Jürgen Beisert <jbe@...> writes:

> 
> Hi Dave,
> 
> On Wednesday 30 October 2013 06:01:40 Dave wrote:
> > Jürgen Beisert <jbe <at> ...> writes:
> > > On Sunday 20 October 2013 20:56:18 Dave wrote:
> > > > Thought I would try 2013.09.0 with kernel 3.11 and have been trying 
to
> > > > sort a problem as follows:
> > > >
> > > > - at random times the cursor will stop flashing and even though px 
aux
> > > > shows that my application is running it is not.  I can still 
navigate
> > > > directories and look at files, so something is still working.
> > > >
> > > > Nothing appears in the log files in /var/log (daemon.log, 
thttpd.log,
> > > > kern.log and user.log)
> > > >
> > > > Any hints on where to look next would be appreciated.
> > >
> > > You are the second person who reports about a stuck Linux-3.11 kernel. 
It
> > > seems the timer implementation is broken since 3.11 (the timer
> > > implementation was changed on the way to 3.11). The S3C2440 CPU has 
now HRT
> > > (high resolution timer) support.
> > > I have no idea how to trigger the fault. At least on my system it 
didn't
> > > hit me yet.
> > >
> > > Just an idea: when it hits you the next time, please "cat" the content
> > > of "/proc/interrupts" a few times and see if the timer interrupt still
> > > increments or not.
> >
> > Thanks for the hint, I wasn't getting anywhere hoping something would 
turn
> > up in a log file!
> >
> > Just tried your suggestion and while I was repeatedly invoking the 
command
> > the system locked and you are right the samsung_timer_irq is frozen.
> 
> Thanks for the report. I will report this regression on the ALKML mailing 
list.
> 
> Regards,
> Juergen
> 

Hello Jürgen,

I guess I am the third person to report this.

I am seeing the same problem using ptxdist 2013.09.0, with kernel 3.11 and 
the standard configuration (I was just testing and did not do anything else 
yet). After a while the system time stops updating and I can see that timer 
interrupts are not being generated anymore (I can verify by checking 
/proc/interrupts)

Did you manage to report this to the ALKML? Do you have any additional 
information on what may be wrong?

Stella




Reply via email to