[PATCH] D61068: [X86][SSE] Disable shouldFoldConstantShiftPairToMask for btver1/btver2 targets (PR40758)

[Sanjay Patel via Phabricator via llvm-commits]

spatel added inline comments.


================
Comment at: lib/Target/X86/X86.td:427-430
+def FeatureFastShiftMasks
+    : SubtargetFeature<
+        "fast-shift-masks", "HasFastShiftMasks", "true",
+        "Prefer a left/right logical shift pair over a shift+and pair">;
----------------
Is there a possibility that we would use this for scalar transforms too? If not, better to make this explicitly about vectors:
"FeatureFastVectorShifts" ?


Repository:
  rL LLVM

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D61068/new/

https://reviews.llvm.org/D61068