[llvm] AMDGPU: Fix using IRAttribute with nounwind for AMDGPUNoAGPR (PR #161954)

[via llvm-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-backend-amdgpu

Author: Matt Arsenault (arsenm)

<details>
<summary>Changes</summary>

Don't think this did anything harmful, but it doesn't make sense
to report this as implementing nounwind handling.

---
Full diff: https://github.com/llvm/llvm-project/pull/161954.diff


1 Files Affected:

- (modified) llvm/lib/Target/AMDGPU/AMDGPUAttributor.cpp (+3-5) 


``````````diff
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUAttributor.cpp b/llvm/lib/Target/AMDGPU/AMDGPUAttributor.cpp
index 9dd64e0f6b35b..8b72d356ebfeb 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUAttributor.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUAttributor.cpp
@@ -1225,11 +1225,9 @@ static bool inlineAsmUsesAGPRs(const InlineAsm *IA) {
 
 // TODO: Migrate to range merge of amdgpu-agpr-alloc.
 // FIXME: Why is this using Attribute::NoUnwind?
-struct AAAMDGPUNoAGPR
-    : public IRAttribute<Attribute::NoUnwind,
-                         StateWrapper<BooleanState, AbstractAttribute>,
-                         AAAMDGPUNoAGPR> {
-  AAAMDGPUNoAGPR(const IRPosition &IRP, Attributor &A) : IRAttribute(IRP) {}
+struct AAAMDGPUNoAGPR : public StateWrapper<BooleanState, AbstractAttribute> {
+  using Base = StateWrapper<BooleanState, AbstractAttribute>;
+  AAAMDGPUNoAGPR(const IRPosition &IRP, Attributor &A) : Base(IRP) {}
 
   static AAAMDGPUNoAGPR &createForPosition(const IRPosition &IRP,
                                            Attributor &A) {

``````````

</details>


https://github.com/llvm/llvm-project/pull/161954