, the quotient bits are
N
incorrect when performing a reduction of (64 + M) when (N ≥ 1) and M is 1 or 2. This
difference does not affect existing software; software that works around the bug
should not be affected.
19.18.7.5 FUCOM, FUCOMP, and FUCOMPP Instructions
When executing the FUCOM, FUCOMP, and FUCOMPP instructions, the 32-bit x87
FPUs perform unordered compare according to IEEE Standard 754. These instruc-
Vol. 3 19-17
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
tions do not exist on the 16-bit IA-32 math coprocessors. The availability of these
new instructions has no impact on existing software.
19.18.7.6 FPTAN Instruction
On the 32-bit x87 FPUs, the range of the operand for the FPTAN instruction is much
63
less restricted (| ST(0) | < 2 ) than on earlier math coprocessors. The instruction
reduces the operand internally using an internal π/4 constant that is more accurate.
The range of the operand is restricted to (| ST(0) | < π/4) on the 16-bit IA-32 math
coprocessors; the operand must be reduced to this range using FPREM. This change
has no impact on existing software.
19.18.7.7 Stack Overflow
On the 32-bit x87 FPUs, if an FPU stack overflow occurs when the invalid-operation
exception is masked, the FPU returns the real, integer, or BCD-integer indefinite
value to the destination operand, depending on the instruction being executed. On
the 16-bit IA-32 math coprocessors, the original operand remains unchanged
following a stack overflow, but it is loaded into register ST(1). This difference has no
impact on existing software.
19.18.7.8 FSIN, FCOS, and FSINCOS Instructions
On the 32-bit x87 FPUs, these instructions perform three common trigonometric
functions. These instructions do not exist on the 16-bit IA-32 math coprocessors. The
availability of these instructions has no impact on existing software, but using them
provides a performance upgrade.
19.18.7.9 FPATAN Instruction
On the 32-bit x87 FPUs, the range of operands for the FPATAN instruction is unre-
stricted. On the 16-bit IA-32 math coprocessors, the absolute value of the operand in
register ST(0) must be smaller than the absolute value of the operand in register
ST(1). This difference has impact on existing software.
19.18.7.10 F2XM1 Instruction
The 32-bit x87 FPUs support a wider range of operands (–1 < ST (0) < + 1) for the
F2XM1 instruction. The supported operand range for the 16-bit IA-32 math coproces-
sors is (0 ≤ ST(0) ≤ 0.5). This difference has no impact on existing software.
19.18.7.11 FLD Instruction
On the 32-bit x87 FPUs, when using the FLD instruction to load an extended-real
value, a denormal-operand exception is not generated because the instruction is not
19-18 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
arithmetic. The 16-bit IA-32 math coprocessors do report a denormal-operand
exception in this situation. This difference does not affect existing software.
On the 32-bit x87 FPUs, loading a denormal value that is in single- or double-real
format causes the value to be converted to extended-real format. Loading a
denormal value on the 16-bit IA-32 math coprocessors causes the value to be
converted to an unnormal. If the next instruction is FXTRACT or FXAM, the 32-bit x87
FPUs will give a different result than the 16-bit IA-32 math coprocessors. This change
was made for IEEE Standard 754 compatibility.
On the 32-bit x87 FPUs, loading an SNaN that is in single- or double-real format
causes the FPU to generate an invalid-operation exception. The 16-bit IA-32 math
coprocessors do not raise an exception when loading a signaling NaN. The invalid-
operation exception handler for 16-bit math coprocessor software needs to be
updated to handle this condition when porting software to 32-bit FPUs. This change
was made for IEEE Standard 754 compatibility.
19.18.7.12 FXTRACT Instruction
On the 32-bit x87 FPUs, if the operand is 0 for the FXTRACT instruction, the divide-
by-zero exception is reported and –∞ is delivered to register ST(1). If the operand is
+∞, no exception is reported. If the operand is 0 on the 16-bit IA-32 math coproces-
sors, 0 is delivered to register ST(1) and no exception is reported. If the operand is
+∞, the invalid-operation exception is reported. These differences have no impact on
existing software. Software usually bypasses 0 and ∞. This change is due to the IEEE
Standard 754 recommendation to fully support the “logb” function.
19.18.7.13 Load Constant Instructions
On 32-bit x87 FPUs, rounding control is in effect for the load constant instructions.
Rounding control is not in effect for the 16-bit IA-32 math coprocessors. Results for
the FLDPI, FLDLN2, FLDLG2, and FLDL2E instructions are the same as for the 16-bit
IA-32 math coprocessors when rounding control is set to round to nearest or round
to +∞. They are the same for the FLDL2T instruction when rounding control is set to
round to nearest, round to –∞, or round to zero. Results are different from the 16-bit
IA-32 math coprocessors in the least significant bit of the mantissa if rounding
control is set to round to –∞ or round to 0 for the FLDPI, FLDLN2, FLDLG2, and
FLDL2E instructions; they are different for the FLDL2T instruction if round to +∞ is
specified. These changes were implemented for compatibility with IEEE Standard
754 for Floating-Point Arithmetic recommendations.
19.18.7.14 FSETPM Instruction
With the 32-bit x87 FPUs, the FSETPM instruction is treated as NOP (no operation).
This instruction informs the Intel 287 math coprocessor that the processor is in
protected mode. This change has no impact on existing software. The 32-bit x87
Vol. 3 19-19
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
FPUs handle all addressing and exception-pointer information, whether in protected
mode or not.
19.18.7.15 FXAM Instruction
With the 32-bit x87 FPUs, if the FPU encounters an empty register when executing
the FXAM instruction, it not generate combinations of C0 through C3 equal to 1101 or
1111. The 16-bit IA-32 math coprocessors may generate these combinations, among
others. This difference has no impact on existing software; it provides a performance
upgrade to provide repeatable results.
19.18.7.16 FSAVE and FSTENV Instructions
With the 32-bit x87 FPUs, the address of a memory operand pointer stored by FSAVE
or FSTENV is undefined if the previous floating-point instruction did not refer to
memory
19.18.8 Transcendental Instructions
The floating-point results of the P6 family and Pentium processors for transcendental
instructions in the core range may differ from the Intel486 processors by about 2 or
3 ulps (see “Transcendental Instruction Accuracy” in Chapter 8, “Programming with
the x87 FPU,” of the Intel® 64 and IA-32 Architectures Software Developer’s Manual,
Volume 1). Condition code flag C1 of the status word may differ as a result. The exact
threshold for underflow and overflow will vary by a few ulps. The P6 family and
Pentium processors’ results will have a worst case error of less than 1 ulp when
rounding to the nearest-even and less than 1.5 ulps when rounding in other modes.
The transcendental instructions are guaranteed to be monotonic, with respect to the
input operands, throughout the domain supported by the instruction.
Transcendental instructions may generate different results in the round-up flag (C1)
on the 32-bit x87 FPUs. The round-up flag is undefined for these instructions on the
16-bit IA-32 math coprocessors. This difference has no impact on existing software.
19.18.9 Obsolete Instructions
The 8087 math coprocessor instructions FENI and FDISI and the Intel 287 math
coprocessor instruction FSETPM are treated as integer NOP instructions in the 32-bit
x87 FPUs. If these opcodes are detected in the instruction stream, no specific opera-
tion is performed and no internal states are affected.
19-20 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
19.18.10 WAIT/FWAIT Prefix Differences
On the Intel486 processor, when a WAIT/FWAIT instruction precedes a floating-point
instruction (one which itself automatically synchronizes with the previous floating-
point instruction), the WAIT/FWAIT instruction is treated as a no-op. Pending
floating-point exceptions from a previous floating-point instruction are processed not
on the WAIT/FWAIT instruction but on the floating-point instruction following the
WAIT/FWAIT instruction. In such a case, the report of a floating-point exception may
appear one instruction later on the Intel486 processor than on a P6 family or Pentium
FPU, or on Intel 387 math coprocessor.
19.18.11 Operands Split Across Segments and/or Pages
On the P6 family, Pentium, and Intel486 processor FPUs, when the first half of an
operand to be written is inside a page or segment and the second half is outside, a
memory fault can cause the first half to be stored but not the second half. In this situ-
ation, the Intel 387 math coprocessor stores nothing.
19.18.12 FPU Instruction Synchronization
On the 32-bit x87 FPUs, all floating-point instructions are automatically synchro-
nized; that is, the processor automatically waits until the previous floating-point
instruction has completed before completing the next floating-point instruction. No
explicit WAIT/FWAIT instructions are required to assure this synchronization. For the
8087 math coprocessors, explicit waits are required before each floating-point
instruction to ensure synchronization. Although 8087 programs having explicit WAIT
instructions execute perfectly on the 32-bit IA-32 processors without reassembly,
these WAIT instructions are unnecessary.
19.19 SERIALIZING INSTRUCTIONS
Certain instructions have been defined to serialize instruction execution to ensure
that modifications to flags, registers and memory are completed before the next
instruction is executed (or in P6 family processor terminology “committed to machine
state”). Because the P6 family processors use branch-prediction and out-of-order
execution techniques to improve performance, instruction execution is not generally
serialized until the results of an executed instruction are committed to machine state
Architectures Software Developer’s Manual, Volume 1).
As a result, at places in a program or task where it is critical to have execution
completed for all previous instructions before executing the next instruction (for
example, at a branch, at the end of a procedure, or in multiprocessor dependent
code), it is useful to add a serializing instruction. See Section 8.3, “Serializing
Instructions,” for more information on serializing instructions.
Vol. 3 19-21
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
19.20 FPU AND MATH COPROCESSOR INITIALIZATION
Table 9-1 shows the states of the FPUs in the P6 family, Pentium, Intel486 processors
and of the Intel 387 math coprocessor and Intel 287 coprocessor following a power-
up, reset, or INIT, or following the execution of an FINIT/FNINIT instruction. The
following is some additional compatibility information concerning the initialization of
x87 FPUs and math coprocessors.
19.20.1 Intel® 387 and Intel® 287 Math Coprocessor Initialization
Following an Intel386 processor reset, the processor identifies its coprocessor type
®
®
(Intel 287 or Intel 387 DX math coprocessor) by sampling its ERROR# input some
time after the falling edge of RESET# signal and before execution of the first floating-
point instruction. The Intel 287 coprocessor keeps its ERROR# output in inactive
state after hardware reset; the Intel 387 coprocessor keeps its ERROR# output in
active state after hardware reset.
Upon hardware reset or execution of the FINIT/FNINIT instruction, the Intel 387
math coprocessor signals an error condition. The P6 family, Pentium, and Intel486
processors, like the Intel 287 coprocessor, do not.
Initialization
When initializing an Intel486 SX processor and an Intel 487 SX math coprocessor,
the initialization routine should check the presence of the math coprocessor and
should set the FPU related flags (EM, MP, and NE) in control register CR0 accordingly
(see Section 2.5, “Control Registers,” for a complete description of these flags). Table
19-2 gives the recommended settings for these flags when the math coprocessor is
present. The FSTCW instruction will give a value of FFFFH for the Intel486 SX micro-
processor and 037FH for the Intel 487 SX math coprocessor.
Table 19-2. Recommended Values of the EM, MP, and NE Flags for Intel486 SX
Microprocessor/Intel 487 SX Math Coprocessor System
CR0 Flags
Intel486 SX Processor Only
Intel 487 SX Math Coprocessor Present
EM
MP
NE
1
0
1
0
1
0, for MS-DOS* systems
1, for user-defined exception handler
The EM and MP flags in register CR0 are interpreted as shown in Table 19-3.
19-22 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
Table 19-3. EM and MP Flag Interpretation
EM
MP
Interpretation
0
0
Floating-point instructions are passed to FPU; WAIT/FWAIT
and other waiting-type instructions ignore TS.
0
1
1
1
0
1
Floating-point instructions are passed to FPU; WAIT/FWAIT
and other waiting-type instructions test TS.
Floating-point instructions trap to emulator; WAIT/FWAIT and
other waiting-type instructions ignore TS.
Floating-point instructions trap to emulator; WAIT/FWAIT and
other waiting-type instructions test TS.
Following is an example code sequence to initialize the system and check for the
presence of Intel486 SX processor/Intel 487 SX math coprocessor.
fninit
fstcw mem_loc
mov ax, mem_loc
cmp ax, 037fh
jz Intel487_SX_Math_CoProcessor_present
jmp Intel486_SX_microprocessor_present
;ax=037fh
;ax=ffffh
If the Intel 487 SX math coprocessor is not present, the following code can be run to
set the CR0 register for the Intel486 SX processor.
mov eax, cr0
and eax, fffffffdh ;make MP=0
or eax, 0024h
mov cr0, eax
;make EM=1, NE=1
This initialization will cause any floating-point instruction to generate a device not
available exception (#NH), interrupt 7. The software emulation will then take control
to execute these instructions. This code is not required if an Intel 487 SX math
coprocessor is present in the system. In that case, the typical initialization routine for
the Intel486 SX microprocessor will be adequate.
Also, when designing an Intel486 SX processor based system with an Intel 487 SX
math coprocessor, timing loops should be independent of clock speed and clocks per
instruction. One way to attain this is to implement these loops in hardware and not in
software (for example, BIOS).
Vol. 3 19-23
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
19.21 CONTROL REGISTERS
The following sections identify the new control registers and control register flags
and fields that were introduced to the 32-bit IA-32 in various processor families. See
Figure 2-6 for the location of these flags and fields in the control registers.
The Pentium III processor introduced one new control flag in control register CR4:
• OSXMMEXCPT (bit 10) — The OS will set this bit if it supports unmasked SIMD
floating-point exceptions.
The Pentium II processor introduced one new control flag in control register CR4:
state during context switches.
The Pentium Pro processor introduced three new control flags in control register CR4:
• PAE (bit 5) — Physical address extension. Enables paging mechanism to
reference extended physical addresses when set; restricts physical addresses to
32 bits when clear (see also: Section 19.22.1.1, “Physical Memory Addressing
Extension”).
• PGE (bit 7) — Page global enable. Inhibits flushing of frequently-used or shared
pages on CR3 writes (see also: Section 19.22.1.2, “Global Pages”).
• PCE (bit 8) — Performance-monitoring counter enable. Enables execution of the
RDPMC instruction at any protection level.
The content of CR4 is 0H following a hardware reset.
flags that enable certain new extensions provided in the Pentium processor:
• VME — Virtual-8086 mode extensions. Enables support for a virtual interrupt flag
in virtual-8086 mode (see Section 17.3, “Interrupt and Exception Handling in
Virtual-8086 Mode”).
flag in protected mode (see Section 17.4, “Protected-Mode Virtual Interrupts”).
procedures running at privileged level 0.
performance (see Section 19.23.3, “Debug Registers DR4 and DR5”).
• PSE — Page size extensions. Enables 4-MByte pages with 32-bit paging when set
(see Section 4.3, “32-Bit Paging”).
• MCE — Machine-check enable. Enables the machine-check exception, allowing
exception handling for certain hardware error conditions (see Chapter 15,
“Machine-Check Architecture”).
The Intel486 processor introduced five new flags in control register CR0:
19-24 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
• NE — Numeric error. Enables the normal mechanism for reporting floating-point
numeric errors.
• WP — Write protect. Write-protects read-only pages against supervisor-mode
accesses.
• AM — Alignment mask. Controls whether alignment checking is performed.
Operates in conjunction with the AC (Alignment Check) flag.
• NW — Not write-through. Enables write-throughs and cache invalidation cycles
when clear and disables invalidation cycles and write-throughs that hit in the
cache when set.
• CD — Cache disable. Enables the internal cache when clear and disables the
cache when set.
The Intel486 processor introduced two new flags in control register CR3:
• PCD — Page-level cache disable. The state of this flag is driven on the PCD# pin
during bus cycles that are not paged, such as interrupt acknowledge cycles, when
paging is enabled. The PCD# pin is used to control caching in an external cache
on a cycle-by-cycle basis.
• PWT — Page-level write-through. The state of this flag is driven on the PWT# pin
during bus cycles that are not paged, such as interrupt acknowledge cycles, when
paging is enabled. The PWT# pin is used to control write through in an external
cache on a cycle-by-cycle basis.
19.22 MEMORY MANAGEMENT FACILITIES
The following sections describe the new memory management facilities available in
the various IA-32 processors and some compatibility differences.
19.22.1 New Memory Management Control Flags
The Pentium Pro processor introduced three new memory management features:
physical memory addressing extension, the global bit in page-table entries, and
general support for larger page sizes. These features are only available when oper-
ating in protected mode.
19.22.1.1 Physical Memory Addressing Extension
The new PAE (physical address extension) flag in control register CR4, bit 5, may
enable additional address lines on the processor, allowing extended physical
addresses. This option can only be used when paging is enabled, using a new page-
table mechanism provided to support the larger physical address range (see Section
4.1, “Paging Modes and Control Bits”).
Vol. 3 19-25
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
19.22.1.2 Global Pages
The new PGE (page global enable) flag in control register CR4, bit 7, provides a
mechanism for preventing frequently used pages from being flushed from the trans-
lation lookaside buffer (TLB). When this flag is set, frequently used pages (such as
pages containing kernel procedures or common data tables) can be marked global by
setting the global flag in a page-directory or page-table entry.
On a task switch or a write to control register CR3 (which normally causes the TLBs
to be flushed), the entries in the TLB marked global are not flushed. Marking pages
global in this manner prevents unnecessary reloading of the TLB due to TLB misses
on frequently used pages. See Section 4.10, “Caching Translation Information” for a
detailed description of this mechanism.
The P6 family processors support large page sizes. For 32-bit paging, this facility is
enabled with the PSE (page size extension) flag in control register CR4, bit 4. When
this flag is set, the processor supports either 4-KByte or 4-MByte page sizes. PAE
paging and IA-32e paging support 2-MByte pages regardless of the value of CR4.PSE
(see Section 4.4, “PAE Paging” and Section 4.5, “IA-32e Paging”). See Chapter 4,
“Paging,” for more information about large page sizes.
19.22.2 CD and NW Cache Control Flags
The CD and NW flags in control register CR0 were introduced in the Intel486
processor. In the P6 family and Pentium processors, these flags are used to imple-
ment a writeback strategy for the data cache; in the Intel486 processor, they imple-
ment a write-through strategy. See Table 11-5 for a comparison of these bits on the
P6 family, Pentium, and Intel486 processors. For complete information on caching,
see Chapter 11, “Memory Cache Control.”
19.22.3 Descriptor Types and Contents
Operating-system code that manages space in descriptor tables often contains an
invalid value in the access-rights field of descriptor-table entries to identify unused
entries. Access rights values of 80H and 00H remain invalid for the P6 family,
Pentium, Intel486, Intel386, and Intel 286 processors. Other values that were invalid
on the Intel 286 processor may be valid on the 32-bit processors because uses for
these bits have been defined.
19-26 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
19.22.4 Changes in Segment Descriptor Loads
On the Intel386 processor, loading a segment descriptor always causes a locked read
and write to set the accessed bit of the descriptor. On the P6 family, Pentium, and
Intel486 processors, the locked read and write occur only if the bit is not already set.
19.23 DEBUG FACILITIES
The P6 family and Pentium processors include extensions to the Intel486 processor
debugging support for breakpoints. To use the new breakpoint features, it is neces-
sary to set the DE flag in control register CR4.
19.23.1 Differences in Debug Register DR6
It is not possible to write a 1 to reserved bit 12 in debug status register DR6 on the
P6 family and Pentium processors; however, it is possible to write a 1 in this bit on the
Intel486 processor. See Table 9-1 for the different setting of this register following a
power-up or hardware reset.
19.23.2 Differences in Debug Register DR7
The P6 family and Pentium processors determines the type of breakpoint access by
the R/W0 through R/W3 fields in debug control register DR7 as follows:
00
01
10
Break on instruction execution only.
Break on data writes only.
Undefined if the DE flag in control register CR4 is cleared; break on I/O reads
or writes but not instruction fetches if the DE flag in control register CR4 is
set.
11
Break on data reads or writes but not instruction fetches.
On the P6 family and Pentium processors, reserved bits 11, 12, 14 and 15 are hard-
wired to 0. On the Intel486 processor, however, bit 12 can be set. See Table 9-1 for
the different settings of this register following a power-up or hardware reset.
19.23.3 Debug Registers DR4 and DR5
Although the DR4 and DR5 registers are documented as reserved, previous genera-
tions of processors aliased references to these registers to debug registers DR6 and
DR7, respectively. When debug extensions are not enabled (the DE flag in control
register CR4 is cleared), the P6 family and Pentium processors remain compatible
with existing software by allowing these aliased references. When debug extensions
Vol. 3 19-27
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
are enabled (the DE flag is set), attempts to reference registers DR4 or DR5 will
result in an invalid-opcode exception (#UD).
19.24 RECOGNITION OF BREAKPOINTS
For the Pentium processor, it is recommended that debuggers execute the LGDT
instruction before returning to the program being debugged to ensure that break-
points are detected. This operation does not need to be performed on the P6 family,
Intel486, or Intel386 processors.
The implementation of test registers on the Intel486 processor used for testing the
cache and TLB has been redesigned using MSRs on the P6 family and Pentium
processors. (Note that MSRs used for this function are different on the P6 family and
Pentium processors.) The MOV to and from test register instructions generate
invalid-opcode exceptions (#UD) on the P6 family processors.
19.25 EXCEPTIONS AND/OR EXCEPTION CONDITIONS
This section describes the new exceptions and exception conditions added to the 32-
bit IA-32 processors and implementation differences in existing exception handling.
See Chapter 6, “Interrupt and Exception Handling,” for a detailed description of the
IA-32 exceptions.
The PentiumIII processor introduced new state with the XMM registers. Computations
involving data in these registers can produce exceptions. A new MXCSR
control/status register is used to determine which exception or exceptions have
occurred. When an exception associated with the XMM registers occurs, an interrupt
is generated.
• SIMD floating-point exception (#XF, interrupt 19) — New exceptions associated
with the SIMD floating-point registers and resulting computations.
No new exceptions were added with the Pentium Pro and Pentium II processors. The
following exception condition was added to the IA-32 with the Pentium Pro
processor:
• Machine-check exception (#MC, interrupt 18) — New exception conditions. Many
exception conditions have been added to the machine-check exception and a new
architecture has been added for handling and reporting on hardware errors. See
Chapter 15, “Machine-Check Architecture,” for a detailed description of the new
conditions.
The following exceptions and/or exception conditions were added to the IA-32 with
the Pentium processor:
• Machine-check exception (#MC, interrupt 18) — New exception. This exception
reports parity and other hardware errors. It is a model-specific exception and
19-28 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
may not be implemented or implemented differently in future processors. The
MCE flag in control register CR4 enables the machine-check exception. When this
bit is clear (which it is at reset), the processor inhibits generation of the machine-
check exception.
• General-protection exception (#GP, interrupt 13) — New exception condition
added. An attempt to write a 1 to a reserved bit position of a special register
causes a general-protection exception to be generated.
• Page-fault exception (#PF, interrupt 14) — New exception condition added. When
a 1 is detected in any of the reserved bit positions of a page-table entry, page-
directory entry, or page-directory pointer during address translation, a page-fault
exception is generated.
The following exception was added to the Intel486 processor:
• Alignment-check exception (#AC, interrupt 17) — New exception. Reports
unaligned memory references when alignment checking is being performed.
The following exceptions and/or exception conditions were added to the Intel386
processor:
• Divide-error exception (#DE, interrupt 0)
— Change in exception handling. Divide-error exceptions on the Intel386
processors always leave the saved CS:IP value pointing to the instruction that
failed. On the 8086 processor, the CS:IP value points to the next instruction.
— Change in exception handling. The Intel386 processors can generate the
largest negative number as a quotient for the IDIV instruction (80H and
8000H). The 8086 processor generates a divide-error exception instead.
• Invalid-opcode exception (#UD, interrupt 6) — New exception condition added.
Improper use of the LOCK instruction prefix can generate an invalid-opcode
exception.
• Page-fault exception (#PF, interrupt 14) — New exception condition added. If
paging is enabled in a 16-bit program, a page-fault exception can be generated
as follows. Paging can be used in a system with 16-bit tasks if all tasks use the
same page directory. Because there is no place in a 16-bit TSS to store the PDBR
register, switching to a 16-bit task does not change the value of the PDBR
register. Tasks ported from the Intel 286 processor should be given 32-bit TSSs
so they can make full use of paging.
• General-protection exception (#GP, interrupt 13) — New exception condition
added. The Intel386 processor sets a limit of 15 bytes on instruction length. The
only way to violate this limit is by putting redundant prefixes before an
instruction. A general-protection exception is generated if the limit on instruction
length is violated. The 8086 processor has no instruction length limit.
Vol. 3 19-29
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
19.25.1 Machine-Check Architecture
The Pentium Pro processor introduced a new architecture to the IA-32 for handling
and reporting on machine-check exceptions. This machine-check architecture
(described in detail in Chapter 15, “Machine-Check Architecture”) greatly expands
the ability of the processor to report on internal hardware errors.
19.25.2 Priority OF Exceptions
The priority of exceptions are broken down into several major categories:
1. Traps on the previous instruction
2. External interrupts
3. Faults on fetching the next instruction
4. Faults in decoding the next instruction
5. Faults on executing an instruction
There are no changes in the priority of these major categories between the different
processors, however, exceptions within these categories are implementation depen-
dent and may change from processor to processor.
19.26 INTERRUPTS
The following differences in handling interrupts are found among the IA-32
processors.
19.26.1 Interrupt Propagation Delay
External hardware interrupts may be recognized on different instruction boundaries
on the P6 family, Pentium, Intel486, and Intel386 processors, due to the superscaler
designs of the P6 family and Pentium processors. Therefore, the EIP pushed onto the
stack when servicing an interrupt may be different for the P6 family, Pentium,
Intel486, and Intel386 processors.
19.26.2 NMI Interrupts
After an NMI interrupt is recognized by the P6 family, Pentium, Intel486, Intel386,
and Intel 286 processors, the NMI interrupt is masked until the first IRET instruction
is executed, unlike the 8086 processor.
19-30 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
19.26.3 IDT Limit
The LIDT instruction can be used to set a limit on the size of the IDT. A double-fault
exception (#DF) is generated if an interrupt or exception attempts to read a vector
beyond the limit. Shutdown then occurs on the 32-bit IA-32 processors if the double-
fault handler vector is beyond the limit. (The 8086 processor does not have a shut-
down mode nor a limit.)
19.27 ADVANCED PROGRAMMABLE INTERRUPT
CONTROLLER (APIC)
The Advanced Programmable Interrupt Controller (APIC), referred to in this book as
the local APIC, was introduced into the IA-32 processors with the Pentium
processor (beginning with the 735/90 and 815/100 models) and is included in the
Pentium 4, Intel Xeon, and P6 family processors. The features and functions of the
local APIC are derived from the Intel 82489DX external APIC, which was used with
the Intel486 and early Pentium processors. Additional refinements of the local APIC
architecture were incorporated in the Pentium 4 and Intel Xeon processors.
19.27.1 Software Visible Differences Between the Local APIC and
the 82489DX
The following features in the local APIC features differ from those found in the
82489DX external APIC:
• When the local APIC is disabled by clearing the APIC software enable/disable flag
in the spurious-interrupt vector MSR, the state of its internal registers are
unaffected, except that the mask bits in the LVT are all set to block local
interrupts to the processor. Also, the local APIC ceases accepting IPIs except for
INIT, SMI, NMI, and start-up IPIs. In the 82489DX, when the local unit is
disabled, all the internal registers including the IRR, ISR and TMR are cleared and
the mask bits in the LVT are set. In this state, the 82489DX local unit will accept
only the reset deassert message.
• In the local APIC, NMI and INIT (except for INIT deassert) are always treated as
edge triggered interrupts, even if programmed otherwise. In the 82489DX, these
interrupts are always level triggered.
• In the local APIC, IPIs generated through the ICR are always treated as edge
triggered (except INIT Deassert). In the 82489DX, the ICR can be used to
generate either edge or level triggered IPIs.
• In the local APIC, the logical destination register supports 8 bits; in the 82489DX,
it supports 32 bits.
• In the local APIC, the APIC ID register is 4 bits wide; in the 82489DX, it is 8 bits
wide.
Vol. 3 19-31
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
• The remote read delivery mode provided in the 82489DX and local APIC for
Pentium processors is not supported in the local APIC in the Pentium 4, Intel
Xeon, and P6 family processors.
• For the 82489DX, in the lowest priority delivery mode, all the target local APICs
specified by the destination field participate in the lowest priority arbitration. For
the local APIC, only those local APICs which have free interrupt slots will
participate in the lowest priority arbitration.
19.27.2 New Features Incorporated in the Local APIC for the P6
Family and Pentium Processors
The local APIC in the Pentium and P6 family processors have the following new
features not found in the 82489DX external APIC.
• Cluster addressing is supported in logical destination mode.
• Focus processor checking can be enabled/disabled.
• Interrupt input signal polarity can be programmed for the LINT0 and LINT1 pins.
• An SMI IPI is supported through the ICR and I/O redirection table.
• An error status register is incorporated into the LVT to log and report APIC errors.
In the P6 family processors, the local APIC incorporates an additional LVT register to
handle performance monitoring counter interrupts.
19.27.3 New Features Incorporated in the Local APIC of the Pentium
4 and Intel Xeon Processors
The local APIC in the Pentium 4 and Intel Xeon processors has the following new
features not found in the P6 family and Pentium processors and in the 82489DX.
• The local APIC ID is extended to 8 bits.
• An thermal sensor register is incorporated into the LVT to handle thermal sensor
interrupts.
• The the ability to deliver lowest-priority interrupts to a focus processor is no
longer supported.
• The flat cluster logical destination mode is not supported.
19.28 TASK SWITCHING AND TSS
This section identifies the implementation differences of task switching, additions to
the TSS and the handling of TSSs and TSS segment selectors.
19-32 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
19.28.1 P6 Family and Pentium Processor TSS
When the virtual mode extensions are enabled (by setting the VME flag in control
register CR4), the TSS in the P6 family and Pentium processors contain an interrupt
redirection bit map, which is used in virtual-8086 mode to redirect interrupts back to
an 8086 program.
19.28.2 TSS Selector Writes
During task state saves, the Intel486 processor writes 2-byte segment selectors into
a 32-bit TSS, leaving the upper 16 bits undefined. For performance reasons, the P6
family and Pentium processors write 4-byte segment selectors into the TSS, with the
upper 2 bytes being 0. For compatibility reasons, code should not depend on the
value of the upper 16 bits of the selector in the TSS.
19.28.3 Order of Reads/Writes to the TSS
The order of reads and writes into the TSS is processor dependent. The P6 family and
Pentium processors may generate different page-fault addresses in control register
CR2 in the same TSS area than the Intel486 and Intel386 processors, if a TSS
crosses a page boundary (which is not recommended).
19.28.4 Using A 16-Bit TSS with 32-Bit Constructs
Task switches using 16-bit TSSs should be used only for pure 16-bit code. Any new
code written using 32-bit constructs (operands, addressing, or the upper word of the
EFLAGS register) should use only 32-bit TSSs. This is due to the fact that the 32-bit
processors do not save the upper 16 bits of EFLAGS to a 16-bit TSS. A task switch
back to a 16-bit task that was executing in virtual mode will never re-enable the
virtual mode, as this flag was not saved in the upper half of the EFLAGS value in the
TSS. Therefore, it is strongly recommended that any code using 32-bit constructs
use a 32-bit TSS to ensure correct behavior in a multitasking environment.
19.28.5 Differences in I/O Map Base Addresses
The Intel486 processor considers the TSS segment to be a 16-bit segment and wraps
around the 64K boundary. Any I/O accesses check for permission to access this I/O
address at the I/O base address plus the I/O offset. If the I/O map base address
exceeds the specified limit of 0DFFFH, an I/O access will wrap around and obtain the
permission for the I/O address at an incorrect location within the TSS. A TSS limit
violation does not occur in this situation on the Intel486 processor. However, the P6
family and Pentium processors consider the TSS to be a 32-bit segment and a limit
violation occurs when the I/O base address plus the I/O offset is greater than the TSS
limit. By following the recommended specification for the I/O base address to be less
Vol. 3 19-33
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
than 0DFFFH, the Intel486 processor will not wrap around and access incorrect loca-
tions within the TSS for I/O port validation and the P6 family and Pentium processors
will not experience general-protection exceptions (#GP). Figure 19-1 demonstrates
the different areas accessed by the Intel486 and the P6 family and Pentium
processors.
Intel486 Processor
P6 family and Pentium Processors
FFFFH + 10H = Outside Segment
for I/O Validation
FFFFH
I/O Map
FFFFH
I/O Map
Base Addres
FFFFH
FFFFH
Base Addres
FFFFH + 10H = FH
for I/O Validation
0H
0H
I/O access at port 10H checks
bitmap at I/O map base address
FFFFH + 10H = offset 10H.
Offset FH from beginning of
TSS segment results because
wraparound occurs.
I/O access at port 10H checks
bitmap at I/O address FFFFH + 10H,
which exceeds segment limit.
Wrap around does not occur,
general-protection exception (#GP)
occurs.
Figure 19-1. I/O Map Base Address Differences
19.29 CACHE MANAGEMENT
The P6 family processors include two levels of internal caches: L1 (level 1) and L2
(level 2). The L1 cache is divided into an instruction cache and a data cache; the L2
cache is a general-purpose cache. See Section 11.1, “Internal Caches, TLBs, and
Buffers,” for a description of these caches. (Note that although the Pentium II
processor L2 cache is physically located on a separate chip in the cassette, it is
considered an internal cache.)
The Pentium processor includes separate level 1 instruction and data caches. The
data cache supports a writeback (or alternatively write-through, on a line by line
basis) policy for memory updates.
The Intel486 processor includes a single level 1 cache for both instructions and data.
The meaning of the CD and NW flags in control register CR0 have been redefined for
the P6 family and Pentium processors. For these processors, the recommended value
(00B) enables writeback for the data cache of the Pentium processor and for the L1
19-34 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
data cache and L2 cache of the P6 family processors. In the Intel486 processor,
setting these flags to (00B) enables write-through for the cache.
External system hardware can force the Pentium processor to disable caching or to
use the write-through cache policy should that be required. In the P6 family proces-
sors, the MTRRs can be used to override the CD and NW flags (see Table 11-6).
The P6 family and Pentium processors support page-level cache management in the
same manner as the Intel486 processor by using the PCD and PWT flags in control
register CR3, the page-directory entries, and the page-table entries. The Intel486
processor, however, is not affected by the state of the PWT flag since the internal
cache of the Intel486 processor is a write-through cache.
19.29.1 Self-Modifying Code with Cache Enabled
On the Intel486 processor, a write to an instruction in the cache will modify it in both
the cache and memory. If the instruction was prefetched before the write, however,
the old version of the instruction could be the one executed. To prevent this problem,
it is necessary to flush the instruction prefetch unit of the Intel486 processor by
coding a jump instruction immediately after any write that modifies an instruction.
The P6 family and Pentium processors, however, check whether a write may modify
an instruction that has been prefetched for execution. This check is based on the
linear address of the instruction. If the linear address of an instruction is found to be
present in the prefetch queue, the P6 family and Pentium processors flush the
prefetch queue, eliminating the need to code a jump instruction after any writes that
modify an instruction.
Because the linear address of the write is checked against the linear address of the
instructions that have been prefetched, special care must be taken for self-modifying
code to work correctly when the physical addresses of the instruction and the written
data are the same, but the linear addresses differ. In such cases, it is necessary to
execute a serializing operation to flush the prefetch queue after the write and before
executing the modified instruction. See Section 8.3, “Serializing Instructions,” for
more information on serializing instructions.
NOTE
The check on linear addresses described above is not in practice a
concern for compatibility. Applications that include self-modifying
code use the same linear address for modifying and fetching the
instruction. System software, such as a debugger, that might
possibly modify an instruction using a different linear address than
that used to fetch the instruction must execute a serializing
operation, such as IRET, before the modified instruction is executed.
Vol. 3 19-35
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
19.29.2 Disabling the L3 Cache
A unified third-level (L3) cache in processors based on Intel NetBurst microarchitec-
ture (see Section 11.1, “Internal Caches, TLBs, and Buffers”) provides the third-level
cache disable flag, bit 6 of the IA32_MISC_ENABLE MSR. The third-level cache
disable flag allows the L3 cache to be disabled and enabled, independently of the L1
and L2 caches (see Section 11.5.4, “Disabling and Enabling the L3 Cache”). The
third-level cache disable flag applies only to processors based on Intel NetBurst
microarchitecture. Processors with L3 and based on other microarchitectures do not
support the third-level cache disable flag.
19.30 PAGING
This section identifies enhancements made to the paging mechanism and implemen-
tation differences in the paging mechanism for various IA-32 processors.
19.30.1 Large Pages
The Pentium processor extended the memory management/paging facilities of the
IA-32 to allow large (4 MBytes) pages sizes (see Section 4.3, “32-Bit Paging”). The
first P6 family processor (the Pentium Pro processor) added a 2 MByte page size to
the IA-32 in conjunction with the physical address extension (PAE) feature (see
Section 4.4, “PAE Paging”).
The availability of large pages with 32-bit paging on any IA-32 processor can be
determined via feature bit 3 (PSE) of register EDX after the CPUID instruction has
been execution with an argument of 1. (Large pages are always available with PAE
paging and IA-32e paging.) Intel processors that do not support the CPUID instruc-
tion support only 32-bit paging and do not support page size enhancements. (See
“CPUID—CPU Identification” in Chapter 3, “Instruction Set Reference, A-M,” in the
Intel® 64 and IA-32 Architectures Software Developer’s Manual, Volume 2A, and AP-
485, Intel Processor Identification and the CPUID Instruction, for more information
on the CPUID instruction.)
19.30.2 PCD and PWT Flags
The PCD and PWT flags were introduced to the IA-32 in the Intel486 processor to
control the caching of pages:
• PCD (page-level cache disable) flag—Controls caching on a page-by-page basis.
• PWT (page-level write-through) flag—Controls the write-through/writeback
caching policy on a page-by-page basis. Since the internal cache of the Intel486
processor is a write-through cache, it is not affected by the state of the PWT flag.
19-36 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
19.30.3 Enabling and Disabling Paging
Paging is enabled and disabled by loading a value into control register CR0 that modi-
fies the PG flag. For backward and forward compatibility with all IA-32 processors,
Intel recommends that the following operations be performed when enabling or
disabling paging:
1. Execute a MOV CR0, REG instruction to either set (enable paging) or clear
(disable paging) the PG flag.
2. Execute a near JMP instruction.
The sequence bounded by the MOV and JMP instructions should be identity mapped
(that is, the instructions should reside on a page whose linear and physical addresses
are identical).
For the P6 family processors, the MOV CR0, REG instruction is serializing, so the
jump operation is not required. However, for backwards compatibility, the JMP
instruction should still be included.
19.31 STACK OPERATIONS
This section identifies the differences in the stack mechanism for the various IA-32
processors.
19.31.1 Selector Pushes and Pops
When pushing a segment selector onto the stack, the Pentium 4, Intel Xeon, P6
family, and Intel486 processors decrement the ESP register by the operand size and
then write 2 bytes. If the operand size is 32-bits, the upper two bytes of the write are
not modified. The Pentium processor decrements the ESP register by the operand
size and determines the size of the write by the operand size. If the operand size is
32-bits, the upper two bytes are written as 0s.
When popping a segment selector from the stack, the Pentium 4, Intel Xeon, P6
family, and Intel486 processors read 2 bytes and increment the ESP register by the
operand size of the instruction. The Pentium processor determines the size of the
read from the operand size and increments the ESP register by the operand size.
It is possible to align a 32-bit selector push or pop such that the operation generates
an exception on a Pentium processor and not on an Pentium 4, Intel Xeon, P6 family,
or Intel486 processor. This could occur if the third and/or fourth byte of the operation
lies beyond the limit of the segment or if the third and/or fourth byte of the operation
is locate on a non-present or inaccessible page.
For a POP-to-memory instruction that meets the following conditions:
• The stack segment size is 16-bit.
• Any 32-bit addressing form with the SIB byte specifying ESP as the base register.
Vol. 3 19-37
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
• The initial stack pointer is FFFCH (32-bit operand) or FFFEH (16-bit operand) and
will wrap around to 0H as a result of the POP operation.
The result of the memory write is implementation-specific. For example, in P6 family
processors, the result of the memory write is SS:0H plus any scaled index and
displacement. In Pentium processors, the result of the memory write may be either a
stack fault (real mode or protected mode with stack segment size of 64 KByte), or
write to SS:10000H plus any scaled index and displacement (protected mode and
stack segment size exceeds 64 KByte).
19.31.2 Error Code Pushes
The Intel486 processor implements the error code pushed on the stack as a 16-bit
value. When pushed onto a 32-bit stack, the Intel486 processor only pushes 2 bytes
and updates ESP by 4. The P6 family and Pentium processors’ error code is a full 32
bits with the upper 16 bits set to zero. The P6 family and Pentium processors, there-
fore, push 4 bytes and update ESP by 4. Any code that relies on the state of the upper
16 bits may produce inconsistent results.
19.31.3 Fault Handling Effects on the Stack
During the handling of certain instructions, such as CALL and PUSHA, faults may
occur in different sequences for the different processors. For example, during far
calls, the Intel486 processor pushes the old CS and EIP before a possible branch fault
is resolved. A branch fault is a fault from a branch instruction occurring from a
segment limit or access rights violation. If a branch fault is taken, the Intel486 and
P6 family processors will have corrupted memory below the stack pointer. However,
the ESP register is backed up to make the instruction restartable. The P6 family
processors issue the branch before the pushes. Therefore, if a branch fault does
occur, these processors do not corrupt memory below the stack pointer. This imple-
mentation difference, however, does not constitute a compatibility problem, as only
values at or above the stack pointer are considered to be valid. Other operations that
encounter faults may also corrupt memory below the stack pointer and this behavior
may vary on different implementations.
19.31.4 Interlevel RET/IRET From a 16-Bit Interrupt or Call Gate
If a call or interrupt is made from a 32-bit stack environment through a 16-bit gate,
only 16 bits of the old ESP can be pushed onto the stack. On the subsequent
RET/IRET, the 16-bit ESP is popped but the full 32-bit ESP is updated since control is
being resumed in a 32-bit stack environment. The Intel486 processor writes the SS
selector into the upper 16 bits of ESP. The P6 family and Pentium processors write
zeros into the upper 16 bits.
19-38 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
19.32 MIXING 16- AND 32-BIT SEGMENTS
The features of the 16-bit Intel 286 processor are an object-code compatible subset
of those of the 32-bit IA-32 processors. The D (default operation size) flag in
segment descriptors indicates whether the processor treats a code or data segment
as a 16-bit or 32-bit segment; the B (default stack size) flag in segment descriptors
indicates whether the processor treats a stack segment as a 16-bit or 32-bit
segment.
The segment descriptors used by the Intel 286 processor are supported by the 32-bit
IA-32 processors if the Intel-reserved word (highest word) of the descriptor is clear.
On the 32-bit IA-32 processors, this word includes the upper bits of the base address
and the segment limit.
The segment descriptors for data segments, code segments, local descriptor tables
(there are no descriptors for global descriptor tables), and task gates are the same
for the 16- and 32-bit processors. Other 16-bit descriptors (TSS segment, call gate,
interrupt gate, and trap gate) are supported by the 32-bit processors.
The 32-bit processors also have descriptors for TSS segments, call gates, interrupt
gates, and trap gates that support the 32-bit architecture. Both kinds of descriptors
can be used in the same system.
For those segment descriptors common to both 16- and 32-bit processors, clear bits
in the reserved word cause the 32-bit processors to interpret these descriptors
exactly as an Intel 286 processor does, that is:
• Base Address — The upper 8 bits of the 32-bit base address are clear, which limits
base addresses to 24 bits.
• Limit — The upper 4 bits of the limit field are clear, restricting the value of the
limit field to 64 KBytes.
• Granularity bit — The G (granularity) flag is clear, indicating the value of the
16-bit limit is interpreted in units of 1 byte.
• Big bit — In a data-segment descriptor, the B flag is clear in the segment
descriptor used by the 32-bit processors, indicating the segment is no larger than
64 KBytes.
• Default bit — In a code-segment descriptor, the D flag is clear, indicating 16-bit
addressing and operands are the default. In a stack-segment descriptor, the D
flag is clear, indicating use of the SP register (instead of the ESP register) and a
64-KByte maximum segment limit.
For information on mixing 16- and 32-bit code in applications, see Chapter 18,
“Mixing 16-Bit and 32-Bit Code.”
19.33 SEGMENT AND ADDRESS WRAPAROUND
This section discusses differences in segment and address wraparound between the
P6 family, Pentium, Intel486, Intel386, Intel 286, and 8086 processors.
Vol. 3 19-39
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
19.33.1 Segment Wraparound
On the 8086 processor, an attempt to access a memory operand that crosses offset
65,535 or 0FFFFH or offset 0 (for example, moving a word to offset 65,535 or
pushing a word when the stack pointer is set to 1) causes the offset to wrap around
modulo 65,536 or 010000H. With the Intel 286 processor, any base and offset combi-
nation that addresses beyond 16 MBytes wraps around to the 1 MByte of the address
space. The P6 family, Pentium, Intel486, and Intel386 processors in real-address
mode generate an exception in these cases:
• A general-protection exception (#GP) if the segment is a data segment (that is,
if the CS, DS, ES, FS, or GS register is being used to address the segment).
• A stack-fault exception (#SS) if the segment is a stack segment (that is, if the SS
register is being used).
An exception to this behavior occurs when a stack access is data aligned, and the
stack pointer is pointing to the last aligned piece of data that size at the top of the
stack (ESP is FFFFFFFCH). When this data is popped, no segment limit violation
occurs and the stack pointer will wrap around to 0.
The address space of the P6 family, Pentium, and Intel486 processors may wrap-
around at 1 MByte in real-address mode. An external A20M# pin forces wraparound
if enabled. On Intel 8086 processors, it is possible to specify addresses greater than
1 MByte. For example, with a selector value FFFFH and an offset of FFFFH, the effec-
tive address would be 10FFEFH (1 MByte plus 65519 bytes). The 8086 processor,
which can form addresses up to 20 bits long, truncates the uppermost bit, which
“wraps” this address to FFEFH. However, the P6 family, Pentium, and Intel486
processors do not truncate this bit if A20M# is not enabled.
If a stack operation wraps around the address limit, shutdown occurs. (The 8086
processor does not have a shutdown mode or a limit.)
The behavior when executing near the limit of a 4-GByte selector (limit=0xFFFFFFFF)
is different between the Pentium Pro and the Pentium 4 family of processors. On the
Pentium Pro, instructions which cross the limit -- for example, a two byte instruction
such as INC EAX that is encoded as 0xFF 0xC0 starting exactly at the limit faults for
a segment violation (a one byte instruction at 0xFFFFFFFF does not cause an excep-
tion). Using the Pentium 4 microprocessor family, neither of these situations causes
a fault.
Segment wraparound and the functionality of A20M# is used primarily by older oper-
ating systems and not used by modern operating systems. On newer Intel 64 proces-
sors, A20M# may be absent.
19.34 STORE BUFFERS AND MEMORY ORDERING
The Pentium 4, Intel Xeon, and P6 family processors provide a store buffer for
temporary storage of writes (stores) to memory (see Section 11.10, “Store Buffer”).
Writes stored in the store buffer(s) are always written to memory in program order,
19-40 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
with the exception of “fast string” store operations (see Section 8.2.4, “Out-of-Order
Stores For String Operations”).
The Pentium processor has two store buffers, one corresponding to each of the pipe-
lines. Writes in these buffers are always written to memory in the order they were
generated by the processor core.
It should be noted that only memory writes are buffered and I/O writes are not. The
Pentium 4, Intel Xeon, P6 family, Pentium, and Intel486 processors do not synchro-
nize the completion of memory writes on the bus and instruction execution after a
write. An I/O, locked, or serializing instruction needs to be executed to synchronize
The Pentium 4, Intel Xeon, and P6 family processors use processor ordering to main-
tain consistency in the order that data is read (loaded) and written (stored) in a
program and the order the processor actually carries out the reads and writes. With
this type of ordering, reads can be carried out speculatively and in any order, reads
can pass buffered writes, and writes to memory are always carried out in program
order. (See Section 8.2, “Memory Ordering,” for more information about processor
producer and a consumer of data. The SFENCE instruction provides a performance-
efficient way of ensuring ordering between routines that produce weakly-ordered
results and routines that consume this data.
No re-ordering of reads occurs on the Pentium processor, except under the condition
®
®
™
noted in Section 8.2.1, “Memory Ordering in the Intel Pentium and Intel486
Processors,” and in the following paragraph describing the Intel486 processor.
Specifically, the store buffers are flushed before the IN instruction is executed. No
reads (as a result of cache miss) are reordered around previously generated writes
sitting in the store buffers. The implication of this is that the store buffers will be
flushed or emptied before a subsequent bus cycle is run on the external bus.
On both the Intel486 and Pentium processors, under certain conditions, a memory
read will go onto the external bus before the pending memory writes in the buffer
even though the writes occurred earlier in the program execution. A memory read
will only be reordered in front of all writes pending in the buffers if all writes pending
in the buffers are cache hits and the read is a cache miss. Under these conditions, the
Intel486 and Pentium processors will not read from an external memory location that
needs to be updated by one of the pending writes.
During a locked bus cycle, the Intel486 processor will always access external
memory, it will never look for the location in the on-chip cache. All data pending in
the Intel486 processor's store buffers will be written to memory before a locked cycle
is allowed to proceed to the external bus. Thus, the locked bus cycle can be used for
eliminating the possibility of reordering read cycles on the Intel486 processor. The
Pentium processor does check its cache on a read-modify-write access and, if the
cache line has been modified, writes the contents back to memory before locking the
bus. The P6 family processors write to their cache on a read-modify-write operation
(if the access does not split across a cache line) and does not write back to system
Vol. 3 19-41
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
memory. If the access does split across a cache line, it locks the bus and accesses
system memory.
I/O reads are never reordered in front of buffered memory writes on an IA-32
processor. This ensures an update of all memory locations before reading the status
from an I/O device.
19.35 BUS LOCKING
The Intel 286 processor performs the bus locking differently than the Intel P6 family,
Pentium, Intel486, and Intel386 processors. Programs that use forms of memory
locking specific to the Intel 286 processor may not run properly when run on later
processors.
A locked instruction is guaranteed to lock only the area of memory defined by the
destination operand, but may lock a larger memory area. For example, typical 8086
and Intel 286 configurations lock the entire physical memory space. Programmers
should not depend on this.
On the Intel 286 processor, the LOCK prefix is sensitive to IOPL. If the CPL is greater
than the IOPL, a general-protection exception (#GP) is generated. On the Intel386
DX, Intel486, and Pentium, and P6 family processors, no check against IOPL is
performed.
The Pentium processor automatically asserts the LOCK# signal when acknowledging
external interrupts. After signaling an interrupt request, an external interrupt
controller may use the data bus to send the interrupt vector to the processor. After
receiving the interrupt request signal, the processor asserts LOCK# to insure that no
other data appears on the data bus until the interrupt vector is received. This bus
locking does not occur on the P6 family processors.
19.36 BUS HOLD
Unlike the 8086 and Intel 286 processors, but like the Intel386 and Intel486 proces-
sors, the P6 family and Pentium processors respond to requests for control of the bus
from other potential bus masters, such as DMA controllers, between transfers of
parts of an unaligned operand, such as two words which form a doubleword. Unlike
the Intel386 processor, the P6 family, Pentium and Intel486 processors respond to
bus hold during reset initialization.
19.37 MODEL-SPECIFIC EXTENSIONS TO THE IA-32
Certain extensions to the IA-32 are specific to a processor or family of IA-32 proces-
sors and may not be implemented or implemented in the same way in future proces-
19-42 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
sors. The following sections describe these model-specific extensions. The CPUID
instruction indicates the availability of some of the model-specific features.
19.37.1 Model-Specific Registers
The Pentium processor introduced a set of model-specific registers (MSRs) for use in
controlling hardware functions and performance monitoring. To access these MSRs,
two new instructions were added to the IA-32 architecture: read MSR (RDMSR) and
write MSR (WRMSR). The MSRs in the Pentium processor are not guaranteed to be
duplicated or provided in the next generation IA-32 processors.
The P6 family processors greatly increased the number of MSRs available to soft-
ware. See Appendix B, “Model-Specific Registers (MSRs),” for a complete list of the
available MSRs. The new registers control the debug extensions, the performance
counters, the machine-check exception capability, the machine-check architecture,
and the MTRRs. These registers are accessible using the RDMSR and WRMSR instruc-
tions. Specific information on some of these new MSRs is provided in the following
sections. As with the Pentium processor MSR, the P6 family processor MSRs are not
guaranteed to be duplicated or provided in the next generation IA-32 processors.
19.37.2 RDMSR and WRMSR Instructions
The RDMSR (read model-specific register) and WRMSR (write model-specific
register) instructions recognize a much larger number of model-specific registers in
the P6 family processors. (See “RDMSR—Read from Model Specific Register” and
“WRMSR—Write to Model Specific Register” in the Intel® 64 and IA-32 Architectures
Software Developer’s Manual, Volumes 2A & 2B for more information.)
19.37.3 Memory Type Range Registers
Memory type range registers (MTRRs) are a new feature introduced into the IA-32 in
the Pentium Pro processor. MTRRs allow the processor to optimize memory opera-
tions for different types of memory, such as RAM, ROM, frame buffer memory, and
memory-mapped I/O.
MTRRs are MSRs that contain an internal map of how physical address ranges are
mapped to various types of memory. The processor uses this internal memory map
to determine the cacheability of various physical memory locations and the optimal
method of accessing memory locations. For example, if a memory location is speci-
fied in an MTRR as write-through memory, the processor handles accesses to this
location as follows. It reads data from that location in lines and caches the read data
or maps all writes to that location to the bus and updates the cache to maintain cache
coherency. In mapping the physical address space with MTRRs, the processor recog-
nizes five types of memory: uncacheable (UC), uncacheable, speculatable, write-
combining (WC), write-through (WT), write-protected (WP), and writeback (WB).
Vol. 3 19-43
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
Earlier IA-32 processors (such as the Intel486 and Pentium processors) used the
KEN# (cache enable) pin and external logic to maintain an external memory map and
signal cacheable accesses to the processor. The MTRR mechanism simplifies hard-
ware designs by eliminating the KEN# pin and the external logic required to drive it.
See Chapter 9, “Processor Management and Initialization,” and Appendix B, “Model-
Specific Registers (MSRs),” for more information on the MTRRs.
19.37.4 Machine-Check Exception and Architecture
The Pentium processor introduced a new exception called the machine-check excep-
tion (#MC, interrupt 18). This exception is used to detect hardware-related errors,
such as a parity error on a read cycle.
The P6 family processors extend the types of errors that can be detected and that
generate a machine-check exception. It also provides a new machine-check architec-
ture for recording information about a machine-check error and provides extended
recovery capability.
The machine-check architecture provides several banks of reporting registers for
recording machine-check errors. Each bank of registers is associated with a specific
and interconnect operations; however, checks are also made of translation lookaside
buffer (TLB) and cache operations.
The machine-check architecture can correct some errors automatically and allow for
reliable restart of instruction execution. It also collects sufficient information for soft-
ware to use in correcting other machine errors not corrected by hardware.
See Chapter 15, “Machine-Check Architecture,” for more information on the
machine-check exception and the machine-check architecture.
19.37.5 Performance-Monitoring Counters
The P6 family and Pentium processors provide two performance-monitoring counters
for use in monitoring internal hardware operations. The number of performance
monitoring counters and associated programming interfaces may be implementation
specific for Pentium 4 processors, Pentium M processors. Later processors may have
implemented these as part of an architectural performance monitoring feature. The
architectural and non-architectural performance monitoring interfaces for different
processor families are described in Chapter 30, “Performance Monitoring,”. Appendix
A, “Performance-Monitoring Events,” lists all the events that can be counted for
architectural performance monitoring events and non-architectural events. The
counters are set up, started, and stopped using two MSRs and the RDMSR and
WRMSR instructions. For the P6 family processors, the current count for a particular
counter can be read using the new RDPMC instruction.
19-44 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
The performance-monitoring counters are useful for debugging programs, optimizing
code, diagnosing system failures, or refining hardware designs. See Chapter 30,
“Performance Monitoring,” for more information on these counters.
19.38 TWO WAYS TO RUN INTEL 286 PROCESSOR TASKS
When porting 16-bit programs to run on 32-bit IA-32 processors, there are two
approaches to consider:
• Porting an entire 16-bit software system to a 32-bit processor, complete with the
old operating system, loader, and system builder. Here, all tasks will have 16-bit
TSSs. The 32-bit processor is being used as if it were a faster version of the 16-bit
processor.
a 32-bit operating system, loader, and system builder. Here, the TSSs used to
represent 286 tasks should be changed to 32-bit TSSs. It is possible to mix 16
and 32-bit TSSs, but the benefits are small and the problems are great. All tasks
in a 32-bit software system should have 32-bit TSSs. It is not necessary to
change the 16-bit object modules themselves; TSSs are usually constructed by
the operating system, by the loader, or by the system builder. See Chapter 18,
“Mixing 16-Bit and 32-Bit Code,” for more detailed information about mixing
16-bit and 32-bit code.
Because the 32-bit processors use the contents of the reserved word of 16-bit
segment descriptors, 16-bit programs that place values in this word may not run
correctly on the 32-bit processors.
Vol. 3 19-45
Download from Www.Somanuals.com. All Manuals Search And Download.
ARCHITECTURE COMPATIBILITY
19-46 Vol. 3
Download from Www.Somanuals.com. All Manuals Search And Download.
|