Re: [lm-sensors] sensors command now hangs [update 2]

Re: [lm-sensors] sensors command now hangs [update 2]

[David Anderson]

Well, even without ipmi sensors hangs up
at (hand typing this) after just 6 hours:

...
w83795adg-i2c-0-27
Adapter: SMBus I801 at 0400



Google search shows someone using redhat has a hanging
sensors command too :

https://bugzilla.redhat.com/show_bug.cgi?id\x1051274

Note there the apparent workaround of nouveau.runrpm=0
on the kernel command line.
I'm using Nvidia driver  319.60 for the GTX 760

DavidA.

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] sensors command now hangs [update 2]

[Guenter Roeck]

On 02/17/2014 01:17 PM, David Anderson wrote:
> Well, even without ipmi sensors hangs up
> at (hand typing this) after just 6 hours:
>
> ...
> w83795adg-i2c-0-27
> Adapter: SMBus I801 at 0400
>
>
>
> Google search shows someone using redhat has a hanging
> sensors command too :
>
> https://bugzilla.redhat.com/show_bug.cgi?id\x1051274
>
> Note there the apparent workaround of nouveau.runrpm=0
> on the kernel command line.
> I'm using Nvidia driver  319.60 for the GTX 760
>

Hi David,

That looks like a different problem, though, unless you
access the nouveau sensor after all. Is it instantiated
in your system ?

To be sure to exclude this, you could try the command line
workaround. Note, though, that it is nouveau.runpm,
not nouveau.runrpm.

Overall I am quite puzzled, as you mentioned earlier that
you also see the hang with the w83627hf driver loaded.
I can not really imagine why (and how) either of those
drivers can end up in a 'hang' condition.

I think I already asked this - in your original e-mail,
you mentioned that the sensors command 'now hangs', and that
only started happening some time ago. What did you change -
if anything - in your system before this started to happen ?

Thanks,
Guenter


_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] sensors command now hangs [update 2]

[David Anderson]

Not using Noveau driver at all here.
         lsmod |grep nov
shows nothing.  I just thought it was interesting that someone
else had a hanging sensors command.

On 02/17/2014 03:17 PM, Guenter Roeck wrote:

> I think I already asked this - in your original e-mail,
> you mentioned that the sensors command 'now hangs', and that
> only started happening some time ago. What did you change -
> if anything - in your system before this started to happen ?

The only change I can think of
was to move from Ubuntu 13.04 to
Ubuntu 13.10.

There are two proprietary nvidia drivers
available from Ubuntu for 13.10, and only  the 319.60 one
works reliably.   319.32 driver gets in trouble
and X stops (kernel keeps running, I can ssh in).
319.60 seems to be reliable.  Not that anyone can
do much with this info. Just saying.
Using nvidia drivers so I can do GPU computation for boinc projects
(Seti@home and Einstein).

DavidA.

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] sensors command now hangs [update 2]

[Guenter Roeck]

On 02/17/2014 04:43 PM, David Anderson wrote:
> Not using Noveau driver at all here.
>          lsmod |grep nov
> shows nothing.  I just thought it was interesting that someone
> else had a hanging sensors command.
>
Hi David,

That grep wont help much, as it is nouveau (with an added u),
not noveau.

> On 02/17/2014 03:17 PM, Guenter Roeck wrote:
>
>> I think I already asked this - in your original e-mail,
>> you mentioned that the sensors command 'now hangs', and that
>> only started happening some time ago. What did you change -
>> if anything - in your system before this started to happen ?
>
> The only change I can think of
> was to move from Ubuntu 13.04 to
> Ubuntu 13.10.
>
Right now I have no idea what may be different. 13.4 is the 3.8 kernel
if I remember correctly, and 13.10 should be 3.11. I did a quick diff
in the relevant drivers but did not see anything that might cause
the problem.

To find out what exactly is hanging, you might try something like

cd /sys/class/hwmon
for d in *
do
(echo $d:
cd $d
if [ ! -f name ]
then
cd device
fi
ls
grep . *
)
done

The last line of output should show the attribute causing the hang,
or rather the one accessed before it.

Guenter


_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors