<html><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><br><div><div>On Aug 21, 2009, at 11:13 AM, <a href="mailto:Alireza.Moshtaghi@microchip.com">Alireza.Moshtaghi@microchip.com</a> wrote:</div><br class="Apple-interchange-newline"><blockquote type="cite"><div><blockquote type="cite"><blockquote type="cite">Add a pass to do call graph analyis to overlay the autos and frame<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">sections of<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">leaf functions. This pass will be extended to color other nodes of<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">the call tree<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">as well in future.<br></blockquote></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite">Sanjiv, this commit, r79562 and r79563 are not the right way to tackle<br></blockquote><blockquote type="cite">this problem, and are coming in the day of the release branch. Please<br></blockquote><blockquote type="cite">revert them and propose patches for these on llvmdev.<br></blockquote><br>Hi Chris<br><br>The reason why we are doing these in backend is that cloning a function<br>is base on which functions call it and of course in frontend we don't<br>have the view of the entire call graph. We'll be happy to revert these,<br>but I would like to know exactly why in your view. <br>With our current design wrto accessing auto variables, and how they work<br>in the cloned function, this is the best that we can do. Although, it is<br>likely that we change the design after 2.6 but we would prefer to get as<br>much of the functionality in 2.6 as possible because we want to base an<br>internal release of PIC16 compiler on llvm 2.6 <br><br></div></blockquote><div><br></div><div>Checking this in the day of the code freeze is totally unacceptable. These sorts of things need to be checked at the very latest a week before so that potential fallout can be discovered (ie. mingw issue).</div><div><br></div><div>-Tanya</div><br><blockquote type="cite"><div>-Ali<br><br>_______________________________________________<br>llvm-commits mailing list<br><a href="mailto:llvm-commits@cs.uiuc.edu">llvm-commits@cs.uiuc.edu</a><br><a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits">http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits</a><br></div></blockquote></div><br></body></html>