Re: [PATCH 0/3] netfilter : 3 patches to boost ip_tables performance

[Harald Welte <laforge@netfilter.org>]

On Thu, Sep 22, 2005 at 03:03:21PM +0200, Andi Kleen wrote:
> 
> > 1) No more central rwlock protecting each table (filter, nat, mangle, raw),
> >     but one lock per CPU. It avoids cache line ping pongs for each packet.
> 
> Another useful change would be to not take the lock when there are no
> rules. Currently just loading iptables has a large overhead.

This is partially due to the netfilter hooks that are registered (so we
always take nf_hook_slow() in the NF_HOOK() macro).

The default policies inside an iptables chain are internally implemented
as a rule.  Thus, policies as built-in rules have packet/byte counters.

Therefore, without making a semantic change, we cannot do any of the
following optimizations:

1) not take a lock when the chain is empty
2) not register at the netfilter hook when the chain is empty.

This is well-known, but I don't think we can change the semantics for
the user during a stable kernel series.  That's one point where not
having 2.7.x really hurts.

-- 
- 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: pgpL5tLBAF0Oz.pgp
Description: PGP signature