[llvm] [AMDGPU][Attributor] Make `AAAMDFlatWorkGroupSize` honor existing attribute (PR #114357)

[via llvm-commits]

================
@@ -767,13 +802,33 @@ struct AAAMDFlatWorkGroupSize : public AAAMDSizeRangeAttribute {
   void initialize(Attributor &A) override {
     Function *F = getAssociatedFunction();
     auto &InfoCache = static_cast<AMDGPUInformationCache &>(A.getInfoCache());
-    unsigned MinGroupSize, MaxGroupSize;
-    std::tie(MinGroupSize, MaxGroupSize) = InfoCache.getFlatWorkGroupSizes(*F);
-    intersectKnown(
-        ConstantRange(APInt(32, MinGroupSize), APInt(32, MaxGroupSize + 1)));
 
-    if (AMDGPU::isEntryFunctionCC(F->getCallingConv()))
-      indicatePessimisticFixpoint();
+    bool HasAttr = false;
+    auto Range = InfoCache.getDefaultFlatWorkGroupSize(*F);
+    auto MaxRange = InfoCache.getMaximumFlatWorkGroupRange(*F);
+
+    if (auto Attr = InfoCache.getFlatWorkGroupSizeAttr(*F)) {
+      // We will only consider an attribute different from max because the front
+      // end always emits the attribure, unfortunately, and sometimes it emits
----------------
choikwa wrote:

typo

https://github.com/llvm/llvm-project/pull/114357