Re: [PATCH] i2c: remove deprecated callbacks

[Benjamin Herrenschmidt <benh@kernel.crashing.org>]

On Thu, 2011-09-01 at 10:52 +0200, Jean Delvare wrote:
> Hi Cong,
> 
> On Thu,  1 Sep 2011 12:40:27 +0800, WANG Cong wrote:
> > i2c_driver.attach_adapter and i2c_driver.detach_adapter are deprecated
> > and scheduled to be removed in Sep 2011.
> > 
> > Cc: Jean Delvare <khali@linux-fr.org>
> > Signed-off-by: WANG Cong <xiyou.wangcong@gmail.com>
> 
> I'm not sure what you are trying to achieve here.
> 
> I am obviously aware of these functions being deprecated, I am the one
> who added the entry in feature-removal-schedule.txt. And as a matter of
> fact, I have a patch ready for months now.
> 
> I could understand you sending a patch if the entry had expired long
> ago, but this isn't the case: it is marked for September 2011 and you
> send a patch on September 1st. Hardly a coincidence. While there
> certainly is a benefit in people reading feature-removal-schedule.txt
> and trying to clean it up by closing old (presumably forgotten)
> entries, this doesn't apply here.

Unfortunately I'm not finished converting the Mac drivers...

> Furthermore, you didn't even check if the deprecated callbacks were
> still in use. It turns out that there are 9 drivers still using one of
> them (6 macintosh drivers and 3 sound drivers) so plain removing it
> will cause build breakages. We obviously don't want to do that, so I am
> not going to apply your patch (nor mine) for now.
> 
> I know that Benjamin Herrenschmidt is working on converting (some of)
> the macintosh drivers. I don't have any news for some time though. Ben,
> are you done with the conversion by now? It would be great to get it
> merged in kernel 3.2.

I'm just bogged down by too many things, including quite a bit of travel
lately. I have some mostly untested patches converting part of windfarm
(not all of it yet), and therm_adt746x and a patch to add the generic
code to register i2c devices from device-tree properly on powermacs.

I'm missing a couple more drivers & bits of windfarm and therm_pm72
which I'd like to use the opportunity to convert into a windfarm module.

Then there's some sound drivers etc...

I am travelling again tomorrow for 2 weeks, and after that a month in
europe, tho I'll have 10 days in the middle where I _might_ be able to
make progress on that. Else it will have to wait for november.
 
> Cong, do you have a specific interest in these callbacks being removed?
> 
> If you want to help, then instead of sending a patch which I already
> have and virtually everybody could have written, please get the
> remaining 3 drivers converted to the current i2c binding model:
>   sound/ppc/keywest.c
>   sound/aoa/codecs/onyx.c
>   sound/aoa/codecs/tas.c
> THAT would be helpful. At the moment, it only looks like you are trying
> to get your name on a trivial patch, i.e. getting credit for work you
> didn't really do.

Cheers,
Ben.

> > 
> > ---
> > diff --git a/Documentation/feature-removal-schedule.txt b/Documentation/feature-removal-schedule.txt
> > index c4a6e14..89a9695 100644
> > --- a/Documentation/feature-removal-schedule.txt
> > +++ b/Documentation/feature-removal-schedule.txt
> > @@ -486,15 +486,6 @@ Files:	include/linux/netfilter_ipv4/ipt_addrtype.h
> >  
> >  ----------------------------
> >  
> > -What:	i2c_driver.attach_adapter
> > -	i2c_driver.detach_adapter
> > -When:	September 2011
> > -Why:	These legacy callbacks should no longer be used as i2c-core offers
> > -	a variety of preferable alternative ways to instantiate I2C devices.
> > -Who:	Jean Delvare <khali@linux-fr.org>
> > -
> > -----------------------------
> > -
> > (...)
>