Re: [PATCH 0/3] clk: Support spread spectrum and use it in clk-scmi

[Krzysztof Kozlowski <krzk@kernel.org>]

On 25/01/2025 13:58, Dario Binacchi wrote:
> On Fri, Jan 24, 2025 at 2:19 PM Peng Fan (OSS) <peng.fan@oss.nxp.com> wrote:
>>
>> - Introduce clk_set_spread_spectrum to set the parameters for enabling
>>   spread spectrum of a clock.
>> - Parse 'assigned-clock-sscs' and configure it by default before using the
>>   clock. The pull request for this property is at [1]
>>   This property is parsed before parsing clock rate.
>>
>> - Enable this feature for clk-scmi on i.MX95.
>>   This may not the best, since checking machine compatibles.
>>   I am thinking to provide an API scmi_get_vendor_info, then driver
>>   could use it for OEM stuff, such as
>>   if (scmi_get_vendor_info returns NXP_IMX)
>>       ops->set_spread_spectrum = scmi_clk_set_spread_spectrum_imx;
>>
> 
> I wonder if your solution is truly generic or merely a generalization
> of your use case, which seems significantly simpler compared to what

Please come with specific arguments why this is not generic enough, not
just FUD. Does it fit your case? If not, what would had to be changed?
These are the comments needed to actually work on generic solution.

> happens on the i.MX8M platform, as discussed in thread
> https://lore.kernel.org/lkml/PAXPR04MB8459537D7D2A49221D0E890D88E32@PAXPR04MB8459.eurprd04.prod.outlook.com/,
> or on the STM32F platform, where parameters are not written directly
> to registers but are instead used in calculations involving the
> parent_rate and the PLL divider, for example.
> 
> I am the author of the patches that introduced spread spectrum
> management for the AM33x and STM32Fx platforms, as well as the
> series, still pending acceptance, for the i.MX8M.
> From my perspective, this functionality varies significantly
> from platform to platform, with key differences that must be
> considered.

So what exactly varies? Come with specifics.

To me look addressing exactly the same.

Best regards,
Krzysztof