From: Denys Fedoryshchenko <denys@visp.net.lb>
To: Andi Kleen <andi@firstfloor.org>
Cc: Joe Malicki <jmalicki@metacarta.com>,
David Miller <davem@davemloft.net>,
johnpol@2ka.mipt.ru, dada1@cosmosbay.com, netdev@vger.kernel.org,
linux-kernel@vger.kernel.org, juhlenko@akamai.com,
sammy@sammy.net
Subject: Re: loaded router, excessive getnstimeofday in oprofile
Date: Thu, 28 Aug 2008 19:48:52 +0300 [thread overview]
Message-ID: <200808281948.52608.denys@visp.net.lb> (raw)
In-Reply-To: <20080828072218.GI26610@one.firstfloor.org>
My small IMHO regarding SO_TIMESTAMP.
1)Right now i have 400-500 Mbps passing router. If i will run
5 "pings" ,simultaneous ,under _USER_ privileges(i know ping is suid),
instead of free 20% CPU time, i will have 1-2% free CPU time. Sure i know
ping is suid program, but it is has been "like this" since long time. By
security psychos it will be caled DoS.
2)Usefullness of this option. What is a difference if on almost idle machine
timestamp retrieved on higher level or lower level?
And why we need on highly loaded server so high precision timestamp (with
expensive timer), if in my case enabling any socket with SO_TIMESTAMP
creating delays more than 10ms(up to 100ms)?
3)Who is most users of SO_TIMESTAMP? iputils which is installed on almost
_ANY_ linux machine? busybox which is using same option? Many others
userspace multiplatform applications? Or banks? I dont take much in account
dhcpd, who is maybe abusing this option.
So there is few good solutions available (IMHO):
1)Introduce some SO_REALTIMESTAMP (anyway even SO_TIMESTAMP not defined in any
standard) for banks and ntp folks, who need them. And even give them timespec
instead timeval, so they will be even more happy with resolution.
2)Provide sysctl,kernel boot, or even "build time" option for "banks" to have
high resolution(and expensive) SO_TIMESTAMP.
next prev parent reply other threads:[~2008-08-28 16:50 UTC|newest]
Thread overview: 69+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-08-22 1:57 loaded router, excessive getnstimeofday in oprofile Denys Fedoryshchenko
2008-08-22 2:23 ` Denys Fedoryshchenko
2008-08-26 9:51 ` Jarek Poplawski
2008-08-26 10:29 ` Denys Fedoryshchenko
2008-08-26 10:47 ` Jarek Poplawski
2008-08-26 10:49 ` Denys Fedoryshchenko
2008-08-26 11:07 ` Jarek Poplawski
2008-08-26 11:15 ` Jarek Poplawski
2008-08-26 11:16 ` Denys Fedoryshchenko
2008-08-26 11:32 ` Jarek Poplawski
2008-08-26 11:32 ` Denys Fedoryshchenko
2008-08-26 20:14 ` Evgeniy Polyakov
2008-08-26 20:44 ` Eric Dumazet
2008-08-26 20:51 ` Evgeniy Polyakov
2008-08-27 12:09 ` Denys Fedoryshchenko
2008-08-27 12:36 ` Evgeniy Polyakov
2008-08-27 14:00 ` Denys Fedoryshchenko
2008-08-27 14:23 ` Evgeniy Polyakov
2008-08-27 12:54 ` Andi Kleen
2008-08-27 16:07 ` Rick Jones
2008-08-27 16:27 ` Andi Kleen
2008-08-27 16:49 ` Rick Jones
2008-08-27 16:56 ` Andi Kleen
2008-08-27 16:57 ` Rick Jones
2008-08-27 17:27 ` Eric Dumazet
2008-08-27 18:32 ` loaded router, excessive getnstimeofday in oprofile\ Andi Kleen
2008-08-27 22:23 ` David Miller
2008-08-27 22:38 ` Andi Kleen
2008-08-27 22:18 ` loaded router, excessive getnstimeofday in oprofile David Miller
2008-08-27 22:39 ` Andi Kleen
2008-08-28 0:45 ` Nick Piggin
2008-08-28 0:48 ` David Miller
2008-08-28 1:07 ` Nick Piggin
2008-08-27 16:17 ` Stephen Hemminger
2008-08-27 17:14 ` Jarek Poplawski
2008-08-27 21:34 ` David Miller
2008-08-28 2:39 ` Jason Uhlenkott
2008-08-28 3:10 ` David Miller
2008-08-28 6:28 ` Joe Malicki
2008-08-28 7:22 ` Andi Kleen
2008-08-28 15:02 ` Denys Fedoryshchenko
2008-08-28 19:01 ` Ilpo Järvinen
2008-08-28 19:31 ` David Miller
2008-08-28 16:48 ` Denys Fedoryshchenko [this message]
2008-08-28 16:56 ` Andi Kleen
2008-08-28 18:57 ` Eric Dumazet
2008-08-28 19:25 ` Denys Fedoryshchenko
2008-08-28 19:37 ` Eric Dumazet
2008-08-28 19:55 ` Denys Fedoryshchenko
2008-08-29 15:43 ` Stephen Hemminger
2008-08-28 19:36 ` David Miller
2008-08-28 19:59 ` Denys Fedoryshchenko
2008-08-29 15:21 ` Joe Malicki
2008-08-29 15:30 ` Andi Kleen
2008-08-29 15:43 ` Joe Malicki
2008-08-29 20:43 ` Evgeniy Polyakov
2008-08-28 18:00 ` Rick Jones
2008-08-28 19:42 ` David Miller
2008-08-28 20:29 ` Rick Jones
2008-08-28 20:32 ` David Miller
2008-08-28 20:45 ` Rick Jones
2008-08-28 20:47 ` David Miller
2008-09-01 2:39 ` Valdis.Kletnieks
2008-09-01 3:51 ` David Miller
2008-09-01 4:08 ` Valdis.Kletnieks
2008-09-01 4:10 ` David Miller
2008-09-02 17:04 ` Rick Jones
2008-08-28 3:35 ` Stephen Hemminger
2008-08-28 8:49 ` Denys Fedoryshchenko
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=200808281948.52608.denys@visp.net.lb \
--to=denys@visp.net.lb \
--cc=andi@firstfloor.org \
--cc=dada1@cosmosbay.com \
--cc=davem@davemloft.net \
--cc=jmalicki@metacarta.com \
--cc=johnpol@2ka.mipt.ru \
--cc=juhlenko@akamai.com \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=sammy@sammy.net \
/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).