netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jesper Dangaard Brouer <brouer@redhat.com>
To: Tom Herbert <tom@herbertland.com>
Cc: Thomas Graf <tgraf@suug.ch>,
	"David S. Miller" <davem@davemloft.net>,
	Linux Kernel Network Developers <netdev@vger.kernel.org>,
	Kernel Team <kernel-team@fb.com>,
	Tariq Toukan <tariqt@mellanox.com>,
	Brenden Blanco <bblanco@plumgrid.com>,
	Alexei Starovoitov <alexei.starovoitov@gmail.com>,
	Eric Dumazet <eric.dumazet@gmail.com>,
	brouer@redhat.com
Subject: Re: [PATCH RFC 1/3] xdp: Infrastructure to generalize XDP
Date: Thu, 22 Sep 2016 15:14:03 +0200	[thread overview]
Message-ID: <20160922151403.24648381@redhat.com> (raw)
In-Reply-To: <20160921215658.2c61ed5e@redhat.com>

On Wed, 21 Sep 2016 21:56:58 +0200
Jesper Dangaard Brouer <brouer@redhat.com> wrote:

> > > I'm not opposed to running non-BPF code at XDP. I'm against adding
> > > a linked list of hook consumers.  
> 
> I also worry about the performance impact of a linked list.  We should
> simple benchmark it instead of discussing it! ;-)

(Note, there are some stability issue with this RFC patchset, when
removing the xdp program, that I had to workaround/patch)


I've started benchmarking this and I only see added cost of 2.89ns from
these patches, at these crazy speeds it does correspond to -485Kpps.

 I was really expecting to see a higher cost of this approach.

I tested this on two different machines. One was suppose to work with
DDIO, but I could not get DDIO working on that machine (result in max
12.7Mpps drop).  Even-though the mlx5 card does work with DDIO.  Even
removed the mlx5 and used same slot but no luck.   (A side-note: Also
measured a 16ns performance difference between which PCIe slot I'm
using).

The reason I wanted to benchmark this on a DDIO machine is, that I'm
suspecting that the added cost, could be hiding behind the cache miss.

Well, I'm running out-of-time benchmarking this stuff, I must prepare
for my Network Performance Workshop ;-)


(A side-note: my skylake motherboard also had a PCI slot, so I found an
old e1000 NIC in my garage, and it worked!)
-- 
Best regards,
  Jesper Dangaard Brouer
  MSc.CS, Principal Kernel Engineer at Red Hat
  Author of http://www.iptv-analyzer.org
  LinkedIn: http://www.linkedin.com/in/brouer

  reply	other threads:[~2016-09-22 13:14 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-09-20 22:00 [PATCH RFC 0/3] xdp: Generalize XDP Tom Herbert
2016-09-20 22:00 ` [PATCH RFC 1/3] xdp: Infrastructure to generalize XDP Tom Herbert
2016-09-20 22:37   ` Eric Dumazet
2016-09-20 22:40     ` Tom Herbert
2016-09-20 22:44   ` Thomas Graf
2016-09-20 22:49     ` Tom Herbert
2016-09-20 23:09       ` Thomas Graf
2016-09-20 23:18         ` Tom Herbert
2016-09-20 23:43           ` Thomas Graf
2016-09-20 23:59             ` Tom Herbert
2016-09-21  0:13               ` Alexei Starovoitov
2016-09-21 11:55               ` Thomas Graf
2016-09-21 14:19                 ` Tom Herbert
2016-09-21 14:48                   ` Thomas Graf
2016-09-21 15:08                     ` Tom Herbert
2016-09-21 19:56                       ` Jesper Dangaard Brouer
2016-09-22 13:14                         ` Jesper Dangaard Brouer [this message]
2016-09-22 14:46                           ` Eric Dumazet
2016-09-21 15:39                   ` Alexei Starovoitov
2016-09-21 17:26                 ` Jakub Kicinski
2016-09-20 23:22         ` Daniel Borkmann
2016-09-21  0:01   ` Alexei Starovoitov
2016-09-21  6:39     ` Jesper Dangaard Brouer
2016-09-21  8:42       ` Daniel Borkmann
2016-09-21 15:44       ` Alexei Starovoitov
2016-09-21 17:26     ` Jakub Kicinski
2016-09-21 17:39       ` Tom Herbert
2016-09-21 18:45         ` Jakub Kicinski
2016-09-21 18:50           ` Tom Herbert
2016-09-21 18:54             ` Jakub Kicinski
2016-09-21 18:58             ` Thomas Graf
2016-09-23 11:13   ` Jamal Hadi Salim
2016-09-23 13:00     ` Jesper Dangaard Brouer
2016-09-23 14:26       ` Alexei Starovoitov
2016-09-25 11:32       ` Jamal Hadi Salim
2016-09-23 14:14     ` Tom Herbert
2016-09-25 12:29       ` Jamal Hadi Salim
2016-09-20 22:00 ` [PATCH RFC 2/3] mlx4: Change XDP/BPF to use generic XDP infrastructure Tom Herbert
2016-09-20 22:00 ` [PATCH RFC 3/3] netdevice: Remove obsolete xdp_netdev_command Tom Herbert

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=20160922151403.24648381@redhat.com \
    --to=brouer@redhat.com \
    --cc=alexei.starovoitov@gmail.com \
    --cc=bblanco@plumgrid.com \
    --cc=davem@davemloft.net \
    --cc=eric.dumazet@gmail.com \
    --cc=kernel-team@fb.com \
    --cc=netdev@vger.kernel.org \
    --cc=tariqt@mellanox.com \
    --cc=tgraf@suug.ch \
    --cc=tom@herbertland.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).