Hi Biju,
On Mon, Jan 10, 2022 at 11:51 AM Biju Das <biju.das.jz@xxxxxxxxxxxxxx> wrote:
> > Subject: Re: [PATCH v3 2/4] watchdog: rzg2l_wdt: Use force reset for WDT
> > reset
> > On Tue, Jan 4, 2022 at 7:12 PM Biju Das <biju.das.jz@xxxxxxxxxxxxxx>
> > wrote:
> > > This patch uses the force reset(WDTRSTB) for triggering WDT reset for
> > > restart callback. This method is faster compared to the overflow
> > > method for triggering watchdog reset.
> > >
> > > Signed-off-by: Biju Das <biju.das.jz@xxxxxxxxxxxxxx>
> > > --- a/drivers/watchdog/rzg2l_wdt.c
> > > +++ b/drivers/watchdog/rzg2l_wdt.c
> > > #define WDT_DEFAULT_TIMEOUT 60U
> > >
> > > @@ -116,15 +119,11 @@ static int rzg2l_wdt_restart(struct
> > > watchdog_device *wdev, {
> > > struct rzg2l_wdt_priv *priv = watchdog_get_drvdata(wdev);
> > >
> > > - /* Reset the module before we modify any register */
> > > - reset_control_reset(priv->rstc);
> > > - pm_runtime_get_sync(wdev->parent);
> >
> > Why are these no longer needed? Because .probe() takes care of that?
>
> This code is added to modify WDTSET register.
> Once watchdog is started, On the fly, we won't be able to
> update WDTSET register. By resetting(assert/deassert) the module
> we can update the WDTSET register. It takes 34 millisec to trigger reset.
>
> But with PEEN_FORCE, on the fly we can update register and it immediately triggers reset.
>
> Then why do .start() and .stop() have
> > reset and Runtime PM handling, too?
>
> .start-> turns on the Clocks using Runtime PM.
>
> We cannot Update WDTSET/WDTEN registers, once watchdog is started.
> Adding reset and Runtime PM handling in .stop, allows to stop the watchdog.
>
> .stop-> turns off the clock using Runtime PM and does the reset(assert/deassert) of the module
> in order to modify WDTSET/WDTEN registers after stop operation.
>
> May be I should keep pm_runtime_get_sync(wdev->parent) in restart callback,
> To turn on the clocks, If someone calls stop followed by restart
I'm still confused: .probe() turns the clock on through Runtime PM,
so it's always running. Calling .start() merely increases the usage
count, and a subsequent .stop() will decrease it again. But the
clock keeps on running? What am I missing?
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
