B.A.T.M.A.N. V leaves the nest v2
Subsystems (3)
Hi,
after several rounds of code & documentation reviews a number of improvements found their way into the patchset. The changelog is as follows:
v2:
- rebroadcast packet immediately after best neighbor timeout
- add a flag to avoid forwarding the same OGMs multiple times
- move TVLV processing out of update_route() and intobatadv_v_ogm_process_per_outif()
- code comments improvements (typ0s, clarifications, etc)
- mark OGM packet as new when protection window is started
- remove redundant throughput checks
- simplify/fix best neighbor check
Happy routing,
The B.A.T.M.A.N. team
PATCH v2 1/12 batman-adv: Add hard_iface specific sysfs wrapper macros for UINT
From: Linus Luessing <[email protected]>
This allows us to easily add a sysfs parameter for an unsigned int later, which is not for a batman mesh interface (e.g. bat0), but for a common interface instead. It allows reading and writing an atomic_t in hard_iface (instead of bat_priv compared to the mesh variant).
Developed by Linus during a 6 months trainee study period in Ascom (Switzerland) AG.
--- net/batman-adv/sysfs.c | 49 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 49 insertions(+)
net/batman-adv/sysfs.c +49 -0
@@ -242,6 +242,55 @@ ssize_t batadv_show_vlan_##_name(struct kobject *kobj, \ static BATADV_ATTR_VLAN(_name, _mode, batadv_show_vlan_##_name, \ batadv_store_vlan_##_name) +#define BATADV_ATTR_HIF_STORE_UINT(_name, _var, _min, _max, _post_func) \ +ssize_t batadv_store_##_name(struct kobject *kobj, \ + struct attribute *attr, char *buff, \ + size_t count) \ +{ \ + struct net_device *net_dev = batadv_kobj_to_netdev(kobj); \ +struct batadv_hard_iface *hard_iface; \ +ssize_t length; \ + \ + hard_iface = batadv_hardif_get_by_netdev(net_dev); \ + if (!hard_iface) \ + return 0; \ + \ + length = __batadv_store_uint_attr(buff, count, _min, _max, \ + _post_func, attr, \ + &hard_iface->_var, net_dev); \ + \ +batadv_hardif_put (hard_iface); \ + return length; \ +} + +#define BATADV_ATTR_HIF_SHOW_UINT(_name, _var) \ +ssize_t batadv_show_##_name(struct kobject *kobj, \ + struct attribute *attr, char *buff) \ +{ \ +struct net_device *net_dev =batadv_kobj_to_netdev (kobj); \ +struct batadv_hard_iface *hard_iface; \ +ssize_t length; \ + \ + hard_iface =batadv_hardif_get_by_netdev (net_dev); \ +if (!hard_iface) \ +return 0 ; \ + \ + length =sprintf (buff,"%i\n" ,atomic_read (&hard_iface->_var)); \ + \ +batadv_hardif_put (hard_iface); \ +return length; \ +} + +/* Use this, if you are going to set [name] in hard_iface to an +* unsigned integer value +*/ +#define BATADV_ATTR_HIF_UINT (_name, _var, _mode, _min, _max, _post_func)\ + static BATADV_ATTR_HIF_STORE_UINT(_name, _var, _min, \ + _max, _post_func) \ + static BATADV_ATTR_HIF_SHOW_UINT(_name, _var) \ + static BATADV_ATTR(_name, _mode, batadv_show_##_name, \ + batadv_store_##_name) + static int batadv_store_bool_attr(char *buff, size_t count, struct net_device *net_dev, const char *attr_name, atomic_t *attr,
Signature · 1 line
2.7.0.rc3
Marek Lindner
AppliedRe: [B.A.T.M.A.N.] [PATCH v2 01/12] batman-adv: Add hard_iface specific sysfs wrapper macros for UINT
Quoted 11 lines · On Saturday, January 16, 2016 16:40:08 Marek Lindner wrote:
From: Linus Luessing <[email protected]>
This allows us to easily add a sysfs parameter for an unsigned int later, which is not for a batman mesh interface (e.g. bat0), but for a common interface instead. It allows reading and writing an
atomic_tinhard_iface(instead ofbat_privcompared to the mesh variant).Developed by Linus during a 6 months trainee study period in Ascom (Switzerland) AG.
Quoted 8 lines
Signed-off-by Linus Luessing <[email protected]>Signed-off-by Marek Lindner <[email protected]>[[email protected]: rename functions and move macros]Signed-off-by Antonio Quartulli <[email protected]>--- net/batman-adv/sysfs.c | 49+++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 49 insertions(+)
Applied in revision ec46535.
Thanks,
Marek
PATCH v2 2/12 batman-adv: ELP - adding basic infrastructure
In reply to Marek LindnerFrom: Linus Luessing <[email protected]>
The B.A.T.M.A.N. protocol originally only used a single message type (called OGM) to determine the link qualities to the direct neighbors and spreading these link quality information through the whole mesh. This procedure is summarized on the BATMAN concept page and explained in details in the RFC draft published in 2008.
This approach was chosen for its simplicity during the protocol design phase and the implementation. However, it also bears some drawbacks:
* Wireless interfaces usually come with some packet loss, therefore a higher broadcast rate is desirable to allow a fast reaction on flaky connections.
Other interfaces of the same host might be connected to Ethernet LANs / VPNs / etc which rarely exhibit packet loss would benefit from a lower broadcast rate to reduce overhead.
* It generally is more desirable to detect local link quality changes at a faster rate than propagating all these changes through the entire mesh (the far end of the mesh does not need to care about local link quality changes that much). Other optimizations strategies, like reducing overhead, might be possible if OGMs weren't used for all tasks in the mesh at the same time.
As a result detecting local link qualities shall be handled by an independent message type, ELP, whereas the OGM message type remains responsible for flooding the mesh with these link quality information and determining the overall path transmit qualities.
Developed by Linus during a 6 months trainee study period in Ascom (Switzerland) AG.
--- Makefile | 3 + README.external | 1 + gen-compat-autoconf.sh | 1 + net/batman-adv/Kconfig | 14 ++++ net/batman-adv/Makefile | 2 + net/batman-adv/bat_algo.h | 15 +++- net/batman-adv/bat_v.c | 74 +++++++++++++++++ net/batman-adv/bat_v_elp.c | 193 +++++++++++++++++++++++++++++++++++++++++++++ net/batman-adv/bat_v_elp.h | 27 +++++++ net/batman-adv/main.c | 1 + net/batman-adv/packet.h | 20 +++++ net/batman-adv/types.h | 18 +++++ 12 files changed, 368 insertions(+), 1 deletion(-) create mode 100644 net/batman-adv/bat_v.c create mode 100644 net/batman-adv/bat_v_elp.c create mode 100644 net/batman-adv/bat_v_elp.h
Makefile +3 -0
@@ -27,6 +27,8 @@ export CONFIG_BATMAN_ADV_DAT=y export CONFIG_BATMAN_ADV_NC=n # B.A.T.M.A.N. multicast optimizations: export CONFIG_BATMAN_ADV_MCAST=y +# B.A.T.M.A.N. V routing algorithm (experimental): +export CONFIG_BATMAN_ADV_BATMAN_V=n PWD:=$(shell pwd) KERNELPATH ?= /lib/modules/$(shell uname -r)/build
@@ -58,6 +60,7 @@ BUILD_FLAGS := \ CONFIG_BATMAN_ADV_DAT=$(CONFIG_BATMAN_ADV_DAT) \ CONFIG_BATMAN_ADV_NC=$(CONFIG_BATMAN_ADV_NC) \ CONFIG_BATMAN_ADV_MCAST=$(CONFIG_BATMAN_ADV_MCAST) \ + CONFIG_BATMAN_ADV_BATMAN_V=$(CONFIG_BATMAN_ADV_BATMAN_V) \ INSTALL_MOD_DIR=updates/net/batman-adv/ all: config
README.external +1 -0
@@ -40,6 +40,7 @@ module). Available options and their possible values are * CONFIG_BATMAN_ADV_DAT=[y*|n] (B.A.T.M.A.N. Distributed ARP Table) * CONFIG_BATMAN_ADV_MCAST=[y*|n] (B.A.T.M.A.N. multicast optimizations) * CONFIG_BATMAN_ADV_NC=[y|n*] (B.A.T.M.A.N. Network Coding) + * CONFIG_BATMAN_ADV_BATMAN_V=[y|n*] (B.A.T.M.A.N. V routing algorithm) e.g., debugging can be enabled by
gen-compat-autoconf.sh +1 -0
@@ -41,6 +41,7 @@ gen_config 'CONFIG_BATMAN_ADV_BLA' ${CONFIG_BATMAN_ADV_BLA:="y"} >> "${TMP}" gen_config 'CONFIG_BATMAN_ADV_DAT' ${CONFIG_BATMAN_ADV_DAT:="y"} >> "${TMP}" gen_config 'CONFIG_BATMAN_ADV_MCAST' ${CONFIG_BATMAN_ADV_MCAST:="y"} >> "${TMP}" gen_config 'CONFIG_BATMAN_ADV_NC' ${CONFIG_BATMAN_ADV_NC:="n"} >> "${TMP}" +gen_config 'CONFIG_BATMAN_ADV_BATMAN_V' ${CONFIG_BATMAN_ADV_BATMAN_V:="n"} >> "${TMP}" # only regenerate compat-autoconf.h when config was changed diff "${TMP}" "${TARGET}" > /dev/null 2>&1 || cp "${TMP}" "${TARGET}"
net/batman-adv/Kconfig +14 -0
@@ -15,6 +15,20 @@ config BATMAN_ADV https://www.open-mesh.org/ for more information and user space tools. +config BATMAN_ADV_BATMAN_V + bool "B.A.T.M.A.N. V protocol (experimental)" + depends on BATMAN_ADV + default n + help + This option enables the B.A.T.M.A.N. V protocol, the successor + of the currently used B.A.T.M.A.N. IV protocol. The main + changes include splitting of the OGM protocol into a neighbor + discovery protocol (Echo Location Protocol, ELP) and a new OGM + Protocol OGMv2 for flooding protocol information through the + network, as well as a throughput based metric. + B.A.T.M.A.N. V is currently considered experimental and not + compatible to B.A.T.M.A.N. IV networks. + config BATMAN_ADV_BLA bool "Bridge Loop Avoidance" depends on BATMAN_ADV && INET
net/batman-adv/Makefile +2 -0
@@ -18,6 +18,8 @@ obj-$(CONFIG_BATMAN_ADV) += batman-adv.o batman-adv-y += bat_iv_ogm.o +batman-adv-$(CONFIG_BATMAN_ADV_BATMAN_V) += bat_v.o +batman-adv-$(CONFIG_BATMAN_ADV_BATMAN_V) += bat_v_elp.o batman-adv-y += bitarray.o batman-adv-$(CONFIG_BATMAN_ADV_BLA) += bridge_loop_avoidance.o batman-adv-$(CONFIG_DEBUG_FS) += debugfs.o
net/batman-adv/bat_algo.h +14 -1
@@ -1,6 +1,6 @@/* Copyright (C) 2011-2016 B.A.T.M.A.N. contributors: * -* Marek Lindner +* Marek Lindner, Linus Lüssing * * This program is free software; you can redistribute it and/or * modify it under the terms of version 2 of the GNU General Public
@@ -20,4 +20,17 @@int batadv_iv_init (void ); +#ifdef CONFIG_BATMAN_ADV_BATMAN_V + +int batadv_v_init (void ); + +#else + +static inline int batadv_v_init (void ) +{ +return 0 ; +} + +#endif /* CONFIG_BATMAN_ADV_BATMAN_V */ +#endif /* _NET_BATMAN_ADV_BAT_ALGO_H_ */
net/batman-adv/bat_v.c new file +74 -0
74 lines not on this page · raw message
net/batman-adv/bat_v_elp.c new file +193 -0
193 lines not on this page · raw message
net/batman-adv/bat_v_elp.h new file +27 -0
@@ -0,0 +1,27 @@ +/* Copyright (C) 2013-2016 B.A.T.M.A.N. contributors: +* +* Linus Lüssing, Marek Lindner +* +* This program is free software; you can redistribute it and/or +* modify it under the terms of version 2 of the GNU General Public +* License as published by the Free Software Foundation. +* +* This program is distributed in the hope that it will be useful, but +* WITHOUT ANY WARRANTY; without even the implied warranty of +* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +* General Public License for more details. +* +* You should have received a copy of the GNU General Public License +* along with this program; if not, see <http://www.gnu.org/licenses/>. +*/ + +#include "main.h" + +#ifndef _NET_BATMAN_ADV_BAT_V_ELP_H_ +#define _NET_BATMAN_ADV_BAT_V_ELP_H_ + +int batadv_v_elp_iface_enable (struct batadv_hard_iface *hard_iface); +void batadv_v_elp_iface_disable (struct batadv_hard_iface *hard_iface); +void batadv_v_elp_primary_iface_set (struct batadv_hard_iface *primary_iface); + +#endif /* _NET_BATMAN_ADV_BAT_V_ELP_H_ */
net/batman-adv/main.c +1 -0
@@ -87,6 +87,7 @@ static int __init batadv_init(void)batadv_recv_handler_init (); +batadv_v_init ();batadv_iv_init ();batadv_nc_init ();
net/batman-adv/packet.h +20 -0
@@ -26,6 +26,7 @@* @BATADV_IV_OGM: originator messages for B.A.T.M.A.N. IV * @BATADV_BCAST: broadcast packets carrying broadcast payload * @BATADV_CODED: network coded packets +* @BATADV_ELP: echo location packets for B.A.T.M.A.N. V * * @BATADV_UNICAST: unicast packets carrying unicast payload traffic * @BATADV_UNICAST_FRAG: unicast packets carrying a fragment of the original
@@ -40,6 +41,7 @@ enum batadv_packettype {BATADV_IV_OGM =0x00 ,BATADV_BCAST =0x01 ,BATADV_CODED =0x02 , +BATADV_ELP =0x03 ,/* 0x40 - 0x7f: unicast */ #define BATADV_UNICAST_MIN 0x40 BATADV_UNICAST = 0x40,
@@ -235,6 +237,24 @@ struct batadv_ogm_packet {#define BATADV_OGM_HLEN sizeof(struct batadv_ogm_packet)/** +* struct batadv_elp_packet - elp (neighbor discovery) packet +* @packet_type: batman-adv packet type, part of the general header +* @version: batman-adv protocol version, part of the genereal header +* @orig: originator mac address +* @seqno: sequence number +* @elp_interval: currently used ELP sending interval in ms +*/ +struct batadv_elp_packet { +u8 packet_type; +u8 version; +u8 orig[ETH_ALEN ]; +__be32 seqno; +__be32 elp_interval; +}; + +#define BATADV_ELP_HLEN sizeof(struct batadv_elp_packet) + +/** * struct batadv_icmp_header - common members among all the ICMP packets * @packet_type: batman-adv packet type, part of the general header * @version: batman-adv protocol version, part of the genereal header
net/batman-adv/types.h +18 -0
@@ -86,6 +86,20 @@ struct batadv_hard_iface_bat_iv { };/** +* struct batadv_hard_iface_bat_v - per hard-interface B.A.T.M.A.N. V data +* @elp_interval: time interval between two ELP transmissions +* @elp_seqno: current ELP sequence number +* @elp_skb: base skb containing the ELP message to send +* @elp_wq: workqueue used to schedule ELP transmissions +*/ +struct batadv_hard_iface_bat_v { +atomic_t elp_interval; +atomic_t elp_seqno; +struct sk_buff *elp_skb; +struct delayed_work elp_wq; +}; + +/** * struct batadv_hard_iface - network device known to batman-adv * @list: list node for batadv_hardif_list * @if_num: identificator of the interface
@@ -99,6 +113,7 @@ struct batadv_hard_iface_bat_iv {* @soft_iface: the batman-adv interface which uses this network interface * @rcu: struct used for freeing in an RCU-safe manner * @bat_iv: per hard-interface B.A.T.M.A.N. IV data +* @bat_v: per hard-interface B.A.T.M.A.N. V data * @cleanup_work: work queue callback item for hard-interface deinit * @debug_dir: dentry for nc subdir in batman-adv directory in debugfs * @neigh_list: list of unique single hop neighbors via this interface
@@ -116,6 +131,9 @@ struct batadv_hard_iface {struct net_device *soft_iface;struct rcu_head rcu;struct batadv_hard_iface_bat_iv bat_iv; +#ifdef CONFIG_BATMAN_ADV_BATMAN_V +struct batadv_hard_iface_bat_v bat_v; +#endif struct work_struct cleanup_work;struct dentry *debug_dir;struct hlist_head neigh_list;
Signature · 1 line
2.7.0.rc3
Marek Lindner
AppliedRe: [B.A.T.M.A.N.] [PATCH v2 02/12] batman-adv: ELP - adding basic infrastructure
Quoted 50 lines · On Saturday, January 16, 2016 16:40:09 Marek Lindner wrote:
From: Linus Luessing <[email protected]>
The B.A.T.M.A.N. protocol originally only used a single message type (called OGM) to determine the link qualities to the direct neighbors and spreading these link quality information through the whole mesh. This procedure is summarized on the BATMAN concept page and explained in details in the RFC draft published in 2008.
This approach was chosen for its simplicity during the protocol design phase and the implementation. However, it also bears some drawbacks:
* Wireless interfaces usually come with some packet loss, therefore a higher broadcast rate is desirable to allow a fast reaction on flaky connections.
Other interfaces of the same host might be connected to Ethernet LANs / VPNs / etc which rarely exhibit packet loss would benefit from a lower broadcast rate to reduce overhead.
* It generally is more desirable to detect local link quality changes at a faster rate than propagating all these changes through the entire mesh (the far end of the mesh does not need to care about local link quality changes that much). Other optimizations strategies, like reducing overhead, might be possible if OGMs weren't used for all tasks in the mesh at the same time.As a result detecting local link qualities shall be handled by an independent message type, ELP, whereas the OGM message type remains responsible for flooding the mesh with these link quality information and determining the overall path transmit qualities.
Developed by Linus during a 6 months trainee study period in Ascom (Switzerland) AG.
Signed-off-by Linus Luessing <[email protected]>Signed-off-by Marek Lindner <[email protected]>Signed-off-by Antonio Quartulli <[email protected]>--- Makefile | 3 + README.external | 1 + gen-compat-autoconf.sh | 1 + net/batman-adv/Kconfig | 14 ++++ net/batman-adv/Makefile | 2 + net/batman-adv/bat_algo.h | 15 +++- net/batman-adv/bat_v.c | 74 +++++++++++++++++ net/batman-adv/bat_v_elp.c | 193+++++++++++++++++++++++++++++++++++++++++++++
net/batman-adv/bat_v_elp.h
Quoted 8 lines
| 27 +++++++
net/batman-adv/main.c | 1 + net/batman-adv/packet.h | 20 +++++ net/batman-adv/types.h | 18 +++++ 12 files changed, 368 insertions(+), 1 deletion(-) create mode 100644 net/batman-adv/bat_v.c create mode 100644 net/batman-adv/bat_v_elp.c create mode 100644 net/batman-adv/bat_v_elp.h
Applied in revision a4b88af.
Thanks,
Marek
PATCH v2 3/12 batman-adv: ELP - creating neighbor structures
In reply to Marek LindnerFrom: Linus Luessing <[email protected]>
Initially developed by Linus during a 6 months trainee study period in Ascom (Switzerland) AG.
--- compat-include/linux/average.h | 74 ++++++++++++++++++++ net/batman-adv/bat_v.c | 23 ++++++- net/batman-adv/bat_v_elp.c | 152 +++++++++++++++++++++++++++++++++++++++++ net/batman-adv/bat_v_elp.h | 4 ++ net/batman-adv/main.h | 3 + net/batman-adv/types.h | 33 +++++++++ 6 files changed, 288 insertions(+), 1 deletion(-) create mode 100644 compat-include/linux/average.h
compat-include/linux/average.h new file +74 -0
@@ -0,0 +1,74 @@ +/* Copyright (C) 2007-2016 B.A.T.M.A.N. contributors: +* +* Marek Lindner, Simon Wunderlich +* +* This program is free software; you can redistribute it and/or +* modify it under the terms of version 2 of the GNU General Public +* License as published by the Free Software Foundation. +* +* This program is distributed in the hope that it will be useful, but +* WITHOUT ANY WARRANTY; without even the implied warranty of +* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +* General Public License for more details. +* +* You should have received a copy of the GNU General Public License +* along with this program; if not, see <http://www.gnu.org/licenses/>. +* +* This file contains macros for maintaining compatibility with older versions +* of the Linux kernel. +*/ + +#ifndef _NET_BATMAN_ADV_COMPAT_LINUX_AVERAGE_H +#define _NET_BATMAN_ADV_COMPAT_LINUX_AVERAGE_H + +#include <linux/version.h> +#if LINUX_VERSION_CODE >=KERNEL_VERSION (2 ,6 ,38 ) +#include_next <linux/average.h> +#endif + +#if LINUX_VERSION_CODE <KERNEL_VERSION (4 ,3 ,0 ) + +/* Exponentially weighted moving average (EWMA) */ + +#define DECLARE_EWMA (name, _factor, _weight) \ + struct ewma_##name { \ + unsigned long internal; \ + }; \ + static inline void ewma_##name##_init(struct ewma_##name *e) \ + { \ + BUILD_BUG_ON(!__builtin_constant_p(_factor)); \ + BUILD_BUG_ON(!__builtin_constant_p(_weight)); \ + BUILD_BUG_ON_NOT_POWER_OF_2(_factor); \ + BUILD_BUG_ON_NOT_POWER_OF_2(_weight); \ + e->internal = 0; \ + } \ + static inline unsigned long \ + ewma_##name##_read(struct ewma_##name *e) \ + { \ + BUILD_BUG_ON(!__builtin_constant_p(_factor)); \ + BUILD_BUG_ON(!__builtin_constant_p(_weight)); \ + BUILD_BUG_ON_NOT_POWER_OF_2(_factor); \ + BUILD_BUG_ON_NOT_POWER_OF_2(_weight); \ + return e->internal >> ilog2(_factor); \ + } \ + static inline void ewma_##name##_add(struct ewma_##name *e, \ + unsigned long val) \ + { \ + unsigned long internal = ACCESS_ONCE(e->internal); \ + unsigned long weight = ilog2(_weight); \ + unsigned long factor = ilog2(_factor); \ + \ + BUILD_BUG_ON(!__builtin_constant_p(_factor)); \ + BUILD_BUG_ON(!__builtin_constant_p(_weight)); \ + BUILD_BUG_ON_NOT_POWER_OF_2(_factor); \ + BUILD_BUG_ON_NOT_POWER_OF_2(_weight); \ + \ + ACCESS_ONCE(e->internal) = internal ? \ + (((internal << weight) - internal) + \ + (val << factor)) >> weight : \ + (val << factor); \ + } + +#endif /* < KERNEL_VERSION(4, 3, 0) */ + +#endif /* _NET_BATMAN_ADV_COMPAT_LINUX_AVERAGE_H */
net/batman-adv/bat_v.c +22 -1
@@ -22,6 +22,7 @@#include <linux/init.h> #include "bat_v_elp.h" +#include "packet.h" static int batadv_v_iface_enable (struct batadv_hard_iface *hard_iface) {
@@ -42,6 +43,12 @@ static void batadv_v_primary_iface_set(struct batadv_hard_iface *hard_iface)batadv_v_elp_primary_iface_set (hard_iface); } +static void +batadv_v_hardif_neigh_init (struct batadv_hardif_neigh_node *hardif_neigh) +{ +ewma_throughput_init (&hardif_neigh->bat_v.throughput); +} +static void batadv_v_ogm_schedule (struct batadv_hard_iface *hard_iface) { }
@@ -56,6 +63,7 @@ static struct batadv_algo_ops batadv_batman_v __read_mostly = { .bat_iface_disable = batadv_v_iface_disable, .bat_iface_update_mac = batadv_v_iface_update_mac, .bat_primary_iface_set = batadv_v_primary_iface_set, + .bat_hardif_neigh_init = batadv_v_hardif_neigh_init, .bat_ogm_emit = batadv_v_ogm_emit, .bat_ogm_schedule = batadv_v_ogm_schedule, };
@@ -70,5 +78,18 @@ static struct batadv_algo_ops batadv_batman_v __read_mostly = {*/ int __init batadv_v_init (void ) { -return batadv_algo_register (&batadv_batman_v); +int ret; + +/* B.A.T.M.A.N. V echo location protocol packet */ + ret =batadv_recv_handler_register (BATADV_ELP , + batadv_v_elp_packet_recv); +if (ret <0 ) +return ret; + + ret =batadv_algo_register (&batadv_batman_v); + +if (ret <0 ) +batadv_recv_handler_unregister (BATADV_ELP ); + +return ret; }
net/batman-adv/bat_v_elp.c +152 -0
162 lines not on this page · raw message
net/batman-adv/bat_v_elp.h +4 -0
@@ -20,8 +20,12 @@#ifndef _NET_BATMAN_ADV_BAT_V_ELP_H_#define _NET_BATMAN_ADV_BAT_V_ELP_H_ +struct sk_buff ; +int batadv_v_elp_iface_enable (struct batadv_hard_iface *hard_iface);void batadv_v_elp_iface_disable (struct batadv_hard_iface *hard_iface);void batadv_v_elp_primary_iface_set (struct batadv_hard_iface *primary_iface); +int batadv_v_elp_packet_recv (struct sk_buff *skb, +struct batadv_hard_iface *if_incoming);#endif /* _NET_BATMAN_ADV_BAT_V_ELP_H_ */
net/batman-adv/main.h +3 -0
@@ -60,6 +60,9 @@#define BATADV_TQ_LOCAL_BIDRECT_RECV_MINIMUM 1#define BATADV_TQ_TOTAL_BIDRECT_LIMIT 1 +/* B.A.T.M.A.N. V */ +#define BATADV_ELP_MAX_AGE 64 +/* number of OGMs sent with the last tt diff */ #define BATADV_TT_OGM_APPEND_MAX 3
net/batman-adv/types.h +33 -0
@@ -22,6 +22,7 @@#error only "main.h" can be included directly#endif +#include <linux/average.h> #include <linux/bitops.h> #include <linux/compiler.h> #include <linux/if_ether.h>
@@ -364,12 +365,28 @@ struct batadv_gw_node {struct rcu_head rcu; }; +DECLARE_EWMA (throughput,1024 ,8 ) + +/** +* struct batadv_hardif_neigh_node_bat_v - B.A.T.M.A.N. V private neighbor +* information +* @throughput: ewma link throughput towards this neighbor +* @elp_interval: time interval between two ELP transmissions +* @elp_latest_seqno: latest and best known ELP sequence number + */ +struct batadv_hardif_neigh_node_bat_v { +struct ewma_throughput throughput; +u32 elp_interval; +u32 elp_latest_seqno; +}; +/** * struct batadv_hardif_neigh_node - unique neighbor per hard-interface * @list: list node for batadv_hard_iface::neigh_list * @addr: the MAC address of the neighboring interface * @if_incoming: pointer to incoming hard-interface * @last_seen: when last packet via this neighbor was received +* @bat_v: B.A.T.M.A.N. V private data * @refcount: number of contexts the object is used * @rcu: struct used for freeing in a RCU-safe manner */
@@ -378,6 +395,9 @@ struct batadv_hardif_neigh_node {u8 addr[ETH_ALEN ];struct batadv_hard_iface *if_incoming;unsigned long last_seen; +#ifdef CONFIG_BATMAN_ADV_BATMAN_V +struct batadv_hardif_neigh_node_bat_v bat_v; +#endif struct kref refcount;struct rcu_head rcu; };
@@ -425,10 +445,20 @@ struct batadv_neigh_ifinfo_bat_iv { };/** +* struct batadv_neigh_ifinfo_bat_v - neighbor information per outgoing +* interface for B.A.T.M.A.N. V +* @throughput: last throughput metric received from originator via this neigh +*/ +struct batadv_neigh_ifinfo_bat_v { +u32 throughput; +}; + +/** * struct batadv_neigh_ifinfo - neighbor information per outgoing interface * @list: list node for batadv_neigh_node::ifinfo_list * @if_outgoing: pointer to outgoing hard-interface * @bat_iv: B.A.T.M.A.N. IV private structure +* @bat_v: B.A.T.M.A.N. V private data * @last_ttl: last received ttl from this neigh node * @refcount: number of contexts the object is used * @rcu: struct used for freeing in a RCU-safe manner
@@ -437,6 +467,9 @@ struct batadv_neigh_ifinfo {struct hlist_node list;struct batadv_hard_iface *if_outgoing;struct batadv_neigh_ifinfo_bat_iv bat_iv; +#ifdef CONFIG_BATMAN_ADV_BATMAN_V +struct batadv_neigh_ifinfo_bat_v bat_v; +#endif u8 last_ttl;struct kref refcount;struct rcu_head rcu;
Signature · 1 line
2.7.0.rc3
Marek Lindner
AppliedRe: [B.A.T.M.A.N.] [PATCH v2 03/12] batman-adv: ELP - creating neighbor structures
Quoted 8 lines · On Saturday, January 16, 2016 16:40:10 Marek Lindner wrote:
From: Linus Luessing <[email protected]>
Initially developed by Linus during a 6 months trainee study period in Ascom (Switzerland) AG.
Signed-off-by Linus Luessing <[email protected]>Signed-off-by Marek Lindner <[email protected]>Signed-off-by Antonio Quartulli <[email protected]>
Quoted 10 lines
--- compat-include/linux/average.h | 74 ++++++++++++++++++++ net/batman-adv/bat_v.c | 23 ++++++- net/batman-adv/bat_v_elp.c | 152 +++++++++++++++++++++++++++++++++++++++++ net/batman-adv/bat_v_elp.h | 4 ++ net/batman-adv/main.h | 3 + net/batman-adv/types.h | 33 +++++++++ 6 files changed, 288 insertions(+), 1 deletion(-) create mode 100644 compat-include/linux/average.h
Applied in revision 2c2fe0b.
Thanks,
Marek
PATCH v2 4/12 batman-adv: ELP - adding sysfs parameter for elp interval
In reply to Marek LindnerFrom: Linus Luessing <[email protected]>
This parameter can be set individually on each interface and allows the configuration of the elp interval for the link quality measurements during runtime. Usually it is desirable to set it to a higher (= slower) value on interfaces which have a more static characteristic (e.g. wired interfaces) or very dense neighbourhoods to reduce overhead.
Developed by Linus during a 6 months trainee study period in Ascom (Switzerland) AG.
--- Documentation/ABI/testing/sysfs-class-net-batman-adv | 8 +++++++- net/batman-adv/sysfs.c | 7 +++++++ 2 files changed, 14 insertions(+), 1 deletion(-)
Documentation/ABI/testing/sysfs-class-net-batman-adv +7 -1
@@ -1,4 +1,11 @@ +What: /sys/class/net/<iface>/batman-adv/elp_interval +Date: Feb 2014 +Contact: Linus Lüssing <[email protected]> +Description: + Defines the interval in milliseconds in which batman + sends its probing packets for link quality measurements. + What: /sys/class/net/<iface>/batman-adv/iface_status Date: May 2010 Contact: Marek Lindner <[email protected]>
@@ -12,4 +19,3 @@ Description: The /sys/class/net/<iface>/batman-adv/mesh_iface file displays the batman mesh interface this <iface> currently is associated with. -
net/batman-adv/sysfs.c +7 -0
@@ -920,10 +920,17 @@ static ssize_t batadv_show_iface_status(struct kobject *kobj, static BATADV_ATTR(mesh_iface, S_IRUGO | S_IWUSR, batadv_show_mesh_iface, batadv_store_mesh_iface); static BATADV_ATTR(iface_status, S_IRUGO, batadv_show_iface_status, NULL); +#ifdef CONFIG_BATMAN_ADV_BATMAN_V +BATADV_ATTR_HIF_UINT (elp_interval, bat_v.elp_interval,S_IRUGO |S_IWUSR , +2 *BATADV_JITTER ,INT_MAX ,NULL ); +#endif static struct batadv_attribute *batadv_batman_attrs[] = { &batadv_attr_mesh_iface, &batadv_attr_iface_status, +#ifdef CONFIG_BATMAN_ADV_BATMAN_V + &batadv_attr_elp_interval, +#endifNULL , };
Signature · 1 line
2.7.0.rc3
Marek Lindner
AppliedRe: [B.A.T.M.A.N.] [PATCH v2 04/12] batman-adv: ELP - adding sysfs parameter for elp interval
Quoted 12 lines · On Saturday, January 16, 2016 16:40:11 Marek Lindner wrote:
From: Linus Luessing <[email protected]>
This parameter can be set individually on each interface and allows the configuration of the elp interval for the link quality measurements during runtime. Usually it is desirable to set it to a higher (= slower) value on interfaces which have a more static characteristic (e.g. wired interfaces) or very dense neighbourhoods to reduce overhead.
Developed by Linus during a 6 months trainee study period in Ascom (Switzerland) AG.
Quoted 8 lines
Signed-off-by Linus Luessing <[email protected]>Signed-off-by Marek Lindner <[email protected]>[[email protected]: respin on top of the latest master]Signed-off-by Antonio Quartulli <[email protected]>--- Documentation/ABI/testing/sysfs-class-net-batman-adv | 8 +++++++- net/batman-adv/sysfs.c | 7 +++++++ 2 files changed, 14 insertions(+), 1 deletion(-)
Applied in revision 82da7d6.
Thanks,
Marek
PATCH v2 5/12 batman-adv: OGMv2 - add basic infrastructure
In reply to Marek LindnerFrom: Antonio Quartulli <[email protected]>
This is the initial implementation of the new OGM protocol (version 2). It has been designed to work on top of the newly added ELP.
In the previous version the OGM protocol was used to both measure link qualities and flood the network with the metric information. In this version the protocol is in charge of the latter task only, leaving the former to ELP.
This means being able to decouple the interval used by the neighbor discovery from the OGM broadcasting, which revealed to be costly in dense networks and needed to be relaxed so leading to a less responsive routing protocol.
--- net/batman-adv/Makefile | 1 + net/batman-adv/bat_algo.h | 13 +++ net/batman-adv/bat_v.c | 51 ++++++++- net/batman-adv/bat_v_ogm.c | 279 +++++++++++++++++++++++++++++++++++++++++++++ net/batman-adv/bat_v_ogm.h | 32 ++++++ net/batman-adv/main.c | 6 + net/batman-adv/main.h | 1 + net/batman-adv/packet.h | 29 +++++ net/batman-adv/types.h | 18 +++ 9 files changed, 427 insertions(+), 3 deletions(-) create mode 100644 net/batman-adv/bat_v_ogm.c create mode 100644 net/batman-adv/bat_v_ogm.h
net/batman-adv/Makefile +1 -0
@@ -20,6 +20,7 @@ obj-$(CONFIG_BATMAN_ADV) += batman-adv.o batman-adv-y += bat_iv_ogm.o batman-adv-$(CONFIG_BATMAN_ADV_BATMAN_V) += bat_v.o batman-adv-$(CONFIG_BATMAN_ADV_BATMAN_V) += bat_v_elp.o +batman-adv-$(CONFIG_BATMAN_ADV_BATMAN_V) += bat_v_ogm.o batman-adv-y += bitarray.o batman-adv-$(CONFIG_BATMAN_ADV_BLA) += bridge_loop_avoidance.o batman-adv-$(CONFIG_DEBUG_FS) += debugfs.o
net/batman-adv/bat_algo.h +13 -0
@@ -18,11 +18,15 @@#ifndef _NET_BATMAN_ADV_BAT_ALGO_H_#define _NET_BATMAN_ADV_BAT_ALGO_H_ +struct batadv_priv ; +int batadv_iv_init (void );#ifdef CONFIG_BATMAN_ADV_BATMAN_V int batadv_v_init (void ); +int batadv_v_mesh_init (struct batadv_priv *bat_priv); +void batadv_v_mesh_free (struct batadv_priv *bat_priv);#else
@@ -31,6 +35,15 @@ static inline int batadv_v_init(void)return 0 ; } +static inline int batadv_v_mesh_init (struct batadv_priv *bat_priv) +{ +return 0 ; +} + +static inline void batadv_v_mesh_free (struct batadv_priv *bat_priv) +{ +} +#endif /* CONFIG_BATMAN_ADV_BATMAN_V */ #endif /* _NET_BATMAN_ADV_BAT_ALGO_H_ */
net/batman-adv/bat_v.c +48 -3
81 lines not on this page · raw message
net/batman-adv/bat_v_ogm.c new file +279 -0
279 lines not on this page · raw message
net/batman-adv/bat_v_ogm.h new file +32 -0
@@ -0,0 +1,32 @@ +/* Copyright (C) 2013-2016 B.A.T.M.A.N. contributors: +* +* Antonio Quartulli +* +* This program is free software; you can redistribute it and/or +* modify it under the terms of version 2 of the GNU General Public +* License as published by the Free Software Foundation. +* +* This program is distributed in the hope that it will be useful, but +* WITHOUT ANY WARRANTY; without even the implied warranty of +* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +* General Public License for more details. +* +* You should have received a copy of the GNU General Public License +* along with this program; if not, see <http://www.gnu.org/licenses/>. +*/ + +#ifndef _BATMAN_ADV_BATADV_V_OGM_H_ +#define _BATMAN_ADV_BATADV_V_OGM_H_ + +struct batadv_hard_iface ; +struct batadv_priv ; +struct sk_buff ; + +int batadv_v_ogm_init (struct batadv_priv *bat_priv); +void batadv_v_ogm_free (struct batadv_priv *bat_priv); +int batadv_v_ogm_iface_enable (struct batadv_hard_iface *hard_iface); +void batadv_v_ogm_primary_iface_set (struct batadv_hard_iface *primary_iface); +int batadv_v_ogm_packet_recv (struct sk_buff *skb, +struct batadv_hard_iface *if_incoming); + +#endif /* _BATMAN_ADV_BATADV_V_OGM_H_ */
net/batman-adv/main.c +6 -0
@@ -160,6 +160,10 @@ int batadv_mesh_init(struct net_device *soft_iface)INIT_HLIST_HEAD (&bat_priv->tvlv.handler_list);INIT_HLIST_HEAD (&bat_priv->softif_vlan_list); + ret =batadv_v_mesh_init (bat_priv); +if (ret <0 ) + goto err; + ret =batadv_originator_init (bat_priv);if (ret <0 ) goto err;
@@ -202,6 +206,8 @@ void batadv_mesh_free(struct net_device *soft_iface)batadv_purge_outstanding_packets (bat_priv,NULL );batadv_gw_node_free (bat_priv); + +batadv_v_mesh_free (bat_priv);batadv_nc_mesh_free (bat_priv);batadv_dat_free (bat_priv);batadv_bla_free (bat_priv);
net/batman-adv/main.h +1 -0
@@ -30,6 +30,7 @@/* B.A.T.M.A.N. parameters */ #define BATADV_TQ_MAX_VALUE 255 +#define BATADV_THROUGHPUT_MAX_VALUE 0xFFFFFFFF#define BATADV_JITTER 20/* Time To Live of broadcast messages */
net/batman-adv/packet.h +29 -0
@@ -27,6 +27,7 @@* @BATADV_BCAST: broadcast packets carrying broadcast payload * @BATADV_CODED: network coded packets * @BATADV_ELP: echo location packets for B.A.T.M.A.N. V +* @BATADV_OGM2: originator messages for B.A.T.M.A.N. V * * @BATADV_UNICAST: unicast packets carrying unicast payload traffic * @BATADV_UNICAST_FRAG: unicast packets carrying a fragment of the original
@@ -42,6 +43,7 @@ enum batadv_packettype {BATADV_BCAST =0x01 ,BATADV_CODED =0x02 ,BATADV_ELP =0x03 , +BATADV_OGM2 =0x04 ,/* 0x40 - 0x7f: unicast */ #define BATADV_UNICAST_MIN 0x40 BATADV_UNICAST = 0x40,
@@ -237,6 +239,33 @@ struct batadv_ogm_packet {#define BATADV_OGM_HLEN sizeof(struct batadv_ogm_packet)/** +* struct batadv_ogm2_packet - ogm2 (routing protocol) packet +* @packet_type: batman-adv packet type, part of the general header +* @version: batman-adv protocol version, part of the general header +* @ttl: time to live for this packet, part of the general header +* @flags: reseved for routing relevant flags - currently always 0 +* @seqno: sequence number +* @orig: originator mac address +* @tvlv_len: length of the appended tvlv buffer (in bytes) +* @throughput: the currently flooded path throughput +*/ +struct batadv_ogm2_packet { +u8 packet_type; +u8 version; +u8 ttl; +u8 flags; +__be32 seqno; +u8 orig[ETH_ALEN ]; +__be16 tvlv_len; +__be32 throughput; +/* __packed is not needed as the struct size is divisible by 4, +* and the largest data type in this struct has a size of 4. +*/ +}; + +#define BATADV_OGM2_HLEN sizeof(struct batadv_ogm2_packet) + +/** * struct batadv_elp_packet - elp (neighbor discovery) packet * @packet_type: batman-adv packet type, part of the general header * @version: batman-adv protocol version, part of the genereal header
net/batman-adv/types.h +18 -0
@@ -802,6 +802,20 @@ struct batadv_softif_vlan { };/** +* struct batadv_priv_bat_v - B.A.T.M.A.N. V per soft-interface private data +* @ogm_buff: buffer holding the OGM packet +* @ogm_buff_len: length of the OGM packet buffer +* @ogm_seqno: OGM sequence number - used to identify each OGM +* @ogm_wq: workqueue used to schedule OGM transmissions +*/ +struct batadv_priv_bat_v { +unsigned char *ogm_buff; +int ogm_buff_len; +atomic_t ogm_seqno; +struct delayed_work ogm_wq; +}; + +/** * struct batadv_priv - per mesh interface data * @mesh_state: current status of the mesh (inactive/active/deactivating) * @soft_iface: net device which holds this struct as private data
@@ -855,6 +869,7 @@ struct batadv_softif_vlan {* @mcast: multicast data * @network_coding: bool indicating whether network coding is enabled * @nc: network coding data +* @bat_v: B.A.T.M.A.N. V per soft-interface private data */ struct batadv_priv {atomic_t mesh_state;
@@ -920,6 +935,9 @@ struct batadv_priv {atomic_t network_coding;struct batadv_priv_nc nc;#endif /* CONFIG_BATMAN_ADV_NC */ +#ifdef CONFIG_BATMAN_ADV_BATMAN_V +struct batadv_priv_bat_v bat_v; +#endif };/**
Signature · 1 line
2.7.0.rc3
Marek Lindner
AppliedRe: [B.A.T.M.A.N.] [PATCH v2 05/12] batman-adv: OGMv2 - add basic infrastructure
Quoted 17 lines · On Saturday, January 16, 2016 16:40:12 Marek Lindner wrote:
From: Antonio Quartulli <[email protected]>
This is the initial implementation of the new OGM protocol (version 2). It has been designed to work on top of the newly added ELP.
In the previous version the OGM protocol was used to both measure link qualities and flood the network with the metric information. In this version the protocol is in charge of the latter task only, leaving the former to ELP.
This means being able to decouple the interval used by the neighbor discovery from the OGM broadcasting, which revealed to be costly in dense networks and needed to be relaxed so leading to a less responsive routing protocol.
Signed-off-by Antonio Quartulli <[email protected]>
Quoted 14 lines
--- net/batman-adv/Makefile | 1 + net/batman-adv/bat_algo.h | 13 +++ net/batman-adv/bat_v.c | 51 ++++++++- net/batman-adv/bat_v_ogm.c | 279 +++++++++++++++++++++++++++++++++++++++++++++ net/batman-adv/bat_v_ogm.h | 32 ++++++ net/batman-adv/main.c | 6 + net/batman-adv/main.h | 1 + net/batman-adv/packet.h | 29 +++++ net/batman-adv/types.h | 18 +++ 9 files changed, 427 insertions(+), 3 deletions(-) create mode 100644 net/batman-adv/bat_v_ogm.c create mode 100644 net/batman-adv/bat_v_ogm.h
Applied in revision 6328353.
Thanks,
Marek
PATCH v2 6/12 batman-adv: OGMv2 - implement originators logic
In reply to Marek LindnerFrom: Antonio Quartulli <[email protected]>
Add the support for recognising new originators in the network and rebroadcast their OGMs.
---
v2:
- rebroadcast packet immediately after best neighbor timeout
- add a flag to avoid forwarding the same OGMs multiple times
- move TVLV processing out of update_route() and into batadv_v_ogm_process_per_outif()
- code comments improvements (typ0s, clarifications, etc)
- mark OGM packet as new when protection window is started
- remove redundant throughput checks
- simplify/fix best neighbor check
net/batman-adv/bat_v_elp.c | 41 +--- net/batman-adv/bat_v_ogm.c | 556 ++++++++++++++++++++++++++++++++++++++++++++- net/batman-adv/bat_v_ogm.h | 4 + net/batman-adv/main.h | 2 + net/batman-adv/types.h | 4 + 5 files changed, 566 insertions(+), 41 deletions(-)
net/batman-adv/bat_v_elp.c +1 -40
@@ -37,8 +37,8 @@#include <linux/workqueue.h> #include "bat_algo.h" +#include "bat_v_ogm.h" #include "hard-interface.h" -#include "hash.h" #include "originator.h" #include "packet.h" #include "routing.h"
@@ -196,45 +196,6 @@ void batadv_v_elp_primary_iface_set(struct batadv_hard_iface *primary_iface) }/** -* batadv_v_ogm_orig_get - retrieve and possibly create an originator node -* @bat_priv: the bat priv with all the soft interface information -* @addr: the address of the originator -* -* Return: the orig_node corresponding to the specified address. If such object -* does not exist it is allocated here. In case of allocation failure returns -* NULL. -*/ -static struct batadv_orig_node * -batadv_v_ogm_orig_get (struct batadv_priv *bat_priv, -const u8 *addr) -{ -struct batadv_orig_node *orig_node; -int hash_added; - - orig_node =batadv_orig_hash_find (bat_priv, addr); -if (orig_node) -return orig_node; - - orig_node =batadv_orig_node_new (bat_priv, addr); -if (!orig_node) -return NULL ; - - hash_added =batadv_hash_add (bat_priv->orig_hash, batadv_compare_orig, - batadv_choose_orig, orig_node, - &orig_node->hash_entry); -if (hash_added !=0 ) { -/* orig_node->refcounter is initialised to 2 by -* batadv_orig_node_new() -*/ -batadv_orig_node_put (orig_node); -batadv_orig_node_put (orig_node); - orig_node =NULL ; - } - -return orig_node; -} - -/** * batadv_v_elp_neigh_update - update an ELP neighbour node * @bat_priv: the bat priv with all the soft interface information * @neigh_addr: the neighbour interface address
net/batman-adv/bat_v_ogm.c +555 -1
593 lines not on this page · raw message
net/batman-adv/bat_v_ogm.h +4 -0
@@ -18,6 +18,8 @@#ifndef _BATMAN_ADV_BATADV_V_OGM_H_#define _BATMAN_ADV_BATADV_V_OGM_H_ +#include <linux/types.h> +struct batadv_hard_iface ;struct batadv_priv ;struct sk_buff ;
@@ -25,6 +27,8 @@ struct sk_buff;int batadv_v_ogm_init (struct batadv_priv *bat_priv);void batadv_v_ogm_free (struct batadv_priv *bat_priv);int batadv_v_ogm_iface_enable (struct batadv_hard_iface *hard_iface); +struct batadv_orig_node *batadv_v_ogm_orig_get (struct batadv_priv *bat_priv, +const u8 *addr);void batadv_v_ogm_primary_iface_set (struct batadv_hard_iface *primary_iface);int batadv_v_ogm_packet_recv (struct sk_buff *skb,struct batadv_hard_iface *if_incoming);
net/batman-adv/main.h +2 -0
@@ -63,6 +63,8 @@/* B.A.T.M.A.N. V */ #define BATADV_ELP_MAX_AGE 64 +#define BATADV_OGM_MAX_ORIGDIFF 5 +#define BATADV_OGM_MAX_AGE 64/* number of OGMs sent with the last tt diff */ #define BATADV_TT_OGM_APPEND_MAX 3
net/batman-adv/types.h +4 -0
@@ -149,6 +149,7 @@ struct batadv_hard_iface {* @router: router that should be used to reach this originator * @last_real_seqno: last and best known sequence number * @last_ttl: ttl of last received packet +* @last_seqno_forwarded: seqno of the OGM which was forwarded last * @batman_seqno_reset: time when the batman seqno window was reset * @refcount: number of contexts the object is used * @rcu: struct used for freeing in an RCU-safe manner
@@ -159,6 +160,7 @@ struct batadv_orig_ifinfo {struct batadv_neigh_node __rcu *router;/* rcu protected pointer */ u32 last_real_seqno;u8 last_ttl; +u32 last_seqno_forwarded;unsigned long batman_seqno_reset;struct kref refcount;struct rcu_head rcu;
@@ -448,9 +450,11 @@ struct batadv_neigh_ifinfo_bat_iv {* struct batadv_neigh_ifinfo_bat_v - neighbor information per outgoing * interface for B.A.T.M.A.N. V * @throughput: last throughput metric received from originator via this neigh +* @last_seqno: last sequence number known for this neighbor */ struct batadv_neigh_ifinfo_bat_v {u32 throughput; +u32 last_seqno; };/**
Signature · 1 line
2.7.0.rc3
Marek Lindner
AppliedRe: [B.A.T.M.A.N.] [PATCH v2 06/12] batman-adv: OGMv2 - implement originators logic
Quoted 18 lines · On Saturday, January 16, 2016 16:40:13 Marek Lindner wrote:
From: Antonio Quartulli <[email protected]>
Add the support for recognising new originators in the network and rebroadcast their OGMs.
Signed-off-by Antonio Quartulli <[email protected]>---
v2:- rebroadcast packet immediately after best neighbor timeout
- add a flag to avoid forwarding the same OGMs multiple times
- move TVLV processing out ofupdate_route()and intobatadv_v_ogm_process_per_outif()- code comments improvements (typ0s, clarifications, etc)
- mark OGM packet as new when protection window is started
- remove redundant throughput checks
- simplify/fix best neighbor check
Quoted 7 lines
net/batman-adv/bat_v_elp.c | 41 +--- net/batman-adv/bat_v_ogm.c | 556 ++++++++++++++++++++++++++++++++++++++++++++- net/batman-adv/bat_v_ogm.h | 4 + net/batman-adv/main.h | 2 + net/batman-adv/types.h | 4 + 5 files changed, 566 insertions(+), 41 deletions(-)
Applied in revision 667996e.
Thanks,
Marek
PATCH v2 7/12 batman-adv: add throughput override attribute to hard_ifaces
In reply to Marek LindnerFrom: Antonio Quartulli <[email protected]>
This attribute is exported to user space to disable the link throughput auto-detection by setting a fixed value. The throughput override value is used when batman-adv is computing the link throughput towards a neighbour.
If the value is set to 0 then batman-adv will try to detect the throughput by itself.
--- .../ABI/testing/sysfs-class-net-batman-adv | 9 +++ net/batman-adv/bat_v.c | 6 ++ net/batman-adv/gateway_common.c | 4 +- net/batman-adv/gateway_common.h | 2 + net/batman-adv/sysfs.c | 74 ++++++++++++++++++++++ net/batman-adv/types.h | 2 + 6 files changed, 95 insertions(+), 2 deletions(-)
Documentation/ABI/testing/sysfs-class-net-batman-adv +9 -0
@@ -1,4 +1,13 @@ +What: /sys/class/net/<iface>/batman-adv/throughput_override +Date: Feb 2014 +Contact: Antonio Quartulli <[email protected]> +description: + Defines the throughput value to be used by B.A.T.M.A.N. V + when estimating the link throughput using this interface. + If the value is set to 0 then batman-adv will try to + estimate the throughput by itself. + What: /sys/class/net/<iface>/batman-adv/elp_interval Date: Feb 2014 Contact: Linus Lüssing <[email protected]>
net/batman-adv/bat_v.c +6 -0
@@ -18,6 +18,7 @@#include "bat_algo.h" #include "main.h" +#include <linux/atomic.h> #include <linux/cache.h> #include <linux/init.h>
@@ -37,6 +38,11 @@ static int batadv_v_iface_enable(struct batadv_hard_iface *hard_iface)if (ret <0 )batadv_v_elp_iface_disable (hard_iface); +/* enable link throughput auto-detection by setting the throughput +* override to zero +*/ +atomic_set (&hard_iface->bat_v.throughput_override,0 ); +return ret; }
net/batman-adv/gateway_common.c +2 -2
@@ -40,8 +40,8 @@* * Return: false on parse error and true otherwise. */ -static bool batadv_parse_throughput (struct net_device *net_dev,char *buff, -const char *description,u32 *throughput) +bool batadv_parse_throughput (struct net_device *net_dev,char *buff, +const char *description,u32 *throughput) {enum batadv_bandwidth_units bw_unit_type =BATADV_BW_UNIT_KBIT ;u64 lthroughput;
net/batman-adv/gateway_common.h +2 -0
@@ -49,5 +49,7 @@ ssize_t batadv_gw_bandwidth_set(struct net_device *net_dev, char *buff,void batadv_gw_tvlv_container_update (struct batadv_priv *bat_priv);void batadv_gw_init (struct batadv_priv *bat_priv);void batadv_gw_free (struct batadv_priv *bat_priv); +bool batadv_parse_throughput (struct net_device *net_dev,char *buff, +const char *description,u32 *throughput);#endif /* _NET_BATMAN_ADV_GATEWAY_COMMON_H_ */
net/batman-adv/sysfs.c +74 -0
92 lines not on this page · raw message
net/batman-adv/types.h +2 -0
@@ -92,12 +92,14 @@ struct batadv_hard_iface_bat_iv {* @elp_seqno: current ELP sequence number * @elp_skb: base skb containing the ELP message to send * @elp_wq: workqueue used to schedule ELP transmissions +* @throughput_override: throughput override to disable link auto-detection */ struct batadv_hard_iface_bat_v {atomic_t elp_interval;atomic_t elp_seqno;struct sk_buff *elp_skb;struct delayed_work elp_wq; +atomic_t throughput_override; };/**
Signature · 1 line
2.7.0.rc3
Marek Lindner
AppliedRe: [B.A.T.M.A.N.] [PATCH v2 07/12] batman-adv: add throughput override attribute to hard_ifaces
Quoted 11 lines · On Saturday, January 16, 2016 16:40:14 Marek Lindner wrote:
From: Antonio Quartulli <[email protected]>
This attribute is exported to user space to disable the link throughput auto-detection by setting a fixed value. The throughput override value is used when batman-adv is computing the link throughput towards a neighbour.
If the value is set to 0 then batman-adv will try to detect the throughput by itself.
Signed-off-by Antonio Quartulli <[email protected]>
Quoted 9 lines
--- .../ABI/testing/sysfs-class-net-batman-adv | 9 +++ net/batman-adv/bat_v.c | 6 ++ net/batman-adv/gateway_common.c | 4 +- net/batman-adv/gateway_common.h | 2 + net/batman-adv/sysfs.c | 74 ++++++++++++++++++++++ net/batman-adv/types.h | 2 + 6 files changed, 95 insertions(+), 2 deletions(-)
Applied in revision c513176.
Thanks,
Marek
PATCH v2 8/12 batman-adv: keep track of when unicast packets are sent
In reply to Marek LindnerFrom: Antonio Quartulli <[email protected]>
To enable ELP to send probing packets over wireless links only if needed, batman-adv must keep track of the last time it sent a unicast packet towards every neighbour.
For this purpose a 2 main changes are introduced:
1) a new member of the elp_neigh_node structure stores the last time a unicast packet was sent towards this neighbour;
2) a wrapper function for sending unicast packets is implemented. This function will simply update the member describe din point 1) and then forward the packet to the real sending routine.
Point 2) implies that any code-path leading to a unicast sending now has to use the new wrapper.
--- net/batman-adv/bat_iv_ogm.c | 2 +- net/batman-adv/bat_v_elp.c | 2 +- net/batman-adv/bat_v_ogm.c | 2 +- net/batman-adv/distributed-arp-table.c | 4 +-- net/batman-adv/fragmentation.c | 8 ++--- net/batman-adv/icmp_socket.c | 2 +- net/batman-adv/network-coding.c | 22 +++++++------- net/batman-adv/send.c | 55 +++++++++++++++++++++++++++++----- net/batman-adv/send.h | 10 +++++-- net/batman-adv/types.h | 2 ++ 10 files changed, 75 insertions(+), 34 deletions(-)
net/batman-adv/bat_iv_ogm.c +1 -1
@@ -478,7 +478,7 @@ static void batadv_iv_ogm_send_to_if(struct batadv_forw_packet *forw_packet,batadv_inc_counter (bat_priv,BATADV_CNT_MGMT_TX );batadv_add_counter (bat_priv,BATADV_CNT_MGMT_TX_BYTES , skb->len +ETH_HLEN ); -batadv_send_skb_packet (skb, hard_iface, batadv_broadcast_addr); +batadv_send_broadcast_skb (skb, hard_iface); } }
net/batman-adv/bat_v_elp.c +1 -1
@@ -104,7 +104,7 @@ static void batadv_v_elp_periodic_work(struct work_struct *work) hard_iface->net_dev->name, atomic_read(&hard_iface->bat_v.elp_seqno)); -batadv_send_skb_packet (skb, hard_iface, batadv_broadcast_addr); +batadv_send_broadcast_skb (skb, hard_iface);atomic_inc (&hard_iface->bat_v.elp_seqno);
net/batman-adv/bat_v_ogm.c +1 -1
@@ -120,7 +120,7 @@ static void batadv_v_ogm_send_to_if(struct sk_buff *skb,batadv_add_counter (bat_priv,BATADV_CNT_MGMT_TX_BYTES , skb->len +ETH_HLEN ); -batadv_send_skb_packet (skb, hard_iface, batadv_broadcast_addr); +batadv_send_broadcast_skb (skb, hard_iface); }/**
net/batman-adv/distributed-arp-table.c +1 -3
@@ -654,9 +654,7 @@ static bool batadv_dat_send_data(struct batadv_priv *bat_priv, goto free_neigh; } - send_status =batadv_send_skb_packet (tmp_skb, - neigh_node->if_incoming, - neigh_node->addr); + send_status =batadv_send_unicast_skb (tmp_skb, neigh_node);if (send_status ==NET_XMIT_SUCCESS ) {/* count the sent packet */ switch (packet_subtype) {
net/batman-adv/fragmentation.c +3 -5
@@ -378,8 +378,7 @@ bool batadv_frag_skb_fwd(struct sk_buff *skb, skb->len + ETH_HLEN); packet->ttl--; -batadv_send_skb_packet (skb, neigh_node->if_incoming, - neigh_node->addr); +batadv_send_unicast_skb (skb, neigh_node); ret = true; }
@@ -486,8 +485,7 @@ bool batadv_frag_send_packet(struct sk_buff *skb,batadv_inc_counter (bat_priv,BATADV_CNT_FRAG_TX );batadv_add_counter (bat_priv,BATADV_CNT_FRAG_TX_BYTES , skb_fragment->len +ETH_HLEN ); -batadv_send_skb_packet (skb_fragment, neigh_node->if_incoming, - neigh_node->addr); +batadv_send_unicast_skb (skb_fragment, neigh_node); frag_header.no++;/* The initial check in this function should cover this case */
@@ -506,7 +504,7 @@ bool batadv_frag_send_packet(struct sk_buff *skb,batadv_inc_counter (bat_priv,BATADV_CNT_FRAG_TX );batadv_add_counter (bat_priv,BATADV_CNT_FRAG_TX_BYTES , skb->len +ETH_HLEN ); -batadv_send_skb_packet (skb, neigh_node->if_incoming, neigh_node->addr); +batadv_send_unicast_skb (skb, neigh_node); ret = true;
net/batman-adv/icmp_socket.c +1 -1
@@ -278,7 +278,7 @@ static ssize_t batadv_socket_write(struct file *file, const char __user *buff,ether_addr_copy (icmp_header->orig, primary_if->net_dev->dev_addr); -batadv_send_skb_packet (skb, neigh_node->if_incoming, neigh_node->addr); +batadv_send_unicast_skb (skb, neigh_node); goto out; dst_unreach:
net/batman-adv/network-coding.c +10 -12
@@ -575,9 +575,7 @@ batadv_nc_hash_find(struct batadv_hashtable *hash,*/ static void batadv_nc_send_packet (struct batadv_nc_packet *nc_packet) { -batadv_send_skb_packet (nc_packet->skb, - nc_packet->neigh_node->if_incoming, - nc_packet->nc_path->next_hop); +batadv_send_unicast_skb (nc_packet->skb, nc_packet->neigh_node); nc_packet->skb =NULL ;batadv_nc_packet_free (nc_packet); }
@@ -1067,11 +1065,11 @@ static bool batadv_nc_code_packets(struct batadv_priv *bat_priv,struct batadv_unicast_packet *packet1;struct batadv_unicast_packet *packet2;struct batadv_coded_packet *coded_packet; -struct batadv_neigh_node *neigh_tmp, *router_neigh; -struct batadv_neigh_node *router_coding =NULL ; +struct batadv_neigh_node *neigh_tmp, *router_neigh, *first_dest; +struct batadv_neigh_node *router_coding =NULL , *second_dest;struct batadv_neigh_ifinfo *router_neigh_ifinfo =NULL ;struct batadv_neigh_ifinfo *router_coding_ifinfo =NULL ; -u8 *first_source, *first_dest, *second_source, *second_dest; +u8 *first_source, *second_source;__be32 packet_id1, packet_id2;size_t count;bool res = false;
@@ -1114,9 +1112,9 @@ static bool batadv_nc_code_packets(struct batadv_priv *bat_priv,*/ if (tq_weighted_neigh >= tq_weighted_coding) {/* Destination from nc_packet is selected for MAC-header */ - first_dest = nc_packet->nc_path->next_hop; + first_dest = nc_packet->neigh_node; first_source = nc_packet->nc_path->prev_hop; - second_dest = neigh_node->addr; + second_dest = neigh_node; second_source = ethhdr->h_source; packet1 = (struct batadv_unicast_packet *)nc_packet->skb->data; packet2 = (struct batadv_unicast_packet *)skb->data;
@@ -1125,9 +1123,9 @@ static bool batadv_nc_code_packets(struct batadv_priv *bat_priv, skb->data + sizeof(*packet2)); }else {/* Destination for skb is selected for MAC-header */ - first_dest = neigh_node->addr; + first_dest = neigh_node; first_source = ethhdr->h_source; - second_dest = nc_packet->nc_path->next_hop; + second_dest = nc_packet->neigh_node; second_source = nc_packet->nc_path->prev_hop; packet1 = (struct batadv_unicast_packet *)skb->data; packet2 = (struct batadv_unicast_packet *)nc_packet->skb->data;
@@ -1169,7 +1167,7 @@ static bool batadv_nc_code_packets(struct batadv_priv *bat_priv, coded_packet->first_ttvn = packet1->ttvn;/* Info about second unicast packet */ -ether_addr_copy (coded_packet->second_dest, second_dest); +ether_addr_copy (coded_packet->second_dest, second_dest->addr);ether_addr_copy (coded_packet->second_source, second_source);ether_addr_copy (coded_packet->second_orig_dest, packet2->dest); coded_packet->second_crc = packet_id2;
@@ -1224,7 +1222,7 @@ static bool batadv_nc_code_packets(struct batadv_priv *bat_priv,batadv_nc_packet_free (nc_packet);/* Send the coded packet and return true */ -batadv_send_skb_packet (skb_dest, neigh_node->if_incoming, first_dest); +batadv_send_unicast_skb (skb_dest, first_dest); res = true; out:if (router_neigh)
net/batman-adv/send.c +48 -7
86 lines not on this page · raw message
net/batman-adv/send.h +7 -3
@@ -28,12 +28,16 @@struct sk_buff ;struct work_struct ; -int batadv_send_skb_packet (struct sk_buff *skb, -struct batadv_hard_iface *hard_iface, -const u8 *dst_addr);int batadv_send_skb_to_orig (struct sk_buff *skb,struct batadv_orig_node *orig_node,struct batadv_hard_iface *recv_if); +int batadv_send_skb_packet (struct sk_buff *skb, +struct batadv_hard_iface *hard_iface, +const u8 *dst_addr); +int batadv_send_broadcast_skb (struct sk_buff *skb, +struct batadv_hard_iface *hard_iface); +int batadv_send_unicast_skb (struct sk_buff *skb, +struct batadv_neigh_node *neigh_node);void batadv_schedule_bat_ogm (struct batadv_hard_iface *hard_iface); int batadv_add_bcast_packet_to_list(struct batadv_priv *bat_priv, const struct sk_buff *skb,
net/batman-adv/types.h +2 -0
@@ -377,11 +377,13 @@ DECLARE_EWMA(throughput, 1024, 8)* @throughput: ewma link throughput towards this neighbor * @elp_interval: time interval between two ELP transmissions * @elp_latest_seqno: latest and best known ELP sequence number +* @last_unicast_tx: when the last unicast packet has been sent to this neighbor */ struct batadv_hardif_neigh_node_bat_v {struct ewma_throughput throughput;u32 elp_interval;u32 elp_latest_seqno; +unsigned long last_unicast_tx; };/**
Signature · 1 line
2.7.0.rc3
Marek Lindner
AppliedRe: [B.A.T.M.A.N.] [PATCH v2 08/12] batman-adv: keep track of when unicast packets are sent
Quoted 18 lines · On Saturday, January 16, 2016 16:40:15 Marek Lindner wrote:
From: Antonio Quartulli <[email protected]>
To enable ELP to send probing packets over wireless links only if needed, batman-adv must keep track of the last time it sent a unicast packet towards every neighbour.
For this purpose a 2 main changes are introduced:
1) a new member of theelp_neigh_nodestructure stores the last time a unicast packet was sent towards this neighbour;
2) a wrapper function for sending unicast packets is implemented. This function will simply update the member describe din point 1) and then forward the packet to the real sending routine.Point 2) implies that any code-path leading to a unicast sending now has to use the new wrapper.
Signed-off-by Antonio Quartulli <[email protected]>
Quoted 13 lines
--- net/batman-adv/bat_iv_ogm.c | 2 +- net/batman-adv/bat_v_elp.c | 2 +- net/batman-adv/bat_v_ogm.c | 2 +- net/batman-adv/distributed-arp-table.c | 4 +-- net/batman-adv/fragmentation.c | 8 ++--- net/batman-adv/icmp_socket.c | 2 +- net/batman-adv/network-coding.c | 22 +++++++------- net/batman-adv/send.c | 55 +++++++++++++++++++++++++++++----- net/batman-adv/send.h | 10 +++++-- net/batman-adv/types.h | 2 ++ 10 files changed, 75 insertions(+), 34 deletions(-)
Applied in revision cdc4545.
Thanks,
Marek
PATCH v2 9/12 batman-adv: ELP - compute the metric based on the estimated throughput
In reply to Marek LindnerFrom: Antonio Quartulli <[email protected]>
In case of wireless interface retrieve the throughput by querying cfg80211. To perform this call a separate work must be scheduled because the function may sleep and this is not allowed within an RCU protected context (RCU in this case is used to iterate over all the neighbours).
Use ethtool to retrieve information about an Ethernet link like HALF/FULL_DUPLEX and advertised bandwidth (e.g. 100/10Mbps).
The metric is updated each time a new ELP packet is sent, this way it is possible to timely react to a metric variation which can imply (for example) a neighbour disconnection.
--- compat.h | 21 +++++ net/batman-adv/bat_v.c | 3 + net/batman-adv/bat_v_elp.c | 207 +++++++++++++++++++++++++++++++++++++++++++++ net/batman-adv/bat_v_elp.h | 2 + net/batman-adv/bat_v_ogm.c | 2 +- net/batman-adv/main.h | 4 + net/batman-adv/types.h | 16 ++++ 7 files changed, 254 insertions(+), 1 deletion(-)
compat.h +21 -0
@@ -170,6 +170,27 @@ static int __batadv_interface_kill_vid(struct net_device *dev, __be16 proto,\#endif /* < KERNEL_VERSION(3, 10, 0) */ +#if LINUX_VERSION_CODE <KERNEL_VERSION (3 ,15 ,0 ) + +/* the expected behaviour of this function is to return 0 on success, therefore +* it is possible to define it as 1 so that batman-adv thinks like something +* went wrong. It will then decide what to do. +*/ +#define cfg80211_get_station (_a, _b, _c) (1) +/* the following define substitute the expected_throughput field with a random +* one existing in the station_info struct. It can be random because due to the +* define above it will never be used. We need it only to make the code compile +*/ +#define expected_throughput filled + +#ifdef CONFIG_BATMAN_ADV_BATMAN_V + +#warning cfg80211 based throughput metric is only supported with Linux 3.15+ + +#endif + +#endif /* < KERNEL_VERSION(3, 15, 0) */ +#if LINUX_VERSION_CODE <KERNEL_VERSION (4 ,3 ,0 )#define IFF_NO_QUEUE 0; dev->tx_queue_len = 0
net/batman-adv/bat_v.c +3 -0
@@ -21,6 +21,7 @@#include <linux/atomic.h> #include <linux/cache.h> #include <linux/init.h> +#include <linux/workqueue.h> #include "bat_v_elp.h" #include "bat_v_ogm.h"
@@ -65,6 +66,8 @@ static void batadv_v_hardif_neigh_init(struct batadv_hardif_neigh_node *hardif_neigh) {ewma_throughput_init (&hardif_neigh->bat_v.throughput); +INIT_WORK (&hardif_neigh->bat_v.metric_work, + batadv_v_elp_throughput_metric_update); }static void batadv_v_ogm_schedule (struct batadv_hard_iface *hard_iface)
net/batman-adv/bat_v_elp.c +207 -0
250 lines not on this page · raw message
net/batman-adv/bat_v_elp.h +2 -0
@@ -21,11 +21,13 @@#define _NET_BATMAN_ADV_BAT_V_ELP_H_struct sk_buff ; +struct work_struct ;int batadv_v_elp_iface_enable (struct batadv_hard_iface *hard_iface);void batadv_v_elp_iface_disable (struct batadv_hard_iface *hard_iface);void batadv_v_elp_primary_iface_set (struct batadv_hard_iface *primary_iface);int batadv_v_elp_packet_recv (struct sk_buff *skb,struct batadv_hard_iface *if_incoming); +void batadv_v_elp_throughput_metric_update (struct work_struct *work);#endif /* _NET_BATMAN_ADV_BAT_V_ELP_H_ */
net/batman-adv/bat_v_ogm.c +1 -1
@@ -339,7 +339,7 @@ static u32 batadv_v_forward_penalty(struct batadv_priv *bat_priv,*/ if ((throughput >10 ) && (if_incoming == if_outgoing) && - (batadv_is_wifi_netdev (if_incoming->net_dev))) + !(if_incoming->bat_v.flags &BATADV_FULL_DUPLEX ))return throughput /2 ;/* hop penalty of 255 equals 100% */
net/batman-adv/main.h +4 -0
@@ -62,6 +62,10 @@#define BATADV_TQ_TOTAL_BIDRECT_LIMIT 1/* B.A.T.M.A.N. V */ +#define BATADV_THROUGHPUT_DEFAULT_VALUE 10 +#define BATADV_ELP_PROBES_PER_NODE 2 +#define BATADV_ELP_MIN_PROBE_SIZE 200/* bytes */ +#define BATADV_ELP_PROBE_MAX_TX_DIFF 100/* milliseconds */ #define BATADV_ELP_MAX_AGE 64#define BATADV_OGM_MAX_ORIGDIFF 5#define BATADV_OGM_MAX_AGE 64
net/batman-adv/types.h +16 -0
@@ -87,12 +87,25 @@ struct batadv_hard_iface_bat_iv { };/** +* enum batadv_v_hard_iface_flags - interface flags useful to B.A.T.M.A.N. V +* @BATADV_FULL_DUPLEX: tells if the connection over this link is full-duplex +* @BATADV_WARNING_DEFAULT: tells whether we have warned the user that no +* throughput data is available for this interface and that default values are +* assumed. +*/ +enum batadv_v_hard_iface_flags { +BATADV_FULL_DUPLEX =BIT (0 ), +BATADV_WARNING_DEFAULT =BIT (1 ), +}; + +/** * struct batadv_hard_iface_bat_v - per hard-interface B.A.T.M.A.N. V data * @elp_interval: time interval between two ELP transmissions * @elp_seqno: current ELP sequence number * @elp_skb: base skb containing the ELP message to send * @elp_wq: workqueue used to schedule ELP transmissions * @throughput_override: throughput override to disable link auto-detection +* @flags: interface specific flags */ struct batadv_hard_iface_bat_v {atomic_t elp_interval;
@@ -100,6 +113,7 @@ struct batadv_hard_iface_bat_v {struct sk_buff *elp_skb;struct delayed_work elp_wq;atomic_t throughput_override; +u8 flags; };/**
@@ -378,12 +392,14 @@ DECLARE_EWMA(throughput, 1024, 8)* @elp_interval: time interval between two ELP transmissions * @elp_latest_seqno: latest and best known ELP sequence number * @last_unicast_tx: when the last unicast packet has been sent to this neighbor +* @metric_work: work queue callback item for metric update */ struct batadv_hardif_neigh_node_bat_v {struct ewma_throughput throughput;u32 elp_interval;u32 elp_latest_seqno;unsigned long last_unicast_tx; +struct work_struct metric_work; };/**
Signature · 1 line
2.7.0.rc3
Marek Lindner
AppliedRe: [B.A.T.M.A.N.] [PATCH v2 09/12] batman-adv: ELP - compute the metric based on the estimated throughput
Quoted 18 lines · On Saturday, January 16, 2016 16:40:16 Marek Lindner wrote:
From: Antonio Quartulli <[email protected]>
In case of wireless interface retrieve the throughput by querying cfg80211. To perform this call a separate work must be scheduled because the function may sleep and this is not allowed within an RCU protected context (RCU in this case is used to iterate over all the neighbours).
Use ethtool to retrieve information about an Ethernet link like HALF/FULL_DUPLEX and advertised bandwidth (e.g. 100/10Mbps).
The metric is updated each time a new ELP packet is sent, this way it is possible to timely react to a metric variation which can imply (for example) a neighbour disconnection.
Signed-off-by Antonio Quartulli <[email protected]>
Quoted 10 lines
--- compat.h | 21 +++++ net/batman-adv/bat_v.c | 3 + net/batman-adv/bat_v_elp.c | 207 +++++++++++++++++++++++++++++++++++++++++++++ net/batman-adv/bat_v_elp.h | 2 + net/batman-adv/bat_v_ogm.c | 2 +- net/batman-adv/main.h | 4 + net/batman-adv/types.h | 16 ++++ 7 files changed, 254 insertions(+), 1 deletion(-)
Applied in revision 5c32451.
Thanks,
Marek
PATCH v2 10/12 batman-adv: B.A.T.M.A.N. V - implement neighbor comparison API calls
In reply to Marek LindnerFrom: Antonio Quartulli <[email protected]>
--- net/batman-adv/bat_v.c | 38 ++++++++++++++++++++++++++++++++++++++ 1 file changed, 38 insertions(+)
net/batman-adv/bat_v.c +38 -0
@@ -19,12 +19,15 @@#include "main.h" #include <linux/atomic.h> +#include <linux/bug.h> #include <linux/cache.h> #include <linux/init.h> +#include <linux/types.h> #include <linux/workqueue.h> #include "bat_v_elp.h" #include "bat_v_ogm.h" +#include "originator.h" #include "packet.h" static int batadv_v_iface_enable (struct batadv_hard_iface *hard_iface)
@@ -78,6 +81,39 @@ static void batadv_v_ogm_emit(struct batadv_forw_packet *forw_packet) { } +static int batadv_v_neigh_cmp (struct batadv_neigh_node *neigh1, +struct batadv_hard_iface *if_outgoing1, +struct batadv_neigh_node *neigh2, +struct batadv_hard_iface *if_outgoing2) +{ +struct batadv_neigh_ifinfo *ifinfo1, *ifinfo2; + + ifinfo1 =batadv_neigh_ifinfo_get (neigh1, if_outgoing1); + ifinfo2 =batadv_neigh_ifinfo_get (neigh2, if_outgoing2); + +if (WARN_ON (!ifinfo1 || !ifinfo2)) +return 0 ; + +return ifinfo1->bat_v.throughput - ifinfo2->bat_v.throughput; +} + +static bool batadv_v_neigh_is_sob (struct batadv_neigh_node *neigh1, +struct batadv_hard_iface *if_outgoing1, +struct batadv_neigh_node *neigh2, +struct batadv_hard_iface *if_outgoing2) +{ +struct batadv_neigh_ifinfo *ifinfo1, *ifinfo2; +u32 threshold; + + ifinfo1 =batadv_neigh_ifinfo_get (neigh1, if_outgoing1); + ifinfo2 =batadv_neigh_ifinfo_get (neigh2, if_outgoing2); + + threshold = ifinfo1->bat_v.throughput /4 ; + threshold = ifinfo1->bat_v.throughput - threshold; + +return ifinfo2->bat_v.throughput > threshold; +} +static struct batadv_algo_ops batadv_batman_v__read_mostly = { .name ="BATMAN_V" , .bat_iface_enable = batadv_v_iface_enable,
@@ -87,6 +123,8 @@ static struct batadv_algo_ops batadv_batman_v __read_mostly = { .bat_hardif_neigh_init = batadv_v_hardif_neigh_init, .bat_ogm_emit = batadv_v_ogm_emit, .bat_ogm_schedule = batadv_v_ogm_schedule, + .bat_neigh_cmp = batadv_v_neigh_cmp, + .bat_neigh_is_similar_or_better = batadv_v_neigh_is_sob, };/**
Signature · 1 line
2.7.0.rc3
Marek Lindner
AppliedRe: [B.A.T.M.A.N.] [PATCH v2 10/12] batman-adv: B.A.T.M.A.N. V - implement neighbor comparison API calls
Quoted 6 lines · On Saturday, January 16, 2016 16:40:17 Marek Lindner wrote:
From: Antonio Quartulli <[email protected]>
Signed-off-by Antonio Quartulli <[email protected]>--- net/batman-adv/bat_v.c | 38 ++++++++++++++++++++++++++++++++++++++ 1 file changed, 38 insertions(+)
Applied in revision b05bbab.
Thanks,
Marek
PATCH v2 11/12 batman-adv: B.A.T.M.A.N. V - implement bat_orig_print API
In reply to Marek LindnerFrom: Antonio Quartulli <[email protected]>
--- net/batman-adv/bat_v.c | 105 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 105 insertions(+)
net/batman-adv/bat_v.c +105 -0
@@ -22,11 +22,17 @@#include <linux/bug.h> #include <linux/cache.h> #include <linux/init.h> +#include <linux/jiffies.h> +#include <linux/netdevice.h> +#include <linux/rculist.h> +#include <linux/rcupdate.h> +#include <linux/seq_file.h> #include <linux/types.h> #include <linux/workqueue.h> #include "bat_v_elp.h" #include "bat_v_ogm.h" +#include "hash.h" #include "originator.h" #include "packet.h"
@@ -81,6 +87,104 @@ static void batadv_v_ogm_emit(struct batadv_forw_packet *forw_packet) { } +/** +* batadv_v_orig_print_neigh - print neighbors for the originator table +* @orig_node: the orig_node for which the neighbors are printed +* @if_outgoing: outgoing interface for these entries +* @seq: debugfs table seq_file struct +* +* Must be called while holding an rcu lock. +*/ +static void +batadv_v_orig_print_neigh (struct batadv_orig_node *orig_node, +struct batadv_hard_iface *if_outgoing, +struct seq_file *seq) +{ +struct batadv_neigh_node *neigh_node; +struct batadv_neigh_ifinfo *n_ifinfo; + +hlist_for_each_entry_rcu (neigh_node, &orig_node->neigh_list, list) { + n_ifinfo =batadv_neigh_ifinfo_get (neigh_node, if_outgoing); +if (!n_ifinfo) +continue ; + +seq_printf (seq," %pM (%9u.%1u)" , + neigh_node->addr, + n_ifinfo->bat_v.throughput /10 , + n_ifinfo->bat_v.throughput %10 ); + +batadv_neigh_ifinfo_put (n_ifinfo); + } +} + +/** +* batadv_v_orig_print - print the originator table +* @bat_priv: the bat priv with all the soft interface information +* @seq: debugfs table seq_file struct +* @if_outgoing: the outgoing interface for which this should be printed +*/ +static void batadv_v_orig_print (struct batadv_priv *bat_priv, +struct seq_file *seq, +struct batadv_hard_iface *if_outgoing) +{ +struct batadv_neigh_node *neigh_node; +struct batadv_hashtable *hash = bat_priv->orig_hash; +int last_seen_msecs, last_seen_secs; +struct batadv_orig_node *orig_node; +struct batadv_neigh_ifinfo *n_ifinfo; +unsigned long last_seen_jiffies; +struct hlist_head *head; +int batman_count =0 ; +u32 i; + +seq_printf (seq," %-15s %s (%11s) %17s [%10s]: %20s ...\n" , +"Originator" ,"last-seen" ,"throughput" ,"Nexthop" , +"outgoingIF" ,"Potential nexthops" ); + +for (i =0 ; i < hash->size; i++) { + head = &hash->table[i]; + +rcu_read_lock (); +hlist_for_each_entry_rcu (orig_node, head, hash_entry) { + neigh_node =batadv_orig_router_get (orig_node, + if_outgoing); +if (!neigh_node) +continue ; + + n_ifinfo =batadv_neigh_ifinfo_get (neigh_node, + if_outgoing); +if (!n_ifinfo) + goto next; + + last_seen_jiffies = jiffies - orig_node->last_seen; + last_seen_msecs =jiffies_to_msecs (last_seen_jiffies); + last_seen_secs = last_seen_msecs /1000 ; + last_seen_msecs = last_seen_msecs %1000 ; + +seq_printf (seq,"%pM %4i.%03is (%9u.%1u) %pM [%10s]:" , + orig_node->orig, last_seen_secs, + last_seen_msecs, + n_ifinfo->bat_v.throughput /10 , + n_ifinfo->bat_v.throughput %10 , + neigh_node->addr, + neigh_node->if_incoming->net_dev->name); + +batadv_v_orig_print_neigh (orig_node, if_outgoing, seq); +seq_puts (seq,"\n" ); + batman_count++; + +next: +batadv_neigh_node_put (neigh_node); +if (n_ifinfo) +batadv_neigh_ifinfo_put (n_ifinfo); + } +rcu_read_unlock (); + } + +if (batman_count ==0 ) +seq_puts (seq,"No batman nodes in range ...\n" ); +} + static int batadv_v_neigh_cmp(struct batadv_neigh_node *neigh1, struct batadv_hard_iface *if_outgoing1, struct batadv_neigh_node *neigh2,
@@ -123,6 +227,7 @@ static struct batadv_algo_ops batadv_batman_v __read_mostly = { .bat_hardif_neigh_init = batadv_v_hardif_neigh_init, .bat_ogm_emit = batadv_v_ogm_emit, .bat_ogm_schedule = batadv_v_ogm_schedule, + .bat_orig_print = batadv_v_orig_print, .bat_neigh_cmp = batadv_v_neigh_cmp, .bat_neigh_is_similar_or_better = batadv_v_neigh_is_sob, };
Signature · 1 line
2.7.0.rc3
Marek Lindner
AppliedRe: [B.A.T.M.A.N.] [PATCH v2 11/12] batman-adv: B.A.T.M.A.N. V - implement bat_orig_print API
Quoted 7 lines · On Saturday, January 16, 2016 16:40:18 Marek Lindner wrote:
From: Antonio Quartulli <[email protected]>
Signed-off-by Antonio Quartulli <[email protected]>--- net/batman-adv/bat_v.c | 105+++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 105 insertions(+)
Applied in revision 662b7c1.
Thanks,
Marek
PATCH v2 12/12 batman-adv: B.A.T.M.A.N. V - implement bat_neigh_print API
In reply to Marek LindnerFrom: Linus Luessing <[email protected]>
Lists all neighbours detected by the Echo Locating Protocol (ELP) and their throughput metric.
Initially Developed by Linus during a 6 months trainee study period in Ascom (Switzerland) AG.
--- net/batman-adv/bat_v.c | 55 ++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 55 insertions(+)
net/batman-adv/bat_v.c +55 -0
@@ -118,6 +118,60 @@ batadv_v_orig_print_neigh(struct batadv_orig_node *orig_node, }/** +* batadv_v_hardif_neigh_print - print a single ELP neighbour node +* @seq: neighbour table seq_file struct +* @hardif_neigh: hardif neighbour information +*/ +static void +batadv_v_hardif_neigh_print (struct seq_file *seq, +struct batadv_hardif_neigh_node *hardif_neigh) +{ +int last_secs, last_msecs; +u32 throughput; + + last_secs =jiffies_to_msecs (jiffies - hardif_neigh->last_seen) /1000 ; + last_msecs =jiffies_to_msecs (jiffies - hardif_neigh->last_seen) %1000 ; + throughput =ewma_throughput_read (&hardif_neigh->bat_v.throughput); + +seq_printf (seq,"%pM %4i.%03is (%9u.%1u) [%10s]\n" , + hardif_neigh->addr, last_secs, last_msecs, throughput /10 , + throughput %10 , hardif_neigh->if_incoming->net_dev->name); +} + +/** +* batadv_v_neigh_print - print the single hop neighbour list +* @bat_priv: the bat priv with all the soft interface information +* @seq: neighbour table seq_file struct +*/ +static void batadv_v_neigh_print (struct batadv_priv *bat_priv, +struct seq_file *seq) +{ +struct net_device *net_dev = (struct net_device *)seq->private; +struct batadv_hardif_neigh_node *hardif_neigh; +struct batadv_hard_iface *hard_iface; +int batman_count =0 ; + +seq_printf (seq," %-15s %s (%11s) [%10s]\n" ,"Neighbor" , +"last-seen" ,"throughput" ,"IF" ); + +rcu_read_lock (); +list_for_each_entry_rcu (hard_iface, &batadv_hardif_list, list) { +if (hard_iface->soft_iface != net_dev) +continue ; + +hlist_for_each_entry_rcu (hardif_neigh, + &hard_iface->neigh_list, list) { +batadv_v_hardif_neigh_print (seq, hardif_neigh); + batman_count++; + } + } +rcu_read_unlock (); + +if (batman_count ==0 ) +seq_puts (seq,"No batman nodes in range ...\n" ); +} + +/** * batadv_v_orig_print - print the originator table * @bat_priv: the bat priv with all the soft interface information * @seq: debugfs table seq_file struct
@@ -230,6 +284,7 @@ static struct batadv_algo_ops batadv_batman_v __read_mostly = { .bat_orig_print = batadv_v_orig_print, .bat_neigh_cmp = batadv_v_neigh_cmp, .bat_neigh_is_similar_or_better = batadv_v_neigh_is_sob, + .bat_neigh_print = batadv_v_neigh_print, };/**
Signature · 1 line
2.7.0.rc3
Marek Lindner
AppliedRe: [B.A.T.M.A.N.] [PATCH v2 12/12] batman-adv: B.A.T.M.A.N. V - implement bat_neigh_print API
Quoted 8 lines · On Saturday, January 16, 2016 16:40:19 Marek Lindner wrote:
From: Linus Luessing <[email protected]>
Lists all neighbours detected by the Echo Locating Protocol (ELP) and their throughput metric.
Initially Developed by Linus during a 6 months trainee study period in Ascom (Switzerland) AG.
Quoted 7 lines
Signed-off-by Linus Luessing <[email protected]>Signed-off-by Marek Lindner <[email protected]>Signed-off-by Antonio Quartulli <[email protected]>--- net/batman-adv/bat_v.c | 55++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 55 insertions(+)
Applied in revision 342f7f3.
Thanks,
Marek
Just installed in on some hardware (2x OM5P + OpenWrt 44654 + compat-wireless 2015-10-26) and enabled BATMAN_V.
/etc/config/wireless:
====================
config wifi-iface 'wmesh'
option device 'radio0'
option ifname 'adhoc0'
option network 'batnet'
option mode 'adhoc'
option ssid 'mesh'
option 'mcast_rate' '18000'
option bssid '02:CA:FE:CA:CA:40'
/etc/config/network
===================
config interface 'batnet'
option mtu '1532'
option proto 'batadv'
option mesh 'bat0'
option routing_algo 'BATMAN_V'
The result was:
==============
[ 168.540000] CPU 0 Unable to handle kernel paging request at virtual address fffffff0, epc == 86d8301c, ra == 86d8301c
[ 168.550000] Oops[#1]:
[ 168.550000] CPU: 0 PID: 265 Comm: kworker/u2:2 Tainted: P 3.18.8 #1
[ 168.550000] Workqueue: bat_events batadv_v_elp_throughput_metric_update [batman_adv]
[ 168.550000] task: 878826c0 ti: 87a24000 task.ti: 87a24000
[ 168.550000] $ 0 : 00000000 00000001 00000000 fffffffe
[ 168.550000] $ 4 : 86a5cf00 00000000 00000000 00000000
[ 168.550000] $ 8 : 00000011 c4653600 ffffffff 00000001
[ 168.550000] $12 : 00000000 0000000c 00000000 00000000
[ 168.550000] $16 : 00000000 00000000 87804a1e 868f5000
[ 168.550000] $20 : 00000000 00000000 87804a10 87804a00
[ 168.550000] $24 : 00000002 86d1a0e4
[ 168.550000] $28 : 87a24000 87a25e40 00000088 86d8301c
[ 168.550000] Hi : 00000000
[ 168.550000] Lo : 00000000
[ 168.550000] epc : 86d8301c batadv_v_elp_throughput_metric_update+0x20/0x58 [batman_adv]
[ 168.550000] Tainted: P
[ 168.550000] ra : 86d8301c batadv_v_elp_throughput_metric_update+0x20/0x58 [batman_adv]
[ 168.550000] Status: 1100fc03 KERNEL EXL IE
[ 168.550000] Cause : 00800008
[ 168.550000] BadVA : fffffff0
[ 168.550000] PrId : 00019750 (MIPS 74Kc)
[ 168.550000] Modules linked in: iptable_nat ath9k nf_nat_pptp nf_nat_ipv4 nf_nat_amanda nf_conntrack_pptp nf_conntrack_ipv6 nf_conntrack_ipv4 nf_conntrack_amanda ipt_REJECT ipt_MASQUERADE ath9k_common xt_time xt_tcpudp xt_tcpmss xt_string xt_statistic xt_state xt_recent xt_quota xt_pkttype xt_physdev xt_owner xt_nat xt_multiport xt_mark xt_mac xt_limit xt_length xt_id xt_hl xt_helper xt_ecn xt_dscp xt_conntrack xt_connmark xt_connlimit xt_connbytes xt_comment xt_addrtype xt_TCPMSS xt_REDIRECT xt_NETMAP xt_LOG xt_HL xt_DSCP xt_CT xt_CLASSIFY ts_kmp ts_fsm ts_bm nf_reject_ipv4 nf_nat_tftp nf_nat_snmp_basic nf_nat_sip nf_nat_proto_gre nf_nat_masquerade_ipv4 nf_nat_irc nf_nat_h323 nf_nat_ftp nf_nat nf_log_ipv4 nf_defrag_ipv6 nf_defrag_ipv4 nf_conntrack_tftp nf_conntrack_snmp nf_conntrack_sip nf_conntrack_rtcache nf_conntrack_proto_gre nf_conntrack_irc nf_conntrack_h323 nf_conntrack_ftp nf_conntrack_broadcast iptable_raw iptable_mangle iptable_filter ipt_ECN ip_tables crc_ccitt ath9k_hw em_nbyte sch_codel sch_prio sch_dsmark act_ipt sch_gred em_meta sch_htb cls_basic sch_netem act_police em_text sch_tbf sch_sfq em_cmp sch_red act_skbedit act_mirred em_u32 cls_u32 cls_tcindex cls_flow cls_route cls_fw sch_hfsc sch_ingress tmp421 hwmon batman_adv libcrc32c crc16 ath10k_pci ath10k_core ath mac80211 cfg80211 compat UDSMARK udsmac i2c_dev tc_classid_mapper(P) filter_group classifier_dns(P) classifier_netblock(P) classifier_bittorrent(P) classifier_rtmp(P) classifier_ssl(P) classifier_content(P) classifier_skype(P) kernel_classifier nf_conntrack xt_set ip_set_list_set ip_set_hash_netport ip_set_hash_netiface ip_set_hash_net ip_set_hash_ipportnet ip_set_hash_ipportip ip_set_hash_ipport ip_set_hash_ip ip_set_bitmap_port ip_set_bitmap_ipmac ip_set_bitmap_ip ip_set nfnetlink ip6t_REJECT nf_reject_ipv6 nf_log_ipv6 nf_log_common ip6table_raw ip6table_mangle ip6table_filter ip6_tables x_tables ifb tun ipv6 arc4 crypto_blkcipher input_polldev gpio_button_hotplug button_hotplug input_core ag71xx crc32c_generic crypto_hash
[ 168.550000] Process kworker/u2:2 (pid: 265, threadinfo=87a24000, task=878826c0, tls=00000000)
[ 168.550000] Stack : 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
803496de 800665ac 00000000 00000000 00000000 00000000 00000000 00000000
87804a00 879cf198 8030bb3e 00000001 00000000 00000000 00000000 00000000
00000010 80123ad4 879fa100 801232d0 879cf21e 879cf180 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 fffffffe 00000002
...
[ 168.550000] Call Trace:
[ 168.550000] [<86d8301c>] batadv_v_elp_throughput_metric_update+0x20/0x58 [batman_adv]
[ 168.550000]
[ 168.550000][B.A.T.M.A.N.] OpenWrt cfg80211.h station_info incompatibility with kernel+mac80211 [was: B.A.T.M.A.N. V leaves the nest v2]
Quoted 7 lines · On Thursday 21 January 2016 12:56:01 Sven Eckelmann wrote:
Just installed in on some hardware (2x OM5P + OpenWrt 44654 + compat-wireless 2015-10-26) and enabled
BATMAN_V.[ 168.540000] CPU 0 Unable to handle kernel paging request at virtual address fffffff0, epc == 86d8301c, ra == 86d8301c [ 168.550000] Oops[#1]: [ 168.550000] CPU: 0 PID: 265 Comm: kworker/u2:2 Tainted: P 3.18.8 #1 [ 168.550000] Workqueue: bat_events batadv_v_elp_throughput_metric_update [batman_adv]
[..]
Quoted 3 lines
[ 168.550000] epc : 86d8301c batadv_v_elp_throughput_metric_update+0x20/0x58 [batman_adv] [ 168.550000] Tainted: P [ 168.550000] ra : 86d8301c batadv_v_elp_throughput_metric_update+0x20/0x58 [batman_adv]
Who knew that struct station_info changed in v3.18-6219-g6de3980, v3.18-6235-gb51f3be and v3.18-6217-ga76b194? Wieder keiner!? :)
So what basically happened was a stack corruption causing the return address to be bogus. This only happened because the kernel version of cfg80211.h (which is used to compile batman-adv) is different than the version in the mac80211 package (which is used at runtime) I have now synced that structure between the OpenWrt kernel and the used compat-wireless, rebuild my OpenWrt and tried again. I hope this helps anyone trying to use B.A.T.M.A.N. V on OpenWrt.
@Felix, would it be ok for you when I propose a "*-sync-cfg80211-station_info.patch" for the 3.18 generic kernel patches? Or do you have a good idea how to force the openwrt-routing batman-adv package against the cfg80211.h of compat-wireless?
Kind regards,
Sven
From: Sven Eckelmann <[email protected]>
Date: Thu, 21 Jan 2016 14:07:32 +0100
Subject: [PATCH] sync cfg80211 station_info
batman-adv v2016.0 with B.A.T.M.A.N. V enabled will query the cfg80211 module for wireless station information. But the module is compiled against the OpenWrt kernel but will be with the cfg80211 module from the mac80211 package. The result is a stack corruption when the definition of these two headers are different.
But the cfg80211.h from the kernel can be synced easily against the compat-wireless patch because it should not be used by the OpenWrt kernel.
This problem happens for example with the kernel v3.18 and compat-wireless generated from 3.19 or later. The patch tries to avoid this by synching the structure against mac80211 2016-01-10.
--- include/net/cfg80211.h | 47 ++++++++++++++++++++++++++++++++++++----------- 1 file changed, 36 insertions(+), 11 deletions(-)
include/net/cfg80211.h +36 -11
@@ -945,12 +945,14 @@ enum rate_info_flags {* @mcs: mcs index if struct describes a 802.11n bitrate * @legacy: bitrate in 100kbit/s for 802.11abg * @nss: number of streams (VHT only) +* @bw: bandwidth (from &enum rate_info_bw) */ struct rate_info {u8 flags;u8 mcs;u16 legacy;u8 nss; +u8 bw; };/**
@@ -984,6 +986,24 @@ struct sta_bss_parameters {u16 beacon_interval; }; +/** +* struct cfg80211_tid_stats - per-TID statistics +* @filled: bitmap of flags using the bits of &enum nl80211_tid_stats to +* indicate the relevant values in this struct are filled +* @rx_msdu: number of received MSDUs +* @tx_msdu: number of (attempted) transmitted MSDUs +* @tx_msdu_retries: number of retries (not counting the first) for +* transmitted MSDUs +* @tx_msdu_failed: number of failed transmitted MSDUs +*/ +struct cfg80211_tid_stats { +u32 filled; +u64 rx_msdu; +u64 tx_msdu; +u64 tx_msdu_retries; +u64 tx_msdu_failed; +}; +#define IEEE80211_MAX_CHAINS 4/**
@@ -991,11 +1011,12 @@ struct sta_bss_parameters {* * Station information filled by driver for get_station() and dump_station. * -* @filled: bitflag of flags from &enum station_info_flags +* @filled: bitflag of flags using the bits of &enum nl80211_sta_info to +* indicate the relevant values in this struct for them * @connected_time: time(in secs) since a station is last connected * @inactive_time: time since last station activity (tx/rx) in milliseconds -* @rx_bytes: bytes received from this station -* @tx_bytes: bytes transmitted to this station +* @rx_bytes: bytes (size of MPDUs) received from this station +* @tx_bytes: bytes (size of MPDUs) transmitted to this station * @llid: mesh local link id * @plid: mesh peer link id * @plink_state: mesh peer link state
@@ -1008,10 +1029,10 @@ struct sta_bss_parameters {* @chain_signal_avg: per-chain signal strength average in dBm * @txrate: current unicast bitrate from this station * @rxrate: current unicast bitrate to this station -* @rx_packets: packets received from this station -* @tx_packets: packets transmitted to this station -* @tx_retries: cumulative retry counts -* @tx_failed: number of failed transmissions (retries exceeded, no ACK) +* @rx_packets: packets (MSDUs & MMPDUs) received from this station +* @tx_packets: packets (MSDUs & MMPDUs) transmitted to this station +* @tx_retries: cumulative retry counts (MPDUs) +* @tx_failed: number of failed transmissions (MPDUs) (retries exceeded, no ACK) * @rx_dropped_misc: Dropped for un-specified reason. * @bss_param: current BSS parameters * @generation: generation number for nl80211 dumps.
@@ -1031,6 +1052,11 @@ struct sta_bss_parameters {* @nonpeer_pm: non-peer mesh STA power save mode * @expected_throughput: expected throughput in kbps (including 802.11 headers) * towards this station. +* @rx_beacon: number of beacons received from this peer +* @rx_beacon_signal_avg: signal strength average (in dBm) for beacons received +* from this peer +* @pertid: per-TID statistics, see &struct cfg80211_tid_stats, using the last +* (IEEE80211_NUM_TIDS) index for MSDUs not encapsulated in QoS-MPDUs. */ struct station_info {u32 filled;
@@ -1071,10 +1097,9 @@ struct station_info {u32 expected_throughput; -/* -* Note: Add a new enum station_info_flags value for each new field and -* use it to check which fields are initialized. -*/ +u64 rx_beacon; +u8 rx_beacon_signal_avg; +struct cfg80211_tid_stats pertid[IEEE80211_NUM_TIDS +1 ]; };/**
Re: [B.A.T.M.A.N.] OpenWrt cfg80211.h station_info incompatibility with kernel+mac80211 [was: B.A.T.M.A.N. V leaves the nest v2]
Quoted 23 lines · Sven Eckelmann
Quoted 7 lines · On Thursday 21 January 2016 12:56:01 Sven Eckelmann wrote:
Just installed in on some hardware (2x OM5P + OpenWrt 44654 + compat-wireless 2015-10-26) and enabled
BATMAN_V.[ 168.540000] CPU 0 Unable to handle kernel paging request at virtual address fffffff0, epc == 86d8301c, ra == 86d8301c [ 168.550000] Oops[#1]: [ 168.550000] CPU: 0 PID: 265 Comm: kworker/u2:2 Tainted: P 3.18.8 #1 [ 168.550000] Workqueue: bat_events batadv_v_elp_throughput_metric_update [batman_adv][..]
Quoted 3 lines
[ 168.550000] epc : 86d8301c batadv_v_elp_throughput_metric_update+0x20/0x58 [batman_adv] [ 168.550000] Tainted: P [ 168.550000] ra : 86d8301c batadv_v_elp_throughput_metric_update+0x20/0x58 [batman_adv]Who knew that struct
station_infochanged in v3.18-6219-g6de3980, v3.18-6235-gb51f3be and v3.18-6217-ga76b194? Wieder keiner!? :)So what basically happened was a stack corruption causing the return address to be bogus. This only happened because the kernel version of
cfg80211.h(which is used to compile batman-adv) is different than the version in the mac80211 package (which is used at runtime) I have now synced that structure between the OpenWrt kernel and the used compat-wireless, rebuild my OpenWrt and tried again. I hope this helps anyone trying to use B.A.T.M.A.N. V on OpenWrt.
Quoted 4 lines
@Felix, would it be ok for you when I propose a "*-sync-cfg80211-station_info.patch" for the 3.18 generic kernel patches? Or do you have a good idea how to force the openwrt-routing batman-adv package against the
cfg80211.hof compat-wireless?
The openwrt mac80211 package exports its headers to $(STAGING_DIR)/usr/include/mac80211
You could just make the batman-adv package use those includes.
- Felix
Re: [B.A.T.M.A.N.] OpenWrt cfg80211.h station_info incompatibility with kernel+mac80211 [was: B.A.T.M.A.N. V leaves the nest v2]
Quoted 8 lines · On Thursday 21 January 2016 14:55:01 Felix Fietkau wrote:
Quoted 4 lines
@Felix, would it be ok for you when I propose a "*-sync-cfg80211-station_info.patch" for the 3.18 generic kernel patches? Or do you have a good idea how to force the openwrt-routing batman-adv package against the
cfg80211.hof compat-wireless?The openwrt mac80211 package exports its headers to $(STAGING_DIR)/usr/include/mac80211
You could just make the batman-adv package use those includes.
Thanks, missed that. I will prepare a pull request for openwrt-routing.
Regards,
Sven
Re: [B.A.T.M.A.N.] OpenWrt cfg80211.h station_info incompatibility with kernel+mac80211 [was: B.A.T.M.A.N. V leaves the nest v2]
Quoted 11 lines · Sven Eckelmann
Quoted 8 lines · On Thursday 21 January 2016 14:55:01 Felix Fietkau wrote:
Quoted 4 lines
@Felix, would it be ok for you when I propose a "*-sync-cfg80211-station_info.patch" for the 3.18 generic kernel patches? Or do you have a good idea how to force the openwrt-routing batman-adv package against the
cfg80211.hof compat-wireless?The openwrt mac80211 package exports its headers to $(STAGING_DIR)/usr/include/mac80211
You could just make the batman-adv package use those includes.
Thanks, missed that. I will prepare a pull request for openwrt-routing.
In case this was missed: we also need to make sure that batman-adv depends on mac80211 and that it gets compiled first.
Cheers,
Signature · 1 line
Antonio Quartulli
Sven Eckelmann
maintainerRe: [B.A.T.M.A.N.] OpenWrt cfg80211.h station_info incompatibility with kernel+mac80211 [was: B.A.T.M.A.N. V leaves the nest v2]
Quoted 5 lines · On Thursday 21 January 2016 22:02:50 Antonio Quartulli wrote:
Quoted 3 lines · Sven Eckelmann
Quoted 2 lines · On Thursday 21 January 2016 14:55:01 Felix Fietkau wrote:
Quoted 2 lines
@Felix, would it be ok for you when I propose a "*-sync-cfg80211-station_info.patch" for the 3.18 generic kernel
patches?
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Or do you have a good idea how to force the openwrt-routing batman-adv
package
Quoted 10 lines
Quoted 7 lines
Quoted 5 lines
Quoted 1 line
against the
cfg80211.hof compat-wireless?The openwrt mac80211 package exports its headers to $(STAGING_DIR)/usr/include/mac80211
You could just make the batman-adv package use those includes.
Thanks, missed that. I will prepare a pull request for openwrt-routing.
In case this was missed: we also need to make sure that batman-adv depends on mac80211 and that it gets compiled first.
You are right. But the B.A.T.M.A.N. V patches cannot be build with the official openwrt-routing package - so no need to get too eager. All we are doing right now is to prepare the integration (for the next release which will include B.A.T.M.A.N. V).
But if anyone is willing to prepare the changes for the internal branch...
Kind regards,
Sven
Sven Eckelmann
maintainerRe: [B.A.T.M.A.N.] OpenWrt cfg80211.h station_info incompatibility with kernel+mac80211 [was: B.A.T.M.A.N. V leaves the nest v2]
In reply to Sven EckelmannQuoted 3 lines · On Thursday 21 January 2016 15:03:10 Sven Eckelmann wrote:
Quoted 2 lines · On Thursday 21 January 2016 14:55:01 Felix Fietkau wrote:
Quoted 2 lines
@Felix, would it be ok for you when I propose a "*-sync-cfg80211-station_info.patch" for the 3.18 generic kernel
patches?
Quoted 1 line
Quoted 1 line
Quoted 1 line
Or do you have a good idea how to force the openwrt-routing batman-adv
package
Quoted 7 lines
Quoted 5 lines
Quoted 1 line
against the
cfg80211.hof compat-wireless?The openwrt mac80211 package exports its headers to $(STAGING_DIR)/usr/include/mac80211
You could just make the batman-adv package use those includes.
Thanks, missed that. I will prepare a pull request for openwrt-routing.
Hm, didn't work out so well. It basically explodes right away because possible_net_t, possible_read_pnet, possible_write_pnet is not defined on v3.18 when not using the mac80211 backporting headers. And when usign the backporting headers then it will explode because it conflicts with batman- adv's own backporting stuff. And without the batman-adv's backporting headers it will also not compile due to missing backporting hacks.
Kind regards,
Sven
Re: [B.A.T.M.A.N.] OpenWrt cfg80211.h station_info incompatibility with kernel+mac80211 [was: B.A.T.M.A.N. V leaves the nest v2]
Quoted 13 lines · Sven Eckelmann
Quoted 3 lines · On Thursday 21 January 2016 15:03:10 Sven Eckelmann wrote:
Quoted 2 lines · On Thursday 21 January 2016 14:55:01 Felix Fietkau wrote:
Quoted 2 lines
@Felix, would it be ok for you when I propose a "*-sync-cfg80211-station_info.patch" for the 3.18 generic kernel
patches?
Quoted 1 line
Quoted 1 line
Quoted 1 line
Or do you have a good idea how to force the openwrt-routing batman-adv
package
Quoted 6 lines
Quoted 5 lines
Quoted 1 line
against the
cfg80211.hof compat-wireless?The openwrt mac80211 package exports its headers to $(STAGING_DIR)/usr/include/mac80211
You could just make the batman-adv package use those includes.
Quoted 8 lines
Quoted 1 line
Thanks, missed that. I will prepare a pull request for openwrt-routing.
Hm, didn't work out so well. It basically explodes right away because
possible_net_t,possible_read_pnet,possible_write_pnetis not defined on v3.18 when not using the mac80211 backporting headers. And when usign the backporting headers then it will explode because it conflicts with batman- adv's own backporting stuff. And without the batman-adv's backporting headers it will also not compile due to missing backporting hacks.
I think in the short term it would be a good idea to identify what's missing in the mac80211 backport support and make patches for it. You should also rebase those on top of backports.git and submit them upstream.
For a long term solution, you guys should probably start pushing for full batman-adv integration in kernel backports. Might be useful for increasing adoption of the latest stuff even outside of OpenWrt.
- Felix
Re: [B.A.T.M.A.N.] OpenWrt cfg80211.h station_info incompatibility with kernel+mac80211 [was: B.A.T.M.A.N. V leaves the nest v2]
On Thursday 21 January 2016 15:47:02 Felix Fietkau wrote:
[...]
Quoted 10 lines
Quoted 6 lines
Hm, didn't work out so well. It basically explodes right away because
possible_net_t,possible_read_pnet,possible_write_pnetis not defined on v3.18 when not using the mac80211 backporting headers. And when usign the backporting headers then it will explode because it conflicts with batman- adv's own backporting stuff. And without the batman-adv's backporting headers it will also not compile due to missing backporting hacks.I think in the short term it would be a good idea to identify what's missing in the mac80211 backport support and make patches for it. You should also rebase those on top of backports.git and submit them upstream.
Yes, this was already started after/during the linux plumbers conf 2014 [1]. But it died the horrible death of being ignored and neglected - especially after the multicast optimization was integrated.
Someone who want to use the mac80211 stuff with the batman-adv package can use the attached patch as starting point.
Kind regards,
Sven
[1] https://github.com/ordex/batman-adv-backport/commits/TODO-compat
From: Sven Eckelmann <[email protected]>
Date: Thu, 21 Jan 2016 16:04:20 +0100
Subject: [PATCH] batman-adv: Use the mac80211 compat layer
--- batman-adv/Makefile | 41 +++++++++++++++++++++++++++-------------- 1 file changed, 27 insertions(+), 14 deletions(-)
batman-adv/Makefile +27 -14
@@ -27,7 +27,7 @@ define KernelPackage/batman-adv URL:=https://www.open-mesh.org/ MAINTAINER:=Marek Lindner <[email protected]> SUBMENU:=Network Support - DEPENDS:=+KMOD_BATMAN_ADV_BLA:kmod-lib-crc16 +kmod-crypto-core +kmod-crypto-crc32c +kmod-lib-crc32c + DEPENDS:=+KMOD_BATMAN_ADV_BLA:kmod-lib-crc16 +kmod-crypto-core +kmod-crypto-crc32c +kmod-lib-crc32c +kmod-cfg80211 TITLE:=B.A.T.M.A.N. Adv FILES:=$(PKG_BUILD_DIR)/net/batman-adv/batman-adv.$(LINUX_KMOD_SUFFIX) AUTOLOAD:=$(call AutoLoad,50,batman-adv)
@@ -47,20 +47,37 @@ define Package/kmod-batman-adv/conffiles /etc/config/batman-adv endef -MAKE_BATMAN_ADV_ARGS += \ - CROSS_COMPILE="$(TARGET_CROSS)" \ - KERNELPATH="$(LINUX_DIR)" \ - ARCH="$(LINUX_KARCH)" \ - PATH="$(TARGET_PATH)" \ - SUBDIRS="$(PKG_BUILD_DIR)" \ - PWD="$(PKG_BUILD_DIR)" \ - LINUX_VERSION="$(LINUX_VERSION)" \ +PKG_EXTRA_KCONFIG:= \ + CONFIG_BATMAN_ADV=m \ CONFIG_BATMAN_ADV_DEBUG=$(if $(CONFIG_KMOD_BATMAN_ADV_DEBUG_LOG),y,n) \ CONFIG_BATMAN_ADV_BLA=$(if $(CONFIG_KMOD_BATMAN_ADV_BLA),y,n) \ CONFIG_BATMAN_ADV_DAT=$(if $(CONFIG_KMOD_BATMAN_ADV_DAT),y,n) \ CONFIG_BATMAN_ADV_MCAST=$(if $(CONFIG_KMOD_BATMAN_ADV_MCAST),y,n) \ CONFIG_BATMAN_ADV_NC=$(if $(CONFIG_KMOD_BATMAN_ADV_NC),y,n) \ - REVISION="" all + CONFIG_BATMAN_ADV_BATMAN_V=$(if $(CONFIG_KMOD_BATMAN_ADV_BATMAN_V),y,n) \ + +PKG_EXTRA_CFLAGS:= \ + $(patsubst CONFIG_%, -DCONFIG_%=1, $(patsubst %=m,%,$(filter %=m,$(PKG_EXTRA_KCONFIG)))) \ + $(patsubst CONFIG_%, -DCONFIG_%=1, $(patsubst %=y,%,$(filter %=y,$(PKG_EXTRA_KCONFIG)))) \ + +NOSTDINC_FLAGS = \ + -I$(PKG_BUILD_DIR)/net/batman-adv \ + -I$(STAGING_DIR)/usr/include/mac80211-backport/uapi \ + -I$(STAGING_DIR)/usr/include/mac80211-backport \ + -I$(STAGING_DIR)/usr/include/mac80211/uapi \ + -I$(STAGING_DIR)/usr/include/mac80211 \ + -include backport/backport.h + +define Build/Compile + +$(MAKE) $(PKG_JOBS) -C "$(LINUX_DIR)" \ + ARCH="$(LINUX_KARCH)" \ + CROSS_COMPILE="$(TARGET_CROSS)" \ + SUBDIRS="$(PKG_BUILD_DIR)/net/batman-adv" \ + $(PKG_EXTRA_KCONFIG) \ + EXTRA_CFLAGS="$(PKG_EXTRA_CFLAGS)" \ + NOSTDINC_FLAGS="$(NOSTDINC_FLAGS)" \ + modules +endef define Build/Prepare $(call Build/Prepare/Default)
@@ -68,10 +85,6 @@ define Build/Prepare $(PKG_BUILD_DIR)/net/batman-adv/main.h endef -define Build/Compile - $(MAKE) -C "$(PKG_BUILD_DIR)" $(MAKE_BATMAN_ADV_ARGS) -endef - define Build/Clean rm -rf $(BUILD_DIR)/$(PKG_NAME)/ endef