nth, random and fuzzy match

nth, random and fuzzy match

[Patrick McHardy]

Also following the decisions from the netfilter workshop,
I've started merging nth and random into a new "statistic"
match. I won't merge fuzzy since it seems to obscure and
complex and I don't see what it offers beyond nth + random.

The current nth match optionally uses shared state for the
counters (which as we know isn't pretty with the current
iptables userspace interface) and has a compile-time limit
on the number of shared counters. I don't see the need for
this shared state, I think everything expressable today
should also be expressable using just per-match state.

So my questions to the nth users: is anyone using the nth
match in a way that can't be done by just matching on every
nth packet with an optional starting point?

Re: nth, random and fuzzy match

[Harald Welte]

[-- Attachment #1: Type: text/plain, Size: 707 bytes --]

On Thu, Apr 06, 2006 at 07:53:30PM +0200, Patrick McHardy wrote:
> So my questions to the nth users: is anyone using the nth
> match in a way that can't be done by just matching on every
> nth packet with an optional starting point?

I suggest posting this to the users mailinglist, since the question is
addressed to users :)

-- 
- Harald Welte <laforge@netfilter.org>                 http://netfilter.org/
============================================================================
  "Fragmentation is like classful addressing -- an interesting early
   architectural error that shows how much experimentation was going
   on while IP was being designed."                    -- Paul Vixie

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: nth, random and fuzzy match

[Patrick McHardy]

Harald Welte wrote:
> On Thu, Apr 06, 2006 at 07:53:30PM +0200, Patrick McHardy wrote:
> 
>>So my questions to the nth users: is anyone using the nth
>>match in a way that can't be done by just matching on every
>>nth packet with an optional starting point?
> 
> 
> I suggest posting this to the users mailinglist, since the question is
> addressed to users :)

I already decided for myself, the functionality is not affected
by removing the shared counters, it just needs to be used
slightly different in some cases.