[PATCH] D150670: [InstCombine] Disable generation of fshl/fshr for rotates

[Nikita Popov via Phabricator via llvm-commits]

nikic accepted this revision.
nikic added a comment.
This revision is now accepted and ready to land.

LGTM, let's give it a try. The patch description needs an update though.



================
Comment at: llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp:920
+            return I;
+        } else { // fshl is a rotate
+          KnownBits LHSKnown = computeKnownBits(I->getOperand(0), Depth + 1, I);
----------------
Add some more explanation here, something like:
`// Avoid converting rotate into funnel shift. Only simplify if one operand is constant.`


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D150670/new/

https://reviews.llvm.org/D150670