projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 2678d02) | patch
AMDGPU/SI: Don't use non-0 waitcnt values when waiting on Flat instructions
authorTom Stellard <thomas.stellard@amd.com>
Fri, 28 Oct 2016 23:53:48 +0000 (23:53 +0000)
committerTom Stellard <thomas.stellard@amd.com>
Fri, 28 Oct 2016 23:53:48 +0000 (23:53 +0000)
commit6695ba0440f38c4cf0ccddd514b079b96786e352
treeb68c6ba36c018659608394ff1e7ba45a74e18f28tree | snapshot
parent2678d023ca5b02dbecd65a88520366523e1ae6decommit | diff
AMDGPU/SI: Don't use non-0 waitcnt values when waiting on Flat instructions

Summary:
Flat instruction can return out of order, so we need always need to wait
for all the outstanding flat operations.

Reviewers: tony-tye, arsenm

Subscribers: kzhuravl, wdng, nhaehnle, llvm-commits, yaxunl

Differential Revision: https://reviews.llvm.org/D25998

llvm-svn: 285479
llvm/lib/Target/AMDGPU/SIInsertWaits.cpp diff | blob | history
llvm/lib/Target/AMDGPU/SIInstrInfo.cpp diff | blob | history
llvm/lib/Target/AMDGPU/SIInstrInfo.h diff | blob | history
llvm/test/CodeGen/MIR/AMDGPU/waitcnt.mir [new file with mode: 0644] blob
Domain: System / Toolchain;