[PATCH] D89105: [X86] Support Intel avxvnni

[Craig Topper via Phabricator via llvm-commits]

craig.topper accepted this revision.
craig.topper added a comment.

LGTM to me with a couple minors that can be fixed on commit.



================
Comment at: clang/test/Preprocessor/predefined-arch-macros.c:1828
+// CHECK_ADL_M64: #define __AVXVNNI__ 1
 // CHECK_ADL_M64-NOT: AVX512
 // CHECK_ADL_M64: #define __AVX__ 1
----------------
doesn't this NOT need to be above AVXNNI to be effective since that's where AVX512 would go alphabetically?


================
Comment at: llvm/lib/Target/X86/MCTargetDesc/X86InstPrinterCommon.cpp:351
   // These all require a pseudo prefix
-  if (Flags & X86::IP_USE_VEX)
+  if (Flags & X86::IP_USE_VEX || TSFlags & X86II::ExplicitVEXPrefix)
     O << "\t{vex}";
----------------
Put parentheses around the & parts. Similar to lock and notrack above. I don't want to think about operator precedence.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D89105/new/

https://reviews.llvm.org/D89105