[PATCH] D54267: [X86][SSE] Add lowerVectorShuffleAsByteRotateAndPermute (PR39387)

[Craig Topper via Phabricator via llvm-commits]

craig.topper added inline comments.


================
Comment at: lib/Target/X86/X86ISelLowering.cpp:10170
+  unsigned EltSize = VT.getScalarSizeInBits();
+  if (ImmBlends && EltSize == 8 && !canWidenShuffleElements(BlendMask))
+    return SDValue();
----------------
RKSimon wrote:
> craig.topper wrote:
> > If the elements could be widened wouldn't they have already been widened in lowerVectorShuffle?
> You can have cases where the blend is wider than the permute (PBLENDD + PSHUFB etc.)
Oh right that makes sense. BlendMask is different than Mask


Repository:
  rL LLVM

https://reviews.llvm.org/D54267