From: Hans de Goede <j.w.r.degoede@hhs.nl>
To: Eric Dumazet <dada1@cosmosbay.com>
Cc: Dave Jones <davej@redhat.com>, netdev@vger.kernel.org
Subject: Re: cat /proc/net/tcp takes 0.5 seconds on x86_64
Date: Tue, 26 Aug 2008 21:01:01 +0200 [thread overview]
Message-ID: <48B452ED.1000308@hhs.nl> (raw)
In-Reply-To: <48B44C3F.6020006@cosmosbay.com>
Eric Dumazet wrote:
> Dave Jones a écrit :
>> Just had this bug reported against our development tree..
<snip>
>> > [hans@localhost devel]$ time cat /proc/net/tcp
>> > <snip>
>> > real 0m0.520s
>> > user 0m0.000s
>> > sys 0m0.446s
>> > > Thats amazingly slow, esp as I only have 8 tcp connections open.
>> > > Some maybe usefull info: top reports a very high load (50%) from
>> soft IRQ's.
>> > > Anyways changing this to a kernel bug.
>>
>
> I wonder why this qualifies as a "kernel bug". This is a well known
> problem.
>
No its not, /proc/net/tcp may be slow in general but not *this* slow ...
<snip>
>
> Time difference between /proc/net/tcp and netlink on a 4GB x86_64 machine :
>
> # dmesg | grep "TCP established hash"
> TCP established hash table entries: 262144 (order: 10, 4194304 bytes)
> # time cat /proc/net/tcp >/dev/null
>
> real 0m0.091s
> user 0m0.001s
> sys 0m0.090s
As quoted above my idle x86_64, using the exact same hash table size, running
2.6.27-rc2.git1 uses 0.520 seconds for that same command, thats a difference of
more then a factor 50 !!
This is not about /proc/net/tcp not being fast, this is about it haven gotten
slower by a factor of 50!
Also notice that this slowdown does not happen on i386.
Anyways I'll try 2.6.27-rc4 and report back with its results.
Regards,
Hans
next prev parent reply other threads:[~2008-08-26 18:49 UTC|newest]
Thread overview: 39+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <bug-459782-176318@bugzilla.redhat.com>
[not found] ` <200808261549.m7QFnVUN032543@bz-web1.app.phx.redhat.com>
2008-08-26 16:37 ` cat /proc/net/tcp takes 0.5 seconds on x86_64 Dave Jones
2008-08-26 18:32 ` Eric Dumazet
2008-08-26 19:01 ` Hans de Goede [this message]
2008-08-26 20:39 ` Eric Dumazet
2008-08-26 20:58 ` Hans de Goede
2008-08-26 21:27 ` Eric Dumazet
2008-08-27 9:14 ` Hans de Goede
2008-08-27 9:05 ` David Miller
2008-08-27 9:45 ` Hans de Goede
2008-08-27 9:39 ` David Miller
2008-08-27 4:19 ` Herbert Xu
2008-08-27 9:07 ` Hans de Goede
2008-08-27 12:41 ` Andi Kleen
2008-08-27 21:29 ` Trent Piepho
2008-08-27 21:47 ` Andi Kleen
2008-08-27 22:54 ` Andi Kleen
2008-08-27 21:29 ` David Miller
2008-08-27 21:48 ` Stephen Hemminger
2008-08-27 22:09 ` David Miller
2008-08-28 6:20 ` Eric Dumazet
2008-08-28 6:51 ` David Miller
2008-08-28 7:13 ` Eric Dumazet
2008-08-28 7:57 ` David Miller
2008-08-28 9:52 ` Eric Dumazet
2008-08-28 7:26 ` Andi Kleen
2008-08-27 22:34 ` Andi Kleen
2008-08-27 22:39 ` David Miller
2008-08-27 22:57 ` Andi Kleen
2008-08-27 23:07 ` David Miller
2008-08-27 23:09 ` Eric Dumazet
2008-08-27 23:15 ` David Miller
2008-08-27 23:35 ` Andi Kleen
2008-08-27 23:43 ` Eric Dumazet
2008-08-27 23:45 ` David Miller
2008-08-28 0:40 ` Eric Dumazet
2008-08-28 7:45 ` Andi Kleen
2008-08-28 7:59 ` David Miller
2008-08-28 8:12 ` Hans de Goede
2008-08-28 8:04 ` David 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=48B452ED.1000308@hhs.nl \
--to=j.w.r.degoede@hhs.nl \
--cc=dada1@cosmosbay.com \
--cc=davej@redhat.com \
--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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.