[PATCH] X86: Emit Win64 SaveXMM opcodes at the right offset in the right order

[David Majnemer]

Hi rnk,

Walk the instructions marked FrameSetup and consider any stores of XMM
registers to the stack as needing a SaveXMM opcode.

This fixes PR22521.

http://reviews.llvm.org/D7527

Files:
  lib/Target/X86/X86FrameLowering.cpp
  test/CodeGen/X86/win64_eh.ll

Index: lib/Target/X86/X86FrameLowering.cpp
===================================================================
--- lib/Target/X86/X86FrameLowering.cpp
+++ lib/Target/X86/X86FrameLowering.cpp
@@ -909,29 +909,29 @@
           .setMIFlag(MachineInstr::FrameSetup);
   }
 
-  // Skip the rest of register spilling code
-  while (MBBI != MBB.end() && MBBI->getFlag(MachineInstr::FrameSetup))
+  while (MBBI != MBB.end() && MBBI->getFlag(MachineInstr::FrameSetup)) {
+    const MachineInstr *FrameInstr = &*MBBI;
     ++MBBI;
 
-  if (NeedsWinEH) {
-    for (const CalleeSavedInfo &Info : MFI->getCalleeSavedInfo()) {
-      unsigned Reg = Info.getReg();
-      if (X86::GR64RegClass.contains(Reg) || X86::GR32RegClass.contains(Reg))
-        continue;
-      assert(X86::FR64RegClass.contains(Reg) && "Unexpected register class");
-
-      int Offset = getFrameIndexOffset(MF, Info.getFrameIdx());
-      Offset += SEHFrameOffset;
-
-      BuildMI(MBB, MBBI, DL, TII.get(X86::SEH_SaveXMM))
-          .addImm(Reg)
-          .addImm(Offset)
-          .setMIFlag(MachineInstr::FrameSetup);
+    if (NeedsWinEH) {
+      int FI;
+      if (unsigned Reg = TII.isStoreToStackSlot(FrameInstr, FI)) {
+        if (X86::FR64RegClass.contains(Reg)) {
+          int Offset = getFrameIndexOffset(MF, FI);
+          Offset += SEHFrameOffset;
+
+          BuildMI(MBB, MBBI, DL, TII.get(X86::SEH_SaveXMM))
+              .addImm(Reg)
+              .addImm(Offset)
+              .setMIFlag(MachineInstr::FrameSetup);
+        }
+      }
     }
+  }
 
+  if (NeedsWinEH)
     BuildMI(MBB, MBBI, DL, TII.get(X86::SEH_EndPrologue))
         .setMIFlag(MachineInstr::FrameSetup);
-  }
 
   // Realign stack after we spilled callee-saved registers (so that we'll be
   // able to calculate their offsets from the frame pointer).
@@ -1468,8 +1468,7 @@
   // It can be done by spilling XMMs to stack frame.
   for (unsigned i = CSI.size(); i != 0; --i) {
     unsigned Reg = CSI[i-1].getReg();
-    if (X86::GR64RegClass.contains(Reg) ||
-        X86::GR32RegClass.contains(Reg))
+    if (X86::GR64RegClass.contains(Reg) || X86::GR32RegClass.contains(Reg))
       continue;
     // Add the callee-saved register as live-in. It's killed at the spill.
     MBB.addLiveIn(Reg);
Index: test/CodeGen/X86/win64_eh.ll
===================================================================
--- test/CodeGen/X86/win64_eh.ll
+++ test/CodeGen/X86/win64_eh.ll
@@ -155,9 +155,9 @@
 ; WIN64: leaq  128(%rsp), %rbp
 ; WIN64: .seh_setframe 5, 128
 ; WIN64: movaps  %xmm7, -32(%rbp)        # 16-byte Spill
+; WIN64: .seh_savexmm 7, 96
 ; WIN64: movaps  %xmm6, -48(%rbp)        # 16-byte Spill
 ; WIN64: .seh_savexmm 6, 80
-; WIN64: .seh_savexmm 7, 96
 ; WIN64: .seh_endprologue
 ; WIN64: andq  $-64, %rsp
 ; WIN64: movaps  -48(%rbp), %xmm6        # 16-byte Reload

EMAIL PREFERENCES
  http://reviews.llvm.org/settings/panel/emailpreferences/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: D7527.19645.patch
Type: text/x-patch
Size: 2824 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20150210/015c9040/attachment.bin>