From: Andi Kleen <ak@muc.de>
To: davidm@hpl.hp.com
Cc: linux-kernel@vger.kernel.org
Subject: Re: NS83820 2.6.0-test5 driver seems unstable on IA64
Date: Fri, 26 Sep 2003 19:04:30 +0200 [thread overview]
Message-ID: <m37k3viiqp.fsf@averell.firstfloor.org> (raw)
In-Reply-To: <A317.6GH.7@gated-at.bofh.it> (David Mosberger's message of "Fri, 26 Sep 2003 17:50:09 +0200")
David Mosberger <davidm@napali.hpl.hp.com> writes:
> Why would that be? Slower, yes, but very slow?
It depends on your architecture I guess. On K8/x86-64 it isn't that big
a deal (one cycle penalty for unaligned accesses), but if you take an
exception for each unaligned read it will be incredible slow.
Of course the copy in the driver has the same problem, so it's not
much better.
My point was basically that the header access are peanuts compared to
the unaligned copy. It would be much better to optimize the copy than
the header access. The proposal of just copying the header does not
address this. And copying the packet in the driver has the same problem,
so IMHO it's useless.
The solution proposed by Ivan sounds much better. The basic problem
is that the Ethernet header is not a multiple of 4 and that misaligns
everything after it. Use one receive descriptor for the MAC header and
another for the rest (IP/TCP/payload) In the rest everything is aligned and
there are no problems with misaligned data types (ignoring exceptional cases
like broken timestamp options which can be handled slowly). Fixing the
stack to handle separate MAC headers should not be that much work, the
code is fairly limited. Drawback is that it requires bigger changes to the
network drivers and maybe some special case code for non standard MAC
headers.
-Andi
next parent reply other threads:[~2003-09-26 17:05 UTC|newest]
Thread overview: 81+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <A2yd.64p.31@gated-at.bofh.it>
[not found] ` <A2yd.64p.29@gated-at.bofh.it>
[not found] ` <A317.6GH.7@gated-at.bofh.it>
2003-09-26 17:04 ` Andi Kleen [this message]
2003-09-26 17:22 ` NS83820 2.6.0-test5 driver seems unstable on IA64 Chris Friesen
2003-09-26 17:47 ` David Mosberger
2003-09-27 3:57 ` David S. Miller
2003-09-27 4:36 ` Andi Kleen
2003-09-27 4:52 ` David Mosberger
[not found] <A4gy.8wi.13@gated-at.bofh.it>
[not found] ` <A4gy.8wi.15@gated-at.bofh.it>
[not found] ` <A4gy.8wi.17@gated-at.bofh.it>
[not found] ` <A4gy.8wi.11@gated-at.bofh.it>
[not found] ` <A4Tv.Ud.37@gated-at.bofh.it>
[not found] ` <AepM.5Zb.41@gated-at.bofh.it>
2003-09-27 4:26 ` Andi Kleen
2003-09-27 4:34 ` David Mosberger
2003-09-27 23:36 ` David S. Miller
[not found] <zU7D.2Ji.27@gated-at.bofh.it>
2003-09-26 15:14 ` Andi Kleen
2003-09-26 15:41 ` David Mosberger
2003-09-26 6:16 Manfred Spraul
2003-09-26 6:07 ` David S. Miller
2003-09-26 6:41 ` David Mosberger
2003-09-26 14:38 ` Ivan Kokshaysky
2003-09-27 3:12 ` David S. Miller
-- strict thread matches above, loose matches on Subject: below --
2003-09-23 22:58 Luck, Tony
2003-09-23 23:32 ` David S. Miller
2003-09-24 5:18 ` Valdis.Kletnieks
2003-09-23 18:21 Luck, Tony
2003-09-23 18:29 ` Benjamin LaHaise
2003-09-23 18:45 ` David S. Miller
2003-09-23 18:51 ` Grant Grundler
2003-09-23 18:51 ` David S. Miller
2003-09-23 20:38 ` Grant Grundler
2003-09-23 20:45 ` David S. Miller
2003-09-23 22:35 ` Grant Grundler
2003-09-23 23:35 ` David S. Miller
2003-09-24 0:13 ` Grant Grundler
2003-09-24 0:04 ` David S. Miller
2003-09-24 1:58 ` William Lee Irwin III
2003-09-23 19:04 ` Benjamin LaHaise
2003-09-23 18:54 ` Andreas Schwab
2003-09-23 18:52 ` David S. Miller
2003-09-23 19:09 ` Andreas Schwab
2003-09-23 19:01 ` David S. Miller
2003-09-23 19:16 ` Andreas Schwab
2003-09-23 19:08 ` David S. Miller
2003-09-23 19:28 ` Matthew Wilcox
2003-09-23 19:22 ` David S. Miller
2003-09-23 23:15 ` Andrew Morton
2003-09-23 23:37 ` David S. Miller
2003-09-24 0:14 ` Matthew Wilcox
2003-09-24 0:06 ` David S. Miller
2003-09-23 19:10 ` Jeff Garzik
2003-09-23 19:11 ` Benjamin LaHaise
2003-09-23 19:22 ` Jeff Garzik
2003-09-23 21:00 ` Alan Cox
2003-09-23 21:16 ` David Mosberger
2003-09-24 1:47 ` David S. Miller
2003-09-23 17:58 Van Maren, Kevin
2003-09-23 17:57 ` David S. Miller
2003-09-23 18:27 ` David Mosberger
2003-09-23 18:47 ` David S. Miller
2003-09-23 19:17 ` David Mosberger
2003-09-23 19:10 ` David S. Miller
2003-09-23 20:00 ` David Mosberger
2003-09-24 3:08 ` David S. Miller
2003-09-25 16:11 ` Ivan Kokshaysky
2003-09-26 1:12 ` David S. Miller
2003-09-23 18:06 ` David Mosberger
2003-09-23 19:11 ` David S. Miller
2003-09-19 4:16 Peter Chubb
2003-09-19 4:38 ` Grant Grundler
2003-09-19 4:43 ` Andi Kleen
2003-09-19 5:01 ` Peter Chubb
2003-09-19 5:53 ` Andi Kleen
2003-09-19 10:49 ` Benjamin LaHaise
2003-09-23 0:34 ` Peter Chubb
2003-09-23 0:36 ` Benjamin LaHaise
2003-09-23 6:22 ` David S. Miller
2003-09-23 10:40 ` Peter Chubb
2003-09-23 10:51 ` David S. Miller
2003-09-23 14:59 ` David Mosberger
2003-09-23 17:27 ` David S. Miller
2003-09-23 17:56 ` David Mosberger
2003-09-24 2:55 ` David S. Miller
2003-09-23 23:04 ` Ian Wienand
2003-09-23 23:35 ` David S. Miller
2003-09-19 5:04 ` Grant Grundler
2003-09-19 5:11 ` Peter Chubb
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=m37k3viiqp.fsf@averell.firstfloor.org \
--to=ak@muc.de \
--cc=davidm@hpl.hp.com \
--cc=linux-kernel@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 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).