On 07/11/2022 23:36, Hawkins, Nick wrote: > > > > This provides the base registers address, programmable logic registers > > > address, and the function 2 registers to allow control access of the HPE > > > fans on the GXP SoC. > > > What is "This"? If "This patch", then drop it. > > https://elixir.bootlin.com/linux/v5.17.1/source/Documentation/process/submitting-patches.rst#L95 > > > If "This hardware" then please instead describe the hardware, not it > components. What are its features? If it controls the fan, then why > there are no PWM-related cells? How do you set the speed? > > Greetings Krzysztof, > > Thank you for the feedback. The intention was this binding.. however, that was an error on my part, and I will correct it to reflect the hardware situation of the GXP with the fan controller and how each of the mapped registers provide control to the system. To answer your questions: The fans speeds are controlled through an external CPLD device which we provide a PWM value (0-255) using the "base" register to the CIF interface. Wrap your emails, it's impossible to simply reply to it. Then your CIF interface is a PWM device? > This interface provides access to the CPLD. The CPLD then drives the fan. The CPLD can generate up to 8 unique different PWMs to multiple fans. So you have other CPLD (not external) which generates PWM based on first CPLD base register? Hm, I think it's one CPLD. > The CPLD monitors the fans and reports the status back to the SoC through the CIF interface to the "plreg base". The plreg includes the installation, failed, and identification statuses. The function 2 register base is used to check the power state of the system as that influences the PWM values read back. > As the PWM generation happens outside the SoC do we still need pwm-cells? If so, should we have a custom compatible for that? > Depends, if these are actually tightly coupled and you cannot use PWM for anything else, then you do not need. > Thanks, > > -Nick > > Best regards, Krzysztof
