Re: [PATCH 1/2] iio: pressure: mpl115: add system suspend/resume support

[Jonathan Cameron]

On Sun, 22 Feb 2026 19:32:38 +0100
Paul Cercueil <paul@xxxxxxxxxxxxxxx> wrote:

> Hi,
> 
> Le dimanche 22 février 2026 à 16:40 +0000, Jonathan Cameron a écrit :
> > On Sat, 21 Feb 2026 12:26:00 -0500
> > Neel Bullywon <neelb2403@xxxxxxxxx> wrote:
> >   
> > > The driver currently only implements runtime PM callbacks to
> > > control
> > > the shutdown GPIO, but lacks system suspend/resume support. This
> > > means
> > > the device may remain powered during system sleep, wasting energy.
> > > 
> > > Add system sleep PM ops using pm_runtime_force_suspend() and
> > > pm_runtime_force_resume() to properly synchronize the runtime PM
> > > state
> > > across system suspend/resume cycles. This ensures the shutdown GPIO
> > > is
> > > asserted during system sleep when the device has a shutdown GPIO
> > > configured.
> > > 
> > > Signed-off-by: Neel Bullywon <neelb2403@xxxxxxxxx>  
> > 
> > Interestingly for the non exported version we have
> > DEFINE_RUNTIME_DEV_PM_OPS()
> > and the docs for that suggest we should use
> > EXPORT_RUNTIME_DEV_PM_OPS() but as you've realized that doesn't set
> > the
> > suspend and resume callbacks. If that was intentional then the
> > documentation
> > is less than clear (and I don't remember that being the intent)
> > 
> > +CC Paul and Rafael.
> > 
> > We have very few users of these macros, so it should be easy enough
> > to
> > make the export variants consistent if we want to.  
> 
> Looks like the users are all in IIO which makes it even more
> convenient.
> 
> It's been a while so my memory is fuzzy, but it would make sense to
> have the suspend/resume callbacks, indeed.

Ok. Neel, would you mind preparing a set to make 
EXPORT_NS_RUNTIME_DEV_PM_OPS() set the SYSTEM_SLEEP_PM_OPS()
as well?  If you can take a look at the existing callers and
shout if there any problems with this change, even better!

I foolishly always thought they were doing this so I 'think'
it'll be fine to make this change.  Also avoids the license
change thing Paul noted below.

Thanks,

Jonathan
> 
> > Jonathan
> >   
> > > ---
> > >  drivers/iio/pressure/mpl115.c | 6 ++++--
> > >  1 file changed, 4 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/drivers/iio/pressure/mpl115.c
> > > b/drivers/iio/pressure/mpl115.c
> > > index 830a5065c008..29794ce53bc6 100644
> > > --- a/drivers/iio/pressure/mpl115.c
> > > +++ b/drivers/iio/pressure/mpl115.c
> > > @@ -244,8 +244,10 @@ static int mpl115_runtime_resume(struct device
> > > *dev)
> > >  	return 0;
> > >  }
> > >  
> > > -EXPORT_NS_RUNTIME_DEV_PM_OPS(mpl115_dev_pm_ops,
> > > mpl115_runtime_suspend,
> > > -			  mpl115_runtime_resume, NULL,
> > > IIO_MPL115);
> > > +EXPORT_NS_GPL_DEV_PM_OPS(mpl115_dev_pm_ops, IIO_MPL115) = {
> > > +	SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
> > > pm_runtime_force_resume)
> > > +	RUNTIME_PM_OPS(mpl115_runtime_suspend,
> > > mpl115_runtime_resume, NULL)
> > > +};  
> 
> One thing I noticed is that now the symbols are exported as GPL, which
> is probably fine (better than the other way around) but if that was on
> purpose, it should at least be mentioned in the commit message.
> 
> Cheers,
> -Paul