projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 7f63e8c) | patch
[X86][SSE] LowerBuildVectorv4x32 - don't insert MOVQ for undef elts
authorSimon Pilgrim <llvm-dev@redking.me.uk>
Mon, 13 May 2019 16:10:11 +0000 (16:10 +0000)
committerSimon Pilgrim <llvm-dev@redking.me.uk>
Mon, 13 May 2019 16:10:11 +0000 (16:10 +0000)
commit73aee29095fa5715a73b900c7cc223534504b770
treeb1723a082bb543ef13593f4c531bee1737281000tree | snapshot
parent7f63e8c0a69e8a28c221e1bf772d35f25f1f48fccommit | diff
[X86][SSE] LowerBuildVectorv4x32 - don't insert MOVQ for undef elts

Fixes the regression noted in D61782 where a VZEXT_MOVL was being inserted because we weren't discriminating between 'zeroable' and 'all undef' for the upper elts.

Differential Revision: https://reviews.llvm.org/D61782

llvm-svn: 360596
llvm/lib/Target/X86/X86ISelLowering.cpp diff | blob | history
llvm/test/CodeGen/X86/phaddsub-undef.ll diff | blob | history
Domain: System / Toolchain;