[PATCH] D28329: [PowerPC, DAGCombiner] Fold a << (b % (sizeof(a) * 8)) back to a single instruction

[Tim Shen via Phabricator via llvm-commits]

timshen updated this revision to Diff 83203.
timshen marked 2 inline comments as done.
timshen added a comment.

Only optimize legal instructions.

I feel like needing a separate patch to add some test cases for illegal vector instructions (that get lowered to scalar instructions), and those shouldn't be combined. I need to investigate how to make those instructions illegal.


https://reviews.llvm.org/D28329

Files:
  llvm/include/llvm/Target/TargetLowering.h
  llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
  llvm/lib/Target/PowerPC/PPCISelLowering.h
  llvm/test/CodeGen/PowerPC/shift_mask.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D28329.83203.patch
Type: text/x-patch
Size: 8302 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20170105/e3d18906/attachment.bin>