Hi All,<div><br></div><div>"Thanks for the suggestion Matthieu. I spoke to Doug and he recommended using attributes rather than a FunctionDecl bit to represent the fp_contract state. This patch is a first attempt at that. Please review and let me know what I've missed, or what could be done better.</div>
<div><br></div><div>Cheers,</div><div>Lang.<br><br><div class="gmail_quote">
On Wed, May 30, 2012 at 11:14 AM, Matthieu Monrocq <span dir="ltr"><<a href="mailto:matthieu.monrocq@gmail.com" target="_blank">matthieu.monrocq@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br><br><div class="gmail_quote"><div><div>On Wed, May 30, 2012 at 9:20 AM, Lang Hames <span dir="ltr"><<a href="mailto:lhames@gmail.com" target="_blank">lhames@gmail.com</a>></span> wrote:<br></div></div>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><div>
<div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="gmail_quote"><div>Agreed - Legal was a poor choice of terms. I don't know if we want to check TLI.isOperationLegal(...) though - an llvm.fma.* call (which is what we lower to) will always be legal...</div>
</div></blockquote><div><br></div><div>Owen pointed out that I'm finding new and interesting ways to abuse the term 'legal' here. To clarify: If isFMACheap returns true we should legalize the llvm.fmuladd.* to an llvm.fma.* and leave it to the existing legalization code to turn the llvm.fma.* into something sane for the target.</div>
<span><font color="#888888">
<div><br></div><div>- Lang.</div></font></span></div>
<br></div></div><div>_______________________________________________<br>
cfe-commits mailing list<br>
<a href="mailto:cfe-commits@cs.uiuc.edu" target="_blank">cfe-commits@cs.uiuc.edu</a><br>
<a href="http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits" target="_blank">http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits</a><br>
<br></div></blockquote></div><br>Hello Lang,<br><br>I only have one nits: I think you'll need Doug's approval for adding a bit flag into FunctionDecl (especially since this is such a specialized construct).<br><br>
I added him in case he missed the discussion.<span><font color="#888888"><br>
<br>-- Matthieu<br>
</font></span></blockquote></div><br></div>