2 files changed, 11 insertions, 29 deletions

diff --git a/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp b/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
index 92a587b..280fbe2 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
@@ -1384,6 +1384,11 @@ void AMDGPUPassConfig::addCodeGenPrepare() {
   if (TM->getTargetTriple().isAMDGCN() && EnableLowerKernelArguments)
     addPass(createAMDGPULowerKernelArgumentsPass());
 
+  TargetPassConfig::addCodeGenPrepare();
+
+  if (isPassEnabled(EnableLoadStoreVectorizer))
+    addPass(createLoadStoreVectorizerPass());
+
   if (TM->getTargetTriple().isAMDGCN()) {
     // This lowering has been placed after codegenprepare to take advantage of
     // address mode matching (which is why it isn't put with the LDS lowerings).
@@ -1392,15 +1397,6 @@ void AMDGPUPassConfig::addCodeGenPrepare() {
     // but has been put before switch lowering and CFG flattening so that those
     // passes can run on the more optimized control flow this pass creates in
     // many cases.
-    //
-    // FIXME: This should ideally be put after the LoadStoreVectorizer.
-    // However, due to some annoying facts about ResourceUsageAnalysis,
-    // (especially as exercised in the resource-usage-dead-function test),
-    // we need all the function passes codegenprepare all the way through
-    // said resource usage analysis to run on the call graph produced
-    // before codegenprepare runs (because codegenprepare will knock some
-    // nodes out of the graph, which leads to function-level passes not
-    // being run on them, which causes crashes in the resource usage analysis).
     addPass(createAMDGPULowerBufferFatPointersPass());
     addPass(createAMDGPULowerIntrinsicsLegacyPass());
     // In accordance with the above FIXME, manually force all the
@@ -1408,11 +1404,6 @@ void AMDGPUPassConfig::addCodeGenPrepare() {
     addPass(new DummyCGSCCPass());
   }
 
-  TargetPassConfig::addCodeGenPrepare();
-
-  if (isPassEnabled(EnableLoadStoreVectorizer))
-    addPass(createLoadStoreVectorizerPass());
-
   // LowerSwitch pass may introduce unreachable blocks that can
   // cause unexpected behavior for subsequent passes. Placing it
   // here seems better that these blocks would get cleaned up by
@@ -2125,6 +2116,11 @@ void AMDGPUCodeGenPassBuilder::addCodeGenPrepare(AddIRPass &addPass) const {
   if (EnableLowerKernelArguments)
     addPass(AMDGPULowerKernelArgumentsPass(TM));
 
+  Base::addCodeGenPrepare(addPass);
+
+  if (isPassEnabled(EnableLoadStoreVectorizer))
+    addPass(LoadStoreVectorizerPass());
+
   // This lowering has been placed after codegenprepare to take advantage of
   // address mode matching (which is why it isn't put with the LDS lowerings).
   // It could be placed anywhere before uniformity annotations (an analysis
@@ -2132,25 +2128,11 @@ void AMDGPUCodeGenPassBuilder::addCodeGenPrepare(AddIRPass &addPass) const {
   // but has been put before switch lowering and CFG flattening so that those
   // passes can run on the more optimized control flow this pass creates in
   // many cases.
-  //
-  // FIXME: This should ideally be put after the LoadStoreVectorizer.
-  // However, due to some annoying facts about ResourceUsageAnalysis,
-  // (especially as exercised in the resource-usage-dead-function test),
-  // we need all the function passes codegenprepare all the way through
-  // said resource usage analysis to run on the call graph produced
-  // before codegenprepare runs (because codegenprepare will knock some
-  // nodes out of the graph, which leads to function-level passes not
-  // being run on them, which causes crashes in the resource usage analysis).
   addPass(AMDGPULowerBufferFatPointersPass(TM));
   addPass.requireCGSCCOrder();
 
   addPass(AMDGPULowerIntrinsicsPass(TM));
 
-  Base::addCodeGenPrepare(addPass);
-
-  if (isPassEnabled(EnableLoadStoreVectorizer))
-    addPass(LoadStoreVectorizerPass());
-
   // LowerSwitch pass may introduce unreachable blocks that can cause unexpected
   // behavior for subsequent passes. Placing it here seems better that these
   // blocks would get cleaned up by UnreachableBlockElim inserted next in the
diff --git a/llvm/lib/Target/TargetMachine.cpp b/llvm/lib/Target/TargetMachine.cpp
index ad7e503..cf85691 100644
--- a/llvm/lib/Target/TargetMachine.cpp
+++ b/llvm/lib/Target/TargetMachine.cpp
@@ -27,7 +27,7 @@
 #include "llvm/Target/TargetLoweringObjectFile.h"
 using namespace llvm;
 
-cl::opt<bool> NoKernelInfoEndLTO(
+cl::opt<bool> llvm::NoKernelInfoEndLTO(
     "no-kernel-info-end-lto",
     cl::desc("remove the kernel-info pass at the end of the full LTO pipeline"),
     cl::init(false), cl::Hidden);