Re: [PATCH 0/6] mfd: tmio: simplify header and move to platform_data

From: Ulf Hansson <ulf.hansson@linaro.org>
To: Lee Jones <lee@kernel.org>
Cc: Wolfram Sang <wsa+renesas@sang-engineering.com>,
	linux-renesas-soc@vger.kernel.org,
	 linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org,  linux-mmc@vger.kernel.org,
	linux-sh@vger.kernel.org
Subject: Re: [PATCH 0/6] mfd: tmio: simplify header and move to platform_data
Date: Mon, 12 Feb 2024 12:17:03 +0100	[thread overview]
Message-ID: <CAPDyKFpho16DU7OorMgXDqiyfFfgM_tWu+DZZOHd0gbjtBw_Cg@mail.gmail.com> (raw)
In-Reply-To: <20240209132837.GJ689448@google.com>

On Fri, 9 Feb 2024 at 14:28, Lee Jones <lee@kernel.org> wrote:
>
> On Fri, 09 Feb 2024, Ulf Hansson wrote:
>
> > On Fri, 9 Feb 2024 at 02:59, Wolfram Sang
> > <wsa+renesas@sang-engineering.com> wrote:
> > >
> > > The MFD parts of the TMIO have been removed by Arnd, so that only the
> > > SD/MMC related functionality is left. Remove the outdated remains in the
> > > public header file and then move it to platform_data as the data is now
> > > specific for the SD/MMC part.
> > >
> > > Based on 6.8-rc3, build bot is happy. Branch is here:
> > >
> > > git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux.git renesas/sdhi/tmio-simplification
> > >
> > > I'd suggest this goes via the MFD tree, so the series would need acks
> > > from the MMC and SH maintainers. Is that okay with everyone?
> >
> > Wouldn't it be better to funnel this via the mmc tree? In that way, we
> > can easily avoid conflicts with additional renesas-mmc driver changes
> > that we have in pipe.
>
> You could say the same about changes SH, MFD and Platform Data have in
> the pipe.
>
> > Or perhaps there are other changes that make the mfd tree preferred?
>
> MFD is usually preferred since the parent device usually lives there and
> we are well accustomed to merging multi-subsystem related sets.
>
> It doesn't really matter how this is merged.  The only stipulation is
> that whoever applies the set does so on a succinct, immutable, tagged
> branch and sends out a pull-request for everyone else to pull from.

You are right.

Although, in this particular case I thought it could make better sense
to use the mmc tree, because 1) we are only removing a header file
from mfd and 2) I know we have other renesas-mmc changes in the pipe
for the next release. The point is, I wanted us to avoid the need for
using an immutable branch. But nevermind.

>
> If you want to do that, there are no complains from me.

Well, it sounds like we may need the flexibility with the immutable
branch, so I suggest we do the usual thing with the mfd tree.

Please add my ack for the mmc related changes.

[...]

Kind regards
Uffe