[all-commits] [llvm/llvm-project] a58522: Prevent LICM and machineLICM from hoisting converg...

[qcolombet via All-commits]

  Branch: refs/heads/master
  Home:   https://github.com/llvm/llvm-project
  Commit: a585228027a0f025011d77916234cfba12e44309
      https://github.com/llvm/llvm-project/commit/a585228027a0f025011d77916234cfba12e44309
  Author: Quentin Colombet <qcolombet at apple.com>
  Date:   2020-11-06 (Fri, 06 Nov 2020)

  Changed paths:
    M llvm/lib/CodeGen/MachineLICM.cpp
    M llvm/lib/Transforms/Scalar/LICM.cpp
    A llvm/test/CodeGen/AMDGPU/machinelicm-convergent.mir
    A llvm/test/Transforms/LICM/convergent.ll

  Log Message:
  -----------
  Prevent LICM and machineLICM from hoisting convergent operations

Results of convergent operations are implicitly affected by the
enclosing control flows and should not be hoisted out of arbitrary
loops.

Patch by Xiaoqing Wu <xiaoqing_wu at apple.com>

Differential Revision: https://reviews.llvm.org/D90361