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

[Shilei Tian via llvm-commits]

================
@@ -767,11 +805,17 @@ 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)));
+    std::optional<std::pair<unsigned, unsigned>> Attr =
+        InfoCache.getFlatWorkGroupSizeAttr(*F);
+
+    if (Attr.has_value()) {
+      auto [Min, Max] = *Attr;
+      intersectKnown(ConstantRange(APInt(32, Min), APInt(32, Max + 1)));
+      indicatePessimisticFixpoint();
----------------
shiltian wrote:

This pessimistic fixed point will not make the AA invalid because the `known -> assumed` range is no longer a full set.

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