AMDGPU: Stop modifying SP in call sequences
Because the stack growth direction and addressing is done
in the same direction, modifying SP at the beginning of the
call sequence was incorrect. If we had a stack passed argument,
we would end up skipping that number of bytes before pushing
arguments, leaving unused/inconsistent space.
The callee creates fixed stack objects in its frame, so
the space necessary for these is already logically allocated
in the callee, so we just let the callee increment SP if
it really requires it.
llvm-svn: 313279
diff --git a/llvm/lib/Target/AMDGPU/SIISelLowering.cpp b/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
index a5c5ecc..99f7bad 100644
--- a/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
+++ b/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
@@ -2153,7 +2153,7 @@
// Adjust the stack pointer for the new arguments...
// These operations are automatically eliminated by the prolog/epilog pass
if (!IsSibCall) {
- Chain = DAG.getCALLSEQ_START(Chain, NumBytes, 0, DL);
+ Chain = DAG.getCALLSEQ_START(Chain, 0, 0, DL);
unsigned OffsetReg = Info->getScratchWaveOffsetReg();
@@ -2359,8 +2359,8 @@
InFlag = Chain.getValue(1);
}
- uint64_t CalleePopBytes = 0;
- Chain = DAG.getCALLSEQ_END(Chain, DAG.getTargetConstant(NumBytes, DL, MVT::i32),
+ uint64_t CalleePopBytes = NumBytes;
+ Chain = DAG.getCALLSEQ_END(Chain, DAG.getTargetConstant(0, DL, MVT::i32),
DAG.getTargetConstant(CalleePopBytes, DL, MVT::i32),
InFlag, DL);
if (!Ins.empty())