netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Denys Fedoryshchenko <denys@visp.net.lb>
To: netdev@vger.kernel.org
Subject: SFQ depth limit
Date: Sat, 28 Jun 2008 22:02:59 +0300	[thread overview]
Message-ID: <200806282202.59581.denys@visp.net.lb> (raw)

Hi

Is there any particular reason to limit SFQ depth to 127 packets?
#define SFQ_DEPTH               128

Just to buffer data on 1Gbps rate for 1 second i need 87k packets, let's say just for 10ms buffer - 873 packets, which is out of limit.
Also if i will use external SFQ flow classifier made by Patrick McHardy, let's say for 4096 ip's, i am not sure also if i hit this limit. 
As i understand, if i increase it - it will increase also data structures size in memory.

And probably i hit some unknown issue.
Example:
tc qdisc add dev eth0 handle 1 root sfq
tc filter add dev eth0 protocol ip pref 1 parent 1: handle 1 flow hash keys dst divisor 1024
MegaRouterCore-KARAM ~ #tc -s qdisc show dev eth0
qdisc sfq 1: root limit 127p quantum 1514b
 Sent 12062 bytes 89 pkt (dropped 3, overlimits 0 requeues 0)
 rate 0bit 0pps backlog 0b 0p requeues 0
MegaRouterCore-KARAM ~ #tc -s qdisc show dev eth0
qdisc sfq 1: root limit 127p quantum 1514b
 Sent 12824 bytes 94 pkt (dropped 3, overlimits 0 requeues 0)
 rate 0bit 0pps backlog 0b 0p requeues 0
MegaRouterCore-KARAM ~ #tc -s qdisc show dev eth0
qdisc sfq 1: root limit 127p quantum 1514b
 Sent 13586 bytes 99 pkt (dropped 3, overlimits 0 requeues 0)
 rate 0bit 0pps backlog 0b 0p requeues 0
MegaRouterCore-KARAM ~ #tc -s -d qdisc show dev eth0
qdisc sfq 1: root limit 127p quantum 1514b flows 127/1024
 Sent 15552 bytes 114 pkt (dropped 8, overlimits 0 requeues 0)
 rate 0bit 0pps backlog 0b 0p requeues 0
MegaRouterCore-KARAM ~ #tc -s -d qdisc show dev eth0
qdisc sfq 1: root limit 127p quantum 1514b flows 127/1024
 Sent 16232 bytes 118 pkt (dropped 9, overlimits 0 requeues 0)
 rate 0bit 0pps backlog 0b 0p requeues 0

after that host is dead...



-- 
------
Technical Manager
Virtual ISP S.A.L.
Lebanon

             reply	other threads:[~2008-06-28 19:03 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-06-28 19:02 Denys Fedoryshchenko [this message]
2008-06-29 19:11 ` [PATCH] Re: SFQ depth limit Jarek Poplawski
2008-06-29 19:38   ` Denys Fedoryshchenko
2008-07-24  4:34   ` David Miller
2008-07-24  8:45     ` Patrick McHardy
2008-07-24  9:49       ` Jarek Poplawski
2008-07-24  9:57         ` Patrick McHardy
2008-07-24 10:53           ` [PATCH] pkt_sched: sch_sfq: revert "dump a real number of flows" patch Jarek Poplawski
2008-07-26  9:29             ` David Miller
2008-07-26 10:58               ` Jarek Poplawski
2008-07-27  0:24                 ` 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=200806282202.59581.denys@visp.net.lb \
    --to=denys@visp.net.lb \
    --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).