linux-kernel - Re: [PATCH V5 4/4] arm64: dts: qcom: sm8550: Add max-sd-hs-hz property

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: <d12957f0-f8b5-4b29-967c-576dadd565de@oss.qualcomm.com>
Date: Thu, 4 Sep 2025 14:27:04 +0200
From: Konrad Dybcio <konrad.dybcio@....qualcomm.com>
To: Krzysztof Kozlowski <krzk@...nel.org>,
        Sarthak Garg <quic_sartgarg@...cinc.com>,
        Ulf Hansson <ulf.hansson@...aro.org>, Rob Herring <robh@...nel.org>,
        Krzysztof Kozlowski <krzk+dt@...nel.org>,
        Conor Dooley
 <conor+dt@...nel.org>,
        Bjorn Andersson <andersson@...nel.org>,
        Konrad Dybcio <konradybcio@...nel.org>,
        Adrian Hunter <adrian.hunter@...el.com>
Cc: linux-mmc@...r.kernel.org, devicetree@...r.kernel.org,
        linux-kernel@...r.kernel.org, linux-arm-msm@...r.kernel.org,
        quic_nguyenb@...cinc.com, quic_rampraka@...cinc.com,
        quic_pragalla@...cinc.com, quic_sayalil@...cinc.com,
        quic_nitirawa@...cinc.com, quic_bhaskarv@...cinc.com,
        kernel@....qualcomm.com
Subject: Re: [PATCH V5 4/4] arm64: dts: qcom: sm8550: Add max-sd-hs-hz
 property

On 9/4/25 12:52 PM, Krzysztof Kozlowski wrote:
> On 04/09/2025 12:51, Krzysztof Kozlowski wrote:
>> On 04/09/2025 10:36, Konrad Dybcio wrote:
>>> On 9/3/25 10:21 AM, 'Krzysztof Kozlowski' via kernel wrote:
>>>> On 03/09/2025 10:04, Sarthak Garg wrote:
>>>>> Due to board-specific hardware constraints particularly related
>>>>> to level shifter in this case the maximum frequency for SD High-Speed
>>>>> (HS) mode must be limited to 37.5 MHz to ensure reliable operation of SD
>>>>> card in HS mode.
>>>>>
>>>>> This is achieved by introducing the `max-sd-hs-hz` property in the
>>>>> device tree, allowing the controller to operate within safe frequency
>>>>> limits for HS mode.
>>>>>
>>>>
>>>> Probably we will now replicate the same discussion... And it will be
>>>> happening every time you send the same and not reflect it in commit msg.
> 
> Just to emphasize this - it will happen EVERY time.
> 
>>>>
>>>> Bindings say board setup, this commit msg says board config, but the
>>>> patch says SoC. This is not correct.
>>>
>>> Both are correct, looking at the problem from two perspectives.
>>>
>>> The bindings description mentions board-specific limitations (e.g. because
>>> "the board's electrical design does not allow one to achieve the full rated
>>> frequency that the SoC can otherwise do, in a stable way")
>>>
>>> Here the author tries to argue that almost all SM8550 boards are broken
>>> in this sense, because the reference design did not feature the required
>>> passive components, making most (derivative) designs sort of "broken by
>>> default" - and only some (if any?) vendors decided to go with the
>>> additional components required to lift this limitation.
>>>
>>> This in turn makes it fair to assume the developer experience would benefit
>>> from having the SD card high speed modes always work (with the slight speed
>>> cap which may not be required for the 1 or 2 designs that took the extra
>>> step) without each board DT creator having to track down this property
>>> separately.
>>
>> And then if you send same v3, I will ask the same. Can the author
> 
> v3 -> v6

So, would you be accepting of this patch if the commit message was:

arm64: dts: qcom: sm8550: Limit max SD HS mode frequency by default

Due to an implementation detail in this SoC, additional passive
electrical components are required to achieve the maximum rated speed
of the SD controller when paired with a High-Speed SD Card. Without them,
the clock frequency must be limited to 37.5 MHz for link stability.

Because the reference design does not contain these components, most
(derivative) boards do not have them either. To accommodate for that,
apply the frequency limit by default and delegate lifting it to the
odd boards that do contain the necessary onboard hardware.

Konrad