[x86] Fix the new vector shuffle lowering's use of VSELECT for AVX2

lowerings.

This was hopelessly broken. First, the x86 backend wants '-1' to be the
element value representing true in a boolean vector, and second the
operand order for VSELECT is backwards from the actual x86 instructions.
To make matters worse, the backend is just using '-1' as the true value
to get the high bit to be set. It doesn't actually symbolically map the
'-1' to anything. But on x86 this isn't quite how it works: there *only*
the high bit is relevant. As a consequence weird non-'-1' values like
0x80 actually "work" once you flip the operands to be backwards.

Anyways, thanks to Hal for helping me sort out what these *should* be.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@218582 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Target/X86/X86ISelLowering.cpp b/lib/Target/X86/X86ISelLowering.cpp
index c74a84665fb968f1584ffd3ac9802ad20b0906ea..0cda9103337b31b6fc67e95b30f9ae10b40c0e54 100644 (file)
--- a/lib/Target/X86/X86ISelLowering.cpp
+++ b/lib/Target/X86/X86ISelLowering.cpp
@@ -7379,22 +7379,34 @@ static SDValue lowerVectorShuffleAsBlend(SDLoc DL, MVT VT, SDValue V1,
     // FALLTHROUGH
   case MVT::v32i8: {
     assert(Subtarget->hasAVX2() && "256-bit integer blends require AVX2!");
-    SDValue PBLENDVMask[32];
     // Scale the blend by the number of bytes per element.
     int Scale =  VT.getScalarSizeInBits() / 8;
     assert(Mask.size() * Scale == 32 && "Not a 256-bit vector!");
+
+    // Compute the VSELECT mask. Note that VSELECT is really confusing in the
+    // mix of LLVM's code generator and the x86 backend. We tell the code
+    // generator that boolean values in the elements of an x86 vector register
+    // are -1 for true and 0 for false. We then use the LLVM semantics of 'true'
+    // mapping a select to operand #1, and 'false' mapping to operand #2. The
+    // reality in x86 is that vector masks (pre-AVX-512) use only the high bit
+    // of the element (the remaining are ignored) and 0 in that high bit would
+    // mean operand #1 while 1 in the high bit would mean operand #2. So while
+    // the LLVM model for boolean values in vector elements gets the relevant
+    // bit set, it is set backwards and over constrained relative to x86's
+    // actual model.
+    SDValue VSELECTMask[32];
     for (int i = 0, Size = Mask.size(); i < Size; ++i)
       for (int j = 0; j < Scale; ++j)
-        PBLENDVMask[Scale * i + j] =
+        VSELECTMask[Scale * i + j] =
             Mask[i] < 0 ? DAG.getUNDEF(MVT::i8)
-                        : DAG.getConstant(Mask[i] < Size ? 0 : 0x80, MVT::i8);
+                        : DAG.getConstant(Mask[i] < Size ? -1 : 0, MVT::i8);
 
     V1 = DAG.getNode(ISD::BITCAST, DL, MVT::v32i8, V1);
     V2 = DAG.getNode(ISD::BITCAST, DL, MVT::v32i8, V2);
     return DAG.getNode(
         ISD::BITCAST, DL, VT,
         DAG.getNode(ISD::VSELECT, DL, MVT::v32i8,
-                    DAG.getNode(ISD::BUILD_VECTOR, DL, MVT::v32i8, PBLENDVMask),
+                    DAG.getNode(ISD::BUILD_VECTOR, DL, MVT::v32i8, VSELECTMask),
                     V1, V2));
   }
 

diff --git a/test/CodeGen/X86/vector-shuffle-256-v16.ll b/test/CodeGen/X86/vector-shuffle-256-v16.ll
index f9d7299cdc464bab4b2617de5d61cada75fd982c..ad6eca3b926584b7b1042f2645493ab71cb0e9ea 100644 (file)
--- a/test/CodeGen/X86/vector-shuffle-256-v16.ll
+++ b/test/CodeGen/X86/vector-shuffle-256-v16.ll
@@ -166,7 +166,7 @@ define <16 x i16> @shuffle_v16i16_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_0
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm1 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm1 = ymm1[u,u,u,u,u,u,u,u,u,u,u,u,u,u,0,1,16,17,16,17,16,17,16,17,16,17,16,17,16,17,16,17]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[0,1,0,1,0,1,0,1,0,1,0,1,0,1,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 8, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -190,7 +190,7 @@ define <16 x i16> @shuffle_v16i16_00_00_00_00_00_00_09_00_00_00_00_00_00_00_00_0
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm1 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm1 = ymm1[u,u,u,u,u,u,u,u,u,u,u,u,2,3,u,u,16,17,16,17,16,17,16,17,16,17,16,17,16,17,16,17]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[0,1,0,1,0,1,0,1,0,1,0,1,u,u,0,1,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,128,128,128,128,128,0,0,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,0,0,0,0,0,0,0,0,0,0,255,255,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 9, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -215,7 +215,7 @@ define <16 x i16> @shuffle_v16i16_00_00_00_00_00_10_00_00_00_00_00_00_00_00_00_0
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm1 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm1 = ymm1[u,u,u,u,u,u,u,u,u,u,4,5,u,u,u,u,16,17,16,17,16,17,16,17,16,17,16,17,16,17,16,17]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[0,1,0,1,0,1,0,1,0,1,u,u,0,1,0,1,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,128,128,128,0,0,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,0,0,0,0,0,0,0,0,255,255,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 10, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -240,7 +240,7 @@ define <16 x i16> @shuffle_v16i16_00_00_00_00_11_00_00_00_00_00_00_00_00_00_00_0
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm1 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm1 = ymm1[u,u,u,u,u,u,u,u,6,7,u,u,u,u,u,u,16,17,16,17,16,17,16,17,16,17,16,17,16,17,16,17]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[0,1,0,1,0,1,0,1,u,u,0,1,0,1,0,1,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,128,0,0,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,0,0,0,0,0,0,255,255,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 11, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -265,7 +265,7 @@ define <16 x i16> @shuffle_v16i16_00_00_00_12_00_00_00_00_00_00_00_00_00_00_00_0
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm1 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm1 = ymm1[u,u,u,u,u,u,8,9,u,u,u,u,u,u,u,u,16,17,16,17,16,17,16,17,16,17,16,17,16,17,16,17]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[0,1,0,1,0,1,u,u,0,1,0,1,0,1,0,1,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,0,0,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,0,0,0,0,255,255,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 0, i32 0, i32 12, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -290,7 +290,7 @@ define <16 x i16> @shuffle_v16i16_00_00_13_00_00_00_00_00_00_00_00_00_00_00_00_0
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm1 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm1 = ymm1[u,u,u,u,10,11,u,u,u,u,u,u,u,u,u,u,16,17,16,17,16,17,16,17,16,17,16,17,16,17,16,17]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[0,1,0,1,u,u,0,1,0,1,0,1,0,1,0,1,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,128,128,0,0,128,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,0,0,255,255,0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 0, i32 13, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -316,7 +316,7 @@ define <16 x i16> @shuffle_v16i16_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_0
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm1 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm1 = ymm1[u,u,12,13,u,u,u,u,u,u,u,u,u,u,u,u,16,17,16,17,16,17,16,17,16,17,16,17,16,17,16,17]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[0,1,u,u,0,1,0,1,0,1,0,1,0,1,0,1,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,0,0,128,128,128,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,255,255,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 14, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -342,7 +342,7 @@ define <16 x i16> @shuffle_v16i16_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_0
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm1 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm1 = ymm1[14,15,u,u,u,u,u,u,u,u,u,u,u,u,u,u,16,17,16,17,16,17,16,17,16,17,16,17,16,17,16,17]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[u,u,0,1,0,1,0,1,0,1,0,1,0,1,0,1,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,128,128,128,128,128,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [255,255,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 15, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -648,7 +648,7 @@ define <16 x i16> @shuffle_v16i16_00_01_02_03_04_05_06_07_08_09_10_11_12_13_14_3
 ;
 ; AVX2-LABEL: @shuffle_v16i16_00_01_02_03_04_05_06_07_08_09_10_11_12_13_14_31
 ; AVX2:       # BB#0:
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,128,128]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 31>
@@ -665,7 +665,7 @@ define <16 x i16> @shuffle_v16i16_16_01_02_03_04_05_06_07_08_09_10_11_12_13_14_1
 ;
 ; AVX2-LABEL: @shuffle_v16i16_16_01_02_03_04_05_06_07_08_09_10_11_12_13_14_15
 ; AVX2:       # BB#0:
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 16, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
@@ -684,7 +684,7 @@ define <16 x i16> @shuffle_v16i16_00_17_02_19_04_21_06_23_24_09_26_11_28_13_30_1
 ;
 ; AVX2-LABEL: @shuffle_v16i16_00_17_02_19_04_21_06_23_24_09_26_11_28_13_30_15
 ; AVX2:       # BB#0:
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,128,128,0,0,128,128,0,0,128,128,0,0,128,128,128,128,0,0,128,128,0,0,128,128,0,0,128,128,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [255,255,0,0,255,255,0,0,255,255,0,0,255,255,0,0,0,0,255,255,0,0,255,255,0,0,255,255,0,0,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 24, i32 9, i32 26, i32 11, i32 28, i32 13, i32 30, i32 15>
@@ -703,7 +703,7 @@ define <16 x i16> @shuffle_v16i16_16_01_18_03_20_05_22_07_08_25_10_27_12_29_14_3
 ;
 ; AVX2-LABEL: @shuffle_v16i16_16_01_18_03_20_05_22_07_08_25_10_27_12_29_14_31
 ; AVX2:       # BB#0:
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,0,0,128,128,0,0,128,128,0,0,128,128,0,0,0,0,128,128,0,0,128,128,0,0,128,128,0,0,128,128]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,255,255,0,0,255,255,0,0,255,255,0,0,255,255,255,255,0,0,255,255,0,0,255,255,0,0,255,255,0,0]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 16, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>

diff --git a/test/CodeGen/X86/vector-shuffle-256-v32.ll b/test/CodeGen/X86/vector-shuffle-256-v32.ll
index afe04589f15efa9dd23fb23d7c0cb447c4226721..338053ad32c8835fa851410326ea4dd16796740c 100644 (file)
--- a/test/CodeGen/X86/vector-shuffle-256-v32.ll
+++ b/test/CodeGen/X86/vector-shuffle-256-v32.ll
@@ -343,7 +343,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_16_
 ; AVX2-NEXT:    vpshufb %ymm1, %ymm0, %ymm2
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb %ymm1, %ymm0, %ymm0
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm1 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm1, %ymm0, %ymm2, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 16, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -368,7 +368,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_17_00_
 ; AVX2-NEXT:    vpshufb %ymm1, %ymm0, %ymm1
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,u,u,u,u,u,u,u,u,1,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,0,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 17, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -393,7 +393,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_18_00_00_
 ; AVX2-NEXT:    vpshufb %ymm1, %ymm0, %ymm1
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,u,u,u,u,u,u,u,2,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,128,128,128,128,128,128,0,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,0,0,0,0,0,0,0,0,0,0,0,255,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 18, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -418,7 +418,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_19_00_00_00_
 ; AVX2-NEXT:    vpshufb %ymm1, %ymm0, %ymm1
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,u,u,u,u,u,u,3,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,128,128,128,128,128,0,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,0,0,0,0,0,0,0,0,0,0,255,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 19, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -443,7 +443,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_20_00_00_00_00_
 ; AVX2-NEXT:    vpshufb %ymm1, %ymm0, %ymm1
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,u,u,u,u,u,4,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,128,128,128,128,0,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,0,0,0,0,0,0,0,0,0,255,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 20, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -468,7 +468,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_21_00_00_00_00_00_
 ; AVX2-NEXT:    vpshufb %ymm1, %ymm0, %ymm1
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,u,u,u,u,5,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,128,128,128,0,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,0,0,0,0,0,0,0,0,255,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 21, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -493,7 +493,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_22_00_00_00_00_00_00_
 ; AVX2-NEXT:    vpshufb %ymm1, %ymm0, %ymm1
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,u,u,u,6,u,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,128,128,0,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,0,0,0,0,0,0,0,255,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 22, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -518,7 +518,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_23_00_00_00_00_00_00_00_
 ; AVX2-NEXT:    vpshufb %ymm1, %ymm0, %ymm1
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,u,u,7,u,u,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,128,0,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,0,0,0,0,0,0,255,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 23, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -543,7 +543,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_24_00_00_00_00_00_00_00_00_
 ; AVX2-NEXT:    vpshufb %ymm1, %ymm0, %ymm1
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,u,8,u,u,u,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,0,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,0,0,0,0,0,255,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 24, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -568,7 +568,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_25_00_00_00_00_00_00_00_00_00_
 ; AVX2-NEXT:    vpshufb %ymm1, %ymm0, %ymm1
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,9,u,u,u,u,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,0,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,0,0,0,0,255,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 25, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -593,7 +593,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_26_00_00_00_00_00_00_00_00_00_00_
 ; AVX2-NEXT:    vpshufb %ymm1, %ymm0, %ymm1
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,10,u,u,u,u,u,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,0,128,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,0,0,0,255,0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 26, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -618,7 +618,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_27_00_00_00_00_00_00_00_00_00_00_00_
 ; AVX2-NEXT:    vpshufb %ymm1, %ymm0, %ymm1
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,11,u,u,u,u,u,u,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,128,128,0,128,128,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,0,0,255,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 27, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -643,7 +643,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_28_00_00_00_00_00_00_00_00_00_00_00_00_
 ; AVX2-NEXT:    vpshufb %ymm1, %ymm0, %ymm1
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[u,u,u,12,u,u,u,u,u,u,u,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,128,0,128,128,128,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,0,255,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 28, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -668,7 +668,7 @@ define <32 x i8> @shuffle_v32i8_00_00_29_00_00_00_00_00_00_00_00_00_00_00_00_00_
 ; AVX2-NEXT:    vpshufb %ymm1, %ymm0, %ymm1
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[u,u,13,u,u,u,u,u,u,u,u,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,128,0,128,128,128,128,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,255,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 29, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -693,7 +693,7 @@ define <32 x i8> @shuffle_v32i8_00_30_00_00_00_00_00_00_00_00_00_00_00_00_00_00_
 ; AVX2-NEXT:    vpshufb %ymm1, %ymm0, %ymm1
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[u,14,u,u,u,u,u,u,u,u,u,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [128,0,128,128,128,128,128,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,255,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 30, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -723,7 +723,7 @@ define <32 x i8> @shuffle_v32i8_31_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_
 ; AVX2-NEXT:    vinserti128 $0, %xmm2, %ymm3, %ymm2
 ; AVX2-NEXT:    vpshufb %ymm2, %ymm1, %ymm1
 ; AVX2-NEXT:    vpshufb %ymm3, %ymm0, %ymm0
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [255,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 31, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -993,7 +993,7 @@ define <32 x i8> @shuffle_v32i8_00_33_02_35_04_37_06_39_08_41_10_43_12_45_14_47_
 ;
 ; AVX2-LABEL: @shuffle_v32i8_00_33_02_35_04_37_06_39_08_41_10_43_12_45_14_47_16_49_18_51_20_53_22_55_24_57_26_59_28_61_30_63
 ; AVX2:       # BB#0:
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 33, i32 2, i32 35, i32 4, i32 37, i32 6, i32 39, i32 8, i32 41, i32 10, i32 43, i32 12, i32 45, i32 14, i32 47, i32 16, i32 49, i32 18, i32 51, i32 20, i32 53, i32 22, i32 55, i32 24, i32 57, i32 26, i32 59, i32 28, i32 61, i32 30, i32 63>
@@ -1018,7 +1018,7 @@ define <32 x i8> @shuffle_v32i8_32_01_34_03_36_05_38_07_40_09_42_11_44_13_46_15_
 ;
 ; AVX2-LABEL: @shuffle_v32i8_32_01_34_03_36_05_38_07_40_09_42_11_44_13_46_15_48_17_50_19_52_21_54_23_56_25_58_27_60_29_62_31
 ; AVX2:       # BB#0:
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 32, i32 1, i32 34, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 29, i32 62, i32 31>
@@ -1067,7 +1067,7 @@ define <32 x i8> @shuffle_v32i8_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_
 ; AVX2-NEXT:    vpxor %ymm2, %ymm2, %ymm2
 ; AVX2-NEXT:    vpshufb %ymm2, %ymm1, %ymm1
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[0,1,0,1,0,1,0,1,0,1,0,1,0,1,0,1,16,17,16,17,16,17,16,17,16,17,16,17,16,17,16,17]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 16, i32 48, i32 16, i32 48, i32 16, i32 48, i32 16, i32 48, i32 16, i32 48, i32 16, i32 48, i32 16, i32 48, i32 16, i32 48>
@@ -1324,7 +1324,7 @@ define <32 x i8> @shuffle_v32i8_00_32_01_33_02_34_03_35_04_36_05_37_06_38_07_39_
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[0,u,1,u,2,u,3,u,4,u,5,u,6,u,7,u,24,u,25,u,26,u,27,u,28,u,29,u,30,u,31,u]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm1 = ymm1[u,0,u,1,u,2,u,3,u,4,u,5,u,6,u,7,u,24,u,25,u,26,u,27,u,28,u,29,u,30,u,31]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 32, i32 1, i32 33, i32 2, i32 34, i32 3, i32 35, i32 4, i32 36, i32 5, i32 37, i32 6, i32 38, i32 7, i32 39, i32 24, i32 56, i32 25, i32 57, i32 26, i32 58, i32 27, i32 59, i32 28, i32 60, i32 29, i32 61, i32 30, i32 62, i32 31, i32 63>
@@ -1350,7 +1350,7 @@ define <32 x i8> @shuffle_v32i8_08_40_09_41_10_42_11_43_12_44_13_45_14_46_15_47_
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[8,u,9,u,10,u,11,u,12,u,13,u,14,u,15,u,16,u,17,u,18,u,19,u,20,u,21,u,22,u,23,u]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm1 = ymm1[u,8,u,9,u,10,u,11,u,12,u,13,u,14,u,15,u,16,u,17,u,18,u,19,u,20,u,21,u,22,u,23]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 8, i32 40, i32 9, i32 41, i32 10, i32 42, i32 11, i32 43, i32 12, i32 44, i32 13, i32 45, i32 14, i32 46, i32 15, i32 47, i32 16, i32 48, i32 17, i32 49, i32 18, i32 50, i32 19, i32 51, i32 20, i32 52, i32 21, i32 53, i32 22, i32 54, i32 23, i32 55>
@@ -1594,13 +1594,13 @@ define <32 x i8> @shuffle_v32i8_42_45_12_13_35_35_60_40_17_22_29_44_33_12_48_51_
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm2 = ymm1[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm2 = ymm2[u,u,u,u,u,u,12,u,u,u,u,u,u,u,0,3,u,u,u,u,u,u,21,16,u,26,u,u,20,18,20,23]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm1 = ymm1[10,13,u,u,3,3,u,8,u,u,u,12,1,u,u,u,u,u,20,u,17,22,u,u,16,u,27,u,u,u,u,u]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm3 = <0,0,u,u,0,0,128,0,u,u,u,0,0,u,128,128,u,u,0,u,0,0,128,128,0,128,0,u,128,128,128,128>
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm3 = <255,255,u,u,255,255,0,255,u,u,u,255,255,u,0,0,u,u,255,u,255,255,0,0,255,0,255,u,0,0,0,0>
 ; AVX2-NEXT:    vpblendvb %ymm3, %ymm1, %ymm2, %ymm1
 ; AVX2-NEXT:    vperm2i128 {{.*}} # ymm2 = ymm0[2,3,0,1]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm2 = ymm2[u,u,u,u,u,u,u,u,1,6,13,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,23,u,u,u,u]
 ; AVX2-NEXT:    vpshufb {{.*}} # ymm0 = ymm0[u,u,12,13,u,u,u,u,u,u,u,u,u,12,u,u,20,19,u,19,u,u,u,u,u,u,u,u,u,u,u,u]
 ; AVX2-NEXT:    vpblendd {{.*}} # ymm0 = ymm0[0,1],ymm2[2],ymm0[3,4,5],ymm2[6],ymm0[7]
-; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [0,0,128,128,0,0,0,0,128,128,128,0,0,128,0,0,128,128,0,128,0,0,0,0,0,0,0,128,0,0,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # ymm2 = [255,255,0,0,255,255,255,255,0,0,0,255,255,0,255,255,0,0,255,0,255,255,255,255,255,255,255,0,255,255,255,255]
 ; AVX2-NEXT:    vpblendvb %ymm2, %ymm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 42, i32 45, i32 12, i32 13, i32 35, i32 35, i32 60, i32 40, i32 17, i32 22, i32 29, i32 44, i32 33, i32 12, i32 48, i32 51, i32 20, i32 19, i32 52, i32 19, i32 49, i32 54, i32 37, i32 32, i32 48, i32 42, i32 59, i32 7, i32 36, i32 34, i32 36, i32 39>