summaryrefslogtreecommitdiffstats
path: root/llvm/lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp
diff options
context:
space:
mode:
Diffstat (limited to 'llvm/lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp')
-rw-r--r--llvm/lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp4
1 files changed, 2 insertions, 2 deletions
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp b/llvm/lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp
index 5f05ce7d2a2..a2d62c73404 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp
@@ -244,10 +244,10 @@ AMDGPUPromoteAlloca::getLocalSizeYZ(IRBuilder<> &Builder) {
// We could do a single 64-bit load here, but it's likely that the basic
// 32-bit and extract sequence is already present, and it is probably easier
// to CSE this. The loads should be mergable later anyway.
- Value *GEPXY = Builder.CreateConstInBoundsGEP1_64(CastDispatchPtr, 1);
+ Value *GEPXY = Builder.CreateConstInBoundsGEP1_64(I32Ty, CastDispatchPtr, 1);
LoadInst *LoadXY = Builder.CreateAlignedLoad(I32Ty, GEPXY, 4);
- Value *GEPZU = Builder.CreateConstInBoundsGEP1_64(CastDispatchPtr, 2);
+ Value *GEPZU = Builder.CreateConstInBoundsGEP1_64(I32Ty, CastDispatchPtr, 2);
LoadInst *LoadZU = Builder.CreateAlignedLoad(I32Ty, GEPZU, 4);
MDNode *MD = MDNode::get(Mod->getContext(), None);
OpenPOWER on IntegriCloud