Skip to contentSkip to conversation

Basic Multicast Optimizations

B.A.T.M.A.N. Superseded Linus Lüssing 4 participants 12 messages May 24, 2013 – May 28, 2013 Kbuild

v2 of 14 Superseded Latest is v14 →

lore
Share & tools
lore
#0 lore raw

This is the second revision of the basic multicast optimization patches. It includes one functional fix and many style fixes, thanks to Antonios feedback.

Three things were not included from these suggestions:

* The default mcast TVLV flags were left as it is not as easy to switch. (see answer/explanation to [PATCH 2/3])
* for_each_pmc_rcu() was not moved to a commen net header file, only a TODO was added. Will do that after this code might have landed in net.

Additionally the limitation of 255 multicast listeners per originator was removed (it was more a relic of the pre-TT announcement mechanism and such limitations should be taken care of by the current TT infrastructure instead).

Cheers, Linus

#1 lore raw

PATCH v2 1/3 batman-adv: Multicast Listener Announcements via Translation Table

With this patch a node which has no bridge interface on top of its soft interface announces its local multicast listeners via the translation table.

---
 Makefile               |    2 +
 Makefile.kbuild        |    1 +
 compat.c               |   21 +++++
 compat.h               |   18 ++++
 gen-compat-autoconf.sh |    1 +
 main.c                 |    6 ++
 main.h                 |    9 ++
 multicast.c            |  213 ++++++++++++++++++++++++++++++++++++++++++++++++
 multicast.h            |   43 ++++++++++
 soft-interface.c       |    3 +
 sysfs.c                |    6 ++
 translation-table.c    |   22 ++++-
 types.h                |   12 +++
 13 files changed, 353 insertions(+), 4 deletions(-)
 create mode 100644 multicast.c
 create mode 100644 multicast.h
Makefile +2 -0
index 407cdc4..d7c6fa6 100644
@@ -27,6 +27,8 @@ export CONFIG_BATMAN_ADV_BLA=y
 export CONFIG_BATMAN_ADV_DAT=y
 # B.A.T.M.A.N network coding (catwoman):
 export CONFIG_BATMAN_ADV_NC=n
+# B.A.T.M.A.N. multicast optimizations:
+export CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS=y
 
 PWD:=$(shell pwd)
 KERNELPATH ?= /lib/modules/$(shell uname -r)/build
Makefile.kbuild +1 -0
index 8ddbfe6..d3efe3a 100644
@@ -38,3 +38,4 @@ batman-adv-y += soft-interface.o
 batman-adv-y += sysfs.o
 batman-adv-y += translation-table.o
 batman-adv-y += unicast.o
+batman-adv-$(CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS) += multicast.o
compat.c +21 -0
index 1f3a39d..27df197 100644
@@ -25,6 +25,27 @@
 #include <linux/version.h>
 #include "main.h"
 
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 38)
+
+void batadv_for_each_pmc_rcu_init(struct in_device *in_dev,
+				  struct ip_mc_list **pmc)
+{
+	read_lock(&in_dev->mc_list_lock);
+	*pmc = in_dev->mc_list;
+}
+
+bool batadv_for_each_pmc_rcu_check(struct in_device *in_dev,
+				   struct ip_mc_list *pmc)
+{
+	if (pmc == NULL)
+		read_unlock(&in_dev->mc_list_lock);
+
+	return (pmc != NULL);
+}
+
+#endif /* < KERNEL_VERSION(2, 6, 38) */
+
+
 #if LINUX_VERSION_CODE < KERNEL_VERSION(3, 0, 0)
 
 void batadv_free_rcu_gw_node(struct rcu_head *rcu)
compat.h +18 -0
index 8bd4f62..de1556a 100644
@@ -152,6 +152,24 @@ static inline int batadv_param_set_copystring(const char *val,
 
 #endif /* < KERNEL_VERSION(2, 6, 37) */
 
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 38)
+
+void batadv_for_each_pmc_rcu_init(struct in_device *in_dev,
+				  struct ip_mc_list **pmc);
+bool batadv_for_each_pmc_rcu_check(struct in_device *in_dev,
+				   struct ip_mc_list *pmc);
+
+#undef for_each_pmc_rcu
+#define for_each_pmc_rcu(in_dev, pmc)				\
+	for (batadv_for_each_pmc_rcu_init(in_dev, &pmc);	\
+	     batadv_for_each_pmc_rcu_check(in_dev, pmc);	\
+	     pmc = NULL)					\
+		for (; pmc != NULL; pmc = pmc->next)
+
+#endif /* < KERNEL_VERSION(2, 6, 38) */
+
+
 #if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 39)
 
 #define kstrtoul strict_strtoul
gen-compat-autoconf.sh +1 -0
index 78573e4..b56ac29 100755
@@ -39,6 +39,7 @@ gen_config() {
 gen_config 'CONFIG_BATMAN_ADV_DEBUG' ${CONFIG_BATMAN_ADV_DEBUG:="n"} >> "${TMP}"
 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_OPTIMIZATIONS' ${CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS:="y"} >> "${TMP}"
 gen_config 'CONFIG_BATMAN_ADV_NC' ${CONFIG_BATMAN_ADV_NC:="n"} >> "${TMP}"
 
 # only regenerate compat-autoconf.h when config was changed
main.c +6 -0
index af89c09..1c82c18 100644
@@ -33,6 +33,7 @@
 #include "bridge_loop_avoidance.h"
 #include "distributed-arp-table.h"
 #include "unicast.h"
+#include "multicast.h"
 #include "gateway_common.h"
 #include "hash.h"
 #include "bat_algo.h"
@@ -115,6 +116,9 @@ int batadv_mesh_init(struct net_device *soft_iface)
 	INIT_LIST_HEAD(&bat_priv->tt.changes_list);
 	INIT_LIST_HEAD(&bat_priv->tt.req_list);
 	INIT_LIST_HEAD(&bat_priv->tt.roam_list);
+#ifdef CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS
+	INIT_LIST_HEAD(&bat_priv->mcast.mla_list);
+#endif
 	INIT_HLIST_HEAD(&bat_priv->tvlv.container_list);
 	INIT_HLIST_HEAD(&bat_priv->tvlv.handler_list);
 
@@ -168,6 +172,8 @@ void batadv_mesh_free(struct net_device *soft_iface)
 	batadv_dat_free(bat_priv);
 	batadv_bla_free(bat_priv);
 
+	batadv_mcast_free(bat_priv);
+
 	/* Free the TT and the originator tables only after having terminated
 	 * all the other depending components which may use these structures for
 	 * their purposes.
main.h +9 -0
index 795345f..c865a9c 100644
@@ -141,6 +141,14 @@ enum batadv_uev_type {
 /* Append 'batman-adv: ' before kernel messages */
 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 
+/* identical to the one used in net/ipv4/igmp.c
+ * TODO: remove redundancy, put in some header file
+ */
+#define for_each_pmc_rcu(in_dev, pmc)				\
+	for (pmc = rcu_dereference(in_dev->mc_list);		\
+	     pmc != NULL;					\
+	     pmc = rcu_dereference(pmc->next_rcu))
+
 /* Kernel headers */
 
 #include <linux/mutex.h>	/* mutex */
@@ -155,6 +163,7 @@ enum batadv_uev_type {
 #include <linux/percpu.h>
 #include <linux/slab.h>
 #include <net/sock.h>		/* struct sock */
+#include <net/addrconf.h>	/* ipv6 address stuff */
 #include <net/rtnetlink.h>
 #include <linux/jiffies.h>
 #include <linux/seq_file.h>
multicast.c new file +213 -0

213 lines not on this page · raw message

multicast.h new file +43 -0

43 lines not on this page · raw message

soft-interface.c +3 -0
index a6ab9e4..b97dfa2 100644
@@ -457,6 +457,9 @@ static int batadv_softif_init_late(struct net_device *dev)
 #ifdef CONFIG_BATMAN_ADV_DAT
 	atomic_set(&bat_priv->distributed_arp_table, 1);
 #endif
+#ifdef CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS
+	atomic_set(&bat_priv->mcast_group_awareness, 1);
+#endif
 	atomic_set(&bat_priv->ap_isolation, 0);
 	atomic_set(&bat_priv->gw_mode, BATADV_GW_MODE_OFF);
 	atomic_set(&bat_priv->gw_sel_class, 20);
sysfs.c +6 -0
index cbbb77b..76390d0 100644
@@ -369,6 +369,9 @@ BATADV_ATTR_SIF_UINT(gw_sel_class, S_IRUGO | S_IWUSR, 1, BATADV_TQ_MAX_VALUE,
 		     batadv_post_gw_deselect);
 static BATADV_ATTR(gw_bandwidth, S_IRUGO | S_IWUSR, batadv_show_gw_bwidth,
 		   batadv_store_gw_bwidth);
+#ifdef CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS
+BATADV_ATTR_SIF_BOOL(mcast_group_awareness, S_IRUGO | S_IWUSR, NULL);
+#endif
 #ifdef CONFIG_BATMAN_ADV_DEBUG
 BATADV_ATTR_SIF_UINT(log_level, S_IRUGO | S_IWUSR, 0, BATADV_DBG_ALL, NULL);
 #endif
@@ -386,6 +389,9 @@ static struct batadv_attribute *batadv_mesh_attrs[] = {
 #ifdef CONFIG_BATMAN_ADV_DAT
 	&batadv_attr_distributed_arp_table,
 #endif
+#ifdef CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS
+	&batadv_attr_mcast_group_awareness,
+#endif
 	&batadv_attr_fragmentation,
 	&batadv_attr_ap_isolation,
 	&batadv_attr_routing_algo,
translation-table.c +18 -4
index df6b5cd..be5cfe1 100644
@@ -26,6 +26,7 @@
 #include "originator.h"
 #include "routing.h"
 #include "bridge_loop_avoidance.h"
+#include "multicast.h"
 
 #include <linux/crc32c.h>
 
@@ -274,14 +275,16 @@ void batadv_tt_local_add(struct net_device *soft_iface, const uint8_t *addr,
 {
 	struct batadv_priv *bat_priv = netdev_priv(soft_iface);
 	struct batadv_tt_local_entry *tt_local;
-	struct batadv_tt_global_entry *tt_global;
+	struct batadv_tt_global_entry *tt_global = NULL;
 	struct hlist_head *head;
 	struct batadv_tt_orig_list_entry *orig_entry;
 	int hash_added;
 	bool roamed_back = false;
 
 	tt_local = batadv_tt_local_hash_find(bat_priv, addr);
-	tt_global = batadv_tt_global_hash_find(bat_priv, addr);
+
+	if (!is_multicast_ether_addr(addr))
+		tt_global = batadv_tt_global_hash_find(bat_priv, addr);
 
 	if (tt_local) {
 		tt_local->last_seen = jiffies;
@@ -332,8 +335,11 @@ void batadv_tt_local_add(struct net_device *soft_iface, const uint8_t *addr,
 	tt_local->last_seen = jiffies;
 	tt_local->common.added_at = tt_local->last_seen;
 
-	/* the batman interface mac address should never be purged */
-	if (batadv_compare_eth(addr, soft_iface->dev_addr))
+	/* the batman interface mac and multicast addresses should never be
+	 * purged
+	 */
+	if (batadv_compare_eth(addr, soft_iface->dev_addr) ||
+	    is_multicast_ether_addr(addr))
 		tt_local->common.flags |= BATADV_TT_CLIENT_NOPURGE;
 
 	hash_added = batadv_hash_add(bat_priv->tt.local_hash, batadv_compare_tt,
@@ -919,6 +925,11 @@ add_orig_entry:
 	ret = true;
 
 out_remove:
+	/* Do not remove multicast addresses from the local hash on
+	 * global additions
+	 */
+	if (is_multicast_ether_addr(tt_addr))
+		goto out;
 
 	/* remove address from local hash if present */
 	local_flags = batadv_tt_local_remove(bat_priv, tt_addr,
@@ -2346,6 +2357,9 @@ void batadv_tt_local_commit_changes(struct batadv_priv *bat_priv)
 {
 	uint16_t changed_num = 0;
 
+	/* Update multicast addresses in local translation table */
+	batadv_mcast_mla_tt_update(bat_priv);
+
 	if (atomic_read(&bat_priv->tt.local_changes) < 1) {
 		if (!batadv_atomic_dec_not_zero(&bat_priv->tt.ogm_append_cnt))
 			batadv_tt_tvlv_container_update(bat_priv);
types.h +12 -0
index c84f5cc..5d73a75 100644
@@ -473,6 +473,12 @@ struct batadv_priv_dat {
 };
 #endif
 
+#ifdef CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS
+struct batadv_priv_mcast {
+	struct list_head mla_list;
+};
+#endif
+
 /**
  * struct batadv_priv_nc - per mesh interface network coding private data
  * @work: work queue callback item for cleanup
@@ -561,6 +567,9 @@ struct batadv_priv {
 #ifdef CONFIG_BATMAN_ADV_DAT
 	atomic_t distributed_arp_table;
 #endif
+#ifdef CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS
+	atomic_t mcast_group_awareness;
+#endif
 	atomic_t gw_mode;
 	atomic_t gw_sel_class;
 	atomic_t orig_interval;
@@ -595,6 +604,9 @@ struct batadv_priv {
 #ifdef CONFIG_BATMAN_ADV_DAT
 	struct batadv_priv_dat dat;
 #endif
+#ifdef CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS
+	struct batadv_priv_mcast mcast;
+#endif
 #ifdef CONFIG_BATMAN_ADV_NC
 	atomic_t network_coding;
 	struct batadv_priv_nc nc;
Signature · 1 line
1.7.10.4
#2 lore raw

Re: [B.A.T.M.A.N.] [PATCHv2 1/3] batman-adv: Multicast Listener Announcements via Translation Table

Quoted 62 lines · Linus Lüssing

With this patch a node which has no bridge interface on top of its soft interface announces its local multicast listeners via the translation table.

---
 Makefile               |    2 +
 Makefile.kbuild        |    1 +
 compat.c               |   21 +++++
 compat.h               |   18 ++++
 gen-compat-autoconf.sh |    1 +
 main.c                 |    6 ++
 main.h                 |    9 ++
 multicast.c            |  213 ++++++++++++++++++++++++++++++++++++++++++++++++
 multicast.h            |   43 ++++++++++
 soft-interface.c       |    3 +
 sysfs.c                |    6 ++
 translation-table.c    |   22 ++++-
 types.h                |   12 +++
 13 files changed, 353 insertions(+), 4 deletions(-)
 create mode 100644 multicast.c
 create mode 100644 multicast.h
Makefile +2 -0
index 407cdc4..d7c6fa6 100644
@@ -27,6 +27,8 @@ export CONFIG_BATMAN_ADV_BLA=y
 export CONFIG_BATMAN_ADV_DAT=y
 # B.A.T.M.A.N network coding (catwoman):
 export CONFIG_BATMAN_ADV_NC=n
+# B.A.T.M.A.N. multicast optimizations:
+export CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS=y
 
 PWD:=$(shell pwd)
 KERNELPATH ?= /lib/modules/$(shell uname -r)/build
Makefile.kbuild +1 -0
index 8ddbfe6..d3efe3a 100644
@@ -38,3 +38,4 @@ batman-adv-y += soft-interface.o
 batman-adv-y += sysfs.o
 batman-adv-y += translation-table.o
 batman-adv-y += unicast.o
+batman-adv-$(CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS) += multicast.o
compat.c +9 -0
index 1f3a39d..27df197 100644
@@ -25,6 +25,27 @@
 #include <linux/version.h>
 #include "main.h"
 
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 38)
+
+void batadv_for_each_pmc_rcu_init(struct in_device *in_dev,
+				  struct ip_mc_list **pmc)
+{
+	read_lock(&in_dev->mc_list_lock);
+	*pmc = in_dev->mc_list;
+}
+

compat.c line 46

Quoted 10 lines
compat.c +10 -0
+bool batadv_for_each_pmc_rcu_check(struct in_device *in_dev,
+				   struct ip_mc_list *pmc)
+{
+	if (pmc == NULL)
+		read_unlock(&in_dev->mc_list_lock);
+
+	return (pmc != NULL);
+}
+
+#endif /* < KERNEL_VERSION(2, 6, 38) */

I guess there is some more compat code required here.

For kernel 2.6.32 I get

compat.h:159:14: warning: ‘struct ip_mc_list’ declared inside parameter list [enabled by default] compat.h:159:14: warning: its scope is only this definition or declaration, which is probably not what you want [enabled by default] compat.h:159:14: warning: ‘struct in_device’ declared inside parameter list [enabled by default] compat.h:161:15: warning: ‘struct ip_mc_list’ declared inside parameter list [enabled by default] compat.h:161:15: warning: ‘struct in_device’ declared inside parameter list [enabled by default]

Am I missing config options? These structs are defined in linux/igmp.h though ...

Cheers,
Simon

#3 lore raw

PATCH v2 2/3 batman-adv: Announce new capability via multicast TVLV

In reply to Linus Lüssing

If the soft interface of a node is not part of a bridge then a node announces a new multicast TVLV: The according flag (BATADV_MCAST_LISTENER_ANNOUNCEMENT) signalizes that this node is announcing all of its multicast listeners via the translation table infrastructure. More precisely, all multicast listeners of scope greater than link-local for IPv4 and of scope greater
or equal to link-local for IPv6.

---
 main.c           |    4 ++++
 multicast.c      |   62 ++++++++++++++++++++++++++++++++++++++++++++++++++++--
 multicast.h      |    7 ++++++
 originator.c     |    3 +++
 packet.h         |    7 ++++++
 soft-interface.c |    1 +
 types.h          |    4 ++++
 7 files changed, 86 insertions(+), 2 deletions(-)
main.c +4 -0
index 1c82c18..6ab5b2d 100644
@@ -145,6 +145,10 @@ int batadv_mesh_init(struct net_device *soft_iface)
 	if (ret < 0)
 		goto err;
 
+	ret = batadv_mcast_init(bat_priv);
+	if (ret < 0)
+		goto err;
+
 	ret = batadv_gw_init(bat_priv);
 	if (ret < 0)
 		goto err;
multicast.c +60 -2

90 lines not on this page · raw message

multicast.h +7 -0
index 3b68e3b..9955a18 100644
@@ -24,6 +24,8 @@
 
 void batadv_mcast_mla_tt_update(struct batadv_priv *bat_priv);
 
+int batadv_mcast_init(struct batadv_priv *bat_priv);
+
 void batadv_mcast_free(struct batadv_priv *bat_priv);
 
 #else
@@ -33,6 +35,11 @@ static inline void batadv_mcast_mla_tt_update(struct batadv_priv *bat_priv)
 	return;
 }
 
+static inline int batadv_mcast_init(struct batadv_priv *bat_priv)
+{
+	return 0;
+}
+
 static inline void batadv_mcast_free(struct batadv_priv *bat_priv)
 {
 	return;
originator.c +3 -0
index 5d53d2f..acc0c2d 100644
@@ -257,6 +257,9 @@ struct batadv_orig_node *batadv_get_orig_node(struct batadv_priv *bat_priv,
 	reset_time = jiffies - 1 - msecs_to_jiffies(BATADV_RESET_PROTECTION_MS);
 	orig_node->bcast_seqno_reset = reset_time;
 	orig_node->batman_seqno_reset = reset_time;
+#ifdef CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS
+	orig_node->mcast_flags = BATADV_MCAST_LISTENER_ANNOUNCEMENT;
+#endif
 
 	atomic_set(&orig_node->bond_candidates, 0);
 
packet.h +7 -0
index 7d45890..c89df30 100644
@@ -97,6 +97,11 @@ enum batadv_unicast_frag_flags {
 	BATADV_UNI_FRAG_LARGETAIL = BIT(1),
 };
 
+/* multicast capabilities */
+enum batadv_mcast_flags {
+	BATADV_MCAST_LISTENER_ANNOUNCEMENT = BIT(0),
+};
+
 /* tt data subtypes */
 #define BATADV_TT_DATA_TYPE_MASK 0x0F
 
@@ -143,6 +148,7 @@ enum batadv_bla_claimframe {
  * @BATADV_TVLV_NC: network coding tvlv
  * @BATADV_TVLV_TT: translation table tvlv
  * @BATADV_TVLV_ROAM: roaming advertisement tvlv
+ * @BATADV_TVLV_MCAST: multicast capability tvlv
  */
 enum batadv_tvlv_type {
 	BATADV_TVLV_GW		= 0x01,
@@ -150,6 +156,7 @@ enum batadv_tvlv_type {
 	BATADV_TVLV_NC		= 0x03,
 	BATADV_TVLV_TT		= 0x04,
 	BATADV_TVLV_ROAM	= 0x05,
+	BATADV_TVLV_MCAST	= 0x06,
 };
 
 /* the destination hardware field in the ARP frame is used to
soft-interface.c +1 -0
index b97dfa2..8bdd649 100644
@@ -459,6 +459,7 @@ static int batadv_softif_init_late(struct net_device *dev)
 #endif
 #ifdef CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS
 	atomic_set(&bat_priv->mcast_group_awareness, 1);
+	atomic_set(&bat_priv->mcast.num_non_aware, 0);
 #endif
 	atomic_set(&bat_priv->ap_isolation, 0);
 	atomic_set(&bat_priv->gw_mode, BATADV_GW_MODE_OFF);
types.h +4 -0
index 5d73a75..c4e5545 100644
@@ -146,6 +146,9 @@ struct batadv_orig_node {
 	unsigned long last_seen;
 	unsigned long bcast_seqno_reset;
 	unsigned long batman_seqno_reset;
+#ifdef CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS
+	uint8_t mcast_flags;
+#endif
 	uint8_t capabilities;
 	atomic_t last_ttvn;
 	uint32_t tt_crc;
@@ -476,6 +479,7 @@ struct batadv_priv_dat {
 #ifdef CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS
 struct batadv_priv_mcast {
 	struct list_head mla_list;
+	atomic_t num_non_aware;
 };
 #endif
 
Signature · 1 line
1.7.10.4
#4 lore raw

Re: [B.A.T.M.A.N.] [PATCHv2 2/3] batman-adv: Announce new capability via multicast TVLV

Quoted 61 lines · Linus Lüssing

If the soft interface of a node is not part of a bridge then a node announces a new multicast TVLV: The according flag (BATADV_MCAST_LISTENER_ANNOUNCEMENT) signalizes that this node is announcing all of its multicast listeners via the translation table infrastructure. More precisely, all multicast listeners of scope greater than link-local for IPv4 and of scope greater
or equal to link-local for IPv6.

---
 main.c           |    4 ++++
 multicast.c      |   62 ++++++++++++++++++++++++++++++++++++++++++++++++++++--
 multicast.h      |    7 ++++++
 originator.c     |    3 +++
 packet.h         |    7 ++++++
 soft-interface.c |    1 +
 types.h          |    4 ++++
 7 files changed, 86 insertions(+), 2 deletions(-)
main.c +4 -0
index 1c82c18..6ab5b2d 100644
@@ -145,6 +145,10 @@ int batadv_mesh_init(struct net_device *soft_iface)
 	if (ret < 0)
 		goto err;
 
+	ret = batadv_mcast_init(bat_priv);
+	if (ret < 0)
+		goto err;
+
 	ret = batadv_gw_init(bat_priv);
 	if (ret < 0)
 		goto err;
multicast.c +19 -0
index 56a1128..36e4c59 100644
@@ -204,10 +213,59 @@ out:
 }
 
 /**
+ * batadv_mcast_tvlv_ogm_handler_v1 - process incoming multicast tvlv container
+ * @bat_priv: the bat priv with all the soft interface information
+ * @orig: the orig_node of the ogm
+ * @flags: flags indicating the tvlv state (see batadv_tvlv_handler_flags)
+ * @tvlv_value: tvlv buffer containing the multicast data
+ * @tvlv_value_len: tvlv buffer length
+ */
+static void batadv_mcast_tvlv_ogm_handler_v1(struct batadv_priv *bat_priv,
+					     struct batadv_orig_node *orig,
+					     uint8_t flags,
+					     void *tvlv_value,
+					     uint16_t tvlv_value_len)
+{
+	uint8_t mcast_flags = BATADV_NO_FLAGS;
+
+	/* only fetch the tvlv value if the handler wasn't called via the
+	 * CIFNOTFND flag and if there is data to fetch
+	 */
+	if (!(flags & BATADV_TVLV_HANDLER_OGM_CIFNOTFND) &&

multicast.c out: · line 244

Quoted 10 lines
multicast.c +10 -0
+	    (tvlv_value) && (tvlv_value_len == sizeof(mcast_flags)))
+		mcast_flags = *(uint8_t *)tvlv_value;
+
+	if (!(mcast_flags & BATADV_MCAST_LISTENER_ANNOUNCEMENT) &&
+	    orig->mcast_flags & BATADV_MCAST_LISTENER_ANNOUNCEMENT) {
+		atomic_inc(&bat_priv->mcast.num_non_aware);
+	} else if (mcast_flags & BATADV_MCAST_LISTENER_ANNOUNCEMENT &&
+		   !(orig->mcast_flags & BATADV_MCAST_LISTENER_ANNOUNCEMENT)) {
+		atomic_dec(&bat_priv->mcast.num_non_aware);
+	}

What happens if the orig_node is removed/times out? I can't see where num_non_aware is updated in this case.

We do have a work struct touch every orig every second anyway (batadv_purge_orig), maybe put counting multicast aware routers there? Then you don't have to care about synchronizing stuff ...

Quoted 27 lines
multicast.c +20 -0
+
+	orig->mcast_flags = mcast_flags;
+}
+
+/**
+ * batadv_mcast_init - initialize the multicast optimizations structures
+ * @bat_priv: the bat priv with all the soft interface information
+ */
+int batadv_mcast_init(struct batadv_priv *bat_priv)
+{
+	batadv_tvlv_handler_register(bat_priv, batadv_mcast_tvlv_ogm_handler_v1,
+				     NULL, BATADV_TVLV_MCAST, 1,
+				     BATADV_TVLV_HANDLER_OGM_CIFNOTFND);
+	return 0;
+}
+
+/**
  * batadv_mcast_free - free the multicast optimizations structures
  * @bat_priv: the bat priv with all the soft interface information
  */
 void batadv_mcast_free(struct batadv_priv *bat_priv)
 {
+	batadv_tvlv_container_unregister(bat_priv, BATADV_TVLV_MCAST, 1);
+	batadv_tvlv_handler_unregister(bat_priv, BATADV_TVLV_MCAST, 1);
+
 	batadv_mcast_mla_tt_clean(bat_priv, NULL);
 }
#5 lore raw

PATCH v2 3/3 batman-adv: Modified forwarding behaviour for multicast packets

In reply to Linus Lüssing

With this patch a multicast packet is not always simply flooded anymore, the bevahiour for the following cases is changed to reduce unnecessary overhead:

If all nodes within the horizon of a certain node have signalized multicast listener announcement capability (BATADV_MCAST_LISTENER_ANNOUNCEMENT) then an IPv6 multicast packet with a destination of IPv6 link-local scope coming from the upstream of this node...

* ...is dropped if there is no according multicast listener in the translation table,
* ...is forwarded via unicast if there is a single node with interested multicast listeners
* ...and otherwise still gets flooded.

---
 multicast.c         |   43 +++++++++++++++++++++++++++++++++
 multicast.h         |    8 +++++++
 soft-interface.c    |   10 ++++++++
 translation-table.c |   66 +++++++++++++++++++++++++++++++++++++++++++++++++++
 translation-table.h |    1 +
 5 files changed, 128 insertions(+)
multicast.c +43 -0
index 36e4c59..bd55c8f 100644
@@ -213,6 +213,49 @@ out:
 }
 
 /**
+ * batadv_mcast_flood - check on how to forward a multicast packet
+ * @skb: The multicast packet to check
+ * @bat_priv: the bat priv with all the soft interface information
+ *
+ * Return 1 if the packet should be flooded, 0 if it should be forwarded
+ * via unicast or -1 if it should be drooped.
+ */
+int batadv_mcast_flood(struct sk_buff *skb, struct batadv_priv *bat_priv)
+{
+	struct ethhdr *ethhdr = (struct ethhdr *)(skb->data);
+	struct ipv6hdr *ip6hdr;
+	int count, ret = 1;
+
+	if (atomic_read(&bat_priv->mcast_group_awareness) &&
+	    !atomic_read(&bat_priv->mcast.num_non_aware) &&
+	    ntohs(ethhdr->h_proto) == ETH_P_IPV6) {
+		if (!pskb_may_pull(skb, sizeof(*ethhdr) + sizeof(*ip6hdr))) {
+			ret = -1;
+			goto out;
+		}
+
+		ip6hdr = ipv6_hdr(skb);
+
+		/* TODO: Implement Multicast Router Discovery, then add
+		 * scope >= IPV6_ADDR_SCOPE_LINKLOCAL, too
+		 */
+		if (IPV6_ADDR_MC_SCOPE(&ip6hdr->daddr) !=
+		    IPV6_ADDR_SCOPE_LINKLOCAL)
+			goto out;
+
+		count = batadv_tt_global_hash_count(bat_priv, ethhdr->h_dest);
+
+		if (!count)
+			ret = -1;
+		else if (count == 1)
+			ret = 0;
+	}
+
+out:
+	return ret;
+}
+
+/**
  * batadv_mcast_tvlv_ogm_handler_v1 - process incoming multicast tvlv container
  * @bat_priv: the bat priv with all the soft interface information
  * @orig: the orig_node of the ogm
multicast.h +8 -0
index 9955a18..0c2baad 100644
@@ -24,6 +24,8 @@
 
 void batadv_mcast_mla_tt_update(struct batadv_priv *bat_priv);
 
+int batadv_mcast_flood(struct sk_buff *skb, struct batadv_priv *bat_priv);
+
 int batadv_mcast_init(struct batadv_priv *bat_priv);
 
 void batadv_mcast_free(struct batadv_priv *bat_priv);
@@ -35,6 +37,12 @@ static inline void batadv_mcast_mla_tt_update(struct batadv_priv *bat_priv)
 	return;
 }
 
+static inline int batadv_mcast_flood(struct sk_buff *skb,
+				     struct batadv_priv *bat_priv)
+{
+	return 1;
+}
+
 static inline int batadv_mcast_init(struct batadv_priv *bat_priv)
 {
 	return 0;
soft-interface.c +10 -0
index 8bdd649..83e4679 100644
@@ -36,6 +36,7 @@
 #include <linux/if_vlan.h>
 #include <linux/if_ether.h>
 #include "unicast.h"
+#include "multicast.h"
 #include "bridge_loop_avoidance.h"
 #include "network-coding.h"
 
@@ -222,6 +223,15 @@ static int batadv_interface_tx(struct sk_buff *skb,
 		}
 	}
 
+	if (do_bcast && !is_broadcast_ether_addr(ethhdr->h_dest)) {
+		ret = batadv_mcast_flood(skb, bat_priv);
+		if (ret < 0)
+			goto dropped;
+
+		if (!ret)
+			do_bcast = false;
+	}
+
 	/* ethernet packet should be broadcasted */
 	if (do_bcast) {
 		primary_if = batadv_primary_if_get_selected(bat_priv);
translation-table.c +66 -0

72 lines not on this page · raw message

translation-table.h +1 -0
index 015d8b9..5986c57 100644
@@ -31,6 +31,7 @@ int batadv_tt_global_seq_print_text(struct seq_file *seq, void *offset);
 void batadv_tt_global_del_orig(struct batadv_priv *bat_priv,
 			       struct batadv_orig_node *orig_node,
 			       const char *message);
+int batadv_tt_global_hash_count(struct batadv_priv *bat_priv, const void *data);
 struct batadv_orig_node *batadv_transtable_search(struct batadv_priv *bat_priv,
 						  const uint8_t *src,
 						  const uint8_t *addr);
Signature · 1 line
1.7.10.4
#6 lore raw

Re: [B.A.T.M.A.N.] [PATCHv2 3/3] batman-adv: Modified forwarding behaviour for multicast packets

Only a few style suggestions here, nothing critical and you can ignore them if you don't like them. :)

Quoted 39 lines · Linus Lüssing

With this patch a multicast packet is not always simply flooded anymore, the bevahiour for the following cases is changed to reduce unnecessary overhead:

If all nodes within the horizon of a certain node have signalized multicast listener announcement capability (BATADV_MCAST_LISTENER_ANNOUNCEMENT) then an IPv6 multicast packet with a destination of IPv6 link-local scope coming from the upstream of this node...

* ...is dropped if there is no according multicast listener in the translation table,
* ...is forwarded via unicast if there is a single node with interested multicast listeners
* ...and otherwise still gets flooded.

---
 multicast.c         |   43 +++++++++++++++++++++++++++++++++
 multicast.h         |    8 +++++++
 soft-interface.c    |   10 ++++++++
 translation-table.c |   66 +++++++++++++++++++++++++++++++++++++++++++++++++++
 translation-table.h |    1 +
 5 files changed, 128 insertions(+)
multicast.c +6 -0
index 36e4c59..bd55c8f 100644
@@ -213,6 +213,49 @@ out:
 }
 
 /**
+ * batadv_mcast_flood - check on how to forward a multicast packet
+ * @skb: The multicast packet to check
+ * @bat_priv: the bat priv with all the soft interface information
+ *
+ * Return 1 if the packet should be flooded, 0 if it should be forwarded
+ * via unicast or -1 if it should be drooped.

multicast.c out: · line 231

Quoted 10 lines
multicast.c +10 -0
+ */
+int batadv_mcast_flood(struct sk_buff *skb, struct batadv_priv *bat_priv)
+{
+	struct ethhdr *ethhdr = (struct ethhdr *)(skb->data);
+	struct ipv6hdr *ip6hdr;
+	int count, ret = 1;
+
+	if (atomic_read(&bat_priv->mcast_group_awareness) &&
+	    !atomic_read(&bat_priv->mcast.num_non_aware) &&
+	    ntohs(ethhdr->h_proto) == ETH_P_IPV6) {

You can safe an indendation below if you return -1 immediately here if the statement above is false. Also multiple statements might be better for readability and later changes, e.g.

if (!atomic_read(&bat_priv->mcast_group_awareness))

	return 1;

if (atomic_read(&bat_priv->mcast.num_non_aware))

	return 1;

if (ntohs(ethhdr->h_proto) != ETH_P_IPV6)

	return 1;

multicast.c out: · line 235

Quoted 4 lines
multicast.c +4 -0
+		if (!pskb_may_pull(skb, sizeof(*ethhdr) + sizeof(*ip6hdr))) {
+			ret = -1;
+			goto out;
+		}

You could directly return -1 here, the out label is not needed (as we don't unlock/free anything here).

I don't quite understand why you return -1, maybe the packet could still be forwarded even if it could not be pulled?

Quoted 7 lines
multicast.c +7 -0
+
+		ip6hdr = ipv6_hdr(skb);
+
+		/* TODO: Implement Multicast Router Discovery, then add
+		 * scope >= IPV6_ADDR_SCOPE_LINKLOCAL, too
+		 */
+		if (IPV6_ADDR_MC_SCOPE(&ip6hdr->daddr) !=

multicast.c out: · line 252

Quoted 10 lines
multicast.c +10 -0
+		    IPV6_ADDR_SCOPE_LINKLOCAL)
+			goto out;
+
+		count = batadv_tt_global_hash_count(bat_priv, ethhdr->h_dest);
+
+		if (!count)
+			ret = -1;
+		else if (count == 1)
+			ret = 0;
+	}

You could use a switch statement here instead for readability, e.g.:

switch (count) {
case 0:

	return -1;
case 1:
	return 0;
default:
	return 1;
}
Quoted 22 lines
multicast.c +6 -0
+
+out:
+	return ret;
+}
+
+/**
  * batadv_mcast_tvlv_ogm_handler_v1 - process incoming multicast tvlv container
  * @bat_priv: the bat priv with all the soft interface information
  * @orig: the orig_node of the ogm
soft-interface.c +1 -0
index 8bdd649..83e4679 100644
@@ -36,6 +36,7 @@
 #include <linux/if_vlan.h>
 #include <linux/if_ether.h>
 #include "unicast.h"
+#include "multicast.h"
 #include "bridge_loop_avoidance.h"
 #include "network-coding.h"
 

soft-interface.c static int batadv_interface_tx(struct sk_buff *skb, · line 226

Quoted 4 lines
soft-interface.c +1 -0
@@ -222,6 +223,15 @@ static int batadv_interface_tx(struct sk_buff *skb,
 		}
 	}
 
+	if (do_bcast && !is_broadcast_ether_addr(ethhdr->h_dest)) {

I'd suggest to put this inside the "is_multicast_etheraddr()" above to make more clear that this handles multicast packets. I was a little confused by the do_bcast && !is_broadcast_ether_addr() first, but that might just be me.

Cheers,
Simon

#7 lore raw

Re: [B.A.T.M.A.N.] [PATCHv2 3/3] batman-adv: Modified forwarding behaviour for multicast packets

Quoted 4 lines · On Tuesday, May 28, 2013 23:40:54 Simon Wunderlich wrote:

You can safe an indendation below if you return -1 immediately here if the statement above is false. Also multiple statements might be better for readability and later changes, e.g.

Quoted 8 lines

if (!atomic_read(&bat_priv->mcast_group_awareness))

        return 1;

if (atomic_read(&bat_priv->mcast.num_non_aware))

        return 1;

if (ntohs(ethhdr->h_proto) != ETH_P_IPV6)

        return 1;

A quick hint from my end: The kernel maintainer don't like code that returns "random" values like 1/0/-1. Either use bool or meaningful defines instead of raw numbers.

Cheers,
Marek

#8 lore raw

Re: [B.A.T.M.A.N.] [PATCHv2 3/3] batman-adv: Modified forwarding behaviour for multicast packets

Quoted 11 lines · Marek Lindner
Quoted 10 lines · On Tuesday, May 28, 2013 23:40:54 Simon Wunderlich wrote:

You can safe an indendation below if you return -1 immediately here if the statement above is false. Also multiple statements might be better for readability and later changes, e.g.

if (!atomic_read(&bat_priv->mcast_group_awareness))

        return 1;

if (atomic_read(&bat_priv->mcast.num_non_aware))

        return 1;
Quoted 6 lines
Quoted 2 lines

if (ntohs(ethhdr->h_proto) != ETH_P_IPV6)

        return 1;

A quick hint from my end: The kernel maintainer don't like code that returns "random" values like 1/0/-1. Either use bool or meaningful defines instead of raw numbers.

Very good suggestion.
I think you can define an enum for this. It will make the code much easier to understand.

Cheers,

Signature · 4 lines
Antonio Quartulli

..each of us alone is worth nothing..
Ernesto "Che" Guevara
#9 lore raw
In reply to Linus Lüssing
Quoted 11 lines · Linus Lüssing

This is the second revision of the basic multicast optimization patches. It includes one functional fix and many style fixes, thanks to Antonios feedback.

Three things were not included from these suggestions:

* The default mcast TVLV flags were left as it is not as easy to switch. (see answer/explanation to [PATCH 2/3])
* for_each_pmc_rcu() was not moved to a commen net header file, only a TODO was added. Will do that after this code might have landed in net.

* VLAN support was not added yet as this will be some more work. Any vlan frames with a multicast destination will still get flooded for now so at least these patches shouldn't introduce any regressions for those. However I have to admit that I didn't think of VLANs yet and Marek suggested to check how much extra work it might create later compared to adding that right from the start. I'll have a look at that.

Quoted 8 lines

Additionally the limitation of 255 multicast listeners per originator was removed (it was more a relic of the pre-TT announcement mechanism and such limitations should be taken care of by the current TT infrastructure instead).

Cheers, Linus

#10 lore raw
Quoted 12 lines · Linus Lüssing
Quoted 11 lines · Linus Lüssing

This is the second revision of the basic multicast optimization patches. It includes one functional fix and many style fixes, thanks to Antonios feedback.

Three things were not included from these suggestions:

* The default mcast TVLV flags were left as it is not as easy to switch. (see answer/explanation to [PATCH 2/3])
* for_each_pmc_rcu() was not moved to a commen net header file, only a TODO was added. Will do that after this code might have landed in net.

Quoted 7 lines

* VLAN support was not added yet as this will be some more work. Any vlan frames with a multicast destination will still get flooded for now so at least these patches shouldn't introduce any regressions for those. However I have to admit that I didn't think of VLANs yet and Marek suggested to check how much extra work it might create later compared to adding that right from the start. I'll have a look at that.

I think (but I am not sure) that the TT-VLAN feature that I just sent to the ml (I have to send v3..) could probably help in this direction, because a TT entry is not defined by its MAC address only anymore, but by the couple {MAC, vid}.

Since your code uses the TT to store the multicast addresses, I think that VLAN support can be added "easily"[tm] after the TT-VLAN code will be in.

Cheers,

Signature · 4 lines
Antonio Quartulli

..each of us alone is worth nothing..
Ernesto "Che" Guevara
#11 lore raw
In reply to Linus Lüssing

Hi,

Quoted 3 lines · On Friday, May 24, 2013 16:02:25 Linus Lüssing wrote:

* for_each_pmc_rcu() was not moved to a commen net header file, only a TODO was added. Will do that after this code might have landed in net.

we discussed this further and concluded that we can't simply make that decision. The order in which patches are going to be merged into net has to be decided by David Miller. So, please prepare a patch destined for net, send it to netdev (cc the batman ml if you like) to explain what we are planning to do and ask for advice.

Cheers,
Marek