From: Michael Riepe <michael.riepe@googlemail.com>
To: David Dillow <dave@thedillows.org>
Cc: "Michael Buesch" <mb@bu3sch.de>,
"Francois Romieu" <romieu@fr.zoreil.com>,
"Rui Santos" <rsantos@grupopie.com>,
"Michael Büker" <m.bueker@berlin.de>,
linux-kernel@vger.kernel.org, netdev@vger.kernel.org
Subject: Re: [PATCH 2.6.30-rc4] r8169: avoid losing MSI interrupts
Date: Sat, 23 May 2009 18:12:22 +0200 [thread overview]
Message-ID: <4A182066.9030201@googlemail.com> (raw)
In-Reply-To: <1243090308.4217.6.camel@obelisk.thedillows.org>
Hi!
David Dillow wrote:
> I wonder if that is the TCP sawtooth pattern -- run up until we drop
> packets, drop off, repeat. I thought newer congestion algorithms would
> help with that, but I've not kept up, this may be another red-herring --
> like the bisection into genirq.
Actually, I just found out that things are much stranger. A freshly
booted system (I'm using 2.6.29.2 + the r8169 patch sent by Michael
Buesch, by the way) behaves like this:
[ 3] local 192.168.178.206 port 44090 connected with 192.168.178.204
port 5001
[ ID] Interval Transfer Bandwidth
[ 3] 0.0-10.0 sec 483 MBytes 405 Mbits/sec
[ 3] 10.0-20.0 sec 472 MBytes 396 Mbits/sec
[ 3] 20.0-30.0 sec 482 MBytes 404 Mbits/sec
[ 3] 30.0-40.0 sec 483 MBytes 405 Mbits/sec
[ 3] 40.0-50.0 sec 480 MBytes 402 Mbits/sec
[ 3] 50.0-60.0 sec 479 MBytes 402 Mbits/sec
[ 3] 0.0-60.0 sec 2.81 GBytes 402 Mbits/sec
Then I've been running another test, something along the lines of
for dest in host1 host1 host2 host2
do ssh $dest dd of=/dev/null bs=8k count=10240000 </dev/zero &
done
After a while, I killed the ssh processes and ran iperf again. And this
time, I got:
[ 3] local 192.168.178.206 port 58029 connected with 192.168.178.204
port 5001
[ ID] Interval Transfer Bandwidth
[ 3] 0.0-10.0 sec 634 MBytes 531 Mbits/sec
[ 3] 10.0-20.0 sec 740 MBytes 621 Mbits/sec
[ 3] 20.0-30.0 sec 641 MBytes 538 Mbits/sec
[ 3] 30.0-40.0 sec 738 MBytes 619 Mbits/sec
[ 3] 40.0-50.0 sec 742 MBytes 622 Mbits/sec
[ 3] 50.0-60.0 sec 743 MBytes 623 Mbits/sec
[ 3] 0.0-60.0 sec 4.14 GBytes 592 Mbits/sec
Obviously, the high-load ssh test (which would kill the device within a
few seconds without the patch) triggers something here.
A few observations later, however, I was convinced that it's not a TCP
congestion or driver issue. Actually, the throughput depends on the CPU
the benchmark is running on. You can see that in gkrellm - whenever the
process jumps to another CPU, the throughput changes. On the four
(virtual) CPUs of the Atom 330, I get these results:
CPU 0: 0.0-60.0 sec 2.65 GBytes 380 Mbits/sec
CPU 1: 0.0-60.0 sec 4.12 GBytes 590 Mbits/sec
CPU 2: 0.0-60.0 sec 3.79 GBytes 543 Mbits/sec
CPU 3: 0.0-60.0 sec 4.13 GBytes 592 Mbits/sec
CPU 0+2 are on the first core, 1+3 on the second.
If I use two connections (iperf -P2) and nail iperf to both threads of a
single core with taskset (the program is multi-threaded, just in case
you wonder), I get this:
CPU 0+2: 0.0-60.0 sec 4.65 GBytes 665 Mbits/sec
CPU 1+3: 0.0-60.0 sec 6.43 GBytes 920 Mbits/sec
That's quite a difference, isn't it?
Now I wonder what CPU 0 is doing...
--
Michael "Tired" Riepe <michael.riepe@googlemail.com>
X-Tired: Each morning I get up I die a little
next prev parent reply other threads:[~2009-05-23 16:12 UTC|newest]
Thread overview: 106+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-03-04 17:28 2.6.27.19 + 28.7: network timeouts for r8169 and 8139too Michael Büker
2009-03-04 22:43 ` Francois Romieu
2009-03-06 0:17 ` Michael Büker
2009-03-08 10:27 ` Tom Weber
2009-03-10 5:42 ` Tom Weber
2009-03-09 12:07 ` Rui Santos
2009-03-13 18:29 ` Rui Santos
2009-03-16 13:07 ` Rui Santos
2009-03-22 21:12 ` Francois Romieu
2009-03-22 21:19 ` Michael Buesch
2009-03-22 22:00 ` Francois Romieu
2009-03-22 22:09 ` Michael Buesch
2009-03-22 22:27 ` Francois Romieu
2009-03-22 22:38 ` Michael Buesch
2009-03-23 11:47 ` Michael Buesch
2009-03-23 12:47 ` Michael Buesch
2009-03-23 23:47 ` Francois Romieu
2009-03-24 9:43 ` Michael Buesch
2009-03-23 14:29 ` Michael Büker
2009-03-23 14:57 ` Rui Santos
2009-03-23 15:04 ` Michael Büker
2009-03-25 11:40 ` Rui Santos
2009-04-04 17:50 ` Michael Buesch
2009-05-10 13:38 ` Michael Riepe
2009-05-10 15:01 ` Michael S. Zick
2009-05-10 15:10 ` Michael S. Zick
2009-05-10 15:53 ` Michael Buesch
2009-05-10 16:27 ` Michael Riepe
2009-05-10 17:09 ` Michael S. Zick
2009-05-11 0:29 ` David Dillow
2009-05-11 20:48 ` Michael Buesch
2009-05-11 21:10 ` Michael Buesch
2009-05-11 21:29 ` David Dillow
2009-05-11 21:59 ` Michael Buesch
2009-05-12 20:29 ` Michael Riepe
2009-05-14 2:38 ` David Dillow
2009-05-14 18:37 ` Michael Riepe
2009-05-14 19:14 ` David Dillow
2009-05-14 19:42 ` Michael Riepe
2009-05-23 1:29 ` [PATCH 2.6.30-rc4] r8169: avoid losing MSI interrupts David Dillow
2009-05-23 9:24 ` Michael Buesch
2009-05-23 14:35 ` Michael Riepe
2009-05-23 14:44 ` Michael Buesch
2009-05-23 15:01 ` Michael Riepe
2009-05-23 16:40 ` Michael Buesch
2009-05-23 14:51 ` David Dillow
2009-05-23 16:12 ` Michael Riepe [this message]
2009-05-23 16:45 ` Michael Buesch
2009-05-23 16:46 ` David Dillow
2009-05-23 16:50 ` Michael Buesch
2009-05-23 16:53 ` Michael Riepe
2009-05-23 17:03 ` David Dillow
2009-05-24 21:15 ` Francois Romieu
2009-05-24 22:55 ` David Dillow
2009-05-26 5:55 ` David Miller
2009-05-26 18:22 ` Michael Buesch
2009-05-26 21:52 ` David Miller
2009-05-26 22:14 ` David Miller
2009-05-26 22:40 ` Michael Riepe
2009-05-26 22:43 ` David Miller
2009-05-26 23:10 ` David Miller
2009-05-27 16:19 ` Michael Buesch
2009-06-16 19:32 ` Rui Santos
2009-08-21 20:57 ` Eric W. Biederman
2009-08-21 21:22 ` Michael Riepe
2009-08-21 22:59 ` David Dillow
2009-08-21 23:34 ` David Dillow
2009-08-22 0:24 ` Eric W. Biederman
2009-08-22 11:48 ` Eric W. Biederman
2009-08-22 12:07 ` Eric W. Biederman
2009-08-22 20:43 ` David Dillow
2009-08-23 17:17 ` Jarek Poplawski
2009-08-23 17:43 ` Michal Soltys
2009-08-23 17:54 ` Jarek Poplawski
2009-08-24 2:37 ` Eric W. Biederman
2009-08-25 0:51 ` Eric W. Biederman
2009-08-25 2:59 ` David Dillow
2009-08-25 20:22 ` Eric W. Biederman
2009-08-25 20:40 ` David Dillow
2009-08-25 21:24 ` Eric W. Biederman
2009-08-25 21:46 ` David Dillow
2009-08-25 22:19 ` Francois Romieu
2009-08-26 3:47 ` Eric W. Biederman
2009-08-26 7:58 ` [PATCH] r8169: Reduce looping in the interrupt handler Eric W. Biederman
2009-08-26 13:56 ` David Dillow
2009-08-26 13:59 ` David Dillow
2009-08-26 20:02 ` Eric W. Biederman
2009-08-26 21:30 ` Francois Romieu
2009-08-26 21:40 ` Eric W. Biederman
2009-08-27 5:24 ` Francois Romieu
2009-08-27 5:38 ` Eric W. Biederman
2009-08-27 23:20 ` Francois Romieu
2009-08-28 1:17 ` Eric W. Biederman
2009-08-28 1:29 ` David Dillow
2009-08-30 20:37 ` Francois Romieu
2009-08-30 20:53 ` Eric W. Biederman
2009-09-01 3:33 ` David Dillow
2009-09-01 9:20 ` Francois Romieu
2009-08-25 21:37 ` [PATCH 2.6.30-rc4] r8169: avoid losing MSI interrupts Eric W. Biederman
2009-08-25 21:54 ` David Dillow
2009-08-25 23:11 ` Francois Romieu
2009-05-12 11:10 ` 2.6.27.19 + 28.7: network timeouts for r8169 and 8139too Krzysztof Halasa
2009-05-12 21:45 ` Michael Riepe
2009-05-13 6:11 ` Francois Romieu
2009-05-13 6:27 ` Michael Riepe
2009-05-13 19:34 ` Krzysztof Halasa
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=4A182066.9030201@googlemail.com \
--to=michael.riepe@googlemail.com \
--cc=dave@thedillows.org \
--cc=linux-kernel@vger.kernel.org \
--cc=m.bueker@berlin.de \
--cc=mb@bu3sch.de \
--cc=netdev@vger.kernel.org \
--cc=romieu@fr.zoreil.com \
--cc=rsantos@grupopie.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 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.