From: Jakub Kicinski <kuba@kernel.org>
To: Alexander Lobakin <aleksander.lobakin@intel.com>
Cc: Saeed Mahameed <saeed@kernel.org>,
"David S. Miller" <davem@davemloft.net>,
Paolo Abeni <pabeni@redhat.com>,
Eric Dumazet <edumazet@google.com>,
"Saeed Mahameed" <saeedm@nvidia.com>, <netdev@vger.kernel.org>,
Tariq Toukan <tariqt@nvidia.com>, Gal Pressman <gal@nvidia.com>
Subject: Re: [net-next 1/9] net/mlx5e: Switch to using napi_build_skb()
Date: Thu, 16 Feb 2023 09:53:24 -0800 [thread overview]
Message-ID: <20230216095324.4fa4f6fb@kernel.org> (raw)
In-Reply-To: <07a89ee6-2886-65b8-d2cb-ca154f1f1f4f@intel.com>
On Thu, 16 Feb 2023 18:26:19 +0100 Alexander Lobakin wrote:
> > Before: 26.5 Gbits/sec
> > After: 30.1 Gbits/sec (+13.6%)
>
> +14%, gosh! Happy to see more and more vendors switching to it, someone
> told me back then we have so fast RAM nowadays that it won't make any
> sense to directly recycle kmem-cached objects. Maybe it's fast, but
> seems like not *so* fast :D
Interestingly I had a similar patch in my tree when testing the skb_ext
cache and enabling slow_gro kills this gain.
IOW without adding an skb_ext using napi_build_skb() gives me ~12%
boost. If I start adding skb_ext (with the cache and perfect reuse)
I'm back to the baseline (26.5Gbps in this case).
But without using napi_build_skb() adding skb_ext (with the cache)
doesn't change anything, skb_ext or not, I'll get 26.5Gbps.
Very finicky. Not sure why this happens. Perhaps napi_build_skb()
let's us fit under some CPU resource constraint and additional
functionality knocks us back over the line?
next prev parent reply other threads:[~2023-02-16 17:53 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-02-16 0:09 [pull request][net-next 0/9] mlx5 updates 2023-02-15 Saeed Mahameed
2023-02-16 0:09 ` [net-next 1/9] net/mlx5e: Switch to using napi_build_skb() Saeed Mahameed
2023-02-16 14:55 ` Maciej Fijalkowski
2023-02-16 17:26 ` Alexander Lobakin
2023-02-16 17:53 ` Jakub Kicinski [this message]
2023-02-16 17:59 ` Alexander Lobakin
2023-02-16 0:09 ` [net-next 2/9] net/mlx5e: Remove redundant page argument in mlx5e_xmit_xdp_buff() Saeed Mahameed
2023-02-16 15:04 ` Maciej Fijalkowski
2023-02-16 0:09 ` [net-next 3/9] net/mlx5e: Remove redundant page argument in mlx5e_xdp_handle() Saeed Mahameed
2023-02-16 15:13 ` Maciej Fijalkowski
2023-02-16 0:09 ` [net-next 4/9] net/mlx5: Simplify eq list traversal Saeed Mahameed
2023-02-16 15:17 ` Maciej Fijalkowski
2023-02-16 0:09 ` [net-next 5/9] net/mlx5e: Implement CT entry update Saeed Mahameed
2023-02-16 15:51 ` Maciej Fijalkowski
2023-02-16 17:15 ` Vlad Buslov
2023-02-17 11:35 ` Maciej Fijalkowski
2023-02-16 0:09 ` [net-next 6/9] net/mlx5e: Allow offloading of ct 'new' match Saeed Mahameed
2023-02-16 0:09 ` [net-next 7/9] net/mlx5e: Remove unused function mlx5e_sq_xmit_simple Saeed Mahameed
2023-02-16 0:09 ` [net-next 8/9] net/mlx5e: Fix outdated TLS comment Saeed Mahameed
2023-02-16 0:09 ` [net-next 9/9] net/mlx5e: RX, Remove doubtful unlikely call Saeed Mahameed
2023-02-16 16:05 ` Maciej Fijalkowski
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=20230216095324.4fa4f6fb@kernel.org \
--to=kuba@kernel.org \
--cc=aleksander.lobakin@intel.com \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=gal@nvidia.com \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=saeed@kernel.org \
--cc=saeedm@nvidia.com \
--cc=tariqt@nvidia.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).