Re: [PATCH v6 4/9] iio: adc: mcp3911: use resource-managed version of iio_device_register

From: Jonathan Cameron
Date: Mon Sep 19 2022 - 11:36:12 EST


On Mon, 12 Sep 2022 13:02:19 +0200
Marcus Folkesson <marcus.folkesson@xxxxxxxxx> wrote:

> Hi,
>
> On Sat, Aug 20, 2022 at 01:41:50PM +0100, Jonathan Cameron wrote:
> > On Mon, 15 Aug 2022 08:16:20 +0200
> > Marcus Folkesson <marcus.folkesson@xxxxxxxxx> wrote:
> >
> > > Keep using managed resources as much as possible.
> > >
> > > Signed-off-by: Marcus Folkesson <marcus.folkesson@xxxxxxxxx>
> > > Reviewed-by: Andy Shevchenko <andy.shevchenko@xxxxxxxxx>
> > > ---
> > > drivers/iio/adc/mcp3911.c | 53 ++++++++++++---------------------------
> > > 1 file changed, 16 insertions(+), 37 deletions(-)
> > >
> > > diff --git a/drivers/iio/adc/mcp3911.c b/drivers/iio/adc/mcp3911.c
> > > index 890af7dca62d..7e2efe702e57 100644
> > > --- a/drivers/iio/adc/mcp3911.c
> > > +++ b/drivers/iio/adc/mcp3911.c
> > > @@ -258,6 +258,13 @@ static int mcp3911_config(struct mcp3911 *adc)
> > > return mcp3911_write(adc, MCP3911_REG_CONFIG, configreg, 2);
> > > }
> > >
> > > +static void mcp3911_cleanup_regulator(void *_adc)
> >
> > Missed this on previous versions, but why not pass
> > the regulator pointer in as the parameter for the callback?
> >
> > static void mcp391_cleanup_regulator(void *reg)
> > {
> > regulator_disable(adc->vref);
> > }
> >
> > Note this can't use the new devm_regulator_get_enable()
> > because we need access to the regulator within the driver.
> >
> > I can tidy this up whilst applying (or given it's really minor I might
> > not bother :)
> >
> > Note we are stalled at the moment with this series on getting the
> > fixes upstream. I'll probably send that pull request shortly.
>
> Just a friendly reminder to not forget to pick up this series.

If things go according to plan, Greg will take the pull request
I sent the other day and I can fast forward the tree such that
the first 3 patches are in my upstream, then do a second pull
request with these applied. They aim here being to keep the
history reasonably clean rather than spaghetti merges.
Fingers crossed on timing working out. This all got delayed
because I sat on the pull request for a week due to travel and
dodgy airport wifi etc.

Jonathan

>
> Thanks,
>
> /Marcus