[GIT PULL] Allwinner conversion to generic pin properties

[Maxime Ripard]

Hi Olof,

On Mon, Jan 09, 2017 at 07:21:20PM -0800, Olof Johansson wrote:
> On Mon, Dec 26, 2016 at 08:42:14AM +0100, Maxime Ripard wrote:
> > Hi Arnd, Olof,
> > 
> > Like we discussed it last time, here is a pull request with the
> > convertion patches to the generic pin properties for 4.10.
> > 
> > Thanks,
> > Maxime
> > 
> > The following changes since commit 7ce7d89f48834cefece7804d38fc5d85382edf77:
> > 
> >   Linux 4.10-rc1 (2016-12-25 16:13:08 -0800)
> > 
> > are available in the git repository at:
> > 
> >   https://git.kernel.org/pub/scm/linux/kernel/git/mripard/linux.git tags/sunxi-generic-pinconf-for-4.10
> > 
> > for you to fetch changes up to 1edcd36fcb48fe841bcc62eda36c105037d9583c:
> > 
> >   ARM: sunxi: Convert pinctrl nodes to generic bindings (2016-12-26 08:27:11 +0100)
> > 
> > ----------------------------------------------------------------
> > Allwinner conversion to generic pin muxing properties
> > 
> > Here is a list of patches that converts the current DT to the generic pin
> > control and muxing properties, now that the pinctrl driver supports it.
> 
> I dropped the ball on this over the holidays for a few reasons, none of them
> good. Unfortunately we're at -rc3+ now, so I've queued this up in next/dt for
> v4.11 at this point. Hope that's alright.
> 
> You might want to use your branch as a base for new DT contents for
> v4.11 in case it ends up being conflict-heavy, which would work fine at our
> end.

We might have conflicts with this unfortunately.

It turns out that we fixed an issue where the pinctrl driver in 4.10
is now honoring the case where we list our pull-ups as disabled, while
it used to be ignored (which was the bug).

However, some boards were listing improperly in their DT that their
pull-ups were disabled, while they actually needed it, which means
that now the driver will disable them, and things get broken (mostly
MMC for now, but I expect other devices to be broken too).

We have two ways of fixing this.

Option 1 is applying a patch that will enable the pull ups on the MMC
devices (because it's ok to do so) that was queued with 4.11 and will
conflict heavily with this PR.

The other option would be to queue the two first patches of that PR as
fixes that will reintroduce the previous behaviour (or at least, the
one the boards expects). I'm guessing in that case, the merge will
just work (since it's the exact same commits anyway), and we fix all
the devices and not just MMC. This has not been confirmed to work at
the moment though, but should be very soon.

What do you think?
Maxime

-- 
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 801 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20170119/dc253827/attachment.sig>