#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"
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();
unsigned FramePtr = TRI->getFrameRegister(MF);
const unsigned MachineFramePtr =
STI.isTarget64BitILP32()
- ? getX86SubSuperRegister(FramePtr, MVT::i64, false)
- : FramePtr;
- unsigned BasePtr = TRI->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");
uint64_t NumBytes = 0;
int stackGrowth = -SlotSize;
- unsigned RDX = Uses64BitFramePtr ? X86::RDX : X86::EDX;
- if (IsWin64Prologue && IsFunclet) {
- // Immediately spill RDX into the home slot. The runtime cares about this.
+ // 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(RDX)
+ .addReg(Establisher)
.setMIFlag(MachineInstr::FrameSetup);
}
// NumBytes value that we would've used for the parent frame.
unsigned ParentFrameNumBytes = NumBytes;
if (IsFunclet)
- NumBytes = MFI->getMaxCallFrameSize();
+ NumBytes = getWinEHFuncletFrameSize(MF);
// Skip the callee-saved push instructions.
bool PushedRegs = false;
int SEHFrameOffset = 0;
if (IsWin64Prologue && HasFP) {
// Set RBP to a small fixed offset from RSP. In the funclet case, we base
- // this calculation on the incoming RDX, which holds the value of RSP from
- // the parent frame at the end of the prologue.
- unsigned SPOrRDX = !IsFunclet ? StackPtr : RDX;
+ // 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),
- SPOrRDX, false, SEHFrameOffset);
+ SPOrEstablisher, false, SEHFrameOffset);
else
BuildMI(MBB, MBBI, DL, TII.get(X86::MOV64rr), FramePtr)
- .addReg(SPOrRDX);
+ .addReg(SPOrEstablisher);
// If this is not a funclet, emit the CFI describing our frame pointer.
if (NeedsWinCFI && !IsFunclet)
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();
uint64_t NumBytes = 0;
if (MBBI->getOpcode() == X86::CATCHRET) {
- NumBytes = MFI->getMaxCallFrameSize();
+ NumBytes = getWinEHFuncletFrameSize(MF);
assert(hasFP(MF) && "EH funclets without FP not yet implemented");
MachineBasicBlock *TargetMBB = MBBI->getOperand(0).getMBB();
RestoreMBB = TargetMBB;
if (STI.is32Bit()) {
RestoreMBB = MF.CreateMachineBasicBlock(MBB.getBasicBlock());
- MF.insert(TargetMBB, RestoreMBB);
+ MF.insert(TargetMBB->getIterator(), RestoreMBB);
MBB.removeSuccessor(TargetMBB);
MBB.addSuccessor(RestoreMBB);
RestoreMBB->addSuccessor(TargetMBB);
.addMBB(TargetMBB);
}
} else if (MBBI->getOpcode() == X86::CLEANUPRET) {
- NumBytes = MFI->getMaxCallFrameSize();
+ 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)
.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())
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.
+ // 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 =
- !MF.getTarget().getMCAsmInfo()->usesWindowsCFI() &&
- !MF.getMMI().getLandingPads().empty();
+ bool HasDwarfEHHandlers = !WindowsCFI &&
+ !MF.getMMI().getLandingPads().empty();
- if (HasDwarfEHHandlers && !isDestroy &&
+ if (HasDwarfEHHandlers && !isDestroy &&
MF.getInfo<X86MachineFunctionInfo>()->getHasPushSequences())
BuildCFI(MBB, I, DL,
MCCFIInstruction::createGnuArgsSize(nullptr, Amount));
// (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;
- if (!(MF.getFunction()->optForMinSize() &&
+ 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;
}
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;
+}