#include "X86Subtarget.h"
#include "X86TargetMachine.h"
#include "llvm/ADT/SmallSet.h"
+#include "llvm/Analysis/LibCallSemantics.h"
#include "llvm/CodeGen/MachineFrameInfo.h"
#include "llvm/CodeGen/MachineFunction.h"
#include "llvm/CodeGen/MachineInstrBuilder.h"
#include "llvm/CodeGen/MachineModuleInfo.h"
#include "llvm/CodeGen/MachineRegisterInfo.h"
+#include "llvm/CodeGen/WinEHFuncInfo.h"
#include "llvm/IR/DataLayout.h"
#include "llvm/IR/Function.h"
#include "llvm/MC/MCAsmInfo.h"
#include "llvm/MC/MCSymbol.h"
-#include "llvm/Support/CommandLine.h"
#include "llvm/Target/TargetOptions.h"
#include "llvm/Support/Debug.h"
#include <cstdlib>
using namespace llvm;
-// FIXME: completely move here.
-extern cl::opt<bool> ForceStackAlign;
-
X86FrameLowering::X86FrameLowering(const X86Subtarget &STI,
unsigned StackAlignOverride)
: TargetFrameLowering(StackGrowsDown, StackAlignOverride,
STI.is64Bit() ? -8 : -4),
- STI(STI), TII(*STI.getInstrInfo()), RegInfo(STI.getRegisterInfo()) {
+ STI(STI), TII(*STI.getInstrInfo()), TRI(STI.getRegisterInfo()) {
// Cache a bunch of frame-related predicates for this subtarget.
- SlotSize = RegInfo->getSlotSize();
+ SlotSize = TRI->getSlotSize();
Is64Bit = STI.is64Bit();
IsLP64 = STI.isTarget64BitLP64();
// standard x86_64 and NaCl use 64-bit frame/stack pointers, x32 - 32-bit.
Uses64BitFramePtr = STI.isTarget64BitLP64() || STI.isTargetNaCl64();
- StackPtr = RegInfo->getStackRegister();
+ StackPtr = TRI->getStackRegister();
}
bool X86FrameLowering::hasReservedCallFrame(const MachineFunction &MF) const {
bool
X86FrameLowering::canSimplifyCallFramePseudos(const MachineFunction &MF) const {
return hasReservedCallFrame(MF) ||
- (hasFP(MF) && !RegInfo->needsStackRealignment(MF)) ||
- RegInfo->hasBasePointer(MF);
+ (hasFP(MF) && !TRI->needsStackRealignment(MF)) ||
+ TRI->hasBasePointer(MF);
}
// needsFrameIndexResolution - Do we need to perform FI resolution for
const MachineModuleInfo &MMI = MF.getMMI();
return (MF.getTarget().Options.DisableFramePointerElim(MF) ||
- RegInfo->needsStackRealignment(MF) ||
+ TRI->needsStackRealignment(MF) ||
MFI->hasVarSizedObjects() ||
- MFI->isFrameAddressTaken() || MFI->hasInlineAsmWithSPAdjust() ||
+ MFI->isFrameAddressTaken() || MFI->hasOpaqueSPAdjustment() ||
MF.getInfo<X86MachineFunctionInfo>()->getForceFramePointer() ||
- MMI.callsUnwindInit() || MMI.callsEHReturn() ||
+ MMI.callsUnwindInit() || MMI.hasEHFunclets() || MMI.callsEHReturn() ||
MFI->hasStackMap() || MFI->hasPatchPoint());
}
/// to this register without worry about clobbering it.
static unsigned findDeadCallerSavedReg(MachineBasicBlock &MBB,
MachineBasicBlock::iterator &MBBI,
- const TargetRegisterInfo *RegInfo,
+ const TargetRegisterInfo *TRI,
bool Is64Bit) {
const MachineFunction *MF = MBB.getParent();
const Function *F = MF->getFunction();
unsigned Reg = MO.getReg();
if (!Reg)
continue;
- for (MCRegAliasIterator AI(Reg, RegInfo, true); AI.isValid(); ++AI)
+ for (MCRegAliasIterator AI(Reg, TRI, true); AI.isValid(); ++AI)
Uses.insert(*AI);
}
if (isSub && !isEAXLiveIn(*MBB.getParent()))
Reg = (unsigned)(Is64Bit ? X86::RAX : X86::EAX);
else
- Reg = findDeadCallerSavedReg(MBB, MBBI, RegInfo, Is64Bit);
+ Reg = findDeadCallerSavedReg(MBB, MBBI, TRI, Is64Bit);
if (Reg) {
unsigned Opc = Is64Bit ? X86::MOV64ri : X86::MOV32ri;
// Use push / pop instead.
unsigned Reg = isSub
? (unsigned)(Is64Bit ? X86::RAX : X86::EAX)
- : findDeadCallerSavedReg(MBB, MBBI, RegInfo, Is64Bit);
+ : findDeadCallerSavedReg(MBB, MBBI, TRI, Is64Bit);
if (Reg) {
unsigned Opc = isSub
? (Is64Bit ? X86::PUSH64r : X86::PUSH32r)
.addReg(Reg, getDefRegState(!isSub) | getUndefRegState(isSub));
if (isSub)
MI->setFlag(MachineInstr::FrameSetup);
+ else
+ MI->setFlag(MachineInstr::FrameDestroy);
Offset -= ThisVal;
continue;
}
MBB, MBBI, DL, isSub ? -ThisVal : ThisVal, InEpilogue);
if (isSub)
MI.setMIFlag(MachineInstr::FrameSetup);
+ else
+ MI.setMIFlag(MachineInstr::FrameDestroy);
Offset -= ThisVal;
}
return MI;
}
-/// mergeSPUpdatesUp - Merge two stack-manipulating instructions upper iterator.
-static
-void mergeSPUpdatesUp(MachineBasicBlock &MBB, MachineBasicBlock::iterator &MBBI,
- unsigned StackPtr, uint64_t *NumBytes = nullptr) {
- if (MBBI == MBB.begin()) return;
-
- MachineBasicBlock::iterator PI = std::prev(MBBI);
- unsigned Opc = PI->getOpcode();
- if ((Opc == X86::ADD64ri32 || Opc == X86::ADD64ri8 ||
- Opc == X86::ADD32ri || Opc == X86::ADD32ri8 ||
- Opc == X86::LEA32r || Opc == X86::LEA64_32r) &&
- PI->getOperand(0).getReg() == StackPtr) {
- if (NumBytes)
- *NumBytes += PI->getOperand(2).getImm();
- MBB.erase(PI);
- } else if ((Opc == X86::SUB64ri32 || Opc == X86::SUB64ri8 ||
- Opc == X86::SUB32ri || Opc == X86::SUB32ri8) &&
- PI->getOperand(0).getReg() == StackPtr) {
- if (NumBytes)
- *NumBytes -= PI->getOperand(2).getImm();
- MBB.erase(PI);
- }
-}
-
int X86FrameLowering::mergeSPUpdates(MachineBasicBlock &MBB,
MachineBasicBlock::iterator &MBBI,
bool doMergeWithPrevious) const {
const MachineFrameInfo *MFI = MF.getFrameInfo();
uint64_t MaxAlign = MFI->getMaxAlignment(); // Desired stack alignment.
unsigned StackAlign = getStackAlignment();
- if (ForceStackAlign) {
+ if (MF.getFunction()->hasFnAttribute("stackrealign")) {
if (MFI->hasCalls())
MaxAlign = (StackAlign > MaxAlign) ? StackAlign : MaxAlign;
else if (MaxAlign < SlotSize)
X86MachineFunctionInfo *X86FI = MF.getInfo<X86MachineFunctionInfo>();
uint64_t MaxAlign = calculateMaxStackAlign(MF); // Desired stack alignment.
uint64_t StackSize = MFI->getStackSize(); // Number of bytes to allocate.
+ bool IsFunclet = MBB.isEHFuncletEntry();
+ bool IsClrFunclet =
+ IsFunclet &&
+ classifyEHPersonality(Fn->getPersonalityFn()) == EHPersonality::CoreCLR;
bool HasFP = hasFP(MF);
bool IsWin64CC = STI.isCallingConvWin64(Fn->getCallingConv());
bool IsWin64Prologue = MF.getTarget().getMCAsmInfo()->usesWindowsCFI();
bool NeedsWinCFI = IsWin64Prologue && Fn->needsUnwindTableEntry();
bool NeedsDwarfCFI =
!IsWin64Prologue && (MMI.hasDebugInfo() || Fn->needsUnwindTableEntry());
- unsigned FramePtr = RegInfo->getFrameRegister(MF);
+ unsigned FramePtr = TRI->getFrameRegister(MF);
const unsigned MachineFramePtr =
STI.isTarget64BitILP32()
- ? getX86SubSuperRegister(FramePtr, MVT::i64, false)
- : FramePtr;
- unsigned BasePtr = RegInfo->getBaseRegister();
- DebugLoc DL;
-
- // Add RETADDR move area to callee saved frame size.
+ ? getX86SubSuperRegister(FramePtr, MVT::i64, false)\r
+ : FramePtr;\r
+ unsigned BasePtr = TRI->getBaseRegister();\r
+ \r
+ // Debug location must be unknown since the first debug location is used\r
+ // to determine the end of the prologue.\r
+ DebugLoc DL;\r
+\r
+ // Add RETADDR move area to callee saved frame size.\r
int TailCallReturnAddrDelta = X86FI->getTCReturnAddrDelta();
if (TailCallReturnAddrDelta && IsWin64Prologue)
report_fatal_error("Can't handle guaranteed tail call under win64 yet");
// stack pointer (we fit in the Red Zone). We also check that we don't
// push and pop from the stack.
if (Is64Bit && !Fn->hasFnAttribute(Attribute::NoRedZone) &&
- !RegInfo->needsStackRealignment(MF) &&
+ !TRI->needsStackRealignment(MF) &&
!MFI->hasVarSizedObjects() && // No dynamic alloca.
!MFI->adjustsStack() && // No calls.
!IsWin64CC && // Win64 has no Red Zone
uint64_t NumBytes = 0;
int stackGrowth = -SlotSize;
+ // Find the funclet establisher parameter
+ unsigned Establisher = X86::NoRegister;
+ if (IsClrFunclet)
+ Establisher = Uses64BitFramePtr ? X86::RCX : X86::ECX;
+ else if (IsFunclet)
+ Establisher = Uses64BitFramePtr ? X86::RDX : X86::EDX;
+
+ if (IsWin64Prologue && IsFunclet & !IsClrFunclet) {
+ // Immediately spill establisher into the home slot.
+ // The runtime cares about this.
+ // MOV64mr %rdx, 16(%rsp)
+ unsigned MOVmr = Uses64BitFramePtr ? X86::MOV64mr : X86::MOV32mr;
+ addRegOffset(BuildMI(MBB, MBBI, DL, TII.get(MOVmr)), StackPtr, true, 16)
+ .addReg(Establisher)
+ .setMIFlag(MachineInstr::FrameSetup);
+ }
+
if (HasFP) {
// Calculate required stack adjustment.
uint64_t FrameSize = StackSize - SlotSize;
NumBytes = FrameSize - X86FI->getCalleeSavedFrameSize();
// Callee-saved registers are pushed on stack before the stack is realigned.
- if (RegInfo->needsStackRealignment(MF) && !IsWin64Prologue)
+ if (TRI->needsStackRealignment(MF) && !IsWin64Prologue)
NumBytes = RoundUpToAlignment(NumBytes, MaxAlign);
// Get the offset of the stack slot for the EBP register, which is
// guaranteed to be the last slot by processFunctionBeforeFrameFinalized.
// Update the frame offset adjustment.
- MFI->setOffsetAdjustment(-NumBytes);
+ if (!IsFunclet)
+ MFI->setOffsetAdjustment(-NumBytes);
+ else
+ assert(MFI->getOffsetAdjustment() == -(int)NumBytes &&
+ "should calculate same local variable offset for funclets");
// Save EBP/RBP into the appropriate stack slot.
BuildMI(MBB, MBBI, DL, TII.get(Is64Bit ? X86::PUSH64r : X86::PUSH32r))
MCCFIInstruction::createDefCfaOffset(nullptr, 2 * stackGrowth));
// Change the rule for the FramePtr to be an "offset" rule.
- unsigned DwarfFramePtr = RegInfo->getDwarfRegNum(MachineFramePtr, true);
+ unsigned DwarfFramePtr = TRI->getDwarfRegNum(MachineFramePtr, true);
BuildCFI(MBB, MBBI, DL, MCCFIInstruction::createOffset(
nullptr, DwarfFramePtr, 2 * stackGrowth));
}
.setMIFlag(MachineInstr::FrameSetup);
}
- if (!IsWin64Prologue) {
+ if (!IsWin64Prologue && !IsFunclet) {
// Update EBP with the new base value.
BuildMI(MBB, MBBI, DL,
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(MachineFramePtr, true);
- BuildCFI(MBB, MBBI, DL,
- MCCFIInstruction::createDefCfaRegister(nullptr, DwarfFramePtr));
+ if (NeedsDwarfCFI) {
+ // Mark effective beginning of when frame pointer becomes valid.
+ // Define the current CFA to use the EBP/RBP register.
+ unsigned DwarfFramePtr = TRI->getDwarfRegNum(MachineFramePtr, true);
+ BuildCFI(MBB, MBBI, DL, MCCFIInstruction::createDefCfaRegister(
+ nullptr, DwarfFramePtr));
+ }
}
// Mark the FramePtr as live-in in every block.
for (MachineFunction::iterator I = MF.begin(), E = MF.end(); I != E; ++I)
I->addLiveIn(MachineFramePtr);
} else {
+ assert(!IsFunclet && "funclets without FPs not yet implemented");
NumBytes = StackSize - X86FI->getCalleeSavedFrameSize();
}
+ // For EH funclets, only allocate enough space for outgoing calls. Save the
+ // NumBytes value that we would've used for the parent frame.
+ unsigned ParentFrameNumBytes = NumBytes;
+ if (IsFunclet)
+ NumBytes = getWinEHFuncletFrameSize(MF);
+
// Skip the callee-saved push instructions.
bool PushedRegs = false;
int StackOffset = 2 * stackGrowth;
while (MBBI != MBB.end() &&
+ MBBI->getFlag(MachineInstr::FrameSetup) &&
(MBBI->getOpcode() == X86::PUSH32r ||
MBBI->getOpcode() == X86::PUSH64r)) {
PushedRegs = true;
// Realign stack after we pushed callee-saved registers (so that we'll be
// able to calculate their offsets from the frame pointer).
// Don't do this for Win64, it needs to realign the stack after the prologue.
- if (!IsWin64Prologue && RegInfo->needsStackRealignment(MF)) {
+ if (!IsWin64Prologue && !IsFunclet && TRI->needsStackRealignment(MF)) {
assert(HasFP && "There should be a frame pointer if stack is realigned.");
BuildStackAlignAND(MBB, MBBI, DL, MaxAlign);
}
// increments is necessary to ensure that the guard pages used by the OS
// virtual memory manager are allocated in correct sequence.
uint64_t AlignedNumBytes = NumBytes;
- if (IsWin64Prologue && RegInfo->needsStackRealignment(MF))
+ if (IsWin64Prologue && !IsFunclet && TRI->needsStackRealignment(MF))
AlignedNumBytes = RoundUpToAlignment(AlignedNumBytes, MaxAlign);
if (AlignedNumBytes >= StackProbeSize && UseStackProbe) {
// Check whether EAX is livein for this function.
int SEHFrameOffset = 0;
if (IsWin64Prologue && HasFP) {
- SEHFrameOffset = calculateSetFPREG(NumBytes);
+ // Set RBP to a small fixed offset from RSP. In the funclet case, we base
+ // this calculation on the incoming establisher, which holds the value of
+ // RSP from the parent frame at the end of the prologue.
+ unsigned SPOrEstablisher = IsFunclet ? Establisher : StackPtr;
+ SEHFrameOffset = calculateSetFPREG(ParentFrameNumBytes);
if (SEHFrameOffset)
addRegOffset(BuildMI(MBB, MBBI, DL, TII.get(X86::LEA64r), FramePtr),
- StackPtr, false, SEHFrameOffset);
+ SPOrEstablisher, false, SEHFrameOffset);
else
- BuildMI(MBB, MBBI, DL, TII.get(X86::MOV64rr), FramePtr).addReg(StackPtr);
+ BuildMI(MBB, MBBI, DL, TII.get(X86::MOV64rr), FramePtr)
+ .addReg(SPOrEstablisher);
- if (NeedsWinCFI)
+ // If this is not a funclet, emit the CFI describing our frame pointer.
+ if (NeedsWinCFI && !IsFunclet)
BuildMI(MBB, MBBI, DL, TII.get(X86::SEH_SetFrame))
.addImm(FramePtr)
.addImm(SEHFrameOffset)
.setMIFlag(MachineInstr::FrameSetup);
+ } else if (IsFunclet && STI.is32Bit()) {
+ // Reset EBP / ESI to something good for funclets.
+ MBBI = restoreWin32EHStackPointers(MBB, MBBI, DL);
}
while (MBBI != MBB.end() && MBBI->getFlag(MachineInstr::FrameSetup)) {
int FI;
if (unsigned Reg = TII.isStoreToStackSlot(FrameInstr, FI)) {
if (X86::FR64RegClass.contains(Reg)) {
- int Offset = getFrameIndexOffset(MF, FI);
+ unsigned IgnoredFrameReg;
+ int Offset = getFrameIndexReference(MF, FI, IgnoredFrameReg);
Offset += SEHFrameOffset;
BuildMI(MBB, MBBI, DL, TII.get(X86::SEH_SaveXMM))
// Realign stack after we spilled callee-saved registers (so that we'll be
// able to calculate their offsets from the frame pointer).
// Win64 requires aligning the stack after the prologue.
- if (IsWin64Prologue && RegInfo->needsStackRealignment(MF)) {
+ if (IsWin64Prologue && TRI->needsStackRealignment(MF)) {
assert(HasFP && "There should be a frame pointer if stack is realigned.");
BuildStackAlignAND(MBB, MBBI, DL, MaxAlign);
}
// of the stack pointer is at this point. Any variable size objects
// will be allocated after this, so we can still use the base pointer
// to reference locals.
- if (RegInfo->hasBasePointer(MF)) {
+ if (TRI->hasBasePointer(MF)) {
// Update the base pointer with the current stack pointer.
unsigned Opc = Uses64BitFramePtr ? X86::MOV64rr : X86::MOV32rr;
BuildMI(MBB, MBBI, DL, TII.get(Opc), BasePtr)
.addReg(StackPtr)
.setMIFlag(MachineInstr::FrameSetup);
if (X86FI->getRestoreBasePointer()) {
- // Stash value of base pointer. Saving RSP instead of EBP shortens dependence chain.
+ // Stash value of base pointer. Saving RSP instead of EBP shortens
+ // dependence chain. Used by SjLj EH.
unsigned Opm = Uses64BitFramePtr ? X86::MOV64mr : X86::MOV32mr;
addRegOffset(BuildMI(MBB, MBBI, DL, TII.get(Opm)),
FramePtr, true, X86FI->getRestoreBasePointerOffset())
.addReg(StackPtr)
.setMIFlag(MachineInstr::FrameSetup);
}
+
+ if (X86FI->getHasSEHFramePtrSave()) {
+ // Stash the value of the frame pointer relative to the base pointer for
+ // Win32 EH. This supports Win32 EH, which does the inverse of the above:
+ // it recovers the frame pointer from the base pointer rather than the
+ // other way around.
+ unsigned Opm = Uses64BitFramePtr ? X86::MOV64mr : X86::MOV32mr;
+ unsigned UsedReg;
+ int Offset =
+ getFrameIndexReference(MF, X86FI->getSEHFramePtrSaveIndex(), UsedReg);
+ assert(UsedReg == BasePtr);
+ addRegOffset(BuildMI(MBB, MBBI, DL, TII.get(Opm)), UsedReg, true, Offset)
+ .addReg(FramePtr)
+ .setMIFlag(MachineInstr::FrameSetup);
+ }
}
if (((!HasFP && NumBytes) || PushedRegs) && NeedsDwarfCFI) {
return !MF.getTarget().getMCAsmInfo()->usesWindowsCFI() || hasFP(MF);
}
+static bool isFuncletReturnInstr(MachineInstr *MI) {
+ switch (MI->getOpcode()) {
+ case X86::CATCHRET:
+ case X86::CLEANUPRET:
+ return true;
+ default:
+ return false;
+ }
+ llvm_unreachable("impossible");
+}
+
+unsigned X86FrameLowering::getWinEHFuncletFrameSize(const MachineFunction &MF) const {
+ // This is the size of the pushed CSRs.
+ unsigned CSSize =
+ MF.getInfo<X86MachineFunctionInfo>()->getCalleeSavedFrameSize();
+ // This is the amount of stack a funclet needs to allocate.
+ unsigned MaxCallSize = MF.getFrameInfo()->getMaxCallFrameSize();
+ // RBP is not included in the callee saved register block. After pushing RBP,
+ // everything is 16 byte aligned. Everything we allocate before an outgoing
+ // call must also be 16 byte aligned.
+ unsigned FrameSizeMinusRBP =
+ RoundUpToAlignment(CSSize + MaxCallSize, getStackAlignment());
+ // Subtract out the size of the callee saved registers. This is how much stack
+ // each funclet will allocate.
+ return FrameSizeMinusRBP - CSSize;
+}
+
void X86FrameLowering::emitEpilogue(MachineFunction &MF,
MachineBasicBlock &MBB) const {
const MachineFrameInfo *MFI = MF.getFrameInfo();
DL = MBBI->getDebugLoc();
// standard x86_64 and NaCl use 64-bit frame/stack pointers, x32 - 32-bit.
const bool Is64BitILP32 = STI.isTarget64BitILP32();
- unsigned FramePtr = RegInfo->getFrameRegister(MF);
+ unsigned FramePtr = TRI->getFrameRegister(MF);
unsigned MachineFramePtr =
Is64BitILP32 ? getX86SubSuperRegister(FramePtr, MVT::i64, false)
: FramePtr;
bool IsWin64Prologue = MF.getTarget().getMCAsmInfo()->usesWindowsCFI();
bool NeedsWinCFI =
IsWin64Prologue && MF.getFunction()->needsUnwindTableEntry();
+ bool IsFunclet = isFuncletReturnInstr(MBBI);
+ MachineBasicBlock *RestoreMBB = nullptr;
// Get the number of bytes to allocate from the FrameInfo.
uint64_t StackSize = MFI->getStackSize();
unsigned CSSize = X86FI->getCalleeSavedFrameSize();
uint64_t NumBytes = 0;
- if (hasFP(MF)) {
+ if (MBBI->getOpcode() == X86::CATCHRET) {
+ NumBytes = getWinEHFuncletFrameSize(MF);
+ assert(hasFP(MF) && "EH funclets without FP not yet implemented");
+ MachineBasicBlock *TargetMBB = MBBI->getOperand(0).getMBB();
+
+ // If this is SEH, this isn't really a funclet return.
+ bool IsSEH = isAsynchronousEHPersonality(
+ classifyEHPersonality(MF.getFunction()->getPersonalityFn()));
+ if (IsSEH) {
+ if (STI.is32Bit())
+ restoreWin32EHStackPointers(MBB, MBBI, DL, /*RestoreSP=*/true);
+ BuildMI(MBB, MBBI, DL, TII.get(X86::JMP_4)).addMBB(TargetMBB);
+ MBBI->eraseFromParent();
+ return;
+ }
+
+ // For 32-bit, create a new block for the restore code.
+ RestoreMBB = TargetMBB;
+ if (STI.is32Bit()) {
+ RestoreMBB = MF.CreateMachineBasicBlock(MBB.getBasicBlock());
+ MF.insert(TargetMBB->getIterator(), RestoreMBB);
+ MBB.removeSuccessor(TargetMBB);
+ MBB.addSuccessor(RestoreMBB);
+ RestoreMBB->addSuccessor(TargetMBB);
+ MBBI->getOperand(0).setMBB(RestoreMBB);
+ }
+
+ // Pop EBP.
+ BuildMI(MBB, MBBI, DL, TII.get(Is64Bit ? X86::POP64r : X86::POP32r),
+ MachineFramePtr)
+ .setMIFlag(MachineInstr::FrameDestroy);
+
+ // Insert frame restoration code in a new block.
+ if (STI.is32Bit()) {
+ auto RestoreMBBI = RestoreMBB->begin();
+ restoreWin32EHStackPointers(*RestoreMBB, RestoreMBBI, DL,
+ /*RestoreSP=*/true);
+ BuildMI(*RestoreMBB, RestoreMBBI, DL, TII.get(X86::JMP_4))
+ .addMBB(TargetMBB);
+ }
+ } else if (MBBI->getOpcode() == X86::CLEANUPRET) {
+ NumBytes = getWinEHFuncletFrameSize(MF);
+ assert(hasFP(MF) && "EH funclets without FP not yet implemented");
+ BuildMI(MBB, MBBI, DL, TII.get(Is64Bit ? X86::POP64r : X86::POP32r),
+ MachineFramePtr)
+ .setMIFlag(MachineInstr::FrameDestroy);
+ } else if (hasFP(MF)) {
// Calculate required stack adjustment.
uint64_t FrameSize = StackSize - SlotSize;
NumBytes = FrameSize - CSSize;
// Callee-saved registers were pushed on stack before the stack was
// realigned.
- if (RegInfo->needsStackRealignment(MF) && !IsWin64Prologue)
+ if (TRI->needsStackRealignment(MF) && !IsWin64Prologue)
NumBytes = RoundUpToAlignment(FrameSize, MaxAlign);
// Pop EBP.
BuildMI(MBB, MBBI, DL,
- TII.get(Is64Bit ? X86::POP64r : X86::POP32r), MachineFramePtr);
+ TII.get(Is64Bit ? X86::POP64r : X86::POP32r), MachineFramePtr)
+ .setMIFlag(MachineInstr::FrameDestroy);
} else {
NumBytes = StackSize - CSSize;
}
MachineBasicBlock::iterator PI = std::prev(MBBI);
unsigned Opc = PI->getOpcode();
- if (Opc != X86::POP32r && Opc != X86::POP64r && Opc != X86::DBG_VALUE &&
- !PI->isTerminator())
+ if ((Opc != X86::POP32r || !PI->getFlag(MachineInstr::FrameDestroy)) &&
+ (Opc != X86::POP64r || !PI->getFlag(MachineInstr::FrameDestroy)) &&
+ Opc != X86::DBG_VALUE && !PI->isTerminator())
break;
--MBBI;
}
MachineBasicBlock::iterator FirstCSPop = MBBI;
+ if (RestoreMBB) {
+ // Fill EAX/RAX with the address of the target block.
+ unsigned ReturnReg = STI.is64Bit() ? X86::RAX : X86::EAX;
+ if (STI.is64Bit()) {
+ // LEA64r RestoreMBB(%rip), %rax
+ BuildMI(MBB, FirstCSPop, DL, TII.get(X86::LEA64r), ReturnReg)
+ .addReg(X86::RIP)
+ .addImm(0)
+ .addReg(0)
+ .addMBB(RestoreMBB)
+ .addReg(0);
+ } else {
+ // MOV32ri $RestoreMBB, %eax
+ BuildMI(MBB, FirstCSPop, DL, TII.get(X86::MOV32ri))
+ .addReg(ReturnReg)
+ .addMBB(RestoreMBB);
+ }
+ // Record that we've taken the address of RestoreMBB and no longer just
+ // reference it in a terminator.
+ RestoreMBB->setHasAddressTaken();
+ }
+
if (MBBI != MBB.end())
DL = MBBI->getDebugLoc();
// If there is an ADD32ri or SUB32ri of ESP immediately before this
// instruction, merge the two instructions.
if (NumBytes || MFI->hasVarSizedObjects())
- mergeSPUpdatesUp(MBB, MBBI, StackPtr, &NumBytes);
+ NumBytes += mergeSPUpdates(MBB, MBBI, true);
// If dynamic alloca is used, then reset esp to point to the last callee-saved
// slot before popping them off! Same applies for the case, when stack was
- // realigned.
- if (RegInfo->needsStackRealignment(MF) || MFI->hasVarSizedObjects()) {
- if (RegInfo->needsStackRealignment(MF))
+ // realigned. Don't do this if this was a funclet epilogue, since the funclets
+ // will not do realignment or dynamic stack allocation.
+ if ((TRI->needsStackRealignment(MF) || MFI->hasVarSizedObjects()) &&
+ !IsFunclet) {
+ if (TRI->needsStackRealignment(MF))
MBBI = FirstCSPop;
unsigned SEHFrameOffset = calculateSetFPREG(SEHStackAllocAmt);
uint64_t LEAAmount =
}
}
-int X86FrameLowering::getFrameIndexOffset(const MachineFunction &MF,
- int FI) const {
+// NOTE: this only has a subset of the full frame index logic. In
+// particular, the FI < 0 and AfterFPPop logic is handled in
+// X86RegisterInfo::eliminateFrameIndex, but not here. Possibly
+// (probably?) it should be moved into here.
+int X86FrameLowering::getFrameIndexReference(const MachineFunction &MF, int FI,
+ unsigned &FrameReg) const {
const MachineFrameInfo *MFI = MF.getFrameInfo();
+
+ // We can't calculate offset from frame pointer if the stack is realigned,
+ // so enforce usage of stack/base pointer. The base pointer is used when we
+ // have dynamic allocas in addition to dynamic realignment.
+ if (TRI->hasBasePointer(MF))
+ FrameReg = TRI->getBaseRegister();
+ else if (TRI->needsStackRealignment(MF))
+ FrameReg = TRI->getStackRegister();
+ else
+ FrameReg = TRI->getFrameRegister(MF);
+
// Offset will hold the offset from the stack pointer at function entry to the
// object.
// We need to factor in additional offsets applied during the prologue to the
}
- if (RegInfo->hasBasePointer(MF)) {
+ if (TRI->hasBasePointer(MF)) {
assert(HasFP && "VLAs and dynamic stack realign, but no FP?!");
if (FI < 0) {
// Skip the saved EBP.
assert((-(Offset + StackSize)) % MFI->getObjectAlignment(FI) == 0);
return Offset + StackSize;
}
- } else if (RegInfo->needsStackRealignment(MF)) {
+ } else if (TRI->needsStackRealignment(MF)) {
if (FI < 0) {
// Skip the saved EBP.
return Offset + SlotSize + FPDelta;
return Offset + FPDelta;
}
-int X86FrameLowering::getFrameIndexReference(const MachineFunction &MF, int FI,
- unsigned &FrameReg) const {
- // We can't calculate offset from frame pointer if the stack is realigned,
- // so enforce usage of stack/base pointer. The base pointer is used when we
- // have dynamic allocas in addition to dynamic realignment.
- if (RegInfo->hasBasePointer(MF))
- FrameReg = RegInfo->getBaseRegister();
- else if (RegInfo->needsStackRealignment(MF))
- FrameReg = RegInfo->getStackRegister();
- else
- FrameReg = RegInfo->getFrameRegister(MF);
- return getFrameIndexOffset(MF, FI);
-}
-
-// Simplified from getFrameIndexOffset keeping only StackPointer cases
-int X86FrameLowering::getFrameIndexOffsetFromSP(const MachineFunction &MF, int FI) const {
+// Simplified from getFrameIndexReference keeping only StackPointer cases
+int X86FrameLowering::getFrameIndexReferenceFromSP(const MachineFunction &MF,
+ int FI,
+ unsigned &FrameReg) const {
const MachineFrameInfo *MFI = MF.getFrameInfo();
// Does not include any dynamic realign.
const uint64_t StackSize = MFI->getStackSize();
// frame). As a result, THE RESULT OF THIS CALL IS MEANINGLESS FOR CSRs
// AND FixedObjects IFF needsStackRealignment or hasVarSizedObject.
- assert(!RegInfo->hasBasePointer(MF) && "we don't handle this case");
+ assert(!TRI->hasBasePointer(MF) && "we don't handle this case");
// We don't handle tail calls, and shouldn't be seeing them
// either.
#endif
}
+ // Fill in FrameReg output argument.
+ FrameReg = TRI->getStackRegister();
+
// This is how the math works out:
//
// %rsp grows (i.e. gets lower) left to right. Each box below is
return Offset + StackSize;
}
-// Simplified from getFrameIndexReference keeping only StackPointer cases
-int X86FrameLowering::getFrameIndexReferenceFromSP(const MachineFunction &MF,
- int FI,
- unsigned &FrameReg) const {
- assert(!RegInfo->hasBasePointer(MF) && "we don't handle this case");
-
- FrameReg = RegInfo->getStackRegister();
- return getFrameIndexOffsetFromSP(MF, FI);
-}
bool X86FrameLowering::assignCalleeSavedSpillSlots(
MachineFunction &MF, const TargetRegisterInfo *TRI,
// Since emitPrologue and emitEpilogue will handle spilling and restoring of
// the frame register, we can delete it from CSI list and not have to worry
// about avoiding it later.
- unsigned FPReg = RegInfo->getFrameRegister(MF);
+ unsigned FPReg = TRI->getFrameRegister(MF);
for (unsigned i = 0; i < CSI.size(); ++i) {
if (TRI->regsOverlap(CSI[i].getReg(),FPReg)) {
CSI.erase(CSI.begin() + i);
if (X86::GR64RegClass.contains(Reg) || X86::GR32RegClass.contains(Reg))
continue;
- const TargetRegisterClass *RC = RegInfo->getMinimalPhysRegClass(Reg);
+ const TargetRegisterClass *RC = TRI->getMinimalPhysRegClass(Reg);
// ensure alignment
SpillSlotOffset -= std::abs(SpillSlotOffset) % RC->getAlignment();
// spill into slot
const TargetRegisterInfo *TRI) const {
DebugLoc DL = MBB.findDebugLoc(MI);
+ // Don't save CSRs in 32-bit EH funclets. The caller saves EBX, EBP, ESI, EDI
+ // for us, and there are no XMM CSRs on Win32.
+ if (MBB.isEHFuncletEntry() && STI.is32Bit() && STI.isOSWindows())
+ return true;
+
// Push GPRs. It increases frame size.
unsigned Opc = STI.is64Bit() ? X86::PUSH64r : X86::PUSH32r;
for (unsigned i = CSI.size(); i != 0; --i) {
if (CSI.empty())
return false;
+ if (isFuncletReturnInstr(MI) && STI.isOSWindows()) {
+ // Don't restore CSRs in 32-bit EH funclets. Matches
+ // spillCalleeSavedRegisters.
+ if (STI.is32Bit())
+ return true;
+ // Don't restore CSRs before an SEH catchret. SEH except blocks do not form
+ // funclets. emitEpilogue transforms these to normal jumps.
+ if (MI->getOpcode() == X86::CATCHRET) {
+ const Function *Func = MBB.getParent()->getFunction();
+ bool IsSEH = isAsynchronousEHPersonality(
+ classifyEHPersonality(Func->getPersonalityFn()));
+ if (IsSEH)
+ return true;
+ }
+ }
+
DebugLoc DL = MBB.findDebugLoc(MI);
// Reload XMMs from stack frame.
!X86::GR32RegClass.contains(Reg))
continue;
- BuildMI(MBB, MI, DL, TII.get(Opc), Reg);
+ BuildMI(MBB, MI, DL, TII.get(Opc), Reg)
+ .setMIFlag(MachineInstr::FrameDestroy);
}
return true;
}
-void
-X86FrameLowering::processFunctionBeforeCalleeSavedScan(MachineFunction &MF,
- RegScavenger *RS) const {
+void X86FrameLowering::determineCalleeSaves(MachineFunction &MF,
+ BitVector &SavedRegs,
+ RegScavenger *RS) const {
+ TargetFrameLowering::determineCalleeSaves(MF, SavedRegs, RS);
+
MachineFrameInfo *MFI = MF.getFrameInfo();
X86MachineFunctionInfo *X86FI = MF.getInfo<X86MachineFunctionInfo>();
}
// Spill the BasePtr if it's used.
- if (RegInfo->hasBasePointer(MF))
- MF.getRegInfo().setPhysRegUsed(RegInfo->getBaseRegister());
+ if (TRI->hasBasePointer(MF)) {
+ SavedRegs.set(TRI->getBaseRegister());
+
+ // Allocate a spill slot for EBP if we have a base pointer and EH funclets.
+ if (MF.getMMI().hasEHFunclets()) {
+ int FI = MFI->CreateSpillStackObject(SlotSize, SlotSize);
+ X86FI->setHasSEHFramePtrSave(true);
+ X86FI->setSEHFramePtrSaveIndex(FI);
+ }
+ }
}
static bool
// The MOV R10, RAX needs to be in a different block, since the RET we emit in
// allocMBB needs to be last (terminating) instruction.
- for (MachineBasicBlock::livein_iterator i = PrologueMBB.livein_begin(),
- e = PrologueMBB.livein_end();
- i != e; i++) {
- allocMBB->addLiveIn(*i);
- checkMBB->addLiveIn(*i);
+ for (const auto &LI : PrologueMBB.liveins()) {
+ allocMBB->addLiveIn(LI);
+ checkMBB->addLiveIn(LI);
}
if (IsNested)
.addImm(StackSize);
BuildMI(allocMBB, DL, TII.get(MOVri), Reg11)
.addImm(X86FI->getArgumentStackSize());
- MF.getRegInfo().setPhysRegUsed(Reg10);
- MF.getRegInfo().setPhysRegUsed(Reg11);
} else {
BuildMI(allocMBB, DL, TII.get(X86::PUSHi32))
.addImm(X86FI->getArgumentStackSize());
MachineBasicBlock *stackCheckMBB = MF.CreateMachineBasicBlock();
MachineBasicBlock *incStackMBB = MF.CreateMachineBasicBlock();
- for (MachineBasicBlock::livein_iterator I = PrologueMBB.livein_begin(),
- E = PrologueMBB.livein_end();
- I != E; I++) {
- stackCheckMBB->addLiveIn(*I);
- incStackMBB->addLiveIn(*I);
+ for (const auto &LI : PrologueMBB.liveins()) {
+ stackCheckMBB->addLiveIn(LI);
+ incStackMBB->addLiveIn(LI);
}
MF.push_front(incStackMBB);
#endif
}
+bool X86FrameLowering::adjustStackWithPops(MachineBasicBlock &MBB,
+ MachineBasicBlock::iterator MBBI, DebugLoc DL, int Offset) const {
+
+ if (Offset <= 0)
+ return false;
+
+ if (Offset % SlotSize)
+ return false;
+
+ int NumPops = Offset / SlotSize;
+ // This is only worth it if we have at most 2 pops.
+ if (NumPops != 1 && NumPops != 2)
+ return false;
+
+ // Handle only the trivial case where the adjustment directly follows
+ // a call. This is the most common one, anyway.
+ if (MBBI == MBB.begin())
+ return false;
+ MachineBasicBlock::iterator Prev = std::prev(MBBI);
+ if (!Prev->isCall() || !Prev->getOperand(1).isRegMask())
+ return false;
+
+ unsigned Regs[2];
+ unsigned FoundRegs = 0;
+
+ auto RegMask = Prev->getOperand(1);
+
+ auto &RegClass =
+ Is64Bit ? X86::GR64_NOREX_NOSPRegClass : X86::GR32_NOREX_NOSPRegClass;
+ // Try to find up to NumPops free registers.
+ for (auto Candidate : RegClass) {
+
+ // Poor man's liveness:
+ // Since we're immediately after a call, any register that is clobbered
+ // by the call and not defined by it can be considered dead.
+ if (!RegMask.clobbersPhysReg(Candidate))
+ continue;
+
+ bool IsDef = false;
+ for (const MachineOperand &MO : Prev->implicit_operands()) {
+ if (MO.isReg() && MO.isDef() && MO.getReg() == Candidate) {
+ IsDef = true;
+ break;
+ }
+ }
+
+ if (IsDef)
+ continue;
+
+ Regs[FoundRegs++] = Candidate;
+ if (FoundRegs == (unsigned)NumPops)
+ break;
+ }
+
+ if (FoundRegs == 0)
+ return false;
+
+ // If we found only one free register, but need two, reuse the same one twice.
+ while (FoundRegs < (unsigned)NumPops)
+ Regs[FoundRegs++] = Regs[0];
+
+ for (int i = 0; i < NumPops; ++i)
+ BuildMI(MBB, MBBI, DL,
+ TII.get(STI.is64Bit() ? X86::POP64r : X86::POP32r), Regs[i]);
+
+ return true;
+}
+
void X86FrameLowering::
eliminateCallFramePseudoInstr(MachineFunction &MF, MachineBasicBlock &MBB,
MachineBasicBlock::iterator I) const {
// If the stack pointer can be changed after prologue, turn the
// adjcallstackup instruction into a 'sub ESP, <amt>' and the
// adjcallstackdown instruction into 'add ESP, <amt>'
- if (Amount == 0)
- return;
// We need to keep the stack aligned properly. To do this, we round the
// amount of space needed for the outgoing arguments up to the next
unsigned StackAlign = getStackAlignment();
Amount = RoundUpToAlignment(Amount, StackAlign);
+ MachineModuleInfo &MMI = MF.getMMI();
+ const Function *Fn = MF.getFunction();
+ bool WindowsCFI = MF.getTarget().getMCAsmInfo()->usesWindowsCFI();
+ bool DwarfCFI = !WindowsCFI &&
+ (MMI.hasDebugInfo() || Fn->needsUnwindTableEntry());
+
+ // If we have any exception handlers in this function, and we adjust
+ // the SP before calls, we may need to indicate this to the unwinder
+ // using GNU_ARGS_SIZE. Note that this may be necessary even when
+ // Amount == 0, because the preceding function may have set a non-0
+ // GNU_ARGS_SIZE.
+ // TODO: We don't need to reset this between subsequent functions,
+ // if it didn't change.
+ bool HasDwarfEHHandlers = !WindowsCFI &&
+ !MF.getMMI().getLandingPads().empty();
+
+ if (HasDwarfEHHandlers && !isDestroy &&
+ MF.getInfo<X86MachineFunctionInfo>()->getHasPushSequences())
+ BuildCFI(MBB, I, DL,
+ MCCFIInstruction::createGnuArgsSize(nullptr, Amount));
+
+ if (Amount == 0)
+ return;
+
// Factor out the amount that gets handled inside the sequence
// (Pushes of argument for frame setup, callee pops for frame destroy)
Amount -= InternalAmt;
+ // If this is a callee-pop calling convention, and we're emitting precise
+ // SP-based CFI, emit a CFA adjust for the amount the callee popped.
+ if (isDestroy && InternalAmt && DwarfCFI && !hasFP(MF) &&
+ MMI.usePreciseUnwindInfo())
+ BuildCFI(MBB, I, DL,
+ MCCFIInstruction::createAdjustCfaOffset(nullptr, -InternalAmt));
+
if (Amount) {
// Add Amount to SP to destroy a frame, and subtract to setup.
int Offset = isDestroy ? Amount : -Amount;
- BuildStackAdjustment(MBB, I, DL, Offset, /*InEpilogue=*/false);
+
+ if (!(Fn->optForMinSize() &&
+ adjustStackWithPops(MBB, I, DL, Offset)))
+ BuildStackAdjustment(MBB, I, DL, Offset, /*InEpilogue=*/false);
}
+
+ if (DwarfCFI && !hasFP(MF)) {
+ // If we don't have FP, but need to generate unwind information,
+ // we need to set the correct CFA offset after the stack adjustment.
+ // How much we adjust the CFA offset depends on whether we're emitting
+ // CFI only for EH purposes or for debugging. EH only requires the CFA
+ // offset to be correct at each call site, while for debugging we want
+ // it to be more precise.
+ int CFAOffset = Amount;
+ if (!MMI.usePreciseUnwindInfo())
+ CFAOffset += InternalAmt;
+ CFAOffset = isDestroy ? -CFAOffset : CFAOffset;
+ BuildCFI(MBB, I, DL,
+ MCCFIInstruction::createAdjustCfaOffset(nullptr, CFAOffset));
+ }
+
return;
}
bool X86FrameLowering::canUseAsEpilogue(const MachineBasicBlock &MBB) const {
assert(MBB.getParent() && "Block is not attached to a function!");
+ // Win64 has strict requirements in terms of epilogue and we are
+ // not taking a chance at messing with them.
+ // I.e., unless this block is already an exit block, we can't use
+ // it as an epilogue.
+ if (MBB.getParent()->getSubtarget<X86Subtarget>().isTargetWin64() &&
+ !MBB.succ_empty() && !MBB.isReturnBlock())
+ return false;
+
if (canUseLEAForSPInEpilogue(*MBB.getParent()))
return true;
// safe to insert the epilogue here.
return !terminatorsNeedFlagsAsInput(MBB);
}
+
+MachineBasicBlock::iterator X86FrameLowering::restoreWin32EHStackPointers(
+ MachineBasicBlock &MBB, MachineBasicBlock::iterator MBBI,
+ DebugLoc DL, bool RestoreSP) const {
+ assert(STI.isTargetWindowsMSVC() && "funclets only supported in MSVC env");
+ assert(STI.isTargetWin32() && "EBP/ESI restoration only required on win32");
+ assert(STI.is32Bit() && !Uses64BitFramePtr &&
+ "restoring EBP/ESI on non-32-bit target");
+
+ MachineFunction &MF = *MBB.getParent();
+ unsigned FramePtr = TRI->getFrameRegister(MF);
+ unsigned BasePtr = TRI->getBaseRegister();
+ MachineModuleInfo &MMI = MF.getMMI();
+ const Function *Fn = MF.getFunction();
+ WinEHFuncInfo &FuncInfo = MMI.getWinEHFuncInfo(Fn);
+ X86MachineFunctionInfo *X86FI = MF.getInfo<X86MachineFunctionInfo>();
+ MachineFrameInfo *MFI = MF.getFrameInfo();
+
+ // FIXME: Don't set FrameSetup flag in catchret case.
+
+ int FI = FuncInfo.EHRegNodeFrameIndex;
+ int EHRegSize = MFI->getObjectSize(FI);
+
+ if (RestoreSP) {
+ // MOV32rm -EHRegSize(%ebp), %esp
+ addRegOffset(BuildMI(MBB, MBBI, DL, TII.get(X86::MOV32rm), X86::ESP),
+ X86::EBP, true, -EHRegSize)
+ .setMIFlag(MachineInstr::FrameSetup);
+ }
+
+ unsigned UsedReg;
+ int EHRegOffset = getFrameIndexReference(MF, FI, UsedReg);
+ int EndOffset = -EHRegOffset - EHRegSize;
+ FuncInfo.EHRegNodeEndOffset = EndOffset;
+
+ if (UsedReg == FramePtr) {
+ // ADD $offset, %ebp
+ unsigned ADDri = getADDriOpcode(false, EndOffset);
+ BuildMI(MBB, MBBI, DL, TII.get(ADDri), FramePtr)
+ .addReg(FramePtr)
+ .addImm(EndOffset)
+ .setMIFlag(MachineInstr::FrameSetup)
+ ->getOperand(3)
+ .setIsDead();
+ assert(EndOffset >= 0 &&
+ "end of registration object above normal EBP position!");
+ } else if (UsedReg == BasePtr) {
+ // LEA offset(%ebp), %esi
+ addRegOffset(BuildMI(MBB, MBBI, DL, TII.get(X86::LEA32r), BasePtr),
+ FramePtr, false, EndOffset)
+ .setMIFlag(MachineInstr::FrameSetup);
+ // MOV32rm SavedEBPOffset(%esi), %ebp
+ assert(X86FI->getHasSEHFramePtrSave());
+ int Offset =
+ getFrameIndexReference(MF, X86FI->getSEHFramePtrSaveIndex(), UsedReg);
+ assert(UsedReg == BasePtr);
+ addRegOffset(BuildMI(MBB, MBBI, DL, TII.get(X86::MOV32rm), FramePtr),
+ UsedReg, true, Offset)
+ .setMIFlag(MachineInstr::FrameSetup);
+ } else {
+ llvm_unreachable("32-bit frames with WinEH must use FramePtr or BasePtr");
+ }
+ return MBBI;
+}
+
+unsigned X86FrameLowering::getWinEHParentFrameOffset(const MachineFunction &MF) const {
+ // RDX, the parent frame pointer, is homed into 16(%rsp) in the prologue.
+ unsigned Offset = 16;
+ // RBP is immediately pushed.
+ Offset += SlotSize;
+ // All callee-saved registers are then pushed.
+ Offset += MF.getInfo<X86MachineFunctionInfo>()->getCalleeSavedFrameSize();
+ // Every funclet allocates enough stack space for the largest outgoing call.
+ Offset += getWinEHFuncletFrameSize(MF);
+ return Offset;
+}
projects / oota-llvm.git / blobdiff