[PATCH] D127766: [InstCombine] Fold memcmp of constant arrays and variable size

[Nikita Popov via Phabricator via llvm-commits]

nikic accepted this revision.
nikic added a comment.

LG. I think this will temporarily regress handling of zeroinitializer, but that should get addressed by D127443 <https://reviews.llvm.org/D127443> soon.



================
Comment at: llvm/test/Transforms/InstCombine/memcmp-6.ll:13
+
+; Exercise memcmp(A, B, N) folding of arrays that differ in a charcater
+; with the opposite sign.
----------------
charcater -> character


================
Comment at: llvm/test/Transforms/InstCombine/memcmp-6.ll:16
+
+define i32 @fold_memcmp_a_b_n(i32* %pcmp, i64 %n) {
+; CHECK-LABEL: @fold_memcmp_a_b_n(
----------------
The pcmp argument is dead.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D127766/new/

https://reviews.llvm.org/D127766