From: jamal <hadi@cyberus.ca>
To: "David S. Miller" <davem@redhat.com>
Cc: <ak@suse.de>, <greearb@candelatech.com>,
<linux-kernel@vger.kernel.org>, <netdev@oss.sgi.com>
Subject: Re: [PATCH] hashed device lookup (Does NOT meet Linus' sumission policy!)
Date: Sun, 7 Jan 2001 10:56:23 -0500 (EST) [thread overview]
Message-ID: <Pine.GSO.4.30.0101071026070.18916-100000@shell.cyberus.ca> (raw)
In-Reply-To: <200101070543.VAA24689@pizda.ninka.net>
On Sat, 6 Jan 2001, David S. Miller wrote:
> I could not agree more. This reminds me to do something I could not
> justify before, making netlink be enabled in the kernel and
> non-configurable.
I always use netlink and friends for something or the other. Route
protocols, traffic control configuration, device configuration,
firewalling; I dont see a problem turning it on. Maybe defaulting it to
on, and then totaly getting rid of the config option.
> I could almost, but not quite, justify it right now just because "ip"
> is becomming standard and needs it.
"ip" has been standard to some people for a few moons now ;->
It is more documneted than ifconfig ;-> and does a lot more
>
> Not to stray from the subject, Ben's effort is still needed. I think real
> numbers are useful instead of claims like it "displayed faster"
>
> See my previous email, if it's just slow because of some poorly coded
> version of ifconfig, it does not justify the patch. If only a
> forcefully created "benchmark" can show some performance problem, that
> is not an acceptable reason to champion this patch. Ok?
>
True.
All the kernel paths affected are really config/control paths as well as
slow paths (eg some device binding to sockets, MSG_DONTROUTE slow path,
traffic control setup, route config, arp config are some that seem useful).
The argumenet for the patch (and it might be weak) If you have 4000
devices, it makes sense for two reasons:
-Some of these paths might hold locks which might stall the fast path.
-Or if you have a busy system, they will contribute to the load.
I used to be against VLANS being devices, i am withdrawing that comment; it's
a lot easier to look on them as devices if you want to run IP on them. And
in this case, it makes sense the possibilirt of over a thousand devices
is good.
So instead of depending what ifconfig does, maybe a better test for Ben is
to measure the kernel level improvement in the lookup for example from
2..6000 devices. Tests with the user space tools will also help. example
to add to Andi's flavor:
"date; time ifconfig -a; date" for each number of devices.
repeat for ip as well ;->
Plot kernel and the two user tools results after several iterations
(number of devices on x-axis and some time/CPU measure on y-axis).
cheers,
jamal
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/
next prev parent reply other threads:[~2001-01-07 15:57 UTC|newest]
Thread overview: 77+ messages / expand[flat|nested] mbox.gz Atom feed top
2001-01-08 8:12 [PATCH] hashed device lookup (New Benchmarks) Ben Greear
2001-01-06 21:33 ` [PATCH] hashed device lookup (Does NOT meet Linus' sumission policy!) Ben Greear
2001-01-06 23:17 ` David S. Miller
2001-01-07 4:06 ` Ben Greear
2001-01-07 13:42 ` [PATCH] hashed device lookup (Does NOT meet Linus' sumission Alan Cox
2001-01-07 15:33 ` Matti Aarnio
2001-01-07 16:46 ` Alan Cox
2001-01-07 17:32 ` Matti Aarnio
2001-01-07 19:02 ` Ben Greear
2001-01-07 18:06 ` Alan Cox
2001-01-07 18:53 ` Matti Aarnio
2001-01-07 19:30 ` Ben Greear
2001-01-07 18:30 ` Alan Cox
2001-01-07 22:40 ` 5116
2001-01-08 2:19 ` David Ford
2001-01-09 20:25 ` Christopher E. Brown
2001-01-10 2:47 ` Ben Greear
2001-01-07 18:21 ` jamal
2001-01-07 19:00 ` Matti Aarnio
2001-01-07 19:10 ` jamal
2001-01-07 19:24 ` Matti Aarnio
2001-01-08 0:21 ` jamal
2001-01-07 19:37 ` Ben Greear
2001-01-07 18:53 ` jamal
2001-01-07 5:36 ` [PATCH] hashed device lookup (Does NOT meet Linus' sumission policy!) David S. Miller
2001-01-07 3:29 ` Chris Wedgwood
2001-01-07 6:15 ` Ben Greear
2001-01-07 10:22 ` David Ford
2001-01-07 12:01 ` David S. Miller
2001-01-08 5:32 ` Andi Kleen
2001-01-08 6:12 ` Chris Wedgwood
2001-01-08 6:26 ` Andi Kleen
2001-01-08 6:57 ` David Ford
2001-01-08 13:08 ` jamal
2001-01-09 13:28 ` Blu3Viper
2001-01-08 6:13 ` Blu3Viper
2001-01-07 12:13 ` Chris Wedgwood
2001-01-07 12:19 ` David Ford
2001-01-07 16:56 ` jamal
2001-01-07 17:37 ` Gleb Natapov
2001-01-07 18:02 ` routable interfaces WAS( " jamal
2001-01-07 19:21 ` routable interfaces WAS( Re: [PATCH] hashed device lookup (DoesNOT " Ben Greear
2001-01-07 18:29 ` jamal
2001-01-07 18:51 ` Gleb Natapov
2001-01-07 19:05 ` jamal
2001-01-07 19:19 ` routable interfaces WAS( Re: [PATCH] hashed device lookup(DoesNOT " Sandy Harris
2001-01-07 20:42 ` Ben Greear
2001-01-08 0:37 ` jamal
2001-01-08 5:25 ` routable interfaces WAS( Re: [PATCH] hashed device lookup(DoesNOTmeet " Ben Greear
2001-01-08 13:05 ` jamal
2001-01-07 3:29 ` [PATCH] hashed device lookup (Does NOT meet " Andi Kleen
2001-01-07 4:00 ` jamal
2001-01-07 4:06 ` Andi Kleen
2001-01-07 5:43 ` David S. Miller
2001-01-07 11:40 ` [little bit OT] ip _IS_ _NOT_ ifconfig and route ! (was Re: [PATCH] hashed device lookup (Does NOT meet Linus' sumission policy!)) Henning P. Schmiedehausen
2001-01-07 11:50 ` David S. Miller
2001-01-07 13:47 ` [PATCH] hashed device lookup (Does NOT meet Linus' sumission Alan Cox
2001-01-07 16:12 ` jamal
2001-01-07 16:51 ` Alan Cox
2001-01-07 15:56 ` jamal [this message]
2001-01-07 16:30 ` [PATCH] hashed device lookup (Does NOT meet Linus' sumission policy!) Gleb Natapov
2001-01-07 16:36 ` jamal
2001-01-07 19:54 ` [PATCH] hashed device lookup (Does NOT meet Linus' sumissionpolicy!) Ben Greear
2001-01-07 6:24 ` Ben Greear
2001-01-07 5:29 ` Andi Kleen
2001-01-07 6:22 ` [PATCH] hashed device lookup (Does NOT meet Linus' sumission policy!) Ben Greear
2001-01-07 5:27 ` Andi Kleen
2001-01-07 8:11 ` [PATCH] hashed device lookup (Does NOT meet Linus' sumission policy!) (Benchmarks) Ben Greear
2001-01-07 7:15 ` Andi Kleen
2001-01-08 7:00 ` [PATCH] hashed device lookup (New Benchmarks) David S. Miller
2001-01-08 16:26 ` Ben Greear
2001-01-08 16:50 ` Andi Kleen
2001-01-09 16:27 ` Ben Greear
2001-01-07 13:50 ` [PATCH] hashed device lookup (Does NOT meet Linus' sumission Alan Cox
2001-01-07 16:44 ` Miquel van Smoorenburg
2001-01-07 19:09 ` Ben Greear
2001-01-07 5:40 ` [PATCH] hashed device lookup (Does NOT meet Linus' sumission policy!) David S. Miller
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=Pine.GSO.4.30.0101071026070.18916-100000@shell.cyberus.ca \
--to=hadi@cyberus.ca \
--cc=ak@suse.de \
--cc=davem@redhat.com \
--cc=greearb@candelatech.com \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@oss.sgi.com \
/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
all inboxes | Powered by JetHome®