This is an archive of the discontinued LLVM Phabricator instance.

[X86] Add support for _MM_ALIGN16
ClosedPublic

Authored by mkuper on Aug 4 2015, 5:30 AM.

Details

Summary

Both MSVC and ICC define this in xmmintrin.h, so, in terms of compatibility, I'd prefer to have it there in clang as well, even though it's MS-specific.

Diff Detail

Repository
rL LLVM

Event Timeline

mkuper updated this revision to Diff 31317.Aug 4 2015, 5:30 AM
mkuper retitled this revision from to [X86] Add support for _MM_ALIGN16.
mkuper updated this object.
mkuper added reviewers: majnemer, echristo.
mkuper added a subscriber: cfe-commits.
echristo accepted this revision.Aug 5 2015, 11:19 AM
echristo edited edge metadata.

Couple of inline comments, otherwise LGTM.

-eric

test/CodeGen/ms-mm-align.c
1 ↗(On Diff #31317)

Any chance of putting thi with the rest of the Intrin.h checks?

2 ↗(On Diff #31317)

Shouldn't need -Oz.

This revision is now accepted and ready to land.Aug 5 2015, 11:19 AM

Thanks, Eric!

test/CodeGen/ms-mm-align.c
1 ↗(On Diff #31317)

I've originally wanted to put it in ms-intinsics.c, but it fails for the "-triple thumbv7--windows" test (as it should - this is an x86-only macro).
If you have a different test in mind, I'll move it there.

2 ↗(On Diff #31317)

Copied that from ms-intrinsics.c as well. Will remove.

echristo added inline comments.Aug 5 2015, 12:50 PM
test/CodeGen/ms-mm-align.c
1 ↗(On Diff #31317)

No that seems reasonable.

2 ↗(On Diff #31317)

Feel free to remove it from the other file as well.

This revision was automatically updated to reflect the committed changes.