From: Andi Kleen <andi@firstfloor.org>
To: Glen Turner <gdt@gdt.id.au>
Cc: netdev@vger.kernel.org
Subject: Re: UDP path MTU discovery
Date: Sun, 28 Mar 2010 10:50:06 +0200 [thread overview]
Message-ID: <877how25kx.fsf@basil.nowhere.org> (raw)
In-Reply-To: <1269561751.2891.8.camel@ilion> (Glen Turner's message of "Fri, 26 Mar 2010 10:32:31 +1030")
Glen Turner <gdt@gdt.id.au> writes:
> In IPv6 routers do not fragment packets, so there is no work
> around. Transactional servers which use UDP over IPv6 encounter
> exponential backoffs within the application and the client
> abandons the transaction. There is no way for the server to
> know that the packet was lost due to Path MTU Discovery and
> to immediately re-transmit it (without an exponential penalty)
> so that the MTU can be probed again.
You can still turn path mtu discovery off and Linux will
fragment based on the known path MTU (I believe when
the too big fragment gets a icmp back the pmtu gets updated)
However you might lose a few packets in the process until the path MTU
is known, but at least it will stay cached (unless you thrash the
routing cache)
In theory one could probably add some hack in the the kernel UDP code
to hold one packet and retransmit it immediately with fragments when
the ICMP comes in. However that would be quite far in behaviour from
traditional UDP and be considered very ugly. It could also mess up
congestion avoidance schemes done by the application.
Still might be preferable over rewriting zillions of applications?
-Andi
--
ak@linux.intel.com -- Speaking for myself only.
next prev parent reply other threads:[~2010-03-28 8:50 UTC|newest]
Thread overview: 34+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-03-26 0:02 UDP path MTU discovery Glen Turner
2010-03-26 0:53 ` Rick Jones
2010-03-26 3:26 ` David Miller
2010-03-26 17:48 ` Rick Jones
2010-03-31 23:42 ` Glen Turner
2010-03-31 23:51 ` Hagen Paul Pfeifer
2010-04-01 0:06 ` Rick Jones
2010-03-26 3:24 ` David Miller
2010-03-28 8:41 ` Andi Kleen
2010-03-31 23:57 ` Glen Turner
2010-04-01 0:57 ` Andi Kleen
2010-03-28 8:50 ` Andi Kleen [this message]
2010-03-29 17:01 ` Rick Jones
2010-03-29 20:14 ` Andi Kleen
2010-03-29 20:25 ` Rick Jones
2010-03-29 20:50 ` Edgar E. Iglesias
2010-03-29 21:01 ` Rick Jones
2010-03-29 21:29 ` Eric Dumazet
2010-03-29 23:38 ` Templin, Fred L
2010-03-30 5:20 ` Andi Kleen
2010-03-30 6:06 ` Eric Dumazet
2010-03-30 6:16 ` Andi Kleen
2010-03-30 6:17 ` UDP path MTU discovery II Andi Kleen
2010-03-30 6:16 ` UDP path MTU discovery Edgar E. Iglesias
2010-03-30 6:19 ` Andi Kleen
2010-03-30 8:20 ` Edgar E. Iglesias
2010-03-30 14:12 ` Andi Kleen
2010-03-30 22:04 ` Edgar E. Iglesias
2010-03-30 15:58 ` Templin, Fred L
2010-03-30 16:06 ` Andi Kleen
2010-03-31 23:43 ` Glen Turner
2010-04-01 0:55 ` Andi Kleen
2010-04-02 5:41 ` Glen Turner
2010-04-04 10:25 ` Andi Kleen
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=877how25kx.fsf@basil.nowhere.org \
--to=andi@firstfloor.org \
--cc=gdt@gdt.id.au \
--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.