diff options
| author | Artem Tamazov <artem.tamazov@amd.com> | 2016-05-19 12:22:39 +0000 |
|---|---|---|
| committer | Artem Tamazov <artem.tamazov@amd.com> | 2016-05-19 12:22:39 +0000 |
| commit | 8ce1f7177b34111466c9d9d7b568e6ec68e1fd53 (patch) | |
| tree | 765545124a0badeea5415fd1b9e7e04d069fd9c0 /clang/lib/Frontend/CompilerInvocation.cpp | |
| parent | 5f94cedeb531f2328403abc2d983a724cf9dcb4a (diff) | |
| download | llvm-8ce1f7177b34111466c9d9d7b568e6ec68e1fd53.zip llvm-8ce1f7177b34111466c9d9d7b568e6ec68e1fd53.tar.gz llvm-8ce1f7177b34111466c9d9d7b568e6ec68e1fd53.tar.bz2 | |
[AMDGPU][llvm-mc] Fixes to support buffer atomics.
Fixes for MUBUF_Atomic instructions to make operand list valid:
- For RTN insns, make a copy of $vdata_in operand as $vdata.
- Do not add operand for GLC, it is hardcoded and comes as a token.
Workaround to avoid adding multiple default optional operands.
Tests added.
Differential Revision: http://reviews.llvm.org/D20257
llvm-svn: 270049
Diffstat (limited to 'clang/lib/Frontend/CompilerInvocation.cpp')
0 files changed, 0 insertions, 0 deletions