[lm-sensors] Any chances to get THMC50 driver into 2.6.22?

[lm-sensors] Any chances to get THMC50 driver into 2.6.22?

[Krzysztof Helt]

Hi,

Are there any chances to get the driver posted here:

http://lists.lm-sensors.org/pipermail/lm-sensors/2007-March/019094.html

into the 2.6.22?

Regards,
Krzysztof

----------------------------------------------------
Nie ka¿da stopka emailowa musi byæ nudna:) 
Shivering Isles - dodatek do Obliviona - najlepszej gry 
komputerowej RPG 2006 roku - zagraj, odmieñ swoje ¿ycie!
http://klik.wp.pl/?adr=http%3A%2F%2Fadv.reklama.wp.pl%2Fas%2Foblivion.html&sid\x1139



_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] Any chances to get THMC50 driver into 2.6.22?

[Jean Delvare]

Hi Krzysztof,

On Tue, 15 May 2007 14:50:25 +0200, Krzysztof Helt wrote:
> Are there any chances to get the driver posted here:
> 
> http://lists.lm-sensors.org/pipermail/lm-sensors/2007-March/019094.html
> 
> into the 2.6.22?

No chance, sorry. The merge window for 2.6.22 is already closed, and
your driver didn't even receive its first review as far as I can see.

I can understand your frustration, but as you can see, we're all pretty
busy, and a driver for a couple rare and old chips can't be given a high
priority :(

What you can do to help your case:
* Find someone to make a first review of your driver. Maybe someone who
  him/herself is waiting for a review. Hans de Goede and Juerg
  Haefliger did such a "cross-review" a few weeks ago, and as a result
  Juerg's dme1737 driver is almost ready for upstream.
* Help the project with something else. The time you save us, we can
  put in reviewing your code.

-- 
Jean Delvare

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] Any chances to get THMC50 driver into 2.6.22?

[Krzysztof Helt]

Dnia 15-05-2007 o godz. 18:38 Jean Delvare napisa³(a):
>
> What you can do to help your case:
> * Find someone to make a first review of your driver. Maybe
someone who
>   him/herself is waiting for a review. Hans de Goede and Juerg
>   Haefliger did such a "cross-review" a few weeks ago, and as a
result
>   Juerg's dme1737 driver is almost ready for upstream.

I'll try.

> * Help the project with something else. The time you save us,
we can
>   put in reviewing your code.
> 

Can you point me these issues?

Regards,
Krzysztof

----------------------------------------------------
NaS - na jedynym koncercie w Polsce. 8.06 20:00 STODO£A 
(ul.Batorego 10) Warszawa. Bilety do nabycia w Ticketpro, 
Shortcut, kasa Stodo³y. Go¶æ spec: Jêdker /WWO/. 
Afterparty - klub Fresh. http://klik.wp.pl/?adr=http%3A%2F%2Fadv.reklama.wp.pl%2Fas%2Fnas.html&sid\x1140



_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] Any chances to get THMC50 driver into 2.6.22?

[Jean Delvare]

Hi Krzysztof,

On Wed, 16 May 2007 08:16:03 +0200, Krzysztof Helt wrote:
> Dnia 15-05-2007 o godz. 18:38 Jean Delvare napisał(a):
> > * Help the project with something else. The time you save us,
> >   we can put in reviewing your code.
> 
> Can you point me these issues?

Well, user support to start with, both on this mailing list and
http://www.lm-sensors.org/report . Some tickets are pretty simple, for
example everyone should be able to solve #2208. And review or testing
of any patch posted to the list, that's always welcome.

You can also test 2.6.22-rc2 or later on your systems and confirm that
there are no regressions concerning lm-sensors.

If you tell us which drivers you can test, I can also look at these
drivers for needed cleanups. For example lots of drivers are still
using ugly macros instead of "dynamic" sysfs attributes so there is
room for improvement.

Then we have a number of big projects in progress at the moment:

* Conversion of i2c-isa-based drivers to platform drivers. Six drivers
remain to be converted. w83627ehf is handled by David Hubbard and is
almost done, it87 and pc87360 are supposedly handled by Corentin Labbe
and Jim Cromie respectively, but I didn't see anything yet. vt8231,
sis5595 and via686a are PCI-based, how we convert them depends on
possible changes to the pci subsystem so I was delaying them a bit, but
we want them in 2.6.23 so I will have to decide quickly. So if you have
access to a device supported by any of these 6 drivers, you could help
with the conversion, either doing it yourself or reviewing or testing.

* Addition of individual alarm files to all hardware monitoring
drivers. A small number of drivers are already converted but there are
still many left. Individual alarm files are required for the future
libsensors. The naming scheme for these files is described in
Documentation/hwmon/sysfs-interface. To figure out the bit mapping of
the alarms, you need to refer to either the datasheet or the 2.4 driver
(in lm-sensors, under kernel/chips).

* We're also working on the new libsensors in branch "3.0.0" of
lm-sensors SVN. The new version scans sysfs for supported files
automatically so you no longer need to add support for every new
driver. Help is needed with reviewing the new code, and testing how it
works with as many drivers and devices as possible. Hans de Goede
posted about this some times ago and didn't receive any answer.

Oh, and if you have experience with writing drivers for serial (RS232)
devices, please let me know, I am trying to write such a driver for an
evaluation module but am going nowhere.

Thanks,
-- 
Jean Delvare

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] Any chances to get THMC50 driver into 2.6.22?

[Krzysztof Helt]

Jean Delvare wrote:
> Hi Krzysztof,

> Well, user support to start with, both on this mailing list and
> http://www.lm-sensors.org/report . Some tickets are pretty simple, for
> example everyone should be able to solve #2208. And review or testing
> of any patch posted to the list, that's always welcome.
> 

I may try fixing them when I do not have any other task at hand (read: I wait for patches accepted in other areas: sparc/fbdev).


> You can also test 2.6.22-rc2 or later on your systems and confirm that
> there are no regressions concerning lm-sensors.
> 

Fine. Unfortunately, I can do this easily only on two computers (A and B, see below).

> If you tell us which drivers you can test, I can also look at these
> drivers for needed cleanups. For example lots of drivers are still
> using ugly macros instead of "dynamic" sysfs attributes so there is
> room for improvement.
> 

I have these three machines:

A. Compaq AP550: i82801AA + dual ADM1022 (thmc50 driver) = 5 temperatures and 2 fans. There is also some LPC chip but it is not used I suppose (no other fans).

B. laptop: sis630 + max1617 (adm1021? driver) + unknown ITE chip (id: ea19). I got 2 temperatures (CPU & MB) but no fan control. I suppose this ITE chip controls the fan (do you know where I can find manual for it?).

C. White box P3: piix4 + as99127f rev 2. This one is my wife's so the access is more limited.

> Then we have a number of big projects in progress at the moment:
> 
> * Conversion of i2c-isa-based drivers to platform drivers. Six drivers
> remain to be converted. w83627ehf is handled by David Hubbard and is
> almost done, it87 and pc87360 are supposedly handled by Corentin Labbe
> and Jim Cromie respectively, but I didn't see anything yet. vt8231,
> sis5595 and via686a are PCI-based, how we convert them depends on
> possible changes to the pci subsystem so I was delaying them a bit, but
> we want them in 2.6.23 so I will have to decide quickly. So if you have
> access to a device supported by any of these 6 drivers, you could help
> with the conversion, either doing it yourself or reviewing or testing.
> 

Sorry, I do not have any of them.

> * Addition of individual alarm files to all hardware monitoring
> drivers. A small number of drivers are already converted but there are
> still many left. Individual alarm files are required for the future
> libsensors. The naming scheme for these files is described in
> Documentation/hwmon/sysfs-interface. To figure out the bit mapping of
> the alarms, you need to refer to either the datasheet or the 2.4 driver
> (in lm-sensors, under kernel/chips).
> 

I can add them for the hardware I have.

> * We're also working on the new libsensors in branch "3.0.0" of
> lm-sensors SVN. The new version scans sysfs for supported files
> automatically so you no longer need to add support for every new
> driver. Help is needed with reviewing the new code, and testing how it
> works with as many drivers and devices as possible. Hans de Goede
> posted about this some times ago and didn't receive any answer.
> 

I can test it with the hardware I have.

> Oh, and if you have experience with writing drivers for serial (RS232)
> devices, please let me know, I am trying to write such a driver for an
> evaluation module but am going nowhere.
> 

I do not know how deep knowledge you need.

Kind regards,
Krzysztof


_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] Any chances to get THMC50 driver into 2.6.22?

[Jean Delvare]

Hi Krzysztof,

On Tue, 22 May 2007 21:25:12 +0200, Krzysztof Helt wrote:
> > Well, user support to start with, both on this mailing list and
> > http://www.lm-sensors.org/report . Some tickets are pretty simple, for
> > example everyone should be able to solve #2208. And review or testing
> > of any patch posted to the list, that's always welcome.
> 
> I may try fixing them when I do not have any other task at hand (read:
> I wait for patches accepted in other areas: sparc/fbdev).

For example, Rudolf Marek just posted a patch adding support for the
IT8726F:
http://lists.lm-sensors.org/pipermail/lm-sensors/2007-May/019767.html
It should be fairly easy to review. Of course I'll end up doing it if
nobody picks it shortly, but it's probably a good one to get started.

> I have these three machines:
> 
> A. Compaq AP550: i82801AA + dual ADM1022 (thmc50 driver) = 5 temperatures
> and 2 fans. There is also some LPC chip but it is not used I suppose (no
> other fans).
> 
> B. laptop: sis630 + max1617 (adm1021? driver) + unknown ITE chip (id: ea19).
> I got 2 temperatures (CPU & MB) but no fan control. I suppose this ITE chip
> controls the fan (do you know where I can find manual for it?).

Are you sure it is an ITE chip? ID 0xea doesn't sound like an ITE chip
at all, more like National Semiconductor. The earlier versions of
sensors-detect might have been confusing in that respect, please check
again with a recent version.

The adm1021 driver is still using macro-generated functions instead of
dynamic callbacks, and doesn't implement individual alarm files. So
these are two things you could work on if you feel like it.

> C. White box P3: piix4 + as99127f rev 2. This one is my wife's so the access
> is more limited.

Same thing for the w83781d driver, dynamic callbacks and individual
alarm files are missing.

> > Oh, and if you have experience with writing drivers for serial (RS232)
> > devices, please let me know, I am trying to write such a driver for an
> > evaluation module but am going nowhere.
> 
> I do not know how deep knowledge you need.

Basically my problem is that I can't get the driver to bind to my
device. I am able to "talk" to the device using minicom so I know the
serial port and the device work OK. But I don't know how to write a
kernel driver which will bind to the serial port and let me read and
write "raw" characters from and to it. I tried copying random examples
from other serial drivers but my .connect() method never gets called.
LDD3 unfortunately doesn't cover legacy serial drivers.

Thanks,
-- 
Jean Delvare

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] Any chances to get THMC50 driver into 2.6.22?

[Jean Delvare]

Hi Krzysztof,

On Wed, 23 May 2007 20:03:41 +0200, Krzysztof Helt wrote:
> Jean Delvare wrote:
> > For example, Rudolf Marek just posted a patch adding support for the
> > IT8726F:
> > http://lists.lm-sensors.org/pipermail/lm-sensors/2007-May/019767.html
> > It should be fairly easy to review. Of course I'll end up doing it if
> > nobody picks it shortly, but it's probably a good one to get started.
> 
> I looked at the changes and I do not see anything wrong. It just additions
> of constants and if clauses extensions.  The only "trick" is that it8726
> identifies the driver as it8716 but it may be ok if they both have exactly
> the same functionality. In short words: it is only addition of the new chip
> id and nothing more.

OK, thanks for the review. Next time, please reply to the post in
question, it makes it easier for me to track how much attention a given
patch received. And Cc'ing the list (as I am doing now) is always
preferred, so that others know what's going on.

> >> B. laptop: sis630 + max1617 (adm1021? driver) + unknown ITE chip (id: ea19).
> >> I got 2 temperatures (CPU & MB) but no fan control. I suppose this ITE chip
> >> controls the fan (do you know where I can find manual for it?).
> > 
> > Are you sure it is an ITE chip? ID 0xea doesn't sound like an ITE chip
> > at all, more like National Semiconductor. The earlier versions of
> > sensors-detect might have been confusing in that respect, please check
> > again with a recent version.
> 
> I am not sure. It is detected as ITE chip as well. It may be the NS chip, though. 

Older versions of sensors-detect would list all possible manufacturers,
which was admittedly confusing. The version in SVN will only list the
correct manufacturer.

-- 
Jean Delvare

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors