[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Z3/lA+w7BoazNeWk@hu-wasimn-hyd.qualcomm.com>
Date: Thu, 9 Jan 2025 20:33:26 +0530
From: Wasim Nazir <quic_wasimn@...cinc.com>
To: Krzysztof Kozlowski <krzk@...nel.org>
CC: Dmitry Baryshkov <dmitry.baryshkov@...aro.org>,
Bjorn Andersson
<andersson@...nel.org>,
Konrad Dybcio <konradybcio@...nel.org>, Rob Herring
<robh@...nel.org>,
Krzysztof Kozlowski <krzk+dt@...nel.org>,
Conor Dooley
<conor+dt@...nel.org>, <linux-arm-msm@...r.kernel.org>,
<devicetree@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
<kernel@...cinc.com>
Subject: Re: [PATCH v5 5/6] arm64: dts: qcom: Add support for QCS9075 Ride &
Ride-r3
On Thu, Jan 09, 2025 at 08:17:54PM +0530, Wasim Nazir wrote:
> On Wed, Jan 08, 2025 at 03:09:09PM +0100, Krzysztof Kozlowski wrote:
> > On 03/01/2025 20:58, Dmitry Baryshkov wrote:
> > >>>>>> Initially, we included the DTS [1] file to avoid duplication. However,
> > >>>>>> based on Krzysztof's previous suggestion [2], we change to this format.
> > >>>>>>
> > >>>>>> Please let us know how to proceed further on this.
> > >>>>>
> > >>>>> Krzysztof asked you to include DTSI files instead of including DTS
> > >>>>> files. Hope this helps.
> > >>>>
> > >>>> Are you suggesting that we should also modify the 9100-ride files to
> > >>>> include DTSI instead of DTS for consistency between QCS9100 and QCS9075?
> > >>>> However, this would result in the duplication of Ethernet nodes in all
> > >>>> the ride board files. Would that be acceptable?
> > >>>
> > >>> git mv foo.dts foo.dtsi
> > >>> echo '#include "foo.dtsi"' > foo.dts
> > >>> git add foo.dts
> > >>> git commit
> > >>>
> > >>
> > >> We cannot convert sa8775p-ride-r3.dts and sa8775p-ride.dts to .dtsi as
> > >> they represent different platforms. In patch [1], we included these DTS
> > >> files to reuse the common hardware nodes.
> > >>
> > >> Could you please advise on how we should proceed with the following
> > >> approaches?
> > >>
> > >> a) Previous approach [1]:
> > >> Include sa8775p-ride-r3.dts and sa8775p-ride.dts in the qcs9075-ride
> > >> platform DTS, similar to the qcs9100-ride platform DTS. This approach
> > >> avoids duplicating Ethernet nodes and maintains uniformity. However, it
> > >> involves including the DTS file directly.
> > >>
> > >> b) Current suggestion:
> > >> Include sa8775p-ride.dtsi in the qcs9075-ride platform DTS and also
> > >> modify the qcs9100-ride platform DTS files to maintain uniformity. This
> > >> approach results in duplicating Ethernet nodes.
> > >>
> > >> Please let us know your recommendation to finalize the DT structure.
> > >
> > > sa8775p.dtsi
> > > `__sa8775p-ride.dtsi
> > > `__sa8775p-ride-r2.dtsi
> > > `__sa8775p-ride.dts
> > > `__qcs9100-ride.dts
> > > `__qcs9075-ride.dts
> > > `__sa8775p-ride-r3.dtsi
> > > `__sa8775p-ride-r3.dts
> > > `__qcs9100-ride-r3.dts
> > > `__qcs9075-ride-r3.dts
> > >
> > Wasim and all other copy-pasters of sa8775p-ride,
> >
> > Just to recap, qcs9100 contributions started this terrible pattern of
> > board including a board. Unfortunately qcs9100 was merged, so that ship
> > has sailed.
> >
> > This patchset was going the same way, because poor choices like to keep
> > spreading, but at one of previous versions I noticed it and objected.
> >
> > This v5 however solves above problem by duplicating the nodes.
> >
> > Apparently all these designs - sa8755p, qcs9100 and qcs9075 - use the
> > same board, but none of this was communicated. I checked all the commit
> > msgs in this patchset and nothing explained about it. What annoys me is
> > that you do not communicate your design forcing us to accept poor DTS or
> > forcing us to guess and make poor judgments.
> >
> > Come with proper hardware description and split out shared parts, like
> > motherboard. Look how other vendors are doing it, e.g. NXP or Renesas.
> > But assuming there are shared parts because I am pretty sure you will
> > pick my comments when it suits you without actually following them fully
> > and without understanding and explaining to us your own hardware.
> >
>
> Hi Krzysztof,
>
> Here is the pictorial flow showing how SoCs are derived and what all boards
> are supported.
>
+-----------------------------------------------------------------------+
| |
| sa8775p |
| | |
| +-----------------------+-----------------------+ |
| | | | |
| v | v |
| qcs9100 | qcs9075 |
| | | | |
| v v v |
| (IOT) (AUTO) (IOT) |
| qcs9100-ride.dts sa8775p-ride.dts qcs9075-ride.dts |
| qcs9100-ride-r3.dts sa8775p-ride-r3.dts qcs9075-ride-r3.dts |
| qcs9075-rb8.dts |
| |
+-----------------------------------------------------------------------+
Updating it as previous one is messed up with whitespaces.
>
> Although we included details about the QCS9075 and QCS9100 in the cover
> letter and commit message, explaining their differences and common
> origin from the SA8775P SOC, the new DT structure suggested by Dmitry
> should make things clearer. This structure properly splits common parts
> and enhances reusability.
>
> > Best regards,
> > Krzysztof
>
> Thanks & Regards,
> Wasim
Powered by blists - more mailing lists