[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <eb9f2d20-9b51-45e1-bf32-504267747f06@collabora.com>
Date: Tue, 3 Sep 2024 01:14:42 +0300
From: Cristian Ciocaltea <cristian.ciocaltea@...labora.com>
To: Shimrra Shai <shimrrashai@...il.com>
Cc: Laurent.pinchart@...asonboard.com, aarnoud@...com, airlied@...il.com,
andrzej.hajda@...el.com, andy.yan@...k-chips.com, conor+dt@...nel.org,
daniel@...ll.ch, devicetree@...r.kernel.org,
dri-devel@...ts.freedesktop.org, heiko@...ech.de, hjc@...k-chips.com,
jernej.skrabec@...il.com, jonas@...boo.se, kernel@...labora.com,
krzk+dt@...nel.org, krzk@...nel.org, ldearquer@...il.com,
linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
linux-rockchip@...ts.infradead.org, maarten.lankhorst@...ux.intel.com,
markyao0591@...il.com, mripard@...nel.org, neil.armstrong@...aro.org,
rfoss@...nel.org, robh@...nel.org, s.hauer@...gutronix.de,
tzimmermann@...e.de
Subject: Re: [PATCH v5 3/4] dt-bindings: display: rockchip: Add schema for
RK3588 HDMI TX Controller
On 9/2/24 4:09 AM, Shimrra Shai wrote:
> Cristian Ciocaltea wrote:
>> On 8/31/24 9:13 AM, Krzysztof Kozlowski wrote:
>>>
>>> Please define all clocks.
>>
>> The other clocks are defined in the common binding, should we reiterate
>> them?
>
> I would suggest yes, they should be reduplicated, if only to maintain
> consistency with all the other docs. A grep through the bridge docs
> shows that there are virtually none which use a "{}" placeholder like
> this.
Are you sure about that? This is precisely the approach followed by the
upstream DW HDMI TX controller binding [1]. Moreover, I've already pointed this
out in [2].
> While it seems kind of like one might worry about "don't
> repeat yourself" syndrome, keep in mind this is not code, but human-
> used documentation. Having all the information available at a glance
> would seem to be the most convenient to the end (developer) user, so
> they aren't having to toggle between two separate files.
I think that's pretty subjective to be stated as a general rule, e.g. I don't
have any problem toggling between multiple files as I regularly keep over 50
files opened in my IDE. Personally, I'd always go for the slightly less
readable approach if I can avoid duplicating content.
I'd suggest to follow the whole thread [2], as this topic has been already
discussed.
> Of course
> there may be some questions regarding docs becoming out of sync, but
> *ideally* we don't want to break backward compatibility with device
> trees (esp. given how I am imagining firmware integration to work on
> these platforms, as the RK3588 is at at least low-end desktop-grade
> performance and UEFI packages have already been built for it), though
> of course that doesn't mean adding new options is off the table.
>
> (FWIW, this is what I did in my now-withdrawn-at-your-request
> re-submission; I reduplicated the bindings as it seemed that's what
> others here were pushing for and thus that felt like the quickest way
> to get this important driver approved.)
This is not really a blocker for the series. Please remember to be patient
while involved (one way or another) in the upstreaming process, as maintainers
need time to review *all* the patches. This might be very important for you,
but there are, usually, tons of other way more important things the maintainers
need to handle in parallel.
[1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/devicetree/bindings/display/rockchip/rockchip,dw-hdmi.yaml#n46
[2] https://lore.kernel.org/lkml/ec84bc0b-c4c2-4735-9f34-52bc3a852aaf@collabora.com/
Powered by blists - more mailing lists