Skip to contentSkip to conversation

pinctrl: Move for_each_maps() to namespace and hide iterator inside

PATCH Superseded Andy Shevchenko 2 participants 4 messages Nov 9, 2022 – Nov 11, 2022

Subsystems (2)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 pinctrl: Move for_each_maps() to namespace and hide iterator inside

First of all, while for_each_maps() is private to pin control subsystem it's still better to have it put into a namespace.

Besides that, users are not relying on iterator variable, so hide it inside for-loop.

---
 drivers/pinctrl/core.c |  6 ++----
 drivers/pinctrl/core.h | 10 +++++-----
 2 files changed, 7 insertions(+), 9 deletions(-)
drivers/pinctrl/core.c +2 -4
index 9e57f4c62e60..f2f99972a0d3 100644
@@ -1028,7 +1028,6 @@ static struct pinctrl *create_pinctrl(struct device *dev,
 	struct pinctrl *p;
 	const char *devname;
 	struct pinctrl_maps *maps_node;
-	int i;
 	const struct pinctrl_map *map;
 	int ret;
 
@@ -1054,7 +1053,7 @@ static struct pinctrl *create_pinctrl(struct device *dev,
 
 	mutex_lock(&pinctrl_maps_mutex);
 	/* Iterate over the pin control maps to locate the right ones */
-	for_each_maps(maps_node, i, map) {
+	for_each_pin_map(maps_node, map) {
 		/* Map must be for this device */
 		if (strcmp(map->dev_name, devname))
 			continue;
@@ -1805,13 +1804,12 @@ static inline const char *map_type(enum pinctrl_map_type type)
 static int pinctrl_maps_show(struct seq_file *s, void *what)
 {
 	struct pinctrl_maps *maps_node;
-	int i;
 	const struct pinctrl_map *map;
 
 	seq_puts(s, "Pinctrl maps:\n");
 
 	mutex_lock(&pinctrl_maps_mutex);
-	for_each_maps(maps_node, i, map) {
+	for_each_pin_map(maps_node, map) {
 		seq_printf(s, "device %s\nstate %s\ntype %s (%d)\n",
 			   map->dev_name, map->name, map_type(map->type),
 			   map->type);
drivers/pinctrl/core.h +5 -5
index 840103c40c14..8b90b4f0bc7e 100644
@@ -242,8 +242,8 @@ extern int pinctrl_force_default(struct pinctrl_dev *pctldev);
 extern struct mutex pinctrl_maps_mutex;
 extern struct list_head pinctrl_maps;
 
-#define for_each_maps(_maps_node_, _i_, _map_) \
-	list_for_each_entry(_maps_node_, &pinctrl_maps, node) \
-		for (_i_ = 0, _map_ = &_maps_node_->maps[_i_]; \
-			_i_ < _maps_node_->num_maps; \
-			_i_++, _map_ = &_maps_node_->maps[_i_])
+#define for_each_pin_map(_maps_node_, _map_)						\
+	list_for_each_entry(_maps_node_, &pinctrl_maps, node)				\
+		for (unsigned int __i = 0;						\
+		     _map_ = &_maps_node_->maps[__i], __i < _maps_node_->num_maps;	\
+		     __i++)
Signature · 1 line
2.35.1
#1 lore raw
Quoted 5 lines · Andy Shevchenko

First of all, while for_each_maps() is private to pin control subsystem it's still better to have it put into a namespace.

Besides that, users are not relying on iterator variable, so hide it inside for-loop.

...

drivers/pinctrl/core.h extern int pinctrl_force_default(struct pinctrl_dev *pctldev); · line 248

Quoted 5 lines
drivers/pinctrl/core.h +4 -0
+#define for_each_pin_map(_maps_node_, _map_)						\
+	list_for_each_entry(_maps_node_, &pinctrl_maps, node)				\
+		for (unsigned int __i = 0;						\
 
+		     _map_ = &_maps_node_->maps[__i], __i < _maps_node_->num_maps;	\

Hmm... I think this is actually not okay, if we have maps be NULL and num_maps = 0, KABOOM is guaranteed.

I will experiment and update this.

Meanwhile, Linus, do you think this change is useful?

Quoted 1 line
drivers/pinctrl/core.h +1 -0
+		     __i++)
Signature · 2 lines
With Best Regards,
Andy Shevchenko
#2 lore raw
Quoted 13 lines · Andy Shevchenko
Quoted 5 lines · Andy Shevchenko

First of all, while for_each_maps() is private to pin control subsystem it's still better to have it put into a namespace.

Besides that, users are not relying on iterator variable, so hide it inside for-loop.

...

Quoted 3 lines
drivers/pinctrl/core.h +3 -0
+#define for_each_pin_map(_maps_node_, _map_)                                         \
+     list_for_each_entry(_maps_node_, &pinctrl_maps, node)                           \
+             for (unsigned int __i = 0;                                              \
Quoted 6 lines
Quoted 1 line

+ _map_ = &_maps_node_->maps[__i], __i < _maps_node_->num_maps; \

Hmm... I think this is actually not okay, if we have maps be NULL and num_maps = 0, KABOOM is guaranteed.

I will experiment and update this.

OK

Quoted 1 line

Meanwhile, Linus, do you think this change is useful?

Even if just a name change, it makes things better by being more readable so yes :)

Yours,
Linus Walleij

#3 lore raw
Quoted 3 lines · Linus Walleij
Quoted 1 line · Andy Shevchenko

On Wed, Nov 09, 2022 at 05:57:24PM +0200, Andy Shevchenko wrote:

...

Quoted 4 lines
Quoted 1 line

Meanwhile, Linus, do you think this change is useful?

Even if just a name change, it makes things better by being more readable so yes :)

v2 has been sent.

Signature · 2 lines
With Best Regards,
Andy Shevchenko