On Thu, May 16, 2019 at 09:55:33AM -0700, Guenter Roeck wrote:
> On 5/16/19 8:07 AM, Linus Torvalds wrote:
> >On Wed, May 15, 2019 at 9:43 PM Eduardo Valentin <edubezval@xxxxxxxxx> wrote:
> >>
> >>- thermal core has a new devm_* API for registering cooling devices, thanks to Guenter R.
> >> I took the entire series, that is why you see changes on drivers/hwmon in this pull.
> >
> >This clashed badly with commit 6b1ec4789fb1 ("hwmon: (pwm-fan) Add RPM
> >support via external interrupt"), which added a timer to the pwm-fan
> >handling.
> >
> >In particular, that timer now needed the same kind of cleanup changes,
> >and I'd like you guys (particularly Guenther, who was involved on both
> >sides) to double-check my merge.
> >
> >The way I solved it was to just make the pwm_fan_pwm_disable()
> >callback do both the pwm_diable() _and_ the del_timer_sync() on the
> >new timer. That seemed to be the simplest solution that meshed with
> >the new devm cleanup model, but while I build-tested the result, I
> >obviously did no actual use testing. And maybe there's some reason why
> >that approach is flawed.
> >
> >Guenther?
>
> Sorry for the trouble. Looks like I did too much cleanup this time around.
>
> Looks ok. I'll have to send a follow-up patch - we should check the
> return value of devm_add_action_or_reset(). No idea why I didn't do that
> in this series. I'll do that after the commit window closes (and after
> I am back from vacation).
OK... From what I could tell, looked fine from a thermal perspective.
>
> Thanks a lot for sorting this out.
>
> Guenter
