[all-commits] [llvm/llvm-project] d9ae85: [AMDGPU] Fix data race in SIInsertWaitcnts

[Jakub Kuderski via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: d9ae852fcc97a6a1b48bf7516d6f3f03a85eed62
      https://github.com/llvm/llvm-project/commit/d9ae852fcc97a6a1b48bf7516d6f3f03a85eed62
  Author: Jakub Kuderski <kubak at google.com>
  Date:   2021-12-18 (Sat, 18 Dec 2021)

  Changed paths:
    M llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp

  Log Message:
  -----------
  [AMDGPU] Fix data race in SIInsertWaitcnts

The race condition happened when two pass managers ran on two different modules but modified/read the global variables.

To address this, I considered using singletons and freestanding functions to allow getting/setting `HardwareLimits` and `RegisterEncoding`, or making it local to the pass. I chose the latter and made it a member of `WaitcntsBrackets`, to minimizes the amount of global state.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D115896