[PATCH] D20036: [InstCombine] Fold icmp eq/ne (udiv i32 A, B), 0 -> icmp ugt/ule B, A

[Chad Rosier via llvm-commits]

mcrosier added a comment.

Committed in r268960.  I did have to move the optimizations after visitICmpInstWithInstAndIntCst() in order to not regress udiv_icmp1 in exact.ll.  Also, I added an additional test case, udiv_icmp2, to provide more context and to improve coverage.


Repository:
  rL LLVM

http://reviews.llvm.org/D20036