B.A.T.M.A.N June 2012

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

17 participants
55 discussions

[B.A.T.M.A.N.] [PATCH] batman-adv: fix skb->data assignment
by Antonio Quartulli 19 Jun '12

19 Jun '12

commit d2b6cc8e460494251442a877fcbc150faa175b4f upstream. skb_linearize(skb) possibly rearranges the skb internal data and then changes the skb->data pointer value. For this reason any other pointer in the code that was assigned skb->data before invoking skb_linearise(skb) must be re-assigned. In the current tt_query message handling code this is not done and therefore, in case of skb linearization, the pointer used to handle the packet header ends up in pointing to poisoned memory. The packet is then dropped but the translation-table mechanism is corrupted. Signed-off-by: Antonio Quartulli <ordex(a)autistici.org> Cc: <stable(a)vger.kernel.org> # 3.3 Signed-off-by: Sven Eckelmann <sven(a)narfation.org> --- net/batman-adv/routing.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/net/batman-adv/routing.c b/net/batman-adv/routing.c index 773e606..7d3ac4e 100644 --- a/net/batman-adv/routing.c +++ b/net/batman-adv/routing.c @@ -622,6 +622,8 @@ int recv_tt_query(struct sk_buff *skb, struct hard_iface *recv_if) * changes */ if (skb_linearize(skb) < 0) goto out; + /* skb_linearize() possibly changed skb->data */ + tt_query = (struct tt_query_packet *)skb->data; tt_len = tt_query->tt_data * sizeof(struct tt_change); -- 1.7.9.4

4 6

[B.A.T.M.A.N.] batman-adv: NS-3
by sebastien bindel 19 Jun '12

19 Jun '12

Hello everybody, I'm a student who implementing batman-adv in the NS-3 simulator. But I have some questions. First : To determine if a link is unidirectionnal or not you use the Echo Quality ? Second : To remove entry in originator table you use a garbage collector like in Batman III ? Third : Why in local and global translation you use roaming, why not just remove and add the node ? Thanks, I will have some other question of course in the futher. Enjoys

2 1

[B.A.T.M.A.N.] [PATCH] batman-adv: fix skb->data assignment
by Antonio Quartulli 19 Jun '12

19 Jun '12

commit d2b6cc8e460494251442a877fcbc150faa175b4f upstream. skb_linearize(skb) possibly rearranges the skb internal data and then changes the skb->data pointer value. For this reason any other pointer in the code that was assigned skb->data before invoking skb_linearise(skb) must be re-assigned. In the current tt_query message handling code this is not done and therefore, in case of skb linearization, the pointer used to handle the packet header ends up in pointing to poisoned memory. The packet is then dropped but the translation-table mechanism is corrupted. Signed-off-by: Antonio Quartulli <ordex(a)autistici.org> Cc: <stable(a)vger.kernel.org> # 3.1 Cc: <stable(a)vger.kernel.org> # 3.2 Signed-off-by: Sven Eckelmann <sven(a)narfation.org> --- net/batman-adv/routing.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/net/batman-adv/routing.c b/net/batman-adv/routing.c index 0f32c81..55136e5 100644 --- a/net/batman-adv/routing.c +++ b/net/batman-adv/routing.c @@ -1246,6 +1246,8 @@ int recv_tt_query(struct sk_buff *skb, struct hard_iface *recv_if) /* packet needs to be linearised to access the TT changes */ if (skb_linearize(skb) < 0) goto out; + /* skb_linearize() possibly changed skb->data */ + tt_query = (struct tt_query_packet *)skb->data; if (is_my_mac(tt_query->dst)) handle_tt_response(bat_priv, tt_query); -- 1.7.9.4

1 0

[B.A.T.M.A.N.] pull request: batman-adv 2012-06-18
by Antonio Quartulli 19 Jun '12

19 Jun '12

Hello David, here is our first set of changes intended for next-next/linux-3.6. Patch 2 fixes a major bug in the TranslationTable code where the old value of skb->data is used for memory access even if the data was relocated. Patches 4, 10, 11, 13, 14 are endianess-related cleanups that we wrote thanks to Al Viro's advice and help. Thanks to Martin Hundebøll batman-adv now supports the ethtool API and can export several counters that are specific to our module (patch 5). Then patch 16 improves the routing protocol API by making part of the TranslationTable code routing agnostic. The rest are minor fixes and other cleanups. Thank you very much, Antonio The following changes since commit 6fac262526ee91ee66210b8919a4297dcf7d544e: ipv4: Cap ADVMSS metric in the FIB rather than the routing cache. (2012-06-17 19:47:34 -0700) 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 dafe94b278e052c3901b137fe6f666f8f92d839a: batman-adv: only store changed gw_bandwidth values (2012-06-18 18:01:07 +0200) ---------------------------------------------------------------- Included changes: * major skb->data pointer usage fix * interval version update * added get_ethtool_stats() support * endianess clean up * routing protocol API improvement wrt TT commit code * fix locking in hash table code * minor cleanups and fixes ---------------------------------------------------------------- Al Viro (5): batman-adv: get rid of pointless cast in memcpy() batman-adv: trivial endianness annotations batman-adv: keep batman_ogm_packet ->seqno net-endian all along batman-adv: don't bother flipping ->tt_data batman-adv: don't bother flipping ->tt_crc Antonio Quartulli (3): batman-adv: fix skb->data assignment batman-adv: convert bat_priv->tt_crc from atomic_t to uint16_t batman-adv: use DBG_ALL in log_level sysfs definition Marek Lindner (5): batman-adv: avoid characters requiring shell escapes in protocol names batman-adv: ignore trailing CR when comparing protocol names batman-adv: return added entries instead of number of possibly added entries batman-adv: turn tt commit code into routing protocol agnostic API batman-adv: only store changed gw_bandwidth values Martin Hundebøll (1): batman-adv: Add get_ethtool_stats() support Matthias Schiffer (2): batman-adv: fix visualization output without neighbors on the primary interface batman-adv: fix locking in hash_add() Sven Eckelmann (3): batman-adv: update internal version number batman-adv: Initialize lockdep class keys for hashes batman-adv: Return error codes instead of -1 on failures Documentation/networking/batman-adv.txt | 5 + net/batman-adv/bat_debugfs.c | 11 +- net/batman-adv/bat_iv_ogm.c | 63 ++++++------ net/batman-adv/bat_sysfs.c | 4 +- net/batman-adv/bridge_loop_avoidance.c | 27 +++-- net/batman-adv/gateway_common.c | 3 + net/batman-adv/hard-interface.c | 4 +- net/batman-adv/hash.c | 9 ++ net/batman-adv/hash.h | 19 ++-- net/batman-adv/icmp_socket.c | 4 +- net/batman-adv/main.c | 42 +++++--- net/batman-adv/main.h | 29 +++++- net/batman-adv/originator.c | 18 ++-- net/batman-adv/packet.h | 18 ++-- net/batman-adv/routing.c | 23 +++-- net/batman-adv/send.c | 74 +------------- net/batman-adv/soft-interface.c | 66 +++++++++++- net/batman-adv/translation-table.c | 168 ++++++++++++++++++++++++------- net/batman-adv/translation-table.h | 7 +- net/batman-adv/types.h | 22 +++- net/batman-adv/vis.c | 29 ++++-- 21 files changed, 420 insertions(+), 225 deletions(-)

2 20

[B.A.T.M.A.N.] [PATCH] batman-adv: don't redirect tt_request for me
by Antonio Quartulli 17 Jun '12

17 Jun '12

If a tt_request is directed to me, it never has to be redirected Signed-off-by: Antonio Quartulli <ordex(a)autistici.org> --- translation-table.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) diff --git a/translation-table.c b/translation-table.c index 2a6d7d6..c32c5f6 100644 --- a/translation-table.c +++ b/translation-table.c @@ -1648,7 +1648,7 @@ out: } -static bool +static void batadv_send_my_tt_response(struct batadv_priv *bat_priv, struct batadv_tt_query_packet *tt_request) { @@ -1656,7 +1656,6 @@ batadv_send_my_tt_response(struct batadv_priv *bat_priv, struct batadv_neigh_node *neigh_node = NULL; struct batadv_hard_iface *primary_if = NULL; uint8_t my_ttvn, req_ttvn, ttvn; - int ret = false; unsigned char *tt_buff; bool full_table; uint16_t tt_len, tt_tot; @@ -1751,7 +1750,6 @@ batadv_send_my_tt_response(struct batadv_priv *bat_priv, batadv_inc_counter(bat_priv, BATADV_CNT_TT_RESPONSE_TX); batadv_send_skb_packet(skb, neigh_node->if_incoming, neigh_node->addr); - ret = true; goto out; unlock: @@ -1765,8 +1763,6 @@ out: batadv_hardif_free_ref(primary_if); if (!ret) kfree_skb(skb); - /* This packet was for me, so it doesn't need to be re-routed */ - return true; } bool batadv_send_tt_response(struct batadv_priv *bat_priv, @@ -1777,7 +1773,9 @@ bool batadv_send_tt_response(struct batadv_priv *bat_priv, if (batadv_bla_is_backbone_gw_orig(bat_priv, tt_request->src)) return true; - return batadv_send_my_tt_response(bat_priv, tt_request); + batadv_send_my_tt_response(bat_priv, tt_request); + /* This packet was for me, so it never needs to be re-routed */ + return true; } else { return batadv_send_other_tt_response(bat_priv, tt_request); } -- 1.7.9.4

2 9

[B.A.T.M.A.N.] batman-adv 2012.2.0 released
by Marek Lindner 16 Jun '12

16 Jun '12

Today, the B.A.T.M.A.N. team releases batman-adv 2012.2.0, packed with new features and improvements in various subsystems as well as the usual set of fixes and cleanups. As the kernel module always depends on the Linux kernel it was compiled against, it does not make sense to provide binaries on our website. As usual, you will find the signed tarballs in our download section: http://downloads.open-mesh.org/batman/releases/batman-adv-2012.2.0/ as well as prepackaged binaries in your distribution. Important changes ----------------- This release comes with a completely rewritten bridge loop avoidance (also known as bridge loop avoidance II). The concept of the first bridge loop avoidance was simple and worked well in small size LANs while larger networks suffered from the overhead. The new concept is very different and requires changes in your batman-adv configuration. If you were using the bridge loop avoidance you should consult our documentation before upgrading your network(s). The default "per hop" penalty was increased to encourage batman-adv to take shorter routes. If you notice altered routing behavior and are unhappy with the result you should revisit the hop penalty configuration option. Thanks ------ Thanks to all people sending in patches: * Antonio Quartulli <ordex(a)autistici.org> * Danny Kukawka <danny.kukawka(a)bisect.de> * David S. Miller <davem(a)davemloft.net> * Eric Dumazet <eric.dumazet(a)gmail.com> * Linus Luessing <linus.luessing(a)web.de> * Marek Lindner <lindner_marek(a)yahoo.de> * Simon Wunderlich <siwu(a)hrz.tu-chemnitz.de> * Sven Eckelmann <sven(a)narfation.org> * Xabier Rodriguez <xrb(a)kalrong.net> and to all those that supported us with good advice, code review and/or rigorous testing: * Al Viro <viro(a)ZenIV.linux.org.uk> * David Laight <David.Laight(a)aculab.com> * Jo-Philipp Wich <jow(a)openwrt.org> * Martin Hundebøll <martin(a)hundeboll.net> batman-adv ---------- The new bridge loop avoidance certainly is the most prominent of the new features this release has to offer (completely replacing the old bridge loop avoidance mechanism). The major design goals were performance and scalability. Since the old implementation was relying on a single gateway to be the master gateway to the LAN for everybody else it created a performance bottleneck. Furthermore, the LAN was used to send traffic to and receive data from the master gateway. With the number of bridged gateways grew the amount of broadcast traffic in the LAN. The new bridge loop avoidance splits the client responsibility amongst all participating gateways. Each gateway "claims" the clients it feels responsible for and ignores the traffic from all other clients to avoid the bridge loop. It also is able to handle multiple VLANs on top of the batX interface connected to different topologies. All details about the loop avoidance mechanism are explained in our documentation section. The routing code also received lots of attention: The recently added routing protocol abstraction was further polished and extended to better accommodate the needs of alternative routing protocols. B.A.T.M.A.N. IV protocol has been enhanced with an additional flag to apply stricter forwarding rules to OGMs which allows the protocol to avert routing loops in certain corner cases. Also, the B.A.T.M.A.N. IV sequence numbers are now randomized at startup to reduce the probability of a collision and thus, slowing down the protocol in the startup phase. Rerouting of unicast payload packets during a roaming phase is handled with greater efficiency to avoid as much packet loss while roaming as possible. All manual HZ-jiffies-calculations have been replaced with the in-kernel jiffies_to_msecs() function. To facilitate comprehension of the code base the ETH_ALEN macro is used instead of hardcoded numeric constants. The batman-adv internal bitarray operations have been converted to the efficient in-kernel bitmap operations. It was discovered that the TT-Request packet did not always send the tt-crc field in network byte order, thereby invalidating the packet. This has been fixed alongside the suboptimal DHCP option list parser used by the gateway extension. OGM sequence numbers now are always printed as unsigned long to avoid misinterpretation while printing the numbers in the debug log. batctl ------ The batctl utility supports the new bridge loop avoidance by providing an option to conveniently enable/disable the bridge loop avoidance and exports the bridge loop avoidance claim table. It will also warn about features that haven't been compiled into batman-adv such as debug log and bridge loop avoidance. The mini-tcpdump learned to display the newly added 'not best hop flag' when parsing OGMs. Happy routing, The B.A.T.M.A.N. team

2 1

[B.A.T.M.A.N.] [PATCH] batman-adv: fix skb->data assignment
by Antonio Quartulli 15 Jun '12

15 Jun '12

skb_linearize(skb) possibly rearranges the skb internal data and then changes the skb->data pointer value. For this reason any other pointer in the code that was assigned skb->data before invoking skb_linearise(skb) must be re-assigned. In the current tt_query message handling code this is not done and therefore, in case of skb linearization, the pointer used to handle the packet header ends up in pointing to poisoned memory. The packet is then dropped but the translation-table mechanism is corrupted. Signed-off-by: Antonio Quartulli <ordex(a)autistici.org> --- *** this patch is an important fix and it is for maint *** routing.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/routing.c b/routing.c index be068f3..b79e42e 100644 --- a/routing.c +++ b/routing.c @@ -638,6 +638,8 @@ int batadv_recv_tt_query(struct sk_buff *skb, struct batadv_hard_iface *recv_if) */ if (skb_linearize(skb) < 0) goto out; + /* skb_linearize() possibly changed skb->data */ + tt_query = (struct batadv_tt_query_packet *)skb->data; tt_size = batadv_tt_len(ntohs(tt_query->tt_data)); -- 1.7.9.4

3 3

[B.A.T.M.A.N.] [PATCH] batman-adv: Correctly check return value from debugfs_create_dir
by Martin Hundebøll 12 Jun '12

12 Jun '12

The setup code in batadv_socket_setup() and debug_log_setup() wrongly assumes that debugfs_create_file() returns !0 on error. Since it actually returns a pointer on success[1], the following check should be inverted. Also, use the return value from the two setup functions in batadv_debugfs_add_meshif(). [1] http://www.fsl.cs.sunysb.edu/kernel-api/re464.html Signed-off-by: Martin Hundebøll <martin(a)hundeboll.net> --- bat_debugfs.c | 11 +++++++---- icmp_socket.c | 4 ++-- 2 files changed, 9 insertions(+), 6 deletions(-) diff --git a/bat_debugfs.c b/bat_debugfs.c index cd636db..bd9325d 100644 --- a/bat_debugfs.c +++ b/bat_debugfs.c @@ -193,13 +193,13 @@ static int debug_log_setup(struct bat_priv *bat_priv) d = debugfs_create_file("log", S_IFREG | S_IRUSR, bat_priv->debug_dir, bat_priv, &log_fops); - if (d) + if (!d) goto err; return 0; err: - return 1; + return -ENOMEM; } static void debug_log_cleanup(struct bat_priv *bat_priv) @@ -347,8 +347,11 @@ int batadv_debugfs_add_meshif(struct net_device *dev) if (!bat_priv->debug_dir) goto out; - batadv_socket_setup(bat_priv); - debug_log_setup(bat_priv); + if (batadv_socket_setup(bat_priv)) + goto rem_attr; + + if (debug_log_setup(bat_priv)) + goto rem_attr; for (bat_debug = mesh_debuginfos; *bat_debug; ++bat_debug) { file = debugfs_create_file(((*bat_debug)->attr).name, diff --git a/icmp_socket.c b/icmp_socket.c index 3fad5aa..40c5e18 100644 --- a/icmp_socket.c +++ b/icmp_socket.c @@ -283,13 +283,13 @@ int batadv_socket_setup(struct bat_priv *bat_priv) d = debugfs_create_file(ICMP_SOCKET, S_IFREG | S_IWUSR | S_IRUSR, bat_priv->debug_dir, bat_priv, &fops); - if (d) + if (!d) goto err; return 0; err: - return 1; + return -ENOMEM; } static void bat_socket_add_packet(struct socket_client *socket_client, -- 1.7.10.2

3 2

[B.A.T.M.A.N.] [PATCH] batman-adv: Remove bat_ prefix from bat_{debugfs, sysfs}.{c, h}
by Sven Eckelmann 12 Jun '12

12 Jun '12

The "bat_" prefix in the source files implementing the batman-adv sysfs and debugfs interface doesn't have a special meaning and are only used by these files and files that implement the actual B.A.T.M.A.N. path finding algorithm. The prefix is better suited to mark files that are used to implement the main part of the path finding. All other files should not use it and therefore gets renamed. Signed-off-by: Sven Eckelmann <sven(a)narfation.org> --- Makefile.kbuild | 4 ++-- bat_debugfs.c => debugfs.c | 2 +- bat_debugfs.h => debugfs.h | 0 gateway_client.c | 2 +- hard-interface.c | 2 +- main.c | 4 ++-- soft-interface.c | 4 ++-- bat_sysfs.c => sysfs.c | 2 +- bat_sysfs.h => sysfs.h | 0 9 files changed, 10 insertions(+), 10 deletions(-) rename bat_debugfs.c => debugfs.c (99%) rename bat_debugfs.h => debugfs.h (100%) rename bat_sysfs.c => sysfs.c (99%) rename bat_sysfs.h => sysfs.h (100%) diff --git a/Makefile.kbuild b/Makefile.kbuild index 6d5c194..8676d2b 100644 --- a/Makefile.kbuild +++ b/Makefile.kbuild @@ -19,11 +19,10 @@ # obj-$(CONFIG_BATMAN_ADV) += batman-adv.o -batman-adv-y += bat_debugfs.o batman-adv-y += bat_iv_ogm.o -batman-adv-y += bat_sysfs.o batman-adv-y += bitarray.o batman-adv-$(CONFIG_BATMAN_ADV_BLA) += bridge_loop_avoidance.o +batman-adv-y += debugfs.o batman-adv-y += gateway_client.o batman-adv-y += gateway_common.o batman-adv-y += hard-interface.o @@ -35,6 +34,7 @@ batman-adv-y += ring_buffer.o batman-adv-y += routing.o batman-adv-y += send.o batman-adv-y += soft-interface.o +batman-adv-y += sysfs.o batman-adv-y += translation-table.o batman-adv-y += unicast.o batman-adv-y += vis.o diff --git a/bat_debugfs.c b/debugfs.c similarity index 99% rename from bat_debugfs.c rename to debugfs.c index f7d161f..de610ba 100644 --- a/bat_debugfs.c +++ b/debugfs.c @@ -21,7 +21,7 @@ #include <linux/debugfs.h> -#include "bat_debugfs.h" +#include "debugfs.h" #include "translation-table.h" #include "originator.h" #include "hard-interface.h" diff --git a/bat_debugfs.h b/debugfs.h similarity index 100% rename from bat_debugfs.h rename to debugfs.h diff --git a/gateway_client.c b/gateway_client.c index 13bca58..6fe5cc8 100644 --- a/gateway_client.c +++ b/gateway_client.c @@ -18,7 +18,7 @@ */ #include "main.h" -#include "bat_sysfs.h" +#include "sysfs.h" #include "gateway_client.h" #include "gateway_common.h" #include "hard-interface.h" diff --git a/hard-interface.c b/hard-interface.c index f18ccc3..14e17ea 100644 --- a/hard-interface.c +++ b/hard-interface.c @@ -23,7 +23,7 @@ #include "send.h" #include "translation-table.h" #include "routing.h" -#include "bat_sysfs.h" +#include "sysfs.h" #include "originator.h" #include "hash.h" #include "bridge_loop_avoidance.h" diff --git a/main.c b/main.c index 17dcdd9..13c88b2 100644 --- a/main.c +++ b/main.c @@ -18,8 +18,8 @@ */ #include "main.h" -#include "bat_sysfs.h" -#include "bat_debugfs.h" +#include "sysfs.h" +#include "debugfs.h" #include "routing.h" #include "send.h" #include "originator.h" diff --git a/soft-interface.c b/soft-interface.c index 7a7d821..b7c655c 100644 --- a/soft-interface.c +++ b/soft-interface.c @@ -22,12 +22,12 @@ #include "hard-interface.h" #include "routing.h" #include "send.h" -#include "bat_debugfs.h" +#include "debugfs.h" #include "translation-table.h" #include "hash.h" #include "gateway_common.h" #include "gateway_client.h" -#include "bat_sysfs.h" +#include "sysfs.h" #include "originator.h" #include <linux/slab.h> #include <linux/ethtool.h> diff --git a/bat_sysfs.c b/sysfs.c similarity index 99% rename from bat_sysfs.c rename to sysfs.c index ca0bdda..6d519ee 100644 --- a/bat_sysfs.c +++ b/sysfs.c @@ -18,7 +18,7 @@ */ #include "main.h" -#include "bat_sysfs.h" +#include "sysfs.h" #include "translation-table.h" #include "originator.h" #include "hard-interface.h" diff --git a/bat_sysfs.h b/sysfs.h similarity index 100% rename from bat_sysfs.h rename to sysfs.h -- 1.7.10

2 1

[B.A.T.M.A.N.] [PATCH] batman-adv: Fix include of linux/kernel.h
by Sven Eckelmann 12 Jun '12

12 Jun '12

This is a non-fatal include problem that was only detected by cppcheck. The regression was introduced by 4738c60318e757cf0062476e2095caa5b6970a8d Signed-off-by: Sven Eckelmann <sven(a)narfation.org> --- types.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/types.h b/types.h index fd86a44..2141c13 100644 --- a/types.h +++ b/types.h @@ -22,7 +22,7 @@ #include "packet.h" #include "bitarray.h" -#include "linux/kernel.h" +#include <linux/kernel.h> #define BATADV_HEADER_LEN \ (ETH_HLEN + max(sizeof(struct batadv_unicast_packet), \ -- 1.7.10

2 2

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

B.A.T.M.A.N June 2012