VCVTSS2SD requires a strict alignment. Thanks Elena.

[oota-llvm.git] / lib / Target / X86 / X86FrameLowering.cpp

diff --git a/lib/Target/X86/X86FrameLowering.cpp b/lib/Target/X86/X86FrameLowering.cpp
index 369589d469a6ae40d3c6a095cd8585c2eedb188f..63454877514633720eac1366c8f0b79470064de8 100644 (file)
--- a/lib/Target/X86/X86FrameLowering.cpp
+++ b/lib/Target/X86/X86FrameLowering.cpp
@@ -17,18 +17,18 @@
 #include "X86MachineFunctionInfo.h"
 #include "X86Subtarget.h"
 #include "X86TargetMachine.h"
-#include "llvm/Function.h"
+#include "llvm/ADT/SmallSet.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/DataLayout.h"
+#include "llvm/Function.h"
 #include "llvm/MC/MCAsmInfo.h"
 #include "llvm/MC/MCSymbol.h"
-#include "llvm/DataLayout.h"
-#include "llvm/Target/TargetOptions.h"
 #include "llvm/Support/CommandLine.h"
-#include "llvm/ADT/SmallSet.h"
+#include "llvm/Target/TargetOptions.h"
 
 using namespace llvm;
 
@@ -625,6 +625,22 @@ uint32_t X86FrameLowering::getCompactUnwindEncoding(MachineFunction &MF) const {
   return CompactUnwindEncoding;
 }
 
+/// usesTheStack - This function checks if any of the users of EFLAGS
+/// copies the EFLAGS. We know that the code that lowers COPY of EFLAGS has
+/// to use the stack, and if we don't adjust the stack we clobber the first
+/// frame index.
+/// See X86InstrInfo::copyPhysReg.
+static bool usesTheStack(MachineFunction &MF) {
+  MachineRegisterInfo &MRI = MF.getRegInfo();
+
+  for (MachineRegisterInfo::reg_iterator ri = MRI.reg_begin(X86::EFLAGS),
+       re = MRI.reg_end(); ri != re; ++ri)
+    if (ri->isCopy())
+      return true;
+
+  return false;
+}
+
 /// 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
@@ -673,12 +689,14 @@ void X86FrameLowering::emitPrologue(MachineFunction &MF) const {
   // 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
-  // stack pointer (we fit in the Red Zone).
-  if (Is64Bit && !Fn->getFnAttributes().hasAttribute(Attributes::NoRedZone) &&
+  // stack pointer (we fit in the Red Zone). We also check that we don't
+  // push and pop from the stack.
+  if (Is64Bit && !Fn->getFnAttributes().hasAttribute(Attribute::NoRedZone) &&
       !RegInfo->needsStackRealignment(MF) &&
       !MFI->hasVarSizedObjects() &&                     // No dynamic alloca.
       !MFI->adjustsStack() &&                           // No calls.
       !IsWin64 &&                                       // Win64 has no Red Zone
+      !usesTheStack(MF) &&                              // Don't push and pop.
       !MF.getTarget().Options.EnableSegmentedStacks) {  // Regular stack
     uint64_t MinSize = X86FI->getCalleeSavedFrameSize();
     if (HasFP) MinSize += SlotSize;
@@ -1138,7 +1156,7 @@ void X86FrameLowering::emitEpilogue(MachineFunction &MF,
     }
 
     MachineInstr *NewMI = prior(MBBI);
-    NewMI->copyImplicitOps(MBBI);
+    NewMI->copyImplicitOps(MF, MBBI);
 
     // Delete the pseudo instruction TCRETURN.
     MBB.erase(MBBI);