[PATCH] D97982: [MC] Introduce NeverAlign fragment type

[Fangrui Song via Phabricator via llvm-commits]

MaskRay added a comment.

This is performance related. I hope folks like @RKSimon @lebedev.ri can comment.



================
Comment at: llvm/include/llvm/MC/MCFragment.h:37
     FT_Align,
+    FT_NeverAlign,
     FT_Data,
----------------
Perhaps the name should include `X86`? Inserting a one-byte nop isn't something any RISC arch can do.


================
Comment at: llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp:4733
+/// parseDirectiveAvoidEndAlign
+///  ::= .avoid_end_align alignment
+bool X86AsmParser::parseDirectiveAvoidEndAlign(SMLoc L) {
----------------
Are you using the assembly syntax `.avoid_end_align` in tools? Or is it for testing purpose only?


================
Comment at: llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp:4737
+  SMLoc AlignmentLoc;
+  const MCSubtargetInfo STI = getSTI();
+  AlignmentLoc = getTok().getLoc();
----------------
Delete


================
Comment at: llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp:4743
+
+  if (getParser().parseToken(AsmToken::EndOfStatement,
+                             "unexpected token in directive"))
----------------
`if (parseEOL())`


================
Comment at: llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp:4748
+  if (Alignment <= 0) {
+    Error(AlignmentLoc, "directive with non-positive alignment");
+    return false;
----------------
expected a positive alignment


================
Comment at: llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp:4753
+  getParser().getStreamer().emitNeverAlignCodeAtEnd(Alignment);
+
+  return false;
----------------
delete blank line


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D97982/new/

https://reviews.llvm.org/D97982