Hi Álvaro,
On Wed, 2020-06-10 at 08:08 +0200, Álvaro Fernández Rojas wrote:
> Hi Florian,
>
> > El 9 jun 2020, a las 22:17, Florian Fainelli <f.fainelli@xxxxxxxxx> escribió:
> >
> >
> >
> > On 6/9/2020 9:41 AM, Álvaro Fernández Rojas wrote:
> > > > > > If you can do without this, with I think this driver could be made to
> > > > > > use reset-simple.
> > > > >
> > > > > Yes, but only if I can add reset support with a configurable sleep range to reset-simple. Is this possible?
> > > >
> > > > I should have mentioned, support for this is on the reset/next branch:
> > > >
> > > > git://git.pengutronix.de/pza/linux.git reset/next
> > >
> > > Yes, but reset_us was only added to reset_simple_data, so there’s no way to fill that value from reset_simple_devdata or device tree, right?
> >
> > Not that I can see, but you could certainly extend it here:
> >
> > if (devdata) {
> > reg_offset = devdata->reg_offset;
> > if (devdata->nr_resets)
> > data->rcdev.nr_resets = devdata->nr_resets;
> > data->active_low = devdata->active_low;
> > data->status_active_low = devdata->status_active_low;
> > }
>
> Yes, I would extend it there too,
You are right, reset_us is missing from reset_simple_devdata.
> but I was just saying that it’s a bit strange that it was only added
> to reset_simple_data without any way to fill the value.
The patch was added for the benefit of drivers that register their own
reset controller using reset_simple_data/ops, like sunxi or socfpga.
This might be considered an oversight, but until now there was no user
inside reset-simple.c.
regards
Philipp
