Re: [PATCH v12 1/1] serial: core: Start managing serial controllers to enable runtime PM

[AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>]

Il 07/06/23 06:46, Chen-Yu Tsai ha scritto:
> On Tue, Jun 6, 2023 at 8:21 PM Tony Lindgren <tony@atomide.com> wrote:
>>
>> * Chen-Yu Tsai <wenst@chromium.org> [230606 09:17]:
>>> I ended up following 8250_dw's design, which seemed less convoluted.
>>> The original code was waaay too convoluted.
>>
>> OK that looks good to me thanks. Good to hear you got it sorted out.
>>
>> The 8250_dw style runtime PM is a good solution for simple cases. Where
>> it won't work are SoCs where runtime PM calls need to propagate up the
>> bus hierarchy. For example, 8250_omap needs runtime PM calls for the
>> interconnect and power domain to get register access working.
> 
> Good to know. On MediaTek platforms I don't think there are any power
> domains covering the basic peripherals. (Or it's hidden from the kernel.)
> 

On (relatively) new SoCs, basic peripherals are always powered, you're correct.

Cheers,
Angelo

>>> BTW, the Bluetooth breakage seems like a different problem.
>>
>> OK seems like we're good to go then :)
> 
> Yup. After a bit more testing, it seems the Bluetooth problem is more like
> an undervolt issue. If I have WiFi and BT probe at the same time, Bluetooth
> fails. If they probe separately, everything works fine.
> 
> ChenYu