Re: [PATCH nf-next 0/8] netfilter: untangle bridge and bridge netfilter

[David Miller <davem@redhat.com>]

From: Florian Westphal <fw@strlen.de>
Date: Mon, 9 Mar 2015 14:13:56 +0100

> Pablo Neira Ayuso <pablo@netfilter.org> wrote:
> 
> [ CC Andy, see below for ip_fragment api discussion ]
> 
>> > Lets start untangling bridge, bridge netfilter, and the
>> > rest of the ip stack (esp. ip_fragment).
>> > 
>> > This changes ip_fragment() so that bridge netfilter
>> > can pass in the required information as arguments instead
>> > of using skb->nf_bridge to pass some extra information to it.
>> > 
>> > Another problem with br_netfilter and the way its plumbed to
>> > ip/ip6-tables (physdev match) is skb->nf_bridge.
>> > 
>> > nf_bridge is kmalloced blob with some extra information, including
>> > the bridge in and outports (mainly for iptables' physdev match).
>> > It also has various state bits so we know what manipulations
>> > have been performed by bridge netfilter on the skb (e.g.
>> > ppp header stripping).
>> >
>> > nf_bridge also provides scratch space where br_netfilter saves
>> > (and later restores) various things, e.g. ipv4 address for
>> > dnat detection, mac address to fix up ip fragmented skbs, etc.
>> > 
>> > But in almost all cases we can avoid using ->data completely.
>> 
>> I think one of the goals of this patchset is to prepare the removal of
>> that nf_bridge pointer from sk_buff which sounds as a good idea to me.
> 
> The 2nd (and last half) of the set folds nf_bridge_info into skbuff,
> at the end (where its not initialized at allocation time).
> 
> The struct is a lot smaller by then (16 bytes on 64bit, 12 on 32bit)
> so we'd increase skbuff size only by 8.

Sorry, increases in size of any kind of sk_buff are strongly discouraged.

Especially for something like nfbridge.