From: Ben Hutchings <bhutchings@solarflare.com>
To: Bill Fink <billfink@mindspring.com>
Cc: Aviv Greenberg <avivgnet@gmail.com>,
Andrew Gallatin <gallatin@myri.com>,
netdev@vger.kernel.org
Subject: Re: Receive side performance issue with multi-10-GigE and NUMA
Date: Thu, 20 Aug 2009 14:14:16 +0100 [thread overview]
Message-ID: <1250774056.16001.157.camel@localhost> (raw)
In-Reply-To: <20090820032601.03a628cf.billfink@mindspring.com>
On Thu, 2009-08-20 at 03:26 -0400, Bill Fink wrote:
> On Sat, 15 Aug 2009, Aviv Greenberg wrote:
>
> > > There may be something in the chipset
> >
> > shooting in the dark: when you lspci -vvv and check the MaxPayload and
> > MaxReadReq values for the myri devices - what are the values and are
> > they equal? Are they the same on all your platforms?
>
> IIRC, under DevCap they indicated MaxPayload 4096 bytes, and under
> DevCtl they indicated MaxPayload 128 bytes and MaxReadReq 4096 bytes,
> and was the same on both the Asus and SuperMicro systems. I will
> doublecheck tomorrow at work. I am not clear on the meanings of
> the different parameters. And is DevCtl for PCI control messages
> and DevCap for actual data transfers or something else?
DevCap is the capability register, which is read-only; DevCtl is the
control register which holds the actual settings.
MaxPayload is the MTU and MRU for PCIe packets. Each sub-tree of
devices connected to a single PCIe root port needs to have MaxPayload
set consistently. MaxReadReq is the maximum size of any DMA read
request. It is a per-device setting (or possibly per-function; I
forget). It can be much larger than MaxPayload since read completions
can be fragmented.
Ben.
--
Ben Hutchings, Senior Software Engineer, Solarflare Communications
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
next prev parent reply other threads:[~2009-08-20 22:47 UTC|newest]
Thread overview: 95+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-08-07 21:06 Receive side performance issue with multi-10-GigE and NUMA Bill Fink
2009-08-07 21:18 ` Brice Goglin
2009-08-07 21:51 ` Bill Fink
2009-08-07 21:53 ` Brice Goglin
2009-08-07 22:08 ` Bill Fink
2009-08-07 22:17 ` Brice Goglin
2009-08-07 22:55 ` Bill Fink
2009-08-08 1:03 ` Andrew Gallatin
2009-08-08 1:35 ` Bill Fink
2009-08-08 11:08 ` Andrew Gallatin
2009-08-08 11:26 ` Neil Horman
2009-08-08 18:21 ` Andrew Gallatin
2009-08-08 18:32 ` Neil Horman
2009-08-11 7:32 ` Bill Fink
2009-08-11 11:02 ` Neil Horman
2009-08-11 19:15 ` Christoph Lameter
2009-08-11 22:27 ` Andi Kleen
2009-08-12 4:30 ` Bill Fink
2009-08-12 7:21 ` Andi Kleen
[not found] ` <4A856781.2080301@myri.com>
2009-08-14 16:38 ` Bill Fink
2009-08-14 16:55 ` Andrew Gallatin
2009-08-14 21:13 ` Aviv Greenberg
2009-08-20 7:26 ` Bill Fink
2009-08-20 13:14 ` Ben Hutchings [this message]
2009-08-21 4:00 ` Bill Fink
2009-08-20 13:17 ` Aviv Greenberg
2009-08-12 0:02 ` Brandeburg, Jesse
2009-08-12 4:38 ` Bill Fink
2009-08-12 16:00 ` Jesse Barnes
2009-08-14 20:31 ` Bill Fink
2009-08-17 16:53 ` Jesse Barnes
2009-08-18 7:07 ` Bill Fink
2009-08-18 11:54 ` Andrew Gallatin
2009-08-19 17:59 ` Bill Fink
2009-08-07 22:12 ` Neil Horman
2009-08-08 0:54 ` Bill Fink
2009-08-08 1:56 ` Neil Horman
2009-08-14 20:44 ` Bill Fink
2009-08-14 23:25 ` Neil Horman
2009-08-20 7:50 ` Bill Fink
2009-08-20 20:19 ` Neil Horman
2009-08-21 4:14 ` Bill Fink
2009-08-21 15:23 ` Neil Horman
2009-08-21 15:36 ` Andrew Gallatin
2009-08-26 7:10 ` Bill Fink
2009-08-26 11:00 ` Neil Horman
2009-08-26 18:08 ` Neil Horman
2009-08-26 18:15 ` Ingo Molnar
2009-08-26 19:04 ` Neil Horman
2009-08-26 19:08 ` Ingo Molnar
2009-08-26 19:36 ` David Miller
2009-08-26 19:48 ` Ingo Molnar
2009-08-26 20:23 ` Neil Horman
2009-08-26 20:40 ` Ingo Molnar
2009-08-26 22:39 ` Neil Horman
2009-08-26 22:44 ` David Miller
2009-08-26 23:05 ` Ingo Molnar
2009-08-26 23:08 ` David Miller
2009-08-26 23:58 ` Ingo Molnar
2009-08-27 0:05 ` Steven Rostedt
2009-08-27 0:35 ` Christoph Hellwig
2009-08-27 9:28 ` Ingo Molnar
2009-08-26 23:05 ` Steven Rostedt
2009-08-26 23:09 ` David Miller
2009-08-26 23:30 ` Ingo Molnar
2009-08-26 23:23 ` Neil Horman
2009-08-26 23:29 ` David Miller
2009-08-26 23:19 ` Neil Horman
2009-08-26 23:14 ` Ingo Molnar
2009-08-26 23:33 ` Steven Rostedt
2009-08-27 0:14 ` Neil Horman
2009-08-27 0:29 ` Steven Rostedt
2009-08-27 1:17 ` Neil Horman
2009-08-27 9:06 ` Ingo Molnar
2009-08-27 9:34 ` Ingo Molnar
2009-08-27 0:34 ` Christoph Hellwig
2009-08-27 0:30 ` blktrace ftrace plugin, was " Christoph Hellwig
2009-08-27 5:26 ` Jens Axboe
2009-08-27 9:12 ` Ingo Molnar
2009-08-27 9:14 ` Jens Axboe
2009-08-27 13:55 ` Arnaldo Carvalho de Melo
2009-08-28 2:03 ` Li Zefan
2009-08-26 23:46 ` Frederic Weisbecker
2009-08-26 20:28 ` Ingo Molnar
2009-08-26 20:01 ` Neil Horman
2009-08-26 22:57 ` Ingo Molnar
2009-08-27 17:32 ` Bill Fink
2009-09-02 5:28 ` Bill Fink
2009-08-27 17:44 ` Bill Fink
2009-08-27 17:51 ` Neil Horman
2009-09-02 5:11 ` Bill Fink
2009-09-02 10:49 ` Neil Horman
2009-09-02 15:38 ` Bill Fink
2009-08-12 23:29 ` David Miller
2009-08-13 2:35 ` Bill Fink
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=1250774056.16001.157.camel@localhost \
--to=bhutchings@solarflare.com \
--cc=avivgnet@gmail.com \
--cc=billfink@mindspring.com \
--cc=gallatin@myri.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.