Re: [PATCH] pwm: don't allow duty cycle higher than period

From: Boris Brezillon
Date: Fri May 27 2016 - 12:40:27 EST


On Fri, 27 May 2016 18:38:14 +0200
Boris Brezillon <boris.brezillon@xxxxxxxxxxxxxxxxxx> wrote:

> On Fri, 27 May 2016 09:35:33 -0700
> Brian Norris <briannorris@xxxxxxxxxxxx> wrote:
>
> > Hi Boris,
> >
> > On Fri, May 27, 2016 at 09:34:39AM +0200, Boris Brezillon wrote:
> > > On Thu, 26 May 2016 14:05:30 -0700
> > > Brian Norris <briannorris@xxxxxxxxxxxx> wrote:
> > >
> > > > It doesn't make sense to allow the duty cycle to be larger than the
> > > > period. I can see this behavior by, e.g.:
> > > >
> > > > # echo 1 > /sys/class/pwm/pwmchip0/export
> > > > # cat /sys/class/pwm/pwmchip0/pwm1/period
> > > > 100
> > > > # echo 101 > /sys/class/pwm/pwmchip0/pwm1/duty_cycle
> > > > [... driver may or may not reject the value, or trigger some logic bug ...]
> > > >
> > > > It's better to see:
> > > >
> > > > # echo 1 > /sys/class/pwm/pwmchip0/export
> > > > # cat /sys/class/pwm/pwmchip0/pwm1/period
> > > > 100
> > > > # echo 101 > /sys/class/pwm/pwmchip0/pwm1/duty_cycle
> > > > -bash: echo: write error: Invalid argument
> > > >
> > > > Signed-off-by: Brian Norris <briannorris@xxxxxxxxxxxx>
> > > > ---
> > > > drivers/pwm/core.c | 3 +++
> > > > 1 file changed, 3 insertions(+)
> > > >
> > > > diff --git a/drivers/pwm/core.c b/drivers/pwm/core.c
> > > > index dba3843c53b8..9246b60f894a 100644
> > > > --- a/drivers/pwm/core.c
> > > > +++ b/drivers/pwm/core.c
> > > > @@ -463,6 +463,9 @@ int pwm_apply_state(struct pwm_device *pwm, struct pwm_state *state)
> > > > if (!memcmp(state, &pwm->state, sizeof(*state)))
> > > > return 0;
> > > >
> > > > + if (state->duty_cycle > state->period)
> > > > + return -EINVAL;
> > > > +
> > >
> > > Argh, I forgot to move the pwm_config() checks [1] into
> > > pwm_apply_state() :-/.
> >
> > Oh, I didn't actually notice this was a regression.
> >
> > > I think we should check all the corner cases (see this diff [2]),
> >
> > Now that you mention it, I think you've also dropped some signed
> > (negative value) checking in pwm_config(). I'll squash in your diff +
> > some pwm_config() fixes.
>
> ->period and ->duty_cycle are unsigned now ;).
>

Never mind, you're right (thought you were talking about checking the
->period and ->duty_cycle values).

--
Boris Brezillon, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com