Daniel Berlin via llvm-dev
2017-Sep-17 22:53 UTC
[llvm-dev] assertion triggered since update to llvm 5
Thanks. This is a dupe of PR 34452. On Sun, Sep 17, 2017 at 2:40 PM, Andrew Kelley <superjoe30 at gmail.com> wrote:> > On Sun, Sep 17, 2017 at 4:21 PM, Davide Italiano <davide at freebsd.org> > wrote: > >> Dump the IR before newgvn and run through opt -newgvn. It should crash. >> I'm not familiar with your frontend but you might want to use -mllvm >> -opt-bisect-limit. >> > > Thanks for the suggestion! I added -mllvm support to my frontend, did your > suggestion, and now I have LLVM IR that causes the crash. > Reported here: https://bugs.llvm.org/show_bug.cgi?id=34651 > > This is with 5.0.0. I don't have a build of master lying around to test > against. > >-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20170917/ea8494d8/attachment.html>
Andrew Kelley via llvm-dev
2017-Sep-17 22:56 UTC
[llvm-dev] assertion triggered since update to llvm 5
Thanks Daniel. I appreciate your patience and guiding me through the troubleshooting process. I look forward to 5.0.1. On Sun, Sep 17, 2017 at 6:53 PM, Daniel Berlin <dberlin at dberlin.org> wrote:> Thanks. This is a dupe of PR 34452. > > > On Sun, Sep 17, 2017 at 2:40 PM, Andrew Kelley <superjoe30 at gmail.com> > wrote: > >> >> On Sun, Sep 17, 2017 at 4:21 PM, Davide Italiano <davide at freebsd.org> >> wrote: >> >>> Dump the IR before newgvn and run through opt -newgvn. It should crash. >>> I'm not familiar with your frontend but you might want to use -mllvm >>> -opt-bisect-limit. >>> >> >> Thanks for the suggestion! I added -mllvm support to my frontend, did >> your suggestion, and now I have LLVM IR that causes the crash. >> Reported here: https://bugs.llvm.org/show_bug.cgi?id=34651 >> >> This is with 5.0.0. I don't have a build of master lying around to test >> against. >> >> >-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20170917/4b589d99/attachment.html>
Daniel Berlin via llvm-dev
2017-Sep-17 22:59 UTC
[llvm-dev] assertion triggered since update to llvm 5
I haven't actually tried to backport any newgvn stuff into the llvm 5 branch (because i don't have any "llvm stable" customer users until you came along). I'll have to see how behind trunk the version is. Being that it's experimental anyway, i may ask the release manager if i can just shove the latest version in there (they are self contained) On Sun, Sep 17, 2017 at 3:56 PM, Andrew Kelley <superjoe30 at gmail.com> wrote:> Thanks Daniel. I appreciate your patience and guiding me through the > troubleshooting process. I look forward to 5.0.1. > > On Sun, Sep 17, 2017 at 6:53 PM, Daniel Berlin <dberlin at dberlin.org> > wrote: > >> Thanks. This is a dupe of PR 34452. >> >> >> On Sun, Sep 17, 2017 at 2:40 PM, Andrew Kelley <superjoe30 at gmail.com> >> wrote: >> >>> >>> On Sun, Sep 17, 2017 at 4:21 PM, Davide Italiano <davide at freebsd.org> >>> wrote: >>> >>>> Dump the IR before newgvn and run through opt -newgvn. It should crash. >>>> I'm not familiar with your frontend but you might want to use -mllvm >>>> -opt-bisect-limit. >>>> >>> >>> Thanks for the suggestion! I added -mllvm support to my frontend, did >>> your suggestion, and now I have LLVM IR that causes the crash. >>> Reported here: https://bugs.llvm.org/show_bug.cgi?id=34651 >>> >>> This is with 5.0.0. I don't have a build of master lying around to test >>> against. >>> >>> >> >-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20170917/ce460162/attachment-0001.html>