[PATCH] D48639: AMDGPU: Fix AMDGPUCodeGenPrepare using uninitialized AMDGPUAS struct

[Matt Arsenault via Phabricator via llvm-commits]

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

Not sure how this wasn't noticed before.


https://reviews.llvm.org/D48639

Files:
  lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp


Index: lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp
===================================================================
--- lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp
+++ lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp
@@ -577,10 +577,11 @@
   if (!TPC)
     return false;
 
-  const TargetMachine &TM = TPC->getTM<TargetMachine>();
+  const AMDGPUTargetMachine &TM = TPC->getTM<AMDGPUTargetMachine>();
   ST = &TM.getSubtarget<SISubtarget>(F);
   DA = &getAnalysis<DivergenceAnalysis>();
   HasUnsafeFPMath = hasUnsafeFPMath(F);
+  AMDGPUASI = TM.getAMDGPUAS();
 
   bool MadeChange = false;
 


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D48639.153054.patch
Type: text/x-patch
Size: 588 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20180627/1ea0606b/attachment.bin>