All of lore.kernel.org
 help / color / mirror / Atom feed
From: Vinicius Costa Gomes <vinicius.gomes@intel.com>
To: "Stéphane Ancelot" <sancelot@numalliance.com>, netdev@vger.kernel.org
Subject: Re: TSN - tc usage for a tbs setup
Date: Thu, 25 Jul 2019 10:30:58 -0700	[thread overview]
Message-ID: <87ef2ef2ct.fsf@intel.com> (raw)
In-Reply-To: <43c8c7bd-f281-a4dc-badc-9672aaccbd6a@numalliance.com>

Hi,

Stéphane Ancelot <sancelot@numalliance.com> writes:

> Hi,
>
> I am trying to setup my network queue for offline time based configuration.
>
> initial setup is :
>
> tc qdisc show dev eth1:
>
> qdisc mq 0: root
>
> qdisc pfifo_fast 0: parent :1 bands 3 priomap  1 2 2 2 1 2 0 0 1 1 1 1 1 
> 1 1 1
>
>
> I won't need pfifo , I have to send one frame at a precise xmit time 
> (high prio), and then maybe some other frames (with low priority)
>
>
> I want to setup offload time based  xmit.
>
> /sbin/tc qdisc add dev eth1 root handle 100:1 etf delta 100000 clockid 
> CLOCK_REALTIME offload
>

Because the common (expected?) use case for ETF is using it on a system
that is running ptp4l (for example), and so, has the NIC PHC clock using
the TAI clock reference, we only accept the clockid to be CLOCK_TAI.
(Perhaps you are using an old version of iproute2, because a clearer
message should have been printed together with the error as well, anyway
there should be something in dmesg too)

That said, when I need to run some experiments with ETF, and do not care
about having the PHC clock is sync with anything else, I use phc2sys to
force the TAI offset to be zero. Something like this:

$ phc2sys -c $IFACE -s CLOCK_REALTIME -O 0 -m

And install ETF as "usual", something like this:

$ tc qdisc add dev $IFACE root handle 100:1 etf delta 100000 clockid CLOCK_TAI offload

Hope this helps.


Cheers,
--
Vinicius

      reply	other threads:[~2019-07-25 17:31 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-07-25  6:50 TSN - tc usage for a tbs setup Stéphane Ancelot
2019-07-25 17:30 ` Vinicius Costa Gomes [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=87ef2ef2ct.fsf@intel.com \
    --to=vinicius.gomes@intel.com \
    --cc=netdev@vger.kernel.org \
    --cc=sancelot@numalliance.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.