Re: [RFC PATCH v2 6/7] dt-bindings: arm: mediatek: Remove SKU specific compatibles for Google Krane

From: Doug Anderson <dianders@chromium.org>
To: Rob Herring <robh@kernel.org>
Cc: Chen-Yu Tsai <wenst@chromium.org>,
	Frank Rowand <frowand.list@gmail.com>,
	Krzysztof Kozlowski <krzysztof.kozlowski+dt@linaro.org>,
	Conor Dooley <conor+dt@kernel.org>,
	Matthias Brugger <matthias.bgg@gmail.com>,
	AngeloGioacchino Del Regno 
	<angelogioacchino.delregno@collabora.com>,
	Hsin-Yi Wang <hsinyi@chromium.org>,
	Dmitry Torokhov <dmitry.torokhov@gmail.com>,
	andriy.shevchenko@linux.intel.com, Jiri Kosina <jikos@kernel.org>,
	linus.walleij@linaro.org, broonie@kernel.org,
	gregkh@linuxfoundation.org, hdegoede@redhat.com,
	james.clark@arm.com, james@equiv.tech, keescook@chromium.org,
	petr.tesarik.ext@huawei.com, rafael@kernel.org,
	tglx@linutronix.de, Jeff LaBundy <jeff@labundy.com>,
	linux-input@vger.kernel.org, devicetree@vger.kernel.org,
	linux-arm-kernel@lists.infradead.org,
	linux-mediatek@lists.infradead.org, linux-kernel@vger.kernel.org,
	Johan Hovold <johan@kernel.org>
Subject: Re: [RFC PATCH v2 6/7] dt-bindings: arm: mediatek: Remove SKU specific compatibles for Google Krane
Date: Fri, 10 Nov 2023 16:29:08 -0800	[thread overview]
Message-ID: <CAD=FV=XARZpeEZo-i+CNTAH=AQrxPfwnzSMXDWx6Pvj4k6cP8g@mail.gmail.com> (raw)
In-Reply-To: <20231110210443.GA419831-robh@kernel.org>

Hi,

On Fri, Nov 10, 2023 at 1:04 PM Rob Herring <robh@kernel.org> wrote:
>
> On Thu, Nov 09, 2023 at 06:06:03PM +0800, Chen-Yu Tsai wrote:
> > In cases where the same Chromebook model is manufactured with different
> > components (MIPI DSI panels, MIPI CSI camera sensors, or trackpad /
> > touchscreens with conflicting addresses), a different SKU ID is
> > allocated to each specific combination. This SKU ID is exported by the
> > bootloader into the device tree, and can be used to "discover" which
> > combination is present on the current machine. Thus we no longer have
> > to specify separate compatible strings for each of them.
>
> You just broke an existing kernel with a new DT having this change.
>
> Just because you come up with a new way to do things, doesn't mean you
> can remove the old way.

I was wondering about that, actually. My understanding was that what
Chen-Yu was doing here was correct, but I'm happy to be educated.

Specifically, I think that after his series old device trees will
continue to boot just fine. ...so if someone took a device tree from
before his series and booted it on a kernel after his series that
everything would be hunky dory. If that doesn't work then, I agree,
that should be fixed.

However, here, he is documenting what the "latest and greatest" device
tree should look at and that matches what's checked into the "dts"
directory. In general, I thought that yaml files didn't necessarily
always document old/deprecated ways of doing things and just focused
on documenting the new/best way.

Now, obviously, if someone took a new device tree and tried to put it
on an old kernel then it wouldn't work, but I was always under the
impression that wasn't a requirement.

-Doug