On Fri, Jul 14, 2017 at 09:54:23AM -0700, Moritz Fischer wrote:
> Hi Guenter,
>
> On Thu, Jul 13, 2017 at 08:57:52PM -0700, Guenter Roeck wrote:
> > On 07/13/2017 12:54 PM, Moritz Fischer wrote:
> > > From: Moritz Fischer <moritz.fischer@xxxxxxxxx>
> > >
> > > Add support for the Maxim/Dallas DS1374 RTC/WDT with trickle charger.
> > > The device can either be configured as simple RTC, as simple RTC with
> > > Alarm (IRQ) as well as simple RTC with watchdog timer.
> > >
> > > Break up the old monolithic driver in drivers/rtc/rtc-ds1374.c into:
> > > - rtc part in drivers/rtc/rtc-ds1374.c
> > > - watchdog part under drivers/watchdog/ds1374-wdt.c
> > > - mfd part drivers/mfd/ds1374.c
> > >
> > > The MFD part takes care of trickle charging and mode selection,
> > > since the usage modes of a) RTC + Alarm or b) RTC + WDT
> > > are mutually exclusive.
> > >
> > > Signed-off-by: Moritz Fischer <mdf@xxxxxxxxxx>
> >
> > [ Only reviewing watchdog part ]
> >
[ ... ]
> > > +}
> > > +
> > > +static int ds1374_wdt_start(struct watchdog_device *wdog)
> > > +{
> > > + int err;
> > > + struct ds1374_wdt *ds1374_wdt = watchdog_get_drvdata(wdog);
> > > +
> > > + err = ds1374_wdt_set_timeout(wdog, wdog->timeout);
> > > + if (err) {
> > > + dev_err(ds1374_wdt->dev, "%s: failed to set timeout (%d) %u\n",
> > > + __func__, err, wdog->timeout);
> > > + return err;
> > > + }
> > > +
> > > + err = ds1374_wdt_ping(wdog);
> > > + if (err) {
> > > + dev_err(ds1374_wdt->dev, "%s: failed to ping (%d)\n", __func__,
> > > + err);
>
> I assume you'd want to get rid of that one too?
Yes.
> >
> > This may bypass MAGICCLOSE: If the watchdog daemon is killed and the module removed,
> > the watchdog will be stopped. Is this really what you want ? If so, why set MAGICCLOSE
> > in the first place ?
>
> So your suggestion would be:
>
> - if (!nowayout)
> - ds1374_wdt_stop(&priv->wdd)
>
Correct. Just drop the remove function.
Thanks,
Guenter
