System locks up, background operations continue running

[Chris]

On Tue, 2015-01-27 at 14:30 +0000, Colin Law wrote:
> On 27 January 2015 at 13:38, Chris <cpollock at embarqmail.com> wrote:
> > Ubuntu 14.04.1 LTS, Gnome 3.12.2, kernel 3.19.0-031900rc5-generic (I'll
> > explain this later) 4Gb ram, Dell Optiplex 780, BIOS version A015.
> >
> > I placed extra information in the above because possibly some questions
> > on them may be asked. Starting in September of last year and
> > periodically continuing through 17 Jan the system would 'lockup' with
> > this in the syslog:
> >
> > kernel: [173977.808008] [drm:i915_hangcheck_elapsed] *ERROR* Hangcheck
> > timer elapsed... render ring idle
> >
> > There are two bug reports for this. One which I initiated at Ubuntu
> > Launchpad - https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1402331
> > and the other at 'freedesktop.org' which had already been initiated by
> > someone else - https://bugs.freedesktop.org/show_bug.cgi?id=75394 If you
> > take a look at the report on Launchpad you'll notice that firstly I was
> > told my BIOS needed updating. That I did with no change. I have since
> > been asked to do a 'kernel bisection' I've never done that before and to
> > be honest the meds I take screw me up most of the day and I haven't been
> > able to follow the directions very well. Be that as it may, I've
> > experienced this issue in kernel versions 3.13.., 3.14.. and now with
> > the 3.19 version however, there's been a change. No longer do I see the
> > 'Hangcheck.....' entry in my syslog. Since 17 Jan I've had seven lockups
> > with no reason that I can see in my syslog. The reason I'm running the
> > 3.19..kernel is that in the freedesktop.org report I was asked to
> > upgrade to the very latest kernel which happened to be the one above to
> > see if it made any difference. As far as background operations such as
> > Fetchmail, Procmail, SpamAssassin, ClamAV and so forth, they continue to
> > run as if nothing has happened. I can still move the mouse cursor around
> > still 'click' on something if it happens while on the web with Firefox
> > however nothing will happen and the clock which I have set to show
> > seconds will freeze.
> >
> > If anyone has any suggestions on how to figure this out they would be
> > greatly appreciated.
> 
> The reason that performing a git bisect is asked for is to find the
> code that introduce the problem.  This relies on knowing that a
> previous version was ok.  It is suggested that 3.13.0-35 was ok.  I
> did not see in the bug report where that was documented.  Is it
> correct?
> 
> The fact that you are getting a different lockup with the latest
> kernel may not be relevant (to this bug) as it may be a completely
> different issue.
> 
> Have you considered trying Ubuntu 14.10 or 15.04?  It won't help find
> the cause of the bug but may help you.
> 
> At the risk of adding further confusion have you seen this?
> http://askubuntu.com/questions/522680/ubuntu-14-04-randomly-freezes-drm
> 
> Colin
Thanks for the reply Colin, I 'attempted' to install 14.10 when it first
came out luckily I do a nightly backup to a 1Tb drive as it completely
hosed my machine for some reason but I won't get into that. I haven't
even looked at 15.04 yet is it in release stage or just as an RC? I see
now why the bisection is necessary, I've got the kernel bisection
information printed off and will get started on it. I've uploaded the
dmesg file from the 2nd lockup last night - http://pastebin.com/bM9iurrX
in my /etc/default/grub file I have this line -
GRUB_CMDLINE_LINUX_DEFAULT="drm.debug=0x66" yes, the link does add more
confusion and I really don't want to get too far into 'trying' out
different things as even though it's just my home machine it's used
pretty much during the day for bill paying and so forth.

Chris

-- 
Chris
KeyID 0xE372A7DA98E6705C
31.11°N 97.89°W (Elev. 1092 ft)
09:38:23 up 14:22, 1 user, load average: 0.13, 0.34, 0.67
Ubuntu 14.04.1 LTS, kernel 3.19.0-031900rc5-generic