[X86] Don't call validateInstruction from MatchAndEmitInstruction when MatchingInlineAsm is set. The MCInst won't be populated.
Without this we can't parse gather instructions in ms inline asm blocks. The validateInstruction function was introduced in r316700 to check gather constraints.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@317713 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/lib/Target/X86/AsmParser/X86AsmParser.cpp b/lib/Target/X86/AsmParser/X86AsmParser.cpp
index b8ea2f0..2a784c9 100644
--- a/lib/Target/X86/AsmParser/X86AsmParser.cpp
+++ b/lib/Target/X86/AsmParser/X86AsmParser.cpp
@@ -2791,7 +2791,7 @@
isParsingIntelSyntax())) {
default: llvm_unreachable("Unexpected match result!");
case Match_Success:
- if (validateInstruction(Inst, Operands))
+ if (!MatchingInlineAsm && validateInstruction(Inst, Operands))
return true;
// Some instructions need post-processing to, for example, tweak which
// encoding is selected. Loop on it while changes happen so the
@@ -3082,7 +3082,7 @@
// instruction will already have been filled in correctly, since the failing
// matches won't have modified it).
if (NumSuccessfulMatches == 1) {
- if (validateInstruction(Inst, Operands))
+ if (!MatchingInlineAsm && validateInstruction(Inst, Operands))
return true;
// Some instructions need post-processing to, for example, tweak which
// encoding is selected. Loop on it while changes happen so the individual
