[llvm] [AMDGPU] Pre-commit tests for "lshr + mad" fold (PR #119509)

[Vikram Hegde via llvm-commits]

================
@@ -1333,5 +1333,403 @@ define i48 @mad_i48_i48(i48 %arg0, i48 %arg1, i48 %arg2) #0 {
   ret i48 %a
 }
 
+define i64 @lshr_mad_i64_1(ptr addrspace(1) %1, ptr addrspace(1) %2) local_unnamed_addr #0 {
----------------
vikramRH wrote:

done

https://github.com/llvm/llvm-project/pull/119509