Re: [PATCH v5 0/2] PCI: mediatek-gen3: Add 2-lanes mode support + clock

[Benjamin Larsson <benjamin.larsson@genexis.eu>]

On 09/08/2026 16:14, Christian Marangi (Ansuel) wrote:
> Il giorno dom 9 ago 2026 alle ore 16:08 Benjamin Larsson
> <benjamin.larsson@genexis.eu> ha scritto:
>>
>> On 09/08/2026 14:12, Christian Marangi (Ansuel) wrote:
>>> Il giorno ven 7 ago 2026 alle ore 17:21 Bjorn Helgaas
>>> <helgaas@kernel.org> ha scritto:
>>>>
>>>> On Fri, Aug 07, 2026 at 11:24:09AM +0800, Chen-Yu Tsai wrote:
>>>>> On Fri, Aug 7, 2026 at 12:53 AM Christian Marangi <ansuelsmth@gmail.com> wrote:
>>>>>> This small series introduce support for 2-lanes mode for Airoha AN7581
>>>>>
>>>>> Just a nitpick, but I would probably name this something else, like
>>>>> "cross-controller lane-bonding mode" or "lane stealing"? PCIe already
>>>>> has 2x lanes as a standard feature, so this naming is a bit confusing.
>>>>> It's not like dual-LVDS display in which LVDS is only a single lane.
>>>>
>>>> I suggested the "2-lane" and "x2" terminology because I assumed the
>>>> result is what the PCIe spec would describe as a "x2 Link" consisting
>>>> of two Lanes.
>>>>
>>>> If that's not the case, maybe "cross-controller lane-bonding mode" or
>>>> "lane stealing" would be more accurate, but I don't know what those
>>>> mean, so if we use them I would also like to know what the result
>>>> looks like in standard PCIe terms.
>>>>
>>>
>>> Mhhh I don't really like the term lane stealing. Also cross-controller
>>> lane-bonding
>>> might be a first. Even if correct it would complicate identification
>>> of the feature
>>> that at the end of the day configures the HW to provide a 2 lanes PCIe.
>>>
>>> Consider that in such mode, the other PCIe controller gets disabled (this is
>>> handled in DT) so it's effectively enabling the standard PCIe 2x lanes
>>> and apply the HW configuration for it.
>>>
>>>>>> SoC. This is needed for correctly functionality of Eagle WiFi Card
>>>>>> normally attached to this SoC that require a 2-line PCIe card to
>>>>>> correctly work (and give the proper performance)
>>>>>>
>>>>>> The first 2 patch address a limitation of the PCIe implementation
>>>>>> where the PERSTOUT reset were indirectly asserted and deasserted
>>>>>> all at the same time (for all the 3 PCIe card) with PCIe
>>>>>> enable and disable.
>>>>>> The 2 patch address this and introduce correct reset to control
>>>>>> reset line for the relevant PCIe line.
>>>>>>
>>>>>> The last 2 patch add additional logic and support to assert
>>>>>> and deassert the PERSTOUT and also apply the required configuration
>>>>>> for 2-lanes mode.
>>>>>>
>>>>>> 2-lanes mode is implemented in DT by adding the required property
>>>>>> and by defining the "num-lanes" to 2.
>>>
>>
>> Hi. Isnt this just pcie bifurcation?
>>
>> Logically bifurcation would need to be disabled when using 2 lanes in
>> one slot and enabled when 2 lanes are split between 2 pcie slots.
>>
>> But I'm not really sure what value a bifurcation property would add.
>> Isnt the current schema enough?
>>
>> The mt7987 will need the same logic as it can also bifurcate one pcie slot.
>>
> 
> The documentation is not so kind on these kind of details... and no register
> for bifurcation... maybe it's the mux one? But the mux settings comes from
> reverse as in documentation that SCU register is a good 7:0 bits of
> data value...

Hi, yeah I think the SCU documentation is wrong for the AN7581. At least 
for the W1700k v1 device SoC version. It works without the SCU PCIC bit 
set there IIRC.

https://forum.openwrt.org/t/quantum-fiber-w1700k-support/222776/280

The vendor driver has this:
https://github.com/merbanan/airoha_pcie/blob/main/airoha_pcie/en7581/pcie-ecnt-phy_7581.c#L1671

Maybe this is the only bit that control the bifurcation?

MvH
Benjamin Larsson