Re: [PATCH v4] tcp: harmonize tcp_vx_rcv header length assumptions

[Patrick McHardy <kaber@trash.net>]

Joe Perches wrote:
> On Wed, 2010-01-13 at 16:13 -0500, William Allen Simpson wrote:
>> I did today's tests on tcp_ipv4, but part of this harmonization is to
>> make v4 and v6 match up better.  As I was examining line by line, one of
>> the things that leaps out is that ipv4 has nf_reset(skb):
>>
>> tcp_ipv4:
>> 	if (!xfrm4_policy_check(sk, XFRM_POLICY_IN, skb))
>> 		goto discard_and_relse;
>> 	nf_reset(skb);
>>
>> 	if (sk_filter(sk, skb))
>> 		goto discard_and_relse;
>>
>> tcp_ipv6:
>> 	if (!xfrm6_policy_check(sk, XFRM_POLICY_IN, skb))
>> 		goto discard_and_relse;
>>
>> 	if (sk_filter(sk, skb))
>> 		goto discard_and_relse;
>>
>> Does anybody know why?  Should ipv6 have it?  Or at least a comment
>> explaining the reasoning for the omission?
> 
> You should ask the person that put the line in, but
> it might be because no IPv6 NAT support exists.
> 
> $ git blame -L 1658,1658 net/ipv4/tcp_ipv4.c
> b59c2701 (Patrick McHardy 2006-01-06 23:06:10 -0800 1658)       nf_reset(skb);
> 
> And
> 
> $ git log -p -1 b59c2701
> commit b59c270104f03960069596722fea70340579244d
> Author: Patrick McHardy <kaber@trash.net>
> Date:   Fri Jan 6 23:06:10 2006 -0800
> 
>     [NETFILTER]: Keep conntrack reference until IPsec policy checks are done
>     
>     Keep the conntrack reference until policy checks have been performed for
>     IPsec NAT support. The reference needs to be dropped before a packet is
>     queued to avoid having the conntrack module unloadable.

In IPv6 it is released in ip6_input.c before the packet is handed to
the protocol handler.