linux-kernel - Re: [PATCH 3/3] arm64: dts: mediatek: mt8365-pumpkin: Add overlays for thp7312 cameras

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20230907150448.GD17610@pendragon.ideasonboard.com>
Date:   Thu, 7 Sep 2023 18:04:48 +0300
From:   Laurent Pinchart <laurent.pinchart@...asonboard.com>
To:     Paul Elder <paul.elder@...asonboard.com>
Cc:     Kieran Bingham <kieran.bingham@...asonboard.com>,
        Krzysztof Kozlowski <krzysztof.kozlowski@...aro.org>,
        linux-media@...r.kernel.org,
        Mauro Carvalho Chehab <mchehab@...nel.org>,
        Rob Herring <robh+dt@...nel.org>,
        Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
        Conor Dooley <conor+dt@...nel.org>,
        Hans Verkuil <hverkuil-cisco@...all.nl>,
        devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-arm-kernel@...ts.infradead.org,
        linux-mediatek@...ts.infradead.org
Subject: Re: [PATCH 3/3] arm64: dts: mediatek: mt8365-pumpkin: Add overlays
 for thp7312 cameras

On Thu, Sep 07, 2023 at 06:04:09PM +0300, Laurent Pinchart wrote:
> On Thu, Sep 07, 2023 at 11:55:13PM +0900, Paul Elder wrote:
> > On Wed, Sep 06, 2023 at 02:14:29PM +0300, Laurent Pinchart wrote:
> > > On Wed, Sep 06, 2023 at 12:01:43PM +0100, Kieran Bingham wrote:
> > > > Quoting Laurent Pinchart (2023-09-06 10:35:31)
> > > > > On Wed, Sep 06, 2023 at 11:21:31AM +0200, Krzysztof Kozlowski wrote:
> > > > > > On 06/09/2023 11:00, Laurent Pinchart wrote:
> > > > > > >>> has a regulator@0. There are similar instances for clocks.
> > > > > > >>>
> > > > > > >>> I understand why it may not be a good idea, and how the root node is
> > > > > > >>> indeed not a bus. In some cases, those regulators and clocks are grouped
> > > > > > >>> in a regulators or clocks node that has a "simple-bus" compatible. I'm
> > > > > > >>> not sure if that's a good idea, but at least it should validate.
> > > > > > >>>
> > > > > > >>> What's the best practice for discrete board-level clocks and regulators
> > > > > > >>> in overlays ? How do we ensure that their node name will not conflict
> > > > > > >>> with the board to which the overlay is attached ?
> > > > > > >>
> > > > > > >> Top-level nodes (so under /) do not have unit addresses. If they have -
> > > > > > >> it's an error, because it is not a bus. Also, unit address requires reg.
> > > > > > >> No reg? No unit address. DTC reports this as warnings as well.
> > > > > > > 
> > > > > > > I agree with all that, but what's the recommended practice to add
> > > > > > > top-level clocks and regulators in overlays, in a way that avoids
> > > > > > > namespace clashes with the base board ?
> > > > > > 
> > > > > > Whether you use regulator@0 or regulator-0, you have the same chances of
> > > > > > clash.
> > > > > 
> > > > > No disagreement there. My question is whether there's a recommended
> > > > > practice to avoid clashes, or if it's an unsolved problem that gets
> > > > > ignored for now because there's only 36h in a day and there are more
> > > > > urgent things to do.
> > > > 
> > > > Should an overlay add these items to a simple-bus specific to that
> > > > overlay/device that is being supported?
> > > > 
> > > > That would 'namespace' the added fixed-clocks/fixed-regulators etc...
> > > > 
> > > > But maybe it's overengineering or mis-using the simple-bus.
> > > 
> > > You would still need to name the node that groups the regulators and
> > > clocks in a way that wouldn't clash between multiple overlays and the
> > > base board. It would be nice to have nodes that are "private" to an
> > > overlay.
> > 
> > What's the best solution to this then :/
> 
> It seems we don't have a good solution. For now, I'd recommend just
> picking a name for the regulator that has a high chance to be unique,
> like reg-thp7312-1v2 for instance.

Or reg-cam-1v2, or ... The name doesn't matter much really, as long as
it's not extremely generic with a high risk of conflict.

> > > > And the items are still not on a 'bus' with an address - they just exist
> > > > on a presumably externally provided board....

-- 
Regards,

Laurent Pinchart