Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:5 as permitted sender) client-ip=2620:137:e000::3:5;
MIME-Version: 1.0
References: <20231207125737.5e7553e3@canb.auug.org.au> <CAL_JsqKXo+Cr=9s=dt1kCQeMadJ_cnuSpm06zmvK8yd-vd2X3g@mail.gmail.com>
 <20231208075847.6bbd23b8@canb.auug.org.au>
In-Reply-To: <20231208075847.6bbd23b8@canb.auug.org.au>
From:   Rob Herring <robh@kernel.org>
Date:   Thu, 7 Dec 2023 16:05:27 -0600
Message-ID: <CAL_JsqKyqe827YRyqZnqcVghhuAuvWoiS4U7uR8wwYoGj3=9KA@mail.gmail.com>
Subject: Re: linux-next: build failure after merge of the devicetree tree
To:     Stephen Rothwell <sfr@canb.auug.org.au>
Cc:     Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Linux Next Mailing List <linux-next@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Thu, Dec 7, 2023 at 2:58=E2=80=AFPM Stephen Rothwell <sfr@canb.auug.org.=
au> wrote:
>
> Hi Rob,
>
> On Thu, 7 Dec 2023 09:11:22 -0600 Rob Herring <robh@kernel.org> wrote:
> >
> > I'm sending out fixes for all these. I want to get the final patch
> > ("of: Stop circularly including of_device.h and of_platform.h") for
> > all this in next to get some better build coverage and catch any new
> > drivers added. But if it is dropped for every new driver that breaks,
> > I'll never get it in. Can you fix these up or just leave them broken?
> > I can keep the fixes in my tree until they get applied by the
> > corresponding subsystem.
>
> These dependencies between trees are impossible to handle.  Please if
> you really need the final patch in, then you must put all the necessary
> fixes in the same branch.  There is no telling what order Linus (or I)
> will merge the interdependent branches.
>
> The alternative is to spray the needed fixes out to the other
> subsystems and then put the final patch in after the merge window
> closes or the next release.

Yes, that's what I've been doing since July and plan to do for the
final patch. The final patch can't wait a cycle because then we'll
just have more moles to wack. There's a couple of dependencies that
still haven't been picked up (seems Sparc is unmaintained), so I'm
going to send those along. I'll carry any further fixes until they I
see they are applied.

> I cannot "just leave them broken" because that will interfere with
> other's trying to get their work done. I will try fix up the newly
> added drivers if they are obvious, but in the case of these include file
> cleanups, that can be quite difficult sometimes.

The simple (but not completely correct) fix is usually just add of.h
and platform_device.h. There could be a case of needing of_platform.h
instead of of_device.h or vice-versa, but those should all be fixed
and rare in new code.

Thanks,
Rob