From: Jesper Dangaard Brouer <brouer@redhat.com>
To: Ioana Ciocoi Radulescu <ruxandra.radulescu@nxp.com>
Cc: Ilias Apalodimas <ilias.apalodimas@linaro.org>,
"netdev@vger.kernel.org" <netdev@vger.kernel.org>,
"davem@davemloft.net" <davem@davemloft.net>,
Ioana Ciornei <ioana.ciornei@nxp.com>,
brouer@redhat.com
Subject: Re: [PATCH net-next 1/4] dpaa2-eth: Use a single page per Rx buffer
Date: Wed, 6 Feb 2019 20:49:56 +0100 [thread overview]
Message-ID: <20190206204956.533326c2@carbon> (raw)
In-Reply-To: <AM0PR04MB49941DCEB3094958E371F960946F0@AM0PR04MB4994.eurprd04.prod.outlook.com>
On Wed, 6 Feb 2019 15:36:33 +0000 Ioana Ciocoi Radulescu <ruxandra.radulescu@nxp.com> wrote:
> > From: Ilias Apalodimas <ilias.apalodimas@linaro.org>
> >
> > Can you share any results on XDP (XDP_DROP is usually useful for the
> > hardware capabilities).
>
> XDP numbers are pretty much the same as before this patch:
>
> On a LS2088A with A72 cores @2GHz (numbers in Mpps):
> 1core 8cores
> -------------------------------------------------------------------------
> XDP_DROP (no touching data) 5.37 29.6 (linerate)
> XDP_DROP (xdp1 sample) 3.14 24.22
It is interesting/problematic to see that the cost of touching the data
is so high 5.37Mpps -> 3.14Mpps. The Intel CPUs have solved this in
hardware with DDIO, which delivers frame in L3-cache. I have some ideas
on how to improve this on ARM (or CPUs without DDIO). I've previous
implemented this as RFC on mlx4 tested on a CPU without DDIO, with
great success 10mpps -> 20Mpps (but it was shutdown, as newer Intel
HW solved the issue). The basic idea is to have an array of frames,
that you start an L2/L3-prefetch on, before going "back" and process
them for XDP or netstack. (p.s. this is the same DPDK does)
--Jesper
next prev parent reply other threads:[~2019-02-06 19:50 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-02-04 17:00 [PATCH net-next 0/4] dpaa2-eth: Driver updates Ioana Ciocoi Radulescu
2019-02-04 17:00 ` [PATCH net-next 1/4] dpaa2-eth: Use a single page per Rx buffer Ioana Ciocoi Radulescu
2019-02-05 9:35 ` Ilias Apalodimas
2019-02-06 15:36 ` Ioana Ciocoi Radulescu
2019-02-06 19:49 ` Jesper Dangaard Brouer [this message]
2019-02-08 18:19 ` Ioana Ciornei
2019-02-26 9:14 ` Ioana Ciornei
2019-02-04 17:00 ` [PATCH net-next 2/4] dpaa2-eth: Use napi_consume_skb() Ioana Ciocoi Radulescu
2019-02-04 17:00 ` [PATCH net-next 3/4] dpaa2-eth: Use FQ-based DPIO enqueue API Ioana Ciocoi Radulescu
2019-02-04 17:00 ` [PATCH net-next 4/4] dpaa2-eth: Update buffer pool refill threshold Ioana Ciocoi Radulescu
2019-02-06 18:50 ` [PATCH net-next 0/4] dpaa2-eth: Driver updates David Miller
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=20190206204956.533326c2@carbon \
--to=brouer@redhat.com \
--cc=davem@davemloft.net \
--cc=ilias.apalodimas@linaro.org \
--cc=ioana.ciornei@nxp.com \
--cc=netdev@vger.kernel.org \
--cc=ruxandra.radulescu@nxp.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).