[lm-sensors] Problem with it87 module from Linux 2.6.26 onward
johannes.truschnigg at gmx.at
Sun Oct 19 17:02:26 CEST 2008
Hello Jean, Andrew and List!
On Saturday 18 October 2008 21:18:07 Jean Delvare wrote:
> Hallo Johannes,
> Adding the lm-sensors list in Cc, as well as Andrew Paprocki who has
> been working on the it87 driver lately.
Great, thanks for redirecting the noise I generated to the proper location :)
> On Sat, 18 Oct 2008 19:18:15 +0200, Johannes Truschnigg wrote:
> > Dear Linux kernel module authors,
> > I'm the owner of an ASUS M2NPV-VM mainboard which features a chip driven
> > by the it87 module. I used the PWM capabilities of this chip to reduce my
> > system's noise level, by having its fan speeds regulated via `fancontrol`
> > from lm_sensors, as the BIOS-provided PWM regulation isn't good enough
> > (the fans spin at full speed despite CPUtemp being lower than 40°C,
> > etc.). However, this mechanism stopped working as soon as I upgraded to
> > the 2.6.26 kernel. I was able to dig up a commentary on some mailing list
> > that this "bug" I hit was introduced by choice, and that I could probably
> > work around it by specifying "fix_pwm_polarity=1" as a parameter to the
> > module. I suppose it would work, if the debug ringbuffer wouldn't contain
> > ### snip ###
> > it87 it87.656: PWM configuration looks sane, won't touch
> > ### snip ###
> > after loading the module accordingly. Is there a way to _really_ force
> > 2.6.25's behavior in this manner? Or am I doing it wrong, and there's
> > another way I can get what I want, which is PWM working the way I would
> > like it to?
> There's no way to force the polarity change, for your own good. If you
> did not need to pass fix_pwm_polarity=1 before, then passing it now
> will not solve your problem.
I see, thanks for clearing that up.
> > I'd really appreciate getting an answer on that matter - thank you in any
> > way for taking your time!
> The it87 driver did not change between kernel 2.6.25 and kernel 2.6.26:
Hmm, that indeed is interesting...
> So, I have a hard time believing that fan speed control is working with
> kernel 2.6.25 and doesn't work with kernel 2.6.26 without any other
> change to your system. You'll have to do more tests to figure out what
> exactly broke fan speed control on your system.
You are right indeed; I had 2.6.25 running with the Gentoo patchset applied (
http://dev.gentoo.org/~dsd/genpatches/ ), and switched to vanilla for 2.6.26.
The file /sys/class/hwmon/hwmon1/device/pwm1 disappeared in turn,
and "fancontrol" finds itself unable to regulate the system fans according to
my preferences. Skimming over the patches' descriptions, I couldn't find
anything that'd denote changed to it87, or sth. else that could be related to
my problem, however. I'm not at the machine right now, and would not want to
risk switching kernels and reboot the box blindly without a chance of getting
near it the next two weeks. I'll check asap whether or not fancontrol works
again with gentoo-sources (2.6.27), and let you all know about success or
failure, if you don't mind.
> Please also explain in more details what the actual problem is. "It
> stopped working" doesn't tell us much.
Hope my above paragraph serves a bit better than my initial report - for sake
of completeness, when starting fancontrol, I now get:
Enabling PWM on fans...
/usr/sbin/fancontrol: line 232: hwmon1/device/pwm1: No such file or directory
Error enabling PWM on /sys/class/hwmon/hwmon1/device/pwm1
Aborting, restoring fans...
/usr/sbin/fancontrol: line 189: hwmon1/device/pwm1: No such file or directory
Verify fans have returned to full speed
Thank you very much for your kind and prompt reply! FOSS support actually
with best regards:
- Johannes Truschnigg ( johannes.truschnigg at gmx.at )
phone: +43 650 2 133337
jabber: johannes.truschnigg at gmail.com
Please do not bother me with HTML-eMail or attachments. Thank you.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 189 bytes
Desc: This is a digitally signed message part.
Url : http://lists.lm-sensors.org/pipermail/lm-sensors/attachments/20081019/88cd7355/attachment.bin
More information about the lm-sensors