Re: [PATCH v2 2/2] hwmon: (gpd-fan) Register both fans on the GPD Win 5

[Alexei Solcanu]

On Sat, 26 Sept 2026 at 22:32, Antheas Kapenekakis <lkml@xxxxxxxxxxx> wrote:
> in my view adding all these functions is unneeded complexity. There
> should be some simpler ternary way.

Ok, I'll inline them as ternaries in v3.

> Are you sure the correct number of fans are populated for each device
> because of the visible entry? If yes, it is fine for this hunk. If no,
> needs dynamic population for the struct in probe.

Yes. The hwmon core only creates the attributes for which is_visible()
returns non-zero. On the Win 5 both channels show up and work
independently. Loading the driver on it with gpd_fan_board=win_mini
works as expected: only fan1_input, pwm1 and pwm1_enable are created.

> You also add a board match for win5, but doesnt the duo have two fans?
[...]
> Verify which duo-like devices implement two fans, it should be all of
> them. I think Cryolitia has a Duo. Win Max 3 according to my GPD
> contact has a duo layout. Unfortunately, I have no users so no DMI. So
> it is at least Duo, Win Max 3, and Win5. Then make it so only these 3
> devices (which should need no separation) report two fans and two fan
> speeds throughout the driver.

Agreed. In v3 I'll drop the separate win5 board and give the duo board
data two channels instead, so the Duo and the Win 5 share it.

I don't have a Win Max 3, so I can't add its DMI string, but I'll add
the duo board data to gpd_module_drvdata, so gpd_fan_board=duo can be
used on it until it has a DMI entry.

I only have a Win 5, so I can't verify the Duo. On the Win 5 the
second fan's RPM is at 0x0476/0x0477. Cryolitia, could you confirm
that the Duo uses the same register?