[PATCH] riscv: dts: sifive: unmatched: Add PWM controlled fans
Conor Dooley
conor at kernel.org
Sun Nov 23 11:59:25 PST 2025
On Sun, Nov 23, 2025 at 08:16:52PM +0100, René Rebe wrote:
> Hi,
>
> On Sun, 23 Nov 2025 19:00:21 +0000, Conor Dooley <conor at kernel.org> wrote:
>
> > > ---
> > > Tested on SiFive HiFive Unmatched A00 running T2/Linux.
> >
> > Was this generated on top of a tree that includes 7dbc4432ea6b ("pwm:
> > sifive: Fix PWM algorithm and clarify inverted compare behavior")?
> > That'd be 6.17-rc1 or later.
>
> Yes, linus' 6.18-rc... tip
Cool, I'll pick this up then.
>
> > If not, the polarity of the pwms will need to be switched because the
> > original driver author intentionally got it wrong and so 6.17-rc1 and
> > later have the correct polarity...
>
> Indeed, I wondered about this and just inverted it yesterday to
> upstream. I had this pwm fans patched for a year already and likewise
> inverted them before.
I don't recall the exact details, but I looked into it when the first
version of the fix appeared. The submitter effectively decided, in
response to review feedback about the pwm duty selection algorithm being
incorrect in their code, to invert it, presumably because that was easier
than actually implementing the feedback. Hilariously, they claim "this isn't
a bug in the driver code" when someone commented on this back in 2019:
https://lore.kernel.org/linux-pwm/CAJ2_jOG2M03aLBgUOgGjWH9CUxq2aTG97eSX70=UaSbGCMMF_g@mail.gmail.com/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 228 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-riscv/attachments/20251123/a81e67c3/attachment.sig>
More information about the linux-riscv
mailing list