[ Date Index ] [ Thread Index ] [ <= Previous by date / thread ] [ Next by date / thread => ]
I've had a quick look and i'm confused already. start_secondary() is the entry point to start up the non boot CPU on SMP systems, and i can't see where cpu_startup_entry() manages to call schedule_preempt_disabled() its probably inline code somewhere as that is part of a mutex locking operation. I can't see directly a driver involved. So this looks like a kernel issue for SMP systems? Which begs the question why are you getting this multiple times, this is a CPU startup function? Any clever power management that can sleep CPUs/cores and this is wake up related? You could try installing the full kernel symbols package, it may give a slightly better stack trace On 17 June 2016 at 14:45, Adrian Midgley <amidgley@xxxxxxxxx> wrote: > Kernel, or a driver for something, if so how should I approach identifying > the culprit, please? > > I've had a few sudden halts, associated with moving the edge of a frame of > eg Iceweasel --> Firefox-esr or Gimp in the Gnome desktop with the (USB) > mouse. > I think they are less with the current version of NVidia's proprietary > driver controlling the 4-headed video card and four monitors you may recall > me mentioning earlier. > > (BTW, if anyone wants a couple of 2-headed NVidia cards which are very nice > but don't go up to the largest resolutions with DVI, a very small token > would secure them.) > > On Fri, 17 Jun 2016 at 14:41 Robin Cornelius <robin.cornelius@xxxxxxxxx> > wrote: >> >> Driver bug, Its a warning that the kernel driver is attempting to >> sleep in the idle thread (which should be non blocking code only) and >> not try to sleep etc. >> >> It is only a warning so is non fatal but i guessing its a ripe way to >> a deadlock situation with the idle thread stalled waiting for >> something else to occur. >> >> >> >> On 17 June 2016 at 13:30, Simon Waters <simon@xxxxxxxxxxxxxx> wrote: >> > If it works worry a little, if it breaks worry a lot. >> > >> > I'd probably try a newer kernel package, or an older one, first. See if >> > it persists, ilunless you know what triggers it. >> > >> > -- >> > The Mailing List for the Devon & Cornwall LUG >> > https://mailman.dclug.org.uk/listinfo/list >> > FAQ: http://www.dcglug.org.uk/listfaq >> >> -- >> The Mailing List for the Devon & Cornwall LUG >> https://mailman.dclug.org.uk/listinfo/list >> FAQ: http://www.dcglug.org.uk/listfaq > > > -- > The Mailing List for the Devon & Cornwall LUG > https://mailman.dclug.org.uk/listinfo/list > FAQ: http://www.dcglug.org.uk/listfaq > -- The Mailing List for the Devon & Cornwall LUG https://mailman.dclug.org.uk/listinfo/list FAQ: http://www.dcglug.org.uk/listfaq