Re: [PATCH v3 7/9] dt-bindings: arm: stm32: Switch st,stm32mp23/25-syscfg into simple-mfd

From: Marek Vasut <marex@nabladev.com>
To: Krzysztof Kozlowski <krzk@kernel.org>
Cc: linux-usb@vger.kernel.org, fabrice.gasnier@foss.st.com,
	Alexandre Torgue <alexandre.torgue@foss.st.com>,
	Christian Bruel <christian.bruel@foss.st.com>,
	Conor Dooley <conor+dt@kernel.org>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Krzysztof Kozlowski <krzk+dt@kernel.org>,
	Maxime Coquelin <mcoquelin.stm32@gmail.com>,
	Neil Armstrong <neil.armstrong@linaro.org>,
	Pankaj Dev <pankaj.dev@st.com>,
	Rahul Kumar <rahul.kumar05@st.com>, Rob Herring <robh@kernel.org>,
	Rosen Penev <rosenp@gmail.com>,
	Thinh Nguyen <Thinh.Nguyen@synopsys.com>,
	Vinod Koul <vkoul@kernel.org>,
	devicetree@vger.kernel.org, kernel@dh-electronics.com,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, linux-phy@lists.infradead.org,
	linux-stm32@st-md-mailman.stormreply.com
Subject: Re: [PATCH v3 7/9] dt-bindings: arm: stm32: Switch st,stm32mp23/25-syscfg into simple-mfd
Date: Thu, 27 Aug 2026 17:04:40 +0200	[thread overview]
Message-ID: <9b197208-dbb4-44dd-b0c0-b4dcaa57c403@nabladev.com> (raw)
In-Reply-To: <20260827-skilled-optimal-shellfish-c548e5@quoll>

On 8/27/26 3:07 PM, Krzysztof Kozlowski wrote:
> On Sat, Aug 22, 2026 at 09:46:40AM +0200, Marek Vasut wrote:
>> The st,stm32mp23-syscfg and st,stm32mp25-syscfg can have subnodes, which
>> represent the USB 2.0 FEMTO-PHY. Turn the syscfg into "simple-mfd" so the
>> PHYs would get populated by the OS.
> 
> Sashiko comment is partially valid - you change the ABI in incompatible
> way, by requiring a new compatible. Commit msg should briefly explain
> the impact, e.g. that there is no impact because phy was not
> implemented by Linux and was not described in the binding.
> 
> OTOH, adding simple-mfd means children have no dependency on the parent,
> so this should be considered or even explicitly mentioned in commit msg
> (parent clock is irrelevant to child, right?)
> 
> OTOH2, you have now patchset dependency which is nowwhere explained.
> Merging constraints are the most important part of cover letter so must
> be clearly stated. IMO, should be stated also here, because many
> maintainers do not read cover letters. Or just decouple dependencies
> with using 'compatibles' pattern (see Qualcomm msm/mdss display
> bindings).
I wonder, wouldn't it be better to write a driver which binds to 
st,stm32mp25-syscfg , handles the SYSCFG clock, and does some 
of_platform_populate() to bring up the PHYs (DT subnodes) ?

I think that would resolve the ABI break problem, since we wouldn't need 
the simple-mfd compatible anymore. And it would also correctly handle 
the SYSCFG clock.

What do you think ?

-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy