On Thu, Sep 11, 2014 at 21:17:34 +0200, repron wrote: > Dnia 9 września 2014 5:13 Michel Dänzer <mic...@daenzer.net> napisał(a): > > > On 08.09.2014 20:32, repron wrote: > > > > Last several freezes of Xorg, coused that Xorg.o.log file ended > > > > suddenly,> without SIGSEGV. > > > I am still trying to reproduce exactly the same state of "freeze" as> > > > mentioned in the bug. > > > Of course it can be a bug in kernel or somewhere else but if I had> > > > uninstalled packages > > > xserver-xorg-video-radeon, xserver-xorg-video-iti > > > and installed in that place packages: > > > fglrx-modules-dkms and firmware-linux-nonfree > > > the error of "freezing" did not appeared. > > > > > (Note that the firmware-linux-nonfree package is required for the radeon > > driver, not for fglrx, but the information you've provided shows that it > > was installed) > > > > > > It seems like the crashes you're experiencing are more or less random > > and all over the place. I can think of two basic explanations for that: > > Either the GPU is scribbling over random system memory locations, or > > there is a problem with the RAM or some other hardware component in the > > system. > > > > You can test for the latter with the memtest86(+) or memtester package. > > > > > > -- > > Earthling Michel Dänzer | http://www.amd.com > > Libre software enthusiast | Mesa and X developer > > > > Thanks much for suggestions concerning memtest. > In both cases (radeon or fglrx) I had firmware-linux-nonfree > package installed. > The last freeze was coused by "SIGPIPE broken" and I was able > to get a full backtrace with gdb - enclosed here as a txt file: > sigpipe_backtrace.txt. > Could you have a look at it and say somthing about it. ?
SIGPIPE in the X server is not an issue, it's expected whenever a client goes away. You should continue past that signal, or tell gdb not to stop on it. Cheers, Julien
signature.asc
Description: Digital signature
