Re: [PATCH 0/7] ARM: dts: stm32: clk: Switch ETHRX clock parent from ETHCK_K to MCO2 on DHCOM SoM

From: Marek Vasut <marex@denx.de>
To: Alexandre TORGUE <alexandre.torgue@foss.st.com>,
	linux-arm-kernel@lists.infradead.org
Cc: Christophe Roullier <christophe.roullier@foss.st.com>,
	Gabriel Fernandez <gabriel.fernandez@foss.st.com>,
	Patrice Chotard <patrice.chotard@foss.st.com>,
	Patrick Delaunay <patrick.delaunay@foss.st.com>,
	Stephen Boyd <swboyd@chromium.org>,
	linux-clk@vger.kernel.org,
	linux-stm32@st-md-mailman.stormreply.com,
	Stephen Boyd <sboyd@kernel.org>
Subject: Re: [PATCH 0/7] ARM: dts: stm32: clk: Switch ETHRX clock parent from ETHCK_K to MCO2 on DHCOM SoM
Date: Mon, 12 Apr 2021 20:44:17 +0200	[thread overview]
Message-ID: <2c410c1b-2a1d-b291-a128-c5d5979be1ef@denx.de> (raw)
In-Reply-To: <b81d6a69-713f-eda7-0837-d6e80d691c6a@foss.st.com>

On 4/12/21 10:09 AM, Alexandre TORGUE wrote:
> Hi Marek

Hello Alex,

[...]

>> All the above still only discusses the clock part of the problem. Even if
>> the clock cyclic dependencies could be solved, it would be necessary to
>> resolve legacy dwmac st,eth-clk-sel and st,eth-ref-clk-sel DT properties
>> and avoid DT ABI break.
> 
> Thanks for those clear explanations and for this series. As discussed, 
> this approach looks good to me as it doesn't break our current strategy 
> for dwmac clock integration. I don't know if those cyclic redundancies 
> will be fixed one day but we can have a look on dwmac DT properties (the 
> gain to change them, the effort to keep the backward compatibility, code 
> readability, ...).
> 
> Your DT patches looks good. I'll merge them soon.
+CC Stephen ; the DT patches depend on the clock driver changes. Would 
it make sense to pick the clock patches through the same tree or how 
should that be handled ?

[...]