From: linas@austin.ibm.com (Linas Vepstas)
To: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Arnd Bergmann <arnd@arndb.de>,
netdev@vger.kernel.org, James K Lewis <jklewis@us.ibm.com>,
linux-kernel@vger.kernel.org, linuxppc-dev@ozlabs.org,
Jens Osterkamp <Jens.Osterkamp@de.ibm.com>
Subject: Re: [PATCH 2/4]: powerpc/cell spidernet low watermark patch.
Date: Fri, 18 Aug 2006 14:23:56 -0500 [thread overview]
Message-ID: <20060818192356.GD26889@austin.ibm.com> (raw)
In-Reply-To: <1155771820.11312.116.camel@localhost.localdomain>
On Thu, Aug 17, 2006 at 01:43:40AM +0200, Benjamin Herrenschmidt wrote:
>
> Sounds good (without actually looking at the code though :), that was a
> long required improvement to that driver. Also, we should probably look
> into using NAPI polling for tx completion queue as well, no ?
Just for a lark, I tried using NAPI polling, while disabling all TX
interrupts. Performance was a disaster: 8Mbits/sec, fom which I conclude
that the tcp ack packets do not flow back fast enough to allw reliance
on NAPI polling for transmit.
I was able to get as high as 960 Mbits/sec in unusal circumstances,
at 100% cpu usage. Oprofile indicates that the next major improvement
would be to add scatter/gather, which I'll take a shot at next week,
if I don't get interrupted. However, I'm getting interrupted a lot these
days.
--linas
WARNING: multiple messages have this Message-ID (diff)
From: linas@austin.ibm.com (Linas Vepstas)
To: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
linuxppc-dev@ozlabs.org,
Jens Osterkamp <Jens.Osterkamp@de.ibm.com>,
James K Lewis <jklewis@us.ibm.com>, Arnd Bergmann <arnd@arndb.de>
Subject: Re: [PATCH 2/4]: powerpc/cell spidernet low watermark patch.
Date: Fri, 18 Aug 2006 14:23:56 -0500 [thread overview]
Message-ID: <20060818192356.GD26889@austin.ibm.com> (raw)
In-Reply-To: <1155771820.11312.116.camel@localhost.localdomain>
On Thu, Aug 17, 2006 at 01:43:40AM +0200, Benjamin Herrenschmidt wrote:
>
> Sounds good (without actually looking at the code though :), that was a
> long required improvement to that driver. Also, we should probably look
> into using NAPI polling for tx completion queue as well, no ?
Just for a lark, I tried using NAPI polling, while disabling all TX
interrupts. Performance was a disaster: 8Mbits/sec, fom which I conclude
that the tcp ack packets do not flow back fast enough to allw reliance
on NAPI polling for transmit.
I was able to get as high as 960 Mbits/sec in unusal circumstances,
at 100% cpu usage. Oprofile indicates that the next major improvement
would be to add scatter/gather, which I'll take a shot at next week,
if I don't get interrupted. However, I'm getting interrupted a lot these
days.
--linas
next prev parent reply other threads:[~2006-08-18 19:24 UTC|newest]
Thread overview: 91+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-08-11 17:03 [PATCH 0/4]: powerpc/cell spidernet ethernet driver fixes Linas Vepstas
2006-08-11 17:03 ` Linas Vepstas
2006-08-11 17:06 ` [PATCH 1/4]: powerpc/cell spidernet burst alignment patch Linas Vepstas
2006-08-11 17:06 ` Linas Vepstas
2006-08-11 17:08 ` [PATCH 2/4]: powerpc/cell spidernet low watermark patch Linas Vepstas
2006-08-11 17:08 ` Linas Vepstas
2006-08-16 23:43 ` Benjamin Herrenschmidt
2006-08-16 23:43 ` Benjamin Herrenschmidt
2006-08-18 19:23 ` Linas Vepstas [this message]
2006-08-18 19:23 ` Linas Vepstas
2006-08-18 21:25 ` David Miller
2006-08-18 21:25 ` David Miller
2006-08-18 22:46 ` Linas Vepstas
2006-08-18 22:46 ` Linas Vepstas
2006-08-18 22:51 ` David Miller
2006-08-18 22:51 ` David Miller
2006-08-18 23:29 ` Linas Vepstas
2006-08-18 23:29 ` Linas Vepstas
2006-08-18 23:45 ` Linas Vepstas
2006-08-18 23:45 ` Linas Vepstas
2006-08-19 4:33 ` Benjamin Herrenschmidt
2006-08-19 4:33 ` Benjamin Herrenschmidt
2006-08-22 0:13 ` Linas Vepstas
2006-08-22 0:13 ` Linas Vepstas
2006-08-22 0:30 ` David Miller
2006-08-22 0:30 ` David Miller
2006-08-19 4:31 ` Benjamin Herrenschmidt
2006-08-19 4:31 ` Benjamin Herrenschmidt
2006-08-11 17:09 ` [PATCH 3/4]: powerpc/cell spidernet stop error printing patch Linas Vepstas
2006-08-11 17:09 ` Linas Vepstas
2006-08-11 17:11 ` [PATCH 4/4]: powerpc/cell spidernet ethtool -i version number info Linas Vepstas
2006-08-11 17:11 ` Linas Vepstas
2006-08-11 18:00 ` Olof Johansson
2006-08-11 18:00 ` Olof Johansson
2006-08-11 18:50 ` James K Lewis
2006-08-11 18:50 ` James K Lewis
2006-08-11 19:46 ` Linas Vepstas
2006-08-11 19:46 ` Linas Vepstas
2006-08-15 19:05 ` Olof Johansson
2006-08-15 19:05 ` Olof Johansson
2006-08-16 0:29 ` Michael Ellerman
2006-08-16 0:29 ` Michael Ellerman
2006-08-11 17:42 ` [PATCH 0/4]: powerpc/cell spidernet ethernet driver fixes jschopp
2006-08-11 17:42 ` jschopp
2006-08-11 17:44 ` Sam Ravnborg
2006-08-11 17:44 ` Sam Ravnborg
2006-08-11 19:31 ` Linas Vepstas
2006-08-11 19:31 ` Linas Vepstas
2006-08-11 20:27 ` Arnd Bergmann
2006-08-11 20:27 ` Arnd Bergmann
2006-08-16 16:18 ` [PATCH 1/2]: powerpc/cell spidernet bottom half Linas Vepstas
2006-08-16 16:18 ` Linas Vepstas
2006-08-16 16:30 ` Jeff Garzik
2006-08-16 16:30 ` Jeff Garzik
2006-08-16 20:30 ` Linas Vepstas
2006-08-16 20:30 ` Linas Vepstas
2006-08-16 20:34 ` Jeff Garzik
2006-08-16 20:34 ` Jeff Garzik
2006-08-16 20:46 ` David Miller
2006-08-16 20:46 ` David Miller
2006-08-16 21:24 ` Arnd Bergmann
2006-08-16 21:24 ` Arnd Bergmann
2006-08-16 21:24 ` Arnd Bergmann
2006-08-16 21:32 ` David Miller
2006-08-16 21:32 ` David Miller
2006-08-16 22:16 ` Arnd Bergmann
2006-08-16 22:16 ` Arnd Bergmann
2006-08-16 22:16 ` Arnd Bergmann
2006-08-16 22:29 ` David Miller
2006-08-16 22:29 ` David Miller
2006-08-16 22:47 ` Arnd Bergmann
2006-08-16 22:47 ` Arnd Bergmann
2006-08-16 23:30 ` Linas Vepstas
2006-08-16 23:30 ` Linas Vepstas
2006-08-16 23:32 ` David Miller
2006-08-16 23:32 ` David Miller
2006-08-17 0:23 ` Linas Vepstas
2006-08-17 0:23 ` Linas Vepstas
2006-08-16 23:24 ` Linas Vepstas
2006-08-16 23:24 ` Linas Vepstas
2006-08-16 22:55 ` Linas Vepstas
2006-08-16 22:55 ` Linas Vepstas
2006-08-16 23:03 ` Arnd Bergmann
2006-08-16 23:03 ` Arnd Bergmann
2006-08-16 23:47 ` Linas Vepstas
2006-08-16 23:47 ` Linas Vepstas
2006-08-16 23:08 ` Rick Jones
2006-08-16 23:08 ` Rick Jones
2006-08-16 21:58 ` Linas Vepstas
2006-08-16 21:58 ` Linas Vepstas
2006-08-16 16:23 ` [PATCH 2/2]: powerpc/cell spidernet refine locking Linas Vepstas
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=20060818192356.GD26889@austin.ibm.com \
--to=linas@austin.ibm.com \
--cc=Jens.Osterkamp@de.ibm.com \
--cc=arnd@arndb.de \
--cc=benh@kernel.crashing.org \
--cc=jklewis@us.ibm.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linuxppc-dev@ozlabs.org \
--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.