public inbox for b.a.t.m.a.n@lists.open-mesh.org
 help / color / mirror / Atom feed
From: Simon Wunderlich <simon.wunderlich@s2003.tu-chemnitz.de>
To: The list for a Better Approach To Mobile Ad-hoc Networking
	<b.a.t.m.a.n@lists.open-mesh.org>
Subject: Re: [B.A.T.M.A.N.] batman adv unicast fragmentation
Date: Sun, 4 Jul 2010 23:57:41 +0200	[thread overview]
Message-ID: <20100704215741.GA28501@pandem0nium> (raw)
In-Reply-To: <20100630205846.230b2112@rechenknecht>

[-- Attachment #1: Type: text/plain, Size: 2154 bytes --]

Hello Andreas,

thank you very much for your patch. I've tested the patchset in my KVM setup.
The fragmentation seems to work fine over one or multiple hops if i use the
same MTUs on all hosts, but some corner cases might not bet handled yet. 

For example:

Host 1 has one interface with MTU 1500 to host 2.
Host 2 has two interfaces, one with MTU 1500 (to host 1) and one interface
with MTU 1470 (to host 3), which might be a VPN connection.
Host 3 has one interface with MTU 1470 to host 2.

Now i can send packets with size 1470 or 1440 from host 1 to host 3 without 
problems, but packets with size 1450 fail. The reason for this might be that 
host 1 sends the packet as is without fragmentation, but host 2 can not 
forward it because the MTU (1470) is too low, but does not fragment it. 
The packets therefore get dropped. 
(note: when i say packets of size xxxx i do: ping -M do -s xxxx)

I've also tried to do a similar setup with MTUs 1530 and 1500 instead
of 1500 and 1470, which should simulate a wifi network connected to a 
fast ethernet network, but unfortunately my KVM crashed with high MTUs,
so i could not verify this setup yet.

However, i think that scenarios like these, a high MTU medium connected
to a low MTU medium are quite common. Linus has interconnected WiFi
networks with VPN, and i've also used some WiFi network interconnected
with fast Ethernet. If forwarding hosts could fragment packets on their
own, this problem could be solved, right? What do you (or others) think?

Just as a minor comment, the copyright year in your new files should just
be 2010, not 2007 - 2010.

best regards,
	Simon

On Wed, Jun 30, 2010 at 08:58:46PM +0200, Andreas Langer wrote:
> Hi,
> 
> things that changed since the last patch:
> - with fragmentation enabled the mtu of batX is always ETH_DATA_LEN
> - fragment if mtu of the outgoing interface is smaller the needed size
> - add a new packet type for fragmentation
> - new recv function for fragmented packets
> - new route unicast packet function to share routing code
> - increase compat number to 12
> 
> 
> regards,
> andreas
> 

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 197 bytes --]

  parent reply	other threads:[~2010-07-04 21:57 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-06-30 18:58 [B.A.T.M.A.N.] batman adv unicast fragmentation Andreas Langer
2010-06-30 19:00 ` [B.A.T.M.A.N.] [PATCH 1/2] batctl: layer2 unicast packet fragmentation Andreas Langer
2010-07-04 22:40   ` Sven Eckelmann
2010-06-30 19:00 ` [B.A.T.M.A.N.] [PATCH 2/2] batman-adv: " Andreas Langer
2010-07-05  0:37   ` Sven Eckelmann
2010-07-05 12:53   ` Sven Eckelmann
2010-07-04 21:57 ` Simon Wunderlich [this message]
2010-07-05  9:24   ` [B.A.T.M.A.N.] batman adv unicast fragmentation Sven Eckelmann

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=20100704215741.GA28501@pandem0nium \
    --to=simon.wunderlich@s2003.tu-chemnitz.de \
    --cc=b.a.t.m.a.n@lists.open-mesh.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