[PATCH] D54158: AMDGPU: Add an option amdgpu-enable-promote-alloca to enable/disable alloca

[Yaxun Liu via Phabricator via llvm-commits]

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

Add this option for debugging and providing workaround.

By default it is enabled so no behavior change in backend.


https://reviews.llvm.org/D54158

Files:
  lib/Target/AMDGPU/AMDGPUTargetMachine.cpp


Index: lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
===================================================================
--- lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
+++ lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
@@ -62,6 +62,12 @@
   cl::ReallyHidden,
   cl::init(true));
 
+static cl::opt<bool> EnablePromoteAlloca(
+  "amdgpu-enable-promote-alloca",
+  cl::desc("Run promote alloca pass"),
+  cl::ReallyHidden,
+  cl::init(true));
+
 static cl::opt<bool>
 EnableEarlyIfConversion("amdgpu-early-ifcvt", cl::Hidden,
                         cl::desc("Run early if-conversion"),
@@ -653,7 +659,8 @@
 
   if (TM.getOptLevel() > CodeGenOpt::None) {
     addPass(createInferAddressSpacesPass());
-    addPass(createAMDGPUPromoteAlloca());
+    if (EnablePromoteAlloca)
+      addPass(createAMDGPUPromoteAlloca());
 
     if (EnableSROA)
       addPass(createSROAPass());


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D54158.172781.patch
Type: text/x-patch
Size: 864 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20181106/5bf26fa7/attachment.bin>