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.] Basic Multicast Optimizations
by Linus Lüssing 15 Aug '13

15 Aug '13

This is the eighth revision of the basic multicast optimization patches. It adds many style + functional changes and fixes some bugs. It also adds a new, important fifth patch, which addresses an issue that I would have overlooked if it weren't for the extensive offline talks with Simon about the interaction of batman-adv and various multicast related RFCs. # Changelog v7 -> v8 ## Style improvements * renamed batadv_mcast_tt_clean() tobatadv_mcast_tt_retract() * renamed batadv_mcast_tvlv_update_flag_counter() to batadv_mcast_update_counter() * renamed batadv_mcast_mla_tt_update() to batadv_mcast_mla_update() * renamed batadv_mcast_update_counter() to batadv_mcast_counter_update() * renamed BATADV_MCAST_HAS_NO_BRIDGE to BATADV_MCAST_WANT_ALL_UNSNOOPABLES (and reverse its logic) * change value of BATADV_UNINIT_FLAGS from -1 to an unused bit, BIT(15) -> no more need to check BATADV_UNINIT_FLAGS in batadv_mcast_counter_update() * moved code from batadv_mcast_mla_tt_update() to newly introduced batadv_mcast_mla_tvlv_update * moved introduction of multicast_mode sysfs file to the third patch * added kernel doc for struct batadv_priv_mcast and enum batadv_mcast_flag ## Functional improvements * changed a '== -EINVAL' to '< 0' * check for a bridge anywhere on top of bat0 (like br0 on bond0 on bat0, not just br0 on bat0) (as discussed on IRC we don't need to check for bridges on top of vlans yet as multicast optimizations are only enabled for non-VLAN packets at the moment) * removed sysfs file control over multicast TT+TVLV announcements, moved sysfs file introduction to the third patch * removed BATADV_MCAST_LISTENER_ANNOUNCEMENT flag: its purpose is now provided by checking the existence of the multicast tvlv instead * made the tvlv value length check in the multicast tvlv handler to not only accept a length of one byte but anything greater, too * New Patch: batman-adv: Send multicast packets to nodes with a WANT_ALL flag ## Bugs * converted the static 'enabled' variable in batadv_mcast_mla_tt_update() to a per bat_priv variable, using bat_priv->mcast.flags (fixes a bug when using multiple batX interfaces) * the announced tvlv was missing the BATADV_MCAST_HAS_NO_BRIDGE flag (not needed anymore with the reversed logic) * batadv_mcast_update_counter() was missing a case which would have needed a BATADV_UNINIT_FLAGS check (not needed anymore with reversed logic and new value for BATADV_UNINIT_FLAGS) * added missing update for num_want_all_unsnoopables (former num_has_bridge) counter in batadv_mcast_purge_orig() Cheers, Linus

2 9

[B.A.T.M.A.N.] DAT issue
by Fernando Pizarro 14 Aug '13

14 Aug '13

Hi all! I'm playing with OpenWRT and BATMAN advance and I had configured one essid to serve dynamic vlans via freeradius reply attributes. I have a bridge (br0) with bat0, wireless and ethernet interface, and hostapd creates a new bridge by vlan (brvlanX) with vlanX and wlan1.X succefully. Br0 has configured a VLAN and IP to connect with every other nodes. There is the same explanation, but schematic: Wireless: phy0 -> wlan0 -> bat0 phy1 -> wlan1 -> wlan1.1 \ --> wlan1.N Hostapd bridge: wlan1 -> brvlan1 -> {wlan1.1; vlan1} \ --> brvlanN -> {wlan1.N; vlanN} Ethernet bridge: br0 -> {bat0; eth0; wlan1} \ --> br0.X == IP/NETMASK When I look at batman dat table there aren't any registry, but variuos stations are connected. Why is that? What's wrong in my config? Thanks a lot. Fernando.

3 8

[B.A.T.M.A.N.] [PATCHv2 0/9] Improving the routing protocol abstraction
by Antonio Quartulli 13 Aug '13

13 Aug '13

Hello list, This is the second version of the "improved routing abstraction" patchset. In this second version I simply removed the bat_metric_compare() API since it was not really needed. The reason is that we are now assuming that all the metrics are represented by a uint32_t and they can be generically compared with a simple comparison (<, == or >). This patchset is introducing a new set of routing API functions meant to improve the routing protocol abstraction. Also the orig_node and the neigh_node structure have been heavily modified in order to split their generic members from those which are metric related. (the rest of the message is what you already had in the RFC cover letter). This changes have been written while developing batman V. The latter helped me in understanding what batman iv and v have in common and what not. The main problem was the metric: the two protocols use different metric domains and different semantics. Therefore all the functions handling/printing the metric needed to be generalised and rearranged to let the protocols decide what to do. Another issue was the way routing protocols handle the orig and neigh node structures. Also these two have been changed and some small APIs have been provided as well. Moreover, after Simon's RFC about the new multi-interface optimisation, we saw the need for a better abstraction so that mechanisms like that could easily be re-used by new algorithms (like batman v) with little effort. This is the second version of this RFC where I introduced some changes to the names of some functions and some other minor changes discussed in the previous RFC thread. + metric related: - bat_metric_get - bat_metric_is_equiv_or_better + orig_node related: - bat_orig_print: print the originator table - bat_orig_free - bat_orig_add_if - bat_orig_del_if Changes from v1: - removed bat_metric_compare() API. This is not really useful since we are assuming all the metrics are in the same domain (uint32_t) so a simple comparison is more than enough. Cheers, *** BLURB HERE *** Antonio Quartulli (9): batman-adv: make struct batadv_neigh_node algorithm agnostic batman-adv: make struct batadv_orig_node algorithm agnostic batman-adv: add bat_orig_print function API batman-adv: add bat_metric_get API function batman-adv: add bat_metric_is_equiv_or_better API function batman-adv: adapt bonding to use the new API functions batman-adv: adapt the neighbor purging routine to use the new API functions batman-adv: provide orig_node routing API batman-adv: adapt the TT component to use the new API functions bat_iv_ogm.c | 344 +++++++++++++++++++++++++++++++++++++++++++--------- gateway_client.c | 16 +-- main.c | 4 +- main.h | 6 + network-coding.c | 8 +- originator.c | 228 +++++++--------------------------- originator.h | 5 +- routing.c | 34 ++++-- routing.h | 3 +- translation-table.c | 31 +++-- types.h | 85 ++++++++----- 11 files changed, 457 insertions(+), 307 deletions(-) -- 1.8.1.5

2 15

[B.A.T.M.A.N.] [PATCH] alfred: handle failing write() in unix sockets
by Simon Wunderlich 12 Aug '13

12 Aug '13

This is not really critical, but some compilers seem to complain about that. Reported-by: Braden Aran <bradenaran(a)gmail.com> Signed-off-by: Simon Wunderlich <siwu(a)hrz.tu-chemnitz.de> --- unix_sock.c | 16 ++++++++++------ 1 file changed, 10 insertions(+), 6 deletions(-) diff --git a/unix_sock.c b/unix_sock.c index b29f635..c778a61 100644 --- a/unix_sock.c +++ b/unix_sock.c @@ -158,7 +158,7 @@ static int unix_sock_req_data_reply(struct globals *globals, int client_sock, struct alfred_push_data_v0 *push; struct hash_it_t *hashit = NULL; uint8_t buf[MAX_PAYLOAD]; - uint16_t seqno = 0; + uint16_t seqno = 0, ret = 0; /* send some data back through the unix socket */ @@ -184,12 +184,15 @@ static int unix_sock_req_data_reply(struct globals *globals, int client_sock, push->header.length = htons(len); push->tx.seqno = htons(seqno++); - write(client_sock, buf, sizeof(push->header) + len); + if (write(client_sock, buf, sizeof(push->header) + len) < 0) { + ret = -1; + break; + } } close(client_sock); - return 0; + return ret; } static int unix_sock_req_data(struct globals *globals, @@ -234,7 +237,7 @@ int unix_sock_req_data_finish(struct globals *globals, struct transaction_head *head) { struct alfred_status_v0 status; - int send_data = 1; + int ret = 0, send_data = 1; int client_sock; uint16_t id; uint8_t requested_type; @@ -258,10 +261,11 @@ int unix_sock_req_data_finish(struct globals *globals, status.header.length = htons(sizeof(status) - sizeof(status.header)); status.tx.id = htons(id); status.tx.seqno = 1; - write(client_sock, &status, sizeof(status)); + if (write(client_sock, &status, sizeof(status)) < 0) + ret = -1; close(client_sock); - return 0; + return ret; } int unix_sock_read(struct globals *globals) -- 1.7.10.4

1 1

[B.A.T.M.A.N.] pull request for net: batman-adv 2013-08-10
by Antonio Quartulli 10 Aug '13

10 Aug '13

Hello David, here you have one fix intended for net/linux-3.11. To avoid kernel paging errors, this patch takes care of re-assigning the value of skb->data to the related pointers after that the skb has been re-allocated. Please pull or let me know of any problem! Thank you, Antonio The following changes since commit 645359930231d5e78fd3296a38b98c1a658a7ade: rtnetlink: Fix inverted check in ndo_dflt_fdb_del() (2013-08-10 01:24:12 -0700) are available in the git repository at: git://git.open-mesh.org/linux-merge.git tags/batman-adv-fix-for-davem for you to fetch changes up to 9d2c9488cedb666bc8206fbdcdc1575e0fbc5929: batman-adv: fix potential kernel paging errors for unicast transmissions (2013-08-10 22:55:42 +0200) ---------------------------------------------------------------- Included change: - reassign pointers to data after skb reallocation to avoid kernel paging errors ---------------------------------------------------------------- Linus Lüssing (1): batman-adv: fix potential kernel paging errors for unicast transmissions net/batman-adv/bridge_loop_avoidance.c | 2 ++ net/batman-adv/gateway_client.c | 13 ++++++++++++- net/batman-adv/gateway_client.h | 3 +-- net/batman-adv/soft-interface.c | 9 ++++++++- net/batman-adv/unicast.c | 13 ++++++++++--- 5 files changed, 33 insertions(+), 7 deletions(-)

2 2

[B.A.T.M.A.N.] [PATCH 00/10] Improving the routing protocol abstraction
by Antonio Quartulli 10 Aug '13

10 Aug '13

Hello list, After two iterations of this patchset as RFC, I'm here sending the first real PATCH series :) This patchset is introducing a new set of routing API functions meant to improve the routing protocol abstraction. Also the orig_node and the neigh_node structure have been heavily modified in order to split their generic members from those which are metric related. (the rest of the message is what you already had in the RFC cover letter). This changes have been written while developing batman V. The latter helped me in understanding what batman iv and v have in common and what not. The main problem was the metric: the two protocols use different metric domains and different semantics. Therefore all the functions handling/printing the metric needed to be generalised and rearranged to let the protocols decide what to do. Another issue was the way routing protocols handle the orig and neigh node structures. Also these two have been changed and some small APIs have been provided as well. Moreover, after Simon's RFC about the new multi-interface optimisation, we saw the need for a better abstraction so that mechanisms like that could easily be re-used by new algorithms (like batman v) with little effort. This is the second version of this RFC where I introduced some changes to the names of some functions and some other minor changes discussed in the previous RFC thread. + metric related: - bat_metric_get - bat_metric_is_equiv_or_better - bat_metric_compare + orig_node related: - bat_orig_print: print the originator table - bat_orig_free - bat_orig_add_if - bat_orig_del_if Cheers, Antonio Quartulli (10): batman-adv: make struct batadv_neigh_node algorithm agnostic batman-adv: make struct batadv_orig_node algorithm agnostic batman-adv: add bat_orig_print function API batman-adv: add bat_metric_get API function batman-adv: add bat_metric_is_equiv_or_better API function batman-adv: add bat_metric_compare API function batman-adv: adapt bonding to use the new API functions batman-adv: adapt the neighbor purging routine to use the new API functions batman-adv: provide orig_node routing API batman-adv: adapt the TT component to use the new API functions bat_iv_ogm.c | 358 +++++++++++++++++++++++++++++++++++++++++++--------- gateway_client.c | 16 +-- main.c | 5 +- main.h | 6 + network-coding.c | 8 +- originator.c | 230 +++++++-------------------------- originator.h | 5 +- routing.c | 35 +++-- routing.h | 3 +- translation-table.c | 31 +++-- types.h | 85 +++++++++---- 11 files changed, 476 insertions(+), 306 deletions(-) -- 1.8.1.5

1 10

[B.A.T.M.A.N.] [RFCv2 00/10] Improving the routing protocol abstraction
by Antonio Quartulli 10 Aug '13

10 Aug '13

Hello list, with this RFC I'd like to introduce some new routing API functions meant to improve the routing protocol abstraction. This changes have been written while developing batman V. The latter helped me in understanding what batman iv and v have in common and what not. The main problem was the metric: the two protocols use different metric domains and different semantics. Therefore all the functions handling/printing the metric needed to be generalised and rearranged to let the protocols decide what to do. Another issue was the way routing protocols handle the orig and neigh node structures. Also these two have been changed and some small APIs have been provided as well. Moreover, after Simon's RFC about the new multi-interface optimisation, we saw the need for a better abstraction so that mechanisms like that could easily be re-used by new algorithms (like batman v) with little effort. This is the second version of this RFC where I introduced some changes to the names of some functions and some other minor changes discussed in the previous RFC thread. The new API functions are the following: + metric related: - bat_metric_get - bat_metric_is_similar - bat_metric_compare + orig_node related: - bat_orig_print: print the originator table - bat_orig_add_if - bat_orig_del_if - bat_orig_free Any feedback will surely be welcome :-) Cheers, Antonio Quartulli (10): batman-adv: make struct batadv_neigh_node algorithm agnostic batman-adv: make struct batadv_orig_node algorithm agnostic batman-adv: add bat_orig_print function API batman-adv: add bat_metric_get API function batman-adv: add bat_metric_is_equiv_or_better API function batman-adv: add bat_metric_compare API function batman-adv: adapt bonding to use the new API functions batman-adv: adapt the neighbor purging routine to use the new API functions batman-adv: provide orig_node routing API batman-adv: adapt the TT component to use the new API functions bat_iv_ogm.c | 359 +++++++++++++++++++++++++++++++++++++++++++--------- gateway_client.c | 16 +-- main.c | 5 +- main.h | 6 + network-coding.c | 8 +- originator.c | 230 +++++++-------------------------- originator.h | 5 +- routing.c | 37 ++++-- routing.h | 3 +- translation-table.c | 31 +++-- types.h | 86 +++++++++---- 11 files changed, 480 insertions(+), 306 deletions(-) -- 1.8.1.5

2 18

[B.A.T.M.A.N.] Alfred, Unix connections randomly refusing a connection
by Joshua Head 09 Aug '13

09 Aug '13

Hey, Just querying for some information on the best way to debug Alfred. It's currently at random points in time, independent of each node refusing to connect to a Linux socket. We have 4 nodes and the refusal is only localised that particular node. Restarting Alfred helps. But it's not a long term solution. We are very interested in Alfred. It solves a lot of our problems with our scripts. Basically if you have any direction for debugging it, or need any dumps. Feel free to ask. As far as I'm aware the Alfred is working, it's updating our test script. Just randomly stops accepting connections. Thanks, Joshua Head

2 1

[B.A.T.M.A.N.] [PATCHv2 0/2] TT-VLAN: BLA2 integration
by Antonio Quartulli 08 Aug '13

08 Aug '13

Hello people, this is the last patchset of the TT-VLAN series and requires the last one to be applied (per-VLAN TT CRC). In a few words... These changes enable BLA2 to filter only those TT entries belonging to the VLAN where BLA discovered other backbone nodes. The current code instead assumes that _all_ the local clients are shared among the backbones, despite of the VLAN used to interconnect them. Example. Imagine 2 nodes having several vlans over the wired connection. Node1 having eth0.1 and eth0.2 Node2 having eth0.1 and eth0.3 Node1 and Node2 are connected using a cable and so they hear each other on eth0.1. Clients connected to eth0.2 and eth0.3 are reachable only through the respective nodes therefore, if Node1 wanted to talk to a client connected to eth0.3 needs to know that it is served by Node2. In the current implementation this is not possible because BLA would consider all the clients reachable through any node of the LAN and will reject all the global TT entries from any other backbone, so losing this TT information. With this patchset, instead, BLA will detect that the backbone connection is only over eth0.1 and so will continue to store all the global clients belonging to the other VLANs. I hope the example did not create more confusion :) Cheers, Changes from v1: - fixed kernel doc Antonio Quartulli (2): batman-adv: make the TT global purge routine VLAN specific batman-adv: make the backbone gw check VLAN specific bridge_loop_avoidance.c | 23 +++++++++++--------- bridge_loop_avoidance.h | 10 +++++---- originator.c | 2 +- routing.c | 2 +- translation-table.c | 58 ++++++++++++++++++++++--------------------------- translation-table.h | 2 +- 6 files changed, 48 insertions(+), 49 deletions(-) -- 1.8.1.5

2 4

[B.A.T.M.A.N.] [PATCH next] batman-adv: use proper network order conversion
by Antonio Quartulli 08 Aug '13

08 Aug '13

Introduced by eb7643628a608cb338d48d298ce7ca15fadfb239 ("batman-adv: set skb priority according to content") Signed-off-by: Antonio Quartulli <ordex(a)autistici.org> --- main.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/main.c b/main.c index ed97382..58c98b6 100644 --- a/main.c +++ b/main.c @@ -283,7 +283,7 @@ void batadv_skb_set_priority(struct sk_buff *skb, int offset) sizeof(*vhdr), &vhdr_tmp); if (!vhdr) return; - prio = htons(vhdr->h_vlan_TCI) & VLAN_PRIO_MASK; + prio = ntohs(vhdr->h_vlan_TCI) & VLAN_PRIO_MASK; prio = prio >> VLAN_PRIO_SHIFT; break; case htons(ETH_P_IP): -- 1.8.1.5

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