On Wed, Nov 12, 2014 at 1:55 PM, Steven Noonan <ste...@uplinklabs.net> wrote: > On Wed, Nov 12, 2014 at 11:22 AM, "Jóhann B. Guðmundsson" > <johan...@gmail.com> wrote: >> >> On 11/12/2014 07:00 PM, Steven Noonan wrote: >>> >>> >What's the best way to approach debugging this issue? Should I do a >>> >debug build to get a better core dump and wait for this to happen >>> >again? >>> > >> >> >> Looks like a regression or dupe of [1] not sure why this bug has not been >> reopen since there is a recent comment there. >> >> Add your info and coredump there > > Thanks, I've done so. > > I'd like to understand in general how one debugs issues like this with > systemd. Given that it's /sbin/init, it's not easy to just attach a > debugger to it and poke around. Do people operate primarily on core > dumps of -O0 -g builds or something?
Well I've worked out some of the logic that leads to this assertion failure: https://bugs.freedesktop.org/show_bug.cgi?id=67427#c21 I also attached a patch. It does stop the reproducible systemd assertion failure as intended, but I'm not sure I'm doing this change in the right place. The comments around JOB_NOP's declaration suggest that since it's not a mergeable job, it's special-cased in a few places. Perhaps this is just one of those places where special-casing was missed? _______________________________________________ systemd-devel mailing list systemd-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/systemd-devel