Skip to contentSkip to conversation

net/rose: Remove unnecessary if check in rose_dev_first()

PATCH In review Thorsten Blum 3 participants 3 messages Jul 4, 2025 – Jul 14, 2025

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 net/rose: Remove unnecessary if check in rose_dev_first()

dev_hold() already checks if its argument is NULL.

---
 net/rose/rose_route.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
net/rose/rose_route.c +1 -2
index b72bf8a08d48..35e21a2bec9c 100644
@@ -608,8 +608,7 @@ struct net_device *rose_dev_first(void)
 			if (first == NULL || strncmp(dev->name, first->name, 3) < 0)
 				first = dev;
 	}
-	if (first)
-		dev_hold(first);
+	dev_hold(first);
 	rcu_read_unlock();
 
 	return first;
Signature · 1 line
2.50.0
reviewer
#1 lore raw
Quoted 3 lines · Thorsten Blum

dev_hold() already checks if its argument is NULL.

Hi Thorsten,

I agree that this is correct. But I think that cleanup like this needs to be in the context of other changes to make it worthwhile.

Quoting documentation:

Clean-up patches
~~~~~~~~~~~~~~~~

Netdev discourages patches which perform simple clean-ups, which are not in the context of other work. For example:

* Addressing ``checkpatch.pl`` warnings
* Addressing :ref:Local variable ordering<rcs> issues
* Conversions to device-managed APIs (``devm_`` helpers)

This is because it is felt that the churn that such changes produce comes at a greater cost than the value of such clean-ups.

Conversely, spelling and grammar fixes are not discouraged.

See: https://www.kernel.org/doc/html/next/process/maintainer-netdev.html#clean-up-patches

--
pw-bot: cr

#2 lore raw
In reply to Thorsten Blum
Quoted 12 lines · Thorsten Blum

dev_hold() already checks if its argument is NULL.

---
 net/rose/rose_route.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
net/rose/rose_route.c +0 -0
index b72bf8a08d48..35e21a2bec9c 100644

net/rose/rose_route.c struct net_device *rose_dev_first(void) · line 611

Quoted 6 lines
net/rose/rose_route.c +1 -2
@@ -608,8 +608,7 @@ struct net_device *rose_dev_first(void)
 			if (first == NULL || strncmp(dev->name, first->name, 3) < 0)
 				first = dev;
 	}
-	if (first)
-		dev_hold(first);
+	dev_hold(first);

I'm not a fan of these sorts of "remove the NULL check" patches in general. Sure it removes a line of code, but does it really improve readability? I feel like someone reading this code might think a NULL check was required.

I guess there is also an argument that this is a tiny speedup. That could be a valid argument especially if we had benchmarking data to back it up.

Of course, if you're planning to take over this code and be the maintainer of it, then you get to do whatever you feel is best. So if this change were part of a larger change where you were taking over then that's fine.

regards,
dan carpenter