[PATCH] D87947: [AMDGPU] Make ds fp atomics overloadable

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added inline comments.


================
Comment at: clang/lib/CodeGen/CGBuiltin.cpp:14772
+    llvm::Type *PTy = FTy->getParamType(0);
+    Src0 = Builder.CreatePointerBitCastOrAddrSpaceCast(Src0, PTy);
+    return Builder.CreateCall(F, { Src0, Src1, Src2, Src3, Src4 });
----------------
rampitec wrote:
> arsenm wrote:
> > I don't think you need a cast here (at least an addrspacecast)
> If removed builtins-amdgcn.cu fails. It is CUDA with LDS pointer passed as flat. I.e. it comes as cast from addrspace(3) to flat. Generic builtin handling below in this file does the same.
I thought these casts would be present in the AST?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D87947/new/

https://reviews.llvm.org/D87947