This is an archive of the discontinued LLVM Phabricator instance.

AMDGPU/SI: Fix s_waitcnt insertion for flat instructions
ClosedPublic

Authored by tstellarAMD on Feb 17 2016, 8:16 PM.

Details

Summary

This was broken in r260694 which swapped the address and data operands
for flat store instructions. The code in SIInsertWaits assumes
that the data operand always comes before the address operand, so
we need to add a special case for flat.

Diff Detail

Repository
rL LLVM

Event Timeline

tstellarAMD retitled this revision from to AMDGPU/SI: Fix s_waitcnt insertion for flat instructions.
tstellarAMD updated this object.
tstellarAMD added a reviewer: arsenm.
tstellarAMD added a subscriber: llvm-commits.
arsenm accepted this revision.Feb 17 2016, 8:55 PM
arsenm edited edge metadata.

LGTM. A test seems hard, but do you have a small reduced testcase?

This revision is now accepted and ready to land.Feb 17 2016, 8:55 PM

LGTM. A test seems hard, but do you have a small reduced testcase?

No, one of the C++AMP test hit this and the test case is really huge. I can try to write a small one.

This revision was automatically updated to reflect the committed changes.