[PATCH] D140599: AMDGPU: Promote array alloca if used by memmove/memcpy

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added inline comments.


================
Comment at: llvm/lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp:502
+
+      auto getPointerIndexOfAlloca = [&](Value *Ptr) -> ConstantInt * {
+        GetElementPtrInst *GEP = dyn_cast<GetElementPtrInst>(Ptr);
----------------
[=]?


================
Comment at: llvm/lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp:593-595
+          case Intrinsic::memcpy:
+          case Intrinsic::memcpy_inline:
+          case Intrinsic::memmove: {
----------------
probably should dyn_cast to MemIntrinsicInst here to defend agains these becoming out of sync


================
Comment at: llvm/lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp:611
+            Type *VecPtrTy = VectorTy->getPointerTo(Alloca->getAddressSpace());
+            Value *BitCast = Builder.CreateBitCast(Alloca, VecPtrTy);
+            Value *VecValue = Builder.CreateLoad(VectorTy, BitCast);
----------------
No reason to care about type pointers anymore


================
Comment at: llvm/test/CodeGen/AMDGPU/promote-alloca-array-aggregate.ll:219
+
+declare void @llvm.memcpy.p5i8.p5i8.i32(ptr addrspace(5) nocapture writeonly, ptr addrspace(5) nocapture readonly, i32, i1 immarg)
+declare void @llvm.memmove.p5i8.p5i8.i32(ptr addrspace(5) nocapture writeonly, ptr addrspace(5) nocapture readonly, i32, i1 immarg)
----------------
Needs tests for the different address space cases you mentioned in the todo 


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D140599/new/

https://reviews.llvm.org/D140599