[all-commits] [llvm/llvm-project] d04e64: [InstCombine] Add tests for PR44242

[Connor Abbott via All-commits]

  Branch: refs/heads/master
  Home:   https://github.com/llvm/llvm-project
  Commit: d04e64a25a58f63af862df9fa305723560d5fded
      https://github.com/llvm/llvm-project/commit/d04e64a25a58f63af862df9fa305723560d5fded
  Author: Connor Abbott <cwabbott0 at gmail.com>
  Date:   2019-12-31 (Tue, 31 Dec 2019)

  Changed paths:
    A llvm/test/Transforms/InstCombine/pr44242.ll

  Log Message:
  -----------
  [InstCombine] Add tests for PR44242

Differential Revision: https://reviews.llvm.org/D71260


  Commit: fb114694e939c0204ac356fc0e830332175cd008
      https://github.com/llvm/llvm-project/commit/fb114694e939c0204ac356fc0e830332175cd008
  Author: Connor Abbott <cwabbott0 at gmail.com>
  Date:   2019-12-31 (Tue, 31 Dec 2019)

  Changed paths:
    M llvm/lib/Transforms/InstCombine/InstCombineCasts.cpp
    M llvm/test/Transforms/InstCombine/pr44242.ll

  Log Message:
  -----------
  [InstCombine] Don't rewrite phi-of-bitcast when the phi has other users

Judging by the existing comments, this was the intention, but the
transform never actually checked if the existing phi's would be removed.
See https://bugs.llvm.org/show_bug.cgi?id=44242 for an example where
this causes much worse code generation on AMDGPU.

Differential Revision: https://reviews.llvm.org/D71209


Compare: https://github.com/llvm/llvm-project/compare/e8c9110b56b5...fb114694e939