[PATCH] D111505: [X86] Prefer VEX encoding in X86 assembler.

[Craig Topper via Phabricator via llvm-commits]

craig.topper added a comment.

Did you consider other patches like this which will sort everything X86InstrAVX512.td after X86InstrSSE.td? The current sorting is done by the instruction name I think so VCVTSD2SIZrm_Int sorts before VCVTSD2SIrm_Int.

  diff --git a/llvm/utils/TableGen/AsmMatcherEmitter.cpp b/llvm/utils/TableGen/AsmMatcherEmitter.cpp
  index 00bdd127e3c2..aaf90aa13272 100644
  --- a/llvm/utils/TableGen/AsmMatcherEmitter.cpp
  +++ b/llvm/utils/TableGen/AsmMatcherEmitter.cpp
  @@ -636,7 +636,7 @@ struct MatchableInfo {
       if (RequiredFeatures.size() != RHS.RequiredFeatures.size())
         return RequiredFeatures.size() > RHS.RequiredFeatures.size();
   
  -    return false;
  +    return TheDef->getID() < RHS.TheDef->getID();
     }
   
     /// couldMatchAmbiguouslyWith - Check whether this matchable could


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D111505/new/

https://reviews.llvm.org/D111505