[llvm-branch-commits] [llvm] [AMDGPU][SDAG] DAGCombine PTRADD -> disjoint OR (PR #146075)

Fabian Ritter via llvm-branch-commits llvm-branch-commits at lists.llvm.org
Sun Aug 3 23:56:40 PDT 2025


================
@@ -2767,6 +2767,19 @@ SDValue DAGCombiner::visitPTRADD(SDNode *N) {
     }
   }
 
+  // Transform (ptradd a, b) -> (or disjoint a, b) if it is equivalent and if
+  // that transformation can't block an offset folding at any use of the ptradd.
+  // This should be done late, after legalization, so that it doesn't block
+  // other ptradd combines that could enable more offset folding.
+  if (LegalOperations && DAG.haveNoCommonBitsSet(N0, N1)) {
+    bool TransformCanBreakAddrMode = any_of(N->users(), [&](SDNode *User) {
+      return canFoldInAddressingMode(N, User, DAG, TLI);
+    });
+
+    if (!TransformCanBreakAddrMode)
----------------
ritter-x2a wrote:

Done.

https://github.com/llvm/llvm-project/pull/146075


More information about the llvm-branch-commits mailing list