[llvm] [DAG] Fold (umin (sub a b) a) -> (usubo a b); (select usubo.1 a usubo.0) (PR #161651)

[Chaitanya Koparkar via llvm-commits]

================
@@ -0,0 +1,14 @@
+; RUN: llc < %s -mtriple=aarch64 | FileCheck %s
+
+; GitHub issue #161036
+
+define i64 @underflow_compare_fold(i64 %a, i64 %b) {
+; CHECK-LABEL: underflow_compare_fold
+; CHECK:      // %bb.0:
+; CHECK-NEXT: subs x8, x0, x1
+; CHECK-NEXT: csel x0, x0, x8, lo
+; CHECK-NEXT: ret
+  %sub = sub i64 %a, %b
+  %cond = tail call i64 @llvm.umin.i64(i64 %sub, i64 %a)
+  ret i64 %cond
+}
----------------
ckoparkar wrote:

I've added some tests with vector types, incorrect patterns, etc.

I didn't guard against multiple uses here since I thought the results of sub and umin are still available after the fold. Am I misunderstanding something here and should I be guarding the pattern with m_OneUse?


https://github.com/llvm/llvm-project/pull/161651