[all-commits] [llvm/llvm-project] b62de2: [X86] Move '0-x == y --> x+y == 0' and similar com...

topperc via All-commits all-commits at lists.llvm.org
Mon Feb 17 11:29:16 PST 2020


  Branch: refs/heads/master
  Home:   https://github.com/llvm/llvm-project
  Commit: b62de210cf50ccb6822260e4075dd93333adb23e
      https://github.com/llvm/llvm-project/commit/b62de210cf50ccb6822260e4075dd93333adb23e
  Author: Craig Topper <craig.topper at gmail.com>
  Date:   2020-02-17 (Mon, 17 Feb 2020)

  Changed paths:
    M llvm/lib/Target/X86/X86ISelLowering.cpp

  Log Message:
  -----------
  [X86] Move '0-x == y --> x+y == 0' and similar combines to EmitCmp.

AArch64 handles this pattern in their lowering code. By emitting
CMN. ARM handles it as an isel pattern.


  Commit: 43e948c4b77f1da88511bac38cfcf4752bc19d37
      https://github.com/llvm/llvm-project/commit/43e948c4b77f1da88511bac38cfcf4752bc19d37
  Author: Craig Topper <craig.topper at gmail.com>
  Date:   2020-02-17 (Mon, 17 Feb 2020)

  Changed paths:
    M llvm/lib/Target/X86/X86ISelLowering.cpp
    M llvm/test/CodeGen/X86/flt-rounds.ll

  Log Message:
  -----------
  [X86] Change how the alignment for the stack object is created in LowerFLT_ROUNDS_.

We don't need FrameInfo's concept of the stack alignment. We just
need to tell it the desired alignment. Which in this case is 2.


Compare: https://github.com/llvm/llvm-project/compare/28213680b2a7...43e948c4b77f


More information about the All-commits mailing list