[all-commits] [llvm/llvm-project] 6bf7c3: [X86] Add test case to show incorrect formation of...

[topperc via All-commits]

  Branch: refs/heads/master
  Home:   https://github.com/llvm/llvm-project
  Commit: 6bf7c345f82a06107a541db80f1ae128844c307c
      https://github.com/llvm/llvm-project/commit/6bf7c345f82a06107a541db80f1ae128844c307c
  Author: Craig Topper <craig.topper at intel.com>
  Date:   2019-12-18 (Wed, 18 Dec 2019)

  Changed paths:
    M llvm/test/CodeGen/X86/vector-constrained-fp-intrinsics.ll

  Log Message:
  -----------
  [X86] Add test case to show incorrect formation of masked add from vselect+strict_fadd.

The masked operation will suppress exceptions on masked out lanes.
But that's not what vselect+strict_fadd means.


  Commit: f0df4218b67d0abe96867804b8932b9b88998f51
      https://github.com/llvm/llvm-project/commit/f0df4218b67d0abe96867804b8932b9b88998f51
  Author: Craig Topper <craig.topper at intel.com>
  Date:   2019-12-18 (Wed, 18 Dec 2019)

  Changed paths:
    M llvm/lib/Target/X86/X86ISelDAGToDAG.cpp
    M llvm/test/CodeGen/X86/vector-constrained-fp-intrinsics.ll

  Log Message:
  -----------
  [X86] Add a simple hack to IsProfitableToFold to prevent vselect+strict fp operations from being folded into masked instructions.

We really need to update the isel patterns to prevent this, but
that requires some tablegen de-tangling. So this hack will work
for correctness in the short term.


Compare: https://github.com/llvm/llvm-project/compare/98e8f774eb6c...f0df4218b67d