Re: [PATCH 4/5] watchdog: ziirave_wdt: Don't bail out on unexpected timeout value

[Andrey Smirnov]

On Wed, Jul 31, 2019 at 11:09 AM Guenter Roeck <linux@xxxxxxxxxxxx> wrote:
>
> On Wed, Jul 31, 2019 at 10:42:51AM -0700, Andrey Smirnov wrote:
> > Reprogramming bootloader on watchdog MCU will result in reported
> > default timeout value of "0". That in turn will be unnecesarily
>
> unnecessarily
>
> > rejected by the driver as invalid device (-ENODEV). Simplify probe to
> > just read stored timeout value, clamp it to an acceptable range and
> > program the value unconditionally to fix the above.
> >
> > Signed-off-by: Andrey Smirnov <andrew.smirnov@xxxxxxxxx>
> > Cc: Chris Healy <cphealy@xxxxxxxxx>
> > Cc: Guenter Roeck <linux@xxxxxxxxxxxx>
> > Cc: Rick Ramstetter <rick@xxxxxxxxxxxxxxx>
> > Cc: linux-watchdog@xxxxxxxxxxxxxxx
> > Cc: linux-kernel@xxxxxxxxxxxxxxx
> > ---
> >  drivers/watchdog/ziirave_wdt.c | 22 +++++++++-------------
> >  1 file changed, 9 insertions(+), 13 deletions(-)
> >
> > diff --git a/drivers/watchdog/ziirave_wdt.c b/drivers/watchdog/ziirave_wdt.c
> > index 89ce6982ba53..33c8d2eadada 100644
> > --- a/drivers/watchdog/ziirave_wdt.c
> > +++ b/drivers/watchdog/ziirave_wdt.c
> > @@ -667,22 +667,18 @@ static int ziirave_wdt_probe(struct i2c_client *client,
> >                       return val;
> >               }
> >
> > -             if (val < ZIIRAVE_TIMEOUT_MIN)
> > -                     return -ENODEV;
> > -
> > -             w_priv->wdd.timeout = val;
> > -     } else {
> > -             ret = ziirave_wdt_set_timeout(&w_priv->wdd,
> > -                                           w_priv->wdd.timeout);
> > -             if (ret) {
> > -                     dev_err(&client->dev, "Failed to set timeout\n");
> > -                     return ret;
> > -             }
> > +             w_priv->wdd.timeout = clamp(val, ZIIRAVE_TIMEOUT_MIN,
> > +                                         ZIIRAVE_TIMEOUT_MAX);
>
> Are you sure ? Effectively that will set the timeout to the minimum,
> ie three seconds. It might be better to define and set some default.
> Your call, of course.
>

It doesn't really matter in my use-case (set timeout is a no-op),  but
it sounds like a better approach, so I'll change it in v2.

Thanks,
Andrey Smirnov