<html><body><p><font size="2">Thanks Melanie.</font><br><br><font size="2">I have pushed one fix to test-suite to explicitly add -ffp-contract=off to unblock our internal buildbot for now.</font><br><font size="2"><a href="https://github.com/llvm/llvm-test-suite/commit/c04a7178a3a50fe919964df59b041c5671db50f7">https://github.com/llvm/llvm-test-suite/commit/c04a7178a3a50fe919964df59b041c5671db50f7</a></font><br><br><font size="2">Our buildbot are OK now. </font><br><br><font size="2">I think you can proceed as long as the change is intended and reasonable. Thanks.</font><br><br><font size="2"><br>Best,<br><br>Jinsong Ji (ΌΝ½πΛΙ), PhD.<br><br>XL/LLVM on Power Compiler Development <br>E-mail: jji@us.ibm.com</font><br><br><img width="16" height="16" src="cid:1__=8FBB0F9DDFF9268F8f9e8a93df938690918c8FB@" border="0" alt="Inactive hide details for "Blower, Melanie I" ---02/14/2020 10:34:32 AM---I reverted MaskRay's "reland" since the original patc"><font size="2" color="#424282">"Blower, Melanie I" ---02/14/2020 10:34:32 AM---I reverted MaskRay's "reland" since the original patch is causing trouble on PowerPC, check-all is p</font><br><br><font size="2" color="#5F5F5F">From: </font><font size="2">"Blower, Melanie I" <melanie.blower@intel.com></font><br><font size="2" color="#5F5F5F">To: </font><font size="2">"reviews+D74436+public+e2b40a7853ffb354@reviews.llvm.org" <reviews+D74436+public+e2b40a7853ffb354@reviews.llvm.org>, "lebedev.ri@gmail.com" <lebedev.ri@gmail.com>, "rjmccall@gmail.com" <rjmccall@gmail.com>, "sepavloff@gmail.com" <sepavloff@gmail.com></font><br><font size="2" color="#5F5F5F">Cc: </font><font size="2">"maskray@google.com" <maskray@google.com>, "jji@us.ibm.com" <jji@us.ibm.com>, "david.bolvansky@gmail.com" <david.bolvansky@gmail.com>, "martin@martin.st" <martin@martin.st>, "Wang, Pengfei" <pengfei.wang@intel.com>, "wuzish@cn.ibm.com" <wuzish@cn.ibm.com>, "nemanja.i.ibm@gmail.com" <nemanja.i.ibm@gmail.com>, "kit.barton@gmail.com" <kit.barton@gmail.com>, "cfe-commits@lists.llvm.org" <cfe-commits@lists.llvm.org>, "mlekena@skidmore.edu" <mlekena@skidmore.edu>, "blitzrakete@gmail.com" <blitzrakete@gmail.com>, "shenhan@google.com" <shenhan@google.com>, "peter.waller@arm.com" <peter.waller@arm.com></font><br><font size="2" color="#5F5F5F">Date: </font><font size="2">02/14/2020 10:34 AM</font><br><font size="2" color="#5F5F5F">Subject: </font><font size="2">[EXTERNAL] RE: [PATCH] D74436: Change clang option -ffp-model=precise to select ffp-contract=on</font><br><hr width="100%" size="2" align="left" noshade style="color:#8091A5; "><br><br><br><tt><font size="2">I reverted MaskRay's "reland" since the original patch is causing trouble on PowerPC, check-all is passing on my box. Sorry for the trouble. <br><br>> -----Original Message-----<br>> From: Andy Kaylor via Phabricator <reviews@reviews.llvm.org><br>> Sent: Thursday, February 13, 2020 9:20 PM<br>> To: Blower, Melanie I <melanie.blower@intel.com>; lebedev.ri@gmail.com;<br>> rjmccall@gmail.com; sepavloff@gmail.com<br>> Cc: maskray@google.com; jji@us.ibm.com; david.bolvansky@gmail.com;<br>> martin@martin.st; Wang, Pengfei <pengfei.wang@intel.com>;<br>> wuzish@cn.ibm.com; nemanja.i.ibm@gmail.com; kit.barton@gmail.com; cfe-<br>> commits@lists.llvm.org; mlekena@skidmore.edu; blitzrakete@gmail.com;<br>> shenhan@google.com; peter.waller@arm.com<br>> Subject: [PATCH] D74436: Change clang option -ffp-model=precise to select ffp-<br>> contract=on<br>> <br>> andrew.w.kaylor added a subscriber: MaskRay.<br>> andrew.w.kaylor added a comment.<br>> <br>> In D74436#1875386 <</font></tt><tt><font size="2"><a href="https://reviews.llvm.org/D74436#1875386">https://reviews.llvm.org/D74436#1875386</a></font></tt><tt><font size="2"> >, @thakis<br>> wrote:<br>> <br>> > The revert of this breaks tests everywhere, as far as I can tell.<br>> <br>> <br>> It looks like something strange happened with the revert:<br>> <br>> > clang-11: warning: overriding '-ffp-model=strict' option with '-ffp-<br>> model=strict' [-Woverriding-t-option]<br>> <br>> I believe the problem is that the original change that was being reverted<br>> contained this:<br>> <br>> clang/lib/Driver/ToolChains/Clang.cpp<br>> @@ -2768,7 +2766,7 @@ static void RenderFloatingPointOptions(const<br>> ToolChain &TC, const Driver &D,<br>> !AssociativeMath && !ReciprocalMath &&<br>> SignedZeros && TrappingMath && RoundingFPMath &&<br>> DenormalFPMath != llvm::DenormalMode::getIEEE() &&<br>> + FPContract.empty())<br>> - (FPContract.equals("off") || FPContract.empty()))<br>> <br>> But sometime in the land-revert-land-revert cycle the line above that changed,<br>> causing the merge to miss this change in the most recent revert. I see that<br>> @MaskRay has since re-landed this change set, but it's going to cause problems<br>> for PowerPC. If someone needs to revert this yet again, I think it can be safely<br>> done by recovering the change above.<br>> <br>> Apologies for the mess!<br>> <br>> <br>> Repository:<br>> rG LLVM Github Monorepo<br>> <br>> CHANGES SINCE LAST ACTION<br>> </font></tt><tt><font size="2"><a href="https://reviews.llvm.org/D74436/new/">https://reviews.llvm.org/D74436/new/</a></font></tt><tt><font size="2"> <br>> <br>> </font></tt><tt><font size="2"><a href="https://reviews.llvm.org/D74436">https://reviews.llvm.org/D74436</a></font></tt><tt><font size="2"> <br>> <br>> <br><br></font></tt><br><br><BR>
</body></html>