netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Vladimir Kondratiev <vkondra@mail.ru>
To: netdev@oss.sgi.com
Cc: jamal <hadi@cyberus.ca>
Subject: Re: How device could assemble qdisc for self?
Date: Thu, 17 Jun 2004 00:07:53 +0300	[thread overview]
Message-ID: <200406170007.58652.vkondra@mail.ru> (raw)
In-Reply-To: <200406162357.49366.vkondra@mail.ru>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Forget to mention,
after these 6 qdiscs I described, there is round robin scheduler that do 
actually sends packets to hard_start_xmit.

On Wednesday 16 June 2004 23:57, Vladimir Kondratiev wrote:
> Hi,
> I am a bit confused looking at qdisc related stuff. Could someone please
> point me to right documentation, code to look for etc?
>
> I'd like to employ NIC with many Tx DMA queues, designated for different
> types of traffic. Jamal pointed to qdiscs as a way to do it.
>
> My current idea is as following:
>
> I need to arrange the following qdisc structure:
>
> - total 6 queues (FIFO) maintained;
>
> - 4 queues get packets accordingly to 802.1D priority tags (some simple
> fixed mapping, actually those described in TGe - standard for wireless QoS)
> Table follows, but really it does not matter.
>
> 802.1D   queue(access category)
> 1        AC_BK     Background
> 2        AC_BK
> 0        AC_BE     Best effort
> 3        AC_BE
> 4        AC_VI     Video
> 5        AC_VI
> 6        AC_VO     Voice
> 7        AC_VO
>
> - 1 queue used for all bcast/mcast packets (wireless AP need it)
>
> - 1 queue used for packets marked by application in some specific way
> (don't know how exactly, long story, next time)
>
> Each queue corresponds to one tx DMA queue within NIC. Each one should be
> stopped/started separately. Then, in hard_start_xmit, I will select proper
> DMA queue based on skb->priority. When DMA queue is full, I will stop
> corresponding qdisc.
>
> Does this structure make sense? Could it be done easier?
>
> Vladimir
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFA0Lauqxdj7mhC6o0RAlGNAJ4m9ffMI2ZQ5t74rfVnRh5pVNfpIwCfWm0D
aRk5Ox9agY1yjL6NzC8kOyI=
=k8Eo
-----END PGP SIGNATURE-----

      reply	other threads:[~2004-06-16 21:07 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2004-06-16 20:57 How device could assemble qdisc for self? Vladimir Kondratiev
2004-06-16 21:07 ` Vladimir Kondratiev [this message]

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=200406170007.58652.vkondra@mail.ru \
    --to=vkondra@mail.ru \
    --cc=hadi@cyberus.ca \
    --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;
as well as URLs for NNTP newsgroup(s).