diff options
| author | Simon Pilgrim <llvm-dev@redking.me.uk> | 2016-04-24 15:05:04 +0000 |
|---|---|---|
| committer | Simon Pilgrim <llvm-dev@redking.me.uk> | 2016-04-24 15:05:04 +0000 |
| commit | f379a6c684d47484fb1a0e47df4321fee21f4f6e (patch) | |
| tree | 3f650beb57bb037c5b3be52570c125af5478fa60 | |
| parent | 03a04a58ea4a1200ce6773735f44c2d26e4d88a5 (diff) | |
| download | bcm5719-llvm-f379a6c684d47484fb1a0e47df4321fee21f4f6e.tar.gz bcm5719-llvm-f379a6c684d47484fb1a0e47df4321fee21f4f6e.zip | |
[X86][XOP] Fixed VPPERM permute op decoding (PR27472).
Fixed issue with VPPERM target shuffle mask decoding that was incorrectly masking off the 3-bit permute op with a 2-bit mask.
llvm-svn: 267346
| -rw-r--r-- | llvm/lib/Target/X86/Utils/X86ShuffleDecode.cpp | 2 | ||||
| -rw-r--r-- | llvm/test/CodeGen/X86/vector-shuffle-combining-xop.ll | 2 |
2 files changed, 2 insertions, 2 deletions
diff --git a/llvm/lib/Target/X86/Utils/X86ShuffleDecode.cpp b/llvm/lib/Target/X86/Utils/X86ShuffleDecode.cpp index d8f8f9e7393..472474d566c 100644 --- a/llvm/lib/Target/X86/Utils/X86ShuffleDecode.cpp +++ b/llvm/lib/Target/X86/Utils/X86ShuffleDecode.cpp @@ -367,7 +367,7 @@ void DecodeVPPERMMask(ArrayRef<uint64_t> RawMask, continue; } - uint64_t PermuteOp = (M >> 5) & 0x3; + uint64_t PermuteOp = (M >> 5) & 0x7; if (PermuteOp == 4) { ShuffleMask.push_back(SM_SentinelZero); continue; diff --git a/llvm/test/CodeGen/X86/vector-shuffle-combining-xop.ll b/llvm/test/CodeGen/X86/vector-shuffle-combining-xop.ll index 0fba8319b59..dda93aca9b8 100644 --- a/llvm/test/CodeGen/X86/vector-shuffle-combining-xop.ll +++ b/llvm/test/CodeGen/X86/vector-shuffle-combining-xop.ll @@ -23,7 +23,7 @@ define <16 x i8> @combine_vpperm_identity(<16 x i8> %a0, <16 x i8> %a1) { define <16 x i8> @combine_vpperm_zero(<16 x i8> %a0, <16 x i8> %a1) { ; CHECK-LABEL: combine_vpperm_zero: ; CHECK: # BB#0: -; CHECK-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero +; CHECK-NEXT: vxorps %xmm0, %xmm0, %xmm0 ; CHECK-NEXT: retq %res0 = call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %a0, <16 x i8> %a1, <16 x i8> <i8 128, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0>) %res1 = call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %res0, <16 x i8> undef, <16 x i8> <i8 0, i8 128, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0>) |
