[llvm] [AMDGPU] Fix image intrinsic optimizer on loads from different resources (PR #69355)

[Carl Ritson via llvm-commits]

================
@@ -108,28 +108,29 @@ void addInstToMergeableList(
     if (IIList.front()->getType() != II->getType())
       continue;
 
-    // Check DMask.
-    Value *DMaskList = IIList.front()->getArgOperand(ImageDimIntr->DMaskIndex);
-    Value *DMask = II->getArgOperand(ImageDimIntr->DMaskIndex);
-    if (DMaskList != DMask)
-      continue;
-
-    // Check VAddr (except FragId).
-    int I = ImageDimIntr->VAddrStart;
-    for (; I < ImageDimIntr->VAddrEnd - 1; ++I) {
-      if (IIList.front()->getArgOperand(I) != II->getArgOperand(I))
-        break;
+    // Check all arguments (DMask, VAddr, RSrc etc).
+    bool AllEqual = true;
+    assert(IIList.front()->arg_size() == II->arg_size());
+    for (int I = 1, E = II->arg_size(); I != E; ++I) {
----------------
perlfu wrote:

This is just stylistic, so feel free to ignore me, but if `AllEqual` was part of the for-loop precondition, then you could just assign it comparison results in the loop and remove the `if (...) {AllEqual = true; break}` patterns.

https://github.com/llvm/llvm-project/pull/69355