netdev - Re: [PATCH net-next v2 1/2] pkt_sched: fq: avoid artificial bursts for clocked flows

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <1423075654.31870.4.camel@edumazet-glaptop2.roam.corp.google.com>
Date:	Wed, 04 Feb 2015 10:47:34 -0800
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	Kenneth Klette Jonassen <kennetkl@....uio.no>
Cc:	"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
	David Miller <davem@...emloft.net>
Subject: Re: [PATCH net-next v2 1/2] pkt_sched: fq: avoid artificial bursts
 for clocked flows

On Wed, 2015-02-04 at 17:56 +0000, Kenneth Klette Jonassen wrote:

> 
> If it is a code issue, please refactor it or point out some specifics.
> 
> Thanks for cc’ing me on your related patch (“do not pace pure ack
> packets”). Note that your patch does not address the core issue that
> this patch attempts to fix: fq throttling is just broken for small,
> clocked flows. This not only concerns TCP/SCTP/ARQ acks, but also
> thin-streams like VoIP and gaming.


I already told you this patch was not wanted.

We want a reasonably efficient packet scheduler, allowing TCP pacing,
at the scale of million of flows.

You can reduce quantum to whatever you need.

We have an internal patch doing something really different,
not adding 8 bytes overhead in the flow structure.

We will upstream it when fully tested.

commit 799735e5d08b3a7fc19176eee316ca647c9bad32
Author: Eric Dumazet <edumazet@...gle.com>
Date:   Mon Aug 11 14:35:36 2014 -0700

    net-sched-fq: special case low rate flows
    
    Effort: net-sched-fq
    Change-Id: Ibee17453702cccf729154f6b4db7aeb028fc555a



--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html