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

[Guenter Roeck <linux@roeck-us.net>]

On Mon, Sep 28, 2026 at 02:03:57PM +0200, Antheas Kapenekakis wrote:
> On Mon, 28 Sept 2026 at 13:05, Guenter Roeck <linux@roeck-us.net> wrote:
> >
> > On Sat, Sep 26, 2026 at 04:11:21PM -0400, Alexei Solcanu via B4 Relay wrote:
> > > From: Alexei Solcanu <alexei.solcanu@pm.me>
> > >
> > > The GPD Win 5 uses gpd_duo_drvdata, which exposes a single fan and pwm
> > > channel: pwm1 writes the same duty cycle to both fans, and fan1_input
> > > only reads the first fan's RPM.
> > >
> > > The two fans have independent registers. Fan 1 uses duty cycle 0x047A
> > > and RPM 0x0478/0x0479, fan 2 uses duty cycle 0x047B and RPM
> > > 0x0476/0x0477. A duty cycle of 0 returns that fan to EC control.
> > >
> > > Add separate board data for the Win 5 and register fan1/pwm1 and
> > > fan2/pwm2, one channel per fan. Boards with a single fan are unchanged.
> > >
> > > This changes behaviour on the Win 5: pwm1 and pwm1_enable now control
> > > only the first fan. The second fan stays under EC control unless pwm2
> > > and pwm2_enable are used.
> > >
> > > The GPD Duo keeps a single channel, since it is untested and its
> > > second fan's RPM register is unknown.
> > >
> > > Tested on a GPD Win 5 (G1618-05).
> > >
> > > Suggested-by: Guenter Roeck <linux@roeck-us.net>
> > > Assisted-by: LLM
> > > Signed-off-by: Alexei Solcanu <alexei.solcanu@pm.me>
> >
> > Another patch new revision sent as reply of a previous one.
> >
> > Where does that idea come from ?  Can _someone_ please let me know ?
> > It seems like a contagious disease, but no one is willing to admit
> > where they caught it.
> 
> Maybe the following?
> https://en.andros.dev/blog/75beece9/sending-patches-via-email-with-git/
> https://nickdesaulniers.github.io/blog/2017/05/16/submitting-your-first-patch-to-the-linux-kernel-and-responding-to-feedback/
> 

Sigh. That "keeping together" does cause me to miss new patch revisions
on a regular basis. Oh well. At some point I may just create an
auto-responder telling people that I am not responding to nested
revisions, and copy the authors of those articles.

> I faintly recall something like that when i was sending my first patch
> a few years ago.
> 
> I checked the kernel docs and they are correct now and explicitly say
> do not do this. Perhaps that was a recent change though.

No, that was not a recent change, unless you count commit d7ac8d85d39f4
("Documentation/SubmittingPatches: discuss In-Reply-To") from 2015 as a
"recent change".

Guenter