X86: correct tail return address calculation

Due to the weird and wondeful usual arithmetic conversions, some
calculations involving negative values were getting performed in
uint32_t and then promoted to int64_t, which is really not a good
idea.

Patch by Katsuhiro Ueno.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187703 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Target/X86/X86FrameLowering.cpp b/lib/Target/X86/X86FrameLowering.cpp
index 5db431b60a4f8e968ec4e28ddd9052361900867b..b994e678a8a68b2ad9647f6f5bb36d963f1a7191 100644 (file)
--- a/lib/Target/X86/X86FrameLowering.cpp
+++ b/lib/Target/X86/X86FrameLowering.cpp
@@ -1323,7 +1323,7 @@ X86FrameLowering::processFunctionBeforeCalleeSavedScan(MachineFunction &MF,
   unsigned SlotSize = RegInfo->getSlotSize();
 
   X86MachineFunctionInfo *X86FI = MF.getInfo<X86MachineFunctionInfo>();
-  int32_t TailCallReturnAddrDelta = X86FI->getTCReturnAddrDelta();
+  int64_t TailCallReturnAddrDelta = X86FI->getTCReturnAddrDelta();
 
   if (TailCallReturnAddrDelta < 0) {
     // create RETURNADDR area
@@ -1336,7 +1336,7 @@ X86FrameLowering::processFunctionBeforeCalleeSavedScan(MachineFunction &MF,
     //   }
     //   [EBP]
     MFI->CreateFixedObject(-TailCallReturnAddrDelta,
-                           (-1U*SlotSize)+TailCallReturnAddrDelta, true);
+                           TailCallReturnAddrDelta - SlotSize, true);
   }
 
   if (hasFP(MF)) {

diff --git a/lib/Target/X86/X86ISelLowering.cpp b/lib/Target/X86/X86ISelLowering.cpp
index 1f6c330e464841a32ae1b141bba135fcd6df37f3..dc565473c1d69c00fbb484b97d0c0055085e339c 100644 (file)
--- a/lib/Target/X86/X86ISelLowering.cpp
+++ b/lib/Target/X86/X86ISelLowering.cpp
@@ -2456,7 +2456,8 @@ EmitTailCallStoreRetAddr(SelectionDAG & DAG, MachineFunction &MF,
   if (!FPDiff) return Chain;
   // Calculate the new stack slot for the return address.
   int NewReturnAddrFI =
-    MF.getFrameInfo()->CreateFixedObject(SlotSize, FPDiff-SlotSize, false);
+    MF.getFrameInfo()->CreateFixedObject(SlotSize, (int64_t)FPDiff - SlotSize,
+                                         false);
   SDValue NewRetAddrFrIdx = DAG.getFrameIndex(NewReturnAddrFI, PtrVT);
   Chain = DAG.getStore(Chain, dl, RetAddrFrIdx, NewRetAddrFrIdx,
                        MachinePointerInfo::getFixedStack(NewReturnAddrFI),
@@ -3323,7 +3324,8 @@ SDValue X86TargetLowering::getReturnAddressFrameIndex(SelectionDAG &DAG) const {
   if (ReturnAddrIndex == 0) {
     // Set up a frame object for the return address.
     unsigned SlotSize = RegInfo->getSlotSize();
-    ReturnAddrIndex = MF.getFrameInfo()->CreateFixedObject(SlotSize, -SlotSize,
+    ReturnAddrIndex = MF.getFrameInfo()->CreateFixedObject(SlotSize,
+                                                           -(int64_t)SlotSize,
                                                            false);
     FuncInfo->setRAIndex(ReturnAddrIndex);
   }

diff --git a/test/CodeGen/X86/tailcall-calleesave.ll b/test/CodeGen/X86/tailcall-calleesave.ll
new file mode 100644 (file)
index 0000000..2770d05
--- /dev/null
+++ b/test/CodeGen/X86/tailcall-calleesave.ll
@@ -0,0 +1,19 @@
+; RUN: llc -march=x86 -tailcallopt < %s | FileCheck %s
+
+target triple = "i686-apple-darwin"
+
+declare fastcc void @foo(i32, i32, i32, i32, i32, i32)
+declare i32* @bar(i32*)
+
+define fastcc void @hoge(i32 %b) nounwind {
+; Do not overwrite pushed callee-save registers
+; CHECK: pushl
+; CHECK: subl $[[SIZE:[0-9]+]], %esp
+; CHECK-NOT: [[SIZE]](%esp)
+  %a = alloca i32
+  store i32 0, i32* %a
+  %d = tail call i32* @bar(i32* %a) nounwind
+  store i32 %b, i32* %d
+  tail call fastcc void @foo(i32 1, i32 2, i32 3, i32 4, i32 5, i32 6) nounwind
+  ret void
+}