Re: [PATCH] of: Create function for counting number of phandles in a property

[Grant Likely <grant.likely@secretlab.ca>]

On Mon, 11 Feb 2013 12:26:15 +0100, Andreas Larsson <andreas@gaisler.com> wrote:
> On 2013-02-11 00:58, Grant Likely wrote:
> > diff --git a/drivers/of/base.c b/drivers/of/base.c
> > index 2390ddb..e1120a2 100644
> > --- a/drivers/of/base.c
> > +++ b/drivers/of/base.c
> > @@ -1025,12 +1025,13 @@ EXPORT_SYMBOL(of_parse_phandle);
> >    * To get a device_node of the `node2' node you may call this:
> >    * of_parse_phandle_with_args(node3, "list", "#list-cells", 1, &args);
> >    */
> > -int of_parse_phandle_with_args(const struct device_node *np, const char *list_name,
> > -				const char *cells_name, int index,
> > -				struct of_phandle_args *out_args)
> > +static int __of_parse_phandle_with_args(const struct device_node *np,
> > +					const char *list_name,
> > +					const char *cells_name, int index,
> > +					struct of_phandle_args *out_args)
> >   {
> >   	const __be32 *list, *list_end;
> > -	int size, cur_index = 0;
> > +	int rc = 0, size, cur_index = 0;
> >   	uint32_t count = 0;
> >   	struct device_node *node = NULL;
> >   	phandle phandle;
> > @@ -1059,12 +1060,14 @@ int of_parse_phandle_with_args(const struct device_node *np, const char *list_na
> >   			if (!node) {
> >   				pr_err("%s: could not find phandle\n",
> >   					 np->full_name);
> > +				rc = -EINVAL;
> >   				break;
> >   			}
> >   			if (of_property_read_u32(node, cells_name, &count)) {
> >   				pr_err("%s: could not get %s for %s\n",
> >   					 np->full_name, cells_name,
> >   					 node->full_name);
> > +				rc = -EINVAL;
> >   				break;
> >   			}
> >
> > @@ -1075,6 +1078,7 @@ int of_parse_phandle_with_args(const struct device_node *np, const char *list_na
> >   			if (list + count > list_end) {
> >   				pr_err("%s: arguments longer than property\n",
> >   					 np->full_name);
> > +				rc = -EINVAL;
> >   				break;
> >   			}
> >   		}
> > @@ -1086,8 +1090,10 @@ int of_parse_phandle_with_args(const struct device_node *np, const char *list_na
> >   		 * or return -ENOENT for an empty entry.
> >   		 */
> >   		if (cur_index == index) {
> > -			if (!phandle)
> > -				return -ENOENT;
> > +			if (!phandle) {
> > +				rc = -ENOENT;
> > +				break;
> > +			}
> >
> >   			if (out_args) {
> >   				int i;
> > @@ -1098,22 +1104,54 @@ int of_parse_phandle_with_args(const struct device_node *np, const char *list_na
> >   				for (i = 0; i < count; i++)
> >   					out_args->args[i] = be32_to_cpup(list++);
> >   			}
> > -			return 0;
> > +
> > +			rc = 0;
> > +			break;
> >   		}
> >
> >   		of_node_put(node);
> >   		node = NULL;
> >   		list += count;
> >   		cur_index++;
> > +		rc = cur_index;
> >   	}
> >
> >   	/* Loop exited without finding a valid entry; return an error */
> >   	if (node)
> >   		of_node_put(node);
> > -	return -EINVAL;
> > +	return rc;
> > +}
> > +
> > +int of_parse_phandle_with_args(const struct device_node *np, const char *list_name,
> > +				const char *cells_name, int index,
> > +				struct of_phandle_args *out_args)
> > +{
> > +	return __of_parse_phandle_with_args(np, list_name, cells_name, index, out_args);
> >   }
> >   EXPORT_SYMBOL(of_parse_phandle_with_args);
> 
> Will this not result in a situation where a call to 
> of_parse_phandle_with_args with an out of bounds index returns the 
> number of tuples instead of an error code and possibly some caller that 
> uses the this count as a phandle instead of handling an error?

Yes, you are right about the out of bounds index. I had meant to write
the following in __of_parse_phandle_with_args:

	return (index < 0) ? rc : -ENOENT;

Which should solve that problem.

> 
> Of course of_count_phandle_with_args can be used to make sure that no 
> such call is made in the first place, but that is another story.

Not really, the __ function still has to test for a negative index.
However, it can at least make sure the index passed is not negative so
you can't get count behaviour when calling for a parse. How about:

int of_parse_phandle_with_args(const struct device_node *np, const char *list_name,
				const char *cells_name, int index,
				struct of_phandle_args *out_args)
{
	if (index < 0)
		return -EINVAL;
	return __of_parse_phandle_with_args(np, list_name, cells_name, index, out_args);
}


> 
> Related to this is that Case 7 in of_selftest_parse_phandle_with_args 
> never gets exercised as far as I can see.

You're right. Also a bug. Fixed now.

> 
> 
> > diff --git a/include/linux/of_gpio.h b/include/linux/of_gpio.h
> > index c454f57..bdbe0f3 100644
> > --- a/include/linux/of_gpio.h
> > +++ b/include/linux/of_gpio.h
> > @@ -50,8 +50,28 @@ static inline struct of_mm_gpio_chip *to_of_mm_gpio_chip(struct gpio_chip *gc)
> >   extern int of_get_named_gpio_flags(struct device_node *np,
> >   		const char *list_name, int index, enum of_gpio_flags *flags);
> >
> > -extern unsigned int of_gpio_named_count(struct device_node *np,
> > -					const char* propname);
> > +/**
> > + * of_gpio_named_count - Count GPIOs for a device
> > + * @np:		device node to count GPIOs for
> > + * @propname:	property name containing gpio specifier(s)
> > + *
> > + * The function returns the count of GPIOs specified for a node.
> > + *
> > + * Note that the empty GPIO specifiers counts too. For example,
> > + *
> > + * gpios = <0
> > + *          &pio1 1 2
> > + *          0
> > + *          &pio2 3 4>;
> > + *
> > + * defines four GPIOs (so this function will return 4), two of which
> > + * are not specified. Returns -EINVAL for an incorrectly formed gpios
> > + * property.
> > + */
> > +static int of_gpio_named_count(struct device_node *np, const char* propname)
> > +{
> > +	return of_count_phandle_with_args(np, propname, "#gpio-cells");
> > +}
> 
> Should this be static inline int?

Yes.

> 
> I think it would be good to also document that it also returns -ENOENT 
> when the propname property is missing, which might be an important case 
> to distinguish from the -EINVAL case.

Done. I'll post a new version shortly.

g.