[x86] Fix a bug in the v8i16 shuffling exposed by the new splat-like

lowering for v16i8.

ASan and some bots caught this bug with existing test cases. Fixing it
even fixed a miscompile with one of the test cases. I'm still a bit
suspicious of this test case as I've not taken a proper amount of time
to think about it, but the fix here is strict goodness.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211976 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Target/X86/X86ISelLowering.cpp b/lib/Target/X86/X86ISelLowering.cpp
index 8a00d6e2bc8f30fb0b05f25ccddb836ed2c7b2eb..8a12c03e072efff4e15156cd5270984f4e4b6c32 100644 (file)
--- a/lib/Target/X86/X86ISelLowering.cpp
+++ b/lib/Target/X86/X86ISelLowering.cpp
@@ -7535,7 +7535,7 @@ static SDValue lowerV8I16BasicBlendVectorShuffle(SDLoc DL, SDValue V1,
       } else {
         // Otherwise pin the low inputs.
         for (int GoodInput : GoodInputs)
-          MoveMask[Mask[GoodInput]] = Mask[GoodInput] - MaskOffset;
+          MoveMask[Mask[GoodInput] - MaskOffset] = Mask[GoodInput] - MaskOffset;
       }
 
       int MoveMaskIdx =

diff --git a/test/CodeGen/X86/vector-shuffle-128-v16.ll b/test/CodeGen/X86/vector-shuffle-128-v16.ll
index f3a0931f731207de1fb4932015bfc3f39c09892f..f392a6221c5f71a4a038a4ea442a9e01bc31f072 100644 (file)
--- a/test/CodeGen/X86/vector-shuffle-128-v16.ll
+++ b/test/CodeGen/X86/vector-shuffle-128-v16.ll
@@ -63,14 +63,15 @@ define <16 x i8> @shuffle_v16i8_00_00_00_00_04_04_04_04_08_08_08_08_12_12_12_12(
 ; CHECK-SSE2-NEXT:    pxor %xmm1, %xmm1
 ; CHECK-SSE2-NEXT:    movdqa %xmm0, %xmm2
 ; CHECK-SSE2-NEXT:    punpcklbw %xmm1, %xmm2
-; CHECK-SSE2-NEXT:    pshufd {{.*}} # xmm2 = xmm2[2,1,2,3]
+; CHECK-SSE2-NEXT:    pshufd {{.*}} # xmm2 = xmm2[0,2,2,3]
+; CHECK-SSE2-NEXT:    pshuflw {{.*}} # xmm2 = xmm2[0,2,2,3,4,5,6,7]
 ; CHECK-SSE2-NEXT:    punpckhbw %xmm1, %xmm0
 ; CHECK-SSE2-NEXT:    pshufd {{.*}} # xmm0 = xmm0[0,2,2,3]
 ; CHECK-SSE2-NEXT:    pshuflw {{.*}} # xmm0 = xmm0[0,2,2,3,4,5,6,7]
 ; CHECK-SSE2-NEXT:    punpcklwd %xmm2, %xmm0
-; CHECK-SSE2-NEXT:    pshufd {{.*}} # xmm0 = xmm0[0,1,0,3]
+; CHECK-SSE2-NEXT:    pshufd {{.*}} # xmm0 = xmm0[0,1,2,1]
 ; CHECK-SSE2-NEXT:    pshuflw {{.*}} # xmm0 = xmm0[1,0,2,3,4,5,6,7]
-; CHECK-SSE2-NEXT:    pshufhw {{.*}} # xmm0 = xmm0[0,1,2,3,5,5,6,7]
+; CHECK-SSE2-NEXT:    pshufhw {{.*}} # xmm0 = xmm0[0,1,2,3,7,5,6,7]
 ; CHECK-SSE2-NEXT:    packuswb %xmm0, %xmm0
 ; CHECK-SSE2-NEXT:    punpcklbw %xmm0, %xmm0
 ; CHECK-SSE2-NEXT:    pshufd {{.*}} # xmm0 = xmm0[0,2,0,1]