[llvm] [AMDGPU] Folding imm offset in more cases for scratch access (PR #70634)

[Matt Arsenault via llvm-commits]

================
@@ -1549,7 +1598,7 @@ bool AMDGPUDAGToDAGISel::SelectFlatOffsetImpl(SDNode *N, SDValue Addr,
   if (Subtarget->hasFlatInstOffsets() && !CanHaveFlatSegmentOffsetBug) {
     SDValue N0, N1;
     if (isBaseWithConstantOffset64(Addr, N0, N1) &&
-        isFlatScratchBaseLegal(N0, FlatVariant)) {
+        isFlatScratchBaseLegal(Addr, false, false, FlatVariant)) {
----------------
arsenm wrote:

/*CommentParamName=*/

https://github.com/llvm/llvm-project/pull/70634