#include "llvm/Support/CommandLine.h"
#include "llvm/Target/TargetOptions.h"
#include "llvm/Support/Debug.h"
+#include <cstdlib>
using namespace llvm;
MFI->hasVarSizedObjects() ||
MFI->isFrameAddressTaken() || MFI->hasInlineAsmWithSPAdjust() ||
MF.getInfo<X86MachineFunctionInfo>()->getForceFramePointer() ||
- MMI.callsUnwindInit() || MMI.callsEHReturn());
+ MMI.callsUnwindInit() || MMI.callsEHReturn() ||
+ MFI->hasStackMap() || MFI->hasPatchPoint());
}
static unsigned getSUBriOpcode(unsigned IsLP64, int64_t Imm) {
static
void emitSPUpdate(MachineBasicBlock &MBB, MachineBasicBlock::iterator &MBBI,
unsigned StackPtr, int64_t NumBytes,
- bool Is64Bit, bool IsLP64, bool UseLEA,
+ bool Is64BitTarget, bool Is64BitStackPtr, bool UseLEA,
const TargetInstrInfo &TII, const TargetRegisterInfo &TRI) {
bool isSub = NumBytes < 0;
uint64_t Offset = isSub ? -NumBytes : NumBytes;
unsigned Opc;
if (UseLEA)
- Opc = getLEArOpcode(IsLP64);
+ Opc = getLEArOpcode(Is64BitStackPtr);
else
Opc = isSub
- ? getSUBriOpcode(IsLP64, Offset)
- : getADDriOpcode(IsLP64, Offset);
+ ? getSUBriOpcode(Is64BitStackPtr, Offset)
+ : getADDriOpcode(Is64BitStackPtr, Offset);
uint64_t Chunk = (1LL << 31) - 1;
DebugLoc DL = MBB.findDebugLoc(MBBI);
while (Offset) {
uint64_t ThisVal = (Offset > Chunk) ? Chunk : Offset;
- if (ThisVal == (Is64Bit ? 8 : 4)) {
+ if (ThisVal == (Is64BitTarget ? 8 : 4)) {
// Use push / pop instead.
unsigned Reg = isSub
- ? (unsigned)(Is64Bit ? X86::RAX : X86::EAX)
- : findDeadCallerSavedReg(MBB, MBBI, TRI, Is64Bit);
+ ? (unsigned)(Is64BitTarget ? X86::RAX : X86::EAX)
+ : findDeadCallerSavedReg(MBB, MBBI, TRI, Is64BitTarget);
if (Reg) {
Opc = isSub
- ? (Is64Bit ? X86::PUSH64r : X86::PUSH32r)
- : (Is64Bit ? X86::POP64r : X86::POP32r);
+ ? (Is64BitTarget ? X86::PUSH64r : X86::PUSH32r)
+ : (Is64BitTarget ? X86::POP64r : X86::POP32r);
MachineInstr *MI = BuildMI(MBB, MBBI, DL, TII.get(Opc))
.addReg(Reg, getDefRegState(!isSub) | getUndefRegState(isSub));
if (isSub)
return false;
}
+void X86FrameLowering::getStackProbeFunction(const X86Subtarget &STI,
+ unsigned &CallOp,
+ const char *&Symbol) {
+ CallOp = STI.is64Bit() ? X86::W64ALLOCA : X86::CALLpcrel32;
+
+ if (STI.is64Bit()) {
+ if (STI.isTargetCygMing()) {
+ Symbol = "___chkstk_ms";
+ } else {
+ Symbol = "__chkstk";
+ }
+ } else if (STI.isTargetCygMing())
+ Symbol = "_alloca";
+ else
+ Symbol = "_chkstk";
+}
+
/// emitPrologue - Push callee-saved registers onto the stack, which
/// automatically adjust the stack pointer. Adjust the stack pointer to allocate
/// space for local variables. Also emit labels used by the exception handler to
bool HasFP = hasFP(MF);
const X86Subtarget &STI = MF.getTarget().getSubtarget<X86Subtarget>();
bool Is64Bit = STI.is64Bit();
- bool IsLP64 = STI.isTarget64BitLP64();
+ // standard x86_64 and NaCl use 64-bit frame/stack pointers, x32 - 32-bit.
+ const bool Uses64BitFramePtr = STI.isTarget64BitLP64() || STI.isTargetNaCl64();
bool IsWin64 = STI.isTargetWin64();
bool IsWinEH =
MF.getTarget().getMCAsmInfo()->getExceptionHandlingType() ==
unsigned StackAlign = getStackAlignment();
unsigned SlotSize = RegInfo->getSlotSize();
unsigned FramePtr = RegInfo->getFrameRegister(MF);
+ const unsigned MachineFramePtr = STI.isTarget64BitILP32() ?
+ getX86SubSuperRegister(FramePtr, MVT::i64, false) : FramePtr;
unsigned StackPtr = RegInfo->getStackRegister();
unsigned BasePtr = RegInfo->getBaseRegister();
DebugLoc DL;
X86FI->setCalleeSavedFrameSize(
X86FI->getCalleeSavedFrameSize() - TailCallReturnAddrDelta);
+ bool UseStackProbe = (STI.isOSWindows() && !STI.isTargetMacho());
+
// If this is x86-64 and the Red Zone is not disabled, if we are a leaf
// function, and use up to 128 bytes of stack space, don't have a frame
// pointer, calls, or dynamic alloca then we do not need to adjust the
if (TailCallReturnAddrDelta < 0) {
MachineInstr *MI =
BuildMI(MBB, MBBI, DL,
- TII.get(getSUBriOpcode(IsLP64, -TailCallReturnAddrDelta)),
+ TII.get(getSUBriOpcode(Uses64BitFramePtr, -TailCallReturnAddrDelta)),
StackPtr)
.addReg(StackPtr)
.addImm(-TailCallReturnAddrDelta)
// Save EBP/RBP into the appropriate stack slot.
BuildMI(MBB, MBBI, DL, TII.get(Is64Bit ? X86::PUSH64r : X86::PUSH32r))
- .addReg(FramePtr, RegState::Kill)
+ .addReg(MachineFramePtr, RegState::Kill)
.setMIFlag(MachineInstr::FrameSetup);
if (NeedsDwarfCFI) {
.addCFIIndex(CFIIndex);
// Change the rule for the FramePtr to be an "offset" rule.
- unsigned DwarfFramePtr = RegInfo->getDwarfRegNum(FramePtr, true);
+ unsigned DwarfFramePtr = RegInfo->getDwarfRegNum(MachineFramePtr, true);
CFIIndex = MMI.addFrameInst(
MCCFIInstruction::createOffset(nullptr,
DwarfFramePtr, 2 * stackGrowth));
// Update EBP with the new base value.
BuildMI(MBB, MBBI, DL,
- TII.get(Is64Bit ? X86::MOV64rr : X86::MOV32rr), FramePtr)
+ TII.get(Uses64BitFramePtr ? X86::MOV64rr : X86::MOV32rr), FramePtr)
.addReg(StackPtr)
.setMIFlag(MachineInstr::FrameSetup);
if (NeedsDwarfCFI) {
// Mark effective beginning of when frame pointer becomes valid.
// Define the current CFA to use the EBP/RBP register.
- unsigned DwarfFramePtr = RegInfo->getDwarfRegNum(FramePtr, true);
+ unsigned DwarfFramePtr = RegInfo->getDwarfRegNum(MachineFramePtr, true);
unsigned CFIIndex = MMI.addFrameInst(
MCCFIInstruction::createDefCfaRegister(nullptr, DwarfFramePtr));
BuildMI(MBB, MBBI, DL, TII.get(TargetOpcode::CFI_INSTRUCTION))
// Mark the FramePtr as live-in in every block.
for (MachineFunction::iterator I = MF.begin(), E = MF.end(); I != E; ++I)
- I->addLiveIn(FramePtr);
+ I->addLiveIn(MachineFramePtr);
} else {
NumBytes = StackSize - X86FI->getCalleeSavedFrameSize();
}
assert(HasFP && "There should be a frame pointer if stack is realigned.");
MachineInstr *MI =
BuildMI(MBB, MBBI, DL,
- TII.get(Is64Bit ? X86::AND64ri32 : X86::AND32ri), StackPtr)
+ TII.get(Uses64BitFramePtr ? X86::AND64ri32 : X86::AND32ri), StackPtr)
.addReg(StackPtr)
.addImm(-MaxAlign)
.setMIFlag(MachineInstr::FrameSetup);
// Adjust stack pointer: ESP -= numbytes.
+ static const size_t PageSize = 4096;
+
// Windows and cygwin/mingw require a prologue helper routine when allocating
// more than 4K bytes on the stack. Windows uses __chkstk and cygwin/mingw
// uses __alloca. __alloca and the 32-bit version of __chkstk will probe the
// responsible for adjusting the stack pointer. Touching the stack at 4K
// increments is necessary to ensure that the guard pages used by the OS
// virtual memory manager are allocated in correct sequence.
- if (NumBytes >= 4096 && STI.isOSWindows() && !STI.isTargetMacho()) {
+ if (NumBytes >= PageSize && UseStackProbe) {
const char *StackProbeSymbol;
+ unsigned CallOp;
- if (Is64Bit) {
- if (STI.isTargetCygMing()) {
- StackProbeSymbol = "___chkstk_ms";
- } else {
- StackProbeSymbol = "__chkstk";
- }
- } else if (STI.isTargetCygMing())
- StackProbeSymbol = "_alloca";
- else
- StackProbeSymbol = "_chkstk";
+ getStackProbeFunction(STI, CallOp, StackProbeSymbol);
// Check whether EAX is livein for this function.
bool isEAXAlive = isEAXLiveIn(MF);
}
BuildMI(MBB, MBBI, DL,
- TII.get(Is64Bit ? X86::W64ALLOCA : X86::CALLpcrel32))
+ TII.get(CallOp))
.addExternalSymbol(StackProbeSymbol)
.addReg(StackPtr, RegState::Define | RegState::Implicit)
.addReg(X86::EFLAGS, RegState::Define | RegState::Implicit)
.setMIFlag(MachineInstr::FrameSetup);
}
if (isEAXAlive) {
- // Restore EAX
- MachineInstr *MI = addRegOffset(BuildMI(MF, DL, TII.get(X86::MOV32rm),
- X86::EAX),
- StackPtr, false, NumBytes - 4);
- MI->setFlag(MachineInstr::FrameSetup);
- MBB.insert(MBBI, MI);
+ // Restore EAX
+ MachineInstr *MI = addRegOffset(BuildMI(MF, DL, TII.get(X86::MOV32rm),
+ X86::EAX),
+ StackPtr, false, NumBytes - 4);
+ MI->setFlag(MachineInstr::FrameSetup);
+ MBB.insert(MBBI, MI);
}
} else if (NumBytes) {
- emitSPUpdate(MBB, MBBI, StackPtr, -(int64_t)NumBytes, Is64Bit, IsLP64,
+ emitSPUpdate(MBB, MBBI, StackPtr, -(int64_t)NumBytes, Is64Bit, Uses64BitFramePtr,
UseLEA, TII, *RegInfo);
}
// will restore SP to (BP - SEHFrameOffset)
for (const CalleeSavedInfo &Info : MFI->getCalleeSavedInfo()) {
int offset = MFI->getObjectOffset(Info.getFrameIdx());
- SEHFrameOffset = std::max(SEHFrameOffset, abs(offset));
+ SEHFrameOffset = std::max(SEHFrameOffset, std::abs(offset));
}
SEHFrameOffset += SEHFrameOffset % 16; // ensure alignmant
// to reference locals.
if (RegInfo->hasBasePointer(MF)) {
// Update the base pointer with the current stack pointer.
- unsigned Opc = Is64Bit ? X86::MOV64rr : X86::MOV32rr;
+ unsigned Opc = Uses64BitFramePtr ? X86::MOV64rr : X86::MOV32rr;
BuildMI(MBB, MBBI, DL, TII.get(Opc), BasePtr)
.addReg(StackPtr)
.setMIFlag(MachineInstr::FrameSetup);
DebugLoc DL = MBBI->getDebugLoc();
const X86Subtarget &STI = MF.getTarget().getSubtarget<X86Subtarget>();
bool Is64Bit = STI.is64Bit();
- bool IsLP64 = STI.isTarget64BitLP64();
+ // standard x86_64 and NaCl use 64-bit frame/stack pointers, x32 - 32-bit.
+ const bool Uses64BitFramePtr = STI.isTarget64BitLP64() || STI.isTargetNaCl64();
+ const bool Is64BitILP32 = STI.isTarget64BitILP32();
bool UseLEA = STI.useLeaForSP();
unsigned StackAlign = getStackAlignment();
unsigned SlotSize = RegInfo->getSlotSize();
unsigned FramePtr = RegInfo->getFrameRegister(MF);
+ unsigned MachineFramePtr = Is64BitILP32 ?
+ getX86SubSuperRegister(FramePtr, MVT::i64, false) : FramePtr;
unsigned StackPtr = RegInfo->getStackRegister();
bool IsWinEH =
// Pop EBP.
BuildMI(MBB, MBBI, DL,
- TII.get(Is64Bit ? X86::POP64r : X86::POP32r), FramePtr);
+ TII.get(Is64Bit ? X86::POP64r : X86::POP32r), MachineFramePtr);
} else {
NumBytes = StackSize - CSSize;
}
if (RegInfo->needsStackRealignment(MF))
MBBI = FirstCSPop;
if (CSSize != 0) {
- unsigned Opc = getLEArOpcode(IsLP64);
+ unsigned Opc = getLEArOpcode(Uses64BitFramePtr);
addRegOffset(BuildMI(MBB, MBBI, DL, TII.get(Opc), StackPtr),
FramePtr, false, -CSSize);
--MBBI;
} else {
- unsigned Opc = (Is64Bit ? X86::MOV64rr : X86::MOV32rr);
+ unsigned Opc = (Uses64BitFramePtr ? X86::MOV64rr : X86::MOV32rr);
BuildMI(MBB, MBBI, DL, TII.get(Opc), StackPtr)
.addReg(FramePtr);
--MBBI;
}
} else if (NumBytes) {
// Adjust stack pointer back: ESP += numbytes.
- emitSPUpdate(MBB, MBBI, StackPtr, NumBytes, Is64Bit, IsLP64, UseLEA,
+ emitSPUpdate(MBB, MBBI, StackPtr, NumBytes, Is64Bit, Uses64BitFramePtr, UseLEA,
TII, *RegInfo);
--MBBI;
}
MachineOperand &DestAddr = MBBI->getOperand(0);
assert(DestAddr.isReg() && "Offset should be in register!");
BuildMI(MBB, MBBI, DL,
- TII.get(Is64Bit ? X86::MOV64rr : X86::MOV32rr),
+ TII.get(Uses64BitFramePtr ? X86::MOV64rr : X86::MOV32rr),
StackPtr).addReg(DestAddr.getReg());
} else if (RetOpcode == X86::TCRETURNri || RetOpcode == X86::TCRETURNdi ||
RetOpcode == X86::TCRETURNmi ||
if (Offset) {
// Check for possible merge with preceding ADD instruction.
Offset += mergeSPUpdates(MBB, MBBI, StackPtr, true);
- emitSPUpdate(MBB, MBBI, StackPtr, Offset, Is64Bit, IsLP64,
+ emitSPUpdate(MBB, MBBI, StackPtr, Offset, Is64Bit, Uses64BitFramePtr,
UseLEA, TII, *RegInfo);
}
// Check for possible merge with preceding ADD instruction.
delta += mergeSPUpdates(MBB, MBBI, StackPtr, true);
- emitSPUpdate(MBB, MBBI, StackPtr, delta, Is64Bit, IsLP64, UseLEA, TII,
+ emitSPUpdate(MBB, MBBI, StackPtr, delta, Is64Bit, Uses64BitFramePtr, UseLEA, TII,
*RegInfo);
}
}
// about avoiding it later.
unsigned FPReg = RegInfo->getFrameRegister(MF);
for (unsigned i = 0; i < CSI.size(); ++i) {
- if (CSI[i].getReg() == FPReg) {
+ if (TRI->regsOverlap(CSI[i].getReg(),FPReg)) {
CSI.erase(CSI.begin() + i);
break;
}
const TargetRegisterClass *RC = RegInfo->getMinimalPhysRegClass(Reg);
// ensure alignment
- SpillSlotOffset -= abs(SpillSlotOffset) % RC->getAlignment();
+ SpillSlotOffset -= std::abs(SpillSlotOffset) % RC->getAlignment();
// spill into slot
SpillSlotOffset -= RC->getSize();
int SlotIndex =
/// and the properties of the function either one or two registers will be
/// needed. Set primary to true for the first register, false for the second.
static unsigned
-GetScratchRegister(bool Is64Bit, const MachineFunction &MF, bool Primary) {
+GetScratchRegister(bool Is64Bit, bool IsLP64, const MachineFunction &MF, bool Primary) {
CallingConv::ID CallingConvention = MF.getFunction()->getCallingConv();
// Erlang stuff.
return Primary ? X86::EBX : X86::EDI;
}
- if (Is64Bit)
- return Primary ? X86::R11 : X86::R12;
+ if (Is64Bit) {
+ if (IsLP64)
+ return Primary ? X86::R11 : X86::R12;
+ else
+ return Primary ? X86::R11D : X86::R12D;
+ }
bool IsNested = HasNestArgument(&MF);
uint64_t StackSize;
const X86Subtarget &STI = MF.getTarget().getSubtarget<X86Subtarget>();
bool Is64Bit = STI.is64Bit();
+ const bool IsLP64 = STI.isTarget64BitLP64();
unsigned TlsReg, TlsOffset;
DebugLoc DL;
- unsigned ScratchReg = GetScratchRegister(Is64Bit, MF, true);
+ unsigned ScratchReg = GetScratchRegister(Is64Bit, IsLP64, MF, true);
assert(!MF.getRegInfo().isLiveIn(ScratchReg) &&
"Scratch register is live-in");
}
if (IsNested)
- allocMBB->addLiveIn(X86::R10);
+ allocMBB->addLiveIn(IsLP64 ? X86::R10 : X86::R10D);
MF.push_front(allocMBB);
MF.push_front(checkMBB);
if (Is64Bit) {
if (STI.isTargetLinux()) {
TlsReg = X86::FS;
- TlsOffset = 0x70;
+ TlsOffset = IsLP64 ? 0x70 : 0x40;
} else if (STI.isTargetDarwin()) {
TlsReg = X86::GS;
TlsOffset = 0x60 + 90*8; // See pthread_machdep.h. Steal TLS slot 90.
}
if (CompareStackPointer)
- ScratchReg = X86::RSP;
+ ScratchReg = IsLP64 ? X86::RSP : X86::ESP;
else
- BuildMI(checkMBB, DL, TII.get(X86::LEA64r), ScratchReg).addReg(X86::RSP)
+ BuildMI(checkMBB, DL, TII.get(IsLP64 ? X86::LEA64r : X86::LEA64_32r), ScratchReg).addReg(X86::RSP)
.addImm(1).addReg(0).addImm(-StackSize).addReg(0);
- BuildMI(checkMBB, DL, TII.get(X86::CMP64rm)).addReg(ScratchReg)
+ BuildMI(checkMBB, DL, TII.get(IsLP64 ? X86::CMP64rm : X86::CMP32rm)).addReg(ScratchReg)
.addReg(0).addImm(1).addReg(0).addImm(TlsOffset).addReg(TlsReg);
} else {
if (STI.isTargetLinux()) {
bool SaveScratch2;
if (CompareStackPointer) {
// The primary scratch register is available for holding the TLS offset.
- ScratchReg2 = GetScratchRegister(Is64Bit, MF, true);
+ ScratchReg2 = GetScratchRegister(Is64Bit, IsLP64, MF, true);
SaveScratch2 = false;
} else {
// Need to use a second register to hold the TLS offset
- ScratchReg2 = GetScratchRegister(Is64Bit, MF, false);
+ ScratchReg2 = GetScratchRegister(Is64Bit, IsLP64, MF, false);
// Unfortunately, with fastcc the second scratch register may hold an
// argument.
// Functions with nested arguments use R10, so it needs to be saved across
// the call to _morestack
+ const unsigned RegAX = IsLP64 ? X86::RAX : X86::EAX;
+ const unsigned Reg10 = IsLP64 ? X86::R10 : X86::R10D;
+ const unsigned Reg11 = IsLP64 ? X86::R11 : X86::R11D;
+ const unsigned MOVrr = IsLP64 ? X86::MOV64rr : X86::MOV32rr;
+ const unsigned MOVri = IsLP64 ? X86::MOV64ri : X86::MOV32ri;
+
if (IsNested)
- BuildMI(allocMBB, DL, TII.get(X86::MOV64rr), X86::RAX).addReg(X86::R10);
+ BuildMI(allocMBB, DL, TII.get(MOVrr), RegAX).addReg(Reg10);
- BuildMI(allocMBB, DL, TII.get(X86::MOV64ri), X86::R10)
+ BuildMI(allocMBB, DL, TII.get(MOVri), Reg10)
.addImm(StackSize);
- BuildMI(allocMBB, DL, TII.get(X86::MOV64ri), X86::R11)
+ BuildMI(allocMBB, DL, TII.get(MOVri), Reg11)
.addImm(X86FI->getArgumentStackSize());
- MF.getRegInfo().setPhysRegUsed(X86::R10);
- MF.getRegInfo().setPhysRegUsed(X86::R11);
+ MF.getRegInfo().setPhysRegUsed(Reg10);
+ MF.getRegInfo().setPhysRegUsed(Reg11);
} else {
BuildMI(allocMBB, DL, TII.get(X86::PUSHi32))
.addImm(X86FI->getArgumentStackSize());
->getSlotSize();
const X86Subtarget &STI = MF.getTarget().getSubtarget<X86Subtarget>();
const bool Is64Bit = STI.is64Bit();
+ const bool IsLP64 = STI.isTarget64BitLP64();
DebugLoc DL;
// HiPE-specific values
const unsigned HipeLeafWords = 24;
SPLimitOffset = 0x4c;
}
- ScratchReg = GetScratchRegister(Is64Bit, MF, true);
+ ScratchReg = GetScratchRegister(Is64Bit, IsLP64, MF, true);
assert(!MF.getRegInfo().isLiveIn(ScratchReg) &&
"HiPE prologue scratch register is live-in");
projects / oota-llvm.git / blobdiff