Skip to contentSkip to conversation

leds: is31fl32xx: Fix NULL pointer dereference when binding via I2C ID table

PATCH Superseded Rosen Penev 4 participants 4 messages Oct 5 – Oct 16 1 message not in the archive LED

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 leds: is31fl32xx: Fix NULL pointer dereference when binding via I2C ID table

If the device binds via the I2C ID table (e.g., fallback matching by node name, sysfs new_device, or software node), device_get_match_data() in is31fl32xx_probe() returns NULL because the is31fl32xx_id table entries have no .driver_data.

Without a NULL check, is31fl32xx_parse_dt() will unconditionally dereference cdef when checking cdef->output_frequency_setting_reg, and is31fl32xx_parse_child_dt() will dereference cdef->channels, leading to a NULL pointer dereference and kernel oops.

Add a NULL check for cdef alongside the existing child count check in is31fl32xx_probe(), returning -EINVAL if no match data is available.

Fixes 2779f4724b2f ("leds: various: use device_get_match_data")
Assisted-by LLM
---
 drivers/leds/leds-is31fl32xx.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/leds/leds-is31fl32xx.c +1 -1
index 285085d27453..7bb0295b0b49 100644
@@ -577,7 +577,7 @@ static int is31fl32xx_probe(struct i2c_client *client)
 	cdef = device_get_match_data(dev);
 
 	count = device_get_child_node_count(dev);
-	if (!count)
+	if (!cdef || !count)
 		return -EINVAL;
 
 	priv = devm_kzalloc(dev, struct_size(priv, leds, count),
Signature · 1 line
2.56.0
#1 lore raw
Sashiko has reviewed this patch and found no issues. It looks great!
Quoted 10 lines

If the device binds via the I2C ID table (e.g., fallback matching by node name, sysfs new_device, or software node), device_get_match_data() in is31fl32xx_probe() returns NULL because the is31fl32xx_id table entries have no .driver_data.

Without a NULL check, is31fl32xx_parse_dt() will unconditionally dereference cdef when checking cdef->output_frequency_setting_reg, and is31fl32xx_parse_child_dt() will dereference cdef->channels, leading to a NULL pointer dereference and kernel oops.

Quoted 7 lines

Add a NULL check for cdef alongside the existing child count check in is31fl32xx_probe(), returning -EINVAL if no match data is available.

Fixes 2779f4724b2f ("leds: various: use device_get_match_data")
Assisted-by LLM

Sashiko has reviewed this patch and found no issues. It looks great!

Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1

A message that is not in the archive (sent off-list or not yet received): [email protected]

maintainer
#4 lore raw
Quoted 32 lines · Lee Jones

--- checkpatch.pl: clean (0 issues) ---

Quoted 28 lines · Rosen Penev

If the device binds via the I2C ID table (e.g., fallback matching by node name, sysfs new_device, or software node), device_get_match_data() in is31fl32xx_probe() returns NULL because the is31fl32xx_id table entries have no .driver_data.

Without a NULL check, is31fl32xx_parse_dt() will unconditionally dereference cdef when checking cdef->output_frequency_setting_reg, and is31fl32xx_parse_child_dt() will dereference cdef->channels, leading to a NULL pointer dereference and kernel oops.

Add a NULL check for cdef alongside the existing child count check in is31fl32xx_probe(), returning -EINVAL if no match data is available.

Fixes 2779f4724b2f ("leds: various: use device_get_match_data")
Assisted-by LLM
---
 drivers/leds/leds-is31fl32xx.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/leds/leds-is31fl32xx.c +0 -0
index 285085d27453..7bb0295b0b49 100644
@@ -577,7 +577,7 @@ static int is31fl32xx_probe(struct i2c_client *client)
 	cdef = device_get_match_data(dev);
 
Quoted 4 lines

drivers/leds/leds-is31fl32xx.c static int is31fl32xx_probe(struct i2c_client *client) · line 581

Quoted 4 lines
drivers/leds/leds-is31fl32xx.c +1 -1
 	count = device_get_child_node_count(dev);
-	if (!count)
+	if (!cdef || !count)
 		return -EINVAL;

If it's possible for devices to be autoloaded via is31fl32xx_id, wouldn't it make more sense to populate its .driver_data instead?

Or, if that's not possible for some odd reason (which you must specify), since this is an || it's odd to check the return values like this. It more inline with current coding guidelines to place the check below its caller / allocator.

Quoted 7 lines
Quoted 4 lines
drivers/leds/leds-is31fl32xx.c +0 -0
 	priv = devm_kzalloc(dev, struct_size(priv, leds, count),
Signature · 1 line
2.56.0
Signature · 1 line
Lee Jones
Signature · 1 line
Lee Jones