System V Application Binary Interface: AMD64 Architecture Processor Supplement

128 Pages·2012·0.544 MB·English

by  Michael Matz, Jan Hubicka, Andreas Jaeger, Mark Mitchell

System V Application Binary Interface
AMD64 Architecture Processor Supplement
Draft Version 0.99.6
Edited by
Michael Matz1, Jan Hubicˇka2, Andreas Jaeger3, Mark Mitchell4
July 2, 2012
[email protected]
[email protected]
[email protected]
[email protected]
AMD64ABIDraft0.99.6–July2,2012–17:14
Contents
1 Introduction 8
2 SoftwareInstallation 9
3 LowLevelSystemInformation 10
3.1 MachineInterface . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.1.1 ProcessorArchitecture . . . . . . . . . . . . . . . . . . . 10
3.1.2 DataRepresentation . . . . . . . . . . . . . . . . . . . . 10
3.2 FunctionCallingSequence . . . . . . . . . . . . . . . . . . . . . 14
3.2.1 RegistersandtheStackFrame . . . . . . . . . . . . . . . 14
3.2.2 TheStackFrame . . . . . . . . . . . . . . . . . . . . . . 15
3.2.3 ParameterPassing . . . . . . . . . . . . . . . . . . . . . 16
3.3 OperatingSystemInterface . . . . . . . . . . . . . . . . . . . . . 23
3.3.1 ExceptionInterface . . . . . . . . . . . . . . . . . . . . . 23
3.3.2 VirtualAddressSpace . . . . . . . . . . . . . . . . . . . 23
3.3.3 PageSize . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.3.4 VirtualAddressAssignments . . . . . . . . . . . . . . . . 23
3.4 ProcessInitialization . . . . . . . . . . . . . . . . . . . . . . . . 26
3.4.1 InitialStackandRegisterState . . . . . . . . . . . . . . . 26
3.4.2 ThreadState . . . . . . . . . . . . . . . . . . . . . . . . 29
3.4.3 AuxiliaryVector . . . . . . . . . . . . . . . . . . . . . . 29
3.5 CodingExamples . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.5.1 ArchitecturalConstraints . . . . . . . . . . . . . . . . . . 32
3.5.2 Conventions . . . . . . . . . . . . . . . . . . . . . . . . 34
3.5.3 Position-IndependentFunctionPrologue . . . . . . . . . . 35
3.5.4 DataObjects . . . . . . . . . . . . . . . . . . . . . . . . 36
3.5.5 FunctionCalls . . . . . . . . . . . . . . . . . . . . . . . 44
3.5.6 Branching . . . . . . . . . . . . . . . . . . . . . . . . . . 46
1
AMD64ABIDraft0.99.6–July2,2012–17:14
3.5.7 VariableArgumentLists . . . . . . . . . . . . . . . . . . 49
3.6 DWARFDefinition . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.6.1 DWARFReleaseNumber . . . . . . . . . . . . . . . . . 55
3.6.2 DWARFRegisterNumberMapping . . . . . . . . . . . . 55
3.7 StackUnwindAlgorithm . . . . . . . . . . . . . . . . . . . . . . 55
4 ObjectFiles 59
4.1 ELFHeader . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4.1.1 MachineInformation . . . . . . . . . . . . . . . . . . . . 59
4.1.2 NumberofProgramHeaders . . . . . . . . . . . . . . . . 59
4.2 Sections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
4.2.1 SectionFlags . . . . . . . . . . . . . . . . . . . . . . . . 60
4.2.2 Sectiontypes . . . . . . . . . . . . . . . . . . . . . . . . 61
4.2.3 SpecialSections . . . . . . . . . . . . . . . . . . . . . . 61
4.2.4 EH_FRAMEsections . . . . . . . . . . . . . . . . . . . 62
4.3 SymbolTable . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
4.4 Relocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.4.1 RelocationTypes . . . . . . . . . . . . . . . . . . . . . . 68
4.4.2 LargeModels . . . . . . . . . . . . . . . . . . . . . . . . 73
5 ProgramLoadingandDynamicLinking 74
5.1 ProgramLoading . . . . . . . . . . . . . . . . . . . . . . . . . . 74
5.1.1 Programheader . . . . . . . . . . . . . . . . . . . . . . . 75
5.2 DynamicLinking . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.2.1 ProgramInterpreter . . . . . . . . . . . . . . . . . . . . . 82
5.2.2 InitializationandTerminationFunctions . . . . . . . . . . 82
6 Libraries 83
6.1 CLibrary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
6.1.1 GlobalDataSymbols . . . . . . . . . . . . . . . . . . . . 83
6.1.2 FloatingPointEnvironmentFunctions . . . . . . . . . . . 83
6.2 UnwindLibraryInterface . . . . . . . . . . . . . . . . . . . . . . 84
6.2.1 ExceptionHandlerFramework . . . . . . . . . . . . . . . 85
6.2.2 DataStructures . . . . . . . . . . . . . . . . . . . . . . . 87
6.2.3 ThrowinganException . . . . . . . . . . . . . . . . . . . 89
6.2.4 ExceptionObjectManagement . . . . . . . . . . . . . . . 92
6.2.5 ContextManagement . . . . . . . . . . . . . . . . . . . . 92
6.2.6 PersonalityRoutine . . . . . . . . . . . . . . . . . . . . . 95
2
AMD64ABIDraft0.99.6–July2,2012–17:14
6.3 UnwindingThroughAssemblerCode . . . . . . . . . . . . . . . 99
7 DevelopmentEnvironment 102
8 ExecutionEnvironment 103
9 Conventions 104
9.1 C++ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
9.2 Fortran . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
9.2.1 Names . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
9.2.2 RepresentationofFortranTypes . . . . . . . . . . . . . . 107
9.2.3 ArgumentPassing . . . . . . . . . . . . . . . . . . . . . 108
9.2.4 Functions . . . . . . . . . . . . . . . . . . . . . . . . . . 109
9.2.5 COMMONblocks . . . . . . . . . . . . . . . . . . . . . 110
9.2.6 Intrinsics . . . . . . . . . . . . . . . . . . . . . . . . . . 111
A LinuxConventions 122
A.1 Executionof32-bitPrograms . . . . . . . . . . . . . . . . . . . . 122
A.2 AMD64LinuxKernelConventions . . . . . . . . . . . . . . . . . 122
A.2.1 CallingConventions . . . . . . . . . . . . . . . . . . . . 122
A.2.2 StackLayout . . . . . . . . . . . . . . . . . . . . . . . . 123
A.2.3 RequiredProcessorFeatures . . . . . . . . . . . . . . . . 123
A.2.4 MiscellaneousRemarks . . . . . . . . . . . . . . . . . . 123
3
AMD64ABIDraft0.99.6–July2,2012–17:14
List of Tables
3.1 HardwareExceptionsandSignals . . . . . . . . . . . . . . . . . 24
3.2 Floating-PointExceptions . . . . . . . . . . . . . . . . . . . . . 24
3.3 x87Floating-PointControlWord . . . . . . . . . . . . . . . . . . 26
3.4 MXCSRStatusBits . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.5 rFLAGSBits . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.1 AMD64Identification . . . . . . . . . . . . . . . . . . . . . . . . 59
4.2 AMD64SpecificSectionHeaderFlag,sh_flags . . . . . . . . 60
4.3 SectionHeaderTypes . . . . . . . . . . . . . . . . . . . . . . . . 61
4.4 Specialsections . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.5 AdditionalSpecialSectionsfortheLargeCodeModel . . . . . . 62
4.6 CommonInformationEntry(CIE) . . . . . . . . . . . . . . . . . 64
4.7 CIEAugmentationSectionContent . . . . . . . . . . . . . . . . 65
4.8 FrameDescriptorEntry(FDE) . . . . . . . . . . . . . . . . . . . 66
4.9 FDEAugmentationSectionContent . . . . . . . . . . . . . . . . 67
4.10 RelocationTypes . . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.11 LargeModelRelocationTypes . . . . . . . . . . . . . . . . . . . 73
5.1 ProgramHeaderTypes . . . . . . . . . . . . . . . . . . . . . . . 75
7.1 PredefinedPre-ProcessorSymbols . . . . . . . . . . . . . . . . . 102
9.1 Milintrinsics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
9.2 F77intrinsics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
9.3 F90intrinsics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
9.4 Mathintrinsics . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
9.5 Unixintrinsics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
A.1 RequiredProcessorFeatures . . . . . . . . . . . . . . . . . . . . 124
4
AMD64ABIDraft0.99.6–July2,2012–17:14
List of Figures
3.1 ScalarTypes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.2 Bit-FieldRanges . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.3 StackFramewithBasePointer . . . . . . . . . . . . . . . . . . . 15
3.4 RegisterUsage . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.5 ParameterPassingExample . . . . . . . . . . . . . . . . . . . . . 22
3.6 RegisterAllocationExample . . . . . . . . . . . . . . . . . . . . 22
3.7 VirtualAddressConfiguration . . . . . . . . . . . . . . . . . . . 25
3.8 ConventionalSegmentArrangements . . . . . . . . . . . . . . . 26
3.9 InitialProcessStack . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.10 auxv_tTypeDefinition . . . . . . . . . . . . . . . . . . . . . . 29
3.11 AuxiliaryVectorTypes . . . . . . . . . . . . . . . . . . . . . . . 30
3.12 Position-IndependentFunctionPrologCode . . . . . . . . . . . . 35
3.13 AbsoluteLoadandStore(SmallModel) . . . . . . . . . . . . . . 37
3.14 Position-IndependentLoadandStore(SmallPICModel) . . . . . 38
3.15 AbsoluteLoadandStore(MediumModel) . . . . . . . . . . . . . 39
3.16 Position-IndependentLoadandStore(MediumPICModel) . . . . 40
3.17 Position-IndependentLoadandStore(MediumPICModel),con-
tinued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.18 AbsoluteGlobalDataLoadandStore . . . . . . . . . . . . . . . 42
3.19 FasterAbsoluteGlobalDataLoadandStore . . . . . . . . . . . . 42
3.20 Position-IndependentGlobalDataLoadandStore . . . . . . . . . 43
3.21 FasterPosition-IndependentGlobalDataLoadandStore . . . . . 43
3.22 Position-Independent Direct Function Call (Small and Medium
Model) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.23 Position-IndependentIndirectFunctionCall . . . . . . . . . . . . 44
3.24 AbsoluteDirectandIndirectFunctionCall . . . . . . . . . . . . . 45
3.25 Position-IndependentDirectandIndirectFunctionCall . . . . . . 45
3.26 AbsoluteBranchingCode . . . . . . . . . . . . . . . . . . . . . . 47
5
AMD64ABIDraft0.99.6–July2,2012–17:14
3.27 ImplicitCalculationofTargetAddress . . . . . . . . . . . . . . . 47
3.28 Position-IndependentBranchingCode . . . . . . . . . . . . . . . 48
3.29 AbsoluteSwitchCode . . . . . . . . . . . . . . . . . . . . . . . 48
3.30 Position-IndependentSwitchCode . . . . . . . . . . . . . . . . . 49
3.31 ParameterPassingExamplewithVariable-ArgumentList . . . . . 50
3.32 RegisterAllocationExampleforVariable-ArgumentList . . . . . 50
3.33 RegisterSaveArea . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.34 va_listTypeDeclaration . . . . . . . . . . . . . . . . . . . . 52
3.35 SampleImplementationofva_arg(l, int) . . . . . . . . . . 54
3.36 DWARFRegisterNumberMapping . . . . . . . . . . . . . . . . 56
3.37 PointerEncodingSpecificationByte . . . . . . . . . . . . . . . . 57
4.1 RelocatableFields . . . . . . . . . . . . . . . . . . . . . . . . . . 68
5.1 GlobalOffsetTable . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.2 ProcedureLinkageTable(smallandmediummodels) . . . . . . . 78
5.3 FinalLargeCodeModelPLT . . . . . . . . . . . . . . . . . . . . 81
6.1 ExamplesforUnwindinginAssembler . . . . . . . . . . . . . . . 101
9.1 Examplemappingofnames . . . . . . . . . . . . . . . . . . . . . 107
9.2 MappingofFortrantoCtypes . . . . . . . . . . . . . . . . . . . 107
Revision History
0.99 AdddescriptionofTLSrelocations(thankstoAlexandreOliva)andmention
thedecimalfloatingpointandAVXtypes(thankstoH.J.Lu).
0.98 Various clarifications and fixes according to feedback from Sun, thanks to
TerrenceMiller. DWARFregisternumbersforsomesystemregisters,thanks
toJanBeulich. AddR_X86_64_SIZE32andR_X86_64_SIZE64relo-
cations; extend meaning of e_phnum to handle more than 0xffff program
headers, thanks to Rod Evans. Add footnote about passing of decimal
datatypes. Specifythat_Boolisbooleanizedatthecaller.
0.97 IntegrateFortranABI.
0.96 UseSHF_X86_64_LARGEinsteadSHF_AMD64_LARGE(thankstoEvan-
droMenezes). CorrectvariousgrammaticalerrorsnotedbyMarkF.Haigh,
6
AMD64ABIDraft0.99.6–July2,2012–17:14
whoalsonotedthattherearenoglobalVLAsinC99. ThanksalsotoRobert
R.Henry.
0.95 Include description of the medium PIC memory model (thanks to Jan Hu-
bicˇka)andlargemodel(thankstoEvandroMenezes).
0.94 AddsectionsinDevelopmentEnvironment,ProgramLoading,adescription
ofEH_FRAMEsectionsandgeneralcleanupstomaketextinthisABIself-
contained. ThankstoMichaelWalkerandTerrenceMiller.
0.93 Addsectionsaboutprogramheaders,newsectiontypesandspecialsections
forunwindinginformation. ThankstoMichaelWalker.
0.92 Fixsometypos(thankstoBryanFord),addsectionaboutstacklayoutinthe
Linux kernel. Fix example in figure 3.5 (thanks to Tom Horsley). Add sec-
tion on unwinding through assembler (written by Michal Ludvig). Remove
mmxext feature (thanks to Evandro Menezes). Add section on Fortran (by
StevenBosscher)andstackunwinding(byJanHubicˇka).
0.91 Clarifythatx87isdefaultmode,notMMX(byHansPeterAnvin).
0.90 ChangeDWARFregisternumbersagain;mentionthat__m128needsalign-
ment; fix typo in figure 3.3; add some comments on kernel expectations;
mention TLS extensions; add example for passing of variable-argument
lists; change semantics of %rax in variable-argument lists; improve for-
matting; mention that X87 class is not used for passing; make /lib64 a
Linuxspecificsection;renamex86-64toAMD64;describepassingofcom-
plex types. Special thanks to Andi Kleen, Michal Ludvig, Michael Matz,
DavidO’BrienandEricYoungfortheircomments.
0.21 Define __int128 as class INTEGER in register passing. Mention that
%alisusedforvariadicargumentlists. Fixsometextualproblems. Thanks
toH.PeterAnvin,BoThorsen,andMichaelMatz.
0.20—2002-07-11 Change DWARF register number values of %rbx, %rsi,
%rsi (thanks to Michal Ludvig). Fix footnotes for fundamental types
(thanks to H. Peter Anvin). Specify size_t (thanks to Bo Thorsen and
Andreas Schwab). Add new section on floating point environment func-
tions.
7
AMD64ABIDraft0.99.6–July2,2012–17:14
0.19—2002-03-27 Set name of Linux dynamic linker, mention %fs. Incorpo-
rate changes from H. Peter Anvin <[email protected]> for booleans and de-
finehandlingofsub-64-bitintegertypesinregisters.
8
AMD64ABIDraft0.99.6–July2,2012–17:14
Chapter 1
Introduction
TheAMD641 architecture2 isanextensionofthex86architecture. Anyprocessor
implementingtheAMD64architecturespecificationwillalsoprovidecompatibil-
itymodesforpreviousdescendantsoftheIntel8086architecture,including32-bit
processors such as the Intel 386, Intel Pentium, and AMD K6-2 processor. Oper-
ating systems conforming to the AMD64 ABI may provide support for executing
programsthataredesignedtoexecuteinthesecompatibilitymodes. TheAMD64
ABI does not apply to such programs; this document applies only to programs
runninginthe“long”modeprovidedbytheAMD64architecture.
Exceptwhereotherwisenoted,theAMD64architectureABIfollowsthecon-
ventions described in the Intel386 ABI. Rather than replicate the entire contents
of the Intel386 ABI, the AMD64 ABI indicates only those places where changes
havebeenmadetotheIntel386ABI.
NoattempthasbeenmadetospecifyanABIforlanguagesotherthanC.How-
ever, it is assumed that many programming languages will wish to link with code
writteninC,sothattheABIspecificationsdocumentedhereapplytheretoo.3
1AMD64hasbeenpreviouslycalledx86-64. Thelatternameisusedinanumberofplacesout
ofhistoricalreasonsinsteadofAMD64.
2The architecture specification is available on the web at http://www.x86-64.org/
documentation.
3Seesection9.1fordetailsonC++ABI.
9
AMD64ABIDraft0.99.6–July2,2012–17:14