[AMDGPU] Fix a potential wrong return value indicating whether a pass modifies a function

shiltian · shiltian · commit 3cb7e72549dd · 2024-04-09T17:12:18.000-04:00
When the alloca is too big for vectorization, the function could have already
been modified in previous iteration of the `for` loop.
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp b/llvm/lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp
@@ -336,7 +336,7 @@ bool AMDGPUPromoteAllocaImpl::run(Function &F, bool PromoteToLDS) {
     if (AllocaCost > VectorizationBudget) {
       LLVM_DEBUG(dbgs() << "  Alloca too big for vectorization: " << *AI
                         << "\n");
-      return false;
+      return Changed;
     }
 
     if (tryPromoteAllocaToVector(*AI)) {

Original file line number	Diff line number	Diff line change
`@@ -336,7 +336,7 @@ bool AMDGPUPromoteAllocaImpl::run(Function &F, bool PromoteToLDS) {`
`336`	`336`	`if (AllocaCost > VectorizationBudget) {`
`337`	`337`	`LLVM_DEBUG(dbgs() << " Alloca too big for vectorization: " << *AI`
`338`	`338`	`<< "\n");`
`339`		`- return false;`
	`339`	`+ return Changed;`
`340`	`340`	`}`
`341`	`341`
`342`	`342`	`if (tryPromoteAllocaToVector(*AI)) {`