Re: [PATCH PATCH] arm64: dts: sunxi: nanopi-neo-plus2: Add pio regulators

[Kryštof Černý]

Yes, you are right with vcc-pd, I misunderstood it, thank you. Let me 
explain and ask about the rest:
VDD_SYS_3.3V = reg_vcc3v3
Groups PA, PE, PF are powered from vcc-io, according to Allwinner H5 
datasheet, Vcc-io is connected to VDD_SYS_3.3V, just like mmc0. Should 
those be set or omitted?
vcc-pc (ball G15), which is labeled as vcc_io2 (probably wrong), which 
is also connected to VDD_SYS_3.3V.
vcc-pd (ball J15) is bonded to VDD_PHY_2.5V, which is always on and made 
from VDD_SYS_3.3V, so I should make a new fixed regulator of name 
"reg_gmac_2v5"? Mainline eth driver does not implement this pin, so it 
would be used only for the pio.
vcc-pg (ball H7) is also VDD_SYS_3.3V.
While PL is supplied from vcc-rtc (vcc_rtc, ball K6), it is connected 
directly to the VDD_SYS_3.3V too, should this be the same regulator or a 
new one or omitted too?
Do you agree with my findings? I hope they are 100% now. If so, I will 
try to make V2 with a new fixed regulator of 2.5V for the PD.

Many thanks for your replies, I will do better next time.

Dne 24. 08. 24 v 14:34 Chen-Yu Tsai napsal(a):
> Hi,
>
> On Sat, Aug 24, 2024 at 5:08 PM Kryštof Černý <cleverline1mc@xxxxxxxxx> wrote:
> > I am sorry if the message is wrong, this is my first patch ever sent to
> > the Linux kernel. I have checked the schematic of the board and it
> > shares the same power line with mmc0, so I assumed I can use the same
> > regulator. Thanks for your feedback and I would be glad for your further
> > response.
> So some of the pin groups do have dedicated supplies, and should thus be
> described, but not all of them. The schematic only shows dedicated
> supplies for PD and PG pingroups. So just add those. PD supply is from
> 2.5V ethernet PHY I/O regulator supply, so you would need to add that
> as well.
>
> The datasheet also mentions a separate supply pin for PC pingroup, but
> it is not shown in the schematic. I would just omit that.
>
> And as Krzysztof mentioned, device tree changes should be to model
> the hardware, not to work around some operating system warning. At
> least most of the time that is. So your commit message should also
> be about fixing the description or providing more detail, and not
> about the operating system.
>
>
> ChenYu
>
> > Dne 24. 08. 24 v 9:40 Krzysztof Kozlowski napsal(a):
> > > On 24/08/2024 09:09, Kryštof Černý wrote:
> > > > The board does not have a dedicated regulator for pio and r_pio,
> > > > but this fixes the kernel warning about dummy regulators being used.
> > > > Tested on the actual board.
> > > Judging by commit msg these are not correct regulators. Please do not
> > > add incorrect hardware description to silence some warnings coming from
> > > OS. Either you need proper (correct) hardware description or fix the
> > > problem other way, assuming there is anything to fix in the first place.
> > >
> > > Best regards,
> > > Krzysztof