Hi Michel, ----- Original Message ----- > On 20/02/17 06:44 PM, Olivier Fourdan wrote: > > ----- Original Message ----- > >> I'd like to do another 1.19 soonish, say middle of next week. > > > > That would be great if we could get to the bottom of the random > > crashes that affect 1.19.x first. > > > > Good news is we now have a core file from downstream bug > > https://bugzilla.redhat.com/show_bug.cgi?id=1424644 > > > > I'm struggling to find the problem so if anyone else has any idea, > > I'm all hear. > > I also got stuck analyzing the valgrind logs in > https://bugs.freedesktop.org/show_bug.cgi?id=99164 .
I have a similar bug here: https://bugs.freedesktop.org/show_bug.cgi?id=99887 > One thing that confuses me in the valgrind output below is: > CloseDownConnection (connection.c:919) seems to refer to the > FreeOsBuffers call (specifically, the free(oco) in there). However, > FreeOsBuffers is defined in io.c, so why does the valgrind output look > as if FreeOsBuffers was inlined into CloseDownConnection? Is there some > kind of LTO going on, or are there any patches affecting this applied on > top of upstream? No, we don't have patches on top of the xserver that would explain that... Probably a valgrind or a gcc issue. But I think I might be onto something, I strongly suspect this is a threading issue. I posted some more logs in bug 99887, and will send a possible patch shortly. Cheers, Olivier _______________________________________________ [email protected]: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
