Skip to content

[WIP] Apx requirements for VM and GC stubs #116806

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
wants to merge 3 commits into
base: main
Choose a base branch
from
+229 −42
Draft

[WIP] Apx requirements for VM and GC stubs #116806

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
940607e
Add context tracking structure in backend and decode them in GC
khushal1996 Jun 18, 2025
5936622
Add APX bypass Knob
khushal1996 Mar 13, 2025
da7033a
Fill volatile context when required
khushal1996 Jun 20, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 6 additions & 6 deletions src/coreclr/System.Private.CoreLib/src/System/Runtime/ExceptionServices/AsmOffsets.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,9 +20,9 @@ class AsmOffsets
public const int OFFSETOF__REGDISPLAY__SP = 0x1b78;
public const int OFFSETOF__REGDISPLAY__ControlPC = 0x1b80;
#else // TARGET_UNIX
public const int SIZEOF__REGDISPLAY = 0xbf0;
public const int OFFSETOF__REGDISPLAY__SP = 0xbd8;
public const int OFFSETOF__REGDISPLAY__ControlPC = 0xbe0;
public const int SIZEOF__REGDISPLAY = 0xd70;
public const int OFFSETOF__REGDISPLAY__SP = 0xcd8;
public const int OFFSETOF__REGDISPLAY__ControlPC = 0xce0;
#endif // TARGET_UNIX
#elif TARGET_ARM64
#if TARGET_UNIX
Expand Down Expand Up @@ -82,11 +82,11 @@ class AsmOffsets
// Release build offsets
#if TARGET_AMD64
#if TARGET_UNIX
public const int SIZEOF__REGDISPLAY = 0x1b80;
public const int SIZEOF__REGDISPLAY = 0x1c00;
public const int OFFSETOF__REGDISPLAY__SP = 0x1b70;
public const int OFFSETOF__REGDISPLAY__ControlPC = 0x1b78;
#else // TARGET_UNIX
public const int SIZEOF__REGDISPLAY = 0xbf0;
public const int SIZEOF__REGDISPLAY = 0xc70;
public const int OFFSETOF__REGDISPLAY__SP = 0xbd0;
public const int OFFSETOF__REGDISPLAY__ControlPC = 0xbd8;
#endif // TARGET_UNIX
Expand Down Expand Up @@ -150,7 +150,7 @@ class AsmOffsets
#if TARGET_UNIX
public const int SIZEOF__PAL_LIMITED_CONTEXT = 0xca0;
#else // TARGET_UNIX
public const int SIZEOF__PAL_LIMITED_CONTEXT = 0x4d0;
public const int SIZEOF__PAL_LIMITED_CONTEXT = 0x550;
#endif // TARGET_UNIX
#elif TARGET_ARM64
#if TARGET_UNIX
Expand Down
18 changes: 18 additions & 0 deletions src/coreclr/debug/ee/debugger.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16891,6 +16891,24 @@ void FuncEvalFrame::UpdateRegDisplay_Impl(const PREGDISPLAY pRD, bool updateFloa
pRD->pCurrentContextPointers->R14 = &(pDE->m_context.R14);
pRD->pCurrentContextPointers->R15 = &(pDE->m_context.R15);

// This would mean we need to update winnt.h in windows sdk.
pRD->volatileCurrContextPointers.R16 = &(pDE->m_context.R16);
pRD->volatileCurrContextPointers.R17 = &(pDE->m_context.R17);
pRD->volatileCurrContextPointers.R18 = &(pDE->m_context.R18);
pRD->volatileCurrContextPointers.R19 = &(pDE->m_context.R19);
pRD->volatileCurrContextPointers.R20 = &(pDE->m_context.R20);
pRD->volatileCurrContextPointers.R21 = &(pDE->m_context.R21);
pRD->volatileCurrContextPointers.R22 = &(pDE->m_context.R22);
pRD->volatileCurrContextPointers.R23 = &(pDE->m_context.R23);
pRD->volatileCurrContextPointers.R24 = &(pDE->m_context.R24);
pRD->volatileCurrContextPointers.R25 = &(pDE->m_context.R25);
pRD->volatileCurrContextPointers.R26 = &(pDE->m_context.R26);
pRD->volatileCurrContextPointers.R27 = &(pDE->m_context.R27);
pRD->volatileCurrContextPointers.R28 = &(pDE->m_context.R28);
pRD->volatileCurrContextPointers.R29 = &(pDE->m_context.R29);
pRD->volatileCurrContextPointers.R30 = &(pDE->m_context.R30);
pRD->volatileCurrContextPointers.R31 = &(pDE->m_context.R31);

// SyncRegDisplayToCurrentContext() sets the pRD->SP and pRD->ControlPC on AMD64.
SyncRegDisplayToCurrentContext(pRD);

Expand Down
23 changes: 23 additions & 0 deletions src/coreclr/debug/inc/dbgtargetcontext.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -256,6 +256,29 @@ typedef struct DECLSPEC_ALIGN(16) {
DWORD64 R14;
DWORD64 R15;

//EGPRs for APX
union {
struct {
DWORD64 R16;
DWORD64 R17;
DWORD64 R18;
DWORD64 R19;
DWORD64 R20;
DWORD64 R21;
DWORD64 R22;
DWORD64 R23;
DWORD64 R24;
DWORD64 R25;
DWORD64 R26;
DWORD64 R27;
DWORD64 R28;
DWORD64 R29;
DWORD64 R30;
DWORD64 R31;
};
DWORD64 R[16];
};

DWORD64 Rip;

union {
Expand Down
6 changes: 3 additions & 3 deletions src/coreclr/inc/gcinfotypes.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -600,16 +600,16 @@ struct AMD64GcInfoEncoding {
static const int SIZE_OF_STACK_AREA_ENCBASE = 3;
static const int SIZE_OF_EDIT_AND_CONTINUE_PRESERVED_AREA_ENCBASE = 4;
static const int REVERSE_PINVOKE_FRAME_ENCBASE = 6;
static const int NUM_REGISTERS_ENCBASE = 2;
static const int NUM_REGISTERS_ENCBASE = 3;
static const int NUM_STACK_SLOTS_ENCBASE = 2;
static const int NUM_UNTRACKED_SLOTS_ENCBASE = 1;
static const int NORM_PROLOG_SIZE_ENCBASE = 5;
static const int NORM_EPILOG_SIZE_ENCBASE = 3;
static const int NORM_CODE_OFFSET_DELTA_ENCBASE = 3;
static const int INTERRUPTIBLE_RANGE_DELTA1_ENCBASE = 6;
static const int INTERRUPTIBLE_RANGE_DELTA2_ENCBASE = 6;
static const int REGISTER_ENCBASE = 3;
static const int REGISTER_DELTA_ENCBASE = 2;
static const int REGISTER_ENCBASE = 5;
static const int REGISTER_DELTA_ENCBASE = 5;
static const int STACK_SLOT_ENCBASE = 6;
static const int STACK_SLOT_DELTA_ENCBASE = 4;
static const int NUM_SAFE_POINTS_ENCBASE = 2;
Expand Down
40 changes: 38 additions & 2 deletions src/coreclr/inc/regdisp.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -197,6 +197,34 @@ typedef struct _Arm64VolatileContextPointer
} Arm64VolatileContextPointer;
#endif //TARGET_ARM64

#if defined(TARGET_AMD64)
typedef struct _Amd64VolatileContextPointer
{
union {
struct {
PDWORD64 R16;
PDWORD64 R17;
PDWORD64 R18;
PDWORD64 R19;
PDWORD64 R20;
PDWORD64 R21;
PDWORD64 R22;
PDWORD64 R23;
PDWORD64 R24;
PDWORD64 R25;
PDWORD64 R26;
PDWORD64 R27;
PDWORD64 R28;
PDWORD64 R29;
PDWORD64 R30;
PDWORD64 R31;
//X18 is reserved by OS, in userspace it represents TEB
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
//X18 is reserved by OS, in userspace it represents TEB

};
PDWORD64 R[16];
};
} Amd64VolatileContextPointer;
#endif //TARGET_AMD64

#if defined(TARGET_LOONGARCH64)
typedef struct _LoongArch64VolatileContextPointer
{
Expand Down Expand Up @@ -253,6 +281,10 @@ struct REGDISPLAY : public REGDISPLAY_BASE {
LoongArch64VolatileContextPointer volatileCurrContextPointers;
#endif

#ifdef TARGET_AMD64
Amd64VolatileContextPointer volatileCurrContextPointers;
#endif

#ifdef TARGET_RISCV64
RiscV64VolatileContextPointer volatileCurrContextPointers;
#endif
Expand Down Expand Up @@ -563,7 +595,11 @@ inline void FillRegDisplay(const PREGDISPLAY pRD, PT_CONTEXT pctx, PT_CONTEXT pC
// Fill volatile context pointers. They can be used by GC in the case of the leaf frame
for (int i=0; i < 18; i++)
pRD->volatileCurrContextPointers.X[i] = &pctx->X[i];
#elif defined(TARGET_LOONGARCH64) // TARGET_ARM64
#elif defined(TARGET_AMD64) // TARGET_ARM64
// Fill volatile context pointers. They can be used by GC in the case of the leaf frame
for (int i=0; i < 16; i++)
pRD->volatileCurrContextPointers.R[i] = &pctx->R[i];
#elif defined(TARGET_LOONGARCH64) // TARGET_AMD64
pRD->volatileCurrContextPointers.A0 = &pctx->A0;
pRD->volatileCurrContextPointers.A1 = &pctx->A1;
pRD->volatileCurrContextPointers.A2 = &pctx->A2;
Expand Down Expand Up @@ -664,7 +700,7 @@ inline size_t * getRegAddr (unsigned regNum, PTR_CONTEXT regs)

return (PTR_size_t)(PTR_BYTE(regs) + OFFSET_OF_REGISTERS[regNum]);
#elif defined(TARGET_AMD64)
_ASSERTE(regNum < 16);
_ASSERTE(regNum < 32);
return (size_t *)&regs->Rax + regNum;
#elif defined(TARGET_ARM)
_ASSERTE(regNum < 16);
Expand Down
6 changes: 6 additions & 0 deletions src/coreclr/jit/compiler.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9771,6 +9771,12 @@ class Compiler
//
bool canUseApxEncoding() const
{
#ifdef DEBUG
if (JitConfig.JitBypassApxCheck())
{
return true;
}
#endif
return compOpportunisticallyDependsOn(InstructionSet_APX);
}

Expand Down
1 change: 1 addition & 0 deletions src/coreclr/jit/jitconfigvalues.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -374,6 +374,7 @@ RELEASE_CONFIG_INTEGER(JitNoInline, "JitNoInline", 0)

#if defined(DEBUG)
CONFIG_INTEGER(JitStressRex2Encoding, "JitStressRex2Encoding", 0) // Enable rex2 encoding for compatible instructions.
CONFIG_INTEGER(JitBypassApxCheck, "JitBypassApxCheck", 0) // Enable apx for testing
CONFIG_INTEGER(JitStressPromotedEvexEncoding, "JitStressPromotedEvexEncoding", 0) // Enable promoted EVEX encoding for
// compatible instructions.
#endif
Expand Down
20 changes: 10 additions & 10 deletions src/coreclr/nativeaot/Runtime/amd64/AsmOffsetsCpu.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,28 +8,28 @@
// NOTE: the offsets MUST be in hex notation WITHOUT the 0x prefix

#ifndef UNIX_AMD64_ABI
PLAT_ASM_SIZEOF(250, ExInfo)
PLAT_ASM_SIZEOF(2D0, ExInfo)
PLAT_ASM_OFFSET(0, ExInfo, m_pPrevExInfo)
PLAT_ASM_OFFSET(8, ExInfo, m_pExContext)
PLAT_ASM_OFFSET(10, ExInfo, m_exception)
PLAT_ASM_OFFSET(18, ExInfo, m_kind)
PLAT_ASM_OFFSET(19, ExInfo, m_passNumber)
PLAT_ASM_OFFSET(1c, ExInfo, m_idxCurClause)
PLAT_ASM_OFFSET(20, ExInfo, m_frameIter)
PLAT_ASM_OFFSET(240, ExInfo, m_notifyDebuggerSP)
PLAT_ASM_OFFSET(2C0, ExInfo, m_notifyDebuggerSP)

PLAT_ASM_OFFSET(0, PInvokeTransitionFrame, m_RIP)
PLAT_ASM_OFFSET(8, PInvokeTransitionFrame, m_FramePointer)
PLAT_ASM_OFFSET(10, PInvokeTransitionFrame, m_pThread)
PLAT_ASM_OFFSET(18, PInvokeTransitionFrame, m_Flags)
PLAT_ASM_OFFSET(20, PInvokeTransitionFrame, m_PreservedRegs)

PLAT_ASM_SIZEOF(220, StackFrameIterator)
PLAT_ASM_SIZEOF(2A0, StackFrameIterator)
PLAT_ASM_OFFSET(10, StackFrameIterator, m_FramePointer)
PLAT_ASM_OFFSET(18, StackFrameIterator, m_ControlPC)
PLAT_ASM_OFFSET(20, StackFrameIterator, m_RegDisplay)
PLAT_ASM_OFFSET(210, StackFrameIterator, m_OriginalControlPC)
PLAT_ASM_OFFSET(218, StackFrameIterator, m_pPreviousTransitionFrame)
PLAT_ASM_OFFSET(290, StackFrameIterator, m_OriginalControlPC)
PLAT_ASM_OFFSET(298, StackFrameIterator, m_pPreviousTransitionFrame)

PLAT_ASM_SIZEOF(100, PAL_LIMITED_CONTEXT)
PLAT_ASM_OFFSET(0, PAL_LIMITED_CONTEXT, IP)
Expand All @@ -56,10 +56,10 @@ PLAT_ASM_OFFSET(0d0, PAL_LIMITED_CONTEXT, Xmm13)
PLAT_ASM_OFFSET(0e0, PAL_LIMITED_CONTEXT, Xmm14)
PLAT_ASM_OFFSET(0f0, PAL_LIMITED_CONTEXT, Xmm15)

PLAT_ASM_SIZEOF(130, REGDISPLAY)
PLAT_ASM_OFFSET(78, REGDISPLAY, SP)
PLAT_ASM_OFFSET(80, REGDISPLAY, IP)
PLAT_ASM_OFFSET(88, REGDISPLAY, SSP)
PLAT_ASM_SIZEOF(1B0, REGDISPLAY)
PLAT_ASM_OFFSET(F8, REGDISPLAY, SP)
PLAT_ASM_OFFSET(100, REGDISPLAY, IP)
PLAT_ASM_OFFSET(108, REGDISPLAY, SSP)

PLAT_ASM_OFFSET(18, REGDISPLAY, pRbx)
PLAT_ASM_OFFSET(20, REGDISPLAY, pRbp)
Expand All @@ -69,7 +69,7 @@ PLAT_ASM_OFFSET(58, REGDISPLAY, pR12)
PLAT_ASM_OFFSET(60, REGDISPLAY, pR13)
PLAT_ASM_OFFSET(68, REGDISPLAY, pR14)
PLAT_ASM_OFFSET(70, REGDISPLAY, pR15)
PLAT_ASM_OFFSET(90, REGDISPLAY, Xmm)
PLAT_ASM_OFFSET(110, REGDISPLAY, Xmm)

#else // !UNIX_AMD64_ABI

Expand Down
16 changes: 16 additions & 0 deletions src/coreclr/nativeaot/Runtime/regdisplay.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,6 +27,22 @@ struct REGDISPLAY
PTR_uintptr_t pR13;
PTR_uintptr_t pR14;
PTR_uintptr_t pR15;
PTR_uintptr_t pR16;
PTR_uintptr_t pR17;
PTR_uintptr_t pR18;
PTR_uintptr_t pR19;
PTR_uintptr_t pR20;
PTR_uintptr_t pR21;
PTR_uintptr_t pR22;
PTR_uintptr_t pR23;
PTR_uintptr_t pR24;
PTR_uintptr_t pR25;
PTR_uintptr_t pR26;
PTR_uintptr_t pR27;
PTR_uintptr_t pR28;
PTR_uintptr_t pR29;
PTR_uintptr_t pR30;
PTR_uintptr_t pR31;
#endif // TARGET_AMD64

uintptr_t SP;
Expand Down
38 changes: 21 additions & 17 deletions src/coreclr/pal/inc/pal.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1464,24 +1464,28 @@ typedef struct DECLSPEC_ALIGN(16) _CONTEXT {
M512 Zmm31;
};

struct
union
{
DWORD64 R16;
DWORD64 R17;
DWORD64 R18;
DWORD64 R19;
DWORD64 R20;
DWORD64 R21;
DWORD64 R22;
DWORD64 R23;
DWORD64 R24;
DWORD64 R25;
DWORD64 R26;
DWORD64 R27;
DWORD64 R28;
DWORD64 R29;
DWORD64 R30;
DWORD64 R31;
struct
{
DWORD64 R16;
DWORD64 R17;
DWORD64 R18;
DWORD64 R19;
DWORD64 R20;
DWORD64 R21;
DWORD64 R22;
DWORD64 R23;
DWORD64 R24;
DWORD64 R25;
DWORD64 R26;
DWORD64 R27;
DWORD64 R28;
DWORD64 R29;
DWORD64 R30;
DWORD64 R31;
};
DWORD64 R[16];
};

} CONTEXT, *PCONTEXT, *LPCONTEXT;
Expand Down
2 changes: 1 addition & 1 deletion src/coreclr/unwinder/amd64/unwinder.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -242,7 +242,7 @@ BOOL DacUnwindStackFrame(CONTEXT * pContext, KNONVOLATILE_CONTEXT_POINTERS* pCon

if (res && pContextPointers)
{
for (int i = 0; i < 16; i++)
for (int i = 0; i < 32; i++)
{
*(&pContextPointers->Rax + i) = &pContext->Rax + i;
}
Expand Down
9 changes: 8 additions & 1 deletion src/coreclr/vm/amd64/asmconstants.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -292,7 +292,7 @@ ASMCONSTANTS_C_ASSERT(OFFSETOF__VASigCookie__pNDirectILStub
#else
// Expression is too complicated, is currently:
// (8*6 + 4*2 + 2*6 + 4 + 8*6 + 8*16 + 8 + /*XMM_SAVE_AREA32*/(2*2 + 1*2 + 2 + 4 + 2*2 + 4 + 2*2 + 4*2 + 16*8 + 16*16 + 1*96) + 26*16 + 8 + 8*5)
#define SIZEOF__CONTEXT (1232)
#define SIZEOF__CONTEXT (1360)
#endif
ASMCONSTANTS_C_ASSERT(SIZEOF__CONTEXT
== sizeof(CONTEXT));
Expand Down Expand Up @@ -445,6 +445,13 @@ ASMCONSTANTS_C_ASSERT(OFFSETOF__CONTEXT__Xmm15
ASMCONSTANTS_C_ASSERT(OFFSETOF__CONTEXT__VectorRegister
== offsetof(CONTEXT, VectorRegister[0]));

// TBD: is this needed for unix?
#if defined(UNIX_AMD64_ABI) && !defined(HOST_WINDOWS)
#define OFFSETOF__CONTEXT__R16 (8*6 + 4*2 + 2*6 + 4 + 8*6 + 8*16 + 8 + 2*16 + 8*16 + 16*16 + 96 + 128*26 + 8*8 + 8*8*7 + 8*16*16 + 8*8*8 + 8*32*16 + 8*64*16)
ASMCONSTANTS_C_ASSERT(OFFSETOF__CONTEXT__R16
== offsetof(CONTEXT, R16));
#endif // UNIX_AMD64_ABI && !HOST_WINDOWS

#define SIZEOF__FaultingExceptionFrame (0x20 + SIZEOF__CONTEXT + 16)
ASMCONSTANTS_C_ASSERT(SIZEOF__FaultingExceptionFrame
== sizeof(FaultingExceptionFrame));
Expand Down
Loading
Loading