Hi David,
here are two bugfixes for batman-adv which we would like to see integrated into net.
Please pull or let me know of any problem!
Thank you,
Simon
The following changes since commit 51c6b429c0c95e67edd1cb0b548c5cf6a6604763:
batman-adv: Fix transmission of final, 16th fragment (2017-02-21 18:33:35 +0100)
are available in the git repository at:
git://git.open-mesh.org/linux-merge.git tags/batadv-net-for-davem-20170316
for you to fetch changes up to 1a9070ec91b37234fe915849b767c61584c64a44:
batman-adv: Initialize gw sel_class via batadv_algo (2017-03-04 17:32:06 +0100)
----------------------------------------------------------------
Here are two batman-adv bugfixes:
- Keep fragments equally sized, avoids some problems with too small fragments,
by Sven Eckelmann
- Initialize gateway class correctly when BATMAN V is compiled in,
by Sven Eckelmann
----------------------------------------------------------------
Sven Eckelmann (2):
batman-adv: Keep fragments equally sized
batman-adv: Initialize gw sel_class via batadv_algo
net/batman-adv/bat_iv_ogm.c | 11 +++++++++++
net/batman-adv/bat_v.c | 14 +++++++++++---
net/batman-adv/fragmentation.c | 20 +++++++++++++-------
net/batman-adv/gateway_common.c | 5 +++++
net/batman-adv/soft-interface.c | 1 -
net/batman-adv/types.h | 2 ++
6 files changed, 42 insertions(+), 11 deletions(-)
Hi,
here are some patches which are for maint. These address different
bugs [1,2,3]. The last fragmentation patch is already part of master. But
Martin Weinelt and Matthias Schiffer identified it as a bugfix for a problem
of Freifunk Darmstadt. The commit message of this change was therefore
rewritten to also explain that it fixes this problem.
Kind regards,
Sven
[1] https://www.open-mesh.org/issues/326
[2] https://www.open-mesh.org/issues/327
[3] https://www.open-mesh.org/issues/328
Johannes Berg (1):
batman-adv: average: change to declare precision, not factor
Sven Eckelmann (2):
batman-adv: Keep fragments equally sized
batman-adv: Initialize gw sel_class via batadv_algo
compat-include/linux/average.h | 67 +++++++++++++++++++++++++++--------------
net/batman-adv/bat_iv_ogm.c | 11 +++++++
net/batman-adv/bat_v.c | 14 +++++++--
net/batman-adv/fragmentation.c | 20 +++++++-----
net/batman-adv/gateway_common.c | 5 +++
net/batman-adv/soft-interface.c | 1 -
net/batman-adv/types.h | 4 ++-
7 files changed, 87 insertions(+), 35 deletions(-)
==========================================================
The Wireless Battle Mesh v10
5th - 11th of June 2017, Vienna, Austria
==========================================================
The next 'Wireless Battle of the Mesh' will take place from Mon 5th -
Sun 11th of June in the Austrian Museum of Folk Life And Folk Art
(http://www.volkskundemuseum.at) in Vienna.
The event aims to bring together all people from across the globe that
are interested in wireless mesh network technologies and community
networks in general. 6 days full of expert presentations, late night
hacking sessions, measurement campaigns, protocol discussions, and a
whole lotta other meshy things.
So if you are a mesh networking enthusiast, community networking
activist, or have an interest in mesh networks in general, you have to
check this out!
Continuously updated information about the event:
http://battlemesh.org/BattleMeshV10
Location
========
We are located in the upper floor of very nice historical museum in the
middle of Vienna. One large conference room and two adjacent hacking
areas are at our full disposal there. The whole area offers nice
atmosphere and plenty of room to deploy measurement testbeds. The
adjacent park is a beautiful analog distraction, especially in this time
of the year. The metalab - a very active local hackerspace - is only 5
minutes away.
Accommodation Offering
======================
For those of you who are looking for a convenient and low cost
accommodation option: we have made block reservation for 40 people in a
nice hostel 10-15min walking distance from the Museum. The packages
include 6 nights in a four-bed room incl. breakfast - for a mere €100.
Follow the simple instructions below to get one of these hot deals a
first come first serve scheduler is used.
Registration
============
The event is *free of charge* and registration is optional - but it
makes the organisation much easier if you tell us in advance that you
plan to come.
Please send a mail to v10(a)battlemesh.org that looks somewhat like this:
########
Subject: Registration
Name and/or Nick
Accommodation package: [Yes/No]
T-Shirt Size: [S/M/L/none]
Other details you want to share: (e.g. community, country, URLs, twitter
handle, dietary restrictions, ...)
########
We will then put your name on the public participants list:
http://battlemesh.org/BattleMeshV10/Participants
And we will answer with payment instructions for the accommodation
package to those that are interested in one.
Travel Scholarships
===================
If you want to apply for a travel scholarship (compensating the costs of
a long distance flight to Vienna and back) please explicitly tell us so
in the registration mail. We will then ask you to prepare a short video
message in which you are giving a brief introduction to yourself and
your interests.
Spread the Word
===============
Please help us spreading the word by forwarding this announcement to all
lists and people that might be interested. Blogging about it is also
very appreciated, and if you do so, please add a ping-back to the wiki page:
http://battlemesh.org/BattleMeshV10
The Local Infrastructure Core Team
==================================
Albert Rafetseder (albert.rafetseder at univie.ac.at, key ID 0x90382EC8)
Clemens Hopfer (datacop at wireloss.net, key ID 0x5EBA9D09)
Paul Fuxjaeger (paul.fuxjaeger at gmx.at, key ID 0xB5BB47E7)
Contact
=======
* Web: http://battlemesh.org/BattleMeshV10
* Email: http://ml.ninux.org/mailman/listinfo/battlemesh
* IRC: irc.freenode.net #battlemesh
* Twitter: https://twitter.com/battlemesh/
_______________________________________________
Battlemesh mailing list
Battlemesh(a)ml.ninux.org
http://ml.ninux.org/mailman/listinfo/battlemesh
-----------------------------------------
Hi David,
here are two bugfixes which we would like to see integrated into net.
Please pull or let me know of any problem!
Thank you,
Simon
The following changes since commit 4ea33ef0f9e95b69db9131d7afd98563713e81b0:
batman-adv: Decrease hardif refcnt on fragmentation send error (2017-01-04 08:22:04 +0100)
are available in the git repository at:
git://git.open-mesh.org/linux-merge.git tags/batadv-net-for-davem-20170301
for you to fetch changes up to 51c6b429c0c95e67edd1cb0b548c5cf6a6604763:
batman-adv: Fix transmission of final, 16th fragment (2017-02-21 18:33:35 +0100)
----------------------------------------------------------------
Here are two batman-adv bugfixes:
- fix a potential double free when fragment merges fail,
by Sven Eckelmann
- fix failing tranmission of the 16th (last) fragment if that exists,
by Linus Lüssing
----------------------------------------------------------------
Linus Lüssing (1):
batman-adv: Fix transmission of final, 16th fragment
Sven Eckelmann (1):
batman-adv: Fix double free during fragment merge error
net/batman-adv/fragmentation.c | 20 +++++++++++---------
1 file changed, 11 insertions(+), 9 deletions(-)
With the addition of multicast optimization features in batman-adv
not only unicast but also multicast addresses are now fed into the
translatation table. This approximately increases the number of
entries by a factor of roughly 2 to 3.
Also, the number of client devices in some known, public mesh networks
has grown to nearly 4000.
Therefore this patch inceases the global translation table size from
1024 to 16384 to reduce load in larger mesh networks.
Signed-off-by: Linus Lüssing <linus.luessing(a)c0d3.blue>
---
ToDo (maybe): Gather some statistics on how much this might reduce load
in some particular networks.
This might be a hotfix for some larger Freifunk mesh networks which are
experiencing a significantly higher load level since they switched to
a recent batman-adv version. Validating whether the hashtable size might
be the reason is currently being investigated.
On the long run we should switch to the resizable hash table implementation
the kernel now provides since not that long ago.
---
net/batman-adv/translation-table.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/batman-adv/translation-table.c b/net/batman-adv/translation-table.c
index e75b493..e87825c 100644
--- a/net/batman-adv/translation-table.c
+++ b/net/batman-adv/translation-table.c
@@ -1462,7 +1462,7 @@ static int batadv_tt_global_init(struct batadv_priv *bat_priv)
if (bat_priv->tt.global_hash)
return 0;
- bat_priv->tt.global_hash = batadv_hash_new(1024);
+ bat_priv->tt.global_hash = batadv_hash_new(16384);
if (!bat_priv->tt.global_hash)
return -ENOMEM;
--
2.1.4