[CFT/Review] net byte order for AF_INET

[Gleb Smirnoff]

On Thu, Oct 11, 2012 at 02:06:11AM +0200, Luigi Rizzo wrote:
L> > M> > @@ -504,6 +504,9 @@
L> > M> >  		if (ip->ip_id == 0)
L> > M> >  			ip->ip_id = ip_newid();
L> > M> >  
L> > M> > +		ip->ip_len = htons(ip->ip_len);
L> > M> > +		ip->ip_off = htons(ip->ip_off);
L> > M> > +
L> > M> 
L> > M> So the packet is expected to come into rip_output() from caller 
L> > M> with ip_len/ip_off in host byte order, right?  As already 
L> > M> suggested - it would be good to add a comment explaining this.
L> > 
L> > This is de facto standard for raw sockets in most OS-es. Byte order
L> > in raw socket is host. And this is the same behavior we had before
L> > the patch. So no reason for extra comments.
L> 
L> ??? just the fact that there was a debate on this item is sufficient
L> ground for commenting it. The fact that this is (presumably) one
L> of the only two places (in and out on raw IP sockets ?) where
L> ip_len and ip_off are in host format is another good reason
L> to comment both instances.

Okay, agreed :)

-- 
Totus tuus, Glebius.