B.A.T.M.A.N February 2009

b.a.t.m.a.n@lists.open-mesh.org

14 participants
12 discussions

[B.A.T.M.A.N.] Batman gateway lock ups
by Outback Dingo 21 Jul '09

21 Jul '09

see pastebin http://www.pastebin.ca/1194874 pertinent info dmesg | grep 'batgat loaded' batgat: [init_module:96] batgat loaded rv1025 uname -a Linux nightwing 2.6.23.16 #16 Tue Apr 22 20:00:17 ART 2008 mips unknown root@nightwing:~# batmand -v WARNING: You are using the unstable batman branch. If you are interested in *using* batman get the latest stable release ! B.A.T.M.A.N. 0.3-beta (compatibility version 5) lsmod Module Size Used by Tainted: P sch_htb 14048 2 ath_ahb 103616 0 wlan_xauth 480 0 wlan_wep 4000 0 wlan_tkip 9856 0 wlan_ccmp 5440 2 wlan_acl 1920 0 ath_rate_minstrel 8352 1 ath_hal 136832 3 ath_ahb,ath_rate_minstrel wlan_scan_sta 8768 1 wlan_scan_ap 6656 0 wlan 152464 10 ath_ahb,wlan_xauth,wlan_wep,wlan_tkip,wlan_ccmp,wlan_acl,ath_rate_minstrel,wlan_scan_sta,wlan_scan_ap batgat 10944 1 ipt_iprange 672 0 ipt_TOS 832 0 ipt_TTL 928 0 xt_MARK 960 3 ipt_ECN 1472 0 xt_CLASSIFY 640 0 ipt_ttl 704 0 ipt_tos 544 0 ipt_time 1568 0 xt_tcpmss 1088 0 xt_statistic 832 0 xt_mark 672 7 xt_mac 736 3 xt_length 736 0 ipt_ecn 1024 0 xt_DSCP 1056 0 xt_dscp 832 0 imq 2096 0 ipt_IMQ 672 2 xt_string 896 0 xt_layer7 9840 0 ipt_ipp2p 6784 0 ipt_LOG 4640 0 xt_CHAOS 1792 0 xt_DELUDE 2624 1 xt_TARPIT 2816 1 xt_quota 800 0 xt_portscan 2016 0 xt_pkttype 704 0 xt_physdev 1488 0 ipt_owner 800 0 iptable_raw 832 0 xt_NOTRACK 832 0 xt_CONNMARK 1088 0 ipt_recent 4992 0 xt_helper 992 0 xt_conntrack 1312 0 xt_connmark 832 0 xt_connbytes 1312 0 tun 6592 0

3 8

[B.A.T.M.A.N.] [PATCH] Fix deletion of aggregate_ogm in /proc/net/batman-adv
by Sven Eckelmann 24 Feb '09

24 Feb '09

remove_proc_entry needs as second parameter the parent directory and not the proc_dir_entry of the file to delete. Otherwise nothing will be removed the deletion of /proc/net/batman-adv will fail because it still has child nodes. Signed-off-by: Sven Eckelmann <sven.eckelmann(a)gmx.de> --- batman-adv-kernelland/proc.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/batman-adv-kernelland/proc.c b/batman-adv-kernelland/proc.c index 6c475d4..00e177a 100644 --- a/batman-adv-kernelland/proc.c +++ b/batman-adv-kernelland/proc.c @@ -134,7 +134,7 @@ void cleanup_procfs(void) remove_proc_entry(PROC_FILE_VIS, proc_batman_dir); if (proc_aggr_file) - remove_proc_entry(PROC_FILE_AGGR, proc_aggr_file); + remove_proc_entry(PROC_FILE_AGGR, proc_batman_dir); if (proc_batman_dir) #ifdef __NET_NET_NAMESPACE_H -- 1.6.1.3

1 0

[B.A.T.M.A.N.] Drop Packet, , single packet to large to fit maximum packet size
by Stephan Enderlein (Freifunk Dresden) 18 Feb '09

18 Feb '09

Hi, I very often get the following error in syslog and batman log. Currently I use batman-experimental rev 1214 The error occurres every second. [ 79646204] Error - Drop Packet, single packet to large to fit maximum packet size scheduled time 79646157, now 79646202, agg_size 4, next_len 276 !! Axel: have you read my last email directly sent to you? It was about the hanging batmand problem. Bye Stephan --------------------------------------- Dipl.Informatiker(FH) Stephan Enderlein Freifunk Dresden

1 0

[B.A.T.M.A.N.] B.A.T.M.A.N. on NS-2 ?
by Arnaud Kaiser 17 Feb '09

17 Feb '09

Hi, I wonder if it exists an implementation of B.A.T.M.A.N. on NS-2 simulator. Thanks. Arnaud.

4 3

[B.A.T.M.A.N.] WiMAX
by Tim LePes 16 Feb '09

16 Feb '09

Has anyone been playing with BATMAN on WiMAX hardware? I have a project that I had to back burner for a while and I am getting ready to start active work again. I am looking to see if I can find a mini PCI WiMAX card that I can use in a Mesh AP, where WiMAX will be used for the backbone and both WiFi and 10/100 Ethernet for clients to connect to the AP. The idea is for drop-and-go network coverage in disaster areas or other places where you want to get connectivity up in a hurry. I have a dev board and am finally able to get some parts to help me work on it again. Power supply, PCI to MiniPCI adapter, PoE hardware, enclosure.. and of course a WiMAX card. I am also shopping for a WiFi card too, so recommendations there would be welcome. Though to be fair I haven't been googling on that much yet, the WiMAX backhaul is my main concern at the moment. Doing the WiFi AP part should be relatively easy as it is a mature technology w/r/t parts and drivers availability. Any advice? So far it seems that Intel has the greatest presence in my googling for WiMAX on Linux. Thanks so much! Tim LePes

4 4

[B.A.T.M.A.N.] Using batman L2 on multiple wireless cards
by Gargi Purohit 14 Feb '09

14 Feb '09

Hi I was wondering if I can run L2 batman kernel module on hardware which has 3 wireless cards. Should be process be the same as in adding interfaces e.g. ath0, ath1, ath2 to /proc/net/batman-adv/interfaces (where ath0, ath1, ath2 are different wireless cards) and adding bat0 with eth0 to the bridge? Regards Gargi

4 6

[B.A.T.M.A.N.] B.A.T.M.A.N L2 over a Bluetooth PAN
by Breno Jacinto 14 Feb '09

14 Feb '09

Hello, I'm experimenting on heterogeneous ad hoc networks, which today are still very restricted in terms of hardware availability for off-the-shelf devices. Actually, the two most common technologies available today are 802.11 and Bluetooth, being this last one able to form simple personal area networks in the form of piconets. In practical terms, the BlueZ stack of the Linux allows the formation of Bluetooth PANs consisting of up to 8 nodes, but it's an incredibly difficult process - actually, the piconet master needs to bridge connections so that all other devices in the network, which depend on this master, are able to reach it other directly - or at least feel like they are doing that. In fact, everything is just an emulation, since Bluetooth is a connection-oriented technology. BNEP (Bluetooth Network Encapsulation Protocol) is the piece responsible for emulating an Ethernet network over Bluetooth, and theoretically any IP and above protocols can be used. Trying with simple IP broadcasts over a Bluetooth network shows that it works, but the problem comes when I try to do "straight" layer 2 broadcasts (not coming from the IP layer), such as the ones that BATMAN L2 does. I tried to run BATMAN and add the interfaces used on each device - pan0 for the master, which is a bridge device, and bnep0 for the slave, but the frame dont seem to reach any destination on the network. The originators table remains empty on both sides of the connection. Running Wireshark shows the frame being generated, but never reaching anyone. So, I still cannot figure out why this is happening. I'd like to know if anyone has adventured in doing something similar and succeeded. best regards, -- -- :: Breno Jacinto :: :: breno - at - gprt.ufpe.br :: :: FingerPrint :: 2F15 8A61 F566 E442 8581 E3C0 EFF4 E202 74B7 7484 :: Persistir no difícil é a única maneira de torná-lo fácil algum dia. ::

2 2

Re: [B.A.T.M.A.N.] B.A.T.M.A.N Digest, Vol 26, Issue 7
by Tim LePes 13 Feb '09

13 Feb '09

On Fri, 13 Feb 2009 03:26:13 +0700, Outback Dingo <outbackdingo(a)gmail.com> wrote: > yupp got me kinda curious also, though you can definatley do stable > long distance with wifi and the right antennas, what advantage really > does wimax provide, if any. Overall id be very surprised if it was > even really adopted. there are very few wimax capable laptops even now > let alone desktops. > > On Fri, Feb 13, 2009 at 3:05 AM, Simon Wunderlich > <simon.wunderlich(a)s2003.tu-chemnitz.de > <mailto:simon.wunderlich@s2003.tu-chemnitz.de>> wrote: > > Hello Tim, > > unfortunately i have no experience with WiMAX, but as far as i know: > > * WiMAX is not license-free like WiFi (means you need to BUY a > license, > at least in germany) > * where is the technical advantage of WiMAX over WiFi? > Sure, the advertisements say "50 km or 108 MBit/s", but they > don't say > "both at the same time". > You can reach this with WiFi as well (with the right antennas > and setup). > > So my question is: Where do you see the advantages in WiMAX over WiFi? > > I don't know if there is some mesh technology available (as in > implemented) for WiMAX (as far as i know 802.16d?) or if there are > adhoc-like modes available to let routing protocols like BATMAN work. > > If you know more about WiMAX meshing i'd be very interested to > hear. :) > > best regards > Simon > I understand you can get good range with WiFi using directional antennae, parabolic antennae, "cantennas", etc. I *have* thought about directional antennas for back-haul, but the project idea was geared toward emergency workers being able to set up a network as simply as possible. So I have been leaning toward a good solution to work with omni-directional antennas. No "aiming" needed... plug and play would be ideal. With directional antennas it would take a little more work. But may be a better solution in the end. Nothing is set in stone here. I was going with WiMAX because, to my understanding, I could get a far, far greater range using omni-directional antennas than I could with WiFi. I do understand that the data rates fall off with distance. I am not interested, here, in providing typical municipal wireless solutions where the users are going to be working with rich content a lot. I want to be able to set up a communications network that can help rescue workers get access to logistics and co-ordination web pages. Mainly, I am interested in implementing SAHANA on this network. From their web site, "Sahana is a Free and Open Source Disaster Management system. It is a web based collaboration tool that addresses the common coordination problems during a disaster from finding missing people, managing aid, managing volunteers, tracking camps effectively between Government groups, the civil society (NGOs) and the victims themselves." You can find more info at http://www.sahana.lk/ or http://en.wikipedia.org/wiki/Sahana_FOSS_Disaster_Management_System if you would like to check it out further. So the primary application I want to support is simply a web app on a server, and have it run reasonably well over the mesh. I am not going for VoIP or streaming video, so the top-end bandwidth is not my main goal. Distance is. Ease of set-up is. But, of course, I will eventually have to do real-world testing to see if the throughput would suffice. If I were to do directional antennas, I would look to WiFi or Microwave just as readily as WiMAX. Now as a secondary function, this disaster-relief network could also support access out to the wider web, and it would be nice for rescue workers in the Red Cross tent or what have some PCs connected to my mesh node, using it as an AP into the mesh and on to the web. The idea is not, even then, to support video conferencing or bittorrents. Email to friends and family. Government and NGO resource web pages. Etc. The dev board I have is a 400MHz Freescale MPC5200B PowerPC SoC - based unit called the EFIKA 5200B from Genesi. See http://www.genesi-usa.com/efika for details. I am going to outfit it with a PCI to MiniPCI adapter and should be able to put two miniPCI cards on it. Now here, I will note that I have only been able to find a mini-PCI EXPRESS adapter (from Intel, and it also supports WiFi). Intel's "Echo Peak"... "Intel® WiMAX/WiFi Link 5150(512ANXMMWG) is a 1X2 MIMO (Multiple Input, Multiple Output) 1x2 (one transmit chain, two receive chains) PCIe Minicard supporting IEEE 802.16e and 802.11a/b/g/n wireless network adapters that operate in the 2.4 GHz and 5.0 GHz spectra for WiFi and 2.5 GHz for WiMAX." 802.16d tends to be good for fixed stations. And the idea is that these nodes will pretty much stay put once set up. 802.16e is largely being touted for it's mobile network features... and it is coming to the US. Clearwire wireless broadband (802.11d methinks) is already deployed in Nashville. Clearwire is now backed by Sprint and others. Sprint's XOHM is also picking up steam. This is an 802.16e service that is one "4G" route. Though for the cell companies, LTE is also on the horizon. Sprint are supposed to be carrying the Nokia N810 tablet now with WiMax in it. More and more MIDs, netbooks, smart phones, and eventually laptops and such will be coming with 802.16e chipsets. Especially since major cell carriers have been investing in this technology. Sprint has infrastructure already. I am not so interested in the mobile aspects of 802.16e though. I am interested in the MIMO technology support, however. Licensing issues I will get to later. There is unlicensed spectrum available here in the US. But as this is geared toward emergency use, I imagine that there will be considerations regarding spectrum use for such an application that may be favorable. More research is in order here, I agree. At one point Genesi told me I had the Australian government interested in my project, but at the time I had nothing to show for it yet so I didn't pursue anything with them. Granted, this is a learn-as-you-go project for me. I have Debian running on the device, but that's about it so far. Well, it also runs Apache. I have to get some hardware to be able to do anything constructive. I am not being backed by anyone, other than Genesi who were gracious enough to give me an EFIKA at no cost. Their hope is that I would create a product around the board, and I would like to make good on that. They are a unique company with a strong open-source vision, including open hardware designs. They make the reference design, and are first to be ready to market any products derived from it. But others can implement my project as their own product, either using Genesi for the hardware or going it on their own. I have a LOT to learn about all this stuff... but for me, that is the point. I get to play with cool toys and maybe in the end do something that could be a benefit to others. WiMAX is "cool" to me, so that was my initial attraction. I also am keenly interested in wireless networking. I am no fan of monopolies. And it's also a great option for rural areas and developing countries. BATMAN is also very cool to me and I have been following along here for a while, lurking on and off. But when all is said and done with the project... I don't need a WiMAX mesh node for myself. So I will turn the EFIKA into a car computer when I'm done... there is already another project doing that and I'll just copy off them at that time. But for now I want to get my idea moving forward. I am completely open to suggestions and advice along the way. Thanks for being curious, and I hope you find my idea interesting. Peace! Tim

3 4

[B.A.T.M.A.N.] Stable adapters for Ad Hoc networks
by Breno Jacinto 10 Feb '09

10 Feb '09

Hello everyone, I`m running several experiments involving ad hoc networks. I`ve selected some well known-to-be-compatible adapters, such as Edimax (Ralink shipset), Alfa networks (Realtek) and Atheros (madwifi). Atheros is available only as built-in adapter on the equipement or as a PCI board. So, I'm experiencing some instability regarding the selection of the BSSID used by all the nodes. I start the network with a single node, then another, then another. Now, for a couple of minutes it works, bu then some nodes start losing sight of each and changing their BSSIDs to a different value (and they should always use the same). I'm really running out of explanations of why this is happening, because these nodes are all in-range of each other, and should be "pingable" directly. My main suspect is the driver instability, but then I dont know of any other better alternatives. I've tested the following USB adapters: - Edimax EW-7318USg (rt73 drivers) - Alfa AWUS036H-11g And the builtin adapters: - Atheros chipset AR5007EG best regards, -- -- :: Breno Jacinto :: :: breno - at - gprt.ufpe.br :: :: FingerPrint :: 2F15 8A61 F566 E442 8581 E3C0 EFF4 E202 74B7 7484 :: Persistir no difícil é a única maneira de torná-lo fácil algum dia. ::

4 13

[B.A.T.M.A.N.] [PATCH] Introduce set/get functions for /proc files
by Sven Eckelmann 07 Feb '09

07 Feb '09

get_integer_file can be used to get an integer from the beginning of a file. It does the same as every other get_* function in linux/kernel.c has done before, but introduced extra error checking for fscanf to deal with the undefined behaviour of it in error cases in c89. It fixes also the problems that in future architectures with sizeof(int) > 4 an invalid write would happen when fscanf tries to write it's results into the target or that on big endian architectures with sizof(int) != 4 the result is complete different than the expected one when the content of the file is not 0. set_integer_file does the same for writing a specific integer to the file and thus resolve the problem of invalid read on architectures with sizof(int) > 4 and wrong file content on big endian systems with sizof(int) != 4. As side effect the probability to get attacked by a goto-raptor was reduced by the factor 1.08. Signed-off-by: Sven Eckelmann <sven.eckelmann(a)gmx.de> --- batman/linux/kernel.c | 89 ++++++++++++++++++++++--------------------------- 1 files changed, 40 insertions(+), 49 deletions(-) diff --git a/batman/linux/kernel.c b/batman/linux/kernel.c index 0595899..203fb33 100644 --- a/batman/linux/kernel.c +++ b/batman/linux/kernel.c @@ -27,6 +27,7 @@ #include <unistd.h> #include <sys/ioctl.h> #include <sys/socket.h> +#include <inttypes.h> #include "../os.h" #include "../batman.h" @@ -36,9 +37,41 @@ +static int get_integer_file(const char* filename) +{ + FILE *f; + int32_t integer = 0; + int n; + + if((f = fopen(filename, "r")) == NULL) + return 0; + + n = fscanf(f, "%"SCNd32, &integer); + fclose(f); + + if (n == 0 || n == EOF) + integer = 0; + + return integer; +} + + + +static void set_integer_file(const char* filename, int32_t integer) +{ + FILE *f; + + if ((f = fopen(filename, "w")) == NULL) + return; + + fprintf(f, "%"PRId32, integer); + fclose(f); +} + + + void set_rp_filter(int32_t state, char* dev) { - FILE *f; char filename[100], *colon_ptr; /* if given interface is an alias use parent interface */ @@ -46,14 +79,8 @@ void set_rp_filter(int32_t state, char* dev) *colon_ptr = '\0'; sprintf( filename, "/proc/sys/net/ipv4/conf/%s/rp_filter", dev); + set_integer_file(filename, state); - if((f = fopen(filename, "w")) == NULL) - goto end; - - fprintf(f, "%d", state); - fclose(f); - -end: if ( colon_ptr != NULL ) *colon_ptr = ':'; } @@ -62,7 +89,6 @@ end: int32_t get_rp_filter(char *dev) { - FILE *f; int32_t state = 0; char filename[100], *colon_ptr; @@ -71,14 +97,8 @@ int32_t get_rp_filter(char *dev) *colon_ptr = '\0'; sprintf( filename, "/proc/sys/net/ipv4/conf/%s/rp_filter", dev); + state = get_integer_file(filename); - if((f = fopen(filename, "r")) == NULL) - goto end; - - fscanf(f, "%d", &state); - fclose(f); - -end: if ( colon_ptr != NULL ) *colon_ptr = ':'; @@ -89,7 +109,6 @@ end: void set_send_redirects( int32_t state, char* dev ) { - FILE *f; char filename[100], *colon_ptr; /* if given interface is an alias use parent interface */ @@ -97,14 +116,8 @@ void set_send_redirects( int32_t state, char* dev ) { *colon_ptr = '\0'; sprintf( filename, "/proc/sys/net/ipv4/conf/%s/send_redirects", dev); + set_integer_file(filename, state); - if((f = fopen(filename, "w")) == NULL) - goto end; - - fprintf(f, "%d", state); - fclose(f); - -end: if ( colon_ptr != NULL ) *colon_ptr = ':'; @@ -114,7 +127,6 @@ end: int32_t get_send_redirects( char *dev ) { - FILE *f; int32_t state = 0; char filename[100], *colon_ptr; @@ -123,14 +135,8 @@ int32_t get_send_redirects( char *dev ) { *colon_ptr = '\0'; sprintf( filename, "/proc/sys/net/ipv4/conf/%s/send_redirects", dev); + state = get_integer_file(filename); - if((f = fopen(filename, "r")) == NULL) - goto end; - - fscanf(f, "%d", &state); - fclose(f); - -end: if ( colon_ptr != NULL ) *colon_ptr = ':'; @@ -142,29 +148,14 @@ end: void set_forwarding(int32_t state) { - FILE *f; - - if((f = fopen("/proc/sys/net/ipv4/ip_forward", "w")) == NULL) - return; - - fprintf(f, "%d", state); - fclose(f); + set_integer_file("/proc/sys/net/ipv4/ip_forward", state); } int32_t get_forwarding(void) { - FILE *f; - int32_t state = 0; - - if((f = fopen("/proc/sys/net/ipv4/ip_forward", "r")) == NULL) - return 0; - - fscanf(f, "%d", &state); - fclose(f); - - return state; + return get_integer_file("/proc/sys/net/ipv4/ip_forward"); } -- 1.6.1.2.549.g547e

2 1

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

B.A.T.M.A.N February 2009