[PATCH] D20728: AMDGPU: Disable AMDGPU for shader calling conventions.

[Bas Nieuwenhuizen via llvm-commits]

bnieuwenhuizen created this revision.
bnieuwenhuizen added reviewers: tstellarAMD, arsenm.
bnieuwenhuizen added a subscriber: llvm-commits.
Herald added subscribers: kzhuravl, arsenm.

The work item intrinsics are not available for the shader
calling conventions. And even if we did hook them up most
shader stages haves some extra restrictions on the amount
of available LDS.

http://reviews.llvm.org/D20728

Files:
  lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp

Index: lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp
===================================================================
--- lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp
+++ lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp
@@ -121,7 +121,7 @@
 }
 
 bool AMDGPUPromoteAlloca::runOnFunction(Function &F) {
-  if (!TM || skipFunction(F))
+  if (!TM || skipFunction(F) || AMDGPU::isShader(F.getCallingConv()))
     return false;
 
   FunctionType *FTy = F.getFunctionType();


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D20728.58774.patch
Type: text/x-patch
Size: 457 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20160527/dd104648/attachment.bin>