Manuale d’uso / di manutenzione del prodotto IA-32 del fabbricante Intel
Vai alla pagina of 636
IA-32 In tel ® Ar chitectur e So ftw ar e De v eloper’ s Manual Vo l u m e 3 A : S ystem Pr ogr amming Guide, P art 1 NO TE: The IA-32 Intel Ar chitecture Softwar e Dev eloper's Manual c onsis.
INFORMATION IN THIS DOCUMENT IS PROVIDED IN CONNECTION WITH INTEL PRO DUCTS. NO LICENSE, EX- PRESS OR IMPLIED, BY ESTOPPEL OR OTHERWISE, TO ANY INTELLECTUAL PROPERTY RIGHT S IS GRANTED BY THIS DOCUMENT.
Vol. 3A iii CONTENT S FOR V OLUME 3A AND 3B CHAPTER 1 ABOUT THIS MANUAL 1.1 IA-32 PROCESSORS COVERED IN THIS MANUA L . . . . . . . . . . . . . . . . . . . . . . . 1-1 1.2 OVERVIEW OF THE SYSTEM PROG RAMMING GUIDE . . . . . . . . . . . . . . . . . . . .
CONTENTS iv Vol. 3A PAGE 2.6.7 Readi ng and Writing Model-Specific Registers . . . . . . . . . . . . . . . . . . . . . . . . . . 2 -29 2.6.7.1 R eading and Writing Model-Specific Registers in 64-Bit Mode . . . . . . . . . . . 2-29 CHAPTER 3 PROTECTED-MODE MEMORY MANAGEMENT 3.
Vol. 3A v CONTENTS PAGE CHAPTER 4 PROTECTION 4.1 ENABLING AND DISABLING SEGMENT AND PAGE PROTECTION . . . . . . . . . . 4-1 4.2 FIELDS AND FLAGS USED FOR SEGMENT-LEVEL AND PAGE-LEVEL PROTECTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
CONTENTS vi Vol. 3A PAGE CHAPTER 5 INTERRUPT AND EXCEPTION HANDLING 5.1 INTERRUPT AND EXCEPTION OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1 5.2 EXCEPTION AND INTERRUPT VECTORS . . . . . . . . . . . . . . . . . . . . . . . . .
Vol. 3A vii CONTENTS PAGE Interrupt 16—x87 FPU Floa ting-Po int Error (#MF) . . . . . . . . . . . . . . . . . . . . . . 5-55 Interrupt 17—Al ignment Check Exception (#AC). . . . . . . . . . . . . . . . . . . . . . . . 5-57 Interrupt 18—Machine-Check Exce ption (#MC) .
CONTENTS viii Vol. 3A PAGE 7.5.4 MP Initialization Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-18 7.5.4.1 Typica l BSP Initialization Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Vol. 3A ix CONTENTS PAGE 7.11.6.3 Halt Idle Logical Processors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-52 7.11.6.4 Potential Usa ge of MONITOR/MWAIT in C1 Idle Loops . . . . . . . . . . . . . . . . 7-52 7.11.6.5 Guideline s for Scheduling Threads on Logical Processors Sharing Execution Resources .
CONTENTS x Vol. 3A PAGE 8.10 APIC BUS MESSAGE PASSING MECHANISM AND PROTOCOL (P6 FAMILY, PENTIUM PROCESSORS) . . . . . . . . . . . . . . . . . . . . . 8-42 8.10.1 Bus Message Fo rmats. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Vol. 3A xi CONTENTS PAGE 9.11.6.4 Update in a System Suppo rting Dual-Cor e Technol ogy . . . . . . . . . . . . . . . . 9-46 9.11.6.5 Update Load er Enhance ments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-46 9.11.7 Update Signature and Verification .
CONTENTS xii Vol. 3A PAGE 10.11.3.1 Base and Mask Calculations with Intel EM64T. . . . . . . . . . . . . . . . . . . . . . . 10-33 10.11.4 Range Size and Alignment Requirement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-34 10.11.4.1 MTRR Precedences .
Vol. 3A xiii CONTENTS PAGE CHAPTER 13 POWER AND THERMAL MANAGEMENT 13.1 ENHANCED INTEL SPEEDSTEP ® TECHNOLOGY . . . . . . . . . . . . . . . . . . . . . . . 13-1 13.1.1 Software Interface For Init iating Performance State Transitions . . . . . . . . .
CONTENTS xiv Vol. 3A PAGE 15.2 VIRTUAL-8086 MODE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-7 15.2.1 Enabling Virtual-8086 Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Vol. 3A xv CONTENTS PAGE 17.6. STREAMING SIMD EXTENSIONS (SSE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-3 17.7. STREAMING SIMD EXTENSIONS 2 (SSE2). . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-3 17.8. STREAMING SIMD EXTENSIONS 3 (SSE3).
CONTENTS xvi Vol. 3A PAGE 17.17.7.12. FXTRACT Instruction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-17 17.17.7.13. Load Constant Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Vol. 3A xvii CONTENTS PAGE 17.29.1. Large Pages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-34 17.29.2. PCD and PWT Flags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
CONTENTS xviii Vol. 3A PAGE 18.5.7.1 Last Exception Records and Intel EM64T . . . . . . . . . . . . . . . . . . . . . . . . . . 18-19 18.5.8 Branch Trace Store (BTS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-19 18.
Vol. 3A xix CONTENTS PAGE 18.11 PERFORMANCE MONITORI NG AND HYPER-THREADING TECHNOLOGY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-60 18.11.1 ESCR MSRs . . . . . . . . . . . . . . . . . . . . .
CONTENTS xx Vol. 3A PAGE 20.7 VM-EXIT CONTROL FIELDS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-14 20.7.1 VM-Exit Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Vol. 3A xxi CONTENTS PAGE 22.3.2.1 Loadin g Guest Control Registers, Deb ug Reg isters, and MSRs . . . . . . . . 21-14 22.3.2.2 Loadin g Guest Segment Registers and Descriptor-Table Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
CONTENTS xxii Vol. 3A PAGE 24.3.2 Exiting From SMM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-4 24.4 SMRAM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Vol. 3A xxiii CONTENTS PAGE CHAPTER 25 VIRTUAL-MACHINE MONITOR PR OGRAMMING CONSIDERATIONS 25.1 VMX SYSTEM PROGRAMMING OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-1 25.2 SUPPORTING PROCESSOR OPERATING MODES IN GUEST ENVIRONMENTS .
CONTENTS xxiv Vol. 3A PAGE 26.3.5.1 Initialization of Virtual TLB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-6 26.3.5.2 Response to Pa ge Faults . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Vol. 3A xxv CONTENTS PAGE APPENDIX C MP INITIALIZATION FO R P6 FAMILY PROCESSORS C.1 OVERVIEW OF THE MP INITIALI ZATION PROCESS FOR P6 FAMILY PROCESSORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
CONTENTS xxvi Vol. 3A PAGE H.3.4 32-Bit Host-State Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H-6 H.4 NATURAL-WI DTH FIELDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Vol. 3A xxvii CONTENTS PAGE Figure 3-23. Format of Page-Direct ory Entries for 4-MByte Pages and 36-Bit Physical Addresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-38 Figure 3-24. IA-32e Mode Paging Structures (4-KByte Pages) .
CONTENTS xxviii Vol. 3A PAGE Figure 7-6. Topological Relationships betwee n Hierarchical IDs in a Hypothetical MP Platfor m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-36 Figure 8-1. Relationship of Local APIC and I/O APIC In Singl e-Processor Systems .
Vol. 3A xxix CONTENTS PAGE Figure 11-2. Mapping of MMX Registers to x87 FPU Data Register Stack . . . . . . . . . . . . 11-7 Figure 12-1. Example of Saving the x87 FPU, MMX, SSE, and SSE2 State During an Operating-System Controlled Task Switch . . . .
CONTENTS xxx Vol. 3A PAGE Figure 18-23. MSR_IFSB_CTL6, Address: 10 7D2H ; MSR_IFSB_CNTR7, Address: 107D3H . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-70 Figure 18-24. PerfEvtSel0 and PerfEvtSel1 MSRs . . . . . . . . . . . . . . . . . . .
Vol. 3A xxxi CONTENTS PAGE Table 6-1. Exception Conditions Checked During a Ta sk Switch . . . . . . . . . . . . . . . . . 6-15 Table 6-2. Effect of a Task Switch on Busy Flag, NT Flag, Previous Task Link Field, and TS Flag . . . . . . . . . . . . . .
CONTENTS xxxii Vol. 3A PAGE Table 11-3. Effect of the MMX, x87 FPU, and FXSAVE/FXRS TOR Instructions on the x87 FPU Tag Wo rd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-4 Table 12-1. Action Taken for Combination s of OSFXSR, OSXMMEXCPT, SSE, SSE2, SSE3, EM, MP, and TS1 .
Vol. 3A xxxiii CONTENTS PAGE Table 23-1. Exit Qualification for Debug Exceptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-5 Table 23-2. Exit Qualification for Task Switch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
CONTENTS xxxiv Vol. 3A PAGE Table F-3. Non-Focused Lowest Priority Message (34 Cycles) . . . . . . . . . . . . . . . . . . . . .F-3 Table F-4. APIC Bus Status Cycles Interpretation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .F-5 Table G-1.
1 About This Manual.
.
Vol. 3A 1-1 CHAPTER 1 ABOUT THIS MANUAL The IA-32 Intel® Ar chitectur e Softwar e Developer ’ s Manual, V olume 3A: S ystem Pr ogramming Guide, Part 1 (order num ber 253668) and the IA-32 Intel® A.
1-2 Vol. 3A ABOUT THIS MANUAL 1.2 OVERVIEW OF THE SYST EM PROGRAMMING GUIDE A description of this manual’ s content follows: Chapter 1 — About This Manual. Gives an overview o f all three volumes of t he IA-32 Intel Ar chitectur e Softwar e Developer ’ s Manual .
Vol. 3A 1-3 ABOUT THIS MANUAL level, including: task swi tching, exception handling, and compatibility with existing system environments. Chapter 12 — SSE, SSE2 and SSE3 System Programming.
1-4 Vol. 3A ABOUT THIS MANUAL Chapter 25 — V irtual-Mach ine Monitoring Programming Considerations. Describes programming considerations for VMMs. VMMs manage virtual machines (VMs). Chapter 26 — V irt ualization of System Resources. Describes the virtualization of the system resources.
Vol. 3A 1-5 ABOUT THIS MANUAL 1.3.1 Bit and Byte Order In illustrations of d ata structures in memory , smaller addresses appear toward the botto m of the figure; addresses increase toward the top. Bit po sitions are numbered from right to left. The numerical value of a set bit is equal to two raised to the power of the bit posit ion.
1-6 Vol. 3A ABOUT THIS MANUAL 1.3.3 Instruction Operands When instructions are represen ted symbolically , a subset of the IA-32 assem bly language is used. In this subset, an instruction has the following form at: label: mnemonic argument1, argument2, argument3 where: • A label is an identifier which is followed by a colon.
Vol. 3A 1-7 ABOUT THIS MANUAL 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 di git is a character from the following set: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, A, B, C, D, E, and F .
1-8 Vol. 3A ABOUT THIS MANUAL 1.3.7 Exceptions An exception is an event that typically occurs when an instruct ion causes an erro r . For example, an attempt to divide by zero generates an excep tion. However, some exceptions, such as break- points, occur under other conditions.
Vol. 3A 1-9 ABOUT THIS MANUAL 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 IA-32 processors is listed on-line at this link: http://developer .
1-10 Vol. 3A ABOUT THIS MANUAL.
2 System Ar chitectur e Overview.
.
Vol. 3A 2-1 CHAPTER 2 SYSTEM ARCHITECTURE OVERVIEW IA-32 architecture (beginning with the Intel386 processor family) provides extens ive support for operating-system and system-dev elop ment software. This supp ort offers multiple modes of operation, which include: • Real mode, protected m ode, virtual 8086 m ode, and system m anagement mod e.
2-2 Vol. 3A SYSTEM ARCHITECTURE OVERVIEW 2.1 OVERVIEW OF THE SY STEM-LEVEL ARCHITECTURE IA-32 system-level archit ecture consists of a se t of registers, data st ructures, and instructions designed to.
Vol. 3A 2-3 SYSTEM ARCHITECTURE OVERVIEW Figure 2-1. IA-32 System-Level Registers and Data Structures Local Descriptor T able (LDT) EFLAGS Register Control Registers CR1 CR2 CR3 CR4 CR0 Global Descriptor T able (GDT) Interrupt Descriptor T able (IDT) IDTR GDTR Interrupt Gate T rap Gate LDT Desc.
2-4 Vol. 3A SYSTEM ARCHITECTURE OVERVIEW Figure 2-2. System-Level Registers an d Data Structures in IA-32e Mode Local Descriptor T able (LDT) CR1 CR2 CR3 CR4 CR0 Global Descriptor T able (GDT) Interrupt Descriptor T able (IDT) IDTR GDTR Interrupt Gate T rap Gate LDT Desc.
Vol. 3A 2-5 SYSTEM ARCHITECTURE OVERVIEW 2.1.1 Global and Local Descriptor T ables When operating in pr otected mode, all memory accesses pass through either the global descriptor table (GDT) or an optional local desc riptor table (LDT) as shown in Figure 2-1.
2-6 Vol. 3A SYSTEM ARCHITECTURE OVERVIEW For example, a CALL to a call gate can provide access to a proce dure in a code segment that is at the same or a numerically lowe r privilege leve l (more priv ileged) than the current code segment. T o access a procedure through a call gate, the calling procedure 1 supplies the selector for the call gate.
Vol. 3A 2-7 SYSTEM ARCHITECTURE OVERVIEW A task can also be accessed through a task gate. A task gate is similar to a call g ate, except that it provides access (through a segment selector) to a TSS rather than a code segment. 2.1.3.1 T as k-St ate Segmen ts in IA-32e Mode Hardware task switches are not supported in IA -32e mode.
2-8 Vol. 3A SYSTEM ARCHITECTURE OVERVIEW The location of pages (sometimes called page frames) in physical memory is contained in t wo types of system data structures: page directories and page tables. Both structures reside i n phys- ical memory (see Figure 2-1).
Vol. 3A 2-9 SYSTEM ARCHITECTURE OVERVIEW • The GDTR, LDTR, and IDTR registers contain the linear addresses and sizes (limits) of their respective tables. See also: Section 2.4, “Memory-Mana gement Registers.” • The task register contains th e linear address and size of th e TSS for the current task.
2-10 Vol. 3A SYSTEM ARCHITECTURE OVERVIEW 2.1.7 Other System Resources Besides the system registers and data structures described in the previous secti ons, system archi- tecture provides the fo llowing additional resources: • Operating system instruction s (see also: Section 2.
Vol. 3A 2-11 SYSTEM ARCHITECTURE OVERVIEW The processor is placed in real-address mode following power-up or a reset. The PE flag in control register CR0 then contro ls whether the processor is oper ating in real-address or protected mode. See also: Section 9.
2-12 Vol. 3A SYSTEM ARCHITECTURE OVERVIEW 2.3 SYSTEM FLAGS AN D FIELDS IN THE EFLAGS REGISTER The system flags and IOPL field of the EFLAGS re gister control I/O, ma skable hardware inter- rupts, debugging, task switchi ng, and the virt ual-8086 mode (see Figure 2-4).
Vol. 3A 2-13 SYSTEM ARCHITECTURE OVERVIEW The IOPL is also one of the mechanisms th at controls the modification of the IF flag and the handling of int errupt s in virtual -80 86 m ode when vi rtual m ode extensions are in effect (when CR4.
2-14 Vol. 3A SYSTEM ARCHITECTURE OVERVIEW VIF V irtual Interrupt (bit 19) — Contains a virtual image of the IF flag. This flag is used in conjunction with the VIP flag. The pro cessor only recognizes the VIF flag when either the VME flag or the PVI flag in cont rol register CR4 is set and the IOPL is less than 3.
Vol. 3A 2-15 SYSTEM ARCHITECTURE OVERVIEW 2.4.1 Global Descriptor T able 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 G DT .
2-16 Vol. 3A SYSTEM ARCHITECTURE OVERVIEW 2.4.3 IDTR Interrupt Descriptor T ab le Register The IDTR register holds the base address (32 bits in protected mode; 64 bits in IA-32 e mode) and 16-bit table limit for the IDT .
Vol. 3A 2-17 SYSTEM ARCHITECTURE OVERVIEW The control registers are summar ized below , and each architectur ally defined control field in these control registers are described indi vidually . In Figure 2-6, the width of the regist er in 64-bit mode is indicated in parenthesis (except for CR0).
2-18 Vol. 3A SYSTEM ARCHITECTURE OVERVIEW When loading a control register , reserved bits shou ld always be set to th e values previously read. The flags in control registers are: PG Paging (bit 31 of CR0) — Enables paging when set; disab les paging when clear .
Vol. 3A 2-19 SYSTEM ARCHITECTURE OVERVIEW NW Not Write-th rough (bit 29 of CR0) — When the NW and CD flags are clear , write- back (for Pentium 4, Inte l Xeon, P6 fami ly , and Pentium processors) or write-through (for Intel486 processors) is enabled for writ es that hit the cache and invalidat ion cycles are enabled.
2-20 Vol. 3A SYSTEM ARCHITECTURE OVERVIEW • If the TS flag is set and the MP flag (b it 1 of CR0) and EM flag are clear, an #NM exception is not raised prior to the ex ecution of an x87 FPU W AIT/FW AIT instruction. • If the EM flag is set, the sett ing of th e TS flag has no affect on the execution of x87 FPU/MMX/SSE/SSE2/SSE3 instructions.
Vol. 3A 2-21 SYSTEM ARCHITECTURE OVERVIEW FPU or math coprocessor present in the syst em. T able 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 T able 1 1- 1).
2-22 Vol. 3A SYSTEM ARCHITECTURE OVERVIEW VME V irtual-8086 Mode Extensions (bit 0 of CR4) — Enables interrupt- and exception- handling extensions in virtual-8 086 mode when set; disables the extensions when clear .
Vol. 3A 2-23 SYSTEM ARCHITECTURE OVERVIEW When enabling the global page feat ure, 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 processo r performance will be impacted.
2-24 Vol. 3A SYSTEM ARCHITECTURE OVERVIEW 2.5.1 CPUID Qualification of Control Regi ster Flags The VME, PVI, TSD, DE, PSE, P AE, MCE, PGE, PCE, OSFXSR, and OSXMMEXCP T flags in control register CR4 are mode l specific.
Vol. 3A 2-25 SYSTEM ARCHITECTURE OVERVIEW 2.6.1 Loading and S toring System Registers The GDTR, LDTR, IDTR, and TR registers each ha ve a load and store instruction for loading data into and storing data from the register: • LGDT (Load GDTR Register) — Loads t he GDT base address and limit from memory into the GDTR register .
2-26 Vol. 3A SYSTEM ARCHITECTURE OVERVIEW • SLDT (S tore LDT Register) — Stores the LDT segment se lector from the LDTR register into memory or a general-purpose register . • L T R (Load T ask Register) — Loads seg ment selector and segment descriptor for a TSS from memory into the task register .
Vol. 3A 2-27 SYSTEM ARCHITECTURE OVERVIEW Offset Is W ithin Limits (LSL Instruction),” fo r a detailed explanation of the function and use of this instruction. The VERR (verify for reading) and VER W (verify for writing) instructions verify if a selected segment is readable or writable, respectively , at a given CPL.
2-28 Vol. 3A SYSTEM ARCHITECTURE OVERVIEW Hardware may respond to this signal in a numbe r of ways. An indicato r light on the front panel may be turned on. An NM I interrupt for recording diagnost ic information may be generated . Reset initialization m ay be invoked (note that the BINIT# pin wa s introduced with th e Pentium Pro processor).
Vol. 3A 2-29 SYSTEM ARCHITECTURE OVERVIEW See Section 18.10, “Per formance Monitoring Overview ,” an d Section 18.9, “Time-S tamp Counter ,” for more information about th e perform ance mon itoring and time-stamp cou nters. The RDTSC instruction was introduced into the IA-32 architecture with the Pentium processor .
2-30 Vol. 3A SYSTEM ARCHITECTURE OVERVIEW.
3 Pr otected-Mode Memory Management.
.
Vol. 3A 3-1 CHAPTER 3 PROTECTED-MODE MEMORY MANAGEMENT This chapter describes the IA-32 a rchitecture’ s protected-mode mem ory management facilities, including the physical mem ory requirements, segmentation mechanism , and paging mechanism.
3-2 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT If paging is not used, the linear address space of the proces sor is mapped di rectly into the phys- ical address space of processor . The physical addr ess space is defined as the range of addresses that the processor can generate on its address bus.
Vol. 3A 3-3 PROTECTED-MODE MEMORY MANAGEMENT If the page being accessed is not currently in physical memory , the processor interrupts execu- tion of the program (by generati ng a page-fault exception). The operating system or executive then reads the page into physical memory from the disk and conti nues execut ing the program.
3-4 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT More complexity can be added to this protected flat model to provide more protecti on. For example, for the paging mechanis m to provide isolation bet ween.
Vol. 3A 3-5 PROTECTED-MODE MEMORY MANAGEMENT 3.2.3 Multi-Segment Model A multi-segment model (su ch as the one shown in Figure 3-4) uses the full cap abilities of the segmentation mechanism to provid ed hardware enforced protection of code, data structures, and programs and tasks.
3-6 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT 3.2.4 Segment ation in IA-32e Mode In IA-32e mode, the effects of segmentation depend on whether the processor is running in compatibility mo de or 64-bit mode. In compatibility m ode, segmentation functions just as it does using legacy 16-bit or 32-bit prot ected mode semantics.
Vol. 3A 3-7 PROTECTED-MODE MEMORY MANAGEMENT 3.3.1 Physical Address S p ace for Processors with Intel ® EM64T On processors that su pport Intel EM64T (CPUID.8000000 1.EDX[29] = 1), the size of p hysical address range is impl ementation-specific and indicated by CPU ID.
3-8 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT If paging is not used, the processor maps the lin ear address directly to a ph ysical address (that is, the linear address goes out on the processor ’ s address bus).
Vol. 3A 3-9 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 . Requested Privilege Level (RPL) (Bits 0 and 1) — Specifies the priv ilege le vel of the selector .
3-10 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT can be available for immediate use. Other segm ents can be made available by loading their segment selectors into these re gisters during program execu tion. Every segment register has a “visible” part and a “hidden” part.
Vol. 3A 3-11 PROTECTED-MODE MEMORY MANAGEMENT 3.4.4 Segment Loading Inst ructions in IA-32e Mode Because ES, DS, and SS segment registers are not us ed in 64-bit mode, thei r fields (b ase, limit, and attribute) in segment descri ptor registers are ignored.
3-12 Vol. 3A 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 acce ss control and status information.
Vol. 3A 3-13 PROTECTED-MODE MEMORY MANAGEMENT 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. Of fsets less than the segment limit generate general -protection ex ceptions.
3-14 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT segment. (This flag should always be set to 1 for 32-bit code and d ata segments and to 0 for 16-bit cod e and d a ta seg men ts .
Vol. 3A 3-15 PROTECTED-MODE MEMORY MANAGEMENT L (64-bit code segment) fla g In IA-32e mode, bit 2 1 of the second doublewo rd of the segmen t descriptor indicates whether a code segment contains native 64-bit code. A value of 1 indicates instruction s in this code segment are executed in 64-bit mode.
3-16 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT 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).
Vol. 3A 3-17 PROTECTED-MODE MEMORY MANAGEMENT 3.5 SYSTEM DESCRIPTOR T YPES 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 descrip tors: • Local descriptor-table (LDT) segment descriptor .
3-18 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT See also: Section 3. 5.1, “Segment Descriptor T ables”, and Section 6.2.2, “TSS Descriptor” (for more information on the sy stem-segment descript ors); see Section 4.8.3, “Call Gates”, Section 5.
Vol. 3A 3-19 PROTECTED-MODE MEMORY MANAGEMENT Each system must have one GDT defined, which may be used fo r all programs and tasks in the system. Optionally , one or more LDT s 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 .
3-20 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT 3.5.2 Segment Descriptor T ables in IA-32e Mode In IA-32e mode, a segment d escriptor table can contain up to 8192 (2 13 ) 8-byte descriptors. An entry in the segment descriptor table can be 8 by tes. System descriptors are expanded to 16 bytes (occupying the space of two entries).
Vol. 3A 3-21 PROTECTED-MODE MEMORY MANAGEMENT accessed for a long time. See S ection 3.12, “T ranslation Lookasi de Buffers (TLBs)”, for more information on the TLBs. 3.6.1 Paging Options Paging is controlled by three fl ags in the processor ’ s control registers: • PG (paging) flag.
3-22 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT 3.6.2 Page T ables and Directorie s in the Absence of Intel EM64T The information that the processo r uses to translate linear ad dresses into physical ad.
Vol. 3A 3-23 PROTECTED-MODE MEMORY MANAGEMENT 3.7.1 Linear Address T ransl ation (4-KByte Pages) Figure 3-12 show s the page dir ectory and p age-t able hierarchy wh en mappi ng lin ear ad dresses to 4-KByte pages. The entries in the page director y point to page tables, and the entries in a page table point to pages in physi cal memory .
3-24 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT T o select the vari ous tabl e entries, the lin ear address is divided into three sections: • Page-directory entry — Bits 22 throug h 31 provide an offset to an entry in the page directory . Th e selected entry provides the base physical address of a page table.
Vol. 3A 3-25 PROTECTED-MODE MEMORY MANAGEMENT NOTE (For the Pentium processor onl y .) When enab ling or disabling large page sizes, the TLBs must be invalidated (flu shed) after the PSE flag in control register CR4 has been set or cleared.
3-26 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT 3.7.6 Page-Directory a nd Page-T able Entries Figure 3-14 shows the format for the page-directory and page-table ent ries when 4-KByte pages and 32-bit physical addresses are being used.
Vol. 3A 3-27 PROTECTED-MODE MEMORY MANAGEMENT (Page-directory entr ies for 4-KByte page tables) — Specifies the physical address of the first byte of a page table. The bits in this field are int erpreted as the 20 most-significant bits of the phy sical address, which forces page tables to be aligned on 4-KByte boundaries.
3-28 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT 3. Invalidate the current page-table entry i n the TLB (see Section 3.12, “T ranslation Lookaside Buffers (TLB s)”, for a discussion of TLBs and how to invalidate th em). 4. Return fro m the page -fault handler to restart the interrupted program (or task).
Vol. 3A 3-29 PROTECTED-MODE MEMORY MANAGEMENT This flag is a “sticky” flag, meani ng that once set, the processor does not implicitly clear it. Only software can clear this flag. The accessed and dirty flags are provided for use by memory m anagement software t o manage the transfer of pages and page tables into and out of physical memory .
3-30 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT in the TLB when register CR3 is loaded or a task switch occurs. This flag is provided to prevent frequently used pages (such as pages that contain kernel or other operating system or executive cod e ) from being flushed from the TLB .
Vol. 3A 3-31 PROTECTED-MODE MEMORY MANAGEMENT When the P AE paging mechanism is enabled, the processor supports two sizes of pages: 4-KByte and 2-MByte.
3-32 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT T o select the vari ous tabl e entries, the lin ear address is divided into three sections: • Page-directory-pointer-table entry—Bits 30 and 31 provide an offset to one of the 4 entries in the page-directory-pointer ta ble.
Vol. 3A 3-33 PROTECTED-MODE MEMORY MANAGEMENT CR4 has no affect on the page size when P AE is en abled.) W ith the PS flag set, the linear address is divided into three sections: • Page-directory-pointer-table entry—Bits 30 an d 31 provide an offset to an entry in the page-directory-pointer table.
3-34 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT 3.8.5 Page-Directory and Page-T a ble Entries With Extended Addressing Enabled Figure 3-20 shows the format for the page-d irectory-pointer -table, page-directory , and page-table entries when 4 -KByte pages and 36 -bit extended ph ysical addresses are being used.
Vol. 3A 3-35 PROTECTED-MODE MEMORY MANAGEMENT Figure 3-20. Format of Page-Directo ry-Po inter-T able, Page-Directory , and Page-T able Entries for 4-KByte Pa ges with P AE Enabled 63 36 35 32 Base Reserved (set to 0) Page-Directory-Pointe r-T able Entry 31 12 11 9 8 543 2 0 P C D P W T Ava il Page-Directory Base Address Addr .
3-36 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT The base physical address in an entry specifies the following, depending on the type of entry: • Page-directory-pointer -table entry — the physical address of the first byte of a 4-KByte page directory .
Vol. 3A 3-37 PROTECTED-MODE MEMORY MANAGEMENT Access (A) and dirty (D) flags (bits 5 and 6) are provided for table en tries that point to pages. Bits 9, 10, and 11 in all the table entries for the physical address extension are available for use by software.
3-38 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT Fi g ur e 3- 23 s ho ws th e fo rm at fo r t h e p ag e- directory entries when 4-MByte pages and 36-bit physical addresses are being us ed. Section 3.7.6, “Page-Dir ectory and Page-T able Entries” describes the functions of the flags and fields in bits 0 through 1 1.
Vol. 3A 3-39 PROTECTED-MODE MEMORY MANAGEMENT 3.10 P AE-ENABLED PAGI NG IN IA-32E MODE Intel EM64T 64-bit extensions expand physical add ress extension (P AE) paging structures to potentially support mapping a 64-b it linear address to a 52-bit physical address.
3-40 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT 3.10.2 IA-32e Mode Linear Address T ranslatio n (2-MByte Pages) Figure 3-25 shows the PML4 tab le, page-direct ory-pointer , and page-di rectory hi erarchy w hen mapping linear addresses to 2-MByte page s in IA-32e mode.
Vol. 3A 3-41 PROTECTED-MODE MEMORY MANAGEMENT • Page-director y entry — Bits 29:21 provide an offset to an entry in the page directory . The selected entry provides the base physical address of a 2-MByte page. • Page offset — Bits 20:0 provides an offset to a physical address in the page.
3-42 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT Except for bit 63, functions of the flags in these entries are as described in Section 3.7.6, “Page- Directory and Page-T able En tries”. The dif ferences are: • A PML4 table entry and a page-direct ory-pointer-table entry are added.
Vol. 3A 3-43 PROTECTED-MODE MEMORY MANAGEMENT • The base physical address fiel d in each entry is extended to 28 bits if the processor ’ s implementation su pports a 40-bit physical address. • Bits 62:52 are available for use by system programmers.
3-44 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT If the execute disable bit is enabled in an IA-32 processor , the reserved bits in paging data struc- tures for legacy 32-bit mode an d 64-bit mode are shown in T abl e 3-5.
Vol. 3A 3-45 PROTECTED-MODE MEMORY MANAGEMENT 3.1 1 MAPPING SEGMENT S TO PAGES The segmentation and paging mechanism s prov ide in the IA-32 architecture support a wide variety of approaches to memory management. When segmen tation and paging is combined, segments can be mapped to pages i n several ways.
3-46 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT 3.12 T RANSLATION LOOKASIDE BUFFERS (TLBS) The processor stores the most recently used pa ge-directory and page-tab le entries in on-chip caches called translation lookaside buffers or TLB s. The P6 family and Pentium processors have separate TLBs for the data and instruction caches.
Vol. 3A 3-47 PROTECTED-MODE MEMORY MANAGEMENT • Implicitly by executing a task switch, which automat ically changes the contents of the CR3 register .
3-48 Vol. 3A PROTECTED-MODE MEMORY MANAGEMENT.
4 Pr otection.
.
Vol. 3A 4-1 CHAPTER 4 PROTECTION In protected mode, the IA-32 architecture provid es a protection mechanism that operates at both the segment level and the page level.
4-2 Vol. 3A PROTECTION that is based on privilege levels can essentially be disabled while still in protected mode by ass ig nin g a p rivilege lev el of 0 (most privil eged) to all segment selecto rs and segment descrip- tors.
Vol. 3A 4-3 PROTECTION • Read/write (R/W) f lag — (Bit 1 of a page-d irectory or page-table entry .) Determines the type of access allowed to a page: read only or re ad-write.
4-4 Vol. 3A PROTECTION Many different styles of prot ection schemes can be implemente d with these fields and flags. When the operating system creat es a descriptor , it places values in these fields and flags in keeping with the particular prot ection style chosen for an operat ing system or executive.
Vol. 3A 4-5 PROTECTION 4.3 LIMIT CHECKING The limit field of a segment descriptor preven ts programs or procedures from addressing memory locations outside the segm ent. The effective value of the limit depends on the sett ing of the G (granularity) flag (see Figure 4-1).
4-6 Vol. 3A PROTECTION 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 .
Vol. 3A 4-7 PROTECTION • When a segment selector is l oaded into a segment register — Certain segment registers can contain only certain desc riptor types, for example: — The CS register only can be loaded with a selector for a code segment.
4-8 Vol. 3A PROTECTION — On a call or jump through a call gate (or on an interrupt- or exception-handler call through a trap or interru pt gate), the pro cessor automatically checks that the segment descriptor being pointed to by the gate is for a code segment.
Vol. 3A 4-9 PROTECTION The processor uses privilege leve ls 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 viol ation, it generates a general-protection excep- tion (#GP).
4-10 Vol. 3A PROTECTION — 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 pro grams running at a CPL of 0 can access the segment.
Vol. 3A 4-11 PROTECTION 4.6 PRIVILEGE LEVEL CHECKI NG WHEN ACCESSING DATA SEGMENT S T o access operands in a data segment, the segment selector for the data segment must be loaded into the data-segment registers (DS, ES, FS, o r GS) or into the stack-s egment register (SS).
4-12 Vol. 3A PROTECTION 4. The procedure in code segm ent 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.
Vol. 3A 4-13 PROTECTION 4.6.1 Accessing Dat a in Code Segment s In some instances it may be desirable to access data structures that are contained in a code segment. The following meth ods of accessing data in code segments are possible: • Load a data-segment register with a segmen t selector for a nonconf orming, readable, code segment.
4-14 Vol. 3A PROTECTION A JMP or CALL instruction can reference another code segment in any of four ways: • The target operand contains the segment selector for the tar get code segment. • The target operand points to a call-gate descri ptor , whi ch contains th e segment selector for the target code segment.
Vol. 3A 4-15 PROTECTION • The DPL of the segment descriptor for the de stination code segmen t that contains the called procedure. • The RPL of the segment selector of the destination code segment.
4-16 Vol. 3A PROTECTION The RPL of the segment selector th at points to a nonconforming co de segment has a limited effect on the privilege check. The RPL must be nu merically less than or equal to the CPL of the calling procedure for a successful c ontrol transfer to occur .
Vol. 3A 4-17 PROTECTION In the example in Figure 4-7, code segment D is a conforming code segment. Therefore, calling procedures in both code segment A and B can access code segment D (us ing either segment selector D1 or D2, respectively), because they both have CPLs th at are greater than or equal to the DPL of the conforming code segment.
4-18 Vol. 3A PROTECTION 4.8.3 Call Gates Call gates facilitate controlled transfers of program control be tween dif ferent privilege levels. They are typically used only in operating systems or e xecutive s that use the privilege-level protection mechanism.
Vol. 3A 4-19 PROTECTION Note that the P flag in a gate descriptor is normally always set to 1. If it i s set to 0, a not present (#NP) exception is generated when a program at tempts to access the descriptor . The operating system can use the P flag for special purposes.
4-20 Vol. 3A PROTECTION • T arget code segments referenced by a 64-b it call gate must be 64-bit code segments (CS.L = 1, CS.D = 0). If not, the referen ce generates a general-protection exception, #GP (CS selector). • Only 64-bit mode call g ates can be referenced in IA-32e mo de (64-bit mode and com pati- bility mode).
Vol. 3A 4-21 PROTECTION Figure 4-10. Call-Gate Mech anism Figure 4-1 1. Privilege Check for Control T ransfer with Call Gate Offset Segment Selector Far Pointer to Call Gate Required but not used by p.
4-22 Vol. 3A PROTECTION The privilege checking rules are dif ferent depending on whether the control transfer was initi- ated with a CALL or a JMP instruction, as shown in T able 4-1.
Vol. 3A 4-23 PROTECTION Call gates allow a single code segment to have pr ocedures that can be accessed at dif ferent priv- ilege levels. For examp le, an operating system located in a code segment ma.
4-24 Vol. 3A 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.
Vol. 3A 4-25 PROTECTION 4. T emporarily saves the current valu es of the SS and ESP registers. 5. Loads the segment selector an d stack pointer for the new stack in the SS and ESP registers. 6. Pushes the temporarily saved val ues for the SS and ESP regist ers (for the calling procedure) onto the ne w stack (see Figure 4-13).
4-26 Vol. 3A PROTECTION 4.8.5.1 St ack Switching in 64-bit Mode Although protection-ch eck rules for call gates are unchanged from 32-bit mode, stack-switch changes in 64-bit mode are different.
Vol. 3A 4-27 PROTECTION from the stack into the EIP regi ster , it checks that the pointer does not exceed the limit of the current code segment. On a far return at the same p rivilege level, the processor pops both a segment selecto r for the code segment being returned to and a return instruct ion pointer from the stack.
4-28 Vol. 3A PROTECTION new CPL (excluding conforming code segments), the segment register is loaded with a null segment selector . See the description of the RET instruction in Chap ter 3, Instructio.
Vol. 3A 4-29 PROTECTION 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 thro ugh programm ing conventions.
4-30 Vol. 3A PROTECTION When SYSEXIT transfers contro l to compatibility mode user code when the operand size attribute is 32 bits, the following fields are generated and bits set: • T a rget code segment — Computed by adding 16 to the value in IA32_SYSENTER_CS.
Vol. 3A 4-31 PROTECTION When SYSRET transfers control to 64-bit mode us er code using REX.W , the processor gets the privilege level 3 target instruction and stack pointer from: • T arget code segment — Reads a non-NULL selector from IA32_ST AR[63:48] + 16.
4-32 Vol. 3A PROTECTION 4.9 PRIVILEGED INSTRUCTIONS Some of the system instru ctions (called “privileged instructi ons”) are protected from use by applicatio n pr ogr ams. Th e pri vil ege d i nst ruct ion s control system functions (such as the loading of system registers).
Vol. 3A 4-33 PROTECTION 3. Checking if the pointer of fse t exceeds the segment limit. 4. Check ing if the supp lier of the point er is allowed to access the segment. 5. Checking the of fset alignmen t. The processor automa tically performs fi rst, s econd, and third checks du ring instruction execu- tion.
4-34 Vol. 3A PROTECTION 4.10.2 Checking Read/Write Right s (VERR and VER W Instructions) When the processor accesses any code or data segment it checks the read/write privileges assigned to the segment to verify that the inte nded read or write opera tion is allowed.
Vol. 3A 4-35 PROTECTION 5. If the privi lege level and type checks pass, loads the unscramb led limit (the limit scaled according to the setting of the G flag in the se gment descriptor) into the destination register and sets the ZF flag in the EF LAGS register .
4-36 Vol. 3A PROTECTION Now assume that instead of setti ng the RPL of the segment selector to 3, th e applicatio n program sets the RPL to 0 (segment se lector D2). The opera ting system can now access da ta segment D, because its CPL and the RPL of segm ent selector D2 are both equal to the DPL of data segment D.
Vol. 3A 4-37 PROTECTION application program (represented by the code-seg m ent selector pushed o nto the stack). If the RPL is less than application program’ s privilege level, th e ARPL instruction chang es the RPL of the segment selector to match the privilege level o f the app lic ati on pr ogr am ( seg me nt selector D1).
4-38 Vol. 3A PROTECTION 4.1 1.1 Page-Protection Flags Protection inform ation for pages is contained in tw o flags in a page-directory or page-table entry (see Figure 3-1 4): the read/write flag (b it 1) and the user/supervisor flag (bit 2).
Vol. 3A 4-39 PROTECTION read/write accessible. User -mode pages which are read/write or read-only are readable; super- visor-mode pages are neither readable nor writable from user mode. A page-fault exception is generated on any attempt to violate the protection rules.
4-40 Vol. 3A PROTECTION Page-level protection can be used to enhance se gment-level protection. For example, if a lar ge read-write data segment is paged, the page-pro tection mechanism can be used to write-protect individual pages. NOTE: * If CR0.WP = 1, access type is determined by the R/ W flags of the page-directory and page-table entries.
Vol. 3A 4-41 PROTECTION While the execute disable bit capabi lity does not introduce new instructio ns, it does require operating systems to use a P AE-enabled environm ent and establish a page-granular protection policy for memory pages. If the execute disable bi t of a memory page is set, that page can be used only as data.
4-42 Vol. 3A PROTECTION tures. Execute-disable bit protection can be activ ated using the execute-disable bit at any level of the pagin g structure, irresp ective of the corresponding entry in other levels. When execute- disable-bit protection is not activated, the page can be used as code or data.
Vol. 3A 4-43 PROTECTION 4.13.3 Reserved Bit Checking The processor enforces reserved bit checking in paging data structu re entries. The bits being checked varies with pa ging mode and may vary with the size of ph ysical address space. T able 4-9 shows the reserved bits that are checked when the execu te disable bit capability is enabled (CR4.
4-44 Vol. 3A PROTECTION T abl e 4-10. Re served Bit Checking WIth Ex ecute-Disable Bit Cap a bility Not Enabled 4.13.4 Exception Handling When execute disable bit capab ility is enabled (IA32_EFER.
5 Interrupt and Exception Handling.
.
Vol. 3A 5-1 CHAPTER 5 INTERRUPT AND EXCEPTION HANDLING This chapter describes the processor ’ s interr upt and ex ception-handling mechanism wh en oper- ating in protected mode. Most of the information pro vided here also applies to interrupt and exception mechanisms used in real-address, virtual-8086 mode, and 64-bit mode.
5-2 Vol. 3A INTERRUPT AND EXCEPTION HANDLING 5.2 EXCEPTION AND INTERRUPT V ECTORS T o aid in handling exceptions and interrupts , each IA-32 architectur e-defined exception and each interrupt condition th at requires special handling by the processor is assigned a unique identification number , called a vect or .
Vol. 3A 5-3 INTERRUPT AND EXCEPTION HANDLING T able 5-1. Protected-Mod e Exceptions and Interrup ts V ector No. Mne- monic Description T ype Error Code Source 0 #DE Divide Error Fault No DIV and IDIV instructions. 1 #DB RESERVED Fault/ Tra p No For Intel use only .
5-4 Vol. 3A INTERRUPT AND EXCEPTION HANDLING 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 lo cal APIC through the system bus (Pentium 4 an d Intel Xeon processo rs) or the APIC serial bus (P6 family and Pentium processors).
Vol. 3A 5-5 INTERRUPT AND EXCEPTION HANDLING 5.4 SOURCES OF EXCEPTIONS The processor receives excep tions from three sources: • Processor -detected pr ogram-error exceptions.
5-6 Vol. 3A INTERRUPT AND EXCEPTION HANDLING • Faults — A fault is an exception that can genera lly be corrected and that, once corrected, allows the program to be restarted with no lo ss of contin uity .
Vol. 3A 5-7 INTERRUPT AND EXCEPTION HANDLING For trap-class exceptions, the return instruction pointer poi nts to the instruction following the trapping instruction. If a trap is detected during an instruc tion which transfers execution, the return instruction pointer reflects the transfer .
5-8 Vol. 3A INTERRUPT AND EXCEPTION HANDLING 5.7 NONMASKABLE INTERRUPT (NMI) The nonmaskable interrupt (NMI) can be ge nerat ed in eith er of tw o ways: • External hardware asserts the NMI pin.
Vol. 3A 5-9 INTERRUPT AND EXCEPTION HANDLING 5.8.1 Masking Maskable Hardware Interrupts The IF flag can disable the servicing of ma skable hardware interrupts received on the processor ’ s INTR pin or through the local APIC (see Section 5. 3.2, “M askable Hardware Inter- rupts”).
5-10 Vol. 3A INTERRUPT AND EXCEPTION HANDLING Manual, V o lume 2A, for a detailed description of the operations these instructions are allowed to perform on the IF flag.
Vol. 3A 5-11 INTERRUPT AND EXCEPTION HANDLING While priority among these classes listed in T abl e 5-2 is consistent th roughout the architecture, exceptions within each class are implementatio n-dependent and may vary from processor to processor .
5-12 Vol. 3A INTERRUPT AND EXCEPTION HANDLING re-generated when the interrupt handler returns ex ecution to the point in t he program or task where the exceptions and/or interrupts occurred.
Vol. 3A 5-13 INTERRUPT AND EXCEPTION HANDLING 5.1 1 IDT DESCRIPTORS The IDT may contain any of th ree ki n ds of gate descriptors: • T ask-gate descriptor • Interrupt-gate descriptor • T rap-gate descriptor Figure 5 -2 shows the form ats for the task-gate, interrupt-gate, and trap -gate descri ptors.
5-14 Vol. 3A INTERRUPT AND EXCEPTION HANDLING 5.12 EXCEPTION AND INTERRUPT HANDL ING The processor handles calls to exception- and interrupt -handlers similar to the way it h andles calls with a CALL in struction to a procedure or a task.
Vol. 3A 5-15 INTERRUPT AND EXCEPTION HANDLING through Section 4.8.6, “Returning from a Called Pr oced ure”). If index poin ts to a task g ate, the processor executes a task switch to the exception- or interrupt -han dler task in a manner similar to a CALL to a task gate (see Section 6.
5-16 Vol. 3A 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 execute d at a numerically lower privilege level, a stack switch occurs. When the stack switch occurs: a.
Vol. 3A 5-17 INTERRUPT AND EXCEPTION HANDLING T o return from an exception- or interrupt-handl er procedure, the handler must use th e IRET (or IRETD) instruction. The IRET instruction is similar to t he RET instruction except that it restores the saved flags into the EFLAGS register .
5-18 Vol. 3A INTERRUPT AND EXCEPTION HANDLING An attempt to violate this rule results in a general-protection exception (#GP). The protection mechanism for exception- and interrupt-handler procedures .
Vol. 3A 5-19 INTERRUPT AND EXCEPTION HANDLING 5.12.2 Interrupt T asks When an exception or interrupt handler is accessed through a task gate in the IDT , a task switch results. Handling an exception or interrupt wit h a separate task offers several advantages: • The entire context of the interrupted pr ogram or task is saved auto mat ically .
5-20 Vol. 3A INTERRUPT AND EXCEPTION HANDLING Figure 5-5. Interrupt T ask Switch IDT T ask Gate TSS for Interrupt- TSS Selector GDT TSS Descriptor Interrupt Ve c t o r TSS Base Address Handling T ask.
Vol. 3A 5-21 INTERRUPT AND EXCEPTION HANDLING 5.13 ERROR CODE When an exception condition is related to a specific segment, the processor pushes an error code onto the stack of the exception ha ndler (whether it is a procedure or task). The error code has the format shown in Figure 5-6.
5-22 Vol. 3A INTERRUPT AND EXCEPTION HANDLING 5.14 EXCEPTION AND INTERRUP T HANDLING IN 64-BIT MODE In 64-bit mode, interrupt and excep tio n handling is similar to what has been described for non- 64-bit modes.
Vol. 3A 5-23 INTERRUPT AND EXCEPTION HANDLING 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 in terrupt gate are similar but not ident ical to legacy 32-bit interrupt gates.
5-24 Vol. 3A INTERRUPT AND EXCEPTION HANDLING 5.14.3 IRET in IA-32e Mode In IA-32e mode, IRET executes with an 8-byte op erand size. There is no thing that fo rces this requirement. The stack is formatted in such a wa y that for actions where IRET is required, the 8-byte IRET operand size works correctly .
Vol. 3A 5-25 INTERRUPT AND EXCEPTION HANDLING In summary , a stack switch in IA -32e mode work s like the legacy stack switch, except that a new SS selector is not loaded from the TSS.
5-26 Vol. 3A INTERRUPT AND EXCEPTION HANDLING The IST mechanism provides up to seven IST poin ters in the TSS. The pointers are referenced by an interru pt-gate descript or in the interrup t-descriptor table (IDT); see Figure 5-7. The gate descriptor cont ains a 3-bit IST index field that pr ovides an offset into the IST section of the TSS.
Vol. 3A 5-27 INTERRUPT AND EXCEPTION HANDLING Interrupt 0—Divide Er ror Exception (#DE) Exception Class Fault. Description Indicates the divisor operan d for a DIV or IDIV inst ruction is 0 or that the result cannot be repre- sented in the number of bits specified for the destination operand.
5-28 Vol. 3A 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 and the ot her deb ug regi sters. Descripti on Indicates that one or more of several debug-ex ception conditions has been detected.
Vol. 3A 5-29 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.
5-30 Vol. 3A INTERRUPT AND EXCEPTION HANDLING Interrupt 3—Breakpoint Exception (#BP) Exception Class Tr a p . Descripti on Indicates that a breakpoint inst ruction (INT 3) w as executed, causing a breakpoint trap to be generated.
Vol. 3A 5-31 INTERRUPT AND EXCEPTION HANDLING Interrupt 4—Overfl ow Exception (#OF) Exception Class Tr a p . Description Indicates that an overflow tr ap occurred when an INTO in struction was executed. The INT O instruction checks the state of the OF flag in the EFLAGS register .
5-32 Vol. 3A INTERRUPT AND EXCEPTION HANDLING Interrupt 5—BOUND Range Exceeded Exception (#BR) Exception Class Fault. Descripti on Indicates that a BOUND-rang e-exceeded fault occurred wh en a BOUND instruction was executed.
Vol. 3A 5-33 INTERRUPT AND EXCEPTION HANDLING Interrupt 6—Invalid Opcode Exception (#UD) Exception Class Fault. Description Indicates that the processor did one of the follo wing things: • Attempted to execute an invalid or reserved opcode.
5-34 Vol. 3A INTERRUPT AND EXCEPTION HANDLING The opcodes D6 and F1 are undefined opcodes that are reserved by the IA-32 architecture. These opcodes, even though undefined, do not generate an inval id opco de excepti on. The UD2 instruction is guaranteed to generate an invalid opcode exception.
Vol. 3A 5-35 INTERRUPT AND EXCEPTION HANDLING Interrupt 7—Device Not A vailable Exception (#NM) Exception Class Fault. Description Indicates one of the following thing s: The device-not-available ex.
5-36 Vol. 3A INTERRUPT AND EXCEPTION HANDLING Saved Instruct ion Pointer The saved contents of CS and EIP registers poi nt to the floating-point instruction or the W AIT/FW A IT instructi on th at generated the exception.
Vol. 3A 5-37 INTERRUPT AND EXCEPTION HANDLING Interrupt 8—Double Fault Exception (#DF) Exception Class Abort. Description Indicates that the processor detected a seco nd except ion wh ile calling an exceptio n h andler for a prior exception.
5-38 Vol. 3A INTERRUPT AND EXCEPTION HANDLING If another exception occurs while attemp ting to call the double-faul t handler , the processor enters shutdown mo de.
Vol. 3A 5-39 INTERRUPT AND EXCEPTION HANDLING Interrupt 9—Coprocessor Segment Overrun Exception Class Abort. (Intel r eserved; do not use. Recent IA-32 processors do not generate this exception.
5-40 Vol. 3A INTERRUPT AND EXCEPTION HANDLING Interrupt 10—Invalid TSS Exception (#TS) Exception Class Fault. Descripti on Indicates that there was an error related to a TS S. Such an error might be detected during a task switch or during the execution of instructions that use informat ion from a TSS.
Vol. 3A 5-41 INTERRUPT AND EXCEPTION HANDLING This exception can generated either in the context of the original task or in the context of the new task (see Section 6.3, “T ask Switching”). Un til the processor has completely verified the presence of the new TSS, the exception is generate d in the context of the original task.
5-42 Vol. 3A INTERRUPT AND EXCEPTION HANDLING Exception Error Co de An error code containing the segm ent selector index for the segm ent descriptor that caused the violation is pushed onto the stack o f the exception handler .
Vol. 3A 5-43 INTERRUPT AND EXCEPTION HANDLING Interrupt 1 1—Segment Not Present (#NP) Exception Class Fault. Description Indicates that the present flag of a segment or g ate 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.
5-44 Vol. 3A INTERRUPT AND EXCEPTION HANDLING Saved Instruct ion Pointer The saved contents of CS and EIP registers norma lly point to the instruct ion that generated the exception.
Vol. 3A 5-45 INTERRUPT AND EXCEPTION HANDLING Interrupt 12—St ack 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 operatio n that refers to the SS register .
5-46 Vol. 3A INTERRUPT AND EXCEPTION HANDLING exception. The stack fau lt handler should thus not rely on being abl e to use the segment select ors fou nd i n t he CS, SS, DS, ES, FS, and GS registers without causing another exception.
Vol. 3A 5-47 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 viol ations called “general- protection violations.
5-48 Vol. 3A INTERRUPT AND EXCEPTION HANDLING • Loading the CR0 register with a se t NW flag and a clear CD flag. • Referencing an entry in the IDT (followin g an interrupt or exception) t hat is not an interrupt, trap, or task gate.
Vol. 3A 5-49 INTERRUPT AND EXCEPTION HANDLING • A selector from a TSS invo lved in a task switch. • IDT vector number . Saved Instruction Pointer The saved contents of CS and EIP registers poin t to the in struction that generated the exception.
5-50 Vol. 3A INTERRUPT AND EXCEPTION HANDLING • If the segment descriptor from a 64-b it call gate is in non-canonical space. • If the DPL from a 64- bit call-gate is less th an 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.
Vol. 3A 5-51 INTERRUPT AND EXCEPTION HANDLING Interrupt 14—Page-Fault Exception (#PF) Exception Class Fault. Description Indicates that, with paging enable d (the PG flag in the CR0 regi st er is se.
5-52 Vol. 3A INTERRUPT AND EXCEPTION HANDLING — The RSVD flag indicates that the processor detected 1s in reserved bits of the page directory , when the PSE or P AE flags in co ntrol register CR4 are set to 1.
Vol. 3A 5-53 INTERRUPT AND EXCEPTION HANDLING Saved Instruction Pointer The saved contents of CS an d EIP registers genera lly point t o the instru ction that generated the exception.
5-54 Vol. 3A INTERRUPT AND EXCEPTION HANDLING When executing this code on one of the 32-bit IA-32 processors, it is possib le to get a page fault, general-protection fault (#G P), or alignment ch eck fault (#AC) after the segment selector has been loaded into the SS register but before the ESP register has b een load ed.
Vol. 3A 5-55 INTERRUPT AND EXCEPTION HANDLING Interrupt 16—x87 FPU Floa ting-Point Error (#MF) Exception Class Fault. Description Indicates that the x87 FPU has detected a floati ng-point error . The NE flag in the register CR0 must be set for an interrupt 16 (floating-point error exceptio n) to be generated.
5-56 Vol. 3A INTERRUPT AND EXCEPTION HANDLING Prior to executing a waiting x87 FPU instruction or the W AIT/FW AIT instruction, the x87 FPU checks for pending x87 FPU floating -point exceptions (as described in step 2 above).
Vol. 3A 5-57 INTERRUPT AND EXCEPTION HANDLING Interrupt 17—Alignment Check Exception (#AC) Exception Class Fault. Description Indicates that the processor detected an una ligned memory operand wh en alignment checking was enabled. Alignment checks are only carried out in data (or stack) accesses (not in code fetches or system segment accesses).
5-58 Vol. 3A INTERRUPT AND EXCEPTION HANDLING Alignment-check excep tions (#AC) are generated only when operating at privilege l evel 3 (user mode). Memory references that default to privileg e level 0, such as segment descriptor loads, do not generate alignment-check exceptions, even when caused by a memory reference made fro m privilege l evel 3.
Vol. 3A 5-59 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 erro r , or tha t an external agent detected a bus error .
5-60 Vol. 3A INTERRUPT AND EXCEPTION HANDLING Program St ate Change The machine-check mechanism is enabled by sett ing the MCE flag in control register CR4. For the Pentium 4, Intel Xeon, P6 family , and Pentium processors, a p rogram-state change always accompanies a machine-check exception, and an abort class exception is generated.
Vol. 3A 5-61 INTERRUPT AND EXCEPTION HANDLING Interrupt 19—SIMD Floati ng-Point Exception (#XF) Exception Class Fault. Description Indicates the processor has detected an SSE/ SSE2/SSE3 SIMD floating-point exceptio n.
5-62 Vol. 3A INTERRUPT AND EXCEPTION HANDLING Note that because SIMD floatin g-point exceptions are precise an d occur immediately , the situ- ation does not arise where an x87 FPU instruct ion, a W AIT/FW AIT inst ruction, or another SSE/SSE2/SSE3 instruction will catch a pen ding unmasked SIMD floatin g-point exception.
Vol. 3A 5-63 INTERRUPT AND EXCEPTION HANDLING Saved Instruction Pointer The saved contents of CS and EIP registers po int to the SSE/SSE2/SSE3 instruction th at was executed when the SIMD floating- point exception was generated. This is the faulting in struction in which the error condition was detect ed.
5-64 Vol. 3A INTERRUPT AND EXCEPTION HANDLING Interrupt s 32 to 255— User Defined Interrupt s Exception Class Not applicable. Descripti on Indicates that the processor did one of the following things: • Executed an INT n instruction where the instruction op erand is one of the vector numbers from 32 thro ugh 255.
6 T ask Management.
.
Vol. 3A 6-1 CHAPTER 6 T ASK MANAGEMENT This chapter describes the IA-32 architecture’ s ta sk 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.
6-2 Vol. 3A T ASK MANAGEMENT 6.1.2 T ask St ate The following items define the stat e of the currentl y 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.
Vol. 3A 6-3 T ASK MANAGEMENT 6.1.3 Executing a T ask 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 instru ction . • A explicit jump to a task with t he JMP instruction.
6-4 Vol. 3A T ASK MANAGEMENT Use of task managem ent facilities for handl ing multitasking appl ications is optional. M ulti- tasking can be handled in software, with each so ftware defined task executed in the context of a single IA-32 architecture task.
Vol. 3A 6-5 T ASK MANAGEMENT The processor updates dynamic fields when a task is suspended during a task switch. The following are d ynamic fields: • General-purpose register fields — State of the EAX, E CX, EDX, EBX, ESP , EBP , ESI, and EDI registers prior to the task switch.
6-6 Vol. 3A T ASK MANAGEMENT • EFLAGS register field — State of the EF AGS register prior to the t ask switch. • EIP (instruction poin ter) field — State of the EIP register prior to the task switch.
Vol. 3A 6-7 T ASK MANAGEMENT 6.2.2 TSS Descriptor The TSS, like all other segments, is defined by a segment descriptor . Figure 6-3 shows the format of a TSS descriptor . TS S descriptors may only be placed in the GDT ; they cannot be placed in an LDT or the IDT .
6-8 Vol. 3A T ASK MANAGEMENT The base, limit, and DPL fields an d the granular ity and present flags have function s similar to their use in data-segment descriptors (see Sect ion 3.
Vol. 3A 6-9 T ASK MANAGEMENT 6.2.4 T ask Register The task register holds the 16-bit segment select or and the entire segment descriptor (32-bit base address, 16-bit segment limit, and descript or attributes) for th e TSS of the current task (see Figure 2-5).
6-10 Vol. 3A T ASK MANAGEMENT The L TR instruction loads a segment selector (s ou rce operand) into the task register that points to a TSS descriptor in the GDT . It then loads the invisi ble portion of the task register with infor- mation from the TSS de scriptor .
Vol. 3A 6-11 T ASK MANAGEMENT 6.2.5 T ask-Gate Descriptor A task-gate descriptor provides an indirect, prot ected reference to a task (see Figure 6-6). It can be placed in the GDT , an LDT , or the IDT . Th e TSS segment selector field in a task-gate descriptor points to a TSS descriptor in the G DT .
6-12 Vol. 3A T ASK MANAGEMENT Figure 6-7 i llustrates how a task gat e in an LDT , a task gate in the GDT , and a task gate in the IDT can all point to the same task.
Vol. 3A 6-13 T ASK MANAGEMENT JMP , CALL, and IRET instructions, as well as inte rrupts and exceptions, are all mechanisms 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 IR ET instruction) determines whether a task switch occurs.
6-14 Vol. 3A T ASK MANAGEMENT 10. I f the task switch was initiated with a CALL instru ctio n, JMP inst ruction , an excepti on, or an interrupt, the processor set s the busy (B) flag in the new task’ s TSS descriptor; if initiated with an IRET instruct ion , the busy (B) flag is left set.
Vol. 3A 6-15 T ASK MANAGEMENT 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 sp ecified in the CPL field of the CS register , which is loaded from the TSS.
6-16 Vol. 3A T ASK MANAGEMENT The TS (task switched) flag in the control regist er CR0 is set every time a task switch occurs. System software uses the TS flag to coordina te the actions of floating-point unit when gener- ating floating-poi nt exceptions with the rest of the processor .
Vol. 3A 6-17 T ASK MANAGEMENT T able 6-2 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 sw itch. The NT flag may be modified by software executing at any privilege level.
6-18 Vol. 3A T ASK MANAGEMENT 6.4.1 Use of Busy Flag T o Pr event Recursive T ask Switching A TSS allows only one con text to be saved for a task; therefore, once a task is called (dispatched), a recursive (or re-entrant) call to th e task would cause the current state of the task to be lost.
Vol. 3A 6-19 T ASK MANAGEMENT In a multiprocessing system, additional sy nchronization and serialization operations must be added to this procedure to insu re that the TSS and its segment descriptor are both locked when the previous task link field is ch anged and the busy flag is cleared.
6-20 Vol. 3A T ASK MANAGEMENT that the mapping of TSS addresses does not chan ge while the processor is reading and updating the TSSs during a task switch. The linear ad dress 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.
Vol. 3A 6-21 T ASK MANAGEMENT • Through segment descriptors in distinct LD T s that are mapped to common addr esses in linear address space — If this common area of the li near address space is mapped to the same area of the physical address space for each task, th ese segment descriptors permit the tasks to share segments.
6-22 Vol. 3A T ASK MANAGEMENT Figure 6-10. 16-Bit TSS Format T ask LDT Sele ctor DS Selector SS Selector CS Selector ES Selector DI SI BP SP BX DX CX AX FLAG Wo rd IP (Entry Point) SS2 SP2 SS1 SP1 SS0.
Vol. 3A 6-23 T ASK MANAGEMENT 6.7 T ASK MANAGEMENT IN 64-B IT MODE In 64-bit mode, task stru cture and task state are simi lar to those in protecte d mode. However , the task switching mechanism av ailable in protected mo de is not supported in 64-bit m ode.
6-24 Vol. 3A T ASK MANAGEMENT Figure 6-1 1. 64-Bit TSS Format 0 31 100 96 92 88 84 80 76 I/O Map Base Address 15 72 68 64 60 56 52 48 44 40 36 32 28 24 20 16 12 8 4 0 RSP0 (lower 32 bits) RSP1 (lower 32 bits) RSP2 (lower 32 bits) Reserved bits. Set to 0.
7 Multiple-Pr ocessor Management.
.
Vol. 3A 7-1 CHAPTER 7 MULTIPLE-PROCESSOR MANAGEMENT The IA-32 architecture pr ovides sever al mechanisms for managing and improving the pe rfor- mance of multiple processors connect ed to the same system bus. These mechanisms include: • Bus locking and/or cache coherency m anagement for performing atom ic operations on system memory .
7-2 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT • T o distribut e interrupt handling among a group o f processors — When several processors are operating in a system in parallel, it is useful to have a cen tralized mechanism for receiving interrupts and distributing them to availa ble processors for servicing.
Vol. 3A 7-3 MULTIPLE-PROCESSOR MANAGEMENT The mechanisms for handling locked atomic operati ons have evolved as the complexity of IA-32 processors has evolved. As such, more recent IA-32 processors (such as the Pentium 4, Intel Xeon, and P6 family processors) provide a more refined locking mechanism than earlier IA-32 processors.
7-4 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT For the Pentium 4, Intel Xeon, and P6 family processors, if the memory area being accessed is cached internally in the processor , the LOCK# si gnal is generally not asserted; instead, locking is only applied to the processor ’ s caches (see Section 7.
Vol. 3A 7-5 MULTIPLE-PROCESSOR MANAGEMENT 7.1.2.2 Software Contro lled Bu s Locking T o explicitly force the LOCK semantics, softwa re can use the LOCK prefix with the following instructions when they are us ed to modi fy a memory locati on.
7-6 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT Locked instructions should not be used to insure that data written can be fetched as instructions. NOTE The locked instructions fo r the current versions of the Pentium 4, Intel X eon, P6 family , Pentium, and Intel486 proce ssors al low data written to be fetched as instructions.
Vol. 3A 7-7 MULTIPLE-PROCESSOR MANAGEMENT T o write cros s-modifying code and insure that it is compliant with curren t and future versions of the IA-32 architecture, the following processor synchroni.
7-8 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT T o all ow optimizing of instruction execution, the IA-32 architecture al lows departures from strong-ordering model called processor ordering in Pentium 4, Intel Xeon, and P6 family processors.
Vol. 3A 7-9 MULTIPLE-PROCESSOR MANAGEMENT 4. Writes can be buffered. 5. Writes are not performed specu latively; they are only performed for instructions that have actually been retired. 6. Data fro m buffered writes can be forwar ded to waiting reads within the proces sor .
7-10 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT 7.2.3 Out-of-Order Stores For S tri ng Operations in Pentium 4, Intel Xeon, and P6 Family Processors The Pentium 4, Intel Xeon, and P6 family proce ssors modify the processors operation during the string store operations (initiated with the M OVS and STOS instructions) to maximi ze perfor- mance.
Vol. 3A 7-11 MULTIPLE-PROCESSOR MANAGEMENT • 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, Pentium 4 and Intel Xeon processors; 32 bytes P6 family and Pentium processors).
7-12 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT Program synchronization can also be carried out with serializin g instru ctions (see Section 7.4). These instructions are typically used at critical procedure or task boundaries to force completion of all previous instructions befo re a jump to a new section of code or a context switch occurs.
Vol. 3A 7-13 MULTIPLE-PROCESSOR MANAGEMENT It is recommended that software written to run on Pentium 4, Intel Xeon, and P6 family proces- sors assume the processor-ordering model or a weaker memory-orderi ng model.
7-14 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT 7.4 SERIALIZING INSTRUCTIONS The IA-32 architecture defines several serializing instructions . These instructions force the processor to complete all modifi .
Vol. 3A 7-15 MULTIPLE-PROCESSOR MANAGEMENT • When an instruction is execute d that enables or disables paging (that is, chang es the PG flag in control register CR0), the instruction should be followed by a jump instruct ion.
7-16 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT • Intel Xeon pr ocessors with family , model, and st epping IDs up to F09H — The selection of the BSP and APs (see Section 7.5. 1, “BSP and AP Processors ”) is handled through arbitration on the system bus, usin g BIPI and FIPI messages (see Secti on 7.
Vol. 3A 7-17 MULTIPLE-PROCESSOR MANAGEMENT • All devices in the system that are capable of delivering interrupts to the processors must be inhibited from doing so for t he duration of the MP initializati on protocol.
7-18 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT • The remainder of the processors (which were not selected as the BSP) are designated as APs. They leave their BSP fl ags in the clear state and enter a “wait- for-SIPI state.
Vol. 3A 7-19 MULTIPLE-PROCESSOR MANAGEMENT The following constants and data definitions are used in the accompanying code examples. They are based on the addresses of the APIC registers as defined in T a bl e 8-1.
7-20 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT space (1-MByte space). For example, a vect or of 0BDH specifies a sta rt-up memory address of 000BD000H. 1 1.
Vol. 3A 7-21 MULTIPLE-PROCESSOR MANAGEMENT 16. W aits for the timer interrupt. 17. Reads and evaluates the COUNT variable and establishes a processor count. 18. If necessary , reconfigures the APIC and c ontinues with the remaining system diagnostics as appropriate.
7-22 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT 7.5.5 Identifying Logical Proc essors in an MP System After the BIOS has completed the MP initiali zation protocol, each logical processor can be uniquely identified by its local APIC ID.
Vol. 3A 7-23 MULTIPLE-PROCESSOR MANAGEMENT For P6 family processors, the APIC ID that is assigned to a processor durin g power-up and initialization is 4 bits (see Figure 7-2). Here, bits 0 an d 1 form a 2-bit processor (or socket ) iden- tifier and bits 2 and 3 form a 2-bit cluster ID.
7-24 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT 7.7 DETECTING HARDWARE MU LTI-THREAD ING SUPPORT AND T OPOLOGY Use the CPUID instruction to detect the presence of ha rdware multi-thread ing support in a ph ys- ical processor . The foll owing can be interpreted : • Hardware Multi-Threading featur e flag (CPUID.
Vol. 3A 7-25 MULTIPLE-PROCESSOR MANAGEMENT 7.7.2 Initializing Dual-Core IA-32 Processors The initialization process fo r an MP system that contains dual-core IA-32 processors is the same as for conventional MP systems (see Section 7.5, “Multiple-Processor (MP) Initialization”).
7-26 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT 7.8 INTEL ® HYPER-THREADING T ECHNOLOGY ARCHITECTURE Figure 7-4 shows a generalized v iew o f an IA-32 processor supportin g Hyper-Threading T ech- nology , using the Intel Xeon processor MP as an exampl e.
Vol. 3A 7-27 MULTIPLE-PROCESSOR MANAGEMENT 7.8.1 St ate of th e Logical Processors The following features are part of the architectural state of l ogical processors within IA-32 processors supporting Hyper-Threading T echnology .
7-28 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT • Debug registers (DR0, DR1, DR2, DR3, DR6, DR7) and the debug control MSRs • Machine check global status (IA32_MC G_ST A TUS) and machine check capabili.
Vol. 3A 7-29 MULTIPLE-PROCESSOR MANAGEMENT of memory , independent of the processor on whic h it is running. See Section 10.11, “Memory T ype Range Registers ( MTRRs),” for inf ormation on setti ng up MTRRs. 7.8.4 Page Attribute T able (P A T) Each logical p rocessor has its own P A T MSR (IA32_CR_ P A T).
7-30 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT The performance counter i nterrupts, events, and precise event monitoring support can be set up and allocated on a per thread (p er logical processor) basis.
Vol. 3A 7-31 MULTIPLE-PROCESSOR MANAGEMENT 7.8.12 Self Modifying Code IA-32 processors supporting Hyper-Threading T echnology support self-modifying code , where data writes modify instructions cached or currentl y in flight.
7-32 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT Entries in the TLBs are tagged wi th an ID that indicates the logi cal processor that init iated the translation. This tag appl ies ev en for translations that are ma rked global using the pag e global feature for memory paging .
Vol. 3A 7-33 MULTIPLE-PROCESSOR MANAGEMENT vector tables for one or both of the logical processors. T ypically in MP systems, the LINT0 and LINT1 pins are not used to deliver interrup ts to the logical processors. Instead all interrup ts are delivered to the local processors through the I/O APIC.
7-34 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT 7.9.2 Memory T ype Ra nge Registers (MTRR) MTRR is shared between two lo gical processors sharing a pr ocessor core if the physical processor supports Hyper-Threading T echnology . MTRR is not shared between logi cal proces- sors located in different cores or different physical packages.
Vol. 3A 7-35 MULTIPLE-PROCESSOR MANAGEMENT 7.10 PROGRAMMING CONSID ERATIONS FOR HARDWARE MULTI-THREADING CAP ABLE PROCESSORS In a multi-threading en vironment, there may be certain ha rdware resources that are physically shared at some level of the hard ware topology .
7-36 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT The value of valid APIC_IDs need not be cont iguo us across package boundary or co re bound- aries. 7.10.2 Identifying Logical Proc essors in an MP System Fo.
Vol. 3A 7-37 MULTIPLE-PROCESSOR MANAGEMENT T able 7-2 sho ws the initial APIC IDs for a hy pothetical situation with a dual processor system. Each physical package providing two processor cores, and each processor core also supporting Hyper-Threading T echnology .
7-38 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT 7.10.3 Algorithm for Three-L evel Mappings of APIC_ID Software can gather the initial APIC_IDs for each logical pro cessor supported by the operating system at runtime 4 and extract identifiers corresponding to the three levels of sharing topology (package, core, and SMT).
Vol. 3A 7-39 MULTIPLE-PROCESSOR MANAGEMENT unsigned int HW MTSupported(void) { try { // verify cpuid in struction is supported execute cpuid with eax = 0 to ge t vendor string execute cpuid with eax =.
7-40 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT store returned value of eax return (unsigned ) ((reg_eax >> 26) +1); } else // must be a single-core processor return 1; } 4.
Vol. 3A 7-41 MULTIPLE-PROCESSOR MANAGEMENT 6. Extract a sub ID given a full ID, maximum sub ID valu e and shift count. // Returns the value of the sub ID, this is not a zero-based value Unsigned char .
7-42 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT 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, P ACKAGE_IDs into arrays for each enabled logical processor .
Vol. 3A 7-43 MULTIPLE-PROCESSOR MANAGEMENT Example 7-3 Compu te the Number of Packag es, Cores, and Proce ssor Relationships in a MP System a) Assemble lists of PACKAGE_ID, CORE_ID, and SMT_ID of each enabl ed logical processors //The BIOS and/or OS may limit the number of logical p rocessors available to app lications // after system boot.
7-44 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT The algorithm below assumes there is symmetry across package boundary if more than one socket is populated in an MP system.
Vol. 3A 7-45 MULTIPLE-PROCESSOR MANAGEMENT If ((PackageID[ProcessorNum] | CoreID[ProcessorNum]) == CoreIDBucket[i]) { CoreProcessorMask[i] |= ProcessorMask; Break; // found in existing bucket, skip to.
7-46 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT 7.1 1.2 P AUSE Instruction The P AUSE instruction improves the performa nce of IA-32 processors supporting Hyp er- Threading T ech nology when executing “spin-wait loo ps” and other routines where one thread is accessing a shared lock or semaphore in a tig ht polling loop.
Vol. 3A 7-47 MULTIPLE-PROCESSOR MANAGEMENT 7.1 1.4 MONIT OR/MW AIT Instruction Operating systems usually im plement idle loops to handle th read synchronization. In a typical idle-loop scenario, there could be several “busy loops” and they would use a set o f memory loca- tions.
7-48 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT Power management related events (such as Thermal Monit or 2 or chipset driven STPCLK# assertion) will not cause th e monitor event pendi ng flag to be cleared. Faults will not cause the monitor event pending flag to be cleared.
Vol. 3A 7-49 MULTIPLE-PROCESSOR MANAGEMENT These above two values bear no relationship to cache line size in the syst em and software should not make any assumptions to th at effect.
7-50 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT 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 Critic al_Section: <critical section code> MOV lockvar, 0 .
Vol. 3A 7-51 MULTIPLE-PROCESSOR MANAGEMENT The MONITOR and MWAIT instructions may be consi dered for use in the C0 i dle state loops, if MONITOR and MWAIT are supported. Example 7-6 An OS Idle Loop with MONIT OR/MW AIT in the C0 Idle Lo op // WorkQueue is a memory locati on indicating there is a thread // ready to run.
7-52 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT other logical processors in the physical package. For this reason, halting idl e logical processors optimizes the performance. 5 If all logical processors within a physical package are halted, the processor will enter a power-saving state.
Vol. 3A 7-53 MULTIPLE-PROCESSOR MANAGEMENT 7.1 1.6.5 Guidelines for Scheduling Threads on Logic al Processors Sharing Execution Resour ces Because the logical processors, the order in which threads are dispatched to logical processors for execution can affect the overa ll efficiency of a system.
7-54 Vol. 3A MULTIPLE-PROCESSOR MANAGEMENT.
8 Advanced Pr ogrammable Interrupt Contr oller (APIC).
.
Vol. 3A 8-1 CHAPTER 8 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) The Advanced Programmable Interrupt Controll er (APIC), referred to in the follo wing sections as the local APIC, was introdu ced into the IA-32 processors with the Pentium processor (see Section 17.
8-2 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) Local APICs can receive interrupt s from the followi ng 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 interrupt pins (LINT0 and LINT1).
Vol. 3A 8-3 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) Xeon processors) or on the APIC bus (for Pentiu m and P6 family processors). See Section 8.2, “System Bus Vs. APIC Bus.” IPIs can be sent to other I A-32 processors in the system or to the originating processor (self- interrupts).
8-4 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) processors through the local inte rrupt pins; however , this mechanism is com monly not used in MP systems. Figure 8-2. Local APICs and I/ O APIC When Intel Xeon Processors Are Used in Multiple- Processor Systems Figure 8- 3.
Vol. 3A 8-5 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) The IPI mechanism is typically used in MP syst ems to send fixed interrupts (interrupts for a specific vector number) and special-purpose inte rrupts to processors on the system bus.
8-6 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 8.4.1 The Local APIC Block Diagram Figure 8-4 gives a functional block diagram for the local APIC.
Vol. 3A 8-7 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) Figure 8-4. Local APIC Structure Current Count Register Initial Count Register Divide Configuration Register V ersion Register Error S tat.
8-8 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) T able 8-1 shows how the APIC registers are mapped into the 4-KByte APIC register s pace. Registers are 32 bits, 64 bits, o r 256 bits in width; all are aligned on 128-bit boun daries. All 32-bit registers should be accessed using 128-bit aligned 32-bit loads or stor es.
Vol. 3A 8-9 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 8.4.2 Presence of the Local APIC Beginning with the P6 family processors, the pr esence or absence of an on-chip local APIC can be detected using the CPUID inst ruction.
8-10 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 8.4.3 Enabling or Disa bling the Local APIC The local APIC can be enabled or disabled in either of two ways: 1.
Vol. 3A 8-11 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 8.4.4 Local APIC St atus and Location The status and location of the local APIC ar e contained in the IA 32_APIC_BASE MSR (see Figure 8-5). MSR bit functions are described belo w: • BSP flag, bit 8 ⎯ Indicat es if the processor is the boot strap processor (BSP).
8-12 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 8.4.6 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).
Vol. 3A 8-13 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 8.4.7.1 Local APIC St ate After Power-Up or Reset Following a power-up or RESET of the processor , the state of local APIC and its regist.
8-14 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 8.4.7.3 Local APIC St ate Af ter an INIT Reset (“W ait-for-SIPI” St ate) An INIT reset of the processor can be initiated in either of two ways: • By asserting the processor ’ s INIT# pin.
Vol. 3A 8-15 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 8.5 HANDLING LOCAL INTERRUPT S The following sections describe facilities that are provided in the local APIC for h andling local interrupts.
8-16 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) monitor register and its associ ated interrupt were introduced in the Pentium 4 and Inte l Xeon processors. As shown in Figures 8-8, some of t hese fields and flags are not availabl e (and reserved ) for some entries.
Vol. 3A 8-17 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) The setup information that can be specified in the registers of the L VT table is as follows: V ector Interrupt vector numb er . Delivery Mode Sp ecifies the type of interrupt to be sent to the processor .
8-18 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) Remote IRR Flag (R ead Only) For fixed mode, level-triggered interrupts; this flag is set when the local APIC accepts the interrupt fo r servicing and is reset when an EOI command is received from the proces sor .
Vol. 3A 8-19 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 8.5.3 Error Handling The local APIC provides an error status register (ESR) that it uses to record errors that it detects when handling interrupts (see Fig ure 8-9). An APIC error interrupt is generated when the local APIC sets one of the error bits in the ESR.
8-20 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 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 regi ster . This initial writ e causes the ESR contents to be updated with the latest error status.
Vol. 3A 8-21 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) The time base for the timer is derived from the processor ’ s bus clock, divided by th e value spec- ified in the divide configuration regi st er . The timer can be configured thr ough the timer L VT entry for one-sho t or periodic operation.
8-22 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 8.5.5 Local Interrupt Accepta nce When a local interrupt is sent to the processo r core, it is subject to the acceptance criteria spec- ified in the interrupt acceptance flow chart in Figure 8-17.
Vol. 3A 8-23 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) The ICR consists of the following fields. V ector The vector number of the interrupt being sent. Delivery Mode Sp ecifies the type of IPI to be sent . This field is also know as the IPI message type field.
8-24 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) send a lowest priority IPI is model specific and should be avoided by B IOS and operating syst em software. 010 (SMI) Delivers an SMI interrupt to the target process or or processors. The vector field must be pro- grammed to 00H for future comp at ibility .
Vol. 3A 8-25 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) Destination Mode Selects either physical (0) or logical (1) destination mo de (see Section 8.
8-26 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) sors and to FFH for Pentium 4 and Intel Xeon pro- cessors. 1 1: (All Excluding Self) The IPI is sent to all pr ocessors in a system with the exception of the processor sending the IPI.
Vol. 3A 8-27 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) All Excluding Self V alid Edge Fixed, Lowest Priority 1 , 4 , NMI, INIT , SMI, Sta rt - Up X All Excluding Self Invalid 2 Level FIxe d, Lowest Priority 4 , NMI, INIT , SMI, Sta rt - Up X NOTES: 1.
8-28 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 8.6.2 Determining IPI Destination The destination of an IPI can be one, all, or a subset (group) of the processors on the system bus.
Vol. 3A 8-29 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) NOTE The number of local APICs that can be addressed on the system bus may be restricted by hardware.
8-30 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) The interpretation of MDA fo r the tw o models is described in the following paragraphs. 1. Flat Mod el — T hi s m o d el is s e l ec te d b y p r og ra m m i ng DF R b i t s 2 8 t h r ou gh 3 1 t o 1111 .
Vol. 3A 8-31 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 8.6.2.3 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 pr ocessors on the system bus and/ or back to itself (see Section 8.
8-32 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) Here, the TPR value is the task priority value in the TPR (see Figure 8-18), the IRR V value is the vector number for th e highest priori.
Vol. 3A 8-33 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) Section 8.10, “APIC Bus Message Passing Mechanism and Protocol (P6 Family , Pentium Processors),” describes the APIC bus arbitration prot ocols and bus message formats, while Section 8.
8-34 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 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 thei r priority and the curr ent task and processor priorities in the TPR and PPR (see Section 8.
Vol. 3A 8-35 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 1. (IPIs only) It examines the IPI message to det ermines if it is the specified destinat ion for the IPI as des cribed in Sectio n 8.
8-36 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 3. If the local APIC determines that it is the desig nated destination for the interrupt but the interrupt request is not one of the inte.
Vol. 3A 8-37 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 8.8.3.1 T ask and Processor Pri orities The local APIC also defines a task priority and a processor priority that it uses in determining the order in which interrupt s should be handled.
8-38 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 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 ISR V val ue is the vector number of the hi ghest priority ISR bit that i s set, or 00H if no ISR bit is set.
Vol. 3A 8-39 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) The IRR contains the active interrupt requests th at have been accepted, but not yet dispatched to the processor for servicing. Wh en the local APIC accepts an interr upt, it sets the bit in the IRR that corresponds the vector of the accepted interrupt .
8-40 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 8.8.5 Signaling Interrupt Servicing Completion For all interrupts except those deliv ered 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 8-21).
Vol. 3A 8-41 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) the TPR. The IC, however , is considered im plementation-dependent with th e under-lying priority mechanisms subject to change. The CR8, by contrast, is part of the Intel EM64T archi- tecture.
8-42 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) The vector number for the spurious -interrupt vector is specified in th e spurious-interrupt vector register (see Figure 8-23).
Vol. 3A 8-43 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) 8.10 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 mechan ism and protocol.
8-44 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) destination and message during device con figuration, allocating one or more non-shared messages to each MSI capab le function.” The capabilities mechanism provided by the PCI Local Bus Specification is used to identify and configure MSI capable PCI devices.
Vol. 3A 8-45 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) • When RH is 1 and the logical destination mode is active in a system using a flat addressing model, the Destination ID field mu st be set so that bits set to 1 identify processors that are present and enabled to receive the interrupt.
8-46 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) Reserved fields are not assumed to be any valu e. Software must preserve their contents on writes. Other fields in the Message Da ta Register are described below . 1. V ector — This 8-bit field contains the interrupt vector associated with the message.
Vol. 3A 8-47 ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC) d. 100B (NMI) — Deliver the signal to all the agents listed in th e destination field. The vector information is ignored. NMI is an edge triggered interrupt regardless of the T rigger Mode Setting.
8-48 Vol. 3A ADVANCED PROGRAMMABLE INTERRUPT CONTROLLER (APIC).
9 Pr ocessor Management and Initialization.
.
Vol. 3A 9-1 CHAPTER 9 PROCESSOR MANAGEMENT AND INITIALIZATION This chapter describes the facilities provi ded for managing processor wide functions and for initializing the processor .
9-2 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION The software-initialization code performs al l system-specific initia lization of the BSP or primary processo r an d the system logic.
Vol. 3A 9-3 PROCESSOR MANAGEMENT AND INITIALIZATION T able 9-1. IA-32 Processor St ates Following Power-up, Reset, or INIT Register Pentium 4 and Intel Xeon Processor P6 Family Processor Pentium Proce.
9-4 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION MXCSR P wr up or Reset: 1F80H INIT : Unchanged Pentium III processor only- Pwr up or Reset: 1F80H INIT : Unchanged NA GDTR, IDTR Base = 00000000H Li.
Vol. 3A 9-5 PROCESSOR MANAGEMENT AND INITIALIZATION 9.1.3 Model and Stepping Information Following a hardware reset, the EDX register contains component iden tification and revision information (see Figure 9-2).
9-6 Vol. 3A 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 byte s below the processor ’ s uppermost physical address.
Vol. 3A 9-7 PROCESSOR MANAGEMENT AND INITIALIZATION The EM flag determines w hether floating-po int instructions are executed by the x87 FPU (EM is cleared) or a device-not-availab le exception (#NM) is generated for all floating-po int instruc- tions so that an exception handler can em ulate the floati ng-point operation (EM = 1).
9-8 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION T o em ulate floating-point instructions, the EM, MP , and NE flag in control register CR0 should be set as shown in T able 9-3. Regardless of the value of the EM bit, the In tel486 SX processor generates a device-not-avail- able exception (#NM) up on encountering any floating-point instru ction.
Vol. 3A 9-9 PROCESSOR MANAGEMENT AND INITIALIZATION 9.4 MODEL-SPECIFIC REGISTERS (MSRS) The Pentium 4, Intel Xeon, P6 family , and Pentium processors contain a model-speci fic registers (MSRs).
9-10 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION 9.6 INITIALIZING SSE/SSE2/SSE3 EXTENSIONS For processors that contain SS E/SSE2/SSE3 extensions, steps must be taken when initializing the processor to allow execu tion of these instructions.
Vol. 3A 9-11 PROCESSOR MANAGEMENT AND INITIALIZATION 9.7.1 Real-Address Mode IDT In real-address mode, the only system data structur e that must be loaded into m emory is the IDT (also called the “interrupt vector table”). By default, the addres s of the base of the IDT is phys- ical address 0H.
9-12 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION • 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 th e necessary interrupt and exception handlers.
Vol. 3A 9-13 PROCESSOR MANAGEMENT AND INITIALIZATION 9.8.2 Initializing Protected-Mode Exceptions and Interrupt s Software init ialization code must at a minim u m load a protected-mode ID T with gate descriptor for each exception vector that the processor can generate.
9-14 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION After the processor has switched to protected m ode , the L TR instruction can be used to load a segment selector for a TSS descri ptor into the task register . This instruction marks the TSS descriptor as busy , but does not perform a task switch.
Vol. 3A 9-15 PROCESSOR MANAGEMENT AND INITIALIZATION 64-bit mode consistency checks fail in the following circumstances: • An attempt is made to enable or disable IA-32e m ode whi le paging is enabled. • IA-32e mode is enabled and an attempt is made to enable paging prior to enabl ing physical-address extensions (P AE).
9-16 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION Compatibility mod e execution is selected on a code-segm ent basis. This mode allows legacy applications to coex ist with 64-bit applications running in 64-bit mode.
Vol. 3A 9-17 PROCESSOR MANAGEMENT AND INITIALIZATION 9.9 MODE SWITCHING T o use the processor i n protected mode after hard ware or software reset, a mode switch must be performed from real-address mo de. Once in protected mo de, software generally does not need to return to real-address mode.
9-18 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION 6. Execute the L TR instruction to lo ad the task register with a segment selecto r 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.
Vol. 3A 9-19 PROCESSOR MANAGEMENT AND INITIALIZATION 4. Load segm ent regist ers SS, DS, ES, FS, and GS with a selector fo r a descri ptor contain ing the following values, which are a ppropriate for .
9-20 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION 9.10 INITIALIZATION AND MODE SWITCHING EXAMPLE This section provides an i nitialization and mode switching example that can be incorporated into an application.
Vol. 3A 9-21 PROCESSOR MANAGEMENT AND INITIALIZATION Figure 9-3. Processor State Af ter Reset T able 9-4. Main Initializat ion Step s in ST ARTUP .ASM Source Listing ST ARTUP .
9-22 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION 9.10.1 Assembler Usage In this example, the Intel assembler ASM386 and build to ols BLD386 are used to assemble and build the initialization code mod ule. The following assumptions are used when using the Intel ASM386 and BLD386 tools.
Vol. 3A 9-23 PROCESSOR MANAGEMENT AND INITIALIZATION 9.10.2 ST ARTUP .ASM Listing Example 9-1 provides high-level sample code designed to mov e the processor into protected mode. This listing does not include any opcode and offset information . Example 9-1.
9-24 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION 32 TSS_INDEX EQU 10 33 34 ; TSS_INDEX is the index of the TSS of the first task to 35 ; run after startup 36 37 38 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;.
Vol. 3A 9-25 PROCESSOR MANAGEMENT AND INITIALIZATION 79 LDT_reg DW ? 80 LDT_h DW ? 81 TRAP_reg DW ? 82 IO_map_base DW ? 83 TASK_STATE ENDS 84 85 ; basic structure of a descrip tor 86 DESC STRUC 87 lim.
9-26 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION 126 127 ; scratch areas for LGDT and LIDT instructions 128 TEMP_GDT_SCRATCH TABLE_REG <> 129 APP_GDT_RAM TABLE_REG <> 130 APP_IDT_RAM T.
Vol. 3A 9-27 PROCESSOR MANAGEMENT AND INITIALIZATION 175 MOV EBX,CR0 176 OR EBX,PE_BIT 177 MOV CR0,EBX 178 179 ; clear prefetch queue 180 JMP CLEAR_LABEL 181 CLEAR_LABEL: 182 183 ; make DS and ES address 4G o f linear memory 184 MOV CX,LINEAR_SEL 185 MOV DS,CX 186 MOV ES,CX 187 188 ; do board specific initiali zation 189 ; 190 ; 191 ; .
9-28 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION 222 MOV ECX, CS_BASE 223 ADD ECX, OFFSET (IDT_E PROM) 224 MOV ESI, [ECX].table_l inear 225 MOV EDI,EAX 226 MOVZX ECX, [ECX].table_l im 227 MOV APP_IDT_ram[EBX].t able_lim,CX 228 INC ECX 229 MOV APP_IDT_ram[EBX].
Vol. 3A 9-29 PROCESSOR MANAGEMENT AND INITIALIZATION 271 272 ;assume no LDT used in t he initial task - if necessary, 273 ;code to move the LDT cou ld be added, and should resemble 274 ;that used to m.
9-30 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION Figure 9-4. Constructin g T emporary GDT and Switching to Pro tected Mode (Lines 162-172 of List File) FFFF FFFFH Base=0, Limit=4G ST AR T : [CS.
Vol. 3A 9-31 PROCESSOR MANAGEMENT AND INITIALIZATION Figure 9-5. Moving the GDT , IDT , and TSS from ROM to RAM (Lines 196-261 of List File) FFFF FFFFH GDT RAM • Move the GDT , IDT , TSS • Fix Ali.
9-32 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION Figure 9-6. T a sk Switching (Lines 282-296 of List File) GDT RAM RAM_ST ART TSS RAM IDT RAM GDT Alias IDT Alias DS EIP EFLAGS CS SS 0 ES ESP • • • • • • SS = TS S.SS ESP = TSS.ESP PUSH TSS.
Vol. 3A 9-33 PROCESSOR MANAGEMENT AND INITIALIZATION 9.10.3 MAIN.ASM Source Code The file MAIN.ASM shown in Example 9-2 defines the data and stack segments for this appli- cation and can be substi tuted with the mai n module task wri tten in a high-lev el language that is invoked by the IRET instruction executed by ST AR TUP .
9-34 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION Example 9-4. Build F ile INIT_BLD_EXAMPLE; SEGMENT *SEGMENTS(DPL = 0) , startup.startup_code(BAS E = 0FFFF0000H) ; TASK BOOT_TASK(OBJECT = startup,.
Vol. 3A 9-35 PROCESSOR MANAGEMENT AND INITIALIZATION 9.1 1 MICROCODE UP DATE FACILITIES The Pentium 4, Intel X eon, and P6 family proces sors have the capability to correct errata by loading an Intel-supplied data blo ck into the pr ocessor . The data block is called a microcode update.
9-36 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION 9.1 1.1 Microcode Up date A microcode update consists of an Intel-su pplied binary that con tains a descriptive header and data. No executable code reside s within the update. Each micr ocode update is tailo red for a specific list of processor signat ures.
Vol. 3A 9-37 PROCESSOR MANAGEMENT AND INITIALIZATION . T able 9-6. Microcode Up date Field Defi nitions Field Name Offset (bytes) Length (bytes) Description Header V ersion 0 4 V ersion number of the update header.
9-38 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION T otal Size 32 4 S pecifies 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.
Vol. 3A 9-39 PROCESSOR MANAGEMENT AND INITIALIZATION Checksum[n] Data Size + 76 + (n * 12) 4 Used by utility software to decompose a microcode update into multiple microcode updates wher e each of the new updates is constructed without the optional Extended Processor Signature T able.
9-40 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION 9.1 1.2 Optional Extended Signature T able The extended signature table is a structure that may be appended to the end of the encrypted data when the encrypted data onl y supports a sing le processor signature (optional case).
Vol. 3A 9-41 PROCESSOR MANAGEMENT AND INITIALIZATION 9.1 1.3 Processor Identification Each microcode update is designed to for a sp eci fic processor or set of processors.
9-42 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION 9.1 1.4 Plat form Identification In addition to verifying t he processor signature, the intended processor platform type m ust be determined to p roperly target the microcode update. The int ended processor platform typ e is determined by reading the IA32 _PLA TFORM_ID register, (MSR 17H).
Vol. 3A 9-43 PROCESSOR MANAGEMENT AND INITIALIZATION Example 9-6. Pseudo Code Example of Processor Flags T est Flag ← 1 << IA32_PLATFORM_ID[52:50] If (Update.HeaderVersion == 00000001h) { If (Update.ProcessorFlags & Flag) { Load Update } Else { // // Assume the Data Size has been used to calculate the // location of Update.
9-44 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION Example 9-7. Pseudo Code Example of Checksum T est N ← 512 If (Update.DataSize != 00000000H) N ← Update.TotalSize / 4 ChkSum ← 0 For (I ← 0; I < N; I++) { ChkSum ← ChkSum + MicrocodeUpdate[I] } If (ChkSum == 00000000H) Success Else Fail 9.
Vol. 3A 9-45 PROCESSOR MANAGEMENT AND INITIALIZATION The loader shown in Example 9-8 assumes that update is the address of a microcode update (header and data) emb edded within the cod e segm ent of the BIOS. It also assumes that the processor is operating i n real mode.
9-46 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION 9.1 1.6.3 Update in a System Supporting Intel Hyper-Thre ading T echnology Intel Hyper-Threading T echnol ogy has implications on th e loading of the microcode u pdate. The update must be loaded for each core in a physical processor .
Vol. 3A 9-47 PROCESSOR MANAGEMENT AND INITIALIZATION 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 th e 64-bit model-sp ecific register at address 08BH (IA32_BIOS_SIGN_ID).
9-48 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION The IA32_BIOS_SIGN_ID register is used to report the m icrocode update signature when CPUID executes.
Vol. 3A 9-49 PROCESSOR MANAGEMENT AND INITIALIZATION 9.1 1.8 Pentium 4, Intel Xeon, and P6 Family Processor Microcode Up date Specifications This section describes the interface that an application can use to dynamically integrate processor- specific updates into th e system B IOS.
9-50 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION update blocks for each microcode upd ate. In a MP system, a common microcode update may be sufficient for each socket in the system. For IA-32 processors ear lier than fami ly 0FH and mo del 03H, the mi crocode update is 2 KBytes.
Vol. 3A 9-51 PROCESSOR MANAGEMENT AND INITIALIZATION { If ((Update.ProcessorSignature[N] == Processor Signature) && (Update.ProcessorFlags[N] & Platform Bits)) { Load Update.UpdateData into the Processor; Verify update was correctly loaded into the processor Go on to next processor Break; } N ← N + 1 } I ← I + (Update.
9-52 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION • The calling program should read an y update data th at already exists in the BIOS in order to make decisions about the appropriaten ess of loading the update. The BIOS must refuse to overwrite a newer update with an older versi on.
Vol. 3A 9-53 PROCESSOR MANAGEMENT AND INITIALIZATION For each processor { If ((this is a unique processor stepping) AND (we have a unique update in the database for this processor)) { Checksum the upd.
9-54 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION } // // Verify the update was loaded correctly // Issue the ReadUpdate function If an error occurred { Display Diagnostic exit } // // Compare the .
Vol. 3A 9-55 PROCESSOR MANAGEMENT AND INITIALIZATION 9.1 1.8.4 IN T 15H-based Interface Intel recommends that a BIOS interface be provided that allo ws additional microcode updates to be added to system flash. The INT15H inte rface is the Intel-defined method for doing this.
9-56 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION Descripti on In order to assure that the BIOS functio n is pr esent, the caller must verify the carry flag, the return code, and the 64-bit si gnature. The update count reflects the nu mber of 2048-byte blocks available for storage within one non-volatile RAM.
Vol. 3A 9-57 PROCESSOR MANAGEMENT AND INITIALIZATION Description The BIOS is responsible for select ing an appropr iate update block in the n on-volatile storage for storing the new update.
9-58 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION If no unused update block s are available and th e above criteria are not met, the BIOS can over- write update block(s) for a processor stepping that is no longer present in the system.
Vol. 3A 9-59 PROCESSOR MANAGEMENT AND INITIALIZATION Figure 9-8. Microcode Up date W rite Operation Flow [1] 1 V a lid U p d ate H eader V er sion? Loader R evis ion M atc h BI OS’s Loader ? D oes U.
9-60 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION Figure 9-9. Microco de Up date Write Opera t io n Flow [2] Ret ur n I NVALI D_RE VI SI ON Yes 1 Update Revis ion Newer Than NVRAM Update? Update Pa.
Vol. 3A 9-61 PROCESSOR MANAGEMENT AND INITIALIZATION 9.1 1 .8.7 Function 02H—Microcode Up date Control This function enables loadin g of binary updates into the processor . T able 9-15 lists the parame- ters and return codes for the f unction. Description This contr ol is provided on a global basis for all updates and processors.
9-62 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION The READ_F AILURE error code returned by this function has meaning only if the control func- tion is implemented in the BIO S NVRAM. The state of this feat ure (enabled/disabled) can also be implemented using CMOS RAM bits wh ere READ failure er rors cannot occur .
Vol. 3A 9-63 PROCESSOR MANAGEMENT AND INITIALIZATION Description The read function enables the caller to read any mic rocode update data that already exists in a BIOS and make decisi ons about the addition of new updates.
9-64 Vol. 3A PROCESSOR MANAGEMENT AND INITIALIZATION UPDA TE_NUM_INV ALID 99H The update number exceeds the maximum numb er of update blocks implemented by the BIOS. NOT_EMPTY 9AH The specified update block is a subseque nt block in use to store a valid microcode update t hat spans multiple blocks.
10 Memory Cache Contr ol.
.
Vol. 3A 10-1 CHAPTER 10 MEMORY CACHE CONTROL This chapter describes the IA-32 architecture’ s memory cache and ca che control mechanisms, the TLBs, and the store buf fer . It also describes th e me m ory ty pe rang e registers (MT RRs) fou nd in the P6 family processors and how they are used to control caching of physica l memory locations.
10-2 Vol. 3A MEMORY CACHE CONTROL T able 10-1. Characteristics of the Caches, TLBs, Store Buffe r, and Write Combining Buffer in IA-32 Processors Cache or Buffer Characte ristics T race Cache 1 - Pentium 4 and Intel Xeon processors: 12 K μ ops, 8-way set associative.
Vol. 3A 10-3 MEMORY CACHE CONTROL The IA-32 processors 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 10-1).
10-4 Vol. 3A MEMORY CACHE CONTROL The trace cache in the Pentium 4 and Intel Xeon pr ocessors is an integral part of the Intel NetBurst microarchitecture and is available in all execution modes: protected mode, sys tem management mode (SMM), and real-address mode.
Vol. 3A 10-5 MEMORY CACHE CONTROL When the processor attempts to write an opera nd to a cacheable area of memory , it first checks if a cache line for that memory location exists in the cache.
10-6 Vol. 3A MEMORY CACHE CONTROL NOTE The behavior of FP and SSE/SSE2 operations on operands in UC memory is implementation dependent. In so me implementations, accesses to UC memory may occur more than once.
Vol. 3A 10-7 MEMORY CACHE CONTROL memory . When writing through to memory , in valid cache lines are never filled, and valid cache lines are either filled or invalidated.
10-8 Vol. 3A MEMORY CACHE CONTROL 10.3.1 Buffering of Write Combining Memory Locations W rites to the WC memory type are not cached in the typical sense of th e word cached. They are retained in an inter nal write combi ning buf fer (WC bu f fer) that is sep arate from the in ternal L1, L2, and L3 caches and the store buf fer .
Vol. 3A 10-9 MEMORY CACHE CONTROL The only elements of WC propagation to the syst em bus that are guaranteed are those provided by transaction atomicity . Fo r example, with a P6 family processor , a completely full WC buffer will always be propagated as a single 32-bit bur st transaction using any chunk order .
10-10 Vol. 3A MEMORY CACHE CONTROL For a description of th ese instructions and there intended use, see Section 10.5.5, “Cache Management I nstructions.” 10.4 CACHE CONTROL PROTOCOL The following section describes th e cache control protocol curren tly defined for the I A-32 archi- tecture.
Vol. 3A 10-11 MEMORY CACHE CONTROL • Cache control and memory ordering instructions — The IA-32 architecture provides several instructions that control the caching of data, the ordering of memory reads and writes, and the prefetching of data.
10-12 Vol. 3A MEMORY CACHE CONTROL Figure 10-2. Cache-Control R egisters an d Bit s Available in IA-32 Processors Page-Directory or Page-T able Entry TLBs MTRRs 3 Physical Memory 0 FFFFFFFFH 2 control.
Vol. 3A 10-13 MEMORY CACHE CONTROL T a ble 10-5. Cach e Operating Modes CD NW Caching and Read/Write Policy L 1 L2/L3 1 0 0 Normal Cache Mode. Highes t performance cache operation. - Read hits access the cache; read misses may cause replacement. - Write hit s update the cache.
10-14 Vol. 3A MEMORY CACHE CONTROL • NW flag, bit 29 of control register CR0 — Controls th e writ e policy fo r system m emory locations (see Section 2.
Vol. 3A 10-15 MEMORY CACHE CONTROL • Memory type range r egisters (MTRRs) (i ntroduced in P6 family pr ocessors) — Control the type of cachin g used in specific regions o f physical memory . Any of the caching types described in Section 10.3, “Methods of Caching A vailable,” can be selected.
10-16 Vol. 3A MEMORY CACHE CONTROL 10.5.2.1 Selecting Memory T ypes for Pentium Pro an d Pentium II Processors The Pentium Pro and Pentium II processors do not support the P A T . Here, the effective memory type for a page is selected with the MTRRs and the PCD and PWT bits in the page-t able or page- directory entry for the page.
Vol. 3A 10-17 MEMORY CACHE CONTROL 4. Setting th e PCD and PWT flags to opposite valu es is considered model-specific for the WP and WC memory types and architecturally -defined for the WB, WT , and UC memory types.
10-18 Vol. 3A MEMORY CACHE CONTROL 10.5.2.3 Writing V alues Acro ss Pag es with Differ ent Memory T ypes If two adjoining pages in m emory have different memory types, and a word o r longer operand is written to a mem ory location that crosses the page boundary between tho se two pages, the operand might be w ritten to memory twice.
Vol. 3A 10-19 MEMORY CACHE CONTROL 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 discuss ion of the TYPE field and the E flag in Section 10.11.2.1, “IA32_MTRR_DEF_TYPE MSR”).
10-20 Vol. 3A MEMORY CACHE CONTROL modified lines (such as, d uring testing or fa ult recovery where cache coherency with main memory is not a concern), software should use the WBINVD instruction. The WBINVD instruction first wr ites back any modified lines in all the internal caches, then invalidates the contents of both the L1, L2, and L3 caches.
Vol. 3A 10-21 MEMORY CACHE CONTROL 10.5.6.1 Adaptive Mode Adaptive mode facilitates L1 data cache sharin g between logical processors. When running in adaptive mode, the L1 data cache is shared acr oss logical processors in the same core if: • CR3 control registers for logical processors sharing the cache are identical.
10-22 Vol. 3A MEMORY CACHE CONTROL 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 t he instruc- tion could be the one executed.
Vol. 3A 10-23 MEMORY CACHE CONTROL cache hierarchy now or as soon as possible, in an ticipation of its use. Th e instructions provide different variations of the hint th at allow selection of the cache leve l into which data will be read.
10-24 Vol. 3A MEMORY CACHE CONTROL 10.10 STORE BUFFER IA-32 processors temporarily st ore each write (store) to memory in a store buffer . The store buffer improves processor perf ormance by allow ing the processor to continue ex ecuting instruc- tions without having to wait until a write to memory and/or to a cache is complete.
Vol. 3A 10-25 MEMORY CACHE CONTROL ization software should then se t the MTRRs to a specific, syst em-defined memory map. T ypi- cally , the BIOS (basic input/output system) so ftware configures the MTRRs. The operating system or executive is then fr ee to modify the memory map us ing the normal page-level cache- ability attributes.
10-26 Vol. 3A MEMORY CACHE CONTROL 10.1 1.1 MTRR Feature Identification The availability of the MTRR feature is model- specific. Software can dete rmine if MTRRs ar e supported on a processor by executing the CPUID instruction and reading the state of the MTRR flag (bit 12) in the feature information register (E DX).
Vol. 3A 10-27 MEMORY CACHE CONTROL • WC (write combining) fla g, bit 10 — The write-combining (WC) memory type is supported when set; t he WC type is not sup ported when clear .
10-28 Vol. 3A MEMORY CACHE CONTROL • 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 MTR Rs when overlaps in ranges occur .
Vol. 3A 10-29 MEMORY CACHE CONTROL For the P6 family processors, the prefix for the fixed range MTRRs is MTRRfix. 10.1 1.2.3 V ariable R ange MTR Rs The Pentium 4, Intel Xeon, an d P6 family processors permit software to specify th e memory type for eight variable-size address ranges, using a pair of MTRRs for each range.
10-30 Vol. 3A MEMORY CACHE CONTROL • PhysBase field, bits 12 through (MAXPHY ADDR-1) — Specifies the base address of the address range. This 24-bit value, in the case where MAXPHY ADDR is 36 bits, is extended by 12 bits at the low end to form the base address (this au toma tically aligns the address on a 4-KByte boundary).
Vol. 3A 10-31 MEMORY CACHE CONTROL All other bits in the IA32_MTRR _PHYSBASE n and IA32_MTRR_PHYSMASK n registers are reserved; the processor generates a general-prot ection excepti on (#GP) if software at tempts to write to them. Some mask values can result in ranges that ar e not continuous.
10-32 Vol. 3A MEMORY CACHE CONTROL 10.1 1.3 Example Base a nd Mask Calculations The examples in this section apply to processo rs that support a maximu m physical address size of 36 bits. The base and m ask values entered in variable-range MTRR pairs are 24-b it values that the processor extends to 36-bits.
Vol. 3A 10-33 MEMORY CACHE CONTROL The following settings for the MTRRs will yield the proper mappin g of the physical address space for this syst em configuration. IA32_MTRR_PHYSBASE0 = 0000 0000 0000 0006H IA32_MTRR_PHYSMASK0 = 0000 000F FC00 0800H Caches 0- 64 MByte as WB c ache type.
10-34 Vol. 3A MEMORY CACHE CONTROL Caches 96-100 MByte as WB cache type. IA32_MTRR_PHYSBASE3 = 0000 0000 0400 0000H IA32_MTRR_PHYSMASK3 = 000 0 00FF FFC0 0800H Caches 64-68 MByte as U C cache type. IA32_MTRR_PHYSBASE4 = 0000 0000 00F0 0000H IA32_MTRR_PHYSMASK4 = 0000 00FF FFF0 0800H Caches 15-16 MByte as U C cache type.
Vol. 3A 10-35 MEMORY CACHE CONTROL d. If two or more variabl e memory ranges match and the memory types are WT and WB, the WT memory type is used. e. For overlaps not defined by the above rules, processor behavior is undefined. 3. If no fixed or variable memory range matche s, the processor uses th e default memory ty pe.
10-36 Vol. 3A MEMORY CACHE CONTROL 10.1 1.7 MTRR Maintenan ce Programming Interface The operating system maintains th e MTRRs after booting and sets up or changes t he memory types for memory-mapped devices. The operating system should provide a driver and applica- tion programming interface (API) to access and set the MTRRs.
Vol. 3A 10-37 MEMORY CACHE CONTROL The pseudocode for the Get4KMem T ype() fun ction in Example 10-1 7 obtains the mem ory type for a single 4-KByte range at a given physical a ddress.
10-38 Vol. 3A MEMORY CACHE CONTROL FI; IF IA32_MTRRCAP.FIX is set AND range can be mapped using a fixed-rang e MTRR THEN pre_mtrr_change(); update affected MTRR; post_mtrr_change(); FI; ELSE (* try to map using a variable MTRR pair *) IF IA32_MTRRCAP.
Vol. 3A 10-39 MEMORY CACHE CONTROL The physical address to variab le range mapping algorithm in the MemT ypeSet function detects conflicts with current variable range regi sters by cycling through them and determining whether the physical address in quest ion matches any of the current ranges.
10-40 Vol. 3A MEMORY CACHE CONTROL 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 regi ster 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.
Vol. 3A 10-41 MEMORY CACHE CONTROL The Pentium 4, Intel Xeon, and P6 family processors provide special support for the physical memory range from 0 to 4 MBytes, which is po tentially mapped by both the fixed and v ari- able MTRRs.
10-42 Vol. 3A MEMORY CACHE CONTROL 10.12.2 IA32_CR_P A T MSR The IA32_CR_P A T MSR is located at MSR addre ss 277H (see to App endix B, “Model-Specific Registers (MSRs),” and this add ress will remain at the sam e address on future IA-32 processors that support the P A T feature.
Vol. 3A 10-43 MEMORY CACHE CONTROL 10.12.3 Selecting a Memory T ype from the P A T T o select a memory type fo r a page from the P A T , a 3-bit index made up of the P A T , PCD, and PWT bits must be encoded in the page-table or page-directory entry for the page.
10-44 Vol. 3A MEMORY CACHE CONTROL The values in all the entries of the P A T can be changed by writing to the IA32_CR_P A T MSR using the WRMSR instruction. The IA32_CR_P A T MSR is read and write accessible (use of the RDMSR and WRMSR instructions, respectively) to so ftware operating at a CPL of 0.
Vol. 3A 10-45 MEMORY CACHE CONTROL 10.12.5 P A T Compatibility wi th Earlier IA -32 Processors For IA-32 processors that supp ort the P A T , th e IA32_CR_P A T MSR is always active.
10-46 Vol. 3A MEMORY CACHE CONTROL.
11 Intel ® MMX ™ T echnology System Pr ogramming.
.
Vol. 3A 11-1 CHAPTER 1 1 INTEL ® MMX ™ T ECHNOLOGY SYSTEM PROGRAMMING This chapter describes those features of the Intel ® MMX™ technology that must be considered when designing or enhancin g an operating syst em to support MMX technology .
11-2 Vol. 3A INTEL® MMX™ T ECHNOLOGY SYSTEM PROGRAM MING When a value is written into an MMX register us ing an MMX instru ction, the value also appears in the corresponding floating-point register in b its 0 through 63 .
Vol. 3A 11-3 INTEL® MMX™ T E CHNOLOGY SYSTEM PROGRAMMING Execution of MMX instru ctions does not affect the other bits in the x87 FPU status word (bi ts 0 through 10 and bits 14 and 15) or the cont.
11-4 Vol. 3A INTEL® MMX™ T ECHNOLOGY SYSTEM PROGRAM MING 1 1 .3 S AVING AND RESTOR ING THE MMX ST ATE AND REGISTERS Because the MMX registers are aliased to the x87 FPU data registers, the MMX stat.
Vol. 3A 11-5 INTEL® MMX™ T E CHNOLOGY SYSTEM PROGRAMMING NOTE The IA-32 architecture does not support scann ing the x87 FPU tag word and then only saving valid entries. 1 1.4 SAVING MMX ST ATE ON T ASK OR CONTEXT SWITCHES When switching from one task or context to another , it is often necessary to save the MMX state.
11-6 Vol. 3A INTEL® MMX™ T ECHNOLOGY SYSTEM PROGRAM MING • Other exceptions can occur indi rectly due to the faulty ex ecution of the exception hand lers for the above exceptions.
Vol. 3A 11-7 INTEL® MMX™ T E CHNOLOGY SYSTEM PROGRAMMING Figure 1 1-2. Mapping of MMX Registe rs to x87 FPU Dat a Register St ack MM0 MM1 MM2 MM3 MM4 MM5 MM6 MM7 ST1 ST2 ST7 ST0 ST6 ST7 ST1 TOS TOS.
11-8 Vol. 3A INTEL® MMX™ T ECHNOLOGY SYSTEM PROGRAM MING.
12 SSE, SSE2 and SSE3 System Pr ogramming.
.
Vol. 3A 12-1 CHAPTER 12 SSE, SSE2 AND SSE3 SYSTEM PROGRAMMING This chapter describes features of the streaming SIMD exte nsions (SSE), streaming SIMD extensions 2 (SSE2) and streaming SIMD extens ions 3 (SSE3) that must be considered when designing or enhancing an operating system to supp ort the Pentium II I , Pentium 4, and Intel Xeon processors.
12-2 Vol. 3A SSE, SSE2 AND SSE3 SYSTEM PROGRAMMING 12.1.2 Checking for SSE/SSE2/SSE3 Extension Support If the processor attempts to execute an uns upported SSE/SSE2/SSE3 instruction , the processor will generate an invalid-op code exception (#UD).
Vol. 3A 12-3 SSE, SSE2 AND SSE3 SYSTEM PROGRAMMING NOTE The OSFXSR and OSXMMEXCP T bits in control register CR4 must be set by the operating system. The processor h as no other way of detecting operating-system support for the FXSA VE and FXRSTOR instructions or for handling SIMD floating-point except ions.
12-4 Vol. 3A SSE, SSE2 AND SSE3 SYSTEM PROGRAMMING The SIMD floating-p oint exception mask bits (bits 7 through 12), the flush-to-zero flag (bit 15), the denormals-are-zero flag (b it 6), and the roundi ng control fiel d (bits 13 and 14) in the MXCSR register should be left in their default va lues of 0.
Vol. 3A 12-5 SSE, SSE2 AND SSE3 SYSTEM PROGRAMMING • System Exceptions: — Inval id-opcode exception (#UD). This exception is generated when executing SSE/SSE2/SSE3 instructions under the following conditions: • SSE/SSE2/SSE3 feature flags returned by CPUID are set to 0.
12-6 Vol. 3A SSE, SSE2 AND SSE3 SYSTEM PROGRAMMING same conditions th at cause x87 FPU float ing-point error exceptio ns (#MF) to be generated for x87 FPU instruction s. Each of these exceptions can be masked, in which case the processor returns a reasona ble result to the destinat ion operand without i nvoking an exception handler .
Vol. 3A 12-7 SSE, SSE2 AND SSE3 SYSTEM PROGRAMMING In some cases, applications can only save the XMM and MXCSR registers in the following way: • Execute eight MOVDQ instructions to save the contents of the XMM0 through XMM7 registers to memory . • Execute a STMXCSR instr ucti on to save the state of the MXCSR register to memory .
12-8 Vol. 3A SSE, SSE2 AND SSE3 SYSTEM PROGRAMMING • The operating system can take the respo nsibility for automatically saving th e x87 FPU, MMX, XXM, and MXCSR registers as part of the task switch.
Vol. 3A 12-9 SSE, SSE2 AND SSE3 SYSTEM PROGRAMMING On a task switch, the operatin g system task switching code must execute the fol lowing pseudo- code to set the TS flag according to the cu rrent owner of the x8 7 FPU/MMX/SSE/SSE2/SSE3 state.
12-10 Vol. 3A SSE, SSE2 AND SSE3 SYSTEM PROGRAMMING • Restores the x87 FPU, MMX, XMM, or MXCSR registers from the new task’ s save area for the x87 FPU/MMX/SSE/SSE2/SSE3 state. • Updates the current x87 FPU/MMX/SSE/SSE2/SSE3 state owner to be the curren t task.
13 Power and Thermal Management.
.
Vol. 3A 13-1 CHAPTER 13 POWER AND THERMAL MANAGEMENT This chapter describes facilities of IA-32 arch itecture used for power management and thermal monitoring.
13-2 Vol. 3A POWER AND THERMAL MANAGEMENT 13.2 P-ST ATE HARDWARE COORDINATION The Advanced Configuration and Power Interface (ACPI) defines performance states (P-state) that are used facilitate syst em software’ s ability to manage processor power consum ption.
Vol. 3A 13-3 POWER AND THERMAL MANAGEMENT If P-states are exposed by the BI OS as hardware coordinated, so ftware is expected to confirm processor suppo rt for P-state hardw are coordina tion feedback and use the feedback mechanism to make P-state decisions.
13-4 Vol. 3A POWER AND THERMAL MANAGEMENT 13.3 MW AIT EXTENSIONS FOR ADVANCED POWER MANAGEMENT IA-32 processors may support a number of C-state 1 that reduce power co nsumption f or inacti ve states.
Vol. 3A 13-5 POWER AND THERMAL MANAGEMENT 13.4 THERMAL MONITORI NG AND PROTECTION The IA-32 architecture provides the follow ing mechanisms for monito ring temperature and controlling thermal po wer: 1. The catast rophic shutdown det ecto r forces processor execution to stop if the processor ’ s core temperature rises above a preset limi t.
13-6 Vol. 3A POWER AND THERMAL MANAGEMENT 13.4.1 Catastrophic Shut down Detector P6 family pr ocessors introduce d a thermal sens or that acts as a catastroph ic shutdown detector . This catastrophic shutdown detector was also i mplemented in Pentium 4, Intel Xeon and Pentium M processors.
Vol. 3A 13-7 POWER AND THERMAL MANAGEMENT MSR_THERM2_CTL register is set to 1 (Fi gure 13-3) 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.
13-8 Vol. 3A POWER AND THERMAL MANAGEMENT • 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 comm ence after the TCC has disengaged.
Vol. 3A 13-9 POWER AND THERMAL MANAGEMENT • High-T emperature Interru pt Enable flag, bit 0 — Enables an i nterrupt to be generated on the transition from a low -temperature to a high-temperature when set; disables the interrupt when clear .(R/W).
13-10 Vol. 3A POWER AND THERMAL MANAGEMENT The IA32_CLOCK_MODULA TION MSR contains the following flag and field used to enable software-controlled clock modulation and to select th e clock modulation .
Vol. 3A 13-11 POWER AND THERMAL MANAGEMENT 13.4.4 Detection of T hermal Moni tor and Sof tware Controlled Clock Modulation Facilities The ACPI flag (bit 22) of the CPUID f eature flags indicates the presence of the IA32_THERM_ST A TUS, IA32_THERM_IN TERRUP T , IA32_CLOCK_MODULA TION MSRs, and the xAPIC thermal L VT entry .
13-12 Vol. 3A POWER AND THERMAL MANAGEMENT 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 FO RCEPR# Event (bit 2, RO) — Indicates whet her PROCHOT# or FORCEPR# is b eing asserted.
Vol. 3A 13-13 POWER AND THERMAL MANAGEMENT • Thermal Threshold #2 Log (bit 9, R/WC0) — Sticky bit that i ndicates whether the Thermal Threshold #2 has been reached since th e last clearing of this bit or a reset. If bit 9 = 1, the Thermal Threshold #2 has been reached.
13-14 Vol. 3A POWER AND THERMAL MANAGEMENT • THERMTRIP# Interrupt Enable (bit 2, R/W) — When a catastroph ic cooling failure occurs, the processor will automatically shut down.
14 Machine Check Ar chitectur e.
.
Vol. 3A 14-1 CHAPTER 14 MACHINE-CHECK ARCHITECTURE This chapter describes the m achine-check architecture and ma chine-check exception mecha- nism found in the Pentiu m 4, Intel Xeon, and P6 family processors. See Chapter 5, “Interrupt 18—Machine-Check Exception (#MC),” for more information on machine- check exceptions.
14-2 Vol. 3A MACHINE-CHECK ARCHITECTURE 14.3 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 14-1).
Vol. 3A 14-3 MACHINE-CHECK ARCHITECTURE Where: • Count field, bi ts 0 through 7 — Indicates the number of ha rdware unit er ror-reporting banks available in a particul ar processor implementation.
14-4 Vol. 3A MACHINE-CHECK ARCHITECTURE Where: • Count field, bits 0 thr ough 7 — Indicat es the number of hardware unit error-reporting banks available in a particular processor im plem entation. • MCG_CTL_P (register pr esent) flag, bit 8 — Indicates that the MCG_CTL register is present when set and absent when clear .
Vol. 3A 14-5 MACHINE-CHECK ARCHITECTURE 14.3.1.4 IA32_MCG_CTL MSR The IA32_MCG_CTL MSR (called the MCG_CTL MS R in P6 fami ly processors) is present if the capability flag MCG_CTL_P is set in the IA32_ MCG_CAP MSR (or the MCG_CAP MSR). IA32_MCG_CTL (or MCG_CTL) controls the rep orting of machine-check exceptions.
14-6 Vol. 3A MACHINE-CHECK ARCHITECTURE 14.3.2.2 IA32_MCi_ST A TUS MSRs Each IA32_MC i _ST A TUS MSR (called MC i _ST A TUS in P6 family processors) contains in for- mation related to a machine-check error if its V AL (valid) f lag is set (see Figure 14-6).
Vol. 3A 14-7 MACHINE-CHECK ARCHITECTURE where the error occurred . Do not read these registers if they are not impl emented in the processor . • MISCV (IA32_MC i _MISC register valid) flag, bit 59 — Indicates (when set) that the IA32_MC i _MISC register contains additional inform ation regarding the error .
14-8 Vol. 3A MACHINE-CHECK ARCHITECTURE 14.3.2.4 IA32_MCi_MISC MSRs The IA32_MC i _MISC MSR (called the MC i _MISC MSR in the P6 family processors) contains additional information describing the machin e-check error if the MISCV flag in the IA32_MC i _ST A TUS register i s set.
Vol. 3A 14-9 MACHINE-CHECK ARCHITECTURE In processors with support for Intel EM64T , 64-bit machine check state MSRs are aliased to the legacy MSRs. In addition, there m ay be registers beyond IA32_MC G_MISC. These may include up to five reserved MSRs (IA32_MCG _RESER VED[1:5]) and save-st ate MSRs for registers introduced in 64-bit mo de.
14-10 Vol. 3A MACHINE-CHECK ARCHITECTURE When a machine-check error is detected on a Pe ntium 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 .
Vol. 3A 14-11 MACHINE-CHECK ARCHITECTURE 14.3.3 Mapping of the Pentium Processor Machine-Check Errors to the Machine-Check Architecture The Pentium processo r reports machine-check errors using tw o registers: P5_MC_TYPE and P5_MC_ADDR.
14-12 Vol. 3A MACHINE-CHECK ARCHITECTURE Example 14-19. Machine-Check Initializa tio n Pseudocode Check CPUID Feature Flags for MCE and MCA support IF CPU supports MCE THEN IF CPU supports MCA THEN IF (IA32_MCG_CAP.
Vol. 3A 14-13 MACHINE-CHECK ARCHITECTURE FOR error-reporting ba nks (0 through MAX_BANK_NUMBER) DO (Optional for BIOS and OS) Log valid errors (OS only) IA32_MCi_STATUS ← 0; OD FI FI FI Setup the Machine Check Exception (#MC) handl er for vector 18 in IDT Set the MCE bit (bit 6) in CR4 register to enable Machine-Chec k Exceptio ns FI 14.
14-14 Vol. 3A MACHINE-CHECK ARCHITECTURE 14.6.2 Compound Error Codes Compound error codes describe errors related to the TLBs, memory , c aches, bus and intercon- nect logic, and internal timer . A set of sub-fi elds is common to all of compound errors.
Vol. 3A 14-15 MACHINE-CHECK ARCHITECTURE For example, the error code ICACHEL1_R D_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. The 2-bit TT sub-field (T able 14 -5) indicates the type of transaction (dat a, instruction, or generic).
14-16 Vol. 3A MACHINE-CHECK ARCHITECTURE The 4-bit RRRR sub-field (see T able 14-7) indicates th e type of action asso ciated with the error . Actions include read and write operations, pr efetches, cache evictions, and snoops. Generic error is returned when the type of error canno t be determin ed.
Vol. 3A 14-17 MACHINE-CHECK ARCHITECTURE 14.6.3 Machine-Check Erro r Codes Interpretation Appendix E, “Inter preting Machine-Check Error Cod es,” provides information on interpretin g the MCA error code, model-specific error code, and other information error code fields.
14-18 Vol. 3A MACHINE-CHECK ARCHITECTURE 14.7.1 Machine-Check Exception Handler The machine-check exception (#MC) corresp onds to vector 18. T o serv ice machine-check exceptions, a trap gate must be added to th e IDT . The pointer in the trap gate must point to a machine-check exceptio n handler .
Vol. 3A 14-19 MACHINE-CHECK ARCHITECTURE • The MCIP flag in the IA32_MCG_ST A TUS re gister indicates whether a machine-check exception was ge nerated. Before retu rning from the machine-ch eck exception handler , software should clear this flag so that it can be used reliably by an error logging utility .
14-20 Vol. 3A MACHINE-CHECK ARCHITECTURE 14.7.3 Pentium Processor Machine-Check Exception Handling T o mak e the machine-check exception handler portable to th e Pentium 4, Intel Xeon, P6 family , and Pentium processors, checks can be made (usi ng CPUID) to determine the processor type.
Vol. 3A 14-21 MACHINE-CHECK ARCHITECTURE AND RIPV flag in IA32_MCG_STATUS = 0 (* execution is not restartable *) THEN RESTARTABILITY = FALSE; return RESTARTABILITY to calli ng proced ure; FI; Save time-stamp counter and processor ID; Set IA32_MC i _STATUS to all 0 s; Execute serializing instruction (i.
14-22 Vol. 3A MACHINE-CHECK ARCHITECTURE The basic algorithm given in Example 14-2 1 can be modi fied to provi de more ro bust recovery techniques. For example, software has the flexibility to attempt recovery using information unavailable to the hardware.
15 8086 Emulation.
.
Vol. 3A 15-1 CHAPTER 15 8086 EMULATION IA-32 processors (begin ning with th e Intel386 processor) prov ide two wa ys to execute new or legacy programs that are assembled and/or compiled to run on an Intel 80 86 processor: • Real-address mode. • V irtual-8086 mode.
15-2 Vol. 3A 8086 EMULATION The following is a summary of the core features of the real-address mo de execution environment as would be seen by a program written fo r the 8086: • The processor supports a nomin al 1-MByte physical address space (see Section 15.
Vol. 3A 15-3 8086 EMULATION 8-byte entries) u sed when handling pro tected-mode interrupts and excep tions. Interrupt and exception ve ctor numbers pr ovide an inde x to entries in th e interrupt table. Each entry provides a pointer (called a “vector”) to an interrupt - or exception-han dling procedure.
15-4 Vol. 3A 8086 EMULATION behavior of the 8086 processor .) Care should be take to en sure that A20M# based address wrap - ping is handled correctly in multipro cessor based system.
Vol. 3A 15-5 8086 EMULATION • Logical instructions AND, OR, XOR, and NOT . • Decimal instructions DAA, D AS, AAA, AAS, AAM, an d AAD. • Stack instructions PUSH and POP (to g eneral-purpose registers and segment registers). • T ype conversion in structions CWD, CDQ, CBW , and CWDE.
15-6 Vol. 3A 8086 EMULATION • ENTER and LEA VE control instructions. • BOUND instruction. • CPU identification (CPUID) instr uction. • System instructions CL TS, INVD, WI NVD, INVLPG , LGDT , SGD T , LIDT , SIDT , LMSW , SMSW , RDMSR, WRMSR, RDTSC, and RDPMC.
Vol. 3A 15-7 8086 EMULATION (For backward compat ibility to Intel 808 6 proce ssors, the default base address and limit of the interrupt vector table shoul d not be chang e d.) T able 15-1 shows the interrupt and exception vector s that can be generated in real-address mode and virtual-8086 mode, and in the Intel 8086 pro cesso r .
15-8 Vol. 3A 8086 EMULATION T able 15-1. Real-Addre ss Mode Exceptions and Interrupt s V ector No. Desc ription Real-Address Mode Virtual-8086 Mode Intel 8086 Processor 0 Divide Error (#DE) Y es Y es .
Vol. 3A 15-9 8086 EMULATION 15.2.1 Enabling Virtual-8086 Mode The processor runs in virtual-8086 mode when the VM (virtual machin e) flag in the EFLAGS register is set. This flag can only be set wh en the processor switches to a new protected-mode task or resumes virtual-8086 mode via an IRET instruction.
15-10 Vol. 3A 8086 EMULATION The 8086 operating-system servi ces consists of a kernel and/or operating-system procedures that the 8086 pr ogram makes calls to. These serv ices can be implemented in either of the following two ways: • They can be included in the 8086 program.
Vol. 3A 15-11 8086 EMULATION • 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 me mo ry -mapped I/O devices. 15.
15-12 Vol. 3A 8086 EMULATION Figure 15-3. Entering and Lea ving Virtual-8086 Mode Monitor Virtual-8086 Real Mode Code Protected- Mode T asks Virtual-8086 Mode T asks (8086 Programs) Protected- Mode In.
Vol. 3A 15-13 8086 EMULATION 15.2.6 Leaving Virtual-8086 Mode The processor can leave the virtu al-8086 mode only through an interrupt or exception . The following are situations where an interrupt or.
15-14 Vol. 3A 8086 EMULATION 15.2.7 Sensitive Instructions When an IA-32 processor is running in virtua l-808 6 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.
Vol. 3A 15-15 8086 EMULATION 15.2.8.2 Memory-Mapped I/O In systems which use memory-map ped I/O, the paging facilities o f the processor can be used to generate exceptions for attempts to access I/O ports.
15-16 Vol. 3A 8086 EMULATION The method the proc essor uses to handle class 2 and 3 i nterrupts depends on the setting of the following flags and fields: • IOPL field (bits 12 and 13 in the EFLAGS register) — Contr ols how class 3 softw are interrupts are handled when the processor is in virtual-808 6 mode (see Section 2.
Vol. 3A 15-17 8086 EMULATION 15.3.1 Class 1—Hardware Inte rrupt and Exception Handlin g in Virtual-8086 Mode In virtual-8086 mode , the Pentium, P6 family , Pentium 4, and Intel Xeon processors handle hardware interrupts and exceptions in the same manner as they are handled by the Intel486 and Intel386 processors.
15-18 Vol. 3A 8086 EMULATION Interrupt and exception handlers can examine the VM flag on the stack to determine if the inter- rupted proc edure was running in vi rtual-8086 mode.
Vol. 3A 15-19 8086 EMULATION The virtual-8086 monitor runs at privilege level 0, like the pro tected-mode interrupt and excep- tion handlers. It is common ly closely tied to the protected-mode gene ral-protection exception (#GP , vector 13) handler .
15-20 Vol. 3A 8086 EMULATION 15.3.1.3 Handling an Interrupt or Exception Through a T ask Gate When an interrupt or exception vector poi nts to a task gate in the IDT , the processor performs a task switch to the selected in terrupt- or exception-handl ing task.
Vol. 3A 15-21 8086 EMULATION available or not enabled, maskable hardware interrupts are handled as class 1 interrupts. Here, if VIF and VIP flag s are needed, the virtual-80 86 monitor can implement them in software.
15-22 Vol. 3A 8086 EMULATION 3. The virtual-808 6 monitor shoul d read the VIF flag in the EFLAGS register . — If the VIF flag is clear, the virtual-8086 monit or 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 .
Vol. 3A 15-23 8086 EMULATION 15.3.3 Class 3—Software Interrupt Handling in V irtual-8086 Mode When the processor receives a software inte rrupt (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.
15-24 Vol. 3A 8086 EMULATION T abl e 15-2. Software Interrupt Handling Methods While in Virtual-8086 Mode Method VME IOPL Bit in Redir . Bitmap* Processor Action 10 3 X Interrupt directed to a protect.
Vol. 3A 15-25 8086 EMULATION Redirecting software interrupts back to th e 8086 program potentially speeds up in terrupt handling because a switch back and forth between virtual-8086 mode and protected mode is not required.
15-26 Vol. 3A 8086 EMULATION 15.3.3.2 Methods 2 and 3: Sof tware Interrupt Handling When a software interrupt occurs in vi rtual-8086 mode and the metho d 2 or 3 conditions are present, the processor generates a general-pr otection exception (#GP). Method 2 is enabled when the VME flag is set to 0 and the IOPL value is less than 3.
Vol. 3A 15-27 8086 EMULATION 6. Loads the CS and EIP register s with values from the interrupt vect or table entry pointed to by the interrupt vector number . Only the 16 low-order bits of the EIP are loaded and t he 16 high-order bits are set to 0. The interrupt vecto r table is assumed to be at linear address 0 of the current virtual-8086 task.
15-28 Vol. 3A 8086 EMULATION 15.4 PROTECTED-MODE VIRTUAL INTERRUPT S The IA-32 processors (beginning with the Pent ium processo r) also support the VIF and VIP flags in the EFLAGS register in protected mode by sett ing the PVI (protected-mode virt ual interrupt) flag in the CR4 register .
16 Mixing 16-Bit and 32-Bit Code.
.
Vol. 3A 16-1 CHAPTER 16 MIXING 16-BIT AND 32-BIT CODE Program modules written to run on IA-3 2 processo rs can be either 16-bi t modules or 32-bit modules. T able 16-1 shows the characteristic of 16 -bit and 32-bit modules. The IA-32 processors function most ef ficiently when executing 32-bit program modules.
16-2 Vol. 3A MIXING 16-BIT AND 32-BIT CODE 16.1 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 segmen ts and operation s : • The D (default operand and address size) flag in code-segment descriptors.
Vol. 3A 16-3 MIXING 16-BIT AND 3 2-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 .
16-4 Vol. 3A MIXING 16-BIT AND 32-BIT CODE A stack that spans less than 64 KBytes can be sh ared by both 16- and 32-b it code segments. This class of stacks includes: • Stacks in expand-up segments with the G (granularity) and B (big) flags in the stack- segment descriptor clear .
Vol. 3A 16-5 MIXING 16-BIT AND 3 2-BIT CODE These methods of transferring program control overcome t he following architectural lim itations imposed on calls between 16-bit and 32-bit code segment s: .
16-6 Vol. 3A MIXING 16-BIT AND 32-BIT CODE While executing 32-b it 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 cal led cod.
Vol. 3A 16-7 MIXING 16-BIT AND 3 2-BIT CODE 16.4.2.1 Controlling the Operand-Size Att ribute For a Call Three things can determine the operand-size of a call: • The D flag in the segment descriptor for the callin g code segment. • An operand-size instruction prefix.
16-8 Vol. 3A MIXING 16-BIT AND 32-BIT CODE 16.4.3 Interrupt Control T ransfers A program-control transfer caused by an exception or interrupt is always carried out through an interrupt or trap gate (located in the IDT).
Vol. 3A 16-9 MIXING 16-BIT AND 3 2-BIT CODE 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 beyon d FFFFH, the interface procedure will need to prov ide the offset to the entry point.
16-10 Vol. 3A MIXING 16-BIT AND 32-BIT CODE.
17 IA-32 Ar chitectur e Compatibility.
.
Vol. 3A 17-1 CHAPTER 17 IA-32 ARCHITECTURE COMP ATIBILITY All IA-32 processors are binary compatible. Compatibili ty means that, within certain limited constraints, programs that execu te on previous generations of IA-32 processors wi ll produce identical results when executed on later IA-32 processors.
17-2 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY 17.2. RESERVED BITS Throughout this manual, certa in bits are marked as reserved in many register and mem ory layout descriptions.
Vol. 3A 17-3 IA-32 ARCHITECTURE COMPATIBILITY 2. Execute the CPUID instruction. The CPUID instruction (added to the IA-32 in the Pen tium processor) indicates the presen ce of new features directly .
17-4 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY ming for conversion to integer . The remaining two instructions (MONIT OR and MW AIT) accelerate synchronization of threads.
Vol. 3A 17-5 IA-32 ARCHITECTURE COMPATIBILITY 17.12.1 Instructions Added Prio r to the Pentiu m Processor The following instructions were added in the Intel486 processor: • BSW AP (byte swap) instruction. • XADD (exchange and add) instruction. • CMPXCHG (compare and ex change) instruction.
17-6 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY • Bit scan instructions. • Double-shift instructio ns. • Byte set on condition instruct ion . • Move with sign/zero extension. • Generalized multi ply instruction. • MOV to and from control registers.
Vol. 3A 17-7 IA-32 ARCHITECTURE COMPATIBILITY • VIP (virtual interrupt pending), bit 20. • ID (identification flag), bit 21. The AC flag (bit 18) was added to the EF LAGS register in the Intel486 processor .
17-8 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY 17.16.2 EFLAGS Pushed on the St ack The setting of the stored values of bits 12 through 15 (which includes the IOPL fi eld and the NT flag) in the EFLAGS .
Vol. 3A 17-9 IA-32 ARCHITECTURE COMPATIBILITY As on the Intel 286 and Intel38 6 processors, the MP (monitor coprocessor) flag (bit 1 of register CR0) determines whether the W AIT/FW AIT instructions or w aiting-type floating-point instruc- tions trap when the context of the x87 FPU is different from that of the currently -executing task.
17-10 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY is reserved on these processors. The additio n o f the SF flag on a 32-bit x87 FPU h as no impact on software. Existing exception h andlers need not chan ge, bu t may be upgraded to take advan- tage of the additional in formation.
Vol. 3A 17-11 IA-32 ARCHITECTURE COMPATIBILITY 17.17.5.1 NANS The 32-bit x87 FPUs disti nguish between signaling NaNs (SNaNs) and quiet NaNs (QNaN s). These x87 FPUs only generat e QNaNs and normally do not generate an ex ception upon encoun- tering a QNaN.
17-12 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY 17.17.6.2 NUMERIC OVERFLOW EXCEPTION (#O) On the 32-bit x87 FPUs, wh en the numeric overflow exceptio n is masked and the roundi ng mode is set to chop (toward 0), the resu lt is the largest positive or smallest negative number .
Vol. 3A 17-13 IA-32 ARCHITECTURE COMPATIBILITY 16-bit IA-32 math coprocessors, it takes precedence over all other exceptions. This difference causes no impact on existing software, but some unneed ed normalization of denormalized oper- ands is prevented on the Intel486 processor and Intel 387 math coprocessor .
17-14 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY 17.17.6.8 INVALI D OPERATION EXCEP TION ON DENOR MALS An invalid-operation exception is not ge nerated on the 32-bit x87 FPUs upon encountering a denormal value when executing a FSQR T , FDIV , or FPREM instruction or upon conversion to BCD or to integer .
Vol. 3A 17-15 IA-32 ARCHITECTURE COMPATIBILITY 17.17.6.14 FLOATING-POIN T ERROR EXCEPTION (#MF) In real mode and protected mode (not inclu ding virtual-8086 mode), interrupt vect or 16 must point to the floatin g-point exception handler .
17-16 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY 17.17.7.5 FUCOM, FUCOMP , AND FUCOMPP INSTRUCTIONS When executing the FUCOM, FUCOMP , and FU COMPP instruction s, the 32-bit x87 FPUs perform unordered comp are according to IEEE Stan dard 754. These instructions do not exist on the 16-bit IA-32 math coprocessors.
Vol. 3A 17-17 IA-32 ARCHITECTURE COMPATIBILITY 16-bit IA-32 math coprocessors do report a deno rmal-operand ex ception in this situ ation. This difference does not af fect existing software. On the 32-bit x87 FPUs, loading a denormal value that is in singl e- or double-real format causes the value to be converted to extended-real format.
17-18 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY 17.17.7.15 FXAM INSTRUCTION W ith the 32-bit x87 FPUs, if the FPU encounters an empty register when executing the FXAM i ns tr u c ti o n , i t n o t g e n e ra t e co m b i na t i o ns o f C0 t h ro ug h C3 e q u al t o 110 1 o r 1111 .
Vol. 3A 17-19 IA-32 ARCHITECTURE COMPATIBILITY 17.17.1 1 Operands S plit Across Segment s and/or Pages On the P6 family , Pentium, and Intel486 p rocessor 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 no t the second half.
17-20 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY coprocessor keeps its ERROR# out put in inactive state after hardware reset; the Intel 387 copro- cessor keeps its ERROR# output in act ive state after hardware reset. Upon hardware reset or executi on of the FINIT/FNINIT i nstruction, the Intel 387 math copro- cessor signals an error conditio n.
Vol. 3A 17-21 IA-32 ARCHITECTURE COMPATIBILITY cmp ax, 037fh jz Intel487_SX_Math_CoProcessor_present;ax=037fh jmp Intel486_SX_microprocessor_prese nt;ax=ffffh If the Intel 487 SX math coprocessor is not presen t, the following code can be run to set the CR0 register for the Intel486 S X pro c essor .
17-22 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY The content of CR4 is 0H following a hardware reset. Control register CR4 was introduced in the Pentiu m processor . This register contains flags that enable certain new extensions provided in th e Penti um processor: • VME — V irtual-8086 mode extensions.
Vol. 3A 17-23 IA-32 ARCHITECTURE COMPATIBILITY 17.21. MEMORY MANAG EMENT FACILITIES The following sections describe the new m emory management facilities avail able in the various IA-32 processors and some comp atib ility differences.
17-24 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY the data cache; in the Intel486 processor , they implement a wr ite-through strategy . See T able 10-5 for a comparison of these bi ts on t he P6 family , Pentium, and Intel486 processo rs. For complete information on caching, see Chapter 10, “Memory Cache Control.
Vol. 3A 17-25 IA-32 ARCHITECTURE COMPATIBILITY On the P6 family and Pentium p rocessors, reserved bits 1 1, 12, 14 and 15 are hard-wired to 0. On the Intel486 processor, however , bit 12 can be set. See T able 9-1 for th e dif ferent settings of this register following a power-up or hardware reset.
17-26 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY tecture has been added for handling and reporting on hardware errors. See Chapter 14, “Machine-Check Architecture,” for a detail ed descrip tion of the new conditions.
Vol. 3A 17-27 IA-32 ARCHITECTURE COMPATIBILITY 17.24.1 Machine-Ch eck Architecture The Pentium Pro processor intro duced a new architecture to the IA-32 for handling and reporting on machine-ch eck exceptions.
17-28 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY 17.25.3 IDT Limit The LIDT instruction can be used to set a lim it 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.
Vol. 3A 17-29 IA-32 ARCHITECTURE COMPATIBILITY • For the 82489DX, in the lowest pri ority delivery mode, all the target local APICs specified by the destination fi eld participate in the lowest p riority arbitration. For the local APIC, only those local APICs which have free interrupt slots will participate in the lowest priority arbitration.
17-30 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY 17.27.1 P6 F amily and Pentium Processor TSS When the virtual mo de extensions are enabled (by setting the VME fl ag in control register CR4), the TSS in.
Vol. 3A 17-31 IA-32 ARCHITECTURE COMPATIBILITY general-protection exceptions (# GP). Figure 17-1 demonstrates the different areas accessed by the Intel486 and the P6 family and Pent ium processors. 17.28. CACHE MANAGEMENT The P6 family processors include two levels of internal caches: L1 (level 1) and L2 (level 2).
17-32 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY External system hardware can force the Pentium processor to disable cachin g or to use the write- through cache policy should that be required. In the P6 family processors, the MTRRs can be used to override the CD and NW flags (see T able 10-6).
Vol. 3A 17-33 IA-32 ARCHITECTURE COMPATIBILITY cache to be disabled and enabled, independently of the L1 and L2 caches (see Section 10.5.4, “Disabling and Enabling the L3 Cache”). 17.29. PAGING This section identifies enhancements made to the paging mechanism and implementation differ- ences in the paging mechanism for various IA-32 processors.
17-34 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY The sequence bounded by the MOV and JMP instructions shoul d be identity mapped (that is, the instructions should reside on a page whos e linear and physical addresses are identical). For the P6 family processors, the MOV CR0, REG instruction is serializing, so the jump oper- ation is not required.
Vol. 3A 17-35 IA-32 ARCHITECTURE COMPATIBILITY 17.30.2 Error Code Pushes The Intel486 processor implements the error co de pushed on the stack as a 16-bit value. When pushed onto a 32-bit stack, t he Intel486 processor only pushes 2 bytes and updates ESP by 4.
17-36 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY The 32-bit processors also have descripto rs for TSS segments, call gates, interrupt gates, and trap gates that supp ort the 32-bit architecture. Both kinds of desc riptors can be used in the same system.
Vol. 3A 17-37 IA-32 ARCHITECTURE COMPATIBILITY An exception to this behavior occurs when a st ack 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 popp ed, no segment limit vi olation occurs and the stack pointer will wrap around to 0.
17-38 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY 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 7.
Vol. 3A 17-39 IA-32 ARCHITECTURE COMPATIBILITY bus to send the interrupt vector to the processor . After receiving the interrupt request signal, the processor asserts LOCK# to insure that no othe r data appears on the data b us until the interrupt vector is received.
17-40 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY 17.36.3 Memory T ype Range Registers Memory type range registers (MTRRs) are a ne w feature introduced in to the IA-32 in the Pentium Pro processor .
Vol. 3A 17-41 IA-32 ARCHITECTURE COMPATIBILITY 17.36.5 Performance-M onitoring Counters The P6 family and Pentium pro cessors provide two performance-monit oring counters for use in monitoring inte rnal hardware operatio ns.
17-42 Vol. 3A IA-32 ARCHITECTURE COMPATIBILITY.
INTEL SALES OFFICES ASIA P ACIFIC Australia Intel Corp. Level 2 448 St Kilda Road Melbourne VI C 3004 Australia Fax:613- 9862 5599 China Intel Corp. Rm 709, Shaanxi Zhongda Int'l Bldg No.30 Nandajie Street Xian AX71000 2 China Fax:(86 29) 7203 356 Intel Corp.
Intel Corp. 999 CANADA PLACE , Suite 404,#1 1 Va n c ou v e r B C V6C 3E2 Canada Fax:604- 844-2813 Intel Corp. 2650 Quee nsview Dr ive, Suite 25 0 Ottawa ON K2B 8H6 Canada Fax:613- 820-5936 Intel Corp. 190 Attwell Drive, Suite 50 0 Rexcdale ON M9W 6H8 Canada Fax:416- 675-2438 Intel Corp.
Un punto importante, dopo l’acquisto del dispositivo (o anche prima di acquisto) è quello di leggere il manuale. Dobbiamo farlo per diversi motivi semplici:
Se non hai ancora comprato il Intel IA-32 è un buon momento per familiarizzare con i dati di base del prodotto. Prime consultare le pagine iniziali del manuale d’uso, che si trova al di sopra. Dovresti trovare lì i dati tecnici più importanti del Intel IA-32 - in questo modo è possibile verificare se l’apparecchio soddisfa le tue esigenze. Esplorando le pagine segenti del manuali d’uso Intel IA-32 imparerai tutte le caratteristiche del prodotto e le informazioni sul suo funzionamento. Le informazioni sul Intel IA-32 ti aiuteranno sicuramente a prendere una decisione relativa all’acquisto.
In una situazione in cui hai già il Intel IA-32, ma non hai ancora letto il manuale d’uso, dovresti farlo per le ragioni sopra descritte. Saprai quindi se hai correttamente usato le funzioni disponibili, e se hai commesso errori che possono ridurre la durata di vita del Intel IA-32.
Tuttavia, uno dei ruoli più importanti per l’utente svolti dal manuale d’uso è quello di aiutare a risolvere i problemi con il Intel IA-32. Quasi sempre, ci troverai Troubleshooting, cioè i guasti più frequenti e malfunzionamenti del dispositivo Intel IA-32 insieme con le istruzioni su come risolverli. Anche se non si riesci a risolvere il problema, il manuale d’uso ti mostrerà il percorso di ulteriori procedimenti – il contatto con il centro servizio clienti o il servizio più vicino.
