From: Leon Romanovsky <leon@kernel.org>
To: Steffen Klassert <steffen.klassert@secunet.com>
Cc: Jakub Kicinski <kuba@kernel.org>,
Jason Gunthorpe <jgg@nvidia.com>,
"David S . Miller" <davem@davemloft.net>,
Herbert Xu <herbert@gondor.apana.org.au>,
netdev@vger.kernel.org, Raed Salem <raeds@nvidia.com>,
ipsec-devel <devel@linux-ipsec.org>
Subject: Re: [PATCH xfrm-next v2 0/6] Extend XFRM core to allow full offload configuration
Date: Mon, 22 Aug 2022 11:54:42 +0300 [thread overview]
Message-ID: <YwNEUguW7aTXC2Vs@unreal> (raw)
In-Reply-To: <20220822084105.GI2602992@gauss3.secunet.de>
On Mon, Aug 22, 2022 at 10:41:05AM +0200, Steffen Klassert wrote:
> On Fri, Aug 19, 2022 at 10:53:56AM -0700, Jakub Kicinski wrote:
> > On Fri, 19 Aug 2022 13:01:22 -0300 Jason Gunthorpe wrote:
> > > Regardless, RDMA doesn't really intersect with this netdev work for
> > > XFRM beyond the usual ways that RDMA IP traffic can be captured by or
> > > run parallel to netdev.
> > >
> > > A significant use case here is for switchdev modes where the switch
> > > will subject traffic from a switch port to ESP, not unlike it already
> > > does with vlan, vxlan, etc and other already fully offloaded switching
> > > transforms.
> >
> > Yup, that's what I thought you'd say. Can't argue with that use case
> > if Steffen is satisfied with the technical aspects.
>
> Yes, everything that can help to overcome the performance problems
> can help and I'm interested in this type of offload. But we need to
> make sure the API is usable by the whole community, so I don't
> want an API for some special case one of the NIC vendors is
> interested in.
BTW, we have a performance data, I planned to send it as part of cover
letter for v3, but it is worth to share it now.
================================================================================
Performance results:
TCP multi-stream, using iperf3 instance per-CPU.
+----------------------+--------+--------+--------+--------+---------+---------+
| | 1 CPU | 2 CPUs | 4 CPUs | 8 CPUs | 16 CPUs | 32 CPUs |
| +--------+--------+--------+--------+---------+---------+
| | BW (Gbps) |
+----------------------+--------+--------+-------+---------+---------+---------+
| Baseline | 27.9 | 59 | 93.1 | 92.8 | 93.7 | 94.4 |
+----------------------+--------+--------+-------+---------+---------+---------+
| Software IPsec | 6 | 11.9 | 23.3 | 45.9 | 83.8 | 91.8 |
+----------------------+--------+--------+-------+---------+---------+---------+
| IPsec crypto offload | 15 | 29.7 | 58.5 | 89.6 | 90.4 | 90.8 |
+----------------------+--------+--------+-------+---------+---------+---------+
| IPsec full offload | 28 | 57 | 90.7 | 91 | 91.3 | 91.9 |
+----------------------+--------+--------+-------+---------+---------+---------+
IPsec full offload mode behaves as baseline and reaches linerate with same amount
of CPUs.
Setups details (similar for both sides):
* NIC: ConnectX6-DX dual port, 100 Gbps each.
Single port used in the tests.
* CPU: Intel(R) Xeon(R) Platinum 8380 CPU @ 2.30GHz
Thanks
next prev parent reply other threads:[~2022-08-22 8:54 UTC|newest]
Thread overview: 45+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-08-16 8:59 [PATCH xfrm-next v2 0/6] Extend XFRM core to allow full offload configuration Leon Romanovsky
2022-08-16 8:59 ` [PATCH xfrm-next v2 1/6] xfrm: add new full offload flag Leon Romanovsky
2022-08-16 8:59 ` [PATCH xfrm-next v2 2/6] xfrm: allow state full offload mode Leon Romanovsky
2022-08-18 10:12 ` Steffen Klassert
2022-08-18 13:28 ` Leon Romanovsky
2022-08-22 8:01 ` Steffen Klassert
2022-08-22 8:46 ` Leon Romanovsky
2022-08-16 8:59 ` [PATCH xfrm-next v2 3/6] xfrm: add an interface to offload policy Leon Romanovsky
2022-08-16 8:59 ` [PATCH xfrm-next v2 4/6] xfrm: add TX datapath support for IPsec full offload mode Leon Romanovsky
2022-08-18 10:24 ` Steffen Klassert
2022-08-18 13:34 ` Leon Romanovsky
2022-08-22 8:04 ` Steffen Klassert
2022-08-22 8:50 ` Leon Romanovsky
2022-08-16 8:59 ` [PATCH xfrm-next v2 5/6] xfrm: add RX datapath protection " Leon Romanovsky
2022-08-18 10:27 ` Steffen Klassert
2022-08-18 13:36 ` Leon Romanovsky
2022-08-22 8:06 ` Steffen Klassert
2022-08-22 9:35 ` Leon Romanovsky
2022-08-16 8:59 ` [PATCH xfrm-next v2 6/6] xfrm: enforce separation between priorities of HW/SW policies Leon Romanovsky
2022-08-17 2:54 ` [PATCH xfrm-next v2 0/6] Extend XFRM core to allow full offload configuration Jakub Kicinski
2022-08-17 5:22 ` Leon Romanovsky
2022-08-17 18:10 ` Jakub Kicinski
2022-08-18 5:24 ` Leon Romanovsky
2022-08-18 10:10 ` Steffen Klassert
2022-08-18 12:51 ` Leon Romanovsky
2022-08-19 1:54 ` Jakub Kicinski
2022-08-19 2:34 ` Jakub Kicinski
2022-08-19 5:52 ` Leon Romanovsky
2022-08-19 15:47 ` Jakub Kicinski
2022-08-19 16:01 ` Jason Gunthorpe
2022-08-19 17:53 ` Jakub Kicinski
2022-08-22 8:41 ` Steffen Klassert
2022-08-22 8:54 ` Leon Romanovsky [this message]
2022-08-22 16:33 ` Jakub Kicinski
2022-08-22 21:27 ` Saeed Mahameed
2022-08-23 0:17 ` Jakub Kicinski
2022-08-23 5:22 ` Steffen Klassert
2022-08-23 14:06 ` Leon Romanovsky
2022-08-23 4:48 ` Leon Romanovsky
2022-08-26 12:20 ` Jason Gunthorpe
2022-08-23 5:34 ` Leon Romanovsky
2022-08-18 10:09 ` Steffen Klassert
2022-08-18 13:26 ` Leon Romanovsky
2022-08-22 8:34 ` Steffen Klassert
2022-08-22 9:34 ` Leon Romanovsky
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=YwNEUguW7aTXC2Vs@unreal \
--to=leon@kernel.org \
--cc=davem@davemloft.net \
--cc=devel@linux-ipsec.org \
--cc=herbert@gondor.apana.org.au \
--cc=jgg@nvidia.com \
--cc=kuba@kernel.org \
--cc=netdev@vger.kernel.org \
--cc=raeds@nvidia.com \
--cc=steffen.klassert@secunet.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).