Re: [PATCH] gpio: New driver for the Intel 82801 (ICH) GPIO pins

From: Peter Tyser <ptyser@xes-inc.com>
To: guenter.roeck@ericsson.com
Cc: Jean Delvare <khali@linux-fr.org>,
	Grant Likely <grant.likely@secretlab.ca>,
	LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] gpio: New driver for the Intel 82801 (ICH) GPIO pins
Date: Thu, 02 Feb 2012 13:56:55 -0600	[thread overview]
Message-ID: <1328212615.12622.193.camel@petert> (raw)
In-Reply-To: <1328204105.2261.140.camel@groeck-laptop>

<snip>

> > > >  drivers/gpio/Kconfig     |    7
> > > >  drivers/gpio/Makefile    |    1
> > > >  drivers/gpio/i801_gpio.c |  432 ++++++++++++++++++++++++++++++++++++++++++++++
> > > >  3 files changed, 440 insertions(+)
> > >
> > > Did this or an alternate patch for gpio support on the recent Intel ICHs go anywhere ?
> > > 
> > > I found another patch which is using a different approach, but I don't see anything
> > > in the latest kernel.
> > 
> > I don't think any driver made it into the kernel, at least I can't find
> > any in 3.3-rc2 nor linux-next. I'm not surprised, as neither driver was
> > using the MFD framework while this would really be the right thing to
> > do here. I never took the time to convert my driver to MFD and
> > apparently Peter did not either. This is still on my to-do list but you
> > know how long it is. If anyone is faster than me at getting the code is
> > a shape suitable for upstream, I'll be happy to help with review and
> > testing.
> > 
> I'd love to take this on, but unfortunately my task list isn't getting
> shorter either. Just bad that this doesn't seem to make it in.
> 
> One reason of course may be that at least Peter's version tried to
> accomplish too much. Looking through the comments, seems there was a
> disagreement on unrelated issues such as if there should be a new
> "unknown" gpio direction or not. Maybe it would make more sense to
> separate the core patch to add ICH gpio support from the rest of Peter's
> proposed changes.

I agree with Jean, I believe the main issue it wasn't accepted was the
fact that it didn't use the MFD framework.  The driver I submitted
follows the model of other drivers that share the same PCI device (iTCO,
esb2rom, etc), so is functionally OK to use as is, it just doesn't use
the recommend MFD framework, which was gating its acceptance.

> Peter, do you have any plans to work on this in the near future ?

I too have been caught up in other tasks.  A co-worker recently created
a patch to transition the ICHx gpio driver I wrote and other related
drivers (esb2rom, iTCO, etc) to the MFD structure with the intention
that it would eventually be merged upstream.  The patch is still under
review internally, but I'll talk to him about submitting it upstream.

Best,
Peter