[patch] [media] rc: double unlock in rc_register_device()

[patch] [media] rc: double unlock in rc_register_device()

[Dan Carpenter]

If change_protocol() fails and we goto out_raw, then it calls unlock
twice.  I noticed that the other time we called change_protocol() we
held the &dev->lock, so I changed it to hold it here too.

Signed-off-by: Dan Carpenter <error27@gmail.com>
---
Compile tested only.

diff --git a/drivers/media/rc/rc-main.c b/drivers/media/rc/rc-main.c
index f57cd56..136a6c6 100644
--- a/drivers/media/rc/rc-main.c
+++ b/drivers/media/rc/rc-main.c
@@ -1097,7 +1097,6 @@ int rc_register_device(struct rc_dev *dev)
 		if (rc < 0)
 			goto out_input;
 	}
-	mutex_unlock(&dev->lock);
 
 	if (dev->change_protocol) {
 		rc = dev->change_protocol(dev, rc_map->rc_type);
@@ -1105,6 +1104,8 @@ int rc_register_device(struct rc_dev *dev)
 			goto out_raw;
 	}
 
+	mutex_unlock(&dev->lock);
+
 	IR_dprintk(1, "Registered rc%ld (driver: %s, remote: %s, mode %s)\n",
 		   dev->devno,
 		   dev->driver_name ? dev->driver_name : "unknown",

Re: [patch] [media] rc: double unlock in rc_register_device()

[Jarod Wilson]

Dan Carpenter wrote:
> If change_protocol() fails and we goto out_raw, then it calls unlock
> twice.

Gah, good catch, I gotta quit adding new bugs... ;)


> I noticed that the other time we called change_protocol() we
> held the &dev->lock, so I changed it to hold it here too.
>
> Signed-off-by: Dan Carpenter<error27@gmail.com>
> ---
> Compile tested only.


I've sanity-checked the code, and yeah, calling change_protocol() 
function pointers with the lock held should be perfectly fine here too. 
The change_protocol functions are device-driver-specific, and don't 
touch the core rc device lock.

Reviewed-by: Jarod Wilson <jarod@redhat.com>
Acked-by: Jarod Wilson <jarod@redhat.com>

-- 
Jarod Wilson
jarod@redhat.com