[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <cb6a3d87-756f-1d97-a29d-8535796cf652@gmail.com>
Date: Mon, 11 Mar 2019 08:41:23 -0700
From: Eric Dumazet <eric.dumazet@...il.com>
To: Olga Albisser <olgabnd@...il.com>, netdev@...r.kernel.org
Cc: Olga Albisser <olga@...isser.org>,
Koen De Schepper <koen.de_schepper@...ia-bell-labs.com>,
Oliver Tilmans <olivier.tilmans@...ia-bell-labs.com>,
Bob Briscoe <research@...briscoe.net>,
Henrik Steen <henrist@...rist.net>
Subject: Re: [PATCH net-next] sched: add dualpi2 scheduler module
FYI, net-next tree is currently closed.
On 03/11/2019 08:14 AM, Olga Albisser wrote:
> DUALPI2 provides extremely low latency & loss to traffic that uses a
> scalable congestion controller (e.g. L4S, DCTCP) without degrading the
> performance of 'classic' traffic (e.g. Reno, Cubic etc.). It is intended
> to be the reference implementation of the IETF's DualQ Coupled AQM.
> +
> +static int dualpi2_qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch,
> + struct sk_buff **to_free)
> +{
> + struct dualpi2_sched_data *q = qdisc_priv(sch);
> + u32 ecn = get_ecn_field(skb);
> + int err;
> +
> + /* set to the time the HTQ packet is in the Q */
> + __net_timestamp(skb);
> +
You can not do that in a qdisc ( blindly overwriting skb->tstamp )
EDT model was adopted in TCP in linux-4.20
You will need to use part of skb->cb instead.
BTW, prefer CLOCK_MONOTONIC, this can avoid all sort of surprises...
Powered by blists - more mailing lists