From: Jakub Kicinski <jakub.kicinski@netronome.com>
To: Alexei Starovoitov <alexei.starovoitov@gmail.com>
Cc: netdev@vger.kernel.org, ast@kernel.org, daniel@iogearbox.net,
dinan.gunawardena@netronome.com
Subject: Re: [RFC 12/12] nfp: bpf: add denser mode of execution
Date: Wed, 1 Jun 2016 23:47:23 +0100 [thread overview]
Message-ID: <20160601234723.098ef2af@jkicinski-Precision-T1700> (raw)
In-Reply-To: <20160601220114.GC24671@ast-mbp.thefacebook.com>
On Wed, 1 Jun 2016 15:01:16 -0700, Alexei Starovoitov wrote:
> On Wed, Jun 01, 2016 at 05:50:14PM +0100, Jakub Kicinski wrote:
> > If BPF uses less than 7 registers programmable engines
> > can process twice as many packets in parallel. Signal
> > this denser mode of operation to FW by setting the lowest
> > bit in DMA address of the machine code buffer.
> >
> > Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
> > Reviewed-by: Dinan Gunawardena <dgunawardena@netronome.com>
> > Reviewed-by: Simon Horman <simon.horman@netronome.com>
>
> wow. that sounds pretty cool.
> I was about to ask whether we can help HW to be more efficient
> by doing something on bpf side like annotating the registers or
> adding 'hw_only' registers...
> but looks like less registers is actually good, since NFP jit
> can parallelize it? Truly wow.
> if you can share the hw architecture details and explain more
> on how this 'dense_mode' works, would be awesome.
I think the best resource of information about the card and its
internals would be the open-nfp website [1]. Regarding optimizations
there are definitely things which could be done. The most obvious is
helping 32bit JITs in general. Annotations which registers are
32bit-only, which operations need zero-extending etc. There is some
state in the verifier that would be useful here as well. Knowing which
registers contain the skb pointer for instance would help to ignore any
operations on them (since as exemplified by simple optimization in
patch 10 the skb pointer has no meaning for the HW). I'm sure more
such things will come up with time.
[1] http://open-nfp.org/resources/
prev parent reply other threads:[~2016-06-01 22:47 UTC|newest]
Thread overview: 54+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-06-01 16:50 [RFC 00/12] BPF hardware offload via cls_bpf Jakub Kicinski
2016-06-01 16:50 ` [RFC 01/12] add basic register-field manipulation macros Jakub Kicinski
2016-06-01 20:15 ` Hannes Frederic Sowa
2016-06-01 23:08 ` Jakub Kicinski
2016-06-02 12:01 ` Hannes Frederic Sowa
2016-06-01 16:50 ` [RFC 02/12] net: cls_bpf: add hardware offload Jakub Kicinski
2016-06-01 17:13 ` John Fastabend
2016-06-01 20:59 ` Jakub Kicinski
2016-06-01 19:34 ` Daniel Borkmann
2016-06-02 7:17 ` Jiri Pirko
2016-06-02 12:07 ` Jakub Kicinski
2016-06-01 16:50 ` [RFC 03/12] net: cls_bpf: limit hardware offload by software-only flag Jakub Kicinski
2016-06-01 17:16 ` John Fastabend
2016-06-01 17:16 ` John Fastabend
2016-06-01 19:40 ` Daniel Borkmann
2016-06-01 21:05 ` Jakub Kicinski
2016-06-01 21:21 ` Daniel Borkmann
2016-06-01 21:26 ` Jakub Kicinski
2016-06-01 21:31 ` Daniel Borkmann
2016-06-02 7:24 ` Jiri Pirko
2016-06-01 16:50 ` [RFC 04/12] net: cls_bpf: add support for marking filters as hardware-only Jakub Kicinski
2016-06-01 17:19 ` John Fastabend
2016-06-01 19:57 ` Daniel Borkmann
2016-06-01 16:50 ` [RFC 05/12] nfp: add BPF to NFP code translator Jakub Kicinski
2016-06-01 20:03 ` Daniel Borkmann
2016-06-01 20:09 ` John Fastabend
2016-06-01 20:15 ` Alexei Starovoitov
2016-06-01 21:23 ` Jakub Kicinski
2016-06-02 16:21 ` John Fastabend
2016-06-01 16:50 ` [RFC 06/12] nfp: add hardware cls_bpf offload Jakub Kicinski
2016-06-01 20:20 ` Daniel Borkmann
2016-06-01 20:52 ` Alexei Starovoitov
2016-06-01 21:15 ` Jakub Kicinski
2016-06-01 21:51 ` Alexei Starovoitov
2016-06-01 21:16 ` Daniel Borkmann
2016-06-01 21:36 ` John Fastabend
2016-06-02 6:57 ` Jiri Pirko
2016-06-02 12:13 ` Jakub Kicinski
2016-06-02 12:30 ` Daniel Borkmann
2016-06-01 23:03 ` Daniel Borkmann
2016-06-01 16:50 ` [RFC 07/12] nfp: add skb mark support to the bpf offload Jakub Kicinski
2016-06-01 21:56 ` Alexei Starovoitov
2016-06-01 22:19 ` Jakub Kicinski
2016-06-01 22:30 ` Daniel Borkmann
2016-06-01 23:01 ` Jakub Kicinski
2016-06-01 16:50 ` [RFC 08/12] net: cls_bpf: allow offloaded filters to update stats Jakub Kicinski
2016-06-01 17:20 ` John Fastabend
2016-06-01 22:09 ` Daniel Borkmann
2016-06-01 16:50 ` [RFC 09/12] nfp: report statistics of offloaded filters Jakub Kicinski
2016-06-01 16:50 ` [RFC 10/12] nfp: bpf: optimize register init Jakub Kicinski
2016-06-01 16:50 ` [RFC 11/12] nfp: bpf: add register rename Jakub Kicinski
2016-06-01 16:50 ` [RFC 12/12] nfp: bpf: add denser mode of execution Jakub Kicinski
2016-06-01 22:01 ` Alexei Starovoitov
2016-06-01 22:47 ` Jakub Kicinski [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=20160601234723.098ef2af@jkicinski-Precision-T1700 \
--to=jakub.kicinski@netronome.com \
--cc=alexei.starovoitov@gmail.com \
--cc=ast@kernel.org \
--cc=daniel@iogearbox.net \
--cc=dinan.gunawardena@netronome.com \
--cc=netdev@vger.kernel.org \
/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