2013-03-09 17:39:26

by Sascha Wilde

[permalink] [raw]
Subject: Re: regression in linux 3.7 - fan speed at 100% after suspend/resume at 100%

Hi *,

same problem here on my HP Compaq 6910p.

I just tested with vanilla linux 3.8.2 -- unfortunately the problem
still persists.[1][2] The last (stable) kernel working for me is 3.6.11.

I'll happily provide any additional information which might help getting
this fixed. Just ask... :)

Please CC me in any response, for I'm not subscribed to lkml.

cheers
sascha

[1] FWIW, self build vanilla linux 3.8.2 with very generic config on
Debian Wheezy 64bit.
[2] FWIW², I'm under the stron impression that the whole fan control
changed in the latest vanilla kernels on my 6910p even before doing
a suspend/resume -- it seems to be more "binary" if you will, in
that the fan is not running at all for a while and then at a 100%
to return back to 0% after doing some serious cooling. With 3.6.11
and older it is more "analogue", with slow fan when only a bit
cooling is needed. No idea if this is related to the issue at
hand.
--
Sascha Wilde

If you think technology can solve your problems you don't understand
technology and you don't understand your problems. (Bruce Schneier)


2013-03-10 15:47:09

by Sascha Wilde

[permalink] [raw]
Subject: Re: regression in linux 3.7 - fan speed at 100% after suspend/resume at 100%

Sascha Wilde <[email protected]> wrote:
[...]
> [2] FWIW², I'm under the stron impression that the whole fan control
> changed in the latest vanilla kernels on my 6910p even before doing
> a suspend/resume -- it seems to be more "binary" if you will, in
> that the fan is not running at all for a while and then at a 100%
> to return back to 0% after doing some serious cooling. With 3.6.11
> and older it is more "analogue", with slow fan when only a bit
> cooling is needed. No idea if this is related to the issue at
> hand.

For that part of the problem I just discovered something that might be a
hint on the cause, there is no governor set for the thermal zones:

/sys/devices/virtual/thermal/thermal_zone0/policy : (null)
/sys/devices/virtual/thermal/thermal_zone1/policy : (null)
/sys/devices/virtual/thermal/thermal_zone2/policy : (null)
/sys/devices/virtual/thermal/thermal_zone3/policy : (null)
/sys/devices/virtual/thermal/thermal_zone4/policy : (null)

to my understanding this policies should default default to the
CONFIG_THERMAL_DEFAULT_GOV_*
configured in the kernel. But it is always (null) for me. And
something like
# echo fair_share >/sys/devices/virtual/thermal/thermal_zone0/policy
does not work either.

Or is this expected behavior?

cheers
sascha
--
Sascha Wilde : "There are 10 types of people in the world.
: Those who understand binary and those who don't."