arp.c duplicate assignment of skb->dev ("cosmetic")

arp.c duplicate assignment of skb->dev ("cosmetic")

[Herbert Rosmanith]

hi,

this is rather cosmetical than functional, but in arp.c,
in arp_send(), on line 489

	skb->dev=dev;

is assigned. in line 563, still in the same routine, it is
assigned again without skb or skb->dev being changed. so I guess
this second assignment is not neccessary. I can't see where
skb->dev is changed within the 80  lines. so, we could remove
the second assignment.

/herp

Re: arp.c duplicate assignment of skb->dev ("cosmetic")

[Bryan Rittmeyer]

Herbert Rosmanith wrote:

> hi,
> 
> this is rather cosmetical than functional, but in arp.c,
> in arp_send(), on line 489
> 
> 	skb->dev=dev;
> 
> is assigned. in line 563, still in the same routine, it is
> assigned again without skb or skb->dev being changed. so I guess
> this second assignment is not neccessary. I can't see where
> skb->dev is changed within the 80  lines. so, we could remove
> the second assignment.


Hi Herbert,

I also noticed this dupe a couple days ago and pointed it out to
the lkml and net code maintainers. Alexey Kuznetsov has
agreed, saying:

"Apparently, this is right. This happened ages ago when eliminating
second argument dev_queue_xmit()."

I assume a fix will be in the next major release.

Regards,

Bryan


-- 
Bryan Rittmeyer
mailto:bryan@ixiacom.com

Re: arp.c duplicate assignment of skb->dev ("cosmetic")

[Herbert Rosmanith]

> > this is rather cosmetical than functional, but in arp.c,
> > in arp_send(), on line 489
> > 
> > 	skb->dev=dev;
> > 
> > is assigned. in line 563, still in the same routine, it is
> > assigned again without skb or skb->dev being changed. so I guess
> 
> Hi Herbert,
> 
> I also noticed this dupe a couple days ago and pointed it out to
> the lkml

eeks. I hope it does not look like I have been posting this only because
I saw your email to lkml. In fact, I've found this a while ago, but was
just too lazy to email it.