From: Andi Kleen <ak@suse.de>
To: David Miller <davem@davemloft.net>
Cc: master@sectorb.msk.ru, hawk@diku.dk, harry@atmos.washington.edu,
linux-kernel@vger.kernel.org, netdev@vger.kernel.org
Subject: Re: Network performance degradation from 2.6.11.12 to 2.6.16.20
Date: Mon, 18 Sep 2006 16:29:53 +0200 [thread overview]
Message-ID: <200609181629.53949.ak@suse.de> (raw)
In-Reply-To: <20060918.070905.98863400.davem@davemloft.net>
>
> People who run tcpdump want "wire" timestamps as close as possible.
> Yes, things get delayed with the IRQ path, DMA delays, IRQ
> mitigation and whatnot, but it's an order of magnitude worse if
> you delay to user read() since that introduces also the delay of
> the packet copies to userspace which are significantly larger than
> these hardware level delays. If tcpdump gets swapped out, the
> timestamp delay can be on the order of several seconds making it
> totally useless.
My proposal wasn't to delay to user read, just to do the time stamp in socket
context. This means as soon as packet or RAW/UDP have looked up the socket and can
check a per socket flag do the time stamp.
The only delay this would add would be the queueing time from the NIC
to the softirq. Do you really think that is that bad?
-Andi
next prev parent reply other threads:[~2006-09-18 14:29 UTC|newest]
Thread overview: 58+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <4492D5D3.4000303@atmos.washington.edu>
2006-06-17 22:35 ` Network performance degradation from 2.6.11.12 to 2.6.16.20 Andrew Morton
2006-06-17 23:23 ` Harry Edmon
2006-06-17 23:56 ` Andrew Morton
2006-06-18 3:16 ` Stephen Hemminger
2006-06-18 23:23 ` Harry Edmon
2006-06-19 13:54 ` Harry Edmon
2006-06-20 2:11 ` Herbert Xu
2006-06-19 14:47 ` Jesper Dangaard Brouer
2006-06-19 15:24 ` Andi Kleen
2006-06-19 17:34 ` Chris Friesen
2006-06-19 20:39 ` Andi Kleen
2006-06-19 18:24 ` Jesper Dangaard Brouer
2006-06-25 21:51 ` Harry Edmon
2006-06-26 4:20 ` Bill Fink
2006-06-25 22:22 ` Willy Tarreau
2006-06-26 5:23 ` Andi Kleen
2006-07-04 11:41 ` Jesper Dangaard Brouer
2006-07-04 11:54 ` Andi Kleen
2006-07-10 10:55 ` Jesper Dangaard Brouer
2006-09-16 12:08 ` Vladimir B. Savkin
2006-09-18 8:35 ` Andi Kleen
2006-09-18 9:03 ` Vladimir B. Savkin
2006-09-18 9:58 ` Andi Kleen
2006-09-18 10:29 ` Vladimir B. Savkin
2006-09-18 11:27 ` Andi Kleen
2006-09-18 15:38 ` Alexey Kuznetsov
2006-09-18 15:54 ` Andi Kleen
2006-09-18 16:28 ` Alexey Kuznetsov
2006-09-18 16:50 ` Andi Kleen
2006-09-18 21:03 ` Alexey Kuznetsov
2006-09-18 21:22 ` David Miller
2006-09-18 21:46 ` Alexey Kuznetsov
2006-09-19 5:55 ` Andi Kleen
2006-09-19 20:31 ` Thomas Graf
2006-09-19 20:43 ` Andi Kleen
2006-09-19 5:52 ` Andi Kleen
2006-09-18 21:18 ` Vladimir B. Savkin
2006-09-18 22:00 ` Alexey Kuznetsov
2006-09-18 21:57 ` David Lang
2006-09-19 19:40 ` David Miller
2006-09-19 19:44 ` Stephen Hemminger
2006-09-18 22:03 ` Vladimir B. Savkin
2006-09-19 19:41 ` David Miller
2006-09-19 19:47 ` David Miller
2006-09-22 15:35 ` Alexey Kuznetsov
2006-09-22 15:43 ` Andi Kleen
2006-09-22 16:51 ` Rick Jones
2007-03-06 13:25 ` Packet timestamps (was: Re: Network performance degradation from 2.6.11.12 to 2.6.16.20) Vladimir B. Savkin
2007-03-06 14:38 ` Eric Dumazet
2007-03-06 14:43 ` Vladimir B. Savkin
2007-03-06 15:16 ` Eric Dumazet
2007-03-06 18:15 ` Vladimir B. Savkin
2006-09-18 21:08 ` Network performance degradation from 2.6.11.12 to 2.6.16.20 Vladimir B. Savkin
2006-09-18 14:09 ` David Miller
2006-09-18 14:29 ` Andi Kleen [this message]
2006-09-18 15:19 ` Alan Cox
2006-09-18 15:19 ` Andi Kleen
2006-06-19 16:40 ` Harry Edmon
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=200609181629.53949.ak@suse.de \
--to=ak@suse.de \
--cc=davem@davemloft.net \
--cc=harry@atmos.washington.edu \
--cc=hawk@diku.dk \
--cc=linux-kernel@vger.kernel.org \
--cc=master@sectorb.msk.ru \
--cc=netdev@vger.kernel.org \
/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).