Merging r371305 and r371307:

------------------------------------------------------------------------
r371305 | nikic | 2019-09-07 14:03:48 +0200 (Sat, 07 Sep 2019) | 1 line

[X86] Add test for PR43230; NFC
------------------------------------------------------------------------

------------------------------------------------------------------------
r371307 | nikic | 2019-09-07 14:13:44 +0200 (Sat, 07 Sep 2019) | 9 lines

[X86] Fix pshuflw formation from repeated shuffle mask (PR43230)

Fix for https://bugs.llvm.org/show_bug.cgi?id=43230.

When creating PSHUFLW from a repeated shuffle mask, we have to apply
the checks to the repeated mask, not the original one. For the test
case from PR43230 the inspected part of the original mask is all undef.

Differential Revision: https://reviews.llvm.org/D67314
------------------------------------------------------------------------


git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_90@371378 91177308-0d34-0410-b5e6-96231b3b80d8

2 files changed, 43 insertions, 2 deletions

diff --git a/lib/Target/X86/X86ISelLowering.cpp b/lib/Target/X86/X86ISelLowering.cpp
index 87ff5a719fca..e67ad2332159 100644
--- a/lib/Target/X86/X86ISelLowering.cpp
+++ b/lib/Target/X86/X86ISelLowering.cpp
@@ -31664,8 +31664,8 @@ static bool matchUnaryPermuteShuffle(MVT MaskVT, ArrayRef<int> Mask,
   if (!ContainsZeros && AllowIntDomain && MaskScalarSizeInBits == 16) {
     SmallVector<int, 4> RepeatedMask;
     if (is128BitLaneRepeatedShuffleMask(MaskEltVT, Mask, RepeatedMask)) {
-      ArrayRef<int> LoMask(Mask.data() + 0, 4);
-      ArrayRef<int> HiMask(Mask.data() + 4, 4);
+      ArrayRef<int> LoMask(RepeatedMask.data() + 0, 4);
+      ArrayRef<int> HiMask(RepeatedMask.data() + 4, 4);
 
       // PSHUFLW: permute lower 4 elements only.
       if (isUndefOrInRange(LoMask, 0, 4) &&
diff --git a/test/CodeGen/X86/vector-shuffle-256-v16.ll b/test/CodeGen/X86/vector-shuffle-256-v16.ll
index 0be1fec5ecd2..001288f87d77 100644
--- a/test/CodeGen/X86/vector-shuffle-256-v16.ll
+++ b/test/CodeGen/X86/vector-shuffle-256-v16.ll
@@ -4754,3 +4754,44 @@ define <16 x i16> @unpckh_v16i16(<16 x i16> %x, <16 x i16> %y) {
   ret <16 x i16> %unpckh
 }
 
+define <16 x i16> @pr43230(<16 x i16> %a, <16 x i16> %b) {
+; AVX1-LABEL: pr43230:
+; AVX1:       # %bb.0:
+; AVX1-NEXT:    vextractf128 $1, %ymm1, %xmm1
+; AVX1-NEXT:    vpsllw $12, %xmm1, %xmm2
+; AVX1-NEXT:    vpsllw $4, %xmm1, %xmm1
+; AVX1-NEXT:    vpor %xmm2, %xmm1, %xmm1
+; AVX1-NEXT:    vpaddw %xmm1, %xmm1, %xmm2
+; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm0
+; AVX1-NEXT:    vpsrlw $8, %xmm0, %xmm3
+; AVX1-NEXT:    vpblendvb %xmm1, %xmm3, %xmm0, %xmm0
+; AVX1-NEXT:    vpsrlw $4, %xmm0, %xmm1
+; AVX1-NEXT:    vpblendvb %xmm2, %xmm1, %xmm0, %xmm0
+; AVX1-NEXT:    vpsrlw $2, %xmm0, %xmm1
+; AVX1-NEXT:    vpaddw %xmm2, %xmm2, %xmm2
+; AVX1-NEXT:    vpblendvb %xmm2, %xmm1, %xmm0, %xmm0
+; AVX1-NEXT:    vpsrlw $1, %xmm0, %xmm1
+; AVX1-NEXT:    vpaddw %xmm2, %xmm2, %xmm2
+; AVX1-NEXT:    vpblendvb %xmm2, %xmm1, %xmm0, %xmm0
+; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
+; AVX1-NEXT:    vandps {{.*}}(%rip), %ymm0, %ymm0
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: pr43230:
+; AVX2:       # %bb.0:
+; AVX2-NEXT:    vpxor %xmm2, %xmm2, %xmm2
+; AVX2-NEXT:    vpunpckhwd {{.*#+}} ymm1 = ymm1[4],ymm2[4],ymm1[5],ymm2[5],ymm1[6],ymm2[6],ymm1[7],ymm2[7],ymm1[12],ymm2[12],ymm1[13],ymm2[13],ymm1[14],ymm2[14],ymm1[15],ymm2[15]
+; AVX2-NEXT:    vpunpckhwd {{.*#+}} ymm0 = ymm2[4],ymm0[4],ymm2[5],ymm0[5],ymm2[6],ymm0[6],ymm2[7],ymm0[7],ymm2[12],ymm0[12],ymm2[13],ymm0[13],ymm2[14],ymm0[14],ymm2[15],ymm0[15]
+; AVX2-NEXT:    vpsrlvd %ymm1, %ymm0, %ymm0
+; AVX2-NEXT:    vpshufb {{.*#+}} ymm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,ymm0[26,27],zero,zero
+; AVX2-NEXT:    retq
+;
+; AVX512VL-LABEL: pr43230:
+; AVX512VL:       # %bb.0:
+; AVX512VL-NEXT:    vpsrlvw %ymm1, %ymm0, %ymm0
+; AVX512VL-NEXT:    vpand {{.*}}(%rip), %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
+  %shr = lshr <16 x i16> %a, %b
+  %shuf = shufflevector <16 x i16> zeroinitializer, <16 x i16> %shr, <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 30, i32 15>
+  ret <16 x i16> %shuf
+}