<html><head></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><br><div><div>On Jun 30, 2011, at 5:16 PM, Rafael Espindola wrote:</div><br class="Apple-interchange-newline"><blockquote type="cite"><span class="Apple-style-span" style="border-collapse: separate; font-family: Optima; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: 2; text-align: -webkit-auto; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; -webkit-border-horizontal-spacing: 0px; -webkit-border-vertical-spacing: 0px; -webkit-text-decorations-in-effect: none; -webkit-text-size-adjust: auto; -webkit-text-stroke-width: 0px; font-size: medium; "><span class="Apple-style-span" style="font-family: monospace; ">+ // FIXME: If "B = X" kills X, we have to move the kill back to its<br>+ // previous use. For now we just avoid the optimization in that case.<br>+ if (MI->getOperand(1).isKill())<br>+ return false;<br></span></span></blockquote></div><br><div>It's not so much the kill flags, they are actually recomputed after register allocation.</div><div><br></div><div>It is important that the live ranges are updated, though. You need to shrink X to its uses when you remove a use.</div><div><br></div><div>There doesn't have to be a kill flag: If you are removing the last use in a loop where X is not live out, it doesn't have to be live in the loop any more.</div><div><br></div><div>/jakob</div><div><br></div></body></html>