[B.A.T.M.A.N.] Basic Multicast Optimizations

[B.A.T.M.A.N.] Basic Multicast Optimizations

[Linus Lüssing]

This is the fourth revision of the basic multicast optimization patches.

It adds a missing enum keyword to the return type of a mcast_forw_mode().

Cheers, Linus

[B.A.T.M.A.N.] [PATCHv4 1/3] batman-adv: Multicast Listener Announcements via Translation Table

[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.

Signed-off-by: Linus Lüssing <linus.luessing@web.de>
---
 Makefile               |    2 +
 Makefile.kbuild        |    1 +
 compat.h               |   20 ++++-
 gen-compat-autoconf.sh |    1 +
 main.c                 |    6 ++
 main.h                 |    1 +
 multicast.c            |  209 ++++++++++++++++++++++++++++++++++++++++++++++++
 multicast.h            |   43 ++++++++++
 soft-interface.c       |    3 +
 sysfs.c                |    6 ++
 translation-table.c    |   22 ++++-
 types.h                |   12 +++
 12 files changed, 321 insertions(+), 5 deletions(-)
 create mode 100644 multicast.c
 create mode 100644 multicast.h

diff --git a/Makefile b/Makefile
index 407cdc4..d7c6fa6 100644
--- a/Makefile
+++ b/Makefile
@@ -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
diff --git a/Makefile.kbuild b/Makefile.kbuild
index 4f4aabb..f154c5b 100644
--- a/Makefile.kbuild
+++ b/Makefile.kbuild
@@ -38,3 +38,4 @@ batman-adv-y += send.o
 batman-adv-y += soft-interface.o
 batman-adv-y += sysfs.o
 batman-adv-y += translation-table.o
+batman-adv-$(CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS) += multicast.o
diff --git a/compat.h b/compat.h
index dbf1926..ddc1b29 100644
--- a/compat.h
+++ b/compat.h
@@ -103,13 +103,31 @@ static inline struct sk_buff *netdev_alloc_skb_ip_align(struct net_device *dev,
 
 #define pr_warn pr_warning
 
+#undef  netdev_for_each_mc_addr
+#define netdev_for_each_mc_addr(mclist, dev) \
+	for (mclist = (struct bat_dev_addr_list *)dev->mc_list; mclist; \
+	     mclist = (struct bat_dev_addr_list *)mclist->next)
+
+/* Note, that this breaks the usage of the normal 'struct netdev_hw_addr'
+ * for kernels < 2.6.35 in batman-adv!
+ */
+#define netdev_hw_addr batadv_dev_addr_list
+struct batadv_dev_addr_list {
+	struct dev_addr_list *next;
+	u8  addr[MAX_ADDR_LEN];
+	u8  da_addrlen;
+	u8  da_synced;
+	int da_users;
+	int da_gusers;
+};
+
 #endif /* < KERNEL_VERSION(2, 6, 35) */
 
 
 #if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 36)
 
 #define __rcu
-#define IFF_BRIDGE_PORT  0 || (hard_iface->net_dev->br_port ? 1 : 0)
+#define IFF_BRIDGE_PORT  0 || (soft_iface->br_port ? 1 : 0)
 
 struct kernel_param_ops {
 	/* Returns 0, or -errno.  arg is in kp->arg. */
diff --git a/gen-compat-autoconf.sh b/gen-compat-autoconf.sh
index 78573e4..b56ac29 100755
--- a/gen-compat-autoconf.sh
+++ b/gen-compat-autoconf.sh
@@ -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
diff --git a/main.c b/main.c
index 5d30fe0..fc5789c 100644
--- a/main.c
+++ b/main.c
@@ -32,6 +32,7 @@
 #include "gateway_client.h"
 #include "bridge_loop_avoidance.h"
 #include "distributed-arp-table.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.
diff --git a/main.h b/main.h
index 1f0b135..87db514 100644
--- a/main.h
+++ b/main.h
@@ -164,6 +164,7 @@ 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>
diff --git a/multicast.c b/multicast.c
new file mode 100644
index 0000000..d6d8b62
--- /dev/null
+++ b/multicast.c
@@ -0,0 +1,209 @@
+/* Copyright (C) 2013 B.A.T.M.A.N. contributors:
+ *
+ * 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
+ * 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, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
+ * 02110-1301, USA
+ */
+
+#include "main.h"
+#include "originator.h"
+#include "hard-interface.h"
+#include "translation-table.h"
+
+struct batadv_hw_addr {
+	struct list_head	list;
+	unsigned char addr[ETH_ALEN];
+};
+
+/**
+ * batadv_mcast_mla_local_collect - collect local multicast listeners
+ * @dev: the device to collect multicast addresses from
+ * @mcast_list: a list to put found addresses into
+ *
+ * Collect multicast addresses of the local multicast listeners
+ * on the given interface, dev, in the given mcast_list.
+ *
+ * Return -ENOMEM on memory allocation error or the number of
+ * items added to the mcast_list otherwise.
+ */
+static int batadv_mcast_mla_local_collect(struct net_device *dev,
+					  struct list_head *mcast_list)
+{
+	struct netdev_hw_addr *mc_list_entry;
+	struct batadv_hw_addr *new;
+	int ret = 0;
+
+	netif_addr_lock_bh(dev);
+	netdev_for_each_mc_addr(mc_list_entry, dev) {
+		new = kmalloc(sizeof(*new), GFP_ATOMIC);
+		if (!new) {
+			ret = -ENOMEM;
+			break;
+		}
+
+		memcpy(&new->addr, &mc_list_entry->addr, ETH_ALEN);
+		list_add(&new->list, mcast_list);
+		ret++;
+	}
+	netif_addr_unlock_bh(dev);
+
+	return ret;
+}
+
+/**
+ * batadv_mcast_mla_is_duplicate - check whether an address is in a list
+ * @mcast_addr: the multicast address to check
+ * @mcast_list: the list with multicast addresses to search in
+ *
+ * Return true if the given address is already in the given list.
+ * Otherwise returns false.
+ */
+static bool batadv_mcast_mla_is_duplicate(uint8_t *mcast_addr,
+					  struct list_head *mcast_list)
+{
+	struct batadv_hw_addr *mcast_entry;
+
+	list_for_each_entry(mcast_entry, mcast_list, list)
+		if (batadv_compare_eth(mcast_entry->addr, mcast_addr))
+			return true;
+
+	return false;
+}
+
+/**
+ * batadv_mcast_mla_collect_free - free a list of multicast addresses
+ * @mcast_list: the list to free
+ *
+ * Remove and free all items in the given mcast_list.
+ */
+void batadv_mcast_mla_collect_free(struct list_head *mcast_list)
+{
+	struct batadv_hw_addr *mcast_entry, *tmp;
+
+	list_for_each_entry_safe(mcast_entry, tmp, mcast_list, list) {
+		list_del(&mcast_entry->list);
+		kfree(mcast_entry);
+	}
+}
+
+/**
+ * batadv_mcast_mla_tt_clean - clean up multicast listener announcements
+ * @bat_priv: the bat priv with all the soft interface information
+ * @mcast_list: a list of addresses which should _not_ be removed
+ *
+ * Retract the announcement of any multicast listener from the
+ * translation table except the ones listed in the given mcast_list.
+ *
+ * If mcast_list is NULL then all are retracted.
+ */
+static void batadv_mcast_mla_tt_clean(struct batadv_priv *bat_priv,
+				      struct list_head *mcast_list)
+{
+	struct batadv_hw_addr *mcast_entry, *tmp;
+
+	list_for_each_entry_safe(mcast_entry, tmp, &bat_priv->mcast.mla_list,
+				 list) {
+		if (mcast_list &&
+		    batadv_mcast_mla_is_duplicate(mcast_entry->addr,
+						  mcast_list))
+			continue;
+
+		batadv_tt_local_remove(bat_priv, mcast_entry->addr,
+				       "mcast TT outdated", false);
+
+		list_del(&mcast_entry->list);
+		kfree(mcast_entry);
+	}
+}
+
+/**
+ * batadv_mcast_mla_tt_add - add multicast listener announcements
+ * @bat_priv: the bat priv with all the soft interface information
+ * @mcast_list: a list of addresses which are going to get added
+ *
+ * Add multicast listener announcements from the given mcast_list to the
+ * translation table if they have not been added yet.
+ */
+static void batadv_mcast_mla_tt_add(struct batadv_priv *bat_priv,
+				    struct list_head *mcast_list)
+{
+	struct batadv_hw_addr *mcast_entry, *tmp;
+
+	if (!mcast_list)
+		return;
+
+	list_for_each_entry_safe(mcast_entry, tmp, mcast_list, list) {
+		if (batadv_mcast_mla_is_duplicate(mcast_entry->addr,
+						  &bat_priv->mcast.mla_list))
+			continue;
+
+		batadv_tt_local_add(bat_priv->soft_iface, mcast_entry->addr,
+				    BATADV_NULL_IFINDEX);
+		list_move_tail(&mcast_entry->list, &bat_priv->mcast.mla_list);
+	}
+}
+
+/**
+ * batadv_mcast_mla_tt_update - update the own MLAs
+ * @bat_priv: the bat priv with all the soft interface information
+ *
+ * Update the own multicast listener announcements in the translation
+ * table. Also take care of registering or unregistering the multicast
+ * tvlv depending on whether the user activated or deactivated
+ * multicast optimizations.
+ */
+void batadv_mcast_mla_tt_update(struct batadv_priv *bat_priv)
+{
+	struct net_device *soft_iface = bat_priv->soft_iface;
+	struct list_head mcast_list;
+	int ret;
+	static bool enabled;
+
+	INIT_LIST_HEAD(&mcast_list);
+
+	/* Avoid attaching MLAs, if multicast optimization is disabled
+	 * or there is a bridge on top of our soft interface (TODO)
+	 */
+	if (!atomic_read(&bat_priv->mcast_group_awareness) ||
+	    bat_priv->soft_iface->priv_flags & IFF_BRIDGE_PORT) {
+		if (enabled)
+			enabled = false;
+
+		goto update;
+	}
+
+	if (!enabled)
+		enabled = true;
+
+	ret = batadv_mcast_mla_local_collect(soft_iface, &mcast_list);
+	if (ret < 0)
+		goto out;
+
+update:
+	batadv_mcast_mla_tt_clean(bat_priv, &mcast_list);
+	batadv_mcast_mla_tt_add(bat_priv, &mcast_list);
+
+out:
+	batadv_mcast_mla_collect_free(&mcast_list);
+}
+
+/**
+ * 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_mcast_mla_tt_clean(bat_priv, NULL);
+}
diff --git a/multicast.h b/multicast.h
new file mode 100644
index 0000000..3b68e3b
--- /dev/null
+++ b/multicast.h
@@ -0,0 +1,43 @@
+/* Copyright (C) 2013 B.A.T.M.A.N. contributors:
+ *
+ * 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
+ * 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, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
+ * 02110-1301, USA
+ */
+
+#ifndef _NET_BATMAN_ADV_MULTICAST_H_
+#define _NET_BATMAN_ADV_MULTICAST_H_
+
+#ifdef CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS
+
+void batadv_mcast_mla_tt_update(struct batadv_priv *bat_priv);
+
+void batadv_mcast_free(struct batadv_priv *bat_priv);
+
+#else
+
+static inline void batadv_mcast_mla_tt_update(struct batadv_priv *bat_priv)
+{
+	return;
+}
+
+static inline void batadv_mcast_free(struct batadv_priv *bat_priv)
+{
+	return;
+}
+
+#endif /* CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS */
+
+#endif /* _NET_BATMAN_ADV_MULTICAST_H_ */
diff --git a/soft-interface.c b/soft-interface.c
index d4fdc61..52244b0 100644
--- a/soft-interface.c
+++ b/soft-interface.c
@@ -456,6 +456,9 @@ 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);
diff --git a/sysfs.c b/sysfs.c
index b70ae52..cc4e7e5 100644
--- a/sysfs.c
+++ b/sysfs.c
@@ -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,
diff --git a/translation-table.c b/translation-table.c
index 7580ee5..0be7b15 100644
--- a/translation-table.c
+++ b/translation-table.c
@@ -26,6 +26,7 @@
 #include "originator.h"
 #include "routing.h"
 #include "bridge_loop_avoidance.h"
+#include "multicast.h"
 
 #include <linux/crc32c.h>
 
@@ -277,14 +278,16 @@ 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;
@@ -335,8 +338,11 @@ 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,
@@ -922,6 +928,11 @@ 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,
@@ -2348,6 +2359,9 @@ 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);
diff --git a/types.h b/types.h
index 304f3d0..8175424 100644
--- a/types.h
+++ b/types.h
@@ -501,6 +501,12 @@ 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
@@ -591,6 +597,9 @@ 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;
@@ -625,6 +634,9 @@ 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;
-- 
1.7.10.4

[B.A.T.M.A.N.] [PATCHv4 2/3] batman-adv: Announce new capability via multicast TVLV

[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.

Signed-off-by: Linus Lüssing <linus.luessing@web.de>
---
 main.c           |    4 ++++
 multicast.c      |   68 ++++++++++++++++++++++++++++++++++++++++++++++++++++--
 multicast.h      |   14 +++++++++++
 originator.c     |    6 +++++
 packet.h         |    7 ++++++
 soft-interface.c |    1 +
 types.h          |    4 ++++
 7 files changed, 102 insertions(+), 2 deletions(-)

diff --git a/main.c b/main.c
index fc5789c..968f77a 100644
--- a/main.c
+++ b/main.c
@@ -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;
diff --git a/multicast.c b/multicast.c
index d6d8b62..6375347 100644
--- a/multicast.c
+++ b/multicast.c
@@ -170,6 +170,7 @@ void batadv_mcast_mla_tt_update(struct batadv_priv *bat_priv)
 	struct list_head mcast_list;
 	int ret;
 	static bool enabled;
+	uint8_t mcast_flags;
 
 	INIT_LIST_HEAD(&mcast_list);
 
@@ -178,14 +179,22 @@ void batadv_mcast_mla_tt_update(struct batadv_priv *bat_priv)
 	 */
 	if (!atomic_read(&bat_priv->mcast_group_awareness) ||
 	    bat_priv->soft_iface->priv_flags & IFF_BRIDGE_PORT) {
-		if (enabled)
+		if (enabled) {
+			batadv_tvlv_container_unregister(bat_priv,
+							 BATADV_TVLV_MCAST, 1);
 			enabled = false;
+		}
 
 		goto update;
 	}
 
-	if (!enabled)
+	if (!enabled) {
+		mcast_flags = BATADV_MCAST_LISTENER_ANNOUNCEMENT;
+		batadv_tvlv_container_register(bat_priv, BATADV_TVLV_MCAST, 1,
+					       &mcast_flags,
+					       sizeof(mcast_flags));
 		enabled = true;
+	}
 
 	ret = batadv_mcast_mla_local_collect(soft_iface, &mcast_list);
 	if (ret < 0)
@@ -200,10 +209,65 @@ 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) &&
+	    (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_no_mla);
+	} else if (mcast_flags & BATADV_MCAST_LISTENER_ANNOUNCEMENT &&
+		   !(orig->mcast_flags & BATADV_MCAST_LISTENER_ANNOUNCEMENT)) {
+		atomic_dec(&bat_priv->mcast.num_no_mla);
+	}
+
+	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);
 }
+
+void batadv_mcast_purge_orig(struct batadv_orig_node *orig_node)
+{
+	if (!(orig_node->mcast_flags & BATADV_MCAST_LISTENER_ANNOUNCEMENT))
+		atomic_dec(&orig_node->bat_priv->mcast.num_no_mla);
+}
diff --git a/multicast.h b/multicast.h
index 3b68e3b..aa58e4b 100644
--- a/multicast.h
+++ b/multicast.h
@@ -24,8 +24,12 @@
 
 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);
 
+void batadv_mcast_purge_orig(struct batadv_orig_node *orig_node);
+
 #else
 
 static inline void batadv_mcast_mla_tt_update(struct batadv_priv *bat_priv)
@@ -33,11 +37,21 @@ 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;
 }
 
+static inline void batadv_mcast_purge_orig(struct batadv_orig_node *orig_node)
+{
+	return;
+}
+
 #endif /* CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS */
 
 #endif /* _NET_BATMAN_ADV_MULTICAST_H_ */
diff --git a/originator.c b/originator.c
index a591dc5..5e827fa 100644
--- a/originator.c
+++ b/originator.c
@@ -29,6 +29,7 @@
 #include "bridge_loop_avoidance.h"
 #include "network-coding.h"
 #include "fragmentation.h"
+#include "multicast.h"
 
 /* hash class keys */
 static struct lock_class_key batadv_orig_hash_lock_class_key;
@@ -143,6 +144,8 @@ static void batadv_orig_node_free_rcu(struct rcu_head *rcu)
 
 	spin_unlock_bh(&orig_node->neigh_list_lock);
 
+	batadv_mcast_purge_orig(orig_node);
+
 	/* Free nc_nodes */
 	batadv_nc_purge_orig(orig_node->bat_priv, orig_node, NULL);
 
@@ -258,6 +261,9 @@ 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);
 
diff --git a/packet.h b/packet.h
index e3ae41a..0d616f9 100644
--- a/packet.h
+++ b/packet.h
@@ -91,6 +91,11 @@ enum batadv_icmp_packettype {
 	BATADV_PARAMETER_PROBLEM       = 12,
 };
 
+/* multicast capabilities */
+enum batadv_mcast_flags {
+	BATADV_MCAST_LISTENER_ANNOUNCEMENT = BIT(0),
+};
+
 /* tt data subtypes */
 #define BATADV_TT_DATA_TYPE_MASK 0x0F
 
@@ -137,6 +142,7 @@ 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,
@@ -144,6 +150,7 @@ 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
diff --git a/soft-interface.c b/soft-interface.c
index 52244b0..bc0cc35 100644
--- a/soft-interface.c
+++ b/soft-interface.c
@@ -458,6 +458,7 @@ 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_no_mla, 0);
 #endif
 	atomic_set(&bat_priv->ap_isolation, 0);
 	atomic_set(&bat_priv->gw_mode, BATADV_GW_MODE_OFF);
diff --git a/types.h b/types.h
index 8175424..0586693 100644
--- a/types.h
+++ b/types.h
@@ -163,6 +163,9 @@ 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;
@@ -504,6 +507,7 @@ struct batadv_priv_dat {
 #ifdef CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS
 struct batadv_priv_mcast {
 	struct list_head mla_list;
+	atomic_t num_no_mla;
 };
 #endif
 
-- 
1.7.10.4

[B.A.T.M.A.N.] [PATCHv4 3/3] batman-adv: Modified forwarding behaviour for multicast packets

[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.

Signed-off-by: Linus Lüssing <linus.luessing@web.de>
---
 multicast.c         |   49 ++++++++++++++++++++++++++++++++++++++
 multicast.h         |   23 ++++++++++++++++++
 soft-interface.c    |   11 +++++++++
 translation-table.c |   66 +++++++++++++++++++++++++++++++++++++++++++++++++++
 translation-table.h |    1 +
 5 files changed, 150 insertions(+)

diff --git a/multicast.c b/multicast.c
index 6375347..e4bcdb2 100644
--- a/multicast.c
+++ b/multicast.c
@@ -21,6 +21,7 @@
 #include "originator.h"
 #include "hard-interface.h"
 #include "translation-table.h"
+#include "multicast.h"
 
 struct batadv_hw_addr {
 	struct list_head	list;
@@ -209,6 +210,54 @@ out:
 }
 
 /**
+ * batadv_mcast_forw_mode - 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 the forwarding mode as enum batadv_forw_mode.
+ */
+enum batadv_forw_mode
+batadv_mcast_forw_mode(struct sk_buff *skb, struct batadv_priv *bat_priv)
+{
+	struct ethhdr *ethhdr = (struct ethhdr *)(skb->data);
+	struct ipv6hdr *ip6hdr;
+	int count;
+
+	if (!atomic_read(&bat_priv->mcast_group_awareness))
+		return BATADV_FORW_ALL;
+
+	if (atomic_read(&bat_priv->mcast.num_no_mla))
+		return BATADV_FORW_ALL;
+
+	if (ntohs(ethhdr->h_proto) != ETH_P_IPV6)
+		return BATADV_FORW_ALL;
+
+	/* We might fail due to out-of-memory -> drop it */
+	if (!pskb_may_pull(skb, sizeof(*ethhdr) + sizeof(*ip6hdr)))
+		return BATADV_FORW_NONE;
+
+	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)
+		return BATADV_FORW_ALL;
+
+	count = batadv_tt_global_hash_count(bat_priv, ethhdr->h_dest);
+
+	switch (count) {
+	case 0:
+		return BATADV_FORW_NONE;
+	case 1:
+		return BATADV_FORW_SINGLE;
+	default:
+		return BATADV_FORW_ALL;
+	}
+}
+
+/**
  * 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
diff --git a/multicast.h b/multicast.h
index aa58e4b..7b291d5 100644
--- a/multicast.h
+++ b/multicast.h
@@ -20,10 +20,27 @@
 #ifndef _NET_BATMAN_ADV_MULTICAST_H_
 #define _NET_BATMAN_ADV_MULTICAST_H_
 
+/**
+ * batadv_forw_mode - the way a packet should be forwarded as
+ * @BATADV_FORW_ALL: forward the packet to all nodes
+ *  (currently via classic flooding)
+ * @BATADV_FORW_SINGLE: forward the packet to a single node
+ *  (currently via the BATMAN_IV unicast routing protocol)
+ * @BATADV_FORW_NONE: don't forward, drop it
+ */
+enum batadv_forw_mode {
+	BATADV_FORW_ALL,
+	BATADV_FORW_SINGLE,
+	BATADV_FORW_NONE,
+};
+
 #ifdef CONFIG_BATMAN_ADV_MCAST_OPTIMIZATIONS
 
 void batadv_mcast_mla_tt_update(struct batadv_priv *bat_priv);
 
+enum batadv_forw_mode
+batadv_mcast_forw_mode(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);
@@ -37,6 +54,12 @@ static inline void batadv_mcast_mla_tt_update(struct batadv_priv *bat_priv)
 	return;
 }
 
+static inline enum batadv_forw_mode
+batadv_mcast_forw_mode(struct sk_buff *skb, struct batadv_priv *bat_priv)
+{
+	return BATADV_FORW_ALL;
+}
+
 static inline int batadv_mcast_init(struct batadv_priv *bat_priv)
 {
 	return 0;
diff --git a/soft-interface.c b/soft-interface.c
index bc0cc35..7ff5a02 100644
--- a/soft-interface.c
+++ b/soft-interface.c
@@ -34,6 +34,7 @@
 #include <linux/ethtool.h>
 #include <linux/etherdevice.h>
 #include <linux/if_vlan.h>
+#include "multicast.h"
 #include "bridge_loop_avoidance.h"
 #include "network-coding.h"
 
@@ -156,6 +157,7 @@ static int batadv_interface_tx(struct sk_buff *skb,
 	bool do_bcast = false;
 	uint32_t seqno;
 	unsigned long brd_delay = 1;
+	enum batadv_forw_mode mode;
 
 	if (atomic_read(&bat_priv->mesh_state) != BATADV_MESH_ACTIVE)
 		goto dropped;
@@ -218,6 +220,15 @@ static int batadv_interface_tx(struct sk_buff *skb,
 		default:
 			break;
 		}
+
+		if (do_bcast && !is_broadcast_ether_addr(ethhdr->h_dest)) {
+			mode = batadv_mcast_forw_mode(skb, bat_priv);
+			if (mode == BATADV_FORW_NONE)
+				goto dropped;
+
+			if (mode == BATADV_FORW_SINGLE)
+				do_bcast = false;
+		}
 	}
 
 	/* ethernet packet should be broadcasted */
diff --git a/translation-table.c b/translation-table.c
index 0be7b15..d14ee64 100644
--- a/translation-table.c
+++ b/translation-table.c
@@ -83,6 +83,72 @@ batadv_tt_hash_find(struct batadv_hashtable *hash, const void *data)
 	return tt_common_entry_tmp;
 }
 
+/**
+ * batadv_tt_orig_entries_count - count the number of originators
+ * @head: a list of originators
+ *
+ * Return the number of originator entries in the given list.
+ *
+ * The caller needs to hold the rcu_read_lock().
+ */
+static int batadv_tt_orig_entries_count(struct hlist_head *head)
+{
+	struct batadv_tt_orig_list_entry *orig_entry;
+	int count = 0;
+
+	hlist_for_each_entry_rcu(orig_entry, head, list) {
+		if (!atomic_read(&orig_entry->refcount))
+			continue;
+
+		count++;
+	}
+
+	return count;
+}
+
+/**
+ * batadv_tt_global_hash_count - count the number of orig entries
+ * @hash: hash table containing the tt entries
+ * @data: the data to count entries for
+ *
+ * Return the number of originators advertising the given address/data
+ * (excluding ourself).
+ */
+int batadv_tt_global_hash_count(struct batadv_priv *bat_priv, const void *data)
+{
+	struct hlist_head *head, *orig_list;
+	struct batadv_tt_common_entry *tt_common_entry;
+	struct batadv_tt_global_entry *tt_global_entry;
+	uint32_t index;
+	int count = 0;
+
+	if (!bat_priv->tt.global_hash)
+		goto out;
+
+	index = batadv_choose_orig(data, bat_priv->tt.global_hash->size);
+	head = &bat_priv->tt.global_hash->table[index];
+
+	rcu_read_lock();
+	hlist_for_each_entry_rcu(tt_common_entry, head, hash_entry) {
+		if (!batadv_compare_eth(tt_common_entry, data))
+			continue;
+
+		if (!atomic_read(&tt_common_entry->refcount))
+			continue;
+
+		tt_global_entry = container_of(tt_common_entry,
+					       struct batadv_tt_global_entry,
+					       common);
+		orig_list = &tt_global_entry->orig_list;
+		count = batadv_tt_orig_entries_count(orig_list);
+		break;
+	}
+	rcu_read_unlock();
+
+out:
+	return count;
+}
+
 static struct batadv_tt_local_entry *
 batadv_tt_local_hash_find(struct batadv_priv *bat_priv, const void *data)
 {
diff --git a/translation-table.h b/translation-table.h
index 015d8b9..5986c57 100644
--- a/translation-table.h
+++ b/translation-table.h
@@ -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);
-- 
1.7.10.4

Re: [B.A.T.M.A.N.] Basic Multicast Optimizations

[Simon Wunderlich]

[-- Attachment #1: Type: text/plain, Size: 1115 bytes --]

Hello Linus,

I gave it a try - but there seems something off. What I did is:
 * apply your patches on 3b38a80 - multicast was enabled by default
 * start 2 VMs directly connected
 * ping6 -I bat0 ff02::1

I only got a reply locally, but not from the peer. When I disabled
multicast, I got two replies from the local host and the peer.

As far as I have tracked the problem down, it appears that
batadv_mcast_forw_mode() returns 1 correctly, but the packet is dropped
later in the process - I would guess that this happens in
batadv_send_generic_unicast_skb() where we try to select the gateway
when the destination mac is multicast instead of looking it up in the
tt table. But I leave the details to you. :)

The rest looks good, but I'll review more deeply again when it's
actually working. It does compile without problems now.

Cheers,
	Simon

On Mon, Jun 10, 2013 at 09:11:55AM +0200, Linus Lüssing wrote:
> This is the fourth revision of the basic multicast optimization patches.
> 
> It adds a missing enum keyword to the return type of a mcast_forw_mode().
> 
> Cheers, Linus

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [B.A.T.M.A.N.] Basic Multicast Optimizations

[Linus Lüssing]

On Wed, Jun 12, 2013 at 12:14:11PM +0200, Simon Wunderlich wrote:
> Hello Linus,
> 
> I gave it a try - but there seems something off. What I did is:
>  * apply your patches on 3b38a80 - multicast was enabled by default
>  * start 2 VMs directly connected
>  * ping6 -I bat0 ff02::1
> 
> I only got a reply locally, but not from the peer. When I disabled
> multicast, I got two replies from the local host and the peer.
> 
> As far as I have tracked the problem down, it appears that
> batadv_mcast_forw_mode() returns 1 correctly, but the packet is dropped

Ok, found a bug - the return 1 is actually incorrect. You've
chosen the one multicast address where no optimization is
conceptually possible. ff02::1 is the only link-local IPv6
multicast address which should return a BATADV_FORW_ALL (0)
instead.

(Conceptually impossible because for ff02::1 is the one multicast
address which by the IPv6 standard every host listens to, without
performing any MLD.)

I'll add another check next to the scope check in
mcast_forw_mode().

> later in the process - I would guess that this happens in
> batadv_send_generic_unicast_skb() where we try to select the gateway
> when the destination mac is multicast instead of looking it up in the
> tt table. But I leave the details to you. :)

Hm, but still, you're right, with this single destination, the
other VM, the ICMPv6 request and reply should have arrived because
the according MAC (33:33:00:00:00:01) is in the global translation table.

I didn't have that issue in my tests so far, I'll try to reproduce
that issue.

> 
> The rest looks good, but I'll review more deeply again when it's
> actually working. It does compile without problems now.
> 
> Cheers,
> 	Simon
> 

Thanks for the feedback again!

Cheers, Linus

Re: [B.A.T.M.A.N.] Basic Multicast Optimizations

[Simon Wunderlich]

[-- Attachment #1: Type: text/plain, Size: 2002 bytes --]

On Wed, Jun 12, 2013 at 02:27:14PM +0200, Linus Lüssing wrote:
> On Wed, Jun 12, 2013 at 12:14:11PM +0200, Simon Wunderlich wrote:
> > Hello Linus,
> > 
> > I gave it a try - but there seems something off. What I did is:
> >  * apply your patches on 3b38a80 - multicast was enabled by default
> >  * start 2 VMs directly connected
> >  * ping6 -I bat0 ff02::1
> > 
> > I only got a reply locally, but not from the peer. When I disabled
> > multicast, I got two replies from the local host and the peer.
> > 
> > As far as I have tracked the problem down, it appears that
> > batadv_mcast_forw_mode() returns 1 correctly, but the packet is dropped
> 
> Ok, found a bug - the return 1 is actually incorrect. You've
> chosen the one multicast address where no optimization is
> conceptually possible. ff02::1 is the only link-local IPv6
> multicast address which should return a BATADV_FORW_ALL (0)
> instead.

Ah, wonderful, and I thought I found a testcase. :D

Can you advise how the feature can be tested then, practically?
I guess I'll need to add some routes and ping6 another address?

> 
> (Conceptually impossible because for ff02::1 is the one multicast
> address which by the IPv6 standard every host listens to, without
> performing any MLD.)
> 
> I'll add another check next to the scope check in
> mcast_forw_mode().
> 
> > later in the process - I would guess that this happens in
> > batadv_send_generic_unicast_skb() where we try to select the gateway
> > when the destination mac is multicast instead of looking it up in the
> > tt table. But I leave the details to you. :)
> 
> Hm, but still, you're right, with this single destination, the
> other VM, the ICMPv6 request and reply should have arrived because
> the according MAC (33:33:00:00:00:01) is in the global translation table.

Yeah, it was, that is what puzzled me.
> 
> I didn't have that issue in my tests so far, I'll try to reproduce
> that issue.

OK.

Thanks,
	Simon

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [B.A.T.M.A.N.] Basic Multicast Optimizations

[Linus Lüssing]

On Wed, Jun 12, 2013 at 02:44:38PM +0200, Simon Wunderlich wrote:
> On Wed, Jun 12, 2013 at 02:27:14PM +0200, Linus Lüssing wrote:
> > On Wed, Jun 12, 2013 at 12:14:11PM +0200, Simon Wunderlich wrote:
> > > Hello Linus,
> > > 
> > > I gave it a try - but there seems something off. What I did is:
> > >  * apply your patches on 3b38a80 - multicast was enabled by default
> > >  * start 2 VMs directly connected
> > >  * ping6 -I bat0 ff02::1
> > > 
> > > I only got a reply locally, but not from the peer. When I disabled
> > > multicast, I got two replies from the local host and the peer.
> > > 
> > > As far as I have tracked the problem down, it appears that
> > > batadv_mcast_forw_mode() returns 1 correctly, but the packet is dropped
> > 
> > Ok, found a bug - the return 1 is actually incorrect. You've
> > chosen the one multicast address where no optimization is
> > conceptually possible. ff02::1 is the only link-local IPv6
> > multicast address which should return a BATADV_FORW_ALL (0)
> > instead.
> 
> Ah, wonderful, and I thought I found a testcase. :D

Well, it was a testcase I hadn't used and you found a bug by that
;).

> 
> Can you advise how the feature can be tested then, practically?
> I guess I'll need to add some routes and ping6 another address?

I was usually using 'ip maddr show' to check which address isn't
in use yet, then I use something like 'ping6 ff02::3%bat0' which should
by default result in no reply.

Then I'd start adding listeners with this tiny C code snippet:
http://pastebin.com/fg9z8z5b on various nodes
(because I didn't find any simple, tiny command line program to do
just that yet).

You'd then get ICMP replies from the according nodes and I checked
via tcpdump that the ICMP packets were forwarded correctly via
unicast or broadcast.


And since "batman-adv: Add dummy soft-interface rx mode handler"
I'm not using that C code snippet that much anymore, but just
adding listeners via "ip maddr add 33:33:00:00:00:03 dev bat0"
for instance. And just checking via tcpdump (since you won't get
any ICMP replies) and the 'batctl tg' output.

[B.A.T.M.A.N.] Basic Multicast Optimizations

[Linus Lüssing]

This is the twelfth revision of the basic multicast optimization patches.

Changes in v12 include (thanks for the feedback again, Marek and Simon):

* added three reserved bytes to the mcast TVLV container to have a four bytes
  alignment (thanks Simon)
* Return value description in kernel docs for batadv_mcast_mla_tvlv_update
  added (thanks Marek)
* Using eth_hdr() instead of skb->data in various places (thanks Marek)
* Switched order of function parameters bat_priv and skb in various places
  for style consistency reasons (thanks Marek)
* renamed one occurence of "BATMAN_IV" to "BATMAN" (thanks Marek)
* kerneldoc for multicast_mode added (thanks Marek)
* rebased on top of master
* now that batadv_tt_local_add() has a potential failure case, a check for
  its return value was added

For Marek's question on gateway code interoperability, I reread that part
and couldn't find any issues with it either. Moreover, any additional
dhcp_rcp variable check seemed redundant to me, so I left that part as is.

Cheers, Linus

[B.A.T.M.A.N.] Basic Multicast Optimizations

[Linus Lüssing]

This is the eleventh revision of the basic multicast optimization patches.

Changes:

* two non-critical compat code fixes (fixes sparse/smatch warnings):
  * typo: 'struct dev_addr_list' vs. 'struct batadv_dev_addr_list'
  * (unreachable code with netdev_master_upper_dev_get_rcu() compat code,
     patch provided by Simon)
* included multicast.h in multicast.c (fixes an sparse warning)
* substituted BATADV_UNINIT_FLAGS with a orig->capa_initialized and
  bat_priv->mcast.enabled as suggested by Marek
  (this adds patch 2/6)
* removed batadv_mcast_counter_update() to increase readability as
  suggested and provided by Marek
* rebased on top of the latest gateway code modifications:
  "batman-adv: send every DHCP packet as bat-unicast"
  (2d5b555644b22e171f3e0486abc0c2e627452d14)
  

[B.A.T.M.A.N.] Basic Multicast Optimizations

[Linus Lüssing]

This is the tenth revision of the basic multicast optimization patches.

No changes were made except dealing with two or three rebase conflicts
while rebasing on top of 24c51de20.

Cheers, Linus

Re: [B.A.T.M.A.N.] Basic Multicast Optimizations

[Simon Wunderlich]

[-- Attachment #1: Type: text/plain, Size: 421 bytes --]

I think this looks good now! Thanks :)

Reviewed/Tested-by: Simon Wunderlich <siwu@hrz.tu-chemnitz.de>

Cheers,
	Simon

On Thu, Aug 15, 2013 at 09:21:14PM +0200, Linus Lüssing wrote:
> This is the ninth revision of the basic multicast optimization patches.
> 
> It fixes a regression in Patch 3/5, introduced in v8 while rebasing:
> An atomic_dec() for the orig_list_count got lost.
> 
> Cheers, Linus
> 

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

[B.A.T.M.A.N.] Basic Multicast Optimizations

[Linus Lüssing]

This is the ninth revision of the basic multicast optimization patches.

It fixes a regression in Patch 3/5, introduced in v8 while rebasing:
An atomic_dec() for the orig_list_count got lost.

Cheers, Linus

Re: [B.A.T.M.A.N.] Basic Multicast Optimizations

[Linus Lüssing]

On Thu, Aug 15, 2013 at 03:56:47PM +0200, Simon Wunderlich wrote:
> Hey Linus,
> 
> I've tested the patchset, and it looks good from my testcases (testing
> multicast to IPv4 and IPv6 addresses). I've not tested the MLD stuff
> though, but from our discussions your implementation looks reasonable.
> Did you get to test it somehow?

Yes, I was testing it with the following patch from the branch
linus/multicast-bridge:

"debug: batman-adv: manual querier flag sysfs switch" (e947548bf)


> 
> From my perspective, if you fix the orig list stuff we could integrate
> your patchset. I'd like to ask others to review the patchset too. :)
> 
> Thanks,
> 	Simon

Re: [B.A.T.M.A.N.] Basic Multicast Optimizations

[Simon Wunderlich]

[-- Attachment #1: Type: text/plain, Size: 424 bytes --]

Hey Linus,

I've tested the patchset, and it looks good from my testcases (testing
multicast to IPv4 and IPv6 addresses). I've not tested the MLD stuff
though, but from our discussions your implementation looks reasonable.
Did you get to test it somehow?

From my perspective, if you fix the orig list stuff we could integrate
your patchset. I'd like to ask others to review the patchset too. :)

Thanks,
	Simon

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

[B.A.T.M.A.N.] Basic Multicast Optimizations

[Linus Lüssing]

This is the eighth revision of the basic multicast optimization patches.

It adds many style + functional changes and fixes some bugs.

It also adds a new, important fifth patch, which addresses an issue
that I would have overlooked if it weren't for the extensive offline
talks with Simon about the interaction of batman-adv and various
multicast related RFCs.


# Changelog v7 -> v8

## Style improvements

* renamed batadv_mcast_tt_clean() tobatadv_mcast_tt_retract()
* renamed batadv_mcast_tvlv_update_flag_counter() to batadv_mcast_update_counter()
* renamed batadv_mcast_mla_tt_update() to batadv_mcast_mla_update()
* renamed batadv_mcast_update_counter() to batadv_mcast_counter_update()

* renamed BATADV_MCAST_HAS_NO_BRIDGE to BATADV_MCAST_WANT_ALL_UNSNOOPABLES (and reverse its logic)
* change value of BATADV_UNINIT_FLAGS from -1 to an unused bit, BIT(15)
  -> no more need to check BATADV_UNINIT_FLAGS in batadv_mcast_counter_update()

* moved code from batadv_mcast_mla_tt_update() to newly introduced batadv_mcast_mla_tvlv_update
* moved introduction of multicast_mode sysfs file to the third patch
* added kernel doc for struct batadv_priv_mcast and enum batadv_mcast_flag

## Functional improvements

* changed a '== -EINVAL' to '< 0'
* check for a bridge anywhere on top of bat0 (like br0 on bond0 on bat0, not just br0 on bat0)
  (as discussed on IRC we don't need to check for bridges on top of vlans yet as multicast
  optimizations are only enabled for non-VLAN packets at the moment)
* removed sysfs file control over multicast TT+TVLV announcements, moved sysfs file introduction to the third patch
* removed BATADV_MCAST_LISTENER_ANNOUNCEMENT flag: its purpose is now provided by checking the existence of the multicast tvlv instead
* made the tvlv value length check in the multicast tvlv handler to not only accept a length of one byte but anything greater, too
* New Patch: batman-adv: Send multicast packets to nodes with a WANT_ALL flag

## Bugs

* converted the static 'enabled' variable in batadv_mcast_mla_tt_update() to a per bat_priv variable, using bat_priv->mcast.flags (fixes a bug when using multiple batX interfaces)
* the announced tvlv was missing the BATADV_MCAST_HAS_NO_BRIDGE flag (not needed anymore with the reversed logic)
* batadv_mcast_update_counter() was missing a case which would have needed a BATADV_UNINIT_FLAGS check (not needed anymore with reversed logic and new value for BATADV_UNINIT_FLAGS)
* added missing update for num_want_all_unsnoopables (former num_has_bridge) counter in batadv_mcast_purge_orig()


Cheers, Linus

Re: [B.A.T.M.A.N.] Basic Multicast Optimizations

[Linus Lüssing]

And the sysfs file 'mcast_group_awareness' got renamed to
'multicast_mode'. The missing ABI documentation for this file was
added, too.

On Thu, Jul 04, 2013 at 12:03:18AM +0200, Linus Lüssing wrote:
> This is the seventh revision of the basic multicast optimization patches.
> 
> It adds the style changes as suggested by Marek via eMail.
> 
> It also adds various suggestion from IRC (for instance adding kerneldoc for the
> functions touched in translation-table.c, making the counter an atomic
> one to be on the safe side for some "exotic" targets).
> 
> 
> PATCHv7 4/4 is a new one and the result of a longer discussion on IRC:
> Whether to add some support for some IPv4 addresses already for the cost of
> a slightly larger complexity both code and documentation wise
> - for a case which is uncommon for most mesh networks anyway: a mesh
> with no bridges on any node involved.
> 
> So the fourth patch is sort of an RFC patch one, too (although the consense/
> tendency on IRC so far was to add and support this case).
> 
> If everyone is fine with that concept, then I can also easily merge it
> into patches 2/4 and 3/4 (if that is desired).
> 
> Cheers, Linus

[B.A.T.M.A.N.] Basic Multicast Optimizations

[Linus Lüssing]

This is the seventh revision of the basic multicast optimization patches.

It adds the style changes as suggested by Marek via eMail.

It also adds various suggestion from IRC (for instance adding kerneldoc for the
functions touched in translation-table.c, making the counter an atomic
one to be on the safe side for some "exotic" targets).


PATCHv7 4/4 is a new one and the result of a longer discussion on IRC:
Whether to add some support for some IPv4 addresses already for the cost of
a slightly larger complexity both code and documentation wise
- for a case which is uncommon for most mesh networks anyway: a mesh
with no bridges on any node involved.

So the fourth patch is sort of an RFC patch one, too (although the consense/
tendency on IRC so far was to add and support this case).

If everyone is fine with that concept, then I can also easily merge it
into patches 2/4 and 3/4 (if that is desired).

Cheers, Linus

Re: [B.A.T.M.A.N.] Basic Multicast Optimizations

[Simon Wunderlich]

[-- Attachment #1: Type: text/plain, Size: 520 bytes --]

Looks good, Tested/Acked-by: Simon Wunderlich <siwu@hrz.tu-chemnitz.de>

Thanks,
	Simon

On Fri, Jun 14, 2013 at 07:50:06PM +0200, Linus Lüssing wrote:
> This is the sixth revision of the basic multicast optimization patches.
> 
> It simplifies and speeds-up the new batadv_tt_global_hash_count():
> 
> * We now use batadv_tt_global_hash_find() instead of our custom look-up.
> * A counter for the orig_list was added, so that we won't need to recount
>   with every multicast packet.
> 
> Cheers, Linus

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

[B.A.T.M.A.N.] Basic Multicast Optimizations

[Linus Lüssing]

This is the sixth revision of the basic multicast optimization patches.

It simplifies and speeds-up the new batadv_tt_global_hash_count():

* We now use batadv_tt_global_hash_find() instead of our custom look-up.
* A counter for the orig_list was added, so that we won't need to recount
  with every multicast packet.

Cheers, Linus

[B.A.T.M.A.N.] Basic Multicast Optimizations

[Linus Lüssing]

This is the fifth revision of the basic multicast optimization patches.

It fixes two bugs:

For one thing packets to the IPv6 all-hosts-link-scope multicast address
(ff02::1) were optimized, too. However this is the one link local address
which we should not never optimize (see RFC4541, section 3 [1]).

For another since commit 87d4a282 ("batman-adv: remove useless find_router look up"
the unicast-forwarding of multicast packets was broken and I had
only runtime tested the patches based on an older commit
(before 87d4a282 such unicast-forwarding was "only" broken if
a gateway-server was present).


Also note that rebasing on top of the TT-VLAN patches wasn't trivial:
Simply resolving the rebase complaints resulted in subtle runtime
bugs. A few lines in the counting methods added to translation-table.c
needed changes.

Cheers, Linus


[1]: Considerations for Internet Group Management Protocol (IGMP)
     and Multicast Listener Discovery (MLD) Snooping Switches
     -> "Packets with the all hosts link-scope address should be forwarded
         on all ports."
     https://tools.ietf.org/html/rfc4541

Re: [B.A.T.M.A.N.] Basic Multicast Optimizations

[Linus Lüssing]

Hrm, forgot to add the v3 an the patch naming. And forgot to
actually add
-static inline batadv_forw_mode
+static inline enum batadv_forw_mode
to the according commit.

Will send a v4 in a sec.

Cheers, Linus

On Mon, Jun 10, 2013 at 08:28:18AM +0200, Linus Lüssing wrote:
> This is the third revision of the basic multicast optimization patches.
> It includes one functional and some compat fixes, and some style improvements,
> thanks to Simons feedback:
> 
> PATCHv3 1/3:
> * the pmc_rcu functions got removed - they weren't actually used in the
>   submitted revision - so there's no more patch for net to export
>   for_each_pmc_rcu() needed anymore. Should also fix Simon's compat error.
> * Compat code for netdev_for_each_mc_addr() added.
>   (I needed to introduce a batadv_hw_addr because I wasn't able to achieve
>   backwards compatibility with a netdev_hw_addr - my "official reasoning" will
>   be that netdev_hw_addr is unnecessarily bloated, struct batadv_hw_addr
>   has just the right size to do the job)
> * Compat fix for IFF_BRIDGE_PORT
> * Removed primary_if->soft_iface in batadv_mcast_mla_tt_update() as we
>   can always use bat_priv->soft_iface instead.
> 
> PATCHv3 2/3
> * Renamed num_non_aware to a more specific num_no_mla - because we will need
>   things like a num_no_tracker for instance with the multicast tracker feature
>   in the future.
> * Fixed removal of orig_node's without BATADV_MCAST_LISTENER_ANNOUNCEMENT
>   flag, update num_no_mla properly now.
> 
> PATCHv3 3/3
> * Renamed batadv_mcast_flood() to batadv_mcast_forw_mode() because it does not
>   only show whether to flood or not (=drop), but also whether to forward via
>   unicast. In the future with the multicast tracker feature a fourth return value
>   is going to be added, making the name "batadv_mcast_flood()" even less fitting.
> * Simon's style suggestions for batadv_mcast_forw_mode() and
>   batadv_interface_tx().
> * batadv_mcast_forw_mode() now returns an enum (thanks to Marek and Antonio for
>   the suggestion).
> 
> I did not change:
> > I don't quite understand why you return -1, maybe the packet could still
> > be forwarded even if it could not be pulled?
> 
> Because if it returns -1 then something is wrong, for instance we could be
> out of memory. If we are out of memory then we probably won't be able to
> forward any packet. Also we shouldn't allocate any more memory for one thing
> but I think it is also better to drop/free packets to save some memory
> to increase the possibility of the system to recover without crashing.
> 
> Cheers, Linus
> 

[B.A.T.M.A.N.] Basic Multicast Optimizations

[Linus Lüssing]

This is the third revision of the basic multicast optimization patches.
It includes one functional and some compat fixes, and some style improvements,
thanks to Simons feedback:

PATCHv3 1/3:
* the pmc_rcu functions got removed - they weren't actually used in the
  submitted revision - so there's no more patch for net to export
  for_each_pmc_rcu() needed anymore. Should also fix Simon's compat error.
* Compat code for netdev_for_each_mc_addr() added.
  (I needed to introduce a batadv_hw_addr because I wasn't able to achieve
  backwards compatibility with a netdev_hw_addr - my "official reasoning" will
  be that netdev_hw_addr is unnecessarily bloated, struct batadv_hw_addr
  has just the right size to do the job)
* Compat fix for IFF_BRIDGE_PORT
* Removed primary_if->soft_iface in batadv_mcast_mla_tt_update() as we
  can always use bat_priv->soft_iface instead.

PATCHv3 2/3
* Renamed num_non_aware to a more specific num_no_mla - because we will need
  things like a num_no_tracker for instance with the multicast tracker feature
  in the future.
* Fixed removal of orig_node's without BATADV_MCAST_LISTENER_ANNOUNCEMENT
  flag, update num_no_mla properly now.

PATCHv3 3/3
* Renamed batadv_mcast_flood() to batadv_mcast_forw_mode() because it does not
  only show whether to flood or not (=drop), but also whether to forward via
  unicast. In the future with the multicast tracker feature a fourth return value
  is going to be added, making the name "batadv_mcast_flood()" even less fitting.
* Simon's style suggestions for batadv_mcast_forw_mode() and
  batadv_interface_tx().
* batadv_mcast_forw_mode() now returns an enum (thanks to Marek and Antonio for
  the suggestion).

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

Because if it returns -1 then something is wrong, for instance we could be
out of memory. If we are out of memory then we probably won't be able to
forward any packet. Also we shouldn't allocate any more memory for one thing
but I think it is also better to drop/free packets to save some memory
to increase the possibility of the system to recover without crashing.

Cheers, Linus

Re: [B.A.T.M.A.N.] Basic Multicast Optimizations

[Marek Lindner]

Hi,

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

Re: [B.A.T.M.A.N.] Basic Multicast Optimizations

[Antonio Quartulli]

[-- Attachment #1: Type: text/plain, Size: 1590 bytes --]

On Fri, May 24, 2013 at 11:00:50AM +0200, Linus Lüssing wrote:
> On Fri, May 24, 2013 at 10:02:25AM +0200, Linus Lüssing wrote:
> > 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.

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,


-- 
Antonio Quartulli

..each of us alone is worth nothing..
Ernesto "Che" Guevara

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [B.A.T.M.A.N.] Basic Multicast Optimizations

[Linus Lüssing]

On Fri, May 24, 2013 at 10:02:25AM +0200, Linus Lüssing wrote:
> 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.

> 
> 
> 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

[B.A.T.M.A.N.] Basic Multicast Optimizations

[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.


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

Re: [B.A.T.M.A.N.] Basic Multicast Optimizations

[Simon Wunderlich]

[-- Attachment #1: Type: text/plain, Size: 10613 bytes --]

Hey Linus,

On Fri, May 17, 2013 at 03:38:56AM +0200, Linus Lüssing wrote:
> > 
> > For the non-link-local IPv4 multicast addresses, could we use the
> > current mechanism (drop/unicast/broadcast) by getting the assigned
> > multicast addresses? (of course, assuming we have no bridge)
> 
> We could - after implementing MRD. The thing is we might have one
> or more IPv4 multicast routers in our network. If we are having
> two multicast listeners on our link, then everything will be fine,
> we will broadcast, multicast routers receive the multicast
> packets, too and the multicast data gets routed further. However
> with no MRD and if there is just one or no multicast listener
> then we'd do the unicast/drop and the multicast router on the link
> and any listener more than one hop away would not receive that data
> anymore.
> 
OK, I see. Well, then we should save that for later ...

> > > For MRD I was thinking about taking a short-cut for now as
> > > implementing would need a few more lines of code. Instead I think
> > > I go for detecting whether a querier is on the link both in
> > > batman-adv and the bridge code for IPv4 or IPv6 and if not disabling
> > > optimizations accordingly and issuing a warning. That way
> > > non-link-local IPv6 multicast traffic could be optimized already
> > > for instance if running an mrd6 instance in userspace which
> > > already performs both MLD querying and MRD.
> > > 
> 
> Sorry, I ment: "For MLD querying I was...". mrd6 does perform MRD,
> so sending things like multicast router advertisements. But
> bridges or batman-adv would still need to parse these
> advertisements and should send multicast router solicitations
> (e.g. when an interface comes up) to be able to quickly determine
> which bridge ports or nodes have multicast routers and want the
> non-link-local multicast traffic.
> 

OK

> > 
> > Hmm, sorry I still get a little confused in that MRD/MLD/IGMP terminology.
> > So as I understand there might be a userspace component which does the MLD/MRD
> > instead of the kernel?
> 
> Yes, the layer 3 multicast routing table is in the kernel, but it
> is configured from userspace from daemons like mrd6. And mrd6 also
> does some non-performance critical protocol stuff like the MLD
> Querier protocol and the MRD multicast router advertisements like
> I said above.
> 

OK

> > Also, is it required to get implement MRD/MLD specifically
> > in batman-adv or bridges, or would it be possible to use a "general" approach
> > which could be used for any interface to emit/exchange these kind of messages?
> 
> The MLD querier is something which only makes sense to me in case
> of multicast snooping bridges and I don't see any benefit in having
> it anywhere else. If there are just the plain interfaces with no
> multicast router then you don't need an MLD querier (and in fact
> no MLD reports from the listeners either). If there's just
> batman-adv with no bridges, then you don't need it either.
> Although...
> 
> There is one use-case where a more general approach might be of
> interest: Wireless interfaces - it might be of interest for the
> mac80211 to know whether there is a listener (or router) behind
> the link provided by your wlan0 device. The wifi driver could then
> decide to refrain from broadcasting a multicast packet or to use
> the (minimum) bitrate of just the(se) listener(s) (/router(s)).

Actually that might be interesting for both batman-adv and "regular"
WiFi setups. I think some commercial vendors already do exactly that
to optimize Multicast. My setup at home is a good example how this
could benefit: 
 * I have 6 APs having mesh and ap bridged
 * Right now, when I multicast stream from my LAN, first batman-adv
   floods all the broadcast through the mesh, then every AP sends the
   broadcast again - all on the lowest rate (at least on the AP).
 --> so if we listen to music using multicast only in the kitchen, 
   the WiFi becomes really slow. :D

With the multicast optimization in batman-adv, we could solve the
broadcast storm in the batman-adv network (only send to the APs where
listener registered). Then still the AP would broadcast these packets
on the lowest rate (1 MBit/s). Not all APs at least, but still the
one where the client is connected. If mac80211 could detect that
and send it via unicast, we could use even HT rates (e.g. 300 Mbit/s)
here. I could enjoy wireless multicast HD video streams - yay. ;)

Probably not the most typical example/use case, but still we can see
how this scenario would benefit from the optimization.

> Or maybe a drop/unicast/flood on an eth0 itself might be nice, too,
> especially if there is a large switch/hub connected to it.
> 
> Although that'd probably be awesome to have, I think it's easier
> to just have the MLD querier in the bridge code for now (especilly
> as there already is some MLD querier code in the bridge - though
> it is so incomplete that it got disabled in April 2012 because of
> causing issues).

Actually I'm not sure about the technical implementation of such
a general approach. Maybe it would be possible to query the bridge
from the various components as well.

> 
> Hm, for the MRD RA parsing and MRD RS I thought about implementing
> that both in the bridge and batman-adv code (it's about 300
> lines of C++ code in mrd6). Not sure how easy it'd be to implement
> a more general approach on top of a Linux netdev for instance.
> Would need to check that.
> 
> > 
> > Sorry if I ask stupid questions. :)
> > 
> > Maybe some example or architecture overview (could be put on the wiki page) would
> > help?
> 
> I guess you mean some visualizaton with things like simple
> devices, bridge, batman-adv, maybe various kinds of multicast
> types (e.g. scope == link-local vs. scope > link-local, IPv6 transient
> vs. IPv6 well-known) and the according RFCs? Hm, not quite sure
> how that could look like would need to think about it - or maybe
> you have some idea about how you think that could/should look
> like?
> 

Hmm, not sure either. Maybe we choose the typical examples (bat0 only,
bat0 + ap + bridge), and show which component sends what. Like having
"blocks" for batman, bridge, userspace mrd, etc, and show which component
sends/querys who for which information.

I know that's pretty vague, but maybe there is a way to bring light in
that for others which are not so experienced in that (me included).

Multicast types (IPv4, IPv6) and their mac-address and RFC recommendations
should probably better go in an extra section. We could make a table of
what types exist, how we handle them (optimize, just broadcast) and why
(RFC, design decisions).

> > OK. We could also increase the version number of the TVLV and interpret
> > both versions in newer multicast implementations, but if we already know
> > that we can define this information in the current implementation.
> 
> But a newer TVLV version would break compatibility for older
> batman-adv versions, they wouldn't recognize the newer TVLV
> version. Or a node would need to add TVLVs for both versions to
> its OGM. But that'd be some more overhead compared to just adding
> another flag.
> 

Yeah you are right. Maybe just add the flag. :D

> > 
> > > > 
> > > > Would you consider Antonios comments and update your patchset? I would like to
> > > > test it in the next days ...
> > > 
> > > Yes, I will. I had written some comments for the comments on IRC,
> > > but I guess it'll be better to write them here on the list again.
> > > 
> > 
> > Mailing list would be better, yes. Thanks. :)
> 
> Done :).
> 

Thanks!

> > 
> > > > 
> > > > Thanks,
> > > > 	Simon
> > > 
> > > Another thing I was thinking about conceptually yesterday was
> > > whether we should use more refined flags instead of just
> > > MULTICAST_LISTENER_ANNOUNCEMENTS for everything (non-link-local
> > > IPv4, all IPv6 except the all-nodes address).
> > > 
> > > That way we could for instance already add some cases for when to
> > > use the multicast optimizations when having a bridge, for
> > > instance:
> > > 
> > > When batman-adv detects that there is an MLD querier and if all
> > > nodes have a MULTICAST_LISTENER_ANNOUNCEMENTS or
> > > MLA_IPV6_TRANSIENT_LINK_LOCAL flag it could  already
> > > optimize link-local, transient IPv6 multicast traffic
> > > without needing to modify anything in the bridge code except the
> > > addition of the export which was already posted as an RFC on the
> > > bridge mailing list.
> > > 
> > > Hm, again will need to think about that. Whether the extra
> > > conceptual complexity is okay because of being able to add some
> > > more use-cases with less code in small chunks already.
> > 
> > Hmm, again I'm not completely sure to follow, but the idea here is
> > to enable functionality when having a userspace MLD instead of
> > the (planned) bridge MLD stuff?
> 
> Yes, that would be the idea.
> 
> > 
> > If we need the bridge MLD stuff anyway (to have the full feature set
> > etc), I'd rather not do more intermediate steps which might be
> > obsolete later. However, if the userspace MLD thing is equivalent
> > feature-wise than this might be interesting to do.
> 
> Hm, don't know. I was thinking that most setups I know of involve
> bridges on top of bat0 and to be able to make use of any multicast
> optimizations in batman-adv in case of such bridging we need an
> MLD querier on the link. Implementing a proper MLD querier might
> need several iterations over the bridge / netdev mailing lists and
> might therefore need quite some time. During that time those
> setups won't be able to make any use of these optimizations.
> 
> Unless taking the suggested short-cut which shouldn't be that
> difficult to implement, I think, then these bat0+bridge setups
> could already enjoy some multicast optimizations and such mesh
> networks might already be able to play with multicast streams as I
> think updating and merging the tracker packet patches will be done
> some time before getting MLD querier bridge code upstream.
> (but maybe I'm estimating all this wrong, dunno)

Actually I did not hear much screaming to have multicast support so far,
so I don't think we "need" an intermediate solution. If you want to it
anyway or need it for something, go ahead, but I'd prefer aiming for
the final implementation from the start without taking detours. :)

Thanks,
	Simon

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [B.A.T.M.A.N.] Basic Multicast Optimizations

[Linus Lüssing]

Hi,

On Thu, May 16, 2013 at 08:31:32PM +0200, Simon Wunderlich wrote:
> On Thu, May 16, 2013 at 07:42:00PM +0200, Linus Lüssing wrote:
> > Hi Simon,
> > 
> > On Thu, May 16, 2013 at 01:51:29PM +0200, Simon Wunderlich wrote:
> > > Hey Linus,
> > > 
> > > thanks for working and integrating this patchset!
> > > 
> > > On Sat, May 11, 2013 at 07:23:24PM +0200, Linus Lüssing wrote:
> > > > This set of patches is the first one for a more efficient, group aware
> > > > multicast forwarding infrastructure in batman-adv.
> > > > 
> > > > This initial set mainly consists of the integration of announcing the
> > > > location of multicast listeners via the translation table mechanism to
> > > > be able to find out which batman-adv nodes are actually interested in
> > > > certain multicast traffic. As well as the signalizing of this
> > > > announcement capability via a new multicast TVLV.
> > > > 
> > > > Finally some basic multicast forwarding opitimizations are introduced:
> > > > If all nodes signalized the MLA capability then link-local IPv6 traffic
> > > > will be dropped if there is no interested listener or gets forwarded
> > > > via a batman-adv unicast packet if there is just one node interested
> > > > in the data.
> > > 
> > > I've seen that you've only implemented that for IPv6, would it be possible
> > > to use it for IPv4 too in the current state? (I use some media streaming
> > > thingy at home where I could test ;])
> > 
> > Hmm, I need to think about it. The thing is it looks like IPv4 does
> > not seem to require IGMP. And according to RFC4541, an
> > informational one, link-local IPv4 multicast shoud be excluded as
> > there might be some applications not issuing IGMP reports.
> > 
> > For non-link-local addresses we'd need MRD first, as we'd
> > otherwise miss to forward things to multicast routers.
> > 
> > However when using no bridge, like it is the case now, RFC4541
> > does not apply and we can in fact reliably get and distribute
> > link-local IPv4 multicast addresses as we do not rely on IGMP
> > here... 
> > 
> > One option would be to introduce another flag. But that might
> > become confusing for users, it might not be easy to explain to
> > users why their 224.0.0.123 would be working on raw bat0 but not
> > when they add a bridge and why this would never work. Instead we
> > could wait for the MRD implementation and tell people that
> > 224.0.0.0/24 won't work.
> > 
> 
> Hmm, as far as I read the RFC 4541) (2.1.2, point 2) we should just
> treat 224.0.0.0/24 (which would translate to 01:00:5e:00:00:xx) as
> broadcast - we should do that in any state of the multicast implementation
> IMHO, and are then on the safe side. For all other IPv4 multicast
> addresses we can assume that IGMP is used.

Ok, yes, that's what I thought, too (although like I said before
technically there is the possibility to safely optimize 224.0.0.0/24, too
if there are no bridges involved - but ok, let's forget that thought).

> 
> For the non-link-local IPv4 multicast addresses, could we use the
> current mechanism (drop/unicast/broadcast) by getting the assigned
> multicast addresses? (of course, assuming we have no bridge)

We could - after implementing MRD. The thing is we might have one
or more IPv4 multicast routers in our network. If we are having
two multicast listeners on our link, then everything will be fine,
we will broadcast, multicast routers receive the multicast
packets, too and the multicast data gets routed further. However
with no MRD and if there is just one or no multicast listener
then we'd do the unicast/drop and the multicast router on the link
and any listener more than one hop away would not receive that data
anymore.

> > > 
> > > > 
> > > > For now, these optimizations only apply if all nodes in the mesh have
> > > > no bridge interface on top their bat0 interface. However making it
> > > > possible with bridges, too, and other features are on the roadmap. See
> > > > the according wiki page for details [1].
> > > 
> > > So from what I understand this means:
> > > 
> > >  * multicast support is working for non-bridged interfaces for now, because
> > >    we lack proper multicast router discovery support (MRD) in the bridge code
> > >  --> you told me someone is working on that, or you'll pick up to get this done?
> > 
> > Yes, I had been exchanging some emails with a group of people a
> > few months ago who wanted to implement MRD and a proper IGMP/MLD
> > querier protocol in the bridge code but didn't hear from them again.
> > So yes, I guess I'll pick up on that.
> 
> OK, cool.
> > 
> > 
> > For MRD I was thinking about taking a short-cut for now as
> > implementing would need a few more lines of code. Instead I think
> > I go for detecting whether a querier is on the link both in
> > batman-adv and the bridge code for IPv4 or IPv6 and if not disabling
> > optimizations accordingly and issuing a warning. That way
> > non-link-local IPv6 multicast traffic could be optimized already
> > for instance if running an mrd6 instance in userspace which
> > already performs both MLD querying and MRD.
> > 

Sorry, I ment: "For MLD querying I was...". mrd6 does perform MRD,
so sending things like multicast router advertisements. But
bridges or batman-adv would still need to parse these
advertisements and should send multicast router solicitations
(e.g. when an interface comes up) to be able to quickly determine
which bridge ports or nodes have multicast routers and want the
non-link-local multicast traffic.

> 
> Hmm, sorry I still get a little confused in that MRD/MLD/IGMP terminology.
> So as I understand there might be a userspace component which does the MLD/MRD
> instead of the kernel?

Yes, the layer 3 multicast routing table is in the kernel, but it
is configured from userspace from daemons like mrd6. And mrd6 also
does some non-performance critical protocol stuff like the MLD
Querier protocol and the MRD multicast router advertisements like
I said above.

> Also, is it required to get implement MRD/MLD specifically
> in batman-adv or bridges, or would it be possible to use a "general" approach
> which could be used for any interface to emit/exchange these kind of messages?

The MLD querier is something which only makes sense to me in case
of multicast snooping bridges and I don't see any benefit in having
it anywhere else. If there are just the plain interfaces with no
multicast router then you don't need an MLD querier (and in fact
no MLD reports from the listeners either). If there's just
batman-adv with no bridges, then you don't need it either.
Although...

There is one use-case where a more general approach might be of
interest: Wireless interfaces - it might be of interest for the
mac80211 to know whether there is a listener (or router) behind
the link provided by your wlan0 device. The wifi driver could then
decide to refrain from broadcasting a multicast packet or to use
the (minimum) bitrate of just the(se) listener(s) (/router(s)).
Or maybe a drop/unicast/flood on an eth0 itself might be nice, too,
especially if there is a large switch/hub connected to it.

Although that'd probably be awesome to have, I think it's easier
to just have the MLD querier in the bridge code for now (especilly
as there already is some MLD querier code in the bridge - though
it is so incomplete that it got disabled in April 2012 because of
causing issues).

Hm, for the MRD RA parsing and MRD RS I thought about implementing
that both in the bridge and batman-adv code (it's about 300
lines of C++ code in mrd6). Not sure how easy it'd be to implement
a more general approach on top of a Linux netdev for instance.
Would need to check that.

> 
> Sorry if I ask stupid questions. :)
> 
> Maybe some example or architecture overview (could be put on the wiki page) would
> help?

I guess you mean some visualizaton with things like simple
devices, bridge, batman-adv, maybe various kinds of multicast
types (e.g. scope == link-local vs. scope > link-local, IPv6 transient
vs. IPv6 well-known) and the according RFCs? Hm, not quite sure
how that could look like would need to think about it - or maybe
you have some idea about how you think that could/should look
like?

> 
> > > 
> > >  * the current optimization is to handle multicast like:
> > >    * no listener - drop
> > >    * one listener - unicast
> > >    * more listener - broadcast
> > >  --> I think that is a good and simple optimization. :) The "tracker" packets
> > >      can come later (as shown in your roadmap). This might require to announce
> > >      that tracker packets are supported in the TVLV?
> > 
> > Yes, we'll probably need to add another capability flag to the
> > multicast TVLV.
> > 
> 
> OK. We could also increase the version number of the TVLV and interpret
> both versions in newer multicast implementations, but if we already know
> that we can define this information in the current implementation.

But a newer TVLV version would break compatibility for older
batman-adv versions, they wouldn't recognize the newer TVLV
version. Or a node would need to add TVLVs for both versions to
its OGM. But that'd be some more overhead compared to just adding
another flag.

> 
> > > 
> > > Would you consider Antonios comments and update your patchset? I would like to
> > > test it in the next days ...
> > 
> > Yes, I will. I had written some comments for the comments on IRC,
> > but I guess it'll be better to write them here on the list again.
> > 
> 
> Mailing list would be better, yes. Thanks. :)

Done :).

> 
> > > 
> > > Thanks,
> > > 	Simon
> > 
> > Another thing I was thinking about conceptually yesterday was
> > whether we should use more refined flags instead of just
> > MULTICAST_LISTENER_ANNOUNCEMENTS for everything (non-link-local
> > IPv4, all IPv6 except the all-nodes address).
> > 
> > That way we could for instance already add some cases for when to
> > use the multicast optimizations when having a bridge, for
> > instance:
> > 
> > When batman-adv detects that there is an MLD querier and if all
> > nodes have a MULTICAST_LISTENER_ANNOUNCEMENTS or
> > MLA_IPV6_TRANSIENT_LINK_LOCAL flag it could  already
> > optimize link-local, transient IPv6 multicast traffic
> > without needing to modify anything in the bridge code except the
> > addition of the export which was already posted as an RFC on the
> > bridge mailing list.
> > 
> > Hm, again will need to think about that. Whether the extra
> > conceptual complexity is okay because of being able to add some
> > more use-cases with less code in small chunks already.
> 
> Hmm, again I'm not completely sure to follow, but the idea here is
> to enable functionality when having a userspace MLD instead of
> the (planned) bridge MLD stuff?

Yes, that would be the idea.

> 
> If we need the bridge MLD stuff anyway (to have the full feature set
> etc), I'd rather not do more intermediate steps which might be
> obsolete later. However, if the userspace MLD thing is equivalent
> feature-wise than this might be interesting to do.

Hm, don't know. I was thinking that most setups I know of involve
bridges on top of bat0 and to be able to make use of any multicast
optimizations in batman-adv in case of such bridging we need an
MLD querier on the link. Implementing a proper MLD querier might
need several iterations over the bridge / netdev mailing lists and
might therefore need quite some time. During that time those
setups won't be able to make any use of these optimizations.

Unless taking the suggested short-cut which shouldn't be that
difficult to implement, I think, then these bat0+bridge setups
could already enjoy some multicast optimizations and such mesh
networks might already be able to play with multicast streams as I
think updating and merging the tracker packet patches will be done
some time before getting MLD querier bridge code upstream.
(but maybe I'm estimating all this wrong, dunno)

> 
> Cheers,
> 	Simon

Cheers, Linus

Re: [B.A.T.M.A.N.] Basic Multicast Optimizations

[Simon Wunderlich]

[-- Attachment #1: Type: text/plain, Size: 7224 bytes --]

On Thu, May 16, 2013 at 07:42:00PM +0200, Linus Lüssing wrote:
> Hi Simon,
> 
> On Thu, May 16, 2013 at 01:51:29PM +0200, Simon Wunderlich wrote:
> > Hey Linus,
> > 
> > thanks for working and integrating this patchset!
> > 
> > On Sat, May 11, 2013 at 07:23:24PM +0200, Linus Lüssing wrote:
> > > This set of patches is the first one for a more efficient, group aware
> > > multicast forwarding infrastructure in batman-adv.
> > > 
> > > This initial set mainly consists of the integration of announcing the
> > > location of multicast listeners via the translation table mechanism to
> > > be able to find out which batman-adv nodes are actually interested in
> > > certain multicast traffic. As well as the signalizing of this
> > > announcement capability via a new multicast TVLV.
> > > 
> > > Finally some basic multicast forwarding opitimizations are introduced:
> > > If all nodes signalized the MLA capability then link-local IPv6 traffic
> > > will be dropped if there is no interested listener or gets forwarded
> > > via a batman-adv unicast packet if there is just one node interested
> > > in the data.
> > 
> > I've seen that you've only implemented that for IPv6, would it be possible
> > to use it for IPv4 too in the current state? (I use some media streaming
> > thingy at home where I could test ;])
> 
> Hmm, I need to think about it. The thing is it looks like IPv4 does
> not seem to require IGMP. And according to RFC4541, an
> informational one, link-local IPv4 multicast shoud be excluded as
> there might be some applications not issuing IGMP reports.
> 
> For non-link-local addresses we'd need MRD first, as we'd
> otherwise miss to forward things to multicast routers.
> 
> However when using no bridge, like it is the case now, RFC4541
> does not apply and we can in fact reliably get and distribute
> link-local IPv4 multicast addresses as we do not rely on IGMP
> here... 
> 
> One option would be to introduce another flag. But that might
> become confusing for users, it might not be easy to explain to
> users why their 224.0.0.123 would be working on raw bat0 but not
> when they add a bridge and why this would never work. Instead we
> could wait for the MRD implementation and tell people that
> 224.0.0.0/24 won't work.
> 

Hmm, as far as I read the RFC 4541) (2.1.2, point 2) we should just
treat 224.0.0.0/24 (which would translate to 01:00:5e:00:00:xx) as
broadcast - we should do that in any state of the multicast implementation
IMHO, and are then on the safe side. For all other IPv4 multicast
addresses we can assume that IGMP is used.

For the non-link-local IPv4 multicast addresses, could we use the
current mechanism (drop/unicast/broadcast) by getting the assigned
multicast addresses? (of course, assuming we have no bridge)
> > 
> > > 
> > > For now, these optimizations only apply if all nodes in the mesh have
> > > no bridge interface on top their bat0 interface. However making it
> > > possible with bridges, too, and other features are on the roadmap. See
> > > the according wiki page for details [1].
> > 
> > So from what I understand this means:
> > 
> >  * multicast support is working for non-bridged interfaces for now, because
> >    we lack proper multicast router discovery support (MRD) in the bridge code
> >  --> you told me someone is working on that, or you'll pick up to get this done?
> 
> Yes, I had been exchanging some emails with a group of people a
> few months ago who wanted to implement MRD and a proper IGMP/MLD
> querier protocol in the bridge code but didn't hear from them again.
> So yes, I guess I'll pick up on that.

OK, cool.
> 
> 
> For MRD I was thinking about taking a short-cut for now as
> implementing would need a few more lines of code. Instead I think
> I go for detecting whether a querier is on the link both in
> batman-adv and the bridge code for IPv4 or IPv6 and if not disabling
> optimizations accordingly and issuing a warning. That way
> non-link-local IPv6 multicast traffic could be optimized already
> for instance if running an mrd6 instance in userspace which
> already performs both MLD querying and MRD.
> 

Hmm, sorry I still get a little confused in that MRD/MLD/IGMP terminology.
So as I understand there might be a userspace component which does the MLD/MRD
instead of the kernel? Also, is it required to get implement MRD/MLD specifically
in batman-adv or bridges, or would it be possible to use a "general" approach
which could be used for any interface to emit/exchange these kind of messages?

Sorry if I ask stupid questions. :)

Maybe some example or architecture overview (could be put on the wiki page) would
help?

> > 
> >  * the current optimization is to handle multicast like:
> >    * no listener - drop
> >    * one listener - unicast
> >    * more listener - broadcast
> >  --> I think that is a good and simple optimization. :) The "tracker" packets
> >      can come later (as shown in your roadmap). This might require to announce
> >      that tracker packets are supported in the TVLV?
> 
> Yes, we'll probably need to add another capability flag to the
> multicast TVLV.
> 

OK. We could also increase the version number of the TVLV and interpret
both versions in newer multicast implementations, but if we already know
that we can define this information in the current implementation.

> > 
> > Would you consider Antonios comments and update your patchset? I would like to
> > test it in the next days ...
> 
> Yes, I will. I had written some comments for the comments on IRC,
> but I guess it'll be better to write them here on the list again.
> 

Mailing list would be better, yes. Thanks. :)

> > 
> > Thanks,
> > 	Simon
> 
> Another thing I was thinking about conceptually yesterday was
> whether we should use more refined flags instead of just
> MULTICAST_LISTENER_ANNOUNCEMENTS for everything (non-link-local
> IPv4, all IPv6 except the all-nodes address).
> 
> That way we could for instance already add some cases for when to
> use the multicast optimizations when having a bridge, for
> instance:
> 
> When batman-adv detects that there is an MLD querier and if all
> nodes have a MULTICAST_LISTENER_ANNOUNCEMENTS or
> MLA_IPV6_TRANSIENT_LINK_LOCAL flag it could  already
> optimize link-local, transient IPv6 multicast traffic
> without needing to modify anything in the bridge code except the
> addition of the export which was already posted as an RFC on the
> bridge mailing list.
> 
> Hm, again will need to think about that. Whether the extra
> conceptual complexity is okay because of being able to add some
> more use-cases with less code in small chunks already.

Hmm, again I'm not completely sure to follow, but the idea here is
to enable functionality when having a userspace MLD instead of
the (planned) bridge MLD stuff?

If we need the bridge MLD stuff anyway (to have the full feature set
etc), I'd rather not do more intermediate steps which might be
obsolete later. However, if the userspace MLD thing is equivalent
feature-wise than this might be interesting to do.

Cheers,
	Simon

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [B.A.T.M.A.N.] Basic Multicast Optimizations

[Linus Lüssing]

Hi Simon,

On Thu, May 16, 2013 at 01:51:29PM +0200, Simon Wunderlich wrote:
> Hey Linus,
> 
> thanks for working and integrating this patchset!
> 
> On Sat, May 11, 2013 at 07:23:24PM +0200, Linus Lüssing wrote:
> > This set of patches is the first one for a more efficient, group aware
> > multicast forwarding infrastructure in batman-adv.
> > 
> > This initial set mainly consists of the integration of announcing the
> > location of multicast listeners via the translation table mechanism to
> > be able to find out which batman-adv nodes are actually interested in
> > certain multicast traffic. As well as the signalizing of this
> > announcement capability via a new multicast TVLV.
> > 
> > Finally some basic multicast forwarding opitimizations are introduced:
> > If all nodes signalized the MLA capability then link-local IPv6 traffic
> > will be dropped if there is no interested listener or gets forwarded
> > via a batman-adv unicast packet if there is just one node interested
> > in the data.
> 
> I've seen that you've only implemented that for IPv6, would it be possible
> to use it for IPv4 too in the current state? (I use some media streaming
> thingy at home where I could test ;])

Hmm, I need to think about it. The thing is it looks like IPv4 does
not seem to require IGMP. And according to RFC4541, an
informational one, link-local IPv4 multicast shoud be excluded as
there might be some applications not issuing IGMP reports.

For non-link-local addresses we'd need MRD first, as we'd
otherwise miss to forward things to multicast routers.

However when using no bridge, like it is the case now, RFC4541
does not apply and we can in fact reliably get and distribute
link-local IPv4 multicast addresses as we do not rely on IGMP
here... 

One option would be to introduce another flag. But that might
become confusing for users, it might not be easy to explain to
users why their 224.0.0.123 would be working on raw bat0 but not
when they add a bridge and why this would never work. Instead we
could wait for the MRD implementation and tell people that
224.0.0.0/24 won't work.

> 
> > 
> > For now, these optimizations only apply if all nodes in the mesh have
> > no bridge interface on top their bat0 interface. However making it
> > possible with bridges, too, and other features are on the roadmap. See
> > the according wiki page for details [1].
> 
> So from what I understand this means:
> 
>  * multicast support is working for non-bridged interfaces for now, because
>    we lack proper multicast router discovery support (MRD) in the bridge code
>  --> you told me someone is working on that, or you'll pick up to get this done?

Yes, I had been exchanging some emails with a group of people a
few months ago who wanted to implement MRD and a proper IGMP/MLD
querier protocol in the bridge code but didn't hear from them again.
So yes, I guess I'll pick up on that.


For MRD I was thinking about taking a short-cut for now as
implementing would need a few more lines of code. Instead I think
I go for detecting whether a querier is on the link both in
batman-adv and the bridge code for IPv4 or IPv6 and if not disabling
optimizations accordingly and issuing a warning. That way
non-link-local IPv6 multicast traffic could be optimized already
for instance if running an mrd6 instance in userspace which
already performs both MLD querying and MRD.

> 
>  * the current optimization is to handle multicast like:
>    * no listener - drop
>    * one listener - unicast
>    * more listener - broadcast
>  --> I think that is a good and simple optimization. :) The "tracker" packets
>      can come later (as shown in your roadmap). This might require to announce
>      that tracker packets are supported in the TVLV?

Yes, we'll probably need to add another capability flag to the
multicast TVLV.

> 
> Would you consider Antonios comments and update your patchset? I would like to
> test it in the next days ...

Yes, I will. I had written some comments for the comments on IRC,
but I guess it'll be better to write them here on the list again.

> 
> Thanks,
> 	Simon

Another thing I was thinking about conceptually yesterday was
whether we should use more refined flags instead of just
MULTICAST_LISTENER_ANNOUNCEMENTS for everything (non-link-local
IPv4, all IPv6 except the all-nodes address).

That way we could for instance already add some cases for when to
use the multicast optimizations when having a bridge, for
instance:

When batman-adv detects that there is an MLD querier and if all
nodes have a MULTICAST_LISTENER_ANNOUNCEMENTS or
MLA_IPV6_TRANSIENT_LINK_LOCAL flag it could  already
optimize link-local, transient IPv6 multicast traffic
without needing to modify anything in the bridge code except the
addition of the export which was already posted as an RFC on the
bridge mailing list.

Hm, again will need to think about that. Whether the extra
conceptual complexity is okay because of being able to add some
more use-cases with less code in small chunks already.


Cheers, Linus

Re: [B.A.T.M.A.N.] Basic Multicast Optimizations

[Simon Wunderlich]

[-- Attachment #1: Type: text/plain, Size: 2144 bytes --]

Hey Linus,

thanks for working and integrating this patchset!

On Sat, May 11, 2013 at 07:23:24PM +0200, Linus Lüssing wrote:
> This set of patches is the first one for a more efficient, group aware
> multicast forwarding infrastructure in batman-adv.
> 
> This initial set mainly consists of the integration of announcing the
> location of multicast listeners via the translation table mechanism to
> be able to find out which batman-adv nodes are actually interested in
> certain multicast traffic. As well as the signalizing of this
> announcement capability via a new multicast TVLV.
> 
> Finally some basic multicast forwarding opitimizations are introduced:
> If all nodes signalized the MLA capability then link-local IPv6 traffic
> will be dropped if there is no interested listener or gets forwarded
> via a batman-adv unicast packet if there is just one node interested
> in the data.

I've seen that you've only implemented that for IPv6, would it be possible
to use it for IPv4 too in the current state? (I use some media streaming
thingy at home where I could test ;])

> 
> For now, these optimizations only apply if all nodes in the mesh have
> no bridge interface on top their bat0 interface. However making it
> possible with bridges, too, and other features are on the roadmap. See
> the according wiki page for details [1].

So from what I understand this means:

 * multicast support is working for non-bridged interfaces for now, because
   we lack proper multicast router discovery support (MRD) in the bridge code
 --> you told me someone is working on that, or you'll pick up to get this done?

 * the current optimization is to handle multicast like:
   * no listener - drop
   * one listener - unicast
   * more listener - broadcast
 --> I think that is a good and simple optimization. :) The "tracker" packets
     can come later (as shown in your roadmap). This might require to announce
     that tracker packets are supported in the TVLV?

Would you consider Antonios comments and update your patchset? I would like to
test it in the next days ...

Thanks,
	Simon

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

[B.A.T.M.A.N.] Basic Multicast Optimizations

[Linus Lüssing]

This set of patches is the first one for a more efficient, group aware
multicast forwarding infrastructure in batman-adv.

This initial set mainly consists of the integration of announcing the
location of multicast listeners via the translation table mechanism to
be able to find out which batman-adv nodes are actually interested in
certain multicast traffic. As well as the signalizing of this
announcement capability via a new multicast TVLV.

Finally some basic multicast forwarding opitimizations are introduced:
If all nodes signalized the MLA capability then link-local IPv6 traffic
will be dropped if there is no interested listener or gets forwarded
via a batman-adv unicast packet if there is just one node interested
in the data.


For now, these optimizations only apply if all nodes in the mesh have
no bridge interface on top their bat0 interface. However making it
possible with bridges, too, and other features are on the roadmap. See
the according wiki page for details [1].


[1]: http://www.open-mesh.org/projects/batman-adv/wiki/Multicast-ideas-updated#Roadmap