Vadim Zeitlin wrote:
> On Sat, 14 Jul 2007 18:19:22 +0200 Hans de Goede <j.w.r.degoede at hhs.nl> wrote:
>
> HdG> > and the AMB sensors 2 and 3 were not being taken into account. Without knowing
> HdG> > anything about libsensors I thought it was because the file names in
> HdG> > /sys/devices/platform/i5k_amb.0 were named {temp0,temp16,temp32,temp48}_xxx and
> HdG> > not temp[1234] as usual
> HdG>
> HdG> Yes, thats a rather severe violation of the hwmon sysfs standard. I do not see
> HdG> this i5k_amb driver in the standard kernel, I assume its an addon driver?
>
> Hello Hans,
>
> This is the patch recently posted to the list:
>
> http://article.gmane.org/gmane.linux.drivers.sensors/14004
>
> sorry for not mentioning it (I forgot I was replying in a different thread
> from my reply in that one).
>
Duh, I actually still have that mail in my Inbox, as I'm planning on reviewing
it as time permits.
> HdG> > and now everything seems to work fine for me except for the 2
> HdG> > problems mentioned in my previous message (long delay when accessing
> HdG> > the sensors and manifestly incorrect values for temp[256]):
> HdG>
> HdG> Does this long delay always happen? If you run sensors in 2 quick
> HdG> successions, the second call should be quite fast.
>
> No, the delay is always present and always the same.
>
Hmm,
I see it refresh the readings every 2 seconds, since reading things takes 1 sec
minimum, I think it would be a good idea to make this somewhat bigger. Darrick,
can you post a new version and or an incremental patch with a slower read
frequency, say once every 5 or 10 seconds?
Also I notice that you are still doing a full read of the chip to data->raw,
this might very well explain some if the slowness. Please remove raw[] from
data and the full read code from the update() function.
Thanks & Regards,
Hans
