Re: [PATCH v2 5/6] gpiolib: consolidate GPIO lookups

[Dmitry Torokhov]

On Wed, Nov 09, 2022 at 01:25:06PM +0200, Andy Shevchenko wrote:
> On Tue, Nov 08, 2022 at 04:26:50PM -0800, Dmitry Torokhov wrote:
> > Ensure that all paths to obtain/look up GPIOD from generic
> > consumer-visible APIs go through the new gpiod_find_and_request()
> > helper, so that we can easily extend it with support for new firmware
> > mechanisms.
> > 
> > The only exception is OF-specific [devm_]gpiod_get_from_of_node() API
> > that is still being used by a couple of drivers and will be removed as
> > soon as patches converting them to use generic fwnode/device APIs are
> > accepted.
> 
> ...
> 
> > +static struct gpio_desc *gpiod_find_by_fwnode(struct fwnode_handle *fwnode,
> > +					      struct device *consumer,
> > +					      const char *con_id,
> > +					      unsigned int idx,
> > +					      enum gpiod_flags *flags,
> > +					      unsigned long *lookupflags)
> >  {
> > -	unsigned long lflags = GPIO_LOOKUP_FLAGS_DEFAULT;
> 
> > -	struct gpio_desc *desc = ERR_PTR(-ENODEV);
> 
> Not sure why this is needed. Now I see that else branch has been changed,
> but looking closer to it, we can drop it completely, while leaving this
> line untouched, correct?

Yes. I believe removing an initializer and doing a series of if/else
if/else was discussed and [soft] agreed-on in the previous review cycle,
but I can change it back.

I think we still need to have it return -ENOENT and not -ENODEV/-EINVAL
so that we can fall back to GPIO lookup tables when dealing with an
unsupported node type.

> 
> > -	int ret;
> > +	struct gpio_desc *desc;
> >  
> > +	dev_dbg(consumer, "GPIO lookup for consumer %s in node '%pfw'\n",
> > +		con_id, fwnode);
> > +
> > +	/* Using device tree? */
> >  	if (is_of_node(fwnode)) {
> > -		desc = gpiod_get_from_of_node(to_of_node(fwnode),
> > -					      propname, index,
> > -					      dflags,
> > -					      label);
> > -		return desc;
> > +		dev_dbg(consumer, "using device tree for GPIO lookup\n");
> > +		desc = of_find_gpio(to_of_node(fwnode),
> > +				    con_id, idx, lookupflags);
> 
> At least con_id can be placed on the previous line.

OK, I made it all 1 line.

> 
> >  	} else if (is_acpi_node(fwnode)) {
> > -		desc = acpi_node_get_gpiod(fwnode, propname, index,
> > -					   &lflags, &dflags);
> > -		if (IS_ERR(desc))
> > -			return desc;
> > +		dev_dbg(consumer, "using ACPI for GPIO lookup\n");
> > +		desc = acpi_find_gpio(fwnode, con_id, idx, flags, lookupflags);
> >  	} else {
> > -		return ERR_PTR(-EINVAL);
> > +		desc = ERR_PTR(-ENOENT);
> >  	}
> >  
> > -	/* Currently only ACPI takes this path */
> > +	return desc;
> > +}
> 
> ...
> 
> > +	struct gpio_desc *desc = ERR_PTR(-ENOENT);
> > +	unsigned long lookupflags;
> > +	int ret;
> 
> > +	if (!IS_ERR_OR_NULL(fwnode))
> 
> I think this is superfluous check.
> 
> Now in the form of this series, you have only a single dev_dbg() that tries to
> dereference it. Do we really need to have it there, since every branch has its
> own dev_dbg() anyway?

As I mentioned, I like to keep this check to show the reader that we
should only descend into gpiod_find_by_fwnode() if we have a valid
fwnode. It is less about code generation and more about the intent.

I did change the logging to remove extra dev_dbg(). We will lose message
when dealing with unsupported node type, but that should not really
happen in practice.

> 
> > +		desc = gpiod_find_by_fwnode(fwnode, consumer, con_id, idx,
> > +					    &flags, &lookupflags);
> 
> > +
> 
> This blank line can be dropped after addressing above.
> 
> > +	if (gpiod_not_found(desc) && platform_lookup_allowed) {
> > +		/*
> > +		 * Either we are not using DT or ACPI, or their lookup did not
> > +		 * return a result. In that case, use platform lookup as a
> > +		 * fallback.
> > +		 */
> > +		dev_dbg(consumer, "using lookup tables for GPIO lookup\n");
> > +		desc = gpiod_find(consumer, con_id, idx, &lookupflags);
> > +	}
> > +
> > +	if (IS_ERR(desc)) {
> > +		dev_dbg(consumer, "No GPIO consumer %s found\n", con_id);
> > +		return desc;
> > +	}
> 
> ...
> 
> > +	return gpiod_find_and_request(NULL, fwnode, con_id, index, flags, label,
> > +				      false);
> 
> One line?

OK :)

> 
> ...
> 
> > +	return gpiod_find_and_request(dev, fwnode, con_id, idx, flags, label,
> > +				      true);
> 
> One line?

OK.

Thanks,

-- 
Dmitry