Re: [PATCH] iio: adc: xilinx-ams: Use device_for_each_child_node_scoped()

[Jonathan Cameron <jic23@kernel.org>]

On Sat, 27 Apr 2024 06:13:01 -0300
Marcelo Schmitt <marcelo.schmitt1@gmail.com> wrote:

> Hi Pedro, Roberto,
> 
> Patch looks overall good except for the _scoped() function name and arguments,
> must have been miss-typed or miss-copied somehow.
> Comment inline.
> 
> Regards,
> Marcelo
> 
> On 04/24, Pedro Mariano wrote:
> > Using device_for_each_child_node_scoped instead of
> > device_for_each_child_node automatically releases the handle on early exit
> > which reduces the chance of bugs that cause resource leaks.
> > 
> > Co-developed-by: Roberto Bolgheroni <robertobolgheroni@usp.br>
> > Signed-off-by: Roberto Bolgheroni <robertobolgheroni@usp.br>
> > Signed-off-by: Pedro Mariano <pedro.mariano@usp.br>
> > ---
> >  drivers/iio/adc/xilinx-ams.c | 7 ++-----
> >  1 file changed, 2 insertions(+), 5 deletions(-)
> > 
> > diff --git a/drivers/iio/adc/xilinx-ams.c b/drivers/iio/adc/xilinx-ams.c
> > index f0b71a122..7f5571d9d 100644
> > --- a/drivers/iio/adc/xilinx-ams.c
> > +++ b/drivers/iio/adc/xilinx-ams.c
> > @@ -1261,7 +1261,6 @@ static int ams_parse_firmware(struct iio_dev *indio_dev)
> >  	struct ams *ams = iio_priv(indio_dev);
> >  	struct iio_chan_spec *ams_channels, *dev_channels;
> >  	struct device *dev = indio_dev->dev.parent;
> > -	struct fwnode_handle *child = NULL;
> >  	struct fwnode_handle *fwnode = dev_fwnode(dev);
> >  	size_t ams_size;
> >  	int ret, ch_cnt = 0, i, rising_off, falling_off;
> > @@ -1283,13 +1282,11 @@ static int ams_parse_firmware(struct iio_dev *indio_dev)
> >  		num_channels += ret;
> >  	}
> >  
> > -	fwnode_for_each_child_node(fwnode, child) {
> > +	fwnode_for_each_child_node_scoped(fwnode, child) {  
> should be
> 	device_for_each_child_node_scoped(dev, child) {

Yes, we didn't bother with a fwnode specific version of this macro because they
aren't nearly as common.  I'm not sure why this driver didn't always use the
device form or why it needs the fwnode_device_is_available()
I suspect this dates back to some confusion on why there were _available variants.

Chances are this driver only cares about DT and in that case the callback used is

static struct fwnode_handle *
of_fwnode_get_next_child_node(const struct fwnode_handle *fwnode,
			      struct fwnode_handle *child)
{
	return of_fwnode_handle(of_get_next_available_child(to_of_node(fwnode),
							    to_of_node(child)));
}

So the fwnode_for_each_child_node() + fwnode_device_is_available() end up doing the
same as device_for_each_child_node().

So I think there are more opportunities to tidy up in here than simply this scoped
change and I'd be keen to see them all done together.

Jonathan

> 
> >  		if (fwnode_device_is_available(child)) {
> >  			ret = ams_init_module(indio_dev, child, ams_channels + num_channels);
> > -			if (ret < 0) {
> > -				fwnode_handle_put(child);
> > +			if (ret < 0)
> >  				return ret;
> > -			}
> >  
> >  			num_channels += ret;
> >  		}
> > -- 
> > 2.44.0
> > 
> >