<html><head><meta http-equiv="Content-Type" content="text/html charset=us-ascii"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><span style="font-family: monospace; ">Hi,</span><div style="font-family: monospace; "><br></div><div style="font-family: monospace; ">For the 3.3 release branch, I would like to propose merging r181620 and r183907:</div><div style="font-family: monospace; "><br></div><div style="font-family: monospace; ">r181620 from llvm trunk (by Chad Rosier):<br><br> [ms-inline asm] Fix a crasher when we fail on a direct match.<br><br> The issue was that the MatchingInlineAsm and VariantID args to the<br> MatchInstructionImpl function weren't being set properly. Specifically, when<br> parsing intel syntax, the parser thought it was parsing inline assembly in the<br> at&t dialect; that will never be the case.<br><br> The crash was caused when the emitter tried to emit the instruction, but the<br> operands weren't set. When parsing inline assembly we only set the opcode, not<br> the operands, which is used to lookup the instruction descriptor.<br> <a href="rdar://13854391">rdar://13854391</a> and PR15945<br><br> Also, this commit reverts r176036. Now that we're correctly parsing the intel<br> syntax the pushad/popad don't match properly. I've reimplemented that fix using<br> a MnemonicAlias.<br><br>r183907 from llvm trunk (by Benjamin Kramer):<br><br> X86: Make the cmov aliases work with intel syntax too.<br><br>Together, these commits make a number of Intel-style inline assembly mnemonics aliases (cmov variants, occurring in several FreeBSD ports) work properly, which could cause assertions (and crashes) otherwise.</div><div style="font-family: monospace; "><br></div><div style="font-family: monospace; ">These commits should apply to the release_33 branch without modification.</div><div style="font-family: monospace; "><br></div><div style="font-family: monospace; ">-Dimitry</div><div><br></div></body></html>