Here is a lonely patch intended for net-next/linux-3.8.
It it simply adapting the batman-adv code to use the new ETH_P_BATMAN define
recently introduced in if_ether.h.
Let me know if there is any problem!
Thank you,
Antonio
The following changes since commit bb728820fe7c42fdb838ab2745fb5fe6b18b5ffa:
core: make GRO methods static. (2012-11-29 13:18:32 -0500)
are available in the git repository at:
git://git.open-mesh.org/linux-merge.git tags/batman-adv-for-davem
for you to fetch changes up to af5d4f7737963f2112f148f97c5820425f050650:
batman-adv: use ETH_P_BATMAN (2012-11-30 10:50:22 +0100)
----------------------------------------------------------------
Included changes:
- Use the new ETH_P_BATMAN define instead of the private BATADV_ETH_P_BATMAN
----------------------------------------------------------------
Antonio Quartulli (1):
batman-adv: use ETH_P_BATMAN
net/batman-adv/hard-interface.c | 3 ++-
net/batman-adv/packet.h | 2 --
net/batman-adv/send.c | 6 ++++--
net/batman-adv/soft-interface.c | 9 +++++----
4 files changed, 11 insertions(+), 9 deletions(-)
Hello:
This is Muye again. I have encountered another problem recently when I
am using B.A.T.M.A.N to do the socket programming.
Currently, I have three nodes. Namely, node A, node B, and node C. The
testing environment is a hall way in our department.
I put node A in one end of the hall way, node C in another end of the
hall way. Node B is in the middle of the hall way.
I can ping node C from node A and vice versa. In node A, I can
traceroute to node C. The traceroute shows it needs to go through node
B.
However, when I send packet from node A to node C. Node C does not
receive anything. But if I send packet from node A to node B, node B
can receive that packet.
Not sure what happened. The link shown below is the program I used to
send packet. Both source and destination are MAC address.
https://lists.open-mesh.org/pipermail/b.a.t.m.a.n/attachments/20111206/52f0…
I greatly appreciate any suggestion, comment, and help.
Thanks
I hereby bring you network coding in batman-adv. This allows a relay
to send two packets (to two destinations) in a single transmission.
Tests[1] have shown an increase in throughput up to 1.6 under the right
circumstances.
For this to work, you will need to compile batman-adv with
CONFIG_BATMAN_ADV_NC=y and a wireless interface with working
promiscuous mode. The rest is taken care of by batman-adv and the
following patches.
[1] http://www.open-mesh.org/projects/open-mesh/wiki/2011-08-18-network-coding-…
Martin Hundebøll (6):
batman-adv: Add the initial code for network coding.
batman-adv: Detect coding nodes and remove these after timeout
batman-adv: Buffer unicast packets before forward.
batman-adv: Code and transmit packets if possible.
batman-adv: Save overheard and tx packets for decoding.
batman-adv: Receive coded packets and decode them.
Makefile | 2 +
Makefile.kbuild | 1 +
bat_iv_ogm.c | 5 +
compat.c | 10 +
compat.h | 1 +
debugfs.c | 18 +
gen-compat-autoconf.sh | 1 +
main.c | 6 +
main.h | 13 +-
network-coding.c | 1818 ++++++++++++++++++++++++++++++++++++++++++++++++
network-coding.h | 122 ++++
originator.c | 6 +
packet.h | 33 +
routing.c | 26 +-
send.c | 5 +
soft-interface.c | 14 +
sysfs-class-net-mesh | 8 +
sysfs.c | 6 +
types.h | 123 ++++
19 files changed, 2209 insertions(+), 9 deletions(-)
create mode 100644 network-coding.c
create mode 100644 network-coding.h
--
1.8.0
hi all,
i am wondering if events in batman-adv can be triggered such as:
- get message when: host/batman-adv lost connection
- get message when: host/batman-adv is connected to min. 1 gateway server
if yes, even whith other events, does anyone have an example how to read the
event in a bash script for further processing?
thanks a lot,
walter.
Hi
I wrote this question to the VillageTelco group, but I think that it
fix better in this list.
"We want make some test with batman-adv and we would like change some
parameter of the code.
So I suppose that I have to add the "new code" to the trunk of openwrt
and compile them together(batman-adv module and openwrt). But this
procedure takes quite time...
I think this because the kernel in which I compile batman-adv must be
the same that the kernel in which I`m going to run the batman-adv
module, mustn't it?But I think that I'm missing something, what am I
wrong?
Is there any way to compile only the kernel module of batman-adv so
that it can work fine with the kernel of openwrt and don't last so
long?
What should we do for try our own changes in the code of batman-adv?
Thank you
Regards"
--
Esteban
Hello David,
here is again our changeset intended for net-next/linux-3.8.
Since the last pull request I removed the patch we were discussing about (the
one exporting the batman-adv compatibility version) as we intend to do not
provide the user the feeling of being "incompatible everyday".
As we discussed some time ago, we are doing what we can to speed our
stabilisation process up and we are not going to break compatibility anymore
before we ultimate such process.
However, I see that every now and then you have the feeling we are not really
doing so. In that case, please, point out what you think is not adequate so that
we can quickly fix it.
Again: we want to stabilise the protocol/messages format and any help/suggestion
from you or the other maintainers is always well accepted (other than being
"rude" only or not answering at all. When you do so, then we have to spend the
entire day in trying to decrypt what was wrong and what we should do to address
the problem, limiting our fun and productivity).
Please pull or let me know if there is any other problem in this batch.
Thank you very much,
Antonio
The following changes since commit de4594a51c904ddcd6c3a6cdd100f7c1d94d3239:
sctp: send abort chunk when max_retrans exceeded (2012-11-20 15:50:37 -0500)
are available in the git repository at:
git://git.open-mesh.org/linux-merge.git tags/batman-adv-for-davem
for you to fetch changes up to e022b956c11084f3ee1c6ece523e14ac07b7c645:
batman-adv: Use packing of 2 for all headers before an ethernet header (2012-11-21 12:35:47 +0100)
----------------------------------------------------------------
Included changes:
- Increase batman-adv version
- Bridge Loop Avoidance: compute checksum (using crc32) on skb fragments instead
of linearising it
- sort the sysfs documentation
- some other minor cleanups
----------------------------------------------------------------
Antonio Quartulli (1):
batman-adv: support array of debugfs general attributes
Marek Lindner (1):
batman-adv: sysfs documentation should keep alphabetical order
Martin Hundebøll (1):
batman-adv: Add wrapper to look up neighbor and send skb
Simon Wunderlich (2):
batman-adv: fix bla compare function
batman-adv: Fix broadcast duplist for fragmentation
Sven Eckelmann (4):
batman-adv: Mark best gateway in transtable_global debugfs
batman-adv: Add function to calculate crc32c for the skb payload
batman-adv: Start new development cycle
batman-adv: Use packing of 2 for all headers before an ethernet header
.../ABI/testing/sysfs-class-net-batman-adv | 11 +-
Documentation/ABI/testing/sysfs-class-net-mesh | 40 +++---
net/batman-adv/Kconfig | 1 +
net/batman-adv/bridge_loop_avoidance.c | 36 +++--
net/batman-adv/bridge_loop_avoidance.h | 6 +-
net/batman-adv/debugfs.c | 34 +++--
net/batman-adv/main.c | 34 +++++
net/batman-adv/main.h | 3 +-
net/batman-adv/packet.h | 16 ++-
net/batman-adv/routing.c | 45 ++----
net/batman-adv/send.c | 33 +++++
net/batman-adv/send.h | 3 +
net/batman-adv/translation-table.c | 155 +++++++++++----------
net/batman-adv/types.h | 2 +-
net/batman-adv/unicast.c | 8 +-
net/batman-adv/vis.c | 35 ++---
16 files changed, 268 insertions(+), 194 deletions(-)