netfilter-devel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Patrick McHardy <kaber@trash.net>
To: Jan Engelhardt <jengelh@medozas.de>
Cc: Simon Lodal <simonl@parknet.dk>, netfilter-devel@vger.kernel.org
Subject: Re: xt2 table core
Date: Mon, 05 Jul 2010 10:55:51 +0200	[thread overview]
Message-ID: <4C319E17.2020908@trash.net> (raw)
In-Reply-To: <alpine.LSU.2.01.1007041440130.12356@obet.zrqbmnf.qr>

Jan Engelhardt wrote:
> On Friday 2010-07-02 05:32, Simon Lodal wrote:
>   
>>> This patchset adds the xtables2 main proper, and incrementally
>>> changes ip6_tables to use it. There are more patches for iptables
>>> and arptables and, but to not make it larger than needed, just this
>>> much for now.
>>>
>>> Previously featured on http://lwn.net/Articles/345176/ .
>>>       
>> Nice to see some new development! Guess we all have a long wishlist that is 
>> hard to implement in the current design, and this may make it easier.
>>
>> But what about performance? I think all these lists must cause a huge amount 
>> of jumps in hot path (eg. filtering packets). The packed blob format is hard to 
>> work with, but once created it is efficient. Do you have any expectations or 
>> even numbers on the performance?
>>     
>
> Your suspicion was right, linked lists seem to have their impact.
>
> The command I used was
>
> 	ping6 localhost -fc 500 -i .001
>
> this sends 500 packets in near-flood fashion (-i0 would lead to
> packet drops).
>
> Xtables1 blob-style 2.6.31-rc4:
> 500 packets transmitted, 500 received, 0% packet loss, time 3532ms
> 500 packets transmitted, 500 received, 0% packet loss, time 3428ms
> 500 packets transmitted, 500 received, 0% packet loss, time 3388ms
> 500 packets transmitted, 500 received, 0% packet loss, time 3339ms
> 500 packets transmitted, 500 received, 0% packet loss, time 3219ms
>
> Xtables1 blob-style 2.6.34:
> 500 packets transmitted, 500 received, 0% packet loss, time 3405ms
> 500 packets transmitted, 500 received, 0% packet loss, time 3525ms
> 500 packets transmitted, 500 received, 0% packet loss, time 3426ms
> 500 packets transmitted, 500 received, 0% packet loss, time 3388ms
> 500 packets transmitted, 500 received, 0% packet loss, time 3328ms
>
> Xtables2 (linked lists) 2.6.35-rc1:
> 500 packets transmitted, 500 received, 0% packet loss, time 14775ms
> 500 packets transmitted, 500 received, 0% packet loss, time 14383ms
> 500 packets transmitted, 500 received, 0% packet loss, time 14381ms
> 500 packets transmitted, 500 received, 0% packet loss, time 15432ms
> 500 packets transmitted, 500 received, 0% packet loss, time 14498ms
>   

While this is probably a pretty bad test, you can't seriously expect
me to merge something that degrades performance by 400% for all users?


  parent reply	other threads:[~2010-07-05  8:55 UTC|newest]

Thread overview: 46+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-06-29  8:42 xt2 table core Jan Engelhardt
2010-06-29  8:42 ` [PATCH 01/56] netfilter: ebtables: simplify a device in/out check Jan Engelhardt
2010-06-29  8:42 ` [PATCH 02/56] netfilter: ebtables: change ebt_basic_match to xt convention Jan Engelhardt
2010-06-29  8:42 ` [PATCH 03/56] netfilter: xtables: move functions around Jan Engelhardt
2010-06-29  8:42 ` [PATCH 04/56] netfilter: xtables: convert basic nfproto match functions into xt matches Jan Engelhardt
2010-06-29  8:42 ` [PATCH 05/56] netfilter: xtables2: initial table skeletal functions Jan Engelhardt
2010-06-29  8:42 ` [PATCH 06/56] netfilter: xtables2: initial chain " Jan Engelhardt
2010-06-29  8:42 ` [PATCH 07/56] netfilter: xtables2: initial rule " Jan Engelhardt
2010-06-29  8:42 ` [PATCH 08/56] netfilter: xtables: alternate size checking in xt_check_match Jan Engelhardt
2010-06-29  8:42 ` [PATCH 09/56] netfilter: xtables: alternate size checking in xt_check_target Jan Engelhardt
2010-06-29  8:42 ` [PATCH 10/56] netfilter: xtables2: per-rule match skeletal functions Jan Engelhardt
2010-06-29  8:42 ` [PATCH 11/56] netfilter: xtables2: per-rule target " Jan Engelhardt
2010-06-29  8:42 ` [PATCH 12/56] netfilter: xtables2: xt_check_target in combination with xt2 contexts Jan Engelhardt
2010-06-29  8:42 ` [PATCH 13/56] netfilter: xtables2: jumpstack (de)allocation functions Jan Engelhardt
2010-06-29  8:42 ` [PATCH 14/56] netfilter: xtables2: table traversal Jan Engelhardt
2010-06-29  8:42 ` [PATCH 15/56] netfilter: xtables: add xt_quota revision 3 Jan Engelhardt
2010-06-29  8:42 ` [PATCH 16/56] netfilter: xtables2: make a copy of the ipv6_filter table Jan Engelhardt
2010-06-29  8:42 ` [PATCH 17/56] netfilter: xtables2: initial xt1->xt2 translation for tables Jan Engelhardt
2010-06-29  8:42 ` [PATCH 18/56] netfilter: xtables2: xt2->xt1 translation - GET_INFO support Jan Engelhardt
2010-06-29  8:42 ` [PATCH 19/56] netfilter: xtables2: xt2->xt1 translation - GET_ENTRIES support Jan Engelhardt
2010-06-29  8:43 ` [PATCH 20/56] netfilter: xtables2: xt1->xt2 translation - SET_REPLACE support Jan Engelhardt
2010-06-29  8:43 ` [PATCH 21/56] netfilter: xtables2: return counters after SET_REPLACE Jan Engelhardt
2010-06-29  8:43 ` [PATCH 22/56] netfilter: xtables2: xt1->xt2 translation - ADD_COUNTERS support Jan Engelhardt
2010-06-29  8:43 ` [PATCH 23/56] netfilter: xtables2: xt2->xt1 translation - compat GET_INFO support Jan Engelhardt
2010-06-29  8:43 ` [PATCH 24/56] netfilter: ip6tables: move mark_chains to xt1_perproto.c Jan Engelhardt
2010-06-29  8:43 ` [PATCH 25/56] netfilter: xtables2: xt2<->xt1 translation - compat GET_ENTRIES/SET_REPLACE support Jan Engelhardt
2010-06-29  8:43 ` [PATCH 26/56] netfilter: xtables2: compat->normal match data translation Jan Engelhardt
2010-06-29  8:43 ` [PATCH 27/56] netfilter: xtables2: compat->normal target " Jan Engelhardt
2010-06-29  8:43 ` [PATCH 28/56] netfilter: xtables2: outsource code into xts_match_to_xt1 function Jan Engelhardt
2010-06-29  8:43 ` [PATCH 29/56] netfilter: xtables2: normal->compat match data translation Jan Engelhardt
2010-06-29  8:43 ` [PATCH 30/56] netfilter: xtables2: normal->compat target " Jan Engelhardt
2010-06-29  8:43 ` [PATCH 31/56] netfilter: xtables2: packet tracing Jan Engelhardt
2010-06-29  8:43 ` [PATCH 32/56] netfilter: xtables: turn procfs entries to walk xt2 table list Jan Engelhardt
2010-06-29  8:43 ` [PATCH 33/56] netfilter: xtables2: switch ip6's tables to the xt2 table format Jan Engelhardt
2010-06-29  8:47 ` xt2 table core [*/33, not */56] Jan Engelhardt
2010-07-02  3:32 ` xt2 table core Simon Lodal
2010-07-04 13:56   ` Jan Engelhardt
2010-07-04 17:22     ` Simon Lodal
2010-07-04 18:00       ` Jan Engelhardt
2010-07-05  8:55     ` Patrick McHardy [this message]
2010-07-05  9:13       ` Jan Engelhardt
2010-07-05  9:15         ` Patrick McHardy
2010-07-05  9:36     ` Eric Dumazet
2010-07-05  9:42       ` Jan Engelhardt
2010-07-05 10:22         ` Eric Dumazet
2010-07-05 10:34           ` Jan Engelhardt

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4C319E17.2020908@trash.net \
    --to=kaber@trash.net \
    --cc=jengelh@medozas.de \
    --cc=netfilter-devel@vger.kernel.org \
    --cc=simonl@parknet.dk \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).