Re: [PATCH v2 1/4] dt-bindings: PCI: rcar-gen4-pci-host: Document optional aux clock

[Marek Vasut <marek.vasut@mailbox.org>]

On 4/23/25 11:38 AM, Geert Uytterhoeven wrote:
> Hi Marek,

Hi,

> On Sun, 13 Apr 2025 at 11:29, Marek Vasut <marek.vasut@mailbox.org> wrote:
>> On 4/10/25 10:48 PM, Rob Herring wrote:
>>> On Sun, Apr 06, 2025 at 04:45:21PM +0200, Marek Vasut wrote:
>>>> Document 'aux' clock which are used to supply the PCIe bus. This
>>>> is useful in case of a hardware setup, where the PCIe controller
>>>> input clock and the PCIe bus clock are supplied from the same
>>>> clock synthesiser, but from different differential clock outputs:
>>>>
>>>>    ____________                    _____________
>>>> | R-Car PCIe |                  | PCIe device |
>>>> |            |                  |             |
>>>> |    PCIe RX<|==================|>PCIe TX     |
>>>> |    PCIe TX<|==================|>PCIe RX     |
>>>> |            |                  |             |
>>>> |   PCIe CLK<|======..  ..======|>PCIe CLK    |
>>>> '------------'      ||  ||      '-------------'
>>>>                       ||  ||
>>>>    ____________       ||  ||
>>>> |  9FGV0441  |      ||  ||
>>>> |            |      ||  ||
>>>> |   CLK DIF0<|======''  ||
>>>> |   CLK DIF1<|==========''
>>>> |   CLK DIF2<|
>>>> |   CLK DIF3<|
>>>> '------------'
>>>>
>>>> The clock are named 'aux' because those are one of the clock listed in
>>>> Documentation/devicetree/bindings/pci/snps,dw-pcie-common.yaml which
>>>> fit closest to the PCIe bus clock. According to that binding document,
>>>> the 'aux' clock describe clock which supply the PMC domain, which is
>>>> likely PCIe Mezzanine Card domain.
>>>
>>> Pretty sure that PMC is "power management controller" given it talks
>>> about low power states.
>>>
>>>
>>>>
>>>> Tested-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
>>>> Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
>>>> ---
>>>> NOTE: Shall we patch Documentation/devicetree/bindings/pci/snps,dw-pcie-common.yaml
>>>>         instead and add 'bus' clock outright ?
>>>
>>> Based on the diagram, this has nothing to do with the specific
>>> controller. It should also probably a root port property, not host
>>> bridge.
>> How would you suggest I describe the clock which supply the PCIe bus
>> clock lane (CLK DIF1 in the diagram) , which have to be enabled together
>> with clock which supply the PCIe controller input clock lane (CLK DIF0) ?
> 
> I think Rob wants you to add clocks/clock-names for this to
> dtschema/schemas/pci/pci-bus-common.yaml.  Then you can have pcie@M,N
> subnode(s) with num-lanes, clock, and clock-names describing the PCIe
> endpoint(s)?
> 
> git grep "pcie*@[0-9],[0-9]" -- $(git grep -l num-lanes -- Documentation/ )
> 
> Does that make sense?

No, not really. There can be any arbitrary PCIe card plugged into the 
M.2 slot, so how can I predict what exactly will be plugged into the 
slot and describe it in DT up front this way ?

-- 
Best regards,
Marek Vasut