[llvm] [AMDGPU] Correctly merge noalias scopes during lowering of LDS data. (PR #131664)

[Sirish Pande via llvm-commits]

================
@@ -244,6 +245,7 @@ template <typename T> std::vector<T> sortByName(std::vector<T> &&V) {
 
 class AMDGPULowerModuleLDS {
   const AMDGPUTargetMachine &TM;
+  friend class ScopedNoAliasAAResult;
----------------
srpande wrote:

No need. I was playing with something else. Does not need it anymore. 

https://github.com/llvm/llvm-project/pull/131664