Re: [PATCH 1/4] Fix expectaction mask dumping, take #3

[Patrick McHardy <kaber@trash.net>]

Yasuyuki KOZAKAI wrote:
> From: Pablo Neira Ayuso <pablo@netfilter.org>
> Date: Wed, 22 Feb 2006 04:20:39 +0100
> 
> 
>>>                                       Moreover, the expectation mask has
>>>l3num value that is different from the l3num in the nfnetlink header,
>>>that's why I introduced this field.
> 
> 
> Yes, but in the current code l3num in expectation mask is always 0xff.
> And even if we port all helpers of ip_conntrack to nf_conntrack,
> they will set 0xff to l3num.
> 
> Which situation kernel wants to set the value except of 0xff to l3num
> in expectation mask ?

Harald and me were actually talking about simplifying the helper lookup
by making the dst part of an expectation fixed (for ip_conntrack). With
nf_conntrack l3num is in the src-part of the tuple, but thats just
for memory-efficiency, it can be treated equally to protonum and the
mask can be eliminated.