From: Willy Tarreau <w@1wt.eu>
To: David Laight <David.Laight@ACULAB.COM>
Cc: "davem@davemloft.net" <davem@davemloft.net>,
"netdev@vger.kernel.org" <netdev@vger.kernel.org>,
Thomas Petazzoni <thomas.petazzoni@free-electrons.com>,
Gregory CLEMENT <gregory.clement@free-electrons.com>
Subject: Re: [PATCH 11/13] net: mvneta: implement rx_copybreak
Date: Thu, 16 Jan 2014 10:36:40 +0100 [thread overview]
Message-ID: <20140116093640.GA16457@1wt.eu> (raw)
In-Reply-To: <063D6719AE5E284EB5DD2968C1650D6D45D866@AcuExch.aculab.com>
On Thu, Jan 16, 2014 at 09:14:00AM +0000, David Laight wrote:
> From: Of Willy Tarreau
> > calling dma_map_single()/dma_unmap_single() is quite expensive compared
> > to copying a small packet. So let's copy short frames and keep the buffers
> > mapped. We set the limit to 256 bytes which seems to give good results both
> > on the XP-GP board and on the AX3/4.
>
> Have you tried something similar for transmits?
> Copying short tx into a preallocated memory area that is dma_mapped
> at driver initialisation?
Not yet by lack of time, but I intend to do so. Both the NIC and the driver
are capable of running at line rate (1.488 Mpps) when filling descriptors
directly, so many improvements are still possible.
Regards,
Willy
next prev parent reply other threads:[~2014-01-16 9:36 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-01-16 7:20 [PATCH 00/13] Assorted mvneta fixes and improvements Willy Tarreau
2014-01-16 7:20 ` [PATCH 01/13] net: mvneta: increase the 64-bit rx/tx stats out of the hot path Willy Tarreau
2014-01-16 7:20 ` [PATCH 02/13] net: mvneta: use per_cpu stats to fix an SMP lock up Willy Tarreau
2014-01-16 7:20 ` [PATCH 03/13] net: mvneta: do not schedule in mvneta_tx_timeout Willy Tarreau
2014-01-16 7:20 ` [PATCH 04/13] net: mvneta: add missing bit descriptions for interrupt masks and causes Willy Tarreau
2014-01-16 7:20 ` [PATCH 05/13] net: mvneta: replace Tx timer with a real interrupt Willy Tarreau
2014-01-16 7:20 ` [PATCH 06/13] net: mvneta: remove tests for impossible cases in the tx_done path Willy Tarreau
2014-01-16 7:20 ` [PATCH 07/13] net: mvneta: factor rx refilling code Willy Tarreau
2014-01-16 7:20 ` [PATCH 08/13] net: mvneta: simplify access to the rx descriptor status Willy Tarreau
2014-01-16 7:20 ` [PATCH 09/13] net: mvneta: prefetch next rx descriptor instead of current one Willy Tarreau
2014-01-16 7:20 ` [PATCH 10/13] net: mvneta: convert to build_skb() Willy Tarreau
2014-01-16 7:20 ` [PATCH 11/13] net: mvneta: implement rx_copybreak Willy Tarreau
2014-01-16 9:14 ` David Laight
2014-01-16 9:36 ` Willy Tarreau [this message]
2014-01-16 19:49 ` David Miller
2014-01-16 20:07 ` Willy Tarreau
2014-01-16 20:11 ` David Miller
2014-01-17 9:28 ` David Laight
2014-01-17 9:48 ` Willy Tarreau
2014-01-17 9:32 ` David Laight
2014-01-16 7:20 ` [PATCH 12/13] net: mvneta: mvneta_tx_done_gbe() cleanups Willy Tarreau
2014-01-16 7:20 ` [PATCH 13/13] net: mvneta: make mvneta_txq_done() return void Willy Tarreau
2014-01-16 7:22 ` [PATCH net-next 00/13] Assorted mvneta fixes and improvements Willy Tarreau
2014-01-16 23:21 ` [PATCH " 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=20140116093640.GA16457@1wt.eu \
--to=w@1wt.eu \
--cc=David.Laight@ACULAB.COM \
--cc=davem@davemloft.net \
--cc=gregory.clement@free-electrons.com \
--cc=netdev@vger.kernel.org \
--cc=thomas.petazzoni@free-electrons.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).