[PATCH] D45993: AMDGPU/SI: Don't promote alloca to vector for AddrSpaceCast instruction.

[Changpeng Fang via Phabricator via llvm-commits]

cfang created this revision.
cfang added reviewers: arsenm, msearles.
Herald added subscribers: t-tye, tpr, dstuttard, yaxunl, nhaehnle, wdng, kzhuravl.

We have no logic to promote alloca to vector for an AddrSpaceCast instruction.


https://reviews.llvm.org/D45993

Files:
  lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp


Index: lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp
===================================================================
--- lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp
+++ lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp
@@ -328,7 +328,6 @@
     return isa<GetElementPtrInst>(LI->getPointerOperand()) && !LI->isVolatile();
   }
   case Instruction::BitCast:
-  case Instruction::AddrSpaceCast:
     return true;
   case Instruction::Store: {
     // Must be the stored pointer operand, not a stored value, plus


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D45993.143659.patch
Type: text/x-patch
Size: 499 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20180423/a262fd61/attachment.bin>