This is an archive of the discontinued LLVM Phabricator instance.

Ignore device-side asm constraint errors while compiling CUDA code for host and vice versa.
ClosedPublic

Authored by tra on Mar 17 2015, 1:38 PM.

Diff Detail

Repository
rL LLVM

Event Timeline

tra updated this revision to Diff 22118.Mar 17 2015, 1:38 PM
tra retitled this revision from to Ignore device-side asm constraint errors while compiling CUDA code for host and vice versa..
tra updated this object.
tra edited the test plan for this revision. (Show Details)
tra added reviewers: echristo, eliben.
tra added a subscriber: Unknown Object (MLST).
eliben added inline comments.Mar 18 2015, 7:54 AM
test/SemaCUDA/asm-constraints.cu
7 ↗(On Diff #22118)

Are there checks anywhere that asm diagnostics are validated within device?

Perhaps a positive test would be helpful (make sure that an invalid constraint produces a diagnostic within a kernel)

tra added a comment.Mar 18 2015, 7:55 AM

Good point. Will do.

echristo accepted this revision.Mar 18 2015, 11:27 AM
echristo edited edge metadata.

As Eli said a positive test would be nice, otherwise LGTM.

-eric

This revision is now accepted and ready to land.Mar 18 2015, 11:27 AM
tra updated this revision to Diff 22211.Mar 18 2015, 2:21 PM
tra edited edge metadata.

Added a test to verify that we do validate asm constraints in CUDA device compilation mode.
Addressed echristo@'s offline style comment.

LGTM.

Thanks!

-eric

eliben accepted this revision.Mar 18 2015, 4:53 PM
eliben edited edge metadata.

LGTM

This revision was automatically updated to reflect the committed changes.