diff options
| author | Matt Arsenault <Matthew.Arsenault@amd.com> | 2019-07-05 20:26:13 +0000 |
|---|---|---|
| committer | Matt Arsenault <Matthew.Arsenault@amd.com> | 2019-07-05 20:26:13 +0000 |
| commit | e7e23e3e91e32263579db5cea49302de7ccf88b2 (patch) | |
| tree | bb52fa61fbc765807ca6f1b89086971979d91d87 /llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp | |
| parent | 8d6ea2d48c87d9ada06c384de3e255a2f6706ac3 (diff) | |
| download | bcm5719-llvm-e7e23e3e91e32263579db5cea49302de7ccf88b2.tar.gz bcm5719-llvm-e7e23e3e91e32263579db5cea49302de7ccf88b2.zip | |
AMDGPU: Make AMDGPUPerfHintAnalysis an SCC pass
Add a string attribute instead of directly setting
MachineFunctionInfo. This avoids trying to get the analysis in the
MachineFunctionInfo in a way that doesn't work with the new pass
manager.
This will also avoid re-visiting the call graph for every single
function.
llvm-svn: 365241
Diffstat (limited to 'llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp')
| -rw-r--r-- | llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp | 2 |
1 files changed, 2 insertions, 0 deletions
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp b/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp index 87a820a395d..7414519aee1 100644 --- a/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp +++ b/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp @@ -740,6 +740,8 @@ void AMDGPUPassConfig::addCodeGenPrepare() { EnableLowerKernelArguments) addPass(createAMDGPULowerKernelArgumentsPass()); + addPass(&AMDGPUPerfHintAnalysisID); + TargetPassConfig::addCodeGenPrepare(); if (EnableLoadStoreVectorizer) |
