B.A.T.M.A.N

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

15 participants
4308 discussions

[B.A.T.M.A.N.] [PATCH maint] batman-adv: increase orig refcount when storing ref in gw_node
by Antonio Quartulli 01 May '14

01 May '14

From: Antonio Quartulli <antonio(a)open-mesh.com> A pointer to the orig_node representing a bat-gateway is stored in the gw_node->orig_node member, but the refcount for such orig_node is never increased. This leads to memory faults when gw_node->orig_node is accessed and the originator has already been freed. Fix this by increasing the refcount on gw_node creation and decreasing it on gw_node free. Signed-off-by: Antonio Quartulli <antonio(a)open-mesh.com> --- gateway_client.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/gateway_client.c b/gateway_client.c index d7fafc1..cfaa2fc 100644 --- a/gateway_client.c +++ b/gateway_client.c @@ -42,8 +42,10 @@ static void batadv_gw_node_free_ref(struct batadv_gw_node *gw_node) { - if (atomic_dec_and_test(&gw_node->refcount)) + if (atomic_dec_and_test(&gw_node->refcount)) { + batadv_orig_node_free_ref(gw_node->orig_node); kfree_rcu(gw_node, rcu); + } } static struct batadv_gw_node * @@ -406,6 +408,9 @@ static void batadv_gw_node_add(struct batadv_priv *bat_priv, if (gateway->bandwidth_down == 0) return; + if (!atomic_inc_not_zero(&orig_node->refcount)) + return; + gw_node = kzalloc(sizeof(*gw_node), GFP_ATOMIC); if (!gw_node) return; -- 1.8.3.2

1 0

[B.A.T.M.A.N.] Batman-adv high TQ and Low throughput
by cmsv 30 Apr '14

30 Apr '14

Recently i have experienced something quite usual to my experience and that was not happening previously in the same scenario where it is happening now. Here is the example: 2 nodes at with adhoc at 2.4ghz at a distance of 415 meters one uses an omni 15 dbi antenna at 18dbi and another 24 dbi grid at 16 dbi. This link was previously providing up to 21 mbit being 17 mbit a the lowest. Batman-adv TQ value was always around 230 being many times above 250. However this has recently changed the same link with 230~250 TQ and the same txpower is now outputting 1.93 Mbits/sec I have used iperf for bandwidth tests. Batman-adv version is 2013.4. Although i have considered several types of scenarios that can be causing this issue i have also tested against some which leaves me puzzled in regards to the presented TQ quality and dBm values as the nodes dBm are sometimes bad (ie: -80 ~ -90) One of the main questions is: shouldn't batman-adv TQ reflect or be influenced by dBm values ? How come with -90 dBm i get batman-adv to show me 230 TQ ? Also any feedback based on your experience regarding this type of scenario is welcome. This situation happened overnight without any access point new configurations or physical setup changes. -- Site: http://wirelesspt.net Mesh: http://tinyurl.com/wirelesspt Admin: http://wirelesspt.net/wiki/Cmsv Twitter: http://twitter.com/wirelesspt Youtube: https://youtube.com/wirelesspt Facebook: https://www.facebook.com/wirelesspt Suporte técnico via sms: 91 19 11 798 Donativos/Paypal: http://tinyurl.com/doar-verba Chave publica PGP/SSH: http://wirelesspt.net/arquivos/pk Email ao abrigo de: https://creativecommons.org/licenses/by-nc-sa/3.0/pt/

4 5

[B.A.T.M.A.N.] MTU issue
by Saurabh Chakrabarti 24 Apr '14

24 Apr '14

Following the Troubleshooting guide, I have changed the MTU for bat0 interface to 1524. I'm getting this message in my syslog: Jan 1 00:00:53 localhost kernel: [ 26.899762] batman_adv: bat0: The MTU of interface wlan0 is too small (1524) to handle the transport of batman-adv packets. Packets going over this interface will be fragmented on layer2 which could impact the performance. Setting the MTU to 1560 would solve the problem. I'm not seeing any apparent performance issues, but do I really need to make the MTU 1560? -saurabh

2 1

[B.A.T.M.A.N.] batman-adv: Multiple NULL derefence bugs (+ 1 bonus)
by Sven Eckelmann 23 Apr '14

23 Apr '14

All taken from batman-adv master (b08dc3d9452d28944e7db09c2005c633b63f55e9) 1. multicast.c (dereference of NULL orig_node on at least the first round): ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ orig_node = NULL before the first entry in hlist_for_each_entry_rcu. NULL dereferenced in atomic_inc_not_zero(...) to get the refcount Was introduced with 45dab5f5509023946793fc9d1d0566c12ca8bb52 ("batman-adv: fix more code style") Code section: ============ static struct batadv_orig_node * batadv_mcast_forw_unsnoop_node_get(struct batadv_priv *bat_priv) { struct batadv_orig_node *tmp_orig_node, *orig_node = NULL; rcu_read_lock(); hlist_for_each_entry_rcu(tmp_orig_node, &bat_priv->mcast.want_all_unsnoopables_list, mcast_want_all_unsnoopables_node) { if (!atomic_inc_not_zero(&orig_node->refcount)) continue; orig_node = tmp_orig_node; break; } rcu_read_unlock(); return orig_node; } Solution: ======== Change atomic_inc_not_zero from orig_node to tmp_orig_node 2. multicast.c (dereference of NULL orig_node on at least the first round): ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ orig_node = NULL before the first entry in hlist_for_each_entry_rcu. NULL dereferenced in atomic_inc_not_zero(...) to get the refcount Was introduced with 45dab5f5509023946793fc9d1d0566c12ca8bb52 ("batman-adv: fix more code style") Code section: ============ static struct batadv_orig_node * batadv_mcast_forw_ipv6_node_get(struct batadv_priv *bat_priv) { struct batadv_orig_node *tmp_orig_node, *orig_node = NULL; rcu_read_lock(); hlist_for_each_entry_rcu(tmp_orig_node, &bat_priv->mcast.want_all_ipv6_list, mcast_want_all_ipv6_node) { if (!atomic_inc_not_zero(&orig_node->refcount)) continue; orig_node = tmp_orig_node; break; } rcu_read_unlock(); return orig_node; } Solution: ======== Change atomic_inc_not_zero from orig_node to tmp_orig_node 3. multicast.c (dereference of NULL orig_node on at least the first round): ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ orig_node = NULL before the first entry in hlist_for_each_entry_rcu. NULL dereferenced in atomic_inc_not_zero(...) to get the refcount Was introduced with 45dab5f5509023946793fc9d1d0566c12ca8bb52 ("batman-adv: fix more code style") Code section: ============ static struct batadv_orig_node * batadv_mcast_forw_ipv4_node_get(struct batadv_priv *bat_priv) { struct batadv_orig_node *tmp_orig_node, *orig_node = NULL; rcu_read_lock(); hlist_for_each_entry_rcu(tmp_orig_node, &bat_priv->mcast.want_all_ipv4_list, mcast_want_all_ipv4_node) { if (!atomic_inc_not_zero(&orig_node->refcount)) continue; orig_node = tmp_orig_node; break; } rcu_read_unlock(); return orig_node; } Solution: ======== Change atomic_inc_not_zero from orig_node to tmp_orig_node 4. originator.c (Indirect hard_iface NULL dereference): ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ hard_iface is NULL and goto out is made. batadv_hardif_free_ref doesn't check for NULL before dereferencing it to get to refcount. Introduced in f13f960797fd1969b3c0470cc97435ddfb6aecb4 ("batman-adv: add debugfs support to view multiif tables"). Code section: ============ int batadv_orig_hardif_seq_print_text(struct seq_file *seq, void *offset) { struct net_device *net_dev = (struct net_device *)seq->private; struct batadv_hard_iface *hard_iface; struct batadv_priv *bat_priv; hard_iface = batadv_hardif_get_by_netdev(net_dev); if (!hard_iface || !hard_iface->soft_iface) { seq_puts(seq, "Interface not known to B.A.T.M.A.N.\n"); goto out; } [...] out: batadv_hardif_free_ref(hard_iface); return 0; } Solution: ======== if (hard_iface) batadv_hardif_free_ref(hard_iface); Related code section: ==================== static inline void batadv_hardif_free_ref_now(struct batadv_hard_iface *hard_iface) { if (atomic_dec_and_test(&hard_iface->refcount)) batadv_hardif_free_rcu(&hard_iface->rcu); } 5. BONUS: fragmentation.c (Impossible code section marked likely): ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ frag_entry_curr is iterator used for list chain->head. This loop is only stopped when the list ends and frag_entry_curr is NULL. The next section following the loop says that it is likely that frag_entry_curr is NOT NULL. This is a contradiction. Introduced in 9b3eab61754d74a93c9840c296013fe3b4a1b606 ("batman-adv: Receive fragmented packets and merge") Code section: ============ static bool batadv_frag_insert_packet(struct batadv_orig_node *orig_node, struct sk_buff *skb, struct hlist_head *chain_out) { [...] /* Find the position for the new fragment. */ hlist_for_each_entry(frag_entry_curr, &chain->head, list) { /* Drop packet if fragment already exists. */ if (frag_entry_curr->no == frag_entry_new->no) goto err_unlock; /* Order fragments from highest to lowest. */ if (frag_entry_curr->no < frag_entry_new->no) { hlist_add_before(&frag_entry_new->list, &frag_entry_curr->list); chain->size += skb->len - hdr_size; chain->timestamp = jiffies; ret = true; goto out; } } /* Reached the end of the list, so insert after 'frag_entry_curr'. */ if (likely(frag_entry_curr)) { hlist_add_after(&frag_entry_curr->list, &frag_entry_new->list); chain->size += skb->len - hdr_size; chain->timestamp = jiffies; ret = true; } [...] } Solution: ======== Rewrite this part to track the last seen batadv_frag_list_entry and add it after that. The empty list should already handled earlier in the scope starting with if (batadv_frag_init_chain(chain, seqno)) { Related code section: ==================== #define hlist_for_each_entry(pos, head, member) \ for (pos = hlist_entry_safe((head)->first, typeof(*(pos)), member);\ pos; \ pos = hlist_entry_safe((pos)->member.next, typeof(*(pos)), member))

1 0

[B.A.T.M.A.N.] [PATCH] batctl: mark multicast-mode switch as potentially compiled out
by Linus Lüssing 21 Apr '14

21 Apr '14

To give users an additional hint if the multicast feature is not compiled in, to let them know what they can do to change that. Signed-off-by: Linus Lüssing <linus.luessing(a)web.de> --- functions.c | 1 + 1 file changed, 1 insertion(+) diff --git a/functions.c b/functions.c index 117dcff..b9ccaff 100644 --- a/functions.c +++ b/functions.c @@ -63,6 +63,7 @@ const char *fs_compile_out_param[] = { batctl_settings[BATCTL_SETTINGS_BLA].sysfs_name, batctl_settings[BATCTL_SETTINGS_DAT].sysfs_name, batctl_settings[BATCTL_SETTINGS_NETWORK_CODING].sysfs_name, + batctl_settings[BATCTL_SETTINGS_MULTICAST_MODE].sysfs_name, batctl_debug_tables[BATCTL_TABLE_BLA_CLAIMS].debugfs_name, batctl_debug_tables[BATCTL_TABLE_BLA_BACKBONES].debugfs_name, batctl_debug_tables[BATCTL_TABLE_DAT].debugfs_name, -- 1.7.10.4

2 1

[B.A.T.M.A.N.] Mesh network splits when a single node goes down
by Saurabh Chakrabarti 19 Apr '14

19 Apr '14

We have a mesh network running on an adhoc wifi network. All the nodes are in static position. The problem is that when one node goes down, the mesh breaks and the bat0 interfaces of the nodes are not reachable (although, strangely, the nodes are reachable via the wlan0 interfaces). Also, the node that has gone down is in one corner of the mesh. When one of the nodes is down, linaro@raspberry:~$ sudo batctl o [B.A.T.M.A.N. adv 2014.1.0, MainIF/MAC: wlan0/10:fe:ed:13:f2:10 (bat0 BATMAN_IV)] Originator last-seen (#/255) Nexthop [outgoingIF]: Potential nexthops ... 10:fe:ed:14:d5:c4 0.090s ( 9) 10:fe:ed:14:d5:c4 [ wlan0]: 10:fe:ed:14:d5:c4 ( 9) c0:4a:00:19:91:ef 0.610s ( 8) c0:4a:00:19:91:ef [ wlan0]: c0:4a:00:19:91:ef ( 8) c0:4a:00:18:db:83 0.410s ( 10) c0:4a:00:18:db:83 [ wlan0]: c0:4a:00:18:db:83 ( 10) 64:70:02:1c:bc:d0 0.090s ( 6) 64:70:02:1c:bc:d0 [ wlan0]: 64:70:02:1c:bc:d0 ( 6) linaro@raspberry:~$ sudo batctl o [B.A.T.M.A.N. adv 2014.1.0, MainIF/MAC: wlan0/c0:4a:00:18:db:83 (bat0 BATMAN_IV)] Originator last-seen (#/255) Nexthop [outgoingIF]: Potential nexthops ... 10:fe:ed:14:d5:c4 0.210s (246) 10:fe:ed:14:d5:c4 [ wlan0]: c0:4a:00:19:91:ef (199) 10:fe:ed:14:d5:c4 (246) c0:4a:00:19:91:ef 0.830s (226) c0:4a:00:19:91:ef [ wlan0]: 10:fe:ed:14:d5:c4 (196) c0:4a:00:19:91:ef (226) 64:70:02:1c:bc:d0 0.520s (163) 64:70:02:1c:bc:d0 [ wlan0]: c0:4a:00:19:91:ef ( 0) 64:70:02:1c:bc:d0 (163) When the node thats down comes back up, linaro@raspberry:~$ sudo batctl o [B.A.T.M.A.N. adv 2014.1.0, MainIF/MAC: wlan0/64:66:b3:18:c9:79 (bat0 BATMAN_IV)] Originator last-seen (#/255) Nexthop [outgoingIF]: Potential nexthops ... 10:fe:ed:13:f2:10 0.890s (255) 10:fe:ed:13:f2:10 [ wlan0]: c0:4a:00:19:91:ef (196) c0:4a:00:19:a9:7f (158) 10:fe:ed:14:d5:c4 ( 0) c0:4a:00:18:db:83 (200) 10:fe:ed:14:95:09 (208) c0:4a:00:18:07:f0 (167) 10:fe:ed:13:f2:10 (255) 10:fe:ed:14:d5:c4 0.220s (181) c0:4a:00:19:91:ef [ wlan0]: 10:fe:ed:14:95:09 ( 0) 10:fe:ed:13:f2:10 (160) c0:4a:00:19:a9:7f (127) c0:4a:00:18:db:83 (163) c0:4a:00:18:07:f0 (141) c0:4a:00:19:91:ef (181) 10:fe:ed:14:d5:c4 ( 0) c0:4a:00:19:a9:7f 0.680s (224) c0:4a:00:19:a9:7f [ wlan0]: 10:fe:ed:14:d5:c4 (100) 10:fe:ed:14:95:09 (168) 10:fe:ed:13:f2:10 (171) c0:4a:00:18:db:83 (200) c0:4a:00:18:07:f0 (179) c0:4a:00:19:91:ef (201) c0:4a:00:19:a9:7f (224) c0:4a:00:19:91:ef 0.440s (233) c0:4a:00:19:91:ef [ wlan0]: c0:4a:00:19:a9:7f (140) 10:fe:ed:14:d5:c4 ( 0) 10:fe:ed:13:f2:10 (156) 10:fe:ed:14:95:09 (194) c0:4a:00:18:07:f0 (159) c0:4a:00:18:db:83 (156) c0:4a:00:19:91:ef (233) c0:4a:00:18:db:83 0.160s (230) c0:4a:00:18:db:83 [ wlan0]: c0:4a:00:19:a9:7f (145) c0:4a:00:18:07:f0 (189) 10:fe:ed:14:d5:c4 ( 0) 10:fe:ed:13:f2:10 (195) 10:fe:ed:14:95:09 (177) c0:4a:00:19:91:ef (169) c0:4a:00:18:db:83 (230) 10:fe:ed:14:95:09 0.110s (235) 10:fe:ed:14:95:09 [ wlan0]: c0:4a:00:19:a9:7f (124) 10:fe:ed:14:d5:c4 ( 0) c0:4a:00:18:db:83 (158) c0:4a:00:18:07:f0 (183) 10:fe:ed:13:f2:10 (190) c0:4a:00:19:91:ef (169) 10:fe:ed:14:95:09 (235) c0:4a:00:18:07:f0 0.040s (253) c0:4a:00:18:07:f0 [ wlan0]: c0:4a:00:19:a9:7f (171) 10:fe:ed:14:d5:c4 ( 0) c0:4a:00:18:db:83 (199) 10:fe:ed:14:95:09 (195) 10:fe:ed:13:f2:10 (199) c0:4a:00:19:91:ef (205) c0:4a:00:18:07:f0 (253) 64:70:02:1c:bc:d0 0.770s (171) c0:4a:00:19:a9:7f [ wlan0]: 10:fe:ed:14:95:09 (128) c0:4a:00:18:07:f0 ( 0) 10:fe:ed:14:d5:c4 ( 0) c0:4a:00:19:91:ef ( 0) c0:4a:00:18:db:83 ( 0) c0:4a:00:19:a9:7f (171) 10:fe:ed:13:f2:10 (134) -saurabh

1 0

[B.A.T.M.A.N.] [RFCv2 0/6] Export the expected throughput
by Antonio Quartulli 11 Apr '14

11 Apr '14

The "Expected throughput" towards a generic wireless peer is a value that can be used by other kernel components to achieve different goals. In particular we have 802.11s and batman-adv which will iuse this estimation to compute their throughput based metric. Therefore with this patchset I wanted to export this result and make it available for future use. Changes included in this patchset: - add the new get_expected_throughput() RC API - export the result of such API via cfg80211::get_station() - implement the get_expected_throughput() RC API for minstrel(_ht) For what concerns Minstrel_HT I decided to directly re-use the throughput value computed by the algorithm itself with minstrel_ht_calc_tp() for max_tp_rate. For the legacy Minstrel instead I computed the expected throughput as the product of the max_tp bitrate multplied by its probability of success. I am not entirely sure about passing sband as third parameter to get_expected_throughput()..maybe somebody can suggest a better solution. Cheers, Antonio Quartulli (6): cfg80211: export expected throughput through get_station() mac80211: add new RC API to retrieve expected throughput mac80211: export expected throughput in set_sta_info() mac80211: minstrel - implement get_expected_throughput() API mac80211: minstrel_ht - implement get_expected_throughput() API cfg80211: implement cfg80211_get_station cfg80211 API include/net/cfg80211.h | 74 +++++++++++++++++++++++--------------- include/net/mac80211.h | 3 ++ net/mac80211/cfg.c | 17 +++++++++ net/mac80211/rc80211_minstrel.c | 15 ++++++++ net/mac80211/rc80211_minstrel_ht.c | 19 ++++++++++ net/wireless/util.c | 18 ++++++++++ 6 files changed, 118 insertions(+), 28 deletions(-) -- 1.8.3.2

4 25

[B.A.T.M.A.N.] can batctl ping but not ping in 2014.1.0
by Gui Iribarren 08 Apr '14

08 Apr '14

Hello again friendly devs, here we are, after a long "running stable" hiatus, back into the bleeding edge for a ride \o/ running a small cloud of recent openwrt trunk (r40361) (OT: kmod-ath9k is running suprisingly smooth! yay!!) with kmod-batman-adv - 3.10.34+2014.1.0-2 and, well... i have bat news :P 1) yesterday i saw something vaguely reminiscent to the old OGM starving issue: in a line of 4 guinea-pig nodes that flow through a river of DeltaLibre, the 4th node would get TQ=1 for the 1st node, and would not even ping it (i can't remember the result of batctl ping, maybe it did), even though the links were really solid (TQ>220 on every one-hop-link of the chain) (the 3rd node was seeing the 1st with TQ>200, and could batctl ping / ping perfectly) at that point i found out kmod-batman-adv was inadvertently compiled without log support :( so that's as much as i can report for now, i'll recompile with that enabled and follow up. 2) this morning, in 2-node cloud testbed at home, uptime=22hs, The Bizarre Behaviour showed up and is sharing breakfast with me. on one side, lying calmly on the floor... root@rockm5:~# batctl o [B.A.T.M.A.N. adv 2014.1.0, MainIF/MAC: wlan0_adhoc.11/dc:9f:db:9c:37:54 (bat0 BATMAN_IV)] Originator last-seen (#/255) Nexthop [outgoingIF]: Potential nexthops ... 64:70:02:ed:f8:ea 0.770s (255) 64:70:02:ed:f8:ea [wlan0_adhoc.11]: 64:70:02:ed:f8:ea (255) 02:00:49:ed:f8:e8 0.320s (255) 64:70:02:ed:f8:ea [wlan0_adhoc.11]: 64:70:02:ed:f8:ea (255) root@rockm5:~# batctl if wlan0_adhoc.11: active 2 meters away, a TL-WDR3600 lurks... root@planit:~# batctl o [B.A.T.M.A.N. adv 2014.1.0, MainIF/MAC: eth0.1.11/02:00:49:ed:f8:e8 (bat0 BATMAN_IV)] Originator last-seen (#/255) Nexthop [outgoingIF]: Potential nexthops ... dc:9f:db:9c:37:54 0.360s (255) dc:9f:db:9c:37:54 [wlan1_adhoc.11]: dc:9f:db:9c:37:54 (255) root@planit:~# batctl if eth0.1.11: active wlan1_adhoc.11: active wlan0_adhoc.11: active ### rockm5 global ip over br-lan: gave its last breath root@planit:~# ip -6 r get 2a00:1508:1:f804::9d:3754/64 2a00:1508:1:f804::9d:3754 from :: dev br-lan src 2a00:1508:1:f804::ed:f8e8 metric 0 root@planit:~# ping 2a00:1508:1:f804::9d:3754 PING 2a00:1508:1:f804::9d:3754 (2a00:1508:1:f804::9d:3754): 56 data bytes --- 2a00:1508:1:f804::9d:3754 ping statistics --- 4 packets transmitted, 0 packets received, 100% packet loss ### rockm5 link-local over br-lan: feeding the daisies root@planit:~# ping6 fe80::de9f:dbff:fe9d:3754%br-lan PING fe80::de9f:dbff:fe9d:3754%br-lan(fe80::de9f:dbff:fe9d:3754) 56 data bytes --- fe80::de9f:dbff:fe9d:3754%br-lan ping statistics --- 3 packets transmitted, 0 received, 100% packet loss, time 2001ms ### lower level link-local works fine (avoiding batman-adv) root@planit:~# ping6 fe80::de9f:dbff:fe9c:3754%wlan1_adhoc.11 PING fe80::de9f:dbff:fe9c:3754%wlan1_adhoc.11(fe80::de9f:dbff:fe9c:3754) 56 data bytes 64 bytes from fe80::de9f:dbff:fe9c:3754: icmp_seq=1 ttl=64 time=2.70 ms 64 bytes from fe80::de9f:dbff:fe9c:3754: icmp_seq=2 ttl=64 time=1.39 ms --- fe80::de9f:dbff:fe9c:3754%wlan1_adhoc.11 ping statistics --- 2 packets transmitted, 2 received, 0% packet loss, time 1001ms rtt min/avg/max/mdev = 1.398/2.053/2.708/0.655 ms ### batctl ping to rockm5 enjoys excellent health root@planit:~# batctl ping dc:9f:db:9c:37:54 PING dc:9f:db:9c:37:54 (dc:9f:db:9c:37:54) 20(48) bytes of data 20 bytes from dc:9f:db:9c:37:54 icmp_seq=1 ttl=50 time=1.16 ms 20 bytes from dc:9f:db:9c:37:54 icmp_seq=2 ttl=50 time=0.90 ms 20 bytes from dc:9f:db:9c:37:54 icmp_seq=3 ttl=50 time=0.90 ms ^C--- dc:9f:db:9c:37:54 ping statistics --- 3 packets transmitted, 3 received, 0% packet loss rtt min/avg/max/mdev = 0.902/0.989/1.162/0.122 ms well, as said before, i have no "batctl l" output to show, but will collect and write chapter two. With a bit of luck, what i described so far rings a bell on someone, and can give an early insight (maybe it's due to the way we are using vlans?) (maybe its because routing_algo = BATMAN_IV?) (maybe the rewritten code is designed to work this way? yay!) (maybe it's our ugly hacky ebtables droppings / anygw magic that are interacting badly in some way? can describe them in detail next time) i must say tho, that this was running fine yesterday, and it broke spontaneously without any manual intervention or config change. oh, BLA2 and DAT are disabled on all nodes. thanks as always, and hope a giggle cheers up your day :) gui

1 1

[B.A.T.M.A.N.] [PATCHv3 maint] batman-adv: fix TT VLAN inconsistency on VLAN re-add
by Antonio Quartulli 06 Apr '14

06 Apr '14

From: Antonio Quartulli <antonio(a)open-mesh.com> When a VLAN interface (on top of batX) is removed and re-added within a short timeframe TT does not have enough time to properly cleanup. This creates an internal TT state mismatch as the newly created softif_vlan will be initialized from scratch with a TT client count of zero (even if TT entries for this VLAN still exist). The resulting TT messages are bogus due to the counter / tt client listing mismatch, thus creating inconsistencies on every node in the network To fix this issue destroy_vlan() has to not free the VLAN object immediately but it has to be kept alive until all the TT entries for this VLAN have been removed. destroy_vlan() still removes the sysfs folder so that the user has the feeling that everything went fine. If the same VLAN is re-added before the old object is free'd, then the latter is resurrected and re-used. Implement such behaviour by increasing the reference counter of a softif_vlan object every time a new local TT entry for such VLAN is created and remove the object from the list only when all the TT entries have been destroyed. Signed-off-by: Antonio Quartulli <antonio(a)open-mesh.com> --- Changes since v2: - remove cleanup_work member that is not needed anymore in this approach - reword commit subject - reword commit message (Thanks Marek!) Changes since v1: - destroy and create vlan sysfs folder within softif_vlan_destroy/create() to avoid lock troubles with soft-interface destruction and delayed jobs. soft-interface.c | 45 ++++++++++++++++++++++++++++++++------------- translation-table.c | 26 ++++++++++++++++++++++++++ types.h | 2 ++ 3 files changed, 60 insertions(+), 13 deletions(-) diff --git a/soft-interface.c b/soft-interface.c index d962363..87e2371 100644 --- a/soft-interface.c +++ b/soft-interface.c @@ -437,10 +437,15 @@ out: * possibly free it * @softif_vlan: the vlan object to release */ -void batadv_softif_vlan_free_ref(struct batadv_softif_vlan *softif_vlan) +void batadv_softif_vlan_free_ref(struct batadv_softif_vlan *vlan) { - if (atomic_dec_and_test(&softif_vlan->refcount)) - kfree_rcu(softif_vlan, rcu); + if (atomic_dec_and_test(&vlan->refcount)) { + spin_lock_bh(&vlan->bat_priv->softif_vlan_list_lock); + hlist_del_rcu(&vlan->list); + spin_unlock_bh(&vlan->bat_priv->softif_vlan_list_lock); + + kfree_rcu(vlan, rcu); + } } /** @@ -494,6 +499,7 @@ int batadv_softif_create_vlan(struct batadv_priv *bat_priv, unsigned short vid) if (!vlan) return -ENOMEM; + vlan->bat_priv = bat_priv; vlan->vid = vid; atomic_set(&vlan->refcount, 1); @@ -505,6 +511,10 @@ int batadv_softif_create_vlan(struct batadv_priv *bat_priv, unsigned short vid) return err; } + spin_lock_bh(&bat_priv->softif_vlan_list_lock); + hlist_add_head_rcu(&vlan->list, &bat_priv->softif_vlan_list); + spin_unlock_bh(&bat_priv->softif_vlan_list_lock); + /* add a new TT local entry. This one will be marked with the NOPURGE * flag */ @@ -512,10 +522,6 @@ int batadv_softif_create_vlan(struct batadv_priv *bat_priv, unsigned short vid) bat_priv->soft_iface->dev_addr, vid, BATADV_NULL_IFINDEX, BATADV_NO_MARK); - spin_lock_bh(&bat_priv->softif_vlan_list_lock); - hlist_add_head_rcu(&vlan->list, &bat_priv->softif_vlan_list); - spin_unlock_bh(&bat_priv->softif_vlan_list_lock); - return 0; } @@ -527,18 +533,13 @@ int batadv_softif_create_vlan(struct batadv_priv *bat_priv, unsigned short vid) static void batadv_softif_destroy_vlan(struct batadv_priv *bat_priv, struct batadv_softif_vlan *vlan) { - spin_lock_bh(&bat_priv->softif_vlan_list_lock); - hlist_del_rcu(&vlan->list); - spin_unlock_bh(&bat_priv->softif_vlan_list_lock); - - batadv_sysfs_del_vlan(bat_priv, vlan); - /* explicitly remove the associated TT local entry because it is marked * with the NOPURGE flag */ batadv_tt_local_remove(bat_priv, bat_priv->soft_iface->dev_addr, vlan->vid, "vlan interface destroyed", false); + batadv_sysfs_del_vlan(bat_priv, vlan); batadv_softif_vlan_free_ref(vlan); } @@ -556,6 +557,8 @@ static int batadv_interface_add_vid(struct net_device *dev, __be16 proto, unsigned short vid) { struct batadv_priv *bat_priv = netdev_priv(dev); + struct batadv_softif_vlan *vlan; + int ret; /* only 802.1Q vlans are supported. * batman-adv does not know how to handle other types @@ -565,6 +568,22 @@ static int batadv_interface_add_vid(struct net_device *dev, __be16 proto, vid |= BATADV_VLAN_HAS_TAG; + /* if a new vlan is getting created, but it is found in the list, it + * means that it was not deleted yet. Make batadv_softif_vlan_get() + * increase the refcounter by one and return + */ + vlan = batadv_softif_vlan_get(bat_priv, vid); + if (vlan) { + ret = 0; + if (!vlan->kobj) { + ret = batadv_sysfs_add_vlan(bat_priv->soft_iface, vlan); + if (ret) + batadv_softif_vlan_free_ref(vlan); + } + + return ret; + } + return batadv_softif_create_vlan(bat_priv, vid); } diff --git a/translation-table.c b/translation-table.c index 959dde7..b689981 100644 --- a/translation-table.c +++ b/translation-table.c @@ -485,6 +485,7 @@ bool batadv_tt_local_add(struct net_device *soft_iface, const uint8_t *addr, struct batadv_priv *bat_priv = netdev_priv(soft_iface); struct batadv_tt_local_entry *tt_local; struct batadv_tt_global_entry *tt_global; + struct batadv_softif_vlan *vlan; struct net_device *in_dev = NULL; struct hlist_head *head; struct batadv_tt_orig_list_entry *orig_entry; @@ -544,6 +545,9 @@ bool batadv_tt_local_add(struct net_device *soft_iface, const uint8_t *addr, if (!tt_local) goto out; + /* increase the refcounter of the related vlan */ + vlan = batadv_softif_vlan_get(bat_priv, vid); + batadv_dbg(BATADV_DBG_TT, bat_priv, "Creating new local tt entry: %pM (vid: %d, ttvn: %d)\n", addr, BATADV_PRINT_VID(vid), @@ -573,6 +577,7 @@ bool batadv_tt_local_add(struct net_device *soft_iface, const uint8_t *addr, if (unlikely(hash_added != 0)) { /* remove the reference for the hash */ batadv_tt_local_entry_free_ref(tt_local); + batadv_softif_vlan_free_ref(vlan); goto out; } @@ -978,6 +983,7 @@ uint16_t batadv_tt_local_remove(struct batadv_priv *bat_priv, { struct batadv_tt_local_entry *tt_local_entry; uint16_t flags, curr_flags = BATADV_NO_FLAGS; + struct batadv_softif_vlan *vlan; tt_local_entry = batadv_tt_local_hash_find(bat_priv, addr, vid); if (!tt_local_entry) @@ -1008,6 +1014,11 @@ uint16_t batadv_tt_local_remove(struct batadv_priv *bat_priv, hlist_del_rcu(&tt_local_entry->common.hash_entry); batadv_tt_local_entry_free_ref(tt_local_entry); + /* decrease the reference held for this vlan */ + vlan = batadv_softif_vlan_get(bat_priv, vid); + batadv_softif_vlan_free_ref(vlan); + batadv_softif_vlan_free_ref(vlan); + out: if (tt_local_entry) batadv_tt_local_entry_free_ref(tt_local_entry); @@ -1080,6 +1091,7 @@ static void batadv_tt_local_table_free(struct batadv_priv *bat_priv) spinlock_t *list_lock; /* protects write access to the hash lists */ struct batadv_tt_common_entry *tt_common_entry; struct batadv_tt_local_entry *tt_local; + struct batadv_softif_vlan *vlan; struct hlist_node *node_tmp; struct hlist_head *head; uint32_t i; @@ -1100,6 +1112,13 @@ static void batadv_tt_local_table_free(struct batadv_priv *bat_priv) tt_local = container_of(tt_common_entry, struct batadv_tt_local_entry, common); + + /* decrease the reference held for this vlan */ + vlan = batadv_softif_vlan_get(bat_priv, + tt_common_entry->vid); + batadv_softif_vlan_free_ref(vlan); + batadv_softif_vlan_free_ref(vlan); + batadv_tt_local_entry_free_ref(tt_local); } spin_unlock_bh(list_lock); @@ -3078,6 +3097,7 @@ static void batadv_tt_local_purge_pending_clients(struct batadv_priv *bat_priv) struct batadv_hashtable *hash = bat_priv->tt.local_hash; struct batadv_tt_common_entry *tt_common; struct batadv_tt_local_entry *tt_local; + struct batadv_softif_vlan *vlan; struct hlist_node *node_tmp; struct hlist_head *head; spinlock_t *list_lock; /* protects write access to the hash lists */ @@ -3106,6 +3126,12 @@ static void batadv_tt_local_purge_pending_clients(struct batadv_priv *bat_priv) tt_local = container_of(tt_common, struct batadv_tt_local_entry, common); + + /* decrease the reference held for this vlan */ + vlan = batadv_softif_vlan_get(bat_priv, tt_common->vid); + batadv_softif_vlan_free_ref(vlan); + batadv_softif_vlan_free_ref(vlan); + batadv_tt_local_entry_free_ref(tt_local); } spin_unlock_bh(list_lock); diff --git a/types.h b/types.h index 78370ab..c4aba2f 100644 --- a/types.h +++ b/types.h @@ -638,6 +638,7 @@ struct batadv_priv_nc { /** * struct batadv_softif_vlan - per VLAN attributes set + * @bat_priv: pointer to the mesh object * @vid: VLAN identifier * @kobj: kobject for sysfs vlan subdirectory * @ap_isolation: AP isolation state @@ -647,6 +648,7 @@ struct batadv_priv_nc { * @rcu: struct used for freeing in a RCU-safe manner */ struct batadv_softif_vlan { + struct batadv_priv *bat_priv; unsigned short vid; struct kobject *kobj; atomic_t ap_isolation; /* boolean */ -- 1.8.3.2

1 0

[B.A.T.M.A.N.] [PATCHv3 maint] batctl: tcpdump - fix IP header parsing
by Antonio Quartulli 06 Apr '14

06 Apr '14

From: Antonio Quartulli <antonio(a)open-mesh.com> Fix the IP header parsing by properly converting the source and the destination addresses to strings. The problem is given by the fact that inet_ntoa() uses a static buffer to store its result, therefore it cannot be invoked twice in a row without saving the first outcome. Moreover inet_ntoa() is deprecated in favor of inet_pton(). Introduced by 35b37756f4a3a230bf7941034d9da178ee5b4948 ("add IPv6 support to tcpdump parser") This patch also slightly beautifies the code. Signed-off-by: Antonio Quartulli <antonio(a)open-mesh.com> --- Change since v2: - fix commit subject Change since v1: - print all errors on stderr tcpdump.c | 73 ++++++++++++++++++++++++++++++++++++--------------------------- 1 file changed, 42 insertions(+), 31 deletions(-) diff --git a/tcpdump.c b/tcpdump.c index e9fac70..94e2a84 100644 --- a/tcpdump.c +++ b/tcpdump.c @@ -258,20 +258,27 @@ static void dump_ipv6(unsigned char *packet_buff, ssize_t buff_len, struct ip6_hdr *iphdr; struct icmp6_hdr *icmphdr; - char ipv6_addr_src[40], ipv6_addr_dst[40]; - char ip_saddr[40], ip_daddr[40], nd_nas_target[40]; + char ipsrc[INET6_ADDRSTRLEN], ipdst[INET6_ADDRSTRLEN]; struct nd_neighbor_solicit *nd_neigh_sol; struct nd_neighbor_advert *nd_advert; + char nd_nas_target[INET6_ADDRSTRLEN]; const char ip_string[] = "IP6"; iphdr = (struct ip6_hdr *)packet_buff; - LEN_CHECK((size_t)buff_len, (size_t)(sizeof(struct ip6_hdr)), "IP6"); + LEN_CHECK((size_t)buff_len, (size_t)(sizeof(struct ip6_hdr)), ip_string); if (!time_printed) print_time(); - inet_ntop(AF_INET6, &(iphdr->ip6_src), ipv6_addr_src, 40); - inet_ntop(AF_INET6, &(iphdr->ip6_dst), ipv6_addr_dst, 40); + if (!inet_ntop(AF_INET6, &iphdr->ip6_src, ipsrc, sizeof(ipsrc))) { + fprintf(stderr, "Cannot decode source IPv6\n"); + return; + } + + if (!inet_ntop(AF_INET6, &iphdr->ip6_dst, ipdst, sizeof(ipdst))) { + fprintf(stderr, "Cannot decode destination IPv6\n"); + return; + } switch (iphdr->ip6_nxt) { case IPPROTO_ICMPV6: @@ -280,7 +287,7 @@ static void dump_ipv6(unsigned char *packet_buff, ssize_t buff_len, icmphdr = (struct icmp6_hdr *)(packet_buff + sizeof(struct ip6_hdr)); - printf("IP6 %s > %s ", ipv6_addr_src, ipv6_addr_dst); + printf("%s %s > %s ", ip_string, ipsrc, ipdst); if (icmphdr->icmp6_type < ICMP6_INFOMSG_MASK && (size_t)(buff_len) > IPV6_MIN_MTU) { fprintf(stderr, @@ -348,16 +355,12 @@ static void dump_ipv6(unsigned char *packet_buff, ssize_t buff_len, } break; case IPPROTO_TCP: - inet_ntop(AF_INET6, &(iphdr->ip6_src), ip_saddr, 40); - inet_ntop(AF_INET6, &(iphdr->ip6_dst), ip_daddr, 40); dump_tcp(ip_string, packet_buff, buff_len, - sizeof(struct ip6_hdr), ip_saddr, ip_daddr); + sizeof(struct ip6_hdr), ipsrc, ipdst); break; case IPPROTO_UDP: - inet_ntop(AF_INET6, &(iphdr->ip6_src), ip_saddr, 40); - inet_ntop(AF_INET6, &(iphdr->ip6_dst), ip_daddr, 40); dump_udp(ip_string, packet_buff, buff_len, - sizeof(struct ip6_hdr), ip_saddr, ip_daddr); + sizeof(struct ip6_hdr), ipsrc, ipdst); break; default: printf(" IPv6 unknown protocol: %i\n", iphdr->ip6_nxt); @@ -367,29 +370,40 @@ static void dump_ipv6(unsigned char *packet_buff, ssize_t buff_len, static void dump_ip(unsigned char *packet_buff, ssize_t buff_len, int time_printed) { + char ipsrc[INET_ADDRSTRLEN], ipdst[INET_ADDRSTRLEN]; struct iphdr *iphdr, *tmp_iphdr; + const char ip_string[] = "IP"; struct udphdr *tmp_udphdr; struct icmphdr *icmphdr; - const char ip_string[] = "IP"; iphdr = (struct iphdr *)packet_buff; - LEN_CHECK((size_t)buff_len, (size_t)(iphdr->ihl * 4), "IP"); + LEN_CHECK((size_t)buff_len, (size_t)(iphdr->ihl * 4), ip_string); if (!time_printed) print_time(); + if (!inet_ntop(AF_INET, &iphdr->saddr, ipsrc, sizeof(ipsrc))) { + fprintf(stderr, "Cannot decode source IP\n"); + return; + } + + if (!inet_ntop(AF_INET, &iphdr->daddr, ipdst, sizeof(ipdst))) { + fprintf(stderr, "Cannot decode destination IP\n"); + return; + } + switch (iphdr->protocol) { case IPPROTO_ICMP: LEN_CHECK((size_t)buff_len - (iphdr->ihl * 4), sizeof(struct icmphdr), "ICMP"); icmphdr = (struct icmphdr *)(packet_buff + (iphdr->ihl * 4)); - printf("IP %s > ", inet_ntoa(*(struct in_addr *)&iphdr->saddr)); + printf("%s %s > ", ip_string, ipsrc); switch (icmphdr->type) { case ICMP_ECHOREPLY: printf("%s: ICMP echo reply, id %hu, seq %hu, length %zu\n", - inet_ntoa(*(struct in_addr *)&iphdr->daddr), - ntohs(icmphdr->un.echo.id), ntohs(icmphdr->un.echo.sequence), + ipdst, ntohs(icmphdr->un.echo.id), + ntohs(icmphdr->un.echo.sequence), (size_t)buff_len - (iphdr->ihl * 4)); break; case ICMP_DEST_UNREACH: @@ -401,46 +415,43 @@ static void dump_ip(unsigned char *packet_buff, ssize_t buff_len, tmp_iphdr = (struct iphdr *)(((char *)icmphdr) + sizeof(struct icmphdr)); tmp_udphdr = (struct udphdr *)(((char *)tmp_iphdr) + (tmp_iphdr->ihl * 4)); - printf("%s: ICMP ", inet_ntoa(*(struct in_addr *)&iphdr->daddr)); + printf("%s: ICMP ", ipdst); printf("%s udp port %hu unreachable, length %zu\n", - inet_ntoa(*(struct in_addr *)&tmp_iphdr->daddr), - ntohs(tmp_udphdr->dest), (size_t)buff_len - (iphdr->ihl * 4)); + ipdst, ntohs(tmp_udphdr->dest), + (size_t)buff_len - (iphdr->ihl * 4)); break; default: printf("%s: ICMP unreachable %hhu, length %zu\n", - inet_ntoa(*(struct in_addr *)&iphdr->daddr), - icmphdr->code, (size_t)buff_len - (iphdr->ihl * 4)); + ipdst, icmphdr->code, + (size_t)buff_len - (iphdr->ihl * 4)); break; } break; case ICMP_ECHO: printf("%s: ICMP echo request, id %hu, seq %hu, length %zu\n", - inet_ntoa(*(struct in_addr *)&iphdr->daddr), - ntohs(icmphdr->un.echo.id), ntohs(icmphdr->un.echo.sequence), + ipdst, ntohs(icmphdr->un.echo.id), + ntohs(icmphdr->un.echo.sequence), (size_t)buff_len - (iphdr->ihl * 4)); break; case ICMP_TIME_EXCEEDED: printf("%s: ICMP time exceeded in-transit, length %zu\n", - inet_ntoa(*(struct in_addr *)&iphdr->daddr), - (size_t)buff_len - (iphdr->ihl * 4)); + ipdst, (size_t)buff_len - (iphdr->ihl * 4)); break; default: printf("%s: ICMP type %hhu, length %zu\n", - inet_ntoa(*(struct in_addr *)&iphdr->daddr), icmphdr->type, + ipdst, icmphdr->type, (size_t)buff_len - (iphdr->ihl * 4)); break; } break; case IPPROTO_TCP: dump_tcp(ip_string, packet_buff, buff_len, iphdr->ihl * 4, - inet_ntoa(*(struct in_addr *)&iphdr->saddr), - inet_ntoa(*(struct in_addr *)&iphdr->daddr)); + ipsrc, ipdst); break; case IPPROTO_UDP: dump_udp(ip_string, packet_buff, buff_len, iphdr->ihl * 4, - inet_ntoa(*(struct in_addr *)&iphdr->saddr), - inet_ntoa(*(struct in_addr *)&iphdr->daddr)); + ipsrc, ipdst); break; default: printf("IP unknown protocol: %i\n", iphdr->protocol); -- 1.8.3.2

2 1

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

B.A.T.M.A.N