2.6.33-rc2: regression: gkrellm no longer shows all the temperatures on thinkpad x60

2.6.33-rc2: regression: gkrellm no longer shows all the temperatures on thinkpad x60

[Pavel Machek]

...I'll have to find out where gkrell got that info. It worked in
2.6.32.

									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [ibm-acpi-devel] 2.6.33-rc2: regression: gkrellm no longer shows all the temperatures on thinkpad x60

[Didier Spaier]

Pavel Machek wrote:
> ...I'll have to find out where gkrell got that info. It worked in
> 2.6.32.
> 
> 									Pavel
Only to make sure... What is the output of cat /proc/acpi/ibm/thermal?

I ask that question because I noticed battery temperature 0° displayed  by conky after updating to 2.6.32.2.
Looking at my .conkyrc I saw I tried to display ${smapi_bat_temp 0} but I had forgotten to re-install the tp_smapi module.

May be gkrellm uses tp_smapi too??

PS what is worst, without tp_smapi hdaps do not work well either so I lost hard-disk parking on my Thinkpad T61 too... Hopefully meanwhile the laptop stayed nicely on my desk ;)

Re: [ibm-acpi-devel] 2.6.33-rc2: regression: gkrellm no longer shows all the temperatures on thinkpad x60

[Pavel Machek]

On Sun 2009-12-27 23:24:15, Didier Spaier wrote:
> Pavel Machek wrote:
> >...I'll have to find out where gkrell got that info. It worked in
> >2.6.32.
> >
> >									Pavel
> Only to make sure... What is the output of cat /proc/acpi/ibm/thermal?
> 
> I ask that question because I noticed battery temperature 0° displayed  by conky after updating to 2.6.32.2.
> Looking at my .conkyrc I saw I tried to display ${smapi_bat_temp 0}
but I had forgotten to re-install the tp_smapi module.

Uhuh, there's no /proc/acpi/ibm:

ls /proc/acpi/ibm         
ls: cannot access /proc/acpi/ibm: No such file or directory

. I guess that's the problem :-(. Config is:

CONFIG_THINKPAD_ACPI=y
# CONFIG_THINKPAD_ACPI_DEBUGFACILITIES is not set
# CONFIG_THINKPAD_ACPI_DEBUG is not set
CONFIG_THINKPAD_ACPI_UNSAFE_LEDS=y
CONFIG_THINKPAD_ACPI_VIDEO=y
# CONFIG_THINKPAD_ACPI_HOTKEY_POLL is not set

...so I do not see "enable legacy interface" option.
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [ibm-acpi-devel] 2.6.33-rc2: regression: gkrellm no longer shows all the temperatures on thinkpad x60

[Didier Spaier]

Pavel Machek wrote:
> On Sun 2009-12-27 23:24:15, Didier Spaier wrote:
>> Pavel Machek wrote:
>>> ...I'll have to find out where gkrell got that info. It worked in
>>> 2.6.32.
>>>
>>> 									Pavel
>> Only to make sure... What is the output of cat /proc/acpi/ibm/thermal?
>>
>> I ask that question because I noticed battery temperature 0° displayed  by conky after updating to 2.6.32.2.
>> Looking at my .conkyrc I saw I tried to display ${smapi_bat_temp 0}
> but I had forgotten to re-install the tp_smapi module.
> 
> Uhuh, there's no /proc/acpi/ibm:
> 
> ls /proc/acpi/ibm         
> ls: cannot access /proc/acpi/ibm: No such file or directory
> 
> . I guess that's the problem :-(. Config is:
> 
> CONFIG_THINKPAD_ACPI=y
> # CONFIG_THINKPAD_ACPI_DEBUGFACILITIES is not set
> # CONFIG_THINKPAD_ACPI_DEBUG is not set
> CONFIG_THINKPAD_ACPI_UNSAFE_LEDS=y
> CONFIG_THINKPAD_ACPI_VIDEO=y
> # CONFIG_THINKPAD_ACPI_HOTKEY_POLL is not set
> 
> ...so I do not see "enable legacy interface" option.
> 									Pavel
bash-3.1$ grep THINKPAD_ACPI /boot/config-2.6.32.2strip
CONFIG_THINKPAD_ACPI=m
CONFIG_THINKPAD_ACPI_DEBUGFACILITIES=y
CONFIG_THINKPAD_ACPI_DEBUG=y
CONFIG_THINKPAD_ACPI_UNSAFE_LEDS=y
CONFIG_THINKPAD_ACPI_VIDEO=y
CONFIG_THINKPAD_ACPI_HOTKEY_POLL=y
bash-3.1$

I suggest you modularize THINKPAD_ACPI.

Re: [ibm-acpi-devel] 2.6.33-rc2: regression: gkrellm no longer shows all the temperatures on thinkpad x60

[Pavel Machek]

On Sun 2009-12-27 23:32:59, Didier Spaier wrote:
> Pavel Machek wrote:
> >On Sun 2009-12-27 23:24:15, Didier Spaier wrote:
> >>Pavel Machek wrote:
> >>>...I'll have to find out where gkrell got that info. It worked in
> >>>2.6.32.
> >>>
> >>>									Pavel
> >>Only to make sure... What is the output of cat /proc/acpi/ibm/thermal?
> >>
> >>I ask that question because I noticed battery temperature 0° displayed  by conky after updating to 2.6.32.2.
> >>Looking at my .conkyrc I saw I tried to display ${smapi_bat_temp 0}
> >but I had forgotten to re-install the tp_smapi module.
> >
> >Uhuh, there's no /proc/acpi/ibm:
> >
> >ls /proc/acpi/ibm         ls: cannot access /proc/acpi/ibm: No
> >such file or directory
> >
> >. I guess that's the problem :-(. Config is:
> >
> >CONFIG_THINKPAD_ACPI=y
> ># CONFIG_THINKPAD_ACPI_DEBUGFACILITIES is not set
> ># CONFIG_THINKPAD_ACPI_DEBUG is not set
> >CONFIG_THINKPAD_ACPI_UNSAFE_LEDS=y
> >CONFIG_THINKPAD_ACPI_VIDEO=y
> ># CONFIG_THINKPAD_ACPI_HOTKEY_POLL is not set
> >
> >...so I do not see "enable legacy interface" option.

> bash-3.1$ grep THINKPAD_ACPI /boot/config-2.6.32.2strip
> CONFIG_THINKPAD_ACPI=m
> CONFIG_THINKPAD_ACPI_DEBUGFACILITIES=y
> CONFIG_THINKPAD_ACPI_DEBUG=y
> CONFIG_THINKPAD_ACPI_UNSAFE_LEDS=y
> CONFIG_THINKPAD_ACPI_VIDEO=y
> CONFIG_THINKPAD_ACPI_HOTKEY_POLL=y
> bash-3.1$
> 
> I suggest you modularize THINKPAD_ACPI.

How would modular version help?

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [ibm-acpi-devel] 2.6.33-rc2: regression: gkrellm no longer shows all the temperatures on thinkpad x60

[Didier Spaier]

Pavel Machek wrote:
> On Sun 2009-12-27 23:32:59, Didier Spaier wrote:
>> Pavel Machek wrote:
>>> On Sun 2009-12-27 23:24:15, Didier Spaier wrote:
>>>> Pavel Machek wrote:
>>>>> ...I'll have to find out where gkrell got that info. It worked in
>>>>> 2.6.32.
>>>>>
>>>>> 									Pavel
>>>> Only to make sure... What is the output of cat /proc/acpi/ibm/thermal?
>>>>
>>>> I ask that question because I noticed battery temperature 0° displayed  by conky after updating to 2.6.32.2.
>>>> Looking at my .conkyrc I saw I tried to display ${smapi_bat_temp 0}
>>> but I had forgotten to re-install the tp_smapi module.
>>>
>>> Uhuh, there's no /proc/acpi/ibm:
>>>
>>> ls /proc/acpi/ibm         ls: cannot access /proc/acpi/ibm: No
>>> such file or directory
>>>
>>> . I guess that's the problem :-(. Config is:
>>>
>>> CONFIG_THINKPAD_ACPI=y
>>> # CONFIG_THINKPAD_ACPI_DEBUGFACILITIES is not set
>>> # CONFIG_THINKPAD_ACPI_DEBUG is not set
>>> CONFIG_THINKPAD_ACPI_UNSAFE_LEDS=y
>>> CONFIG_THINKPAD_ACPI_VIDEO=y
>>> # CONFIG_THINKPAD_ACPI_HOTKEY_POLL is not set
>>>
>>> ...so I do not see "enable legacy interface" option.
> 
>> bash-3.1$ grep THINKPAD_ACPI /boot/config-2.6.32.2strip
>> CONFIG_THINKPAD_ACPI=m
>> CONFIG_THINKPAD_ACPI_DEBUGFACILITIES=y
>> CONFIG_THINKPAD_ACPI_DEBUG=y
>> CONFIG_THINKPAD_ACPI_UNSAFE_LEDS=y
>> CONFIG_THINKPAD_ACPI_VIDEO=y
>> CONFIG_THINKPAD_ACPI_HOTKEY_POLL=y
>> bash-3.1$
>>
>> I suggest you modularize THINKPAD_ACPI.
> 
> How would modular version help?
> 
I can't tell you how exactly.

But I can tell that is *does* work as a module, at least with 2.6.32.2.

Plus, reading Documentation/laptops/thinkpad-acpi make obvious IMHO,
though not explicitly stated, that it is supposed to be built as a module.

For instance 'grep module thinkpad-acpi|wc -l'  returns 17 here ;)

And how the 'modprobe' command (founded twice in the same file) is supposed
to work with a driver built in the kernel image?

But of course deciding to try that or not is up to you :-)

Best regards,

Didier

Re: [ibm-acpi-devel] 2.6.33-rc2: regression: gkrellm no longer shows all the temperatures on thinkpad x60

[Henrique de Moraes Holschuh]

Please apply the latest stack of patches (sent them to acpi-test
yesterday).

It is failing to register the ALSA mixer for some reason, and due
to a bug, it is not loading the module at all.  I will look at the reason it
is failing to register the ALSA mixer soon.  Meanwhile, the patches I sent
to Len make sure the module can still load sucessfully.

-- 
  "One disk to rule them all, One disk to find them. One disk to bring
  them all and in the darkness grind them. In the Land of Redmond
  where the shadows lie." -- The Silicon Valley Tarot
  Henrique Holschuh

Re: [ibm-acpi-devel] 2.6.33-rc2: regression: gkrellm no longer shows all the temperatures on thinkpad x60

[Pavel Machek]

Hi!

> Please apply the latest stack of patches (sent them to acpi-test
> yesterday).
> 
> It is failing to register the ALSA mixer for some reason, and due
> to a bug, it is not loading the module at all.  I will look at the reason it
> is failing to register the ALSA mixer soon.  Meanwhile, the patches I sent
> to Len make sure the module can still load sucessfully.

Well, I'm not sure its completely fixed. I got this in my syslog:

thinkpad_acpi: THERMAL EMERGENCY: a sensor reports something is
extremely hot!
thinkpad_acpi: temperatures (Celsius): 95 47 N/A 89 47 N/A 41 N/A 51
61 N/A N/A N/A N/A N/A N/A

...meaning thinkpad_acpi is active and working, but still gkrellm does
not report temperatures as it used to.

root@amd:~# cat /proc/acpi/ibm/thermal 
temperatures:	74 48 -128 75 47 -128 41 -128 51 58 -128 -128 -128
-128 -128 -128
root@amd:~# 

...is there, so I'm not sure why gkrellm stopped working.
								Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [ibm-acpi-devel] 2.6.33-rc2: regression: gkrellm no longer shows all the temperatures on thinkpad x60

[Henrique de Moraes Holschuh]

On Thu, 07 Jan 2010, Pavel Machek wrote:
> > Please apply the latest stack of patches (sent them to acpi-test
> > yesterday).
> > 
> > It is failing to register the ALSA mixer for some reason, and due
> > to a bug, it is not loading the module at all.  I will look at the reason it
> > is failing to register the ALSA mixer soon.  Meanwhile, the patches I sent
> > to Len make sure the module can still load sucessfully.
> 
> Well, I'm not sure its completely fixed. I got this in my syslog:
> 
> thinkpad_acpi: THERMAL EMERGENCY: a sensor reports something is
> extremely hot!
> thinkpad_acpi: temperatures (Celsius): 95 47 N/A 89 47 N/A 41 N/A 51
> 61 N/A N/A N/A N/A N/A N/A

And the extremely hot sensor is that 89°C, which is the GPU I think (varies
from thinkpad to thinkpad).  That is not normal, and I'd say it needs
repair, probably a badly seated heatsink or failed thermal interface
compound.

> ...meaning thinkpad_acpi is active and working, but still gkrellm does
> not report temperatures as it used to.

Well, gkrellm has always been a hack, it may be objecting to the third N/A
there (which reads as -128 through /proc, and results in ENXIO on hwmon
sysfs).

> root@amd:~# cat /proc/acpi/ibm/thermal 
> temperatures:	74 48 -128 75 47 -128 41 -128 51 58 -128 -128 -128
> -128 -128 -128
> root@amd:~# 

The driver is operating normally (which doesn't surprise me, I didn't touch
that code for quite a while, other than adding the temperature dump so that
we could track what was cooking in your thinkpad :) ).

> ...is there, so I'm not sure why gkrellm stopped working.

I'd take a look on the gkrellm side of things.  Maybe I am doing something
it doesn't like in the hwmon interface, etc.

-- 
  "One disk to rule them all, One disk to find them. One disk to bring
  them all and in the darkness grind them. In the Land of Redmond
  where the shadows lie." -- The Silicon Valley Tarot
  Henrique Holschuh

Re: [ibm-acpi-devel] 2.6.33-rc2: regression: gkrellm no longer shows all the temperatures on thinkpad x60

[Henrique de Moraes Holschuh]

On Sat, 09 Jan 2010, Henrique de Moraes Holschuh wrote:
> On Thu, 07 Jan 2010, Pavel Machek wrote:
> > > Please apply the latest stack of patches (sent them to acpi-test
> > > yesterday).
> > > 
> > > It is failing to register the ALSA mixer for some reason, and due
> > > to a bug, it is not loading the module at all.  I will look at the reason it
> > > is failing to register the ALSA mixer soon.  Meanwhile, the patches I sent
> > > to Len make sure the module can still load sucessfully.
> > 
> > Well, I'm not sure its completely fixed. I got this in my syslog:
> > 
> > thinkpad_acpi: THERMAL EMERGENCY: a sensor reports something is
> > extremely hot!
> > thinkpad_acpi: temperatures (Celsius): 95 47 N/A 89 47 N/A 41 N/A 51
> > 61 N/A N/A N/A N/A N/A N/A
> 
> And the extremely hot sensor is that 89°C, which is the GPU I think (varies
> from thinkpad to thinkpad).  That is not normal, and I'd say it needs
> repair, probably a badly seated heatsink or failed thermal interface
> compound.

Hmm... come to think of it, the CPU is also quite hot (95°C), and it usually
shares the heatsink system with the GPU/north bridge, so the problem might
be on the CPU side of things.

While it can easily be a hardware problem in the heatsink, we could also
have a screwed up ACPI thermal control in our hands once again (either
kernel bug, or bogus ACPI firmware in the laptop).

-- 
  "One disk to rule them all, One disk to find them. One disk to bring
  them all and in the darkness grind them. In the Land of Redmond
  where the shadows lie." -- The Silicon Valley Tarot
  Henrique Holschuh

Re: [ibm-acpi-devel] 2.6.33-rc2: regression: gkrellm no longer shows all the temperatures on thinkpad x60

[Pavel Machek]

Hi!

> > > Well, I'm not sure its completely fixed. I got this in my syslog:
> > > 
> > > thinkpad_acpi: THERMAL EMERGENCY: a sensor reports something is
> > > extremely hot!
> > > thinkpad_acpi: temperatures (Celsius): 95 47 N/A 89 47 N/A 41 N/A 51
> > > 61 N/A N/A N/A N/A N/A N/A
> > 
> > And the extremely hot sensor is that 89°C, which is the GPU I think (varies
> > from thinkpad to thinkpad).  That is not normal, and I'd say it needs
> > repair, probably a badly seated heatsink or failed thermal interface
> > compound.

Well, I'd prefer not to open it unless absolutely neccessary. It is my
primary machine. 

> Hmm... come to think of it, the CPU is also quite hot (95°C), and it usually
> shares the heatsink system with the GPU/north bridge, so the problem might
> be on the CPU side of things.
> 
> While it can easily be a hardware problem in the heatsink, we could also
> have a screwed up ACPI thermal control in our hands once again (either
> kernel bug, or bogus ACPI firmware in the laptop).

Well, galeon (web browser) started misbehaving -- eating 100% cpu --
lately, so maybe that's it.

									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html