Hi Geert,
>
> Hi Fabrizio,
>
> On Thu, Nov 17, 2022 at 12:49 PM Fabrizio Castro
> <fabrizio.castro.jz@xxxxxxxxxxx> wrote:
> > As per section 48.4 of the HW User Manual, IPs in the RZ/V2M
> > SoC need either a TYPE-A reset sequence or a TYPE-B reset
> > sequence. More specifically, the watchdog IP needs a TYPE-B
> > reset sequence.
> >
> > If the proper reset sequence isn't implemented, then resetting
> > IPs may lead to undesired behaviour. In the restart callback of
> > the watchdog driver the reset has basically no effect on the
> > desired funcionality, as the register writes following the reset
> > happen before the IP manages to come out of reset.
> >
> > Implement the TYPE-B reset sequence in the watchdog driver to
> > address the issues with the restart callback on RZ/V2M.
> >
> > Fixes: ec122fd94eeb ("watchdog: rzg2l_wdt: Add rzv2m support")
> > Signed-off-by: Fabrizio Castro <fabrizio.castro.jz@xxxxxxxxxxx>
>
> Reviewed-by: Geert Uytterhoeven <geert+renesas@xxxxxxxxx>
>
> Perhaps this logic can be incorporated into the RZ/V2M reset controller
> driver later, so reset consumers don't have to care about TYPE-A and
> TYPE-B reset, but can just call reset_control_reset()?
> I understand that's not gonna be easy, as it needs to know about the
> relation between resets and clocks, and how to handle both cases (clock
> (not) switched off) for TYPE-B resets.
Yeah, we have been thinking about dealing with this in the reset controller
driver, but as you pointed out it's not going to be simple, and therefore
it'll take some time. This change will guarantee the correct behaviour of
the watchdog for now, we'll tackle the larger issue later on, if that's okay
with you.
Thanks,
Fab
>
> Gr{oetje,eeting}s,
>
> Geert
>
> --
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-
> m68k.org
>
> 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
