Re: therm_adt746x: attach_adapter method is deprecated (fwd)

Re: therm_adt746x: attach_adapter method is deprecated (fwd)

[Jean Delvare]

Hi Christian,

Redirecting to the right mailing list, as 1* this is an i2c issue and
2* the therm_adt746x driver doesn't integrate into the standard
hwmon/libsensors interface anyway.

On Thu, 21 Apr 2011 17:59:50 -0700 (PDT), Christian Kujau wrote:
> Starting with 2.6.38+, I noticed the following messages during bootup:
> 
> i2c i2c-0: attach_adapter method is deprecated
> Please use another way to instantiate your i2c_client
> 
> adt746x seems to continue to work, though. This is on a PowerBook G4.
> 
> Is this something to be worried about?

This will be something to worry about if nobody fixes the code before
September 2011, when the deprecated methods get removed.

I think Ben is working on this. Ben, any progress? Did you manage to
convert the code? If not, is there anything I can do to help?

-- 
Jean Delvare

Re: therm_adt746x: attach_adapter method is deprecated (fwd)

[Christian Kujau]

On Fri, 22 Apr 2011 at 10:31, Jean Delvare wrote:
> Redirecting to the right mailing list, as 1* this is an i2c issue and
> 2* the therm_adt746x driver doesn't integrate into the standard
> hwmon/libsensors interface anyway.

Ah, thanks for redirecting. Yeah, lm-sensors never worked for me and I'm 
still reading fan speeds/temperatures off of /sys/devices/temperatures/*

Christian.
-- 
BOFH excuse #140:

LBNC (luser brain not connected)

Re: therm_adt746x: attach_adapter method is deprecated (fwd)

[Jean Delvare]

On Fri, 22 Apr 2011 10:59:16 -0700 (PDT), Christian Kujau wrote:
> On Fri, 22 Apr 2011 at 10:31, Jean Delvare wrote:
> > Redirecting to the right mailing list, as 1* this is an i2c issue and
> > 2* the therm_adt746x driver doesn't integrate into the standard
> > hwmon/libsensors interface anyway.
> 
> Ah, thanks for redirecting. Yeah, lm-sensors never worked for me and I'm 
> still reading fan speeds/temperatures off of /sys/devices/temperatures/*

Note that I think it would make a lot of sense to expose all the
monitored values at least read-only through a hwmon-style interface in
all these powermac specific drivers. This would let these values be
reported to users by all standard monitoring tools.

Independent from the warnings you were reported, though.

-- 
Jean Delvare

Re: therm_adt746x: attach_adapter method is deprecated (fwd)

[Benjamin Herrenschmidt]

On Fri, 2011-04-22 at 20:18 +0200, Jean Delvare wrote:
> On Fri, 22 Apr 2011 10:59:16 -0700 (PDT), Christian Kujau wrote:
> > On Fri, 22 Apr 2011 at 10:31, Jean Delvare wrote:
> > > Redirecting to the right mailing list, as 1* this is an i2c issue and
> > > 2* the therm_adt746x driver doesn't integrate into the standard
> > > hwmon/libsensors interface anyway.
> > 
> > Ah, thanks for redirecting. Yeah, lm-sensors never worked for me and I'm 
> > still reading fan speeds/temperatures off of /sys/devices/temperatures/*
> 
> Note that I think it would make a lot of sense to expose all the
> monitored values at least read-only through a hwmon-style interface in
> all these powermac specific drivers. This would let these values be
> reported to users by all standard monitoring tools.
> 
> Independent from the warnings you were reported, though.

Ok, I've finally decided to dedicate a couple of days in fixing them,
it's non-trivial, but I'm getting there slowly. In some cases (windfarm)
it's easy and actually removes significant amounts of code from the
drivers. In other cases, it's harder because the drivers are pretty
awful in the first place (therm_adt746x).

I might completely replace therm_pm72 with a windfarm implementation for
those older G5's which would be more consistent (all G5's would thus use
windfarm) since it's cleaner (therm_pm72 hard code i2c addresses etc...
while windfarm is meant for more dynamic probing of the sensors).

Regarding the hwmon interface, well, that's a whole different issue...

Windfarm generically creates sysfs files for all sensors so I suppose it
-could- start exposing them in a more standard interface.

therm_adt746x creates its own special /sys/devices/temperatures with all
the files in it. I don't want to remove it for now but it might be
worthwhile adding an hwmon interface to it, at least I can still test
on it.

therm_windtunnel is similar but I don't have the HW to test so that's
going to be a bit nastier.

I hope to have something before the end of the week for you to review.

Cheers,
Ben. 

Re: therm_adt746x: attach_adapter method is deprecated (fwd)

[Jean Delvare]

Hi Benjamin,

On Tue, 05 Jul 2011 17:20:40 +1000, Benjamin Herrenschmidt wrote:
> On Fri, 2011-04-22 at 20:18 +0200, Jean Delvare wrote:
> > Note that I think it would make a lot of sense to expose all the
> > monitored values at least read-only through a hwmon-style interface in
> > all these powermac specific drivers. This would let these values be
> > reported to users by all standard monitoring tools.
> > 
> > Independent from the warnings you were reported, though.
> 
> Ok, I've finally decided to dedicate a couple of days in fixing them,
> it's non-trivial, but I'm getting there slowly. In some cases (windfarm)
> it's easy and actually removes significant amounts of code from the
> drivers. In other cases, it's harder because the drivers are pretty
> awful in the first place (therm_adt746x).

I am glad to hear that the change will simplify the code in at least
some cases.

> I might completely replace therm_pm72 with a windfarm implementation for
> those older G5's which would be more consistent (all G5's would thus use
> windfarm) since it's cleaner (therm_pm72 hard code i2c addresses etc...
> while windfarm is meant for more dynamic probing of the sensors).
> 
> Regarding the hwmon interface, well, that's a whole different issue...
> 
> Windfarm generically creates sysfs files for all sensors so I suppose it
> -could- start exposing them in a more standard interface.
> 
> therm_adt746x creates its own special /sys/devices/temperatures with all
> the files in it. I don't want to remove it for now but it might be
> worthwhile adding an hwmon interface to it, at least I can still test
> on it.

Sounds good.

> therm_windtunnel is similar but I don't have the HW to test so that's
> going to be a bit nastier.
> 
> I hope to have something before the end of the week for you to review.

Sure, as soon as you have anything ready, I'll be happy to review it.

-- 
Jean Delvare

Re: therm_adt746x: attach_adapter method is deprecated (fwd)

[Christian Kujau]

On Tue, 5 Jul 2011 at 10:27, Jean Delvare wrote:
> > I hope to have something before the end of the week for you to review.
> 
> Sure, as soon as you have anything ready, I'll be happy to review it.

...and I'll be happy to test (on my Powerbook G4) :-)

Thanks,
Christian.
-- 
BOFH excuse #149:

Dew on the telephone lines.