* Re: [PATCH -next] qlcnic: fix compiler warning
From: David Miller @ 2014-01-13 22:46 UTC (permalink / raw)
To: shahed.shaikh
Cc: martin, himanshu.madhani, rajesh.borundia, linux-kernel, trivial,
netdev
In-Reply-To: <262CB373A6D1F14F9B81E82F74F77D5A46F52B8E@avmb2.qlogic.org>
From: Shahed Shaikh <shahed.shaikh@qlogic.com>
Date: Fri, 10 Jan 2014 07:19:19 +0000
>
> Adding netdev.
>
>> -----Original Message-----
>> From: Martin Kaiser,,, [mailto:martin@reykholt.kaiser.cx] On Behalf Of
>> Martin Kaiser
>> Sent: Thursday, January 09, 2014 9:29 PM
>> To: Himanshu Madhani; Rajesh Borundia
>> Cc: linux-kernel; trivial@kernel.org
>> Subject: [PATCH -next] qlcnic: fix compiler warning
>>
>> Add an explicit cast to fix the following warning (seen on Debian Wheezy, gcc
>> 4.7.2)
>>
>> CC [M] drivers/net/wireless/rtlwifi/rtl8192ce/trx.o
>> drivers/net/ethernet/qlogic/qlcnic/qlcnic_io.c: In function
>> ‘qlcnic_send_filter’:
>> drivers/net/ethernet/qlogic/qlcnic/qlcnic_io.c:349:3: warning:
>> passing argument 2 of ‘ether_addr_equal’ from incompatible pointer type
>> [enabled by default]
>> In file included from include/linux/if_vlan.h:16:0,
>> from drivers/net/ethernet/qlogic/qlcnic/qlcnic_io.c:9:
>> include/linux/etherdevice.h:244:20: note: expected ‘const u8 *’ but
>> argument is of type ‘u64 *’
>>
>
> If I am not wrong, this patch should go to David's net-next tree.
>
>> Signed-off-by: Martin Kaiser <martin@kaiser.cx>
>
> Acked-by: Shahed Shaikh <shahed.shaikh@qlogic.com>
Please post the original patch to netdev so that it properly gets
queued up in patchwork, thank you very much.
^ permalink raw reply
* Re: [PATCH net-next] sctp: make sctp_addto_chunk_fixed local
From: David Miller @ 2014-01-13 22:43 UTC (permalink / raw)
To: stephen; +Cc: vyasevich, nhorman, netdev
In-Reply-To: <20140109223111.71e55286@nehalam.linuxnetplumber.net>
From: Stephen Hemminger <stephen@networkplumber.org>
Date: Thu, 9 Jan 2014 22:31:11 -0800
> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Applied, thanks Stephen.
^ permalink raw reply
* Re: [PATCH 0/5] Assorted mvneta fixes
From: Arnaud Ebalard @ 2014-01-13 22:36 UTC (permalink / raw)
To: Willy Tarreau
Cc: davem, netdev, Thomas Petazzoni, Gregory CLEMENT, Eric Dumazet
In-Reply-To: <20140112222233.GG16576@1wt.eu>
Hi,
Willy Tarreau <w@1wt.eu> writes:
>> Funny enough, I spent some time this week-end trying to find the root
>> cause of some kernel freezes and panics appearing randomly after some GB
>> read on a ReadyNAS 102 configured as a NFS server.
>>
>> I tested your fixes and performance series together on top of current
>> 3.13.0-rc7 and I am now unable to reproduce the freeze and panics after
>> having read more than the 300GB of traffic from the NAS: following
>> bandwith with a bwm-ng shows the rate is also far more stable than w/
>> previous driver logic (55MB/sec). So, FWIW:
>>
>> Tested-by: Arnaud Ebalard <arno@natisbad.org>
>
> Thanks for this.
>
> BTW, the "performance" series is not supposed to fix anything,
I was lazy and wanted to give the whole set a try in a single pass.
> and still it seems difficult to me to find what patch might have fixed
> your problem. Maybe the timer used in place of an IRQ has an even
> worse effect than what we could imagine ?
I guess so.
>> Willy, I can extend the test to RN2120 if you think it is useful to also
>> do additional tests on a dual-core armada XP.
>
> It's up to you. These patches have run extensively on my Mirabox (Armada370),
> OpenBlocks AX3 (ArmadaXP dual core) and the XP-GP board (ArmadaXP quad core),
> and fixed the stability issues and performance issues I was facing there. But
> you may be interested in testing them with your workloads (none of my boxes
> is used as an NFS server, NAS or whatever, they mainly see HTTP and very small
> packets used in stress tests).
Well, I spent the evening on my RN104 (Aramda370 w/ 2 GbE ifaces) and my
RN2120 (Dual core ArmadaXP w/ 2GbE ifaces) using one as a router and
serving NFS traffic from the other (and then changing roles). I passed
hundreds of GB of TCP/NFS traffic and did not see any issue.
Additionally, FWIW, testing both using netperf show they easily support
routing traffic w/ line rate perf.
Regarding the patches, the problem they solve impacts all Armada boards
(370 and XP) which are used for network tasks. I think it would be nice
to have those backported to stable. I can commit to do the tests of the
backports both on XP and 370 hardware down to 3.12 or 3.11 kernel if it
can help.
Cheers,
a+
^ permalink raw reply
* [PATCH 09/14] batman-adv: remove returns at the end of void functions
From: Antonio Quartulli @ 2014-01-13 22:31 UTC (permalink / raw)
To: davem; +Cc: netdev, b.a.t.m.a.n, Antonio Quartulli, Marek Lindner
In-Reply-To: <1389652298-472-1-git-send-email-antonio@meshcoding.com>
Return at the end of void functions is not needed.
Since most of the void functions in the code do not do so,
make all the others consistent by removing the useless
returns. Actually all the functions to be "fixed" are in
network-coding.h only.
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
---
net/batman-adv/network-coding.h | 7 -------
1 file changed, 7 deletions(-)
diff --git a/net/batman-adv/network-coding.h b/net/batman-adv/network-coding.h
index ab59b3d..136db6d 100644
--- a/net/batman-adv/network-coding.h
+++ b/net/batman-adv/network-coding.h
@@ -62,7 +62,6 @@ static inline int batadv_nc_mesh_init(struct batadv_priv *bat_priv)
static inline void batadv_nc_mesh_free(struct batadv_priv *bat_priv)
{
- return;
}
static inline void
@@ -72,7 +71,6 @@ batadv_nc_update_nc_node(struct batadv_priv *bat_priv,
struct batadv_ogm_packet *ogm_packet,
int is_single_hop_neigh)
{
- return;
}
static inline void
@@ -81,17 +79,14 @@ batadv_nc_purge_orig(struct batadv_priv *bat_priv,
bool (*to_purge)(struct batadv_priv *,
struct batadv_nc_node *))
{
- return;
}
static inline void batadv_nc_init_bat_priv(struct batadv_priv *bat_priv)
{
- return;
}
static inline void batadv_nc_init_orig(struct batadv_orig_node *orig_node)
{
- return;
}
static inline bool batadv_nc_skb_forward(struct sk_buff *skb,
@@ -104,14 +99,12 @@ static inline void
batadv_nc_skb_store_for_decoding(struct batadv_priv *bat_priv,
struct sk_buff *skb)
{
- return;
}
static inline void
batadv_nc_skb_store_sniffed_unicast(struct batadv_priv *bat_priv,
struct sk_buff *skb)
{
- return;
}
static inline int batadv_nc_nodes_seq_print_text(struct seq_file *seq,
--
1.8.5.2
^ permalink raw reply related
* [PATCH 08/14] batman-adv: add debugfs support to view multiif tables
From: Antonio Quartulli @ 2014-01-13 22:31 UTC (permalink / raw)
To: davem
Cc: netdev, b.a.t.m.a.n, Simon Wunderlich, Marek Lindner,
Antonio Quartulli
In-Reply-To: <1389652298-472-1-git-send-email-antonio@meshcoding.com>
From: Simon Wunderlich <simon@open-mesh.com>
Show tables for the multi interface operation. Originator tables
are added per hard interface.
This patch also changes the API by adding the interface to the
bat_orig_print() parameters.
Signed-off-by: Simon Wunderlich <simon@open-mesh.com>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>
---
net/batman-adv/bat_iv_ogm.c | 12 +++++++-----
net/batman-adv/debugfs.c | 16 +++++++++++++++
net/batman-adv/originator.c | 48 ++++++++++++++++++++++++++++++++++++++++++++-
net/batman-adv/originator.h | 1 +
net/batman-adv/types.h | 3 ++-
5 files changed, 73 insertions(+), 7 deletions(-)
diff --git a/net/batman-adv/bat_iv_ogm.c b/net/batman-adv/bat_iv_ogm.c
index a9b7e82..fb196a0 100644
--- a/net/batman-adv/bat_iv_ogm.c
+++ b/net/batman-adv/bat_iv_ogm.c
@@ -1777,9 +1777,11 @@ batadv_iv_ogm_orig_print_neigh(struct batadv_orig_node *orig_node,
* batadv_iv_ogm_orig_print - print the originator table
* @bat_priv: the bat priv with all the soft interface information
* @seq: debugfs table seq_file struct
+ * @if_outgoing: the outgoing interface for which this should be printed
*/
static void batadv_iv_ogm_orig_print(struct batadv_priv *bat_priv,
- struct seq_file *seq)
+ struct seq_file *seq,
+ struct batadv_hard_iface *if_outgoing)
{
struct batadv_neigh_node *neigh_node;
struct batadv_hashtable *hash = bat_priv->orig_hash;
@@ -1801,12 +1803,12 @@ static void batadv_iv_ogm_orig_print(struct batadv_priv *bat_priv,
rcu_read_lock();
hlist_for_each_entry_rcu(orig_node, head, hash_entry) {
neigh_node = batadv_orig_router_get(orig_node,
- BATADV_IF_DEFAULT);
+ if_outgoing);
if (!neigh_node)
continue;
n_ifinfo = batadv_neigh_ifinfo_get(neigh_node,
- BATADV_IF_DEFAULT);
+ if_outgoing);
if (!n_ifinfo)
goto next;
@@ -1824,8 +1826,8 @@ static void batadv_iv_ogm_orig_print(struct batadv_priv *bat_priv,
neigh_node->addr,
neigh_node->if_incoming->net_dev->name);
- batadv_iv_ogm_orig_print_neigh(orig_node,
- BATADV_IF_DEFAULT, seq);
+ batadv_iv_ogm_orig_print_neigh(orig_node, if_outgoing,
+ seq);
seq_puts(seq, "\n");
batman_count++;
diff --git a/net/batman-adv/debugfs.c b/net/batman-adv/debugfs.c
index 2ea3695..26cbf34 100644
--- a/net/batman-adv/debugfs.c
+++ b/net/batman-adv/debugfs.c
@@ -248,6 +248,19 @@ static int batadv_originators_open(struct inode *inode, struct file *file)
return single_open(file, batadv_orig_seq_print_text, net_dev);
}
+/**
+ * batadv_originators_hardif_open - handles debugfs output for the
+ * originator table of an hard interface
+ * @inode: inode pointer to debugfs file
+ * @file: pointer to the seq_file
+ */
+static int batadv_originators_hardif_open(struct inode *inode,
+ struct file *file)
+{
+ struct net_device *net_dev = (struct net_device *)inode->i_private;
+ return single_open(file, batadv_orig_hardif_seq_print_text, net_dev);
+}
+
static int batadv_gateways_open(struct inode *inode, struct file *file)
{
struct net_device *net_dev = (struct net_device *)inode->i_private;
@@ -383,8 +396,11 @@ struct batadv_debuginfo batadv_hardif_debuginfo_##_name = { \
.release = single_release, \
}, \
};
+static BATADV_HARDIF_DEBUGINFO(originators, S_IRUGO,
+ batadv_originators_hardif_open);
static struct batadv_debuginfo *batadv_hardif_debuginfos[] = {
+ &batadv_hardif_debuginfo_originators,
NULL,
};
diff --git a/net/batman-adv/originator.c b/net/batman-adv/originator.c
index bd88780..5ed037f 100644
--- a/net/batman-adv/originator.c
+++ b/net/batman-adv/originator.c
@@ -940,11 +940,57 @@ int batadv_orig_seq_print_text(struct seq_file *seq, void *offset)
return 0;
}
- bat_priv->bat_algo_ops->bat_orig_print(bat_priv, seq);
+ bat_priv->bat_algo_ops->bat_orig_print(bat_priv, seq,
+ BATADV_IF_DEFAULT);
return 0;
}
+/**
+ * batadv_orig_hardif_seq_print_text - writes originator infos for a specific
+ * outgoing interface
+ * @seq: debugfs table seq_file struct
+ * @offset: not used
+ *
+ * Returns 0
+ */
+int batadv_orig_hardif_seq_print_text(struct seq_file *seq, void *offset)
+{
+ struct net_device *net_dev = (struct net_device *)seq->private;
+ struct batadv_hard_iface *hard_iface;
+ struct batadv_priv *bat_priv;
+
+ hard_iface = batadv_hardif_get_by_netdev(net_dev);
+
+ if (!hard_iface || !hard_iface->soft_iface) {
+ seq_puts(seq, "Interface not known to B.A.T.M.A.N.\n");
+ goto out;
+ }
+
+ bat_priv = netdev_priv(hard_iface->soft_iface);
+ if (!bat_priv->bat_algo_ops->bat_orig_print) {
+ seq_puts(seq,
+ "No printing function for this routing protocol\n");
+ goto out;
+ }
+
+ if (hard_iface->if_status != BATADV_IF_ACTIVE) {
+ seq_puts(seq, "Interface not active\n");
+ goto out;
+ }
+
+ seq_printf(seq, "[B.A.T.M.A.N. adv %s, IF/MAC: %s/%pM (%s %s)]\n",
+ BATADV_SOURCE_VERSION, hard_iface->net_dev->name,
+ hard_iface->net_dev->dev_addr,
+ hard_iface->soft_iface->name, bat_priv->bat_algo_ops->name);
+
+ bat_priv->bat_algo_ops->bat_orig_print(bat_priv, seq, hard_iface);
+
+out:
+ batadv_hardif_free_ref(hard_iface);
+ return 0;
+}
+
int batadv_orig_hash_add_if(struct batadv_hard_iface *hard_iface,
int max_if_num)
{
diff --git a/net/batman-adv/originator.h b/net/batman-adv/originator.h
index 75a4d7b..ad17c3c 100644
--- a/net/batman-adv/originator.h
+++ b/net/batman-adv/originator.h
@@ -53,6 +53,7 @@ batadv_orig_ifinfo_new(struct batadv_orig_node *orig_node,
void batadv_orig_ifinfo_free_ref(struct batadv_orig_ifinfo *orig_ifinfo);
int batadv_orig_seq_print_text(struct seq_file *seq, void *offset);
+int batadv_orig_hardif_seq_print_text(struct seq_file *seq, void *offset);
int batadv_orig_hash_add_if(struct batadv_hard_iface *hard_iface,
int max_if_num);
int batadv_orig_hash_del_if(struct batadv_hard_iface *hard_iface,
diff --git a/net/batman-adv/types.h b/net/batman-adv/types.h
index 529b374..8a1fa16 100644
--- a/net/batman-adv/types.h
+++ b/net/batman-adv/types.h
@@ -1083,7 +1083,8 @@ struct batadv_algo_ops {
struct batadv_neigh_node *neigh2,
struct batadv_hard_iface *if_outgoing2);
/* orig_node handling API */
- void (*bat_orig_print)(struct batadv_priv *priv, struct seq_file *seq);
+ void (*bat_orig_print)(struct batadv_priv *priv, struct seq_file *seq,
+ struct batadv_hard_iface *hard_iface);
void (*bat_orig_free)(struct batadv_orig_node *orig_node);
int (*bat_orig_add_if)(struct batadv_orig_node *orig_node,
int max_if_num);
--
1.8.5.2
^ permalink raw reply related
* [PATCH 07/14] batman-adv: add debugfs structure for information per interface
From: Antonio Quartulli @ 2014-01-13 22:31 UTC (permalink / raw)
To: davem
Cc: netdev, b.a.t.m.a.n, Simon Wunderlich, Marek Lindner,
Antonio Quartulli
In-Reply-To: <1389652298-472-1-git-send-email-antonio@meshcoding.com>
From: Simon Wunderlich <simon@open-mesh.com>
To show information per interface, add a debugfs hardif structure
similar to the system in sysfs. Hard interface folders will be created
in "$debugfs/batman-adv/". Files are not yet added.
Signed-off-by: Simon Wunderlich <simon@open-mesh.com>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>
---
net/batman-adv/debugfs.c | 73 +++++++++++++++++++++++++++++++++++++++++
net/batman-adv/debugfs.h | 2 ++
net/batman-adv/hard-interface.c | 9 +++++
net/batman-adv/types.h | 2 ++
4 files changed, 86 insertions(+)
diff --git a/net/batman-adv/debugfs.c b/net/batman-adv/debugfs.c
index 0adbcee..2ea3695 100644
--- a/net/batman-adv/debugfs.c
+++ b/net/batman-adv/debugfs.c
@@ -369,6 +369,25 @@ static struct batadv_debuginfo *batadv_mesh_debuginfos[] = {
NULL,
};
+#define BATADV_HARDIF_DEBUGINFO(_name, _mode, _open) \
+struct batadv_debuginfo batadv_hardif_debuginfo_##_name = { \
+ .attr = { \
+ .name = __stringify(_name), \
+ .mode = _mode, \
+ }, \
+ .fops = { \
+ .owner = THIS_MODULE, \
+ .open = _open, \
+ .read = seq_read, \
+ .llseek = seq_lseek, \
+ .release = single_release, \
+ }, \
+};
+
+static struct batadv_debuginfo *batadv_hardif_debuginfos[] = {
+ NULL,
+};
+
void batadv_debugfs_init(void)
{
struct batadv_debuginfo **bat_debug;
@@ -396,6 +415,7 @@ void batadv_debugfs_init(void)
return;
err:
debugfs_remove_recursive(batadv_debugfs);
+ batadv_debugfs = NULL;
}
void batadv_debugfs_destroy(void)
@@ -404,6 +424,59 @@ void batadv_debugfs_destroy(void)
batadv_debugfs = NULL;
}
+/**
+ * batadv_debugfs_add_hardif - creates the base directory for a hard interface
+ * in debugfs.
+ * @hard_iface: hard interface which should be added.
+ */
+int batadv_debugfs_add_hardif(struct batadv_hard_iface *hard_iface)
+{
+ struct batadv_debuginfo **bat_debug;
+ struct dentry *file;
+
+ if (!batadv_debugfs)
+ goto out;
+
+ hard_iface->debug_dir = debugfs_create_dir(hard_iface->net_dev->name,
+ batadv_debugfs);
+ if (!hard_iface->debug_dir)
+ goto out;
+
+ for (bat_debug = batadv_hardif_debuginfos; *bat_debug; ++bat_debug) {
+ file = debugfs_create_file(((*bat_debug)->attr).name,
+ S_IFREG | ((*bat_debug)->attr).mode,
+ hard_iface->debug_dir,
+ hard_iface->net_dev,
+ &(*bat_debug)->fops);
+ if (!file)
+ goto rem_attr;
+ }
+
+ return 0;
+rem_attr:
+ debugfs_remove_recursive(hard_iface->debug_dir);
+ hard_iface->debug_dir = NULL;
+out:
+#ifdef CONFIG_DEBUG_FS
+ return -ENOMEM;
+#else
+ return 0;
+#endif /* CONFIG_DEBUG_FS */
+}
+
+/**
+ * batadv_debugfs_del_hardif - delete the base directory for a hard interface
+ * in debugfs.
+ * @hard_iface: hard interface which is deleted.
+ */
+void batadv_debugfs_del_hardif(struct batadv_hard_iface *hard_iface)
+{
+ if (batadv_debugfs) {
+ debugfs_remove_recursive(hard_iface->debug_dir);
+ hard_iface->debug_dir = NULL;
+ }
+}
+
int batadv_debugfs_add_meshif(struct net_device *dev)
{
struct batadv_priv *bat_priv = netdev_priv(dev);
diff --git a/net/batman-adv/debugfs.h b/net/batman-adv/debugfs.h
index 0861b62..981794f 100644
--- a/net/batman-adv/debugfs.h
+++ b/net/batman-adv/debugfs.h
@@ -24,5 +24,7 @@ void batadv_debugfs_init(void);
void batadv_debugfs_destroy(void);
int batadv_debugfs_add_meshif(struct net_device *dev);
void batadv_debugfs_del_meshif(struct net_device *dev);
+int batadv_debugfs_add_hardif(struct batadv_hard_iface *hard_iface);
+void batadv_debugfs_del_hardif(struct batadv_hard_iface *hard_iface);
#endif /* _NET_BATMAN_ADV_DEBUGFS_H_ */
diff --git a/net/batman-adv/hard-interface.c b/net/batman-adv/hard-interface.c
index bebd46ce..3f0e41a 100644
--- a/net/batman-adv/hard-interface.c
+++ b/net/batman-adv/hard-interface.c
@@ -23,6 +23,7 @@
#include "translation-table.h"
#include "routing.h"
#include "sysfs.h"
+#include "debugfs.h"
#include "originator.h"
#include "hash.h"
#include "bridge_loop_avoidance.h"
@@ -539,6 +540,7 @@ static void batadv_hardif_remove_interface_finish(struct work_struct *work)
hard_iface = container_of(work, struct batadv_hard_iface,
cleanup_work);
+ batadv_debugfs_del_hardif(hard_iface);
batadv_sysfs_del_hardif(&hard_iface->hardif_obj);
batadv_hardif_free_ref(hard_iface);
}
@@ -569,6 +571,11 @@ batadv_hardif_add_interface(struct net_device *net_dev)
hard_iface->net_dev = net_dev;
hard_iface->soft_iface = NULL;
hard_iface->if_status = BATADV_IF_NOT_IN_USE;
+
+ ret = batadv_debugfs_add_hardif(hard_iface);
+ if (ret)
+ goto free_sysfs;
+
INIT_LIST_HEAD(&hard_iface->list);
INIT_WORK(&hard_iface->cleanup_work,
batadv_hardif_remove_interface_finish);
@@ -585,6 +592,8 @@ batadv_hardif_add_interface(struct net_device *net_dev)
return hard_iface;
+free_sysfs:
+ batadv_sysfs_del_hardif(&hard_iface->hardif_obj);
free_if:
kfree(hard_iface);
release_dev:
diff --git a/net/batman-adv/types.h b/net/batman-adv/types.h
index 1409279..529b374 100644
--- a/net/batman-adv/types.h
+++ b/net/batman-adv/types.h
@@ -84,6 +84,7 @@ struct batadv_hard_iface_bat_iv {
* @rcu: struct used for freeing in an RCU-safe manner
* @bat_iv: BATMAN IV specific per hard interface data
* @cleanup_work: work queue callback item for hard interface deinit
+ * @debug_dir: dentry for nc subdir in batman-adv directory in debugfs
*/
struct batadv_hard_iface {
struct list_head list;
@@ -98,6 +99,7 @@ struct batadv_hard_iface {
struct rcu_head rcu;
struct batadv_hard_iface_bat_iv bat_iv;
struct work_struct cleanup_work;
+ struct dentry *debug_dir;
};
/**
--
1.8.5.2
^ permalink raw reply related
* [PATCH 06/14] batman-adv: add bonding again
From: Antonio Quartulli @ 2014-01-13 22:31 UTC (permalink / raw)
To: davem
Cc: netdev, b.a.t.m.a.n, Simon Wunderlich, Marek Lindner,
Antonio Quartulli
In-Reply-To: <1389652298-472-1-git-send-email-antonio@meshcoding.com>
From: Simon Wunderlich <simon@open-mesh.com>
With the new interface alternating, the first hop may send packets
in a round robin fashion to it's neighbors because it has multiple
valid routes built by the multi interface optimization. This patch
enables the feature if bonding is selected. Note that unlike the
bonding implemented before, this version is much simpler and may
even enable multi path routing to a certain degree.
Signed-off-by: Simon Wunderlich <simon@open-mesh.com>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>
---
net/batman-adv/originator.c | 4 ++
net/batman-adv/routing.c | 117 ++++++++++++++++++++++++++++++++++++++++++--
net/batman-adv/routing.h | 2 +-
net/batman-adv/types.h | 2 +
4 files changed, 121 insertions(+), 4 deletions(-)
diff --git a/net/batman-adv/originator.c b/net/batman-adv/originator.c
index b8ef416..bd88780 100644
--- a/net/batman-adv/originator.c
+++ b/net/batman-adv/originator.c
@@ -701,6 +701,10 @@ batadv_purge_orig_ifinfo(struct batadv_priv *bat_priv,
hlist_del_rcu(&orig_ifinfo->list);
batadv_orig_ifinfo_free_ref(orig_ifinfo);
+ if (orig_node->last_bonding_candidate == orig_ifinfo) {
+ orig_node->last_bonding_candidate = NULL;
+ batadv_orig_ifinfo_free_ref(orig_ifinfo);
+ }
}
spin_unlock_bh(&orig_node->neigh_list_lock);
diff --git a/net/batman-adv/routing.c b/net/batman-adv/routing.c
index 55e9aeb..153065a 100644
--- a/net/batman-adv/routing.c
+++ b/net/batman-adv/routing.c
@@ -427,16 +427,127 @@ static int batadv_check_unicast_packet(struct batadv_priv *bat_priv,
struct batadv_neigh_node *
batadv_find_router(struct batadv_priv *bat_priv,
struct batadv_orig_node *orig_node,
- const struct batadv_hard_iface *recv_if)
+ struct batadv_hard_iface *recv_if)
{
- struct batadv_neigh_node *router;
+ struct batadv_algo_ops *bao = bat_priv->bat_algo_ops;
+ struct batadv_neigh_node *first_candidate_router = NULL;
+ struct batadv_neigh_node *next_candidate_router = NULL;
+ struct batadv_neigh_node *router, *cand_router = NULL;
+ struct batadv_neigh_node *last_cand_router = NULL;
+ struct batadv_orig_ifinfo *cand, *first_candidate = NULL;
+ struct batadv_orig_ifinfo *next_candidate = NULL;
+ struct batadv_orig_ifinfo *last_candidate;
+ bool last_candidate_found = false;
if (!orig_node)
return NULL;
router = batadv_orig_router_get(orig_node, recv_if);
- /* TODO: fill this later with new bonding mechanism */
+ /* only consider bonding for recv_if == BATADV_IF_DEFAULT (first hop)
+ * and if activated.
+ */
+ if (recv_if == BATADV_IF_DEFAULT || !atomic_read(&bat_priv->bonding) ||
+ !router)
+ return router;
+
+ /* bonding: loop through the list of possible routers found
+ * for the various outgoing interfaces and find a candidate after
+ * the last chosen bonding candidate (next_candidate). If no such
+ * router is found, use the first candidate found (the previously
+ * chosen bonding candidate might have been the last one in the list).
+ * If this can't be found either, return the previously choosen
+ * router - obviously there are no other candidates.
+ */
+ rcu_read_lock();
+ last_candidate = orig_node->last_bonding_candidate;
+ if (last_candidate)
+ last_cand_router = rcu_dereference(last_candidate->router);
+
+ hlist_for_each_entry_rcu(cand, &orig_node->ifinfo_list, list) {
+ /* acquire some structures and references ... */
+ if (!atomic_inc_not_zero(&cand->refcount))
+ continue;
+
+ cand_router = rcu_dereference(cand->router);
+ if (!cand_router)
+ goto next;
+
+ if (!atomic_inc_not_zero(&cand_router->refcount)) {
+ cand_router = NULL;
+ goto next;
+ }
+
+ /* alternative candidate should be good enough to be
+ * considered
+ */
+ if (!bao->bat_neigh_is_equiv_or_better(cand_router,
+ cand->if_outgoing,
+ router, recv_if))
+ goto next;
+
+ /* don't use the same router twice */
+ if (last_cand_router == cand_router)
+ goto next;
+
+ /* mark the first possible candidate */
+ if (!first_candidate) {
+ atomic_inc(&cand_router->refcount);
+ atomic_inc(&cand->refcount);
+ first_candidate = cand;
+ first_candidate_router = cand_router;
+ }
+
+ /* check if the loop has already passed the previously selected
+ * candidate ... this function should select the next candidate
+ * AFTER the previously used bonding candidate.
+ */
+ if (!last_candidate || last_candidate_found) {
+ next_candidate = cand;
+ next_candidate_router = cand_router;
+ break;
+ }
+
+ if (last_candidate == cand)
+ last_candidate_found = true;
+next:
+ /* free references */
+ if (cand_router) {
+ batadv_neigh_node_free_ref(cand_router);
+ cand_router = NULL;
+ }
+ batadv_orig_ifinfo_free_ref(cand);
+ }
+ rcu_read_unlock();
+
+ /* last_bonding_candidate is reset below, remove the old reference. */
+ if (orig_node->last_bonding_candidate)
+ batadv_orig_ifinfo_free_ref(orig_node->last_bonding_candidate);
+
+ /* After finding candidates, handle the three cases:
+ * 1) there is a next candidate, use that
+ * 2) there is no next candidate, use the first of the list
+ * 3) there is no candidate at all, return the default router
+ */
+ if (next_candidate) {
+ batadv_neigh_node_free_ref(router);
+
+ /* remove references to first candidate, we don't need it. */
+ if (first_candidate) {
+ batadv_neigh_node_free_ref(first_candidate_router);
+ batadv_orig_ifinfo_free_ref(first_candidate);
+ }
+ router = next_candidate_router;
+ orig_node->last_bonding_candidate = next_candidate;
+ } else if (first_candidate) {
+ batadv_neigh_node_free_ref(router);
+
+ /* refcounting has already been done in the loop above. */
+ router = first_candidate_router;
+ orig_node->last_bonding_candidate = first_candidate;
+ } else {
+ orig_node->last_bonding_candidate = NULL;
+ }
return router;
}
diff --git a/net/batman-adv/routing.h b/net/batman-adv/routing.h
index 8920d0b..8fa23fc 100644
--- a/net/batman-adv/routing.h
+++ b/net/batman-adv/routing.h
@@ -44,7 +44,7 @@ int batadv_recv_unhandled_unicast_packet(struct sk_buff *skb,
struct batadv_neigh_node *
batadv_find_router(struct batadv_priv *bat_priv,
struct batadv_orig_node *orig_node,
- const struct batadv_hard_iface *recv_if);
+ struct batadv_hard_iface *recv_if);
int batadv_window_protected(struct batadv_priv *bat_priv, int32_t seq_num_diff,
unsigned long *last_reset);
diff --git a/net/batman-adv/types.h b/net/batman-adv/types.h
index c317dfc..1409279 100644
--- a/net/batman-adv/types.h
+++ b/net/batman-adv/types.h
@@ -198,6 +198,7 @@ struct batadv_orig_bat_iv {
* @orig: originator ethernet address
* @primary_addr: hosts primary interface address
* @ifinfo_list: list for routers per outgoing interface
+ * @last_bonding_candidate: pointer to last ifinfo of last used router
* @batadv_dat_addr_t: address of the orig node in the distributed hash
* @last_seen: time when last packet from this node was received
* @bcast_seqno_reset: time when the broadcast seqno window was reset
@@ -238,6 +239,7 @@ struct batadv_orig_node {
uint8_t orig[ETH_ALEN];
uint8_t primary_addr[ETH_ALEN];
struct hlist_head ifinfo_list;
+ struct batadv_orig_ifinfo *last_bonding_candidate;
#ifdef CONFIG_BATMAN_ADV_DAT
batadv_dat_addr_t dat_addr;
#endif
--
1.8.5.2
^ permalink raw reply related
* [PATCH 05/14] batman-adv: consider outgoing interface in OGM sending
From: Antonio Quartulli @ 2014-01-13 22:31 UTC (permalink / raw)
To: davem
Cc: netdev, b.a.t.m.a.n, Simon Wunderlich, Marek Lindner,
Antonio Quartulli
In-Reply-To: <1389652298-472-1-git-send-email-antonio@meshcoding.com>
From: Simon Wunderlich <simon@open-mesh.com>
The current OGM sending an aggregation functionality decides on
which interfaces a packet should be sent when it parses the forward
packet struct. However, with the network wide multi interface
optimization the outgoing interface is decided by the OGM processing
function.
This is reflected by moving the decision in the OGM processing function
and add the outgoing interface in the forwarding packet struct. This
practically implies that an OGM may be added multiple times (once per
outgoing interface), and this also affects aggregation which needs to
consider the outgoing interface as well.
Signed-off-by: Simon Wunderlich <simon@open-mesh.com>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>
---
net/batman-adv/bat_iv_ogm.c | 153 +++++++++++++++++++++++++++-----------------
net/batman-adv/send.c | 19 ++++--
net/batman-adv/types.h | 7 +-
3 files changed, 115 insertions(+), 64 deletions(-)
diff --git a/net/batman-adv/bat_iv_ogm.c b/net/batman-adv/bat_iv_ogm.c
index 789aaa9..a9b7e82 100644
--- a/net/batman-adv/bat_iv_ogm.c
+++ b/net/batman-adv/bat_iv_ogm.c
@@ -466,17 +466,9 @@ static void batadv_iv_ogm_send_to_if(struct batadv_forw_packet *forw_packet,
/* send a batman ogm packet */
static void batadv_iv_ogm_emit(struct batadv_forw_packet *forw_packet)
{
- struct batadv_hard_iface *hard_iface;
struct net_device *soft_iface;
struct batadv_priv *bat_priv;
struct batadv_hard_iface *primary_if = NULL;
- struct batadv_ogm_packet *batadv_ogm_packet;
- unsigned char directlink;
- uint8_t *packet_pos;
-
- packet_pos = forw_packet->skb->data;
- batadv_ogm_packet = (struct batadv_ogm_packet *)packet_pos;
- directlink = (batadv_ogm_packet->flags & BATADV_DIRECTLINK ? 1 : 0);
if (!forw_packet->if_incoming) {
pr_err("Error - can't forward packet: incoming iface not specified\n");
@@ -486,6 +478,12 @@ static void batadv_iv_ogm_emit(struct batadv_forw_packet *forw_packet)
soft_iface = forw_packet->if_incoming->soft_iface;
bat_priv = netdev_priv(soft_iface);
+ if (WARN_ON(!forw_packet->if_outgoing))
+ goto out;
+
+ if (WARN_ON(forw_packet->if_outgoing->soft_iface != soft_iface))
+ goto out;
+
if (forw_packet->if_incoming->if_status != BATADV_IF_ACTIVE)
goto out;
@@ -493,52 +491,35 @@ static void batadv_iv_ogm_emit(struct batadv_forw_packet *forw_packet)
if (!primary_if)
goto out;
- /* multihomed peer assumed
- * non-primary OGMs are only broadcasted on their interface
- */
- if ((directlink && (batadv_ogm_packet->ttl == 1)) ||
- (forw_packet->own && (forw_packet->if_incoming != primary_if))) {
- /* FIXME: what about aggregated packets ? */
- batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
- "%s packet (originator %pM, seqno %u, TTL %d) on interface %s [%pM]\n",
- (forw_packet->own ? "Sending own" : "Forwarding"),
- batadv_ogm_packet->orig,
- ntohl(batadv_ogm_packet->seqno),
- batadv_ogm_packet->ttl,
- forw_packet->if_incoming->net_dev->name,
- forw_packet->if_incoming->net_dev->dev_addr);
-
- /* skb is only used once and than forw_packet is free'd */
- batadv_send_skb_packet(forw_packet->skb,
- forw_packet->if_incoming,
- batadv_broadcast_addr);
- forw_packet->skb = NULL;
-
- goto out;
- }
-
- /* broadcast on every interface */
- rcu_read_lock();
- list_for_each_entry_rcu(hard_iface, &batadv_hardif_list, list) {
- if (hard_iface->soft_iface != soft_iface)
- continue;
-
- batadv_iv_ogm_send_to_if(forw_packet, hard_iface);
- }
- rcu_read_unlock();
+ /* only for one specific outgoing interface */
+ batadv_iv_ogm_send_to_if(forw_packet, forw_packet->if_outgoing);
out:
if (primary_if)
batadv_hardif_free_ref(primary_if);
}
-/* return true if new_packet can be aggregated with forw_packet */
+/**
+ * batadv_iv_ogm_can_aggregate - find out if an OGM can be aggregated on an
+ * existing forward packet
+ * @new_bat_ogm_packet: OGM packet to be aggregated
+ * @bat_priv: the bat priv with all the soft interface information
+ * @packet_len: (total) length of the OGM
+ * @send_time: timestamp (jiffies) when the packet is to be sent
+ * @direktlink: true if this is a direct link packet
+ * @if_incoming: interface where the packet was received
+ * @if_outgoing: interface for which the retransmission should be considered
+ * @forw_packet: the forwarded packet which should be checked
+ *
+ * Returns true if new_packet can be aggregated with forw_packet
+ */
static bool
batadv_iv_ogm_can_aggregate(const struct batadv_ogm_packet *new_bat_ogm_packet,
struct batadv_priv *bat_priv,
int packet_len, unsigned long send_time,
bool directlink,
const struct batadv_hard_iface *if_incoming,
+ const struct batadv_hard_iface *if_outgoing,
const struct batadv_forw_packet *forw_packet)
{
struct batadv_ogm_packet *batadv_ogm_packet;
@@ -572,6 +553,10 @@ batadv_iv_ogm_can_aggregate(const struct batadv_ogm_packet *new_bat_ogm_packet,
if (!primary_if)
goto out;
+ /* packet is not leaving on the same interface. */
+ if (forw_packet->if_outgoing != if_outgoing)
+ goto out;
+
/* packets without direct link flag and high TTL
* are flooded through the net
*/
@@ -613,11 +598,21 @@ out:
return res;
}
-/* create a new aggregated packet and add this packet to it */
+/* batadv_iv_ogm_aggregate_new - create a new aggregated packet and add this
+ * packet to it.
+ * @packet_buff: pointer to the OGM
+ * @packet_len: (total) length of the OGM
+ * @send_time: timestamp (jiffies) when the packet is to be sent
+ * @direct_link: whether this OGM has direct link status
+ * @if_incoming: interface where the packet was received
+ * @if_outgoing: interface for which the retransmission should be considered
+ * @own_packet: true if it is a self-generated ogm
+ */
static void batadv_iv_ogm_aggregate_new(const unsigned char *packet_buff,
int packet_len, unsigned long send_time,
bool direct_link,
struct batadv_hard_iface *if_incoming,
+ struct batadv_hard_iface *if_outgoing,
int own_packet)
{
struct batadv_priv *bat_priv = netdev_priv(if_incoming->soft_iface);
@@ -628,6 +623,9 @@ static void batadv_iv_ogm_aggregate_new(const unsigned char *packet_buff,
if (!atomic_inc_not_zero(&if_incoming->refcount))
return;
+ if (!atomic_inc_not_zero(&if_outgoing->refcount))
+ goto out_free_incoming;
+
/* own packet should always be scheduled */
if (!own_packet) {
if (!batadv_atomic_dec_not_zero(&bat_priv->batman_queue_left)) {
@@ -668,6 +666,7 @@ static void batadv_iv_ogm_aggregate_new(const unsigned char *packet_buff,
forw_packet_aggr->own = own_packet;
forw_packet_aggr->if_incoming = if_incoming;
+ forw_packet_aggr->if_outgoing = if_outgoing;
forw_packet_aggr->num_packets = 0;
forw_packet_aggr->direct_link_flags = BATADV_NO_FLAGS;
forw_packet_aggr->send_time = send_time;
@@ -690,6 +689,8 @@ static void batadv_iv_ogm_aggregate_new(const unsigned char *packet_buff,
return;
out:
+ batadv_hardif_free_ref(if_outgoing);
+out_free_incoming:
batadv_hardif_free_ref(if_incoming);
}
@@ -713,10 +714,21 @@ static void batadv_iv_ogm_aggregate(struct batadv_forw_packet *forw_packet_aggr,
}
}
+/**
+ * batadv_iv_ogm_queue_add - queue up an OGM for transmission
+ * @bat_priv: the bat priv with all the soft interface information
+ * @packet_buff: pointer to the OGM
+ * @packet_len: (total) length of the OGM
+ * @if_incoming: interface where the packet was received
+ * @if_outgoing: interface for which the retransmission should be considered
+ * @own_packet: true if it is a self-generated ogm
+ * @send_time: timestamp (jiffies) when the packet is to be sent
+ */
static void batadv_iv_ogm_queue_add(struct batadv_priv *bat_priv,
unsigned char *packet_buff,
int packet_len,
struct batadv_hard_iface *if_incoming,
+ struct batadv_hard_iface *if_outgoing,
int own_packet, unsigned long send_time)
{
/* _aggr -> pointer to the packet we want to aggregate with
@@ -742,6 +754,7 @@ static void batadv_iv_ogm_queue_add(struct batadv_priv *bat_priv,
bat_priv, packet_len,
send_time, direct_link,
if_incoming,
+ if_outgoing,
forw_packet_pos)) {
forw_packet_aggr = forw_packet_pos;
break;
@@ -765,7 +778,8 @@ static void batadv_iv_ogm_queue_add(struct batadv_priv *bat_priv,
batadv_iv_ogm_aggregate_new(packet_buff, packet_len,
send_time, direct_link,
- if_incoming, own_packet);
+ if_incoming, if_outgoing,
+ own_packet);
} else {
batadv_iv_ogm_aggregate(forw_packet_aggr, packet_buff,
packet_len, direct_link);
@@ -778,7 +792,8 @@ static void batadv_iv_ogm_forward(struct batadv_orig_node *orig_node,
struct batadv_ogm_packet *batadv_ogm_packet,
bool is_single_hop_neigh,
bool is_from_best_next_hop,
- struct batadv_hard_iface *if_incoming)
+ struct batadv_hard_iface *if_incoming,
+ struct batadv_hard_iface *if_outgoing)
{
struct batadv_priv *bat_priv = netdev_priv(if_incoming->soft_iface);
uint16_t tvlv_len;
@@ -823,7 +838,8 @@ static void batadv_iv_ogm_forward(struct batadv_orig_node *orig_node,
batadv_iv_ogm_queue_add(bat_priv, (unsigned char *)batadv_ogm_packet,
BATADV_OGM_HLEN + tvlv_len,
- if_incoming, 0, batadv_iv_ogm_fwd_send_time());
+ if_incoming, if_outgoing, 0,
+ batadv_iv_ogm_fwd_send_time());
}
/**
@@ -868,10 +884,11 @@ static void batadv_iv_ogm_schedule(struct batadv_hard_iface *hard_iface)
struct batadv_priv *bat_priv = netdev_priv(hard_iface->soft_iface);
unsigned char **ogm_buff = &hard_iface->bat_iv.ogm_buff;
struct batadv_ogm_packet *batadv_ogm_packet;
- struct batadv_hard_iface *primary_if;
+ struct batadv_hard_iface *primary_if, *tmp_hard_iface;
int *ogm_buff_len = &hard_iface->bat_iv.ogm_buff_len;
uint32_t seqno;
uint16_t tvlv_len = 0;
+ unsigned long send_time;
primary_if = batadv_primary_if_get_selected(bat_priv);
@@ -894,10 +911,32 @@ static void batadv_iv_ogm_schedule(struct batadv_hard_iface *hard_iface)
atomic_inc(&hard_iface->bat_iv.ogm_seqno);
batadv_iv_ogm_slide_own_bcast_window(hard_iface);
- batadv_iv_ogm_queue_add(bat_priv, hard_iface->bat_iv.ogm_buff,
- hard_iface->bat_iv.ogm_buff_len, hard_iface, 1,
- batadv_iv_ogm_emit_send_time(bat_priv));
+ send_time = batadv_iv_ogm_emit_send_time(bat_priv);
+
+ if (hard_iface != primary_if) {
+ /* OGMs from secondary interfaces are only scheduled on their
+ * respective interfaces.
+ */
+ batadv_iv_ogm_queue_add(bat_priv, *ogm_buff, *ogm_buff_len,
+ hard_iface, hard_iface, 1, send_time);
+ goto out;
+ }
+
+ /* OGMs from primary interfaces are scheduled on all
+ * interfaces.
+ */
+ rcu_read_lock();
+ list_for_each_entry_rcu(tmp_hard_iface, &batadv_hardif_list, list) {
+ if (tmp_hard_iface->soft_iface != hard_iface->soft_iface)
+ continue;
+ batadv_iv_ogm_queue_add(bat_priv, *ogm_buff,
+ *ogm_buff_len, hard_iface,
+ tmp_hard_iface, 1, send_time);
+ }
+ rcu_read_unlock();
+
+out:
if (primary_if)
batadv_hardif_free_ref(primary_if);
}
@@ -1446,6 +1485,10 @@ batadv_iv_ogm_process_per_outif(const struct sk_buff *skb, int ogm_offset,
}
batadv_orig_ifinfo_free_ref(orig_ifinfo);
+ /* only forward for specific interface, not for the default one. */
+ if (if_outgoing == BATADV_IF_DEFAULT)
+ goto out_neigh;
+
/* is single hop (direct) neighbor */
if (is_single_hop_neigh) {
/* OGMs from secondary interfaces should only scheduled once
@@ -1460,7 +1503,8 @@ batadv_iv_ogm_process_per_outif(const struct sk_buff *skb, int ogm_offset,
/* mark direct link on incoming interface */
batadv_iv_ogm_forward(orig_node, ethhdr, ogm_packet,
is_single_hop_neigh,
- is_from_best_next_hop, if_incoming);
+ is_from_best_next_hop, if_incoming,
+ if_outgoing);
batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
"Forwarding packet: rebroadcast neighbor packet with direct link flag\n");
@@ -1480,16 +1524,11 @@ batadv_iv_ogm_process_per_outif(const struct sk_buff *skb, int ogm_offset,
goto out_neigh;
}
- /* only forward the packet on the default interface until the
- * OGM forwarding has been reworked to send on specific interfaces.
- */
- if (if_outgoing != BATADV_IF_DEFAULT)
- goto out_neigh;
batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
"Forwarding packet: rebroadcast originator packet\n");
batadv_iv_ogm_forward(orig_node, ethhdr, ogm_packet,
is_single_hop_neigh, is_from_best_next_hop,
- if_incoming);
+ if_incoming, if_outgoing);
out_neigh:
if ((orig_neigh_node) && (!is_single_hop_neigh))
diff --git a/net/batman-adv/send.c b/net/batman-adv/send.c
index 30d12c4..c93b92d 100644
--- a/net/batman-adv/send.c
+++ b/net/batman-adv/send.c
@@ -387,6 +387,8 @@ static void batadv_forw_packet_free(struct batadv_forw_packet *forw_packet)
kfree_skb(forw_packet->skb);
if (forw_packet->if_incoming)
batadv_hardif_free_ref(forw_packet->if_incoming);
+ if (forw_packet->if_outgoing)
+ batadv_hardif_free_ref(forw_packet->if_outgoing);
kfree(forw_packet);
}
@@ -450,6 +452,7 @@ int batadv_add_bcast_packet_to_list(struct batadv_priv *bat_priv,
forw_packet->skb = newskb;
forw_packet->if_incoming = primary_if;
+ forw_packet->if_outgoing = NULL;
/* how often did we send the bcast packet ? */
forw_packet->num_packets = 0;
@@ -545,11 +548,16 @@ void batadv_send_outstanding_bat_ogm_packet(struct work_struct *work)
bat_priv->bat_algo_ops->bat_ogm_emit(forw_packet);
- /* we have to have at least one packet in the queue
- * to determine the queues wake up time unless we are
- * shutting down
+ /* we have to have at least one packet in the queue to determine the
+ * queues wake up time unless we are shutting down.
+ *
+ * only re-schedule if this is the "original" copy, e.g. the OGM of the
+ * primary interface should only be rescheduled once per period, but
+ * this function will be called for the forw_packet instances of the
+ * other secondary interfaces as well.
*/
- if (forw_packet->own)
+ if (forw_packet->own &&
+ forw_packet->if_incoming == forw_packet->if_outgoing)
batadv_schedule_bat_ogm(forw_packet->if_incoming);
out:
@@ -610,7 +618,8 @@ batadv_purge_outstanding_packets(struct batadv_priv *bat_priv,
* we delete only packets belonging to the given interface
*/
if ((hard_iface) &&
- (forw_packet->if_incoming != hard_iface))
+ (forw_packet->if_incoming != hard_iface) &&
+ (forw_packet->if_outgoing != hard_iface))
continue;
spin_unlock_bh(&bat_priv->forw_bat_list_lock);
diff --git a/net/batman-adv/types.h b/net/batman-adv/types.h
index d3e2bf4..c317dfc 100644
--- a/net/batman-adv/types.h
+++ b/net/batman-adv/types.h
@@ -1018,8 +1018,10 @@ struct batadv_skb_cb {
* @direct_link_flags: direct link flags for aggregated OGM packets
* @num_packets: counter for bcast packet retransmission
* @delayed_work: work queue callback item for packet sending
- * @if_incoming: pointer incoming hard-iface or primary iface if locally
- * generated packet
+ * @if_incoming: pointer to incoming hard-iface or primary iface if
+ * locally generated packet
+ * @if_outgoing: packet where the packet should be sent to, or NULL if
+ * unspecified
*/
struct batadv_forw_packet {
struct hlist_node list;
@@ -1031,6 +1033,7 @@ struct batadv_forw_packet {
uint8_t num_packets;
struct delayed_work delayed_work;
struct batadv_hard_iface *if_incoming;
+ struct batadv_hard_iface *if_outgoing;
};
/**
--
1.8.5.2
^ permalink raw reply related
* [PATCH 03/14] batman-adv: split out router from orig_node
From: Antonio Quartulli @ 2014-01-13 22:31 UTC (permalink / raw)
To: davem
Cc: netdev, b.a.t.m.a.n, Simon Wunderlich, Marek Lindner,
Antonio Quartulli
In-Reply-To: <1389652298-472-1-git-send-email-antonio@meshcoding.com>
From: Simon Wunderlich <simon@open-mesh.com>
For the network wide multi interface optimization there are different
routers for each outgoing interface (outgoing from the OGM perspective,
incoming for payload traffic). To reflect this, change the router and
associated data to a list of routers.
While at it, rename batadv_orig_node_get_router() to
batadv_orig_router_get() to follow the new naming scheme.
Signed-off-by: Simon Wunderlich <simon@open-mesh.com>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>
---
net/batman-adv/bat_iv_ogm.c | 454 +++++++++++++++++++--------------
net/batman-adv/distributed-arp-table.c | 3 +-
net/batman-adv/gateway_client.c | 11 +-
net/batman-adv/icmp_socket.c | 3 +-
net/batman-adv/network-coding.c | 25 +-
net/batman-adv/originator.c | 242 +++++++++++++++++-
net/batman-adv/originator.h | 12 +-
net/batman-adv/routing.c | 38 ++-
net/batman-adv/routing.h | 1 +
net/batman-adv/translation-table.c | 3 +-
net/batman-adv/types.h | 32 ++-
11 files changed, 605 insertions(+), 219 deletions(-)
diff --git a/net/batman-adv/bat_iv_ogm.c b/net/batman-adv/bat_iv_ogm.c
index 157cc11..1f9fe42 100644
--- a/net/batman-adv/bat_iv_ogm.c
+++ b/net/batman-adv/bat_iv_ogm.c
@@ -907,21 +907,21 @@ static void batadv_iv_ogm_schedule(struct batadv_hard_iface *hard_iface)
* originator
* @bat_priv: the bat priv with all the soft interface information
* @orig_node: the orig node who originally emitted the ogm packet
+ * @orig_ifinfo: ifinfo for the outgoing interface of the orig_node
* @ethhdr: Ethernet header of the OGM
* @batadv_ogm_packet: the ogm packet
* @if_incoming: interface where the packet was received
* @if_outgoing: interface for which the retransmission should be considered
- * @tt_buff: pointer to the tt buffer
* @dup_status: the duplicate status of this ogm packet.
*/
static void
batadv_iv_ogm_orig_update(struct batadv_priv *bat_priv,
struct batadv_orig_node *orig_node,
+ struct batadv_orig_ifinfo *orig_ifinfo,
const struct ethhdr *ethhdr,
const struct batadv_ogm_packet *batadv_ogm_packet,
struct batadv_hard_iface *if_incoming,
struct batadv_hard_iface *if_outgoing,
- const unsigned char *tt_buff,
enum batadv_dup_status dup_status)
{
struct batadv_neigh_ifinfo *neigh_ifinfo = NULL;
@@ -1004,14 +1004,14 @@ batadv_iv_ogm_orig_update(struct batadv_priv *bat_priv,
spin_unlock_bh(&neigh_node->ifinfo_lock);
if (dup_status == BATADV_NO_DUP) {
- orig_node->last_ttl = batadv_ogm_packet->ttl;
+ orig_ifinfo->last_ttl = batadv_ogm_packet->ttl;
neigh_ifinfo->last_ttl = batadv_ogm_packet->ttl;
}
/* if this neighbor already is our next hop there is nothing
* to change
*/
- router = batadv_orig_node_get_router(orig_node);
+ router = batadv_orig_router_get(orig_node, if_outgoing);
if (router == neigh_node)
goto out;
@@ -1048,7 +1048,7 @@ batadv_iv_ogm_orig_update(struct batadv_priv *bat_priv,
goto out;
}
- batadv_update_route(bat_priv, orig_node, neigh_node);
+ batadv_update_route(bat_priv, orig_node, if_outgoing, neigh_node);
goto out;
unlock:
@@ -1209,6 +1209,7 @@ batadv_iv_ogm_update_seqnos(const struct ethhdr *ethhdr,
{
struct batadv_priv *bat_priv = netdev_priv(if_incoming->soft_iface);
struct batadv_orig_node *orig_node;
+ struct batadv_orig_ifinfo *orig_ifinfo = NULL;
struct batadv_neigh_node *neigh_node;
struct batadv_neigh_ifinfo *neigh_ifinfo;
int is_dup;
@@ -1225,13 +1226,19 @@ batadv_iv_ogm_update_seqnos(const struct ethhdr *ethhdr,
if (!orig_node)
return BATADV_NO_DUP;
+ orig_ifinfo = batadv_orig_ifinfo_new(orig_node, if_outgoing);
+ if (WARN_ON(!orig_ifinfo)) {
+ batadv_orig_node_free_ref(orig_node);
+ return 0;
+ }
+
spin_lock_bh(&orig_node->bat_iv.ogm_cnt_lock);
- seq_diff = seqno - orig_node->last_real_seqno;
+ seq_diff = seqno - orig_ifinfo->last_real_seqno;
/* signalize caller that the packet is to be dropped. */
if (!hlist_empty(&orig_node->neigh_list) &&
batadv_window_protected(bat_priv, seq_diff,
- &orig_node->batman_seqno_reset)) {
+ &orig_ifinfo->batman_seqno_reset)) {
ret = BATADV_PROTECTED;
goto out;
}
@@ -1245,7 +1252,7 @@ batadv_iv_ogm_update_seqnos(const struct ethhdr *ethhdr,
neigh_addr = neigh_node->addr;
is_dup = batadv_test_bit(neigh_ifinfo->bat_iv.real_bits,
- orig_node->last_real_seqno,
+ orig_ifinfo->last_real_seqno,
seqno);
if (batadv_compare_eth(neigh_addr, ethhdr->h_source) &&
@@ -1273,163 +1280,65 @@ batadv_iv_ogm_update_seqnos(const struct ethhdr *ethhdr,
if (need_update) {
batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
- "updating last_seqno: old %u, new %u\n",
- orig_node->last_real_seqno, seqno);
- orig_node->last_real_seqno = seqno;
+ "%s updating last_seqno: old %u, new %u\n",
+ if_outgoing ? if_outgoing->net_dev->name : "DEFAULT",
+ orig_ifinfo->last_real_seqno, seqno);
+ orig_ifinfo->last_real_seqno = seqno;
}
out:
spin_unlock_bh(&orig_node->bat_iv.ogm_cnt_lock);
batadv_orig_node_free_ref(orig_node);
+ if (orig_ifinfo)
+ batadv_orig_ifinfo_free_ref(orig_ifinfo);
return ret;
}
-static void batadv_iv_ogm_process(const struct ethhdr *ethhdr,
- struct batadv_ogm_packet *batadv_ogm_packet,
- const unsigned char *tt_buff,
- struct batadv_hard_iface *if_incoming)
+
+/**
+ * batadv_iv_ogm_process_per_outif - process a batman iv OGM for an outgoing if
+ * @skb: the skb containing the OGM
+ * @orig_node: the (cached) orig node for the originator of this OGM
+ * @if_incoming: the interface where this packet was received
+ * @if_outgoing: the interface for which the packet should be considered
+ */
+static void
+batadv_iv_ogm_process_per_outif(const struct sk_buff *skb, int ogm_offset,
+ struct batadv_orig_node *orig_node,
+ struct batadv_hard_iface *if_incoming,
+ struct batadv_hard_iface *if_outgoing)
{
struct batadv_priv *bat_priv = netdev_priv(if_incoming->soft_iface);
- struct batadv_hard_iface *hard_iface;
- struct batadv_orig_node *orig_neigh_node, *orig_node, *orig_node_tmp;
struct batadv_neigh_node *router = NULL, *router_router = NULL;
+ struct batadv_orig_node *orig_neigh_node;
+ struct batadv_orig_ifinfo *orig_ifinfo;
struct batadv_neigh_node *orig_neigh_router = NULL;
struct batadv_neigh_ifinfo *router_ifinfo = NULL;
- int has_directlink_flag;
- int is_my_addr = 0, is_my_orig = 0, is_my_oldorig = 0;
- int is_bidirect;
- bool is_single_hop_neigh = false;
- bool is_from_best_next_hop = false;
- int sameseq, similar_ttl;
+ struct batadv_ogm_packet *ogm_packet;
enum batadv_dup_status dup_status;
- uint32_t if_incoming_seqno;
+ bool is_from_best_next_hop = false;
+ bool is_single_hop_neigh = false;
+ bool sameseq, similar_ttl;
+ struct sk_buff *skb_priv;
+ struct ethhdr *ethhdr;
uint8_t *prev_sender;
+ int is_bidirect;
- /* Silently drop when the batman packet is actually not a
- * correct packet.
- *
- * This might happen if a packet is padded (e.g. Ethernet has a
- * minimum frame length of 64 byte) and the aggregation interprets
- * it as an additional length.
- *
- * TODO: A more sane solution would be to have a bit in the
- * batadv_ogm_packet to detect whether the packet is the last
- * packet in an aggregation. Here we expect that the padding
- * is always zero (or not 0x01)
+ /* create a private copy of the skb, as some functions change tq value
+ * and/or flags.
*/
- if (batadv_ogm_packet->packet_type != BATADV_IV_OGM)
+ skb_priv = skb_copy(skb, GFP_ATOMIC);
+ if (!skb_priv)
return;
- /* could be changed by schedule_own_packet() */
- if_incoming_seqno = atomic_read(&if_incoming->bat_iv.ogm_seqno);
+ ethhdr = eth_hdr(skb_priv);
+ ogm_packet = (struct batadv_ogm_packet *)(skb_priv->data + ogm_offset);
- if (batadv_ogm_packet->flags & BATADV_DIRECTLINK)
- has_directlink_flag = 1;
- else
- has_directlink_flag = 0;
-
- if (batadv_compare_eth(ethhdr->h_source, batadv_ogm_packet->orig))
+ dup_status = batadv_iv_ogm_update_seqnos(ethhdr, ogm_packet,
+ if_incoming, if_outgoing);
+ if (batadv_compare_eth(ethhdr->h_source, ogm_packet->orig))
is_single_hop_neigh = true;
- batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
- "Received BATMAN packet via NB: %pM, IF: %s [%pM] (from OG: %pM, via prev OG: %pM, seqno %u, tq %d, TTL %d, V %d, IDF %d)\n",
- ethhdr->h_source, if_incoming->net_dev->name,
- if_incoming->net_dev->dev_addr, batadv_ogm_packet->orig,
- batadv_ogm_packet->prev_sender,
- ntohl(batadv_ogm_packet->seqno), batadv_ogm_packet->tq,
- batadv_ogm_packet->ttl,
- batadv_ogm_packet->version, has_directlink_flag);
-
- rcu_read_lock();
- list_for_each_entry_rcu(hard_iface, &batadv_hardif_list, list) {
- if (hard_iface->if_status != BATADV_IF_ACTIVE)
- continue;
-
- if (hard_iface->soft_iface != if_incoming->soft_iface)
- continue;
-
- if (batadv_compare_eth(ethhdr->h_source,
- hard_iface->net_dev->dev_addr))
- is_my_addr = 1;
-
- if (batadv_compare_eth(batadv_ogm_packet->orig,
- hard_iface->net_dev->dev_addr))
- is_my_orig = 1;
-
- if (batadv_compare_eth(batadv_ogm_packet->prev_sender,
- hard_iface->net_dev->dev_addr))
- is_my_oldorig = 1;
- }
- rcu_read_unlock();
-
- if (is_my_addr) {
- batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
- "Drop packet: received my own broadcast (sender: %pM)\n",
- ethhdr->h_source);
- return;
- }
-
- if (is_my_orig) {
- unsigned long *word;
- int offset;
- int32_t bit_pos;
- int16_t if_num;
- uint8_t *weight;
-
- orig_neigh_node = batadv_iv_ogm_orig_get(bat_priv,
- ethhdr->h_source);
- if (!orig_neigh_node)
- return;
-
- /* neighbor has to indicate direct link and it has to
- * come via the corresponding interface
- * save packet seqno for bidirectional check
- */
- if (has_directlink_flag &&
- batadv_compare_eth(if_incoming->net_dev->dev_addr,
- batadv_ogm_packet->orig)) {
- if_num = if_incoming->if_num;
- offset = if_num * BATADV_NUM_WORDS;
-
- spin_lock_bh(&orig_neigh_node->bat_iv.ogm_cnt_lock);
- word = &(orig_neigh_node->bat_iv.bcast_own[offset]);
- bit_pos = if_incoming_seqno - 2;
- bit_pos -= ntohl(batadv_ogm_packet->seqno);
- batadv_set_bit(word, bit_pos);
- weight = &orig_neigh_node->bat_iv.bcast_own_sum[if_num];
- *weight = bitmap_weight(word,
- BATADV_TQ_LOCAL_WINDOW_SIZE);
- spin_unlock_bh(&orig_neigh_node->bat_iv.ogm_cnt_lock);
- }
-
- batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
- "Drop packet: originator packet from myself (via neighbor)\n");
- batadv_orig_node_free_ref(orig_neigh_node);
- return;
- }
-
- if (is_my_oldorig) {
- batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
- "Drop packet: ignoring all rebroadcast echos (sender: %pM)\n",
- ethhdr->h_source);
- return;
- }
-
- if (batadv_ogm_packet->flags & BATADV_NOT_BEST_NEXT_HOP) {
- batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
- "Drop packet: ignoring all packets not forwarded from the best next hop (sender: %pM)\n",
- ethhdr->h_source);
- return;
- }
-
- orig_node = batadv_iv_ogm_orig_get(bat_priv, batadv_ogm_packet->orig);
- if (!orig_node)
- return;
-
- dup_status = batadv_iv_ogm_update_seqnos(ethhdr, batadv_ogm_packet,
- if_incoming,
- BATADV_IF_DEFAULT);
-
if (dup_status == BATADV_PROTECTED) {
batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
"Drop packet: packet within seqno protection time (sender: %pM)\n",
@@ -1437,29 +1346,28 @@ static void batadv_iv_ogm_process(const struct ethhdr *ethhdr,
goto out;
}
- if (batadv_ogm_packet->tq == 0) {
+ if (ogm_packet->tq == 0) {
batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
"Drop packet: originator packet with tq equal 0\n");
goto out;
}
- router = batadv_orig_node_get_router(orig_node);
+ router = batadv_orig_router_get(orig_node, if_outgoing);
if (router) {
- orig_node_tmp = router->orig_node;
- router_router = batadv_orig_node_get_router(orig_node_tmp);
- router_ifinfo = batadv_neigh_ifinfo_get(router,
- BATADV_IF_DEFAULT);
+ router_router = batadv_orig_router_get(router->orig_node,
+ if_outgoing);
+ router_ifinfo = batadv_neigh_ifinfo_get(router, if_outgoing);
}
- if ((router && router_ifinfo->bat_iv.tq_avg != 0) &&
+ if ((router_ifinfo && router_ifinfo->bat_iv.tq_avg != 0) &&
(batadv_compare_eth(router->addr, ethhdr->h_source)))
is_from_best_next_hop = true;
- prev_sender = batadv_ogm_packet->prev_sender;
+ prev_sender = ogm_packet->prev_sender;
/* avoid temporary routing loops */
if (router && router_router &&
(batadv_compare_eth(router->addr, prev_sender)) &&
- !(batadv_compare_eth(batadv_ogm_packet->orig, prev_sender)) &&
+ !(batadv_compare_eth(ogm_packet->orig, prev_sender)) &&
(batadv_compare_eth(router->addr, router_router->addr))) {
batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
"Drop packet: ignoring all rebroadcast packets that may make me loop (sender: %pM)\n",
@@ -1467,7 +1375,8 @@ static void batadv_iv_ogm_process(const struct ethhdr *ethhdr,
goto out;
}
- batadv_tvlv_ogm_receive(bat_priv, batadv_ogm_packet, orig_node);
+ if (if_outgoing == BATADV_IF_DEFAULT)
+ batadv_tvlv_ogm_receive(bat_priv, ogm_packet, orig_node);
/* if sender is a direct neighbor the sender mac equals
* originator mac
@@ -1483,9 +1392,10 @@ static void batadv_iv_ogm_process(const struct ethhdr *ethhdr,
/* Update nc_nodes of the originator */
batadv_nc_update_nc_node(bat_priv, orig_node, orig_neigh_node,
- batadv_ogm_packet, is_single_hop_neigh);
+ ogm_packet, is_single_hop_neigh);
- orig_neigh_router = batadv_orig_node_get_router(orig_neigh_node);
+ orig_neigh_router = batadv_orig_router_get(orig_neigh_node,
+ if_outgoing);
/* drop packet if sender is not a direct neighbor and if we
* don't route towards it
@@ -1497,25 +1407,41 @@ static void batadv_iv_ogm_process(const struct ethhdr *ethhdr,
}
is_bidirect = batadv_iv_ogm_calc_tq(orig_node, orig_neigh_node,
- batadv_ogm_packet, if_incoming,
- BATADV_IF_DEFAULT);
+ ogm_packet, if_incoming,
+ if_outgoing);
/* update ranking if it is not a duplicate or has the same
* seqno and similar ttl as the non-duplicate
*/
- sameseq = orig_node->last_real_seqno == ntohl(batadv_ogm_packet->seqno);
- similar_ttl = orig_node->last_ttl - 3 <= batadv_ogm_packet->ttl;
+ orig_ifinfo = batadv_orig_ifinfo_new(orig_node, if_outgoing);
+ if (!orig_ifinfo)
+ goto out_neigh;
+
+ sameseq = orig_ifinfo->last_real_seqno == ntohl(ogm_packet->seqno);
+ similar_ttl = (orig_ifinfo->last_ttl - 3) <= ogm_packet->ttl;
+
if (is_bidirect && ((dup_status == BATADV_NO_DUP) ||
- (sameseq && similar_ttl)))
- batadv_iv_ogm_orig_update(bat_priv, orig_node, ethhdr,
- batadv_ogm_packet, if_incoming,
- BATADV_IF_DEFAULT, tt_buff,
- dup_status);
+ (sameseq && similar_ttl))) {
+ batadv_iv_ogm_orig_update(bat_priv, orig_node,
+ orig_ifinfo, ethhdr,
+ ogm_packet, if_incoming,
+ if_outgoing, dup_status);
+ }
+ batadv_orig_ifinfo_free_ref(orig_ifinfo);
/* is single hop (direct) neighbor */
if (is_single_hop_neigh) {
+ /* OGMs from secondary interfaces should only scheduled once
+ * per interface where it has been received, not multiple times
+ */
+ if ((ogm_packet->ttl <= 2) &&
+ (if_incoming != if_outgoing)) {
+ batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
+ "Drop packet: OGM from secondary interface and wrong outgoing interface\n");
+ goto out_neigh;
+ }
/* mark direct link on incoming interface */
- batadv_iv_ogm_forward(orig_node, ethhdr, batadv_ogm_packet,
+ batadv_iv_ogm_forward(orig_node, ethhdr, ogm_packet,
is_single_hop_neigh,
is_from_best_next_hop, if_incoming);
@@ -1537,9 +1463,14 @@ static void batadv_iv_ogm_process(const struct ethhdr *ethhdr,
goto out_neigh;
}
+ /* only forward the packet on the default interface until the
+ * OGM forwarding has been reworked to send on specific interfaces.
+ */
+ if (if_outgoing != BATADV_IF_DEFAULT)
+ goto out_neigh;
batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
"Forwarding packet: rebroadcast originator packet\n");
- batadv_iv_ogm_forward(orig_node, ethhdr, batadv_ogm_packet,
+ batadv_iv_ogm_forward(orig_node, ethhdr, ogm_packet,
is_single_hop_neigh, is_from_best_next_hop,
if_incoming);
@@ -1554,6 +1485,165 @@ out:
if (orig_neigh_router)
batadv_neigh_node_free_ref(orig_neigh_router);
+ kfree_skb(skb_priv);
+}
+
+/**
+ * batadv_iv_ogm_process - process an incoming batman iv OGM
+ * @skb: the skb containing the OGM
+ * @ogm_offset: offset to the OGM which should be processed (for aggregates)
+ * @if_incoming: the interface where this packet was receved
+ */
+static void batadv_iv_ogm_process(const struct sk_buff *skb, int ogm_offset,
+ struct batadv_hard_iface *if_incoming)
+{
+ struct batadv_priv *bat_priv = netdev_priv(if_incoming->soft_iface);
+ struct batadv_orig_node *orig_neigh_node, *orig_node;
+ struct batadv_hard_iface *hard_iface;
+ struct batadv_ogm_packet *ogm_packet;
+ uint32_t if_incoming_seqno;
+ bool has_directlink_flag;
+ struct ethhdr *ethhdr;
+ bool is_my_oldorig = false;
+ bool is_my_addr = false;
+ bool is_my_orig = false;
+
+ ogm_packet = (struct batadv_ogm_packet *)(skb->data + ogm_offset);
+ ethhdr = eth_hdr(skb);
+
+ /* Silently drop when the batman packet is actually not a
+ * correct packet.
+ *
+ * This might happen if a packet is padded (e.g. Ethernet has a
+ * minimum frame length of 64 byte) and the aggregation interprets
+ * it as an additional length.
+ *
+ * TODO: A more sane solution would be to have a bit in the
+ * batadv_ogm_packet to detect whether the packet is the last
+ * packet in an aggregation. Here we expect that the padding
+ * is always zero (or not 0x01)
+ */
+ if (ogm_packet->packet_type != BATADV_IV_OGM)
+ return;
+
+ /* could be changed by schedule_own_packet() */
+ if_incoming_seqno = atomic_read(&if_incoming->bat_iv.ogm_seqno);
+
+ if (ogm_packet->flags & BATADV_DIRECTLINK)
+ has_directlink_flag = true;
+ else
+ has_directlink_flag = false;
+
+ batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
+ "Received BATMAN packet via NB: %pM, IF: %s [%pM] (from OG: %pM, via prev OG: %pM, seqno %u, tq %d, TTL %d, V %d, IDF %d)\n",
+ ethhdr->h_source, if_incoming->net_dev->name,
+ if_incoming->net_dev->dev_addr, ogm_packet->orig,
+ ogm_packet->prev_sender, ntohl(ogm_packet->seqno),
+ ogm_packet->tq, ogm_packet->ttl,
+ ogm_packet->version, has_directlink_flag);
+
+ rcu_read_lock();
+ list_for_each_entry_rcu(hard_iface, &batadv_hardif_list, list) {
+ if (hard_iface->if_status != BATADV_IF_ACTIVE)
+ continue;
+
+ if (hard_iface->soft_iface != if_incoming->soft_iface)
+ continue;
+
+ if (batadv_compare_eth(ethhdr->h_source,
+ hard_iface->net_dev->dev_addr))
+ is_my_addr = true;
+
+ if (batadv_compare_eth(ogm_packet->orig,
+ hard_iface->net_dev->dev_addr))
+ is_my_orig = true;
+
+ if (batadv_compare_eth(ogm_packet->prev_sender,
+ hard_iface->net_dev->dev_addr))
+ is_my_oldorig = true;
+ }
+ rcu_read_unlock();
+
+ if (is_my_addr) {
+ batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
+ "Drop packet: received my own broadcast (sender: %pM)\n",
+ ethhdr->h_source);
+ return;
+ }
+
+ if (is_my_orig) {
+ unsigned long *word;
+ int offset;
+ int32_t bit_pos;
+ int16_t if_num;
+ uint8_t *weight;
+
+ orig_neigh_node = batadv_iv_ogm_orig_get(bat_priv,
+ ethhdr->h_source);
+ if (!orig_neigh_node)
+ return;
+
+ /* neighbor has to indicate direct link and it has to
+ * come via the corresponding interface
+ * save packet seqno for bidirectional check
+ */
+ if (has_directlink_flag &&
+ batadv_compare_eth(if_incoming->net_dev->dev_addr,
+ ogm_packet->orig)) {
+ if_num = if_incoming->if_num;
+ offset = if_num * BATADV_NUM_WORDS;
+
+ spin_lock_bh(&orig_neigh_node->bat_iv.ogm_cnt_lock);
+ word = &(orig_neigh_node->bat_iv.bcast_own[offset]);
+ bit_pos = if_incoming_seqno - 2;
+ bit_pos -= ntohl(ogm_packet->seqno);
+ batadv_set_bit(word, bit_pos);
+ weight = &orig_neigh_node->bat_iv.bcast_own_sum[if_num];
+ *weight = bitmap_weight(word,
+ BATADV_TQ_LOCAL_WINDOW_SIZE);
+ spin_unlock_bh(&orig_neigh_node->bat_iv.ogm_cnt_lock);
+ }
+
+ batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
+ "Drop packet: originator packet from myself (via neighbor)\n");
+ batadv_orig_node_free_ref(orig_neigh_node);
+ return;
+ }
+
+ if (is_my_oldorig) {
+ batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
+ "Drop packet: ignoring all rebroadcast echos (sender: %pM)\n",
+ ethhdr->h_source);
+ return;
+ }
+
+ if (ogm_packet->flags & BATADV_NOT_BEST_NEXT_HOP) {
+ batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
+ "Drop packet: ignoring all packets not forwarded from the best next hop (sender: %pM)\n",
+ ethhdr->h_source);
+ return;
+ }
+
+ orig_node = batadv_iv_ogm_orig_get(bat_priv, ogm_packet->orig);
+ if (!orig_node)
+ return;
+
+ batadv_iv_ogm_process_per_outif(skb, ogm_offset, orig_node,
+ if_incoming, BATADV_IF_DEFAULT);
+
+ rcu_read_lock();
+ list_for_each_entry_rcu(hard_iface, &batadv_hardif_list, list) {
+ if (hard_iface->if_status != BATADV_IF_ACTIVE)
+ continue;
+
+ if (hard_iface->soft_iface != bat_priv->soft_iface)
+ continue;
+
+ batadv_iv_ogm_process_per_outif(skb, ogm_offset, orig_node,
+ if_incoming, hard_iface);
+ }
+ rcu_read_unlock();
+
batadv_orig_node_free_ref(orig_node);
}
@@ -1561,11 +1651,9 @@ static int batadv_iv_ogm_receive(struct sk_buff *skb,
struct batadv_hard_iface *if_incoming)
{
struct batadv_priv *bat_priv = netdev_priv(if_incoming->soft_iface);
- struct batadv_ogm_packet *batadv_ogm_packet;
- struct ethhdr *ethhdr;
- int buff_pos = 0, packet_len;
- unsigned char *tvlv_buff, *packet_buff;
+ struct batadv_ogm_packet *ogm_packet;
uint8_t *packet_pos;
+ int ogm_offset;
bool ret;
ret = batadv_check_management_packet(skb, if_incoming, BATADV_OGM_HLEN);
@@ -1582,24 +1670,19 @@ static int batadv_iv_ogm_receive(struct sk_buff *skb,
batadv_add_counter(bat_priv, BATADV_CNT_MGMT_RX_BYTES,
skb->len + ETH_HLEN);
- packet_len = skb_headlen(skb);
- ethhdr = eth_hdr(skb);
- packet_buff = skb->data;
- batadv_ogm_packet = (struct batadv_ogm_packet *)packet_buff;
+ ogm_offset = 0;
+ ogm_packet = (struct batadv_ogm_packet *)skb->data;
/* unpack the aggregated packets and process them one by one */
- while (batadv_iv_ogm_aggr_packet(buff_pos, packet_len,
- batadv_ogm_packet->tvlv_len)) {
- tvlv_buff = packet_buff + buff_pos + BATADV_OGM_HLEN;
+ while (batadv_iv_ogm_aggr_packet(ogm_offset, skb_headlen(skb),
+ ogm_packet->tvlv_len)) {
+ batadv_iv_ogm_process(skb, ogm_offset, if_incoming);
- batadv_iv_ogm_process(ethhdr, batadv_ogm_packet,
- tvlv_buff, if_incoming);
+ ogm_offset += BATADV_OGM_HLEN;
+ ogm_offset += ntohs(ogm_packet->tvlv_len);
- buff_pos += BATADV_OGM_HLEN;
- buff_pos += ntohs(batadv_ogm_packet->tvlv_len);
-
- packet_pos = packet_buff + buff_pos;
- batadv_ogm_packet = (struct batadv_ogm_packet *)packet_pos;
+ packet_pos = skb->data + ogm_offset;
+ ogm_packet = (struct batadv_ogm_packet *)packet_pos;
}
kfree_skb(skb);
@@ -1661,7 +1744,8 @@ static void batadv_iv_ogm_orig_print(struct batadv_priv *bat_priv,
rcu_read_lock();
hlist_for_each_entry_rcu(orig_node, head, hash_entry) {
- neigh_node = batadv_orig_node_get_router(orig_node);
+ neigh_node = batadv_orig_router_get(orig_node,
+ BATADV_IF_DEFAULT);
if (!neigh_node)
continue;
diff --git a/net/batman-adv/distributed-arp-table.c b/net/batman-adv/distributed-arp-table.c
index 997ae6a..1d214cd 100644
--- a/net/batman-adv/distributed-arp-table.c
+++ b/net/batman-adv/distributed-arp-table.c
@@ -589,7 +589,8 @@ static bool batadv_dat_send_data(struct batadv_priv *bat_priv,
if (cand[i].type == BATADV_DAT_CANDIDATE_NOT_FOUND)
continue;
- neigh_node = batadv_orig_node_get_router(cand[i].orig_node);
+ neigh_node = batadv_orig_router_get(cand[i].orig_node,
+ BATADV_IF_DEFAULT);
if (!neigh_node)
goto free_orig;
diff --git a/net/batman-adv/gateway_client.c b/net/batman-adv/gateway_client.c
index 4e8f5b1..9ece204 100644
--- a/net/batman-adv/gateway_client.c
+++ b/net/batman-adv/gateway_client.c
@@ -146,7 +146,7 @@ batadv_gw_get_best_gw_node(struct batadv_priv *bat_priv)
continue;
orig_node = gw_node->orig_node;
- router = batadv_orig_node_get_router(orig_node);
+ router = batadv_orig_router_get(orig_node, BATADV_IF_DEFAULT);
if (!router)
continue;
@@ -266,7 +266,8 @@ void batadv_gw_election(struct batadv_priv *bat_priv)
if (next_gw) {
sprintf(gw_addr, "%pM", next_gw->orig_node->orig);
- router = batadv_orig_node_get_router(next_gw->orig_node);
+ router = batadv_orig_router_get(next_gw->orig_node,
+ BATADV_IF_DEFAULT);
if (!router) {
batadv_gw_reselect(bat_priv);
goto out;
@@ -335,7 +336,7 @@ void batadv_gw_check_election(struct batadv_priv *bat_priv,
if (!curr_gw_orig)
goto reselect;
- router_gw = batadv_orig_node_get_router(curr_gw_orig);
+ router_gw = batadv_orig_router_get(curr_gw_orig, BATADV_IF_DEFAULT);
if (!router_gw)
goto reselect;
@@ -348,7 +349,7 @@ void batadv_gw_check_election(struct batadv_priv *bat_priv,
if (curr_gw_orig == orig_node)
goto out;
- router_orig = batadv_orig_node_get_router(orig_node);
+ router_orig = batadv_orig_router_get(orig_node, BATADV_IF_DEFAULT);
if (!router_orig)
goto out;
@@ -576,7 +577,7 @@ static int batadv_write_buffer_text(struct batadv_priv *bat_priv,
struct batadv_neigh_ifinfo *router_ifinfo = NULL;
int ret = -1;
- router = batadv_orig_node_get_router(gw_node->orig_node);
+ router = batadv_orig_router_get(gw_node->orig_node, BATADV_IF_DEFAULT);
if (!router)
goto out;
diff --git a/net/batman-adv/icmp_socket.c b/net/batman-adv/icmp_socket.c
index 3c08eee..b384cb8 100644
--- a/net/batman-adv/icmp_socket.c
+++ b/net/batman-adv/icmp_socket.c
@@ -215,7 +215,8 @@ static ssize_t batadv_socket_write(struct file *file, const char __user *buff,
if (!orig_node)
goto dst_unreach;
- neigh_node = batadv_orig_node_get_router(orig_node);
+ neigh_node = batadv_orig_router_get(orig_node,
+ BATADV_IF_DEFAULT);
if (!neigh_node)
goto dst_unreach;
diff --git a/net/batman-adv/network-coding.c b/net/batman-adv/network-coding.c
index aec3fa5..6a0b328 100644
--- a/net/batman-adv/network-coding.c
+++ b/net/batman-adv/network-coding.c
@@ -718,9 +718,21 @@ static bool batadv_can_nc_with_orig(struct batadv_priv *bat_priv,
struct batadv_orig_node *orig_node,
struct batadv_ogm_packet *ogm_packet)
{
- if (orig_node->last_real_seqno != ntohl(ogm_packet->seqno))
+ struct batadv_orig_ifinfo *orig_ifinfo;
+ uint32_t last_real_seqno;
+ uint8_t last_ttl;
+
+ orig_ifinfo = batadv_orig_ifinfo_get(orig_node, BATADV_IF_DEFAULT);
+ if (!orig_ifinfo)
return false;
- if (orig_node->last_ttl != ogm_packet->ttl + 1)
+
+ last_ttl = orig_ifinfo->last_ttl;
+ last_real_seqno = orig_ifinfo->last_real_seqno;
+ batadv_orig_ifinfo_free_ref(orig_ifinfo);
+
+ if (last_real_seqno != ntohl(ogm_packet->seqno))
+ return false;
+ if (last_ttl != ogm_packet->ttl + 1)
return false;
if (!batadv_compare_eth(ogm_packet->orig, ogm_packet->prev_sender))
return false;
@@ -1019,7 +1031,11 @@ static bool batadv_nc_code_packets(struct batadv_priv *bat_priv,
int coded_size = sizeof(*coded_packet);
int header_add = coded_size - unicast_size;
- router_neigh = batadv_orig_node_get_router(neigh_node->orig_node);
+ /* TODO: do we need to consider the outgoing interface for
+ * coded packets?
+ */
+ router_neigh = batadv_orig_router_get(neigh_node->orig_node,
+ BATADV_IF_DEFAULT);
if (!router_neigh)
goto out;
@@ -1029,7 +1045,8 @@ static bool batadv_nc_code_packets(struct batadv_priv *bat_priv,
goto out;
neigh_tmp = nc_packet->neigh_node;
- router_coding = batadv_orig_node_get_router(neigh_tmp->orig_node);
+ router_coding = batadv_orig_router_get(neigh_tmp->orig_node,
+ BATADV_IF_DEFAULT);
if (!router_coding)
goto out;
diff --git a/net/batman-adv/originator.c b/net/batman-adv/originator.c
index 1a4725f..b8ef416 100644
--- a/net/batman-adv/originator.c
+++ b/net/batman-adv/originator.c
@@ -231,14 +231,31 @@ void batadv_neigh_node_free_ref(struct batadv_neigh_node *neigh_node)
call_rcu(&neigh_node->rcu, batadv_neigh_node_free_rcu);
}
-/* increases the refcounter of a found router */
+/**
+ * batadv_orig_node_get_router - router to the originator depending on iface
+ * @orig_node: the orig node for the router
+ * @if_outgoing: the interface where the payload packet has been received or
+ * the OGM should be sent to
+ *
+ * Returns the neighbor which should be router for this orig_node/iface.
+ *
+ * The object is returned with refcounter increased by 1.
+ */
struct batadv_neigh_node *
-batadv_orig_node_get_router(struct batadv_orig_node *orig_node)
+batadv_orig_router_get(struct batadv_orig_node *orig_node,
+ const struct batadv_hard_iface *if_outgoing)
{
- struct batadv_neigh_node *router;
+ struct batadv_orig_ifinfo *orig_ifinfo;
+ struct batadv_neigh_node *router = NULL;
rcu_read_lock();
- router = rcu_dereference(orig_node->router);
+ hlist_for_each_entry_rcu(orig_ifinfo, &orig_node->ifinfo_list, list) {
+ if (orig_ifinfo->if_outgoing != if_outgoing)
+ continue;
+
+ router = rcu_dereference(orig_ifinfo->router);
+ break;
+ }
if (router && !atomic_inc_not_zero(&router->refcount))
router = NULL;
@@ -248,6 +265,86 @@ batadv_orig_node_get_router(struct batadv_orig_node *orig_node)
}
/**
+ * batadv_orig_ifinfo_get - find the ifinfo from an orig_node
+ * @orig_node: the orig node to be queried
+ * @if_outgoing: the interface for which the ifinfo should be acquired
+ *
+ * Returns the requested orig_ifinfo or NULL if not found.
+ *
+ * The object is returned with refcounter increased by 1.
+ */
+struct batadv_orig_ifinfo *
+batadv_orig_ifinfo_get(struct batadv_orig_node *orig_node,
+ struct batadv_hard_iface *if_outgoing)
+{
+ struct batadv_orig_ifinfo *tmp, *orig_ifinfo = NULL;
+
+ rcu_read_lock();
+ hlist_for_each_entry_rcu(tmp, &orig_node->ifinfo_list,
+ list) {
+ if (tmp->if_outgoing != if_outgoing)
+ continue;
+
+ if (!atomic_inc_not_zero(&tmp->refcount))
+ continue;
+
+ orig_ifinfo = tmp;
+ break;
+ }
+ rcu_read_unlock();
+
+ return orig_ifinfo;
+}
+
+/**
+ * batadv_orig_ifinfo_new - search and possibly create an orig_ifinfo object
+ * @orig_node: the orig node to be queried
+ * @if_outgoing: the interface for which the ifinfo should be acquired
+ *
+ * Returns NULL in case of failure or the orig_ifinfo object for the if_outgoing
+ * interface otherwise. The object is created and added to the list
+ * if it does not exist.
+ *
+ * The object is returned with refcounter increased by 1.
+ */
+struct batadv_orig_ifinfo *
+batadv_orig_ifinfo_new(struct batadv_orig_node *orig_node,
+ struct batadv_hard_iface *if_outgoing)
+{
+ struct batadv_orig_ifinfo *orig_ifinfo = NULL;
+ unsigned long reset_time;
+
+ spin_lock_bh(&orig_node->neigh_list_lock);
+
+ orig_ifinfo = batadv_orig_ifinfo_get(orig_node, if_outgoing);
+ if (orig_ifinfo)
+ goto out;
+
+ orig_ifinfo = kzalloc(sizeof(*orig_ifinfo), GFP_ATOMIC);
+ if (!orig_ifinfo)
+ goto out;
+
+ if (if_outgoing != BATADV_IF_DEFAULT &&
+ !atomic_inc_not_zero(&if_outgoing->refcount)) {
+ kfree(orig_ifinfo);
+ orig_ifinfo = NULL;
+ goto out;
+ }
+
+ reset_time = jiffies - 1;
+ reset_time -= msecs_to_jiffies(BATADV_RESET_PROTECTION_MS);
+ orig_ifinfo->batman_seqno_reset = reset_time;
+ orig_ifinfo->if_outgoing = if_outgoing;
+ INIT_HLIST_NODE(&orig_ifinfo->list);
+ atomic_set(&orig_ifinfo->refcount, 2);
+ hlist_add_head_rcu(&orig_ifinfo->list,
+ &orig_node->ifinfo_list);
+out:
+ spin_unlock_bh(&orig_node->neigh_list_lock);
+ return orig_ifinfo;
+}
+
+/**
* batadv_neigh_ifinfo_get - find the ifinfo from an neigh_node
* @neigh_node: the neigh node to be queried
* @if_outgoing: the interface for which the ifinfo should be acquired
@@ -360,11 +457,51 @@ out:
return neigh_node;
}
+/**
+ * batadv_orig_ifinfo_free_rcu - free the orig_ifinfo object
+ * @rcu: rcu pointer of the orig_ifinfo object
+ */
+static void batadv_orig_ifinfo_free_rcu(struct rcu_head *rcu)
+{
+ struct batadv_orig_ifinfo *orig_ifinfo;
+
+ orig_ifinfo = container_of(rcu, struct batadv_orig_ifinfo, rcu);
+
+ if (orig_ifinfo->if_outgoing != BATADV_IF_DEFAULT)
+ batadv_hardif_free_ref_now(orig_ifinfo->if_outgoing);
+
+ kfree(orig_ifinfo);
+}
+
+/**
+ * batadv_orig_ifinfo_free_ref - decrement the refcounter and possibly free
+ * the orig_ifinfo (without rcu callback)
+ * @orig_ifinfo: the orig_ifinfo object to release
+ */
+static void
+batadv_orig_ifinfo_free_ref_now(struct batadv_orig_ifinfo *orig_ifinfo)
+{
+ if (atomic_dec_and_test(&orig_ifinfo->refcount))
+ batadv_orig_ifinfo_free_rcu(&orig_ifinfo->rcu);
+}
+
+/**
+ * batadv_orig_ifinfo_free_ref - decrement the refcounter and possibly free
+ * the orig_ifinfo
+ * @orig_ifinfo: the orig_ifinfo object to release
+ */
+void batadv_orig_ifinfo_free_ref(struct batadv_orig_ifinfo *orig_ifinfo)
+{
+ if (atomic_dec_and_test(&orig_ifinfo->refcount))
+ call_rcu(&orig_ifinfo->rcu, batadv_orig_ifinfo_free_rcu);
+}
+
static void batadv_orig_node_free_rcu(struct rcu_head *rcu)
{
struct hlist_node *node_tmp;
struct batadv_neigh_node *neigh_node;
struct batadv_orig_node *orig_node;
+ struct batadv_orig_ifinfo *orig_ifinfo;
orig_node = container_of(rcu, struct batadv_orig_node, rcu);
@@ -377,6 +514,11 @@ static void batadv_orig_node_free_rcu(struct rcu_head *rcu)
batadv_neigh_node_free_ref_now(neigh_node);
}
+ hlist_for_each_entry_safe(orig_ifinfo, node_tmp,
+ &orig_node->ifinfo_list, list) {
+ hlist_del_rcu(&orig_ifinfo->list);
+ batadv_orig_ifinfo_free_ref_now(orig_ifinfo);
+ }
spin_unlock_bh(&orig_node->neigh_list_lock);
/* Free nc_nodes */
@@ -474,6 +616,7 @@ struct batadv_orig_node *batadv_orig_node_new(struct batadv_priv *bat_priv,
INIT_HLIST_HEAD(&orig_node->neigh_list);
INIT_LIST_HEAD(&orig_node->vlan_list);
+ INIT_HLIST_HEAD(&orig_node->ifinfo_list);
spin_lock_init(&orig_node->bcast_seqno_lock);
spin_lock_init(&orig_node->neigh_list_lock);
spin_lock_init(&orig_node->tt_buff_lock);
@@ -489,13 +632,11 @@ struct batadv_orig_node *batadv_orig_node_new(struct batadv_priv *bat_priv,
orig_node->bat_priv = bat_priv;
memcpy(orig_node->orig, addr, ETH_ALEN);
batadv_dat_init_orig_node_addr(orig_node);
- orig_node->router = NULL;
atomic_set(&orig_node->last_ttvn, 0);
orig_node->tt_buff = NULL;
orig_node->tt_buff_len = 0;
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;
/* create a vlan object for the "untagged" LAN */
vlan = batadv_orig_node_vlan_new(orig_node, BATADV_NO_FLAGS);
@@ -520,6 +661,55 @@ free_orig_node:
}
/**
+ * batadv_purge_orig_ifinfo - purge obsolete ifinfo entries from originator
+ * @bat_priv: the bat priv with all the soft interface information
+ * @orig_node: orig node which is to be checked
+ *
+ * Returns true if any ifinfo entry was purged, false otherwise.
+ */
+static bool
+batadv_purge_orig_ifinfo(struct batadv_priv *bat_priv,
+ struct batadv_orig_node *orig_node)
+{
+ struct batadv_orig_ifinfo *orig_ifinfo;
+ struct batadv_hard_iface *if_outgoing;
+ struct hlist_node *node_tmp;
+ bool ifinfo_purged = false;
+
+ spin_lock_bh(&orig_node->neigh_list_lock);
+
+ /* for all ifinfo objects for this originator */
+ hlist_for_each_entry_safe(orig_ifinfo, node_tmp,
+ &orig_node->ifinfo_list, list) {
+ if_outgoing = orig_ifinfo->if_outgoing;
+
+ /* always keep the default interface */
+ if (if_outgoing == BATADV_IF_DEFAULT)
+ continue;
+
+ /* don't purge if the interface is not (going) down */
+ if ((if_outgoing->if_status != BATADV_IF_INACTIVE) &&
+ (if_outgoing->if_status != BATADV_IF_NOT_IN_USE) &&
+ (if_outgoing->if_status != BATADV_IF_TO_BE_REMOVED))
+ continue;
+
+ batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
+ "router/ifinfo purge: originator %pM, iface: %s\n",
+ orig_node->orig, if_outgoing->net_dev->name);
+
+ ifinfo_purged = true;
+
+ hlist_del_rcu(&orig_ifinfo->list);
+ batadv_orig_ifinfo_free_ref(orig_ifinfo);
+ }
+
+ spin_unlock_bh(&orig_node->neigh_list_lock);
+
+ return ifinfo_purged;
+}
+
+
+/**
* batadv_purge_orig_neighbors - purges neighbors from originator
* @bat_priv: the bat priv with all the soft interface information
* @orig_node: orig node which is to be checked
@@ -607,10 +797,22 @@ batadv_find_best_neighbor(struct batadv_priv *bat_priv,
return best;
}
+/**
+ * batadv_purge_orig_node - purges obsolete information from an orig_node
+ * @bat_priv: the bat priv with all the soft interface information
+ * @orig_node: orig node which is to be checked
+ *
+ * This function checks if the orig_node or substructures of it have become
+ * obsolete, and purges this information if that's the case.
+ *
+ * Returns true if the orig_node is to be removed, false otherwise.
+ */
static bool batadv_purge_orig_node(struct batadv_priv *bat_priv,
struct batadv_orig_node *orig_node)
{
struct batadv_neigh_node *best_neigh_node;
+ struct batadv_hard_iface *hard_iface;
+ bool changed;
if (batadv_has_timed_out(orig_node->last_seen,
2 * BATADV_PURGE_TIMEOUT)) {
@@ -620,15 +822,39 @@ static bool batadv_purge_orig_node(struct batadv_priv *bat_priv,
jiffies_to_msecs(orig_node->last_seen));
return true;
}
- if (!batadv_purge_orig_neighbors(bat_priv, orig_node))
+ changed = batadv_purge_orig_ifinfo(bat_priv, orig_node);
+ changed = changed || batadv_purge_orig_neighbors(bat_priv, orig_node);
+
+ if (!changed)
return false;
+ /* first for NULL ... */
best_neigh_node = batadv_find_best_neighbor(bat_priv, orig_node,
BATADV_IF_DEFAULT);
- batadv_update_route(bat_priv, orig_node, best_neigh_node);
+ batadv_update_route(bat_priv, orig_node, BATADV_IF_DEFAULT,
+ best_neigh_node);
if (best_neigh_node)
batadv_neigh_node_free_ref(best_neigh_node);
+ /* ... then for all other interfaces. */
+ rcu_read_lock();
+ list_for_each_entry_rcu(hard_iface, &batadv_hardif_list, list) {
+ if (hard_iface->if_status != BATADV_IF_ACTIVE)
+ continue;
+
+ if (hard_iface->soft_iface != bat_priv->soft_iface)
+ continue;
+
+ best_neigh_node = batadv_find_best_neighbor(bat_priv,
+ orig_node,
+ hard_iface);
+ batadv_update_route(bat_priv, orig_node, hard_iface,
+ best_neigh_node);
+ if (best_neigh_node)
+ batadv_neigh_node_free_ref(best_neigh_node);
+ }
+ rcu_read_unlock();
+
return false;
}
diff --git a/net/batman-adv/originator.h b/net/batman-adv/originator.h
index 29fa4c4..75a4d7b 100644
--- a/net/batman-adv/originator.h
+++ b/net/batman-adv/originator.h
@@ -34,7 +34,8 @@ batadv_neigh_node_new(struct batadv_hard_iface *hard_iface,
struct batadv_orig_node *orig_node);
void batadv_neigh_node_free_ref(struct batadv_neigh_node *neigh_node);
struct batadv_neigh_node *
-batadv_orig_node_get_router(struct batadv_orig_node *orig_node);
+batadv_orig_router_get(struct batadv_orig_node *orig_node,
+ const struct batadv_hard_iface *if_outgoing);
struct batadv_neigh_ifinfo *
batadv_neigh_ifinfo_new(struct batadv_neigh_node *neigh,
struct batadv_hard_iface *if_outgoing);
@@ -42,6 +43,15 @@ struct batadv_neigh_ifinfo *
batadv_neigh_ifinfo_get(struct batadv_neigh_node *neigh,
struct batadv_hard_iface *if_outgoing);
void batadv_neigh_ifinfo_free_ref(struct batadv_neigh_ifinfo *neigh_ifinfo);
+
+struct batadv_orig_ifinfo *
+batadv_orig_ifinfo_get(struct batadv_orig_node *orig_node,
+ struct batadv_hard_iface *if_outgoing);
+struct batadv_orig_ifinfo *
+batadv_orig_ifinfo_new(struct batadv_orig_node *orig_node,
+ struct batadv_hard_iface *if_outgoing);
+void batadv_orig_ifinfo_free_ref(struct batadv_orig_ifinfo *orig_ifinfo);
+
int batadv_orig_seq_print_text(struct seq_file *seq, void *offset);
int batadv_orig_hash_add_if(struct batadv_hard_iface *hard_iface,
int max_if_num);
diff --git a/net/batman-adv/routing.c b/net/batman-adv/routing.c
index 4fd2687..55e9aeb 100644
--- a/net/batman-adv/routing.c
+++ b/net/batman-adv/routing.c
@@ -33,13 +33,32 @@
static int batadv_route_unicast_packet(struct sk_buff *skb,
struct batadv_hard_iface *recv_if);
+/**
+ * _batadv_update_route - set the router for this originator
+ * @bat_priv: the bat priv with all the soft interface information
+ * @orig_node: orig node which is to be configured
+ * @recv_if: the receive interface for which this route is set
+ * @neigh_node: neighbor which should be the next router
+ *
+ * This function does not perform any error checks
+ */
static void _batadv_update_route(struct batadv_priv *bat_priv,
struct batadv_orig_node *orig_node,
+ struct batadv_hard_iface *recv_if,
struct batadv_neigh_node *neigh_node)
{
+ struct batadv_orig_ifinfo *orig_ifinfo;
struct batadv_neigh_node *curr_router;
- curr_router = batadv_orig_node_get_router(orig_node);
+ orig_ifinfo = batadv_orig_ifinfo_get(orig_node, recv_if);
+ if (!orig_ifinfo)
+ return;
+
+ rcu_read_lock();
+ curr_router = rcu_dereference(orig_ifinfo->router);
+ if (curr_router && !atomic_inc_not_zero(&curr_router->refcount))
+ curr_router = NULL;
+ rcu_read_unlock();
/* route deleted */
if ((curr_router) && (!neigh_node)) {
@@ -69,16 +88,25 @@ static void _batadv_update_route(struct batadv_priv *bat_priv,
neigh_node = NULL;
spin_lock_bh(&orig_node->neigh_list_lock);
- rcu_assign_pointer(orig_node->router, neigh_node);
+ rcu_assign_pointer(orig_ifinfo->router, neigh_node);
spin_unlock_bh(&orig_node->neigh_list_lock);
+ batadv_orig_ifinfo_free_ref(orig_ifinfo);
/* decrease refcount of previous best neighbor */
if (curr_router)
batadv_neigh_node_free_ref(curr_router);
}
+/**
+ * batadv_update_route - set the router for this originator
+ * @bat_priv: the bat priv with all the soft interface information
+ * @orig_node: orig node which is to be configured
+ * @recv_if: the receive interface for which this route is set
+ * @neigh_node: neighbor which should be the next router
+ */
void batadv_update_route(struct batadv_priv *bat_priv,
struct batadv_orig_node *orig_node,
+ struct batadv_hard_iface *recv_if,
struct batadv_neigh_node *neigh_node)
{
struct batadv_neigh_node *router = NULL;
@@ -86,10 +114,10 @@ void batadv_update_route(struct batadv_priv *bat_priv,
if (!orig_node)
goto out;
- router = batadv_orig_node_get_router(orig_node);
+ router = batadv_orig_router_get(orig_node, recv_if);
if (router != neigh_node)
- _batadv_update_route(bat_priv, orig_node, neigh_node);
+ _batadv_update_route(bat_priv, orig_node, recv_if, neigh_node);
out:
if (router)
@@ -406,7 +434,7 @@ batadv_find_router(struct batadv_priv *bat_priv,
if (!orig_node)
return NULL;
- router = batadv_orig_node_get_router(orig_node);
+ router = batadv_orig_router_get(orig_node, recv_if);
/* TODO: fill this later with new bonding mechanism */
diff --git a/net/batman-adv/routing.h b/net/batman-adv/routing.h
index 192f0aa..8920d0b 100644
--- a/net/batman-adv/routing.h
+++ b/net/batman-adv/routing.h
@@ -23,6 +23,7 @@ bool batadv_check_management_packet(struct sk_buff *skb,
int header_len);
void batadv_update_route(struct batadv_priv *bat_priv,
struct batadv_orig_node *orig_node,
+ struct batadv_hard_iface *recv_if,
struct batadv_neigh_node *neigh_node);
int batadv_recv_icmp_packet(struct sk_buff *skb,
struct batadv_hard_iface *recv_if);
diff --git a/net/batman-adv/translation-table.c b/net/batman-adv/translation-table.c
index 22c32ae..ec89a1b 100644
--- a/net/batman-adv/translation-table.c
+++ b/net/batman-adv/translation-table.c
@@ -1400,7 +1400,8 @@ batadv_transtable_best_orig(struct batadv_priv *bat_priv,
head = &tt_global_entry->orig_list;
hlist_for_each_entry_rcu(orig_entry, head, list) {
- router = batadv_orig_node_get_router(orig_entry->orig_node);
+ router = batadv_orig_router_get(orig_entry->orig_node,
+ BATADV_IF_DEFAULT);
if (!router)
continue;
diff --git a/net/batman-adv/types.h b/net/batman-adv/types.h
index 8435d7c..d3e2bf4 100644
--- a/net/batman-adv/types.h
+++ b/net/batman-adv/types.h
@@ -101,6 +101,28 @@ struct batadv_hard_iface {
};
/**
+ * struct batadv_orig_ifinfo - originator info per outgoing interface
+ * @list: list node for orig_node::ifinfo_list
+ * @if_outgoing: pointer to outgoing hard interface
+ * @router: router that should be used to reach this originator
+ * @last_real_seqno: last and best known sequence number
+ * @last_ttl: ttl of last received packet
+ * @batman_seqno_reset: time when the batman seqno window was reset
+ * @refcount: number of contexts the object is used
+ * @rcu: struct used for freeing in an RCU-safe manner
+ */
+struct batadv_orig_ifinfo {
+ struct hlist_node list;
+ struct batadv_hard_iface *if_outgoing;
+ struct batadv_neigh_node __rcu *router; /* rcu protected pointer */
+ uint32_t last_real_seqno;
+ uint8_t last_ttl;
+ unsigned long batman_seqno_reset;
+ atomic_t refcount;
+ struct rcu_head rcu;
+};
+
+/**
* struct batadv_frag_table_entry - head in the fragment buffer table
* @head: head of list with fragments
* @lock: lock to protect the list of fragments
@@ -175,11 +197,10 @@ struct batadv_orig_bat_iv {
* struct batadv_orig_node - structure for orig_list maintaining nodes of mesh
* @orig: originator ethernet address
* @primary_addr: hosts primary interface address
- * @router: router that should be used to reach this originator
+ * @ifinfo_list: list for routers per outgoing interface
* @batadv_dat_addr_t: address of the orig node in the distributed hash
* @last_seen: time when last packet from this node was received
* @bcast_seqno_reset: time when the broadcast seqno window was reset
- * @batman_seqno_reset: time when the batman seqno window was reset
* @capabilities: announced capabilities of this originator
* @last_ttvn: last seen translation table version number
* @tt_buff: last tt changeset this node received from the orig node
@@ -192,8 +213,6 @@ struct batadv_orig_bat_iv {
* made up by two operations (data structure update and metdata -CRC/TTVN-
* recalculation) and they have to be executed atomically in order to avoid
* another thread to read the table/metadata between those.
- * @last_real_seqno: last and best known sequence number
- * @last_ttl: ttl of last received packet
* @bcast_bits: bitfield containing the info which payload broadcast originated
* from this orig node this host already has seen (relative to
* last_bcast_seqno)
@@ -218,13 +237,12 @@ struct batadv_orig_bat_iv {
struct batadv_orig_node {
uint8_t orig[ETH_ALEN];
uint8_t primary_addr[ETH_ALEN];
- struct batadv_neigh_node __rcu *router; /* rcu protected pointer */
+ struct hlist_head ifinfo_list;
#ifdef CONFIG_BATMAN_ADV_DAT
batadv_dat_addr_t dat_addr;
#endif
unsigned long last_seen;
unsigned long bcast_seqno_reset;
- unsigned long batman_seqno_reset;
uint8_t capabilities;
atomic_t last_ttvn;
unsigned char *tt_buff;
@@ -233,8 +251,6 @@ struct batadv_orig_node {
bool tt_initialised;
/* prevents from changing the table while reading it */
spinlock_t tt_lock;
- uint32_t last_real_seqno;
- uint8_t last_ttl;
DECLARE_BITMAP(bcast_bits, BATADV_TQ_LOCAL_WINDOW_SIZE);
uint32_t last_bcast_seqno;
struct hlist_head neigh_list;
--
1.8.5.2
^ permalink raw reply related
* [PATCH 04/14] batman-adv: add WiFi penalty
From: Antonio Quartulli @ 2014-01-13 22:31 UTC (permalink / raw)
To: davem
Cc: netdev, b.a.t.m.a.n, Simon Wunderlich, Marek Lindner,
Antonio Quartulli
In-Reply-To: <1389652298-472-1-git-send-email-antonio@meshcoding.com>
From: Simon Wunderlich <simon@open-mesh.com>
If the same interface is used for sending and receiving, there might be
throughput degradation on half-duplex interfaces such as WiFi. Add a
penalty if the same interface is used to reflect this problem in the
metric. At the same time, change the hop penalty from 30 to 15 so there
will be no change for single wifi mesh network. the effective hop
penalty will stay at 30 due to the new wifi penalty for these networks.
Signed-off-by: Simon Wunderlich <simon@open-mesh.com>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>
---
net/batman-adv/bat_iv_ogm.c | 27 ++++++++++++++++++++++-----
net/batman-adv/hard-interface.h | 1 +
net/batman-adv/soft-interface.c | 2 +-
3 files changed, 24 insertions(+), 6 deletions(-)
diff --git a/net/batman-adv/bat_iv_ogm.c b/net/batman-adv/bat_iv_ogm.c
index 1f9fe42..789aaa9 100644
--- a/net/batman-adv/bat_iv_ogm.c
+++ b/net/batman-adv/bat_iv_ogm.c
@@ -1088,6 +1088,7 @@ static int batadv_iv_ogm_calc_tq(struct batadv_orig_node *orig_node,
unsigned int neigh_rq_inv_cube, neigh_rq_max_cube;
int tq_asym_penalty, inv_asym_penalty, if_num, ret = 0;
unsigned int combined_tq;
+ int tq_iface_penalty;
/* find corresponding one hop neighbor */
rcu_read_lock();
@@ -1169,15 +1170,31 @@ static int batadv_iv_ogm_calc_tq(struct batadv_orig_node *orig_node,
inv_asym_penalty /= neigh_rq_max_cube;
tq_asym_penalty = BATADV_TQ_MAX_VALUE - inv_asym_penalty;
- combined_tq = batadv_ogm_packet->tq * tq_own * tq_asym_penalty;
- combined_tq /= BATADV_TQ_MAX_VALUE * BATADV_TQ_MAX_VALUE;
+ /* penalize if the OGM is forwarded on the same interface. WiFi
+ * interfaces and other half duplex devices suffer from throughput
+ * drops as they can't send and receive at the same time.
+ */
+ tq_iface_penalty = BATADV_TQ_MAX_VALUE;
+ if (if_outgoing && (if_incoming == if_outgoing) &&
+ batadv_is_wifi_netdev(if_outgoing->net_dev))
+ tq_iface_penalty = batadv_hop_penalty(BATADV_TQ_MAX_VALUE,
+ bat_priv);
+
+ combined_tq = batadv_ogm_packet->tq *
+ tq_own *
+ tq_asym_penalty *
+ tq_iface_penalty;
+ combined_tq /= BATADV_TQ_MAX_VALUE *
+ BATADV_TQ_MAX_VALUE *
+ BATADV_TQ_MAX_VALUE;
batadv_ogm_packet->tq = combined_tq;
batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
- "bidirectional: orig = %-15pM neigh = %-15pM => own_bcast = %2i, real recv = %2i, local tq: %3i, asym_penalty: %3i, total tq: %3i\n",
+ "bidirectional: orig = %-15pM neigh = %-15pM => own_bcast = %2i, real recv = %2i, local tq: %3i, asym_penalty: %3i, iface_penalty: %3i, total tq: %3i, if_incoming = %s, if_outgoing = %s\n",
orig_node->orig, orig_neigh_node->orig, total_count,
- neigh_rq_count, tq_own,
- tq_asym_penalty, batadv_ogm_packet->tq);
+ neigh_rq_count, tq_own, tq_asym_penalty, tq_iface_penalty,
+ batadv_ogm_packet->tq, if_incoming->net_dev->name,
+ if_outgoing ? if_outgoing->net_dev->name : "DEFAULT");
/* if link has the minimum required transmission quality
* consider it bidirectional
diff --git a/net/batman-adv/hard-interface.h b/net/batman-adv/hard-interface.h
index 8d4304b..ccd42c2 100644
--- a/net/batman-adv/hard-interface.h
+++ b/net/batman-adv/hard-interface.h
@@ -40,6 +40,7 @@ enum batadv_hard_if_cleanup {
extern struct notifier_block batadv_hard_if_notifier;
bool batadv_is_wifi_netdev(struct net_device *net_device);
+bool batadv_is_wifi_iface(int ifindex);
struct batadv_hard_iface*
batadv_hardif_get_by_netdev(const struct net_device *net_dev);
int batadv_hardif_enable_interface(struct batadv_hard_iface *hard_iface,
diff --git a/net/batman-adv/soft-interface.c b/net/batman-adv/soft-interface.c
index c50f643..2dc2270 100644
--- a/net/batman-adv/soft-interface.c
+++ b/net/batman-adv/soft-interface.c
@@ -697,7 +697,7 @@ static int batadv_softif_init_late(struct net_device *dev)
atomic_set(&bat_priv->gw.bandwidth_down, 100);
atomic_set(&bat_priv->gw.bandwidth_up, 20);
atomic_set(&bat_priv->orig_interval, 1000);
- atomic_set(&bat_priv->hop_penalty, 30);
+ atomic_set(&bat_priv->hop_penalty, 15);
#ifdef CONFIG_BATMAN_ADV_DEBUG
atomic_set(&bat_priv->log_level, 0);
#endif
--
1.8.5.2
^ permalink raw reply related
* [PATCH 02/14] batman-adv: split tq information in neigh_node struct
From: Antonio Quartulli @ 2014-01-13 22:31 UTC (permalink / raw)
To: davem
Cc: netdev, b.a.t.m.a.n, Simon Wunderlich, Marek Lindner,
Antonio Quartulli
In-Reply-To: <1389652298-472-1-git-send-email-antonio@meshcoding.com>
From: Simon Wunderlich <simon@open-mesh.com>
For the network wide multi interface optimization it is required to save
metrics per outgoing interface in one neighbor. Therefore a new type is
introduced to keep interface-specific information. This also requires
some changes in access and list management.
The compare and equiv_or_better API calls are changed to take the
outgoing interface into consideration.
Signed-off-by: Simon Wunderlich <simon@open-mesh.com>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>
---
net/batman-adv/bat_iv_ogm.c | 267 +++++++++++++++++++++++++++++--------
net/batman-adv/gateway_client.c | 75 +++++++++--
net/batman-adv/hard-interface.h | 17 +++
net/batman-adv/main.h | 6 +
net/batman-adv/network-coding.c | 24 +++-
net/batman-adv/originator.c | 229 ++++++++++++++++++++++++++++---
net/batman-adv/originator.h | 7 +
net/batman-adv/translation-table.c | 3 +-
net/batman-adv/types.h | 74 ++++++----
9 files changed, 588 insertions(+), 114 deletions(-)
diff --git a/net/batman-adv/bat_iv_ogm.c b/net/batman-adv/bat_iv_ogm.c
index 48f5d35..157cc11 100644
--- a/net/batman-adv/bat_iv_ogm.c
+++ b/net/batman-adv/bat_iv_ogm.c
@@ -272,7 +272,14 @@ batadv_iv_ogm_neigh_new(struct batadv_hard_iface *hard_iface,
if (!neigh_node)
goto out;
- spin_lock_init(&neigh_node->bat_iv.lq_update_lock);
+ if (!atomic_inc_not_zero(&hard_iface->refcount)) {
+ kfree(neigh_node);
+ neigh_node = NULL;
+ goto out;
+ }
+
+ neigh_node->orig_node = orig_neigh;
+ neigh_node->if_incoming = hard_iface;
batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
"Creating new neighbor %pM for orig_node %pM on interface %s\n",
@@ -895,15 +902,30 @@ static void batadv_iv_ogm_schedule(struct batadv_hard_iface *hard_iface)
batadv_hardif_free_ref(primary_if);
}
+/**
+ * batadv_iv_ogm_orig_update - use OGM to update corresponding data in an
+ * originator
+ * @bat_priv: the bat priv with all the soft interface information
+ * @orig_node: the orig node who originally emitted the ogm packet
+ * @ethhdr: Ethernet header of the OGM
+ * @batadv_ogm_packet: the ogm packet
+ * @if_incoming: interface where the packet was received
+ * @if_outgoing: interface for which the retransmission should be considered
+ * @tt_buff: pointer to the tt buffer
+ * @dup_status: the duplicate status of this ogm packet.
+ */
static void
batadv_iv_ogm_orig_update(struct batadv_priv *bat_priv,
struct batadv_orig_node *orig_node,
const struct ethhdr *ethhdr,
const struct batadv_ogm_packet *batadv_ogm_packet,
struct batadv_hard_iface *if_incoming,
+ struct batadv_hard_iface *if_outgoing,
const unsigned char *tt_buff,
enum batadv_dup_status dup_status)
{
+ struct batadv_neigh_ifinfo *neigh_ifinfo = NULL;
+ struct batadv_neigh_ifinfo *router_ifinfo = NULL;
struct batadv_neigh_node *neigh_node = NULL, *tmp_neigh_node = NULL;
struct batadv_neigh_node *router = NULL;
struct batadv_orig_node *orig_node_tmp;
@@ -931,12 +953,21 @@ batadv_iv_ogm_orig_update(struct batadv_priv *bat_priv,
if (dup_status != BATADV_NO_DUP)
continue;
- spin_lock_bh(&tmp_neigh_node->bat_iv.lq_update_lock);
- batadv_ring_buffer_set(tmp_neigh_node->bat_iv.tq_recv,
- &tmp_neigh_node->bat_iv.tq_index, 0);
- tq_avg = batadv_ring_buffer_avg(tmp_neigh_node->bat_iv.tq_recv);
- tmp_neigh_node->bat_iv.tq_avg = tq_avg;
- spin_unlock_bh(&tmp_neigh_node->bat_iv.lq_update_lock);
+ /* only update the entry for this outgoing interface */
+ neigh_ifinfo = batadv_neigh_ifinfo_get(tmp_neigh_node,
+ if_outgoing);
+ if (!neigh_ifinfo)
+ continue;
+
+ spin_lock_bh(&tmp_neigh_node->ifinfo_lock);
+ batadv_ring_buffer_set(neigh_ifinfo->bat_iv.tq_recv,
+ &neigh_ifinfo->bat_iv.tq_index, 0);
+ tq_avg = batadv_ring_buffer_avg(neigh_ifinfo->bat_iv.tq_recv);
+ neigh_ifinfo->bat_iv.tq_avg = tq_avg;
+ spin_unlock_bh(&tmp_neigh_node->ifinfo_lock);
+
+ batadv_neigh_ifinfo_free_ref(neigh_ifinfo);
+ neigh_ifinfo = NULL;
}
if (!neigh_node) {
@@ -958,20 +989,23 @@ batadv_iv_ogm_orig_update(struct batadv_priv *bat_priv,
"Updating existing last-hop neighbor of originator\n");
rcu_read_unlock();
+ neigh_ifinfo = batadv_neigh_ifinfo_new(neigh_node, if_outgoing);
+ if (!neigh_ifinfo)
+ goto out;
neigh_node->last_seen = jiffies;
- spin_lock_bh(&neigh_node->bat_iv.lq_update_lock);
- batadv_ring_buffer_set(neigh_node->bat_iv.tq_recv,
- &neigh_node->bat_iv.tq_index,
+ spin_lock_bh(&neigh_node->ifinfo_lock);
+ batadv_ring_buffer_set(neigh_ifinfo->bat_iv.tq_recv,
+ &neigh_ifinfo->bat_iv.tq_index,
batadv_ogm_packet->tq);
- tq_avg = batadv_ring_buffer_avg(neigh_node->bat_iv.tq_recv);
- neigh_node->bat_iv.tq_avg = tq_avg;
- spin_unlock_bh(&neigh_node->bat_iv.lq_update_lock);
+ tq_avg = batadv_ring_buffer_avg(neigh_ifinfo->bat_iv.tq_recv);
+ neigh_ifinfo->bat_iv.tq_avg = tq_avg;
+ spin_unlock_bh(&neigh_node->ifinfo_lock);
if (dup_status == BATADV_NO_DUP) {
orig_node->last_ttl = batadv_ogm_packet->ttl;
- neigh_node->last_ttl = batadv_ogm_packet->ttl;
+ neigh_ifinfo->last_ttl = batadv_ogm_packet->ttl;
}
/* if this neighbor already is our next hop there is nothing
@@ -981,14 +1015,23 @@ batadv_iv_ogm_orig_update(struct batadv_priv *bat_priv,
if (router == neigh_node)
goto out;
- /* if this neighbor does not offer a better TQ we won't consider it */
- if (router && (router->bat_iv.tq_avg > neigh_node->bat_iv.tq_avg))
- goto out;
+ if (router) {
+ router_ifinfo = batadv_neigh_ifinfo_get(router, if_outgoing);
+ if (!router_ifinfo)
+ goto out;
+
+ /* if this neighbor does not offer a better TQ we won't
+ * consider it
+ */
+ if (router_ifinfo->bat_iv.tq_avg > neigh_ifinfo->bat_iv.tq_avg)
+ goto out;
+ }
/* if the TQ is the same and the link not more symmetric we
* won't consider it either
*/
- if (router && (neigh_node->bat_iv.tq_avg == router->bat_iv.tq_avg)) {
+ if (router_ifinfo &&
+ (neigh_ifinfo->bat_iv.tq_avg == router_ifinfo->bat_iv.tq_avg)) {
orig_node_tmp = router->orig_node;
spin_lock_bh(&orig_node_tmp->bat_iv.ogm_cnt_lock);
if_num = router->if_incoming->if_num;
@@ -1015,15 +1058,31 @@ out:
batadv_neigh_node_free_ref(neigh_node);
if (router)
batadv_neigh_node_free_ref(router);
+ if (neigh_ifinfo)
+ batadv_neigh_ifinfo_free_ref(neigh_ifinfo);
+ if (router_ifinfo)
+ batadv_neigh_ifinfo_free_ref(router_ifinfo);
}
+/**
+ * batadv_iv_ogm_calc_tq - calculate tq for current received ogm packet
+ * @orig_node: the orig node who originally emitted the ogm packet
+ * @orig_neigh_node: the orig node struct of the neighbor who sent the packet
+ * @batadv_ogm_packet: the ogm packet
+ * @if_incoming: interface where the packet was received
+ * @if_outgoing: interface for which the retransmission should be considered
+ *
+ * Returns 1 if the link can be considered bidirectional, 0 otherwise
+ */
static int batadv_iv_ogm_calc_tq(struct batadv_orig_node *orig_node,
struct batadv_orig_node *orig_neigh_node,
struct batadv_ogm_packet *batadv_ogm_packet,
- struct batadv_hard_iface *if_incoming)
+ struct batadv_hard_iface *if_incoming,
+ struct batadv_hard_iface *if_outgoing)
{
struct batadv_priv *bat_priv = netdev_priv(if_incoming->soft_iface);
struct batadv_neigh_node *neigh_node = NULL, *tmp_neigh_node;
+ struct batadv_neigh_ifinfo *neigh_ifinfo;
uint8_t total_count;
uint8_t orig_eq_count, neigh_rq_count, neigh_rq_inv, tq_own;
unsigned int neigh_rq_inv_cube, neigh_rq_max_cube;
@@ -1068,7 +1127,13 @@ static int batadv_iv_ogm_calc_tq(struct batadv_orig_node *orig_node,
spin_lock_bh(&orig_node->bat_iv.ogm_cnt_lock);
if_num = if_incoming->if_num;
orig_eq_count = orig_neigh_node->bat_iv.bcast_own_sum[if_num];
- neigh_rq_count = neigh_node->bat_iv.real_packet_count;
+ neigh_ifinfo = batadv_neigh_ifinfo_new(neigh_node, if_outgoing);
+ if (neigh_ifinfo) {
+ neigh_rq_count = neigh_ifinfo->bat_iv.real_packet_count;
+ batadv_neigh_ifinfo_free_ref(neigh_ifinfo);
+ } else {
+ neigh_rq_count = 0;
+ }
spin_unlock_bh(&orig_node->bat_iv.ogm_cnt_lock);
/* pay attention to not get a value bigger than 100 % */
@@ -1132,17 +1197,20 @@ out:
* @ethhdr: ethernet header of the packet
* @batadv_ogm_packet: OGM packet to be considered
* @if_incoming: interface on which the OGM packet was received
+ * @if_outgoing: interface for which the retransmission should be considered
*
* Returns duplicate status as enum batadv_dup_status
*/
static enum batadv_dup_status
batadv_iv_ogm_update_seqnos(const struct ethhdr *ethhdr,
const struct batadv_ogm_packet *batadv_ogm_packet,
- const struct batadv_hard_iface *if_incoming)
+ const struct batadv_hard_iface *if_incoming,
+ struct batadv_hard_iface *if_outgoing)
{
struct batadv_priv *bat_priv = netdev_priv(if_incoming->soft_iface);
struct batadv_orig_node *orig_node;
- struct batadv_neigh_node *tmp_neigh_node;
+ struct batadv_neigh_node *neigh_node;
+ struct batadv_neigh_ifinfo *neigh_ifinfo;
int is_dup;
int32_t seq_diff;
int need_update = 0;
@@ -1169,15 +1237,19 @@ batadv_iv_ogm_update_seqnos(const struct ethhdr *ethhdr,
}
rcu_read_lock();
- hlist_for_each_entry_rcu(tmp_neigh_node,
- &orig_node->neigh_list, list) {
- neigh_addr = tmp_neigh_node->addr;
- is_dup = batadv_test_bit(tmp_neigh_node->bat_iv.real_bits,
+ hlist_for_each_entry_rcu(neigh_node, &orig_node->neigh_list, list) {
+ neigh_ifinfo = batadv_neigh_ifinfo_new(neigh_node,
+ if_outgoing);
+ if (!neigh_ifinfo)
+ continue;
+
+ neigh_addr = neigh_node->addr;
+ is_dup = batadv_test_bit(neigh_ifinfo->bat_iv.real_bits,
orig_node->last_real_seqno,
seqno);
if (batadv_compare_eth(neigh_addr, ethhdr->h_source) &&
- tmp_neigh_node->if_incoming == if_incoming) {
+ neigh_node->if_incoming == if_incoming) {
set_mark = 1;
if (is_dup)
ret = BATADV_NEIGH_DUP;
@@ -1188,13 +1260,14 @@ batadv_iv_ogm_update_seqnos(const struct ethhdr *ethhdr,
}
/* if the window moved, set the update flag. */
- bitmap = tmp_neigh_node->bat_iv.real_bits;
+ bitmap = neigh_ifinfo->bat_iv.real_bits;
need_update |= batadv_bit_get_packet(bat_priv, bitmap,
seq_diff, set_mark);
- packet_count = bitmap_weight(tmp_neigh_node->bat_iv.real_bits,
+ packet_count = bitmap_weight(bitmap,
BATADV_TQ_LOCAL_WINDOW_SIZE);
- tmp_neigh_node->bat_iv.real_packet_count = packet_count;
+ neigh_ifinfo->bat_iv.real_packet_count = packet_count;
+ batadv_neigh_ifinfo_free_ref(neigh_ifinfo);
}
rcu_read_unlock();
@@ -1221,6 +1294,7 @@ static void batadv_iv_ogm_process(const struct ethhdr *ethhdr,
struct batadv_orig_node *orig_neigh_node, *orig_node, *orig_node_tmp;
struct batadv_neigh_node *router = NULL, *router_router = NULL;
struct batadv_neigh_node *orig_neigh_router = NULL;
+ struct batadv_neigh_ifinfo *router_ifinfo = NULL;
int has_directlink_flag;
int is_my_addr = 0, is_my_orig = 0, is_my_oldorig = 0;
int is_bidirect;
@@ -1353,7 +1427,8 @@ static void batadv_iv_ogm_process(const struct ethhdr *ethhdr,
return;
dup_status = batadv_iv_ogm_update_seqnos(ethhdr, batadv_ogm_packet,
- if_incoming);
+ if_incoming,
+ BATADV_IF_DEFAULT);
if (dup_status == BATADV_PROTECTED) {
batadv_dbg(BATADV_DBG_BATMAN, bat_priv,
@@ -1372,9 +1447,11 @@ static void batadv_iv_ogm_process(const struct ethhdr *ethhdr,
if (router) {
orig_node_tmp = router->orig_node;
router_router = batadv_orig_node_get_router(orig_node_tmp);
+ router_ifinfo = batadv_neigh_ifinfo_get(router,
+ BATADV_IF_DEFAULT);
}
- if ((router && router->bat_iv.tq_avg != 0) &&
+ if ((router && router_ifinfo->bat_iv.tq_avg != 0) &&
(batadv_compare_eth(router->addr, ethhdr->h_source)))
is_from_best_next_hop = true;
@@ -1420,7 +1497,8 @@ static void batadv_iv_ogm_process(const struct ethhdr *ethhdr,
}
is_bidirect = batadv_iv_ogm_calc_tq(orig_node, orig_neigh_node,
- batadv_ogm_packet, if_incoming);
+ batadv_ogm_packet, if_incoming,
+ BATADV_IF_DEFAULT);
/* update ranking if it is not a duplicate or has the same
* seqno and similar ttl as the non-duplicate
@@ -1431,7 +1509,8 @@ static void batadv_iv_ogm_process(const struct ethhdr *ethhdr,
(sameseq && similar_ttl)))
batadv_iv_ogm_orig_update(bat_priv, orig_node, ethhdr,
batadv_ogm_packet, if_incoming,
- tt_buff, dup_status);
+ BATADV_IF_DEFAULT, tt_buff,
+ dup_status);
/* is single hop (direct) neighbor */
if (is_single_hop_neigh) {
@@ -1527,6 +1606,34 @@ static int batadv_iv_ogm_receive(struct sk_buff *skb,
return NET_RX_SUCCESS;
}
+/* batadv_iv_ogm_orig_print_neigh - print neighbors for the originator table
+ * @orig_node: the orig_node for which the neighbors are printed
+ * @if_outgoing: outgoing interface for these entries
+ * @seq: debugfs table seq_file struct
+ *
+ * Must be called while holding an rcu lock.
+ */
+static void
+batadv_iv_ogm_orig_print_neigh(struct batadv_orig_node *orig_node,
+ struct batadv_hard_iface *if_outgoing,
+ struct seq_file *seq)
+{
+ struct batadv_neigh_node *neigh_node;
+ struct batadv_neigh_ifinfo *n_ifinfo;
+
+ hlist_for_each_entry_rcu(neigh_node, &orig_node->neigh_list, list) {
+ n_ifinfo = batadv_neigh_ifinfo_get(neigh_node, if_outgoing);
+ if (!n_ifinfo)
+ continue;
+
+ seq_printf(seq, " %pM (%3i)",
+ neigh_node->addr,
+ n_ifinfo->bat_iv.tq_avg);
+
+ batadv_neigh_ifinfo_free_ref(n_ifinfo);
+ }
+}
+
/**
* batadv_iv_ogm_orig_print - print the originator table
* @bat_priv: the bat priv with all the soft interface information
@@ -1535,10 +1642,11 @@ static int batadv_iv_ogm_receive(struct sk_buff *skb,
static void batadv_iv_ogm_orig_print(struct batadv_priv *bat_priv,
struct seq_file *seq)
{
- struct batadv_neigh_node *neigh_node, *neigh_node_tmp;
+ struct batadv_neigh_node *neigh_node;
struct batadv_hashtable *hash = bat_priv->orig_hash;
int last_seen_msecs, last_seen_secs;
struct batadv_orig_node *orig_node;
+ struct batadv_neigh_ifinfo *n_ifinfo;
unsigned long last_seen_jiffies;
struct hlist_head *head;
int batman_count = 0;
@@ -1557,7 +1665,12 @@ static void batadv_iv_ogm_orig_print(struct batadv_priv *bat_priv,
if (!neigh_node)
continue;
- if (neigh_node->bat_iv.tq_avg == 0)
+ n_ifinfo = batadv_neigh_ifinfo_get(neigh_node,
+ BATADV_IF_DEFAULT);
+ if (!n_ifinfo)
+ goto next;
+
+ if (n_ifinfo->bat_iv.tq_avg == 0)
goto next;
last_seen_jiffies = jiffies - orig_node->last_seen;
@@ -1567,22 +1680,19 @@ static void batadv_iv_ogm_orig_print(struct batadv_priv *bat_priv,
seq_printf(seq, "%pM %4i.%03is (%3i) %pM [%10s]:",
orig_node->orig, last_seen_secs,
- last_seen_msecs, neigh_node->bat_iv.tq_avg,
+ last_seen_msecs, n_ifinfo->bat_iv.tq_avg,
neigh_node->addr,
neigh_node->if_incoming->net_dev->name);
- hlist_for_each_entry_rcu(neigh_node_tmp,
- &orig_node->neigh_list, list) {
- seq_printf(seq, " %pM (%3i)",
- neigh_node_tmp->addr,
- neigh_node_tmp->bat_iv.tq_avg);
- }
-
+ batadv_iv_ogm_orig_print_neigh(orig_node,
+ BATADV_IF_DEFAULT, seq);
seq_puts(seq, "\n");
batman_count++;
next:
batadv_neigh_node_free_ref(neigh_node);
+ if (n_ifinfo)
+ batadv_neigh_ifinfo_free_ref(n_ifinfo);
}
rcu_read_unlock();
}
@@ -1594,37 +1704,84 @@ next:
/**
* batadv_iv_ogm_neigh_cmp - compare the metrics of two neighbors
* @neigh1: the first neighbor object of the comparison
+ * @if_outgoing1: outgoing interface for the first neighbor
* @neigh2: the second neighbor object of the comparison
+ * @if_outgoing2: outgoing interface for the second neighbor
*
* Returns a value less, equal to or greater than 0 if the metric via neigh1 is
* lower, the same as or higher than the metric via neigh2
*/
static int batadv_iv_ogm_neigh_cmp(struct batadv_neigh_node *neigh1,
- struct batadv_neigh_node *neigh2)
+ struct batadv_hard_iface *if_outgoing1,
+ struct batadv_neigh_node *neigh2,
+ struct batadv_hard_iface *if_outgoing2)
{
+ struct batadv_neigh_ifinfo *neigh1_ifinfo, *neigh2_ifinfo;
uint8_t tq1, tq2;
+ int diff;
- tq1 = neigh1->bat_iv.tq_avg;
- tq2 = neigh2->bat_iv.tq_avg;
+ neigh1_ifinfo = batadv_neigh_ifinfo_get(neigh1, if_outgoing1);
+ neigh2_ifinfo = batadv_neigh_ifinfo_get(neigh2, if_outgoing2);
- return tq1 - tq2;
+ if (!neigh1_ifinfo || !neigh2_ifinfo) {
+ diff = 0;
+ goto out;
+ }
+
+ tq1 = neigh1_ifinfo->bat_iv.tq_avg;
+ tq2 = neigh2_ifinfo->bat_iv.tq_avg;
+ diff = tq1 - tq2;
+
+out:
+ if (neigh1_ifinfo)
+ batadv_neigh_ifinfo_free_ref(neigh1_ifinfo);
+ if (neigh2_ifinfo)
+ batadv_neigh_ifinfo_free_ref(neigh2_ifinfo);
+
+ return diff;
}
/**
* batadv_iv_ogm_neigh_is_eob - check if neigh1 is equally good or better than
* neigh2 from the metric prospective
* @neigh1: the first neighbor object of the comparison
+ * @if_outgoing: outgoing interface for the first neighbor
* @neigh2: the second neighbor object of the comparison
- *
- * Returns true if the metric via neigh1 is equally good or better than the
- * metric via neigh2, false otherwise.
+ * @if_outgoing2: outgoing interface for the second neighbor
+
+ * Returns true if the metric via neigh1 is equally good or better than
+ * the metric via neigh2, false otherwise.
*/
-static bool batadv_iv_ogm_neigh_is_eob(struct batadv_neigh_node *neigh1,
- struct batadv_neigh_node *neigh2)
+static bool
+batadv_iv_ogm_neigh_is_eob(struct batadv_neigh_node *neigh1,
+ struct batadv_hard_iface *if_outgoing1,
+ struct batadv_neigh_node *neigh2,
+ struct batadv_hard_iface *if_outgoing2)
{
- int diff = batadv_iv_ogm_neigh_cmp(neigh1, neigh2);
+ struct batadv_neigh_ifinfo *neigh1_ifinfo, *neigh2_ifinfo;
+ uint8_t tq1, tq2;
+ bool ret;
- return diff > -BATADV_TQ_SIMILARITY_THRESHOLD;
+ neigh1_ifinfo = batadv_neigh_ifinfo_get(neigh1, if_outgoing1);
+ neigh2_ifinfo = batadv_neigh_ifinfo_get(neigh2, if_outgoing2);
+
+ /* we can't say that the metric is better */
+ if (!neigh1_ifinfo || !neigh2_ifinfo) {
+ ret = false;
+ goto out;
+ }
+
+ tq1 = neigh1_ifinfo->bat_iv.tq_avg;
+ tq2 = neigh2_ifinfo->bat_iv.tq_avg;
+ ret = (tq1 - tq2) > -BATADV_TQ_SIMILARITY_THRESHOLD;
+
+out:
+ if (neigh1_ifinfo)
+ batadv_neigh_ifinfo_free_ref(neigh1_ifinfo);
+ if (neigh2_ifinfo)
+ batadv_neigh_ifinfo_free_ref(neigh2_ifinfo);
+
+ return ret;
}
static struct batadv_algo_ops batadv_batman_iv __read_mostly = {
diff --git a/net/batman-adv/gateway_client.c b/net/batman-adv/gateway_client.c
index 4150a64..4e8f5b1 100644
--- a/net/batman-adv/gateway_client.c
+++ b/net/batman-adv/gateway_client.c
@@ -129,6 +129,7 @@ static struct batadv_gw_node *
batadv_gw_get_best_gw_node(struct batadv_priv *bat_priv)
{
struct batadv_neigh_node *router;
+ struct batadv_neigh_ifinfo *router_ifinfo;
struct batadv_gw_node *gw_node, *curr_gw = NULL;
uint32_t max_gw_factor = 0, tmp_gw_factor = 0;
uint32_t gw_divisor;
@@ -149,10 +150,15 @@ batadv_gw_get_best_gw_node(struct batadv_priv *bat_priv)
if (!router)
continue;
+ router_ifinfo = batadv_neigh_ifinfo_get(router,
+ BATADV_IF_DEFAULT);
+ if (!router_ifinfo)
+ goto next;
+
if (!atomic_inc_not_zero(&gw_node->refcount))
goto next;
- tq_avg = router->bat_iv.tq_avg;
+ tq_avg = router_ifinfo->bat_iv.tq_avg;
switch (atomic_read(&bat_priv->gw_sel_class)) {
case 1: /* fast connection */
@@ -197,6 +203,8 @@ batadv_gw_get_best_gw_node(struct batadv_priv *bat_priv)
next:
batadv_neigh_node_free_ref(router);
+ if (router_ifinfo)
+ batadv_neigh_ifinfo_free_ref(router_ifinfo);
}
rcu_read_unlock();
@@ -239,6 +247,7 @@ void batadv_gw_election(struct batadv_priv *bat_priv)
{
struct batadv_gw_node *curr_gw = NULL, *next_gw = NULL;
struct batadv_neigh_node *router = NULL;
+ struct batadv_neigh_ifinfo *router_ifinfo = NULL;
char gw_addr[18] = { '\0' };
if (atomic_read(&bat_priv->gw_mode) != BATADV_GW_MODE_CLIENT)
@@ -262,6 +271,13 @@ void batadv_gw_election(struct batadv_priv *bat_priv)
batadv_gw_reselect(bat_priv);
goto out;
}
+
+ router_ifinfo = batadv_neigh_ifinfo_get(router,
+ BATADV_IF_DEFAULT);
+ if (!router_ifinfo) {
+ batadv_gw_reselect(bat_priv);
+ goto out;
+ }
}
if ((curr_gw) && (!next_gw)) {
@@ -276,7 +292,8 @@ void batadv_gw_election(struct batadv_priv *bat_priv)
next_gw->bandwidth_down / 10,
next_gw->bandwidth_down % 10,
next_gw->bandwidth_up / 10,
- next_gw->bandwidth_up % 10, router->bat_iv.tq_avg);
+ next_gw->bandwidth_up % 10,
+ router_ifinfo->bat_iv.tq_avg);
batadv_throw_uevent(bat_priv, BATADV_UEV_GW, BATADV_UEV_ADD,
gw_addr);
} else {
@@ -286,7 +303,8 @@ void batadv_gw_election(struct batadv_priv *bat_priv)
next_gw->bandwidth_down / 10,
next_gw->bandwidth_down % 10,
next_gw->bandwidth_up / 10,
- next_gw->bandwidth_up % 10, router->bat_iv.tq_avg);
+ next_gw->bandwidth_up % 10,
+ router_ifinfo->bat_iv.tq_avg);
batadv_throw_uevent(bat_priv, BATADV_UEV_GW, BATADV_UEV_CHANGE,
gw_addr);
}
@@ -300,11 +318,15 @@ out:
batadv_gw_node_free_ref(next_gw);
if (router)
batadv_neigh_node_free_ref(router);
+ if (router_ifinfo)
+ batadv_neigh_ifinfo_free_ref(router_ifinfo);
}
void batadv_gw_check_election(struct batadv_priv *bat_priv,
struct batadv_orig_node *orig_node)
{
+ struct batadv_neigh_ifinfo *router_orig_tq = NULL;
+ struct batadv_neigh_ifinfo *router_gw_tq = NULL;
struct batadv_orig_node *curr_gw_orig;
struct batadv_neigh_node *router_gw = NULL, *router_orig = NULL;
uint8_t gw_tq_avg, orig_tq_avg;
@@ -317,6 +339,11 @@ void batadv_gw_check_election(struct batadv_priv *bat_priv,
if (!router_gw)
goto reselect;
+ router_gw_tq = batadv_neigh_ifinfo_get(router_gw,
+ BATADV_IF_DEFAULT);
+ if (!router_gw_tq)
+ goto reselect;
+
/* this node already is the gateway */
if (curr_gw_orig == orig_node)
goto out;
@@ -325,8 +352,13 @@ void batadv_gw_check_election(struct batadv_priv *bat_priv,
if (!router_orig)
goto out;
- gw_tq_avg = router_gw->bat_iv.tq_avg;
- orig_tq_avg = router_orig->bat_iv.tq_avg;
+ router_orig_tq = batadv_neigh_ifinfo_get(router_orig,
+ BATADV_IF_DEFAULT);
+ if (!router_orig_tq)
+ goto out;
+
+ gw_tq_avg = router_gw_tq->bat_iv.tq_avg;
+ orig_tq_avg = router_orig_tq->bat_iv.tq_avg;
/* the TQ value has to be better */
if (orig_tq_avg < gw_tq_avg)
@@ -352,6 +384,10 @@ out:
batadv_neigh_node_free_ref(router_gw);
if (router_orig)
batadv_neigh_node_free_ref(router_orig);
+ if (router_gw_tq)
+ batadv_neigh_ifinfo_free_ref(router_gw_tq);
+ if (router_orig_tq)
+ batadv_neigh_ifinfo_free_ref(router_orig_tq);
return;
}
@@ -537,28 +573,36 @@ static int batadv_write_buffer_text(struct batadv_priv *bat_priv,
{
struct batadv_gw_node *curr_gw;
struct batadv_neigh_node *router;
+ struct batadv_neigh_ifinfo *router_ifinfo = NULL;
int ret = -1;
router = batadv_orig_node_get_router(gw_node->orig_node);
if (!router)
goto out;
+ router_ifinfo = batadv_neigh_ifinfo_get(router, BATADV_IF_DEFAULT);
+ if (!router_ifinfo)
+ goto out;
+
curr_gw = batadv_gw_get_selected_gw_node(bat_priv);
ret = seq_printf(seq, "%s %pM (%3i) %pM [%10s]: %u.%u/%u.%u MBit\n",
(curr_gw == gw_node ? "=>" : " "),
gw_node->orig_node->orig,
- router->bat_iv.tq_avg, router->addr,
+ router_ifinfo->bat_iv.tq_avg, router->addr,
router->if_incoming->net_dev->name,
gw_node->bandwidth_down / 10,
gw_node->bandwidth_down % 10,
gw_node->bandwidth_up / 10,
gw_node->bandwidth_up % 10);
- batadv_neigh_node_free_ref(router);
if (curr_gw)
batadv_gw_node_free_ref(curr_gw);
out:
+ if (router_ifinfo)
+ batadv_neigh_ifinfo_free_ref(router_ifinfo);
+ if (router)
+ batadv_neigh_node_free_ref(router);
return ret;
}
@@ -746,6 +790,7 @@ bool batadv_gw_out_of_range(struct batadv_priv *bat_priv,
struct batadv_neigh_node *neigh_curr = NULL, *neigh_old = NULL;
struct batadv_orig_node *orig_dst_node = NULL;
struct batadv_gw_node *gw_node = NULL, *curr_gw = NULL;
+ struct batadv_neigh_ifinfo *curr_ifinfo, *old_ifinfo;
struct ethhdr *ethhdr = (struct ethhdr *)skb->data;
bool out_of_range = false;
uint8_t curr_tq_avg;
@@ -787,7 +832,14 @@ bool batadv_gw_out_of_range(struct batadv_priv *bat_priv,
if (!neigh_curr)
goto out;
- curr_tq_avg = neigh_curr->bat_iv.tq_avg;
+ curr_ifinfo = batadv_neigh_ifinfo_get(neigh_curr,
+ BATADV_IF_DEFAULT);
+ if (!curr_ifinfo)
+ goto out;
+
+ curr_tq_avg = curr_ifinfo->bat_iv.tq_avg;
+ batadv_neigh_ifinfo_free_ref(curr_ifinfo);
+
break;
case BATADV_GW_MODE_OFF:
default:
@@ -798,8 +850,13 @@ bool batadv_gw_out_of_range(struct batadv_priv *bat_priv,
if (!neigh_old)
goto out;
- if (curr_tq_avg - neigh_old->bat_iv.tq_avg > BATADV_GW_THRESHOLD)
+ old_ifinfo = batadv_neigh_ifinfo_get(neigh_old, BATADV_IF_DEFAULT);
+ if (!old_ifinfo)
+ goto out;
+
+ if ((curr_tq_avg - old_ifinfo->bat_iv.tq_avg) > BATADV_GW_THRESHOLD)
out_of_range = true;
+ batadv_neigh_ifinfo_free_ref(old_ifinfo);
out:
if (orig_dst_node)
diff --git a/net/batman-adv/hard-interface.h b/net/batman-adv/hard-interface.h
index d8ca8ac..8d4304b 100644
--- a/net/batman-adv/hard-interface.h
+++ b/net/batman-adv/hard-interface.h
@@ -51,6 +51,11 @@ int batadv_hardif_min_mtu(struct net_device *soft_iface);
void batadv_update_min_mtu(struct net_device *soft_iface);
void batadv_hardif_free_rcu(struct rcu_head *rcu);
+/**
+ * batadv_hardif_free_ref - decrement the hard interface refcounter and
+ * possibly free it
+ * @hard_iface: the hard interface to free
+ */
static inline void
batadv_hardif_free_ref(struct batadv_hard_iface *hard_iface)
{
@@ -58,6 +63,18 @@ batadv_hardif_free_ref(struct batadv_hard_iface *hard_iface)
call_rcu(&hard_iface->rcu, batadv_hardif_free_rcu);
}
+/**
+ * batadv_hardif_free_ref_now - decrement the hard interface refcounter and
+ * possibly free it (without rcu callback)
+ * @hard_iface: the hard interface to free
+ */
+static inline void
+batadv_hardif_free_ref_now(struct batadv_hard_iface *hard_iface)
+{
+ if (atomic_dec_and_test(&hard_iface->refcount))
+ batadv_hardif_free_rcu(&hard_iface->rcu);
+}
+
static inline struct batadv_hard_iface *
batadv_primary_if_get_selected(struct batadv_priv *bat_priv)
{
diff --git a/net/batman-adv/main.h b/net/batman-adv/main.h
index a468760..49a2394 100644
--- a/net/batman-adv/main.h
+++ b/net/batman-adv/main.h
@@ -72,6 +72,12 @@
#define BATADV_NO_MARK 0
+/* default interface for multi interface operation. The default interface is
+ * used for communication which originated locally (i.e. is not forwarded)
+ * or where special forwarding is not desired/necessary.
+ */
+#define BATADV_IF_DEFAULT ((struct batadv_hard_iface *)NULL)
+
#define BATADV_NUM_WORDS BITS_TO_LONGS(BATADV_TQ_LOCAL_WINDOW_SIZE)
#define BATADV_LOG_BUF_LEN 8192 /* has to be a power of 2 */
diff --git a/net/batman-adv/network-coding.c b/net/batman-adv/network-coding.c
index d6f0be3..aec3fa5 100644
--- a/net/batman-adv/network-coding.c
+++ b/net/batman-adv/network-coding.c
@@ -1008,6 +1008,8 @@ static bool batadv_nc_code_packets(struct batadv_priv *bat_priv,
struct batadv_coded_packet *coded_packet;
struct batadv_neigh_node *neigh_tmp, *router_neigh;
struct batadv_neigh_node *router_coding = NULL;
+ struct batadv_neigh_ifinfo *router_neigh_ifinfo = NULL;
+ struct batadv_neigh_ifinfo *router_coding_ifinfo = NULL;
uint8_t *first_source, *first_dest, *second_source, *second_dest;
__be32 packet_id1, packet_id2;
size_t count;
@@ -1021,15 +1023,25 @@ static bool batadv_nc_code_packets(struct batadv_priv *bat_priv,
if (!router_neigh)
goto out;
+ router_neigh_ifinfo = batadv_neigh_ifinfo_get(router_neigh,
+ BATADV_IF_DEFAULT);
+ if (!router_neigh_ifinfo)
+ goto out;
+
neigh_tmp = nc_packet->neigh_node;
router_coding = batadv_orig_node_get_router(neigh_tmp->orig_node);
if (!router_coding)
goto out;
- tq_tmp = batadv_nc_random_weight_tq(router_neigh->bat_iv.tq_avg);
- tq_weighted_neigh = tq_tmp;
- tq_tmp = batadv_nc_random_weight_tq(router_coding->bat_iv.tq_avg);
- tq_weighted_coding = tq_tmp;
+ router_coding_ifinfo = batadv_neigh_ifinfo_get(router_coding,
+ BATADV_IF_DEFAULT);
+ if (!router_coding_ifinfo)
+ goto out;
+
+ tq_tmp = router_neigh_ifinfo->bat_iv.tq_avg;
+ tq_weighted_neigh = batadv_nc_random_weight_tq(tq_tmp);
+ tq_tmp = router_coding_ifinfo->bat_iv.tq_avg;
+ tq_weighted_coding = batadv_nc_random_weight_tq(tq_tmp);
/* Select one destination for the MAC-header dst-field based on
* weighted TQ-values.
@@ -1153,6 +1165,10 @@ out:
batadv_neigh_node_free_ref(router_neigh);
if (router_coding)
batadv_neigh_node_free_ref(router_coding);
+ if (router_neigh_ifinfo)
+ batadv_neigh_ifinfo_free_ref(router_neigh_ifinfo);
+ if (router_coding_ifinfo)
+ batadv_neigh_ifinfo_free_ref(router_coding_ifinfo);
return res;
}
diff --git a/net/batman-adv/originator.c b/net/batman-adv/originator.c
index 03db87f..1a4725f 100644
--- a/net/batman-adv/originator.c
+++ b/net/batman-adv/originator.c
@@ -148,10 +148,87 @@ err:
return -ENOMEM;
}
+/**
+ * batadv_neigh_ifinfo_free_rcu - free the neigh_ifinfo object
+ * @rcu: rcu pointer of the neigh_ifinfo object
+ */
+static void batadv_neigh_ifinfo_free_rcu(struct rcu_head *rcu)
+{
+ struct batadv_neigh_ifinfo *neigh_ifinfo;
+
+ neigh_ifinfo = container_of(rcu, struct batadv_neigh_ifinfo, rcu);
+
+ if (neigh_ifinfo->if_outgoing != BATADV_IF_DEFAULT)
+ batadv_hardif_free_ref_now(neigh_ifinfo->if_outgoing);
+
+ kfree(neigh_ifinfo);
+}
+
+/**
+ * batadv_neigh_ifinfo_free_now - decrement the refcounter and possibly free
+ * the neigh_ifinfo (without rcu callback)
+ * @neigh_ifinfo: the neigh_ifinfo object to release
+ */
+static void
+batadv_neigh_ifinfo_free_ref_now(struct batadv_neigh_ifinfo *neigh_ifinfo)
+{
+ if (atomic_dec_and_test(&neigh_ifinfo->refcount))
+ batadv_neigh_ifinfo_free_rcu(&neigh_ifinfo->rcu);
+}
+
+/**
+ * batadv_neigh_ifinfo_free_ref - decrement the refcounter and possibly free
+ * the neigh_ifinfo
+ * @neigh_ifinfo: the neigh_ifinfo object to release
+ */
+void batadv_neigh_ifinfo_free_ref(struct batadv_neigh_ifinfo *neigh_ifinfo)
+{
+ if (atomic_dec_and_test(&neigh_ifinfo->refcount))
+ call_rcu(&neigh_ifinfo->rcu, batadv_neigh_ifinfo_free_rcu);
+}
+
+/**
+ * batadv_neigh_node_free_rcu - free the neigh_node
+ * @rcu: rcu pointer of the neigh_node
+ */
+static void batadv_neigh_node_free_rcu(struct rcu_head *rcu)
+{
+ struct hlist_node *node_tmp;
+ struct batadv_neigh_node *neigh_node;
+ struct batadv_neigh_ifinfo *neigh_ifinfo;
+
+ neigh_node = container_of(rcu, struct batadv_neigh_node, rcu);
+
+ hlist_for_each_entry_safe(neigh_ifinfo, node_tmp,
+ &neigh_node->ifinfo_list, list) {
+ batadv_neigh_ifinfo_free_ref_now(neigh_ifinfo);
+ }
+ batadv_hardif_free_ref_now(neigh_node->if_incoming);
+
+ kfree(neigh_node);
+}
+
+/**
+ * batadv_neigh_node_free_ref_now - decrement the neighbors refcounter
+ * and possibly free it (without rcu callback)
+ * @neigh_node: neigh neighbor to free
+ */
+static void
+batadv_neigh_node_free_ref_now(struct batadv_neigh_node *neigh_node)
+{
+ if (atomic_dec_and_test(&neigh_node->refcount))
+ batadv_neigh_node_free_rcu(&neigh_node->rcu);
+}
+
+/**
+ * batadv_neigh_node_free_ref - decrement the neighbors refcounter
+ * and possibly free it
+ * @neigh_node: neigh neighbor to free
+ */
void batadv_neigh_node_free_ref(struct batadv_neigh_node *neigh_node)
{
if (atomic_dec_and_test(&neigh_node->refcount))
- kfree_rcu(neigh_node, rcu);
+ call_rcu(&neigh_node->rcu, batadv_neigh_node_free_rcu);
}
/* increases the refcounter of a found router */
@@ -171,6 +248,84 @@ batadv_orig_node_get_router(struct batadv_orig_node *orig_node)
}
/**
+ * batadv_neigh_ifinfo_get - find the ifinfo from an neigh_node
+ * @neigh_node: the neigh node to be queried
+ * @if_outgoing: the interface for which the ifinfo should be acquired
+ *
+ * The object is returned with refcounter increased by 1.
+ *
+ * Returns the requested neigh_ifinfo or NULL if not found
+ */
+struct batadv_neigh_ifinfo *
+batadv_neigh_ifinfo_get(struct batadv_neigh_node *neigh,
+ struct batadv_hard_iface *if_outgoing)
+{
+ struct batadv_neigh_ifinfo *neigh_ifinfo = NULL,
+ *tmp_neigh_ifinfo;
+
+ rcu_read_lock();
+ hlist_for_each_entry_rcu(tmp_neigh_ifinfo, &neigh->ifinfo_list,
+ list) {
+ if (tmp_neigh_ifinfo->if_outgoing != if_outgoing)
+ continue;
+
+ if (!atomic_inc_not_zero(&tmp_neigh_ifinfo->refcount))
+ continue;
+
+ neigh_ifinfo = tmp_neigh_ifinfo;
+ break;
+ }
+ rcu_read_unlock();
+
+ return neigh_ifinfo;
+}
+
+/**
+ * batadv_neigh_ifinfo_new - search and possibly create an neigh_ifinfo object
+ * @neigh_node: the neigh node to be queried
+ * @if_outgoing: the interface for which the ifinfo should be acquired
+ *
+ * Returns NULL in case of failure or the neigh_ifinfo object for the
+ * if_outgoing interface otherwise. The object is created and added to the list
+ * if it does not exist.
+ *
+ * The object is returned with refcounter increased by 1.
+ */
+struct batadv_neigh_ifinfo *
+batadv_neigh_ifinfo_new(struct batadv_neigh_node *neigh,
+ struct batadv_hard_iface *if_outgoing)
+{
+ struct batadv_neigh_ifinfo *neigh_ifinfo;
+
+ spin_lock_bh(&neigh->ifinfo_lock);
+
+ neigh_ifinfo = batadv_neigh_ifinfo_get(neigh, if_outgoing);
+ if (neigh_ifinfo)
+ goto out;
+
+ neigh_ifinfo = kzalloc(sizeof(*neigh_ifinfo), GFP_ATOMIC);
+ if (!neigh_ifinfo)
+ goto out;
+
+ if (if_outgoing && !atomic_inc_not_zero(&if_outgoing->refcount)) {
+ kfree(neigh_ifinfo);
+ neigh_ifinfo = NULL;
+ goto out;
+ }
+
+ INIT_HLIST_NODE(&neigh_ifinfo->list);
+ atomic_set(&neigh_ifinfo->refcount, 2);
+ neigh_ifinfo->if_outgoing = if_outgoing;
+
+ hlist_add_head_rcu(&neigh_ifinfo->list, &neigh->ifinfo_list);
+
+out:
+ spin_unlock_bh(&neigh->ifinfo_lock);
+
+ return neigh_ifinfo;
+}
+
+/**
* batadv_neigh_node_new - create and init a new neigh_node object
* @hard_iface: the interface where the neighbour is connected to
* @neigh_addr: the mac address of the neighbour interface
@@ -191,6 +346,8 @@ batadv_neigh_node_new(struct batadv_hard_iface *hard_iface,
goto out;
INIT_HLIST_NODE(&neigh_node->list);
+ INIT_HLIST_HEAD(&neigh_node->ifinfo_list);
+ spin_lock_init(&neigh_node->ifinfo_lock);
memcpy(neigh_node->addr, neigh_addr, ETH_ALEN);
neigh_node->if_incoming = hard_iface;
@@ -217,7 +374,7 @@ static void batadv_orig_node_free_rcu(struct rcu_head *rcu)
hlist_for_each_entry_safe(neigh_node, node_tmp,
&orig_node->neigh_list, list) {
hlist_del_rcu(&neigh_node->list);
- batadv_neigh_node_free_ref(neigh_node);
+ batadv_neigh_node_free_ref_now(neigh_node);
}
spin_unlock_bh(&orig_node->neigh_list_lock);
@@ -362,20 +519,23 @@ free_orig_node:
return NULL;
}
+/**
+ * batadv_purge_orig_neighbors - purges neighbors from originator
+ * @bat_priv: the bat priv with all the soft interface information
+ * @orig_node: orig node which is to be checked
+ *
+ * Returns true if any neighbor was purged, false otherwise
+ */
static bool
batadv_purge_orig_neighbors(struct batadv_priv *bat_priv,
- struct batadv_orig_node *orig_node,
- struct batadv_neigh_node **best_neigh)
+ struct batadv_orig_node *orig_node)
{
- struct batadv_algo_ops *bao = bat_priv->bat_algo_ops;
struct hlist_node *node_tmp;
struct batadv_neigh_node *neigh_node;
bool neigh_purged = false;
unsigned long last_seen;
struct batadv_hard_iface *if_incoming;
- *best_neigh = NULL;
-
spin_lock_bh(&orig_node->neigh_list_lock);
/* for all neighbors towards this originator ... */
@@ -405,13 +565,6 @@ batadv_purge_orig_neighbors(struct batadv_priv *bat_priv,
hlist_del_rcu(&neigh_node->list);
batadv_neigh_node_free_ref(neigh_node);
- } else {
- /* store the best_neighbour if this is the first
- * iteration or if a better neighbor has been found
- */
- if (!*best_neigh ||
- bao->bat_neigh_cmp(neigh_node, *best_neigh) > 0)
- *best_neigh = neigh_node;
}
}
@@ -419,6 +572,41 @@ batadv_purge_orig_neighbors(struct batadv_priv *bat_priv,
return neigh_purged;
}
+/**
+ * batadv_find_best_neighbor - finds the best neighbor after purging
+ * @bat_priv: the bat priv with all the soft interface information
+ * @orig_node: orig node which is to be checked
+ * @if_outgoing: the interface for which the metric should be compared
+ *
+ * Returns the current best neighbor, with refcount increased.
+ */
+static struct batadv_neigh_node *
+batadv_find_best_neighbor(struct batadv_priv *bat_priv,
+ struct batadv_orig_node *orig_node,
+ struct batadv_hard_iface *if_outgoing)
+{
+ struct batadv_neigh_node *best = NULL, *neigh;
+ struct batadv_algo_ops *bao = bat_priv->bat_algo_ops;
+
+ rcu_read_lock();
+ hlist_for_each_entry_rcu(neigh, &orig_node->neigh_list, list) {
+ if (best && (bao->bat_neigh_cmp(neigh, if_outgoing,
+ best, if_outgoing) <= 0))
+ continue;
+
+ if (!atomic_inc_not_zero(&neigh->refcount))
+ continue;
+
+ if (best)
+ batadv_neigh_node_free_ref(best);
+
+ best = neigh;
+ }
+ rcu_read_unlock();
+
+ return best;
+}
+
static bool batadv_purge_orig_node(struct batadv_priv *bat_priv,
struct batadv_orig_node *orig_node)
{
@@ -431,12 +619,15 @@ static bool batadv_purge_orig_node(struct batadv_priv *bat_priv,
orig_node->orig,
jiffies_to_msecs(orig_node->last_seen));
return true;
- } else {
- if (batadv_purge_orig_neighbors(bat_priv, orig_node,
- &best_neigh_node))
- batadv_update_route(bat_priv, orig_node,
- best_neigh_node);
}
+ if (!batadv_purge_orig_neighbors(bat_priv, orig_node))
+ return false;
+
+ best_neigh_node = batadv_find_best_neighbor(bat_priv, orig_node,
+ BATADV_IF_DEFAULT);
+ batadv_update_route(bat_priv, orig_node, best_neigh_node);
+ if (best_neigh_node)
+ batadv_neigh_node_free_ref(best_neigh_node);
return false;
}
diff --git a/net/batman-adv/originator.h b/net/batman-adv/originator.h
index 99ed3c6..29fa4c4 100644
--- a/net/batman-adv/originator.h
+++ b/net/batman-adv/originator.h
@@ -35,6 +35,13 @@ batadv_neigh_node_new(struct batadv_hard_iface *hard_iface,
void batadv_neigh_node_free_ref(struct batadv_neigh_node *neigh_node);
struct batadv_neigh_node *
batadv_orig_node_get_router(struct batadv_orig_node *orig_node);
+struct batadv_neigh_ifinfo *
+batadv_neigh_ifinfo_new(struct batadv_neigh_node *neigh,
+ struct batadv_hard_iface *if_outgoing);
+struct batadv_neigh_ifinfo *
+batadv_neigh_ifinfo_get(struct batadv_neigh_node *neigh,
+ struct batadv_hard_iface *if_outgoing);
+void batadv_neigh_ifinfo_free_ref(struct batadv_neigh_ifinfo *neigh_ifinfo);
int batadv_orig_seq_print_text(struct seq_file *seq, void *offset);
int batadv_orig_hash_add_if(struct batadv_hard_iface *hard_iface,
int max_if_num);
diff --git a/net/batman-adv/translation-table.c b/net/batman-adv/translation-table.c
index 63d2570..22c32ae 100644
--- a/net/batman-adv/translation-table.c
+++ b/net/batman-adv/translation-table.c
@@ -1405,7 +1405,8 @@ batadv_transtable_best_orig(struct batadv_priv *bat_priv,
continue;
if (best_router &&
- bao->bat_neigh_cmp(router, best_router) <= 0) {
+ bao->bat_neigh_cmp(router, BATADV_IF_DEFAULT,
+ best_router, BATADV_IF_DEFAULT) <= 0) {
batadv_neigh_node_free_ref(router);
continue;
}
diff --git a/net/batman-adv/types.h b/net/batman-adv/types.h
index 0f68845..8435d7c 100644
--- a/net/batman-adv/types.h
+++ b/net/batman-adv/types.h
@@ -289,47 +289,62 @@ struct batadv_gw_node {
};
/**
- * struct batadv_neigh_bat_iv - B.A.T.M.A.N. IV specific structure for single
- * hop neighbors
+ * struct batadv_neigh_node - structure for single hops neighbors
+ * @list: list node for batadv_orig_node::neigh_list
+ * @orig_node: pointer to corresponding orig_node
+ * @addr: the MAC address of the neighboring interface
+ * @ifinfo_list: list for routing metrics per outgoing interface
+ * @ifinfo_lock: lock protecting private ifinfo members and list
+ * @if_incoming: pointer to incoming hard interface
+ * @last_seen: when last packet via this neighbor was received
+ * @last_ttl: last received ttl from this neigh node
+ * @rcu: struct used for freeing in an RCU-safe manner
+ * @bat_iv: B.A.T.M.A.N. IV private structure
+ */
+struct batadv_neigh_node {
+ struct hlist_node list;
+ struct batadv_orig_node *orig_node;
+ uint8_t addr[ETH_ALEN];
+ struct hlist_head ifinfo_list;
+ spinlock_t ifinfo_lock; /* protects ifinfo_list and its members */
+ struct batadv_hard_iface *if_incoming;
+ unsigned long last_seen;
+ atomic_t refcount;
+ struct rcu_head rcu;
+};
+
+/* struct batadv_neigh_node_bat_iv - neighbor information per outgoing
+ * interface for BATMAN IV
* @tq_recv: ring buffer of received TQ values from this neigh node
* @tq_index: ring buffer index
* @tq_avg: averaged tq of all tq values in the ring buffer (tq_recv)
* @real_bits: bitfield containing the number of OGMs received from this neigh
* node (relative to orig_node->last_real_seqno)
* @real_packet_count: counted result of real_bits
- * @lq_update_lock: lock protecting tq_recv & tq_index
*/
-struct batadv_neigh_bat_iv {
+struct batadv_neigh_ifinfo_bat_iv {
uint8_t tq_recv[BATADV_TQ_GLOBAL_WINDOW_SIZE];
uint8_t tq_index;
uint8_t tq_avg;
DECLARE_BITMAP(real_bits, BATADV_TQ_LOCAL_WINDOW_SIZE);
uint8_t real_packet_count;
- spinlock_t lq_update_lock; /* protects tq_recv & tq_index */
};
-/**
- * struct batadv_neigh_node - structure for single hops neighbors
- * @list: list node for batadv_orig_node::neigh_list
- * @orig_node: pointer to corresponding orig_node
- * @addr: the MAC address of the neighboring interface
- * @if_incoming: pointer to incoming hard interface
- * @last_seen: when last packet via this neighbor was received
+/* struct batadv_neigh_ifinfo - neighbor information per outgoing interface
+ * @list: list node for batadv_neigh_node::ifinfo_list
+ * @if_outgoing: pointer to outgoing hard interface
+ * @bat_iv: B.A.T.M.A.N. IV private structure
* @last_ttl: last received ttl from this neigh node
* @refcount: number of contexts the object is used
- * @rcu: struct used for freeing in an RCU-safe manner
- * @bat_iv: B.A.T.M.A.N. IV private structure
+ * @rcu: struct used for freeing in a RCU-safe manner
*/
-struct batadv_neigh_node {
+struct batadv_neigh_ifinfo {
struct hlist_node list;
- struct batadv_orig_node *orig_node;
- uint8_t addr[ETH_ALEN];
- struct batadv_hard_iface *if_incoming;
- unsigned long last_seen;
+ struct batadv_hard_iface *if_outgoing;
+ struct batadv_neigh_ifinfo_bat_iv bat_iv;
uint8_t last_ttl;
atomic_t refcount;
struct rcu_head rcu;
- struct batadv_neigh_bat_iv bat_iv;
};
/**
@@ -1013,9 +1028,11 @@ struct batadv_forw_packet {
* @bat_primary_iface_set: called when primary interface is selected / changed
* @bat_ogm_schedule: prepare a new outgoing OGM for the send queue
* @bat_ogm_emit: send scheduled OGM
- * @bat_neigh_cmp: compare the metrics of two neighbors
- * @bat_neigh_is_equiv_or_better: check if neigh1 is equally good or
- * better than neigh2 from the metric prospective
+ * @bat_neigh_cmp: compare the metrics of two neighbors for their respective
+ * outgoing interfaces
+ * @bat_neigh_is_equiv_or_better: check if neigh1 is equally good or better
+ * than neigh2 for their respective outgoing interface from the metric
+ * prospective
* @bat_orig_print: print the originator table (optional)
* @bat_orig_free: free the resources allocated by the routing algorithm for an
* orig_node object
@@ -1034,9 +1051,14 @@ struct batadv_algo_ops {
void (*bat_ogm_schedule)(struct batadv_hard_iface *hard_iface);
void (*bat_ogm_emit)(struct batadv_forw_packet *forw_packet);
int (*bat_neigh_cmp)(struct batadv_neigh_node *neigh1,
- struct batadv_neigh_node *neigh2);
- bool (*bat_neigh_is_equiv_or_better)(struct batadv_neigh_node *neigh1,
- struct batadv_neigh_node *neigh2);
+ struct batadv_hard_iface *if_outgoing1,
+ struct batadv_neigh_node *neigh2,
+ struct batadv_hard_iface *if_outgoing2);
+ bool (*bat_neigh_is_equiv_or_better)
+ (struct batadv_neigh_node *neigh1,
+ struct batadv_hard_iface *if_outgoing1,
+ struct batadv_neigh_node *neigh2,
+ struct batadv_hard_iface *if_outgoing2);
/* orig_node handling API */
void (*bat_orig_print)(struct batadv_priv *priv, struct seq_file *seq);
void (*bat_orig_free)(struct batadv_orig_node *orig_node);
--
1.8.5.2
^ permalink raw reply related
* [PATCH 01/14] batman-adv: remove bonding and interface alternating
From: Antonio Quartulli @ 2014-01-13 22:31 UTC (permalink / raw)
To: davem
Cc: netdev, b.a.t.m.a.n, Simon Wunderlich, Marek Lindner,
Antonio Quartulli
In-Reply-To: <1389652298-472-1-git-send-email-antonio@meshcoding.com>
From: Simon Wunderlich <simon@open-mesh.com>
Remove bonding and interface alternating code - it will be replaced
by a new, network-wide multi interface optimization which enables
both bonding and interface alternating in a better way.
Keep the sysfs and find router function though, this will be needed
later.
Signed-off-by: Simon Wunderlich <simon@open-mesh.com>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>
---
net/batman-adv/bat_iv_ogm.c | 5 -
net/batman-adv/originator.c | 15 +--
net/batman-adv/routing.c | 300 ++------------------------------------------
net/batman-adv/routing.h | 9 --
net/batman-adv/types.h | 10 +-
5 files changed, 12 insertions(+), 327 deletions(-)
diff --git a/net/batman-adv/bat_iv_ogm.c b/net/batman-adv/bat_iv_ogm.c
index 278d78b..48f5d35 100644
--- a/net/batman-adv/bat_iv_ogm.c
+++ b/net/batman-adv/bat_iv_ogm.c
@@ -974,8 +974,6 @@ batadv_iv_ogm_orig_update(struct batadv_priv *bat_priv,
neigh_node->last_ttl = batadv_ogm_packet->ttl;
}
- batadv_bonding_candidate_add(bat_priv, orig_node, neigh_node);
-
/* if this neighbor already is our next hop there is nothing
* to change
*/
@@ -1424,9 +1422,6 @@ static void batadv_iv_ogm_process(const struct ethhdr *ethhdr,
is_bidirect = batadv_iv_ogm_calc_tq(orig_node, orig_neigh_node,
batadv_ogm_packet, if_incoming);
- batadv_bonding_save_primary(orig_node, orig_neigh_node,
- batadv_ogm_packet);
-
/* update ranking if it is not a duplicate or has the same
* seqno and similar ttl as the non-duplicate
*/
diff --git a/net/batman-adv/originator.c b/net/batman-adv/originator.c
index 04a6d51..03db87f 100644
--- a/net/batman-adv/originator.c
+++ b/net/batman-adv/originator.c
@@ -196,8 +196,6 @@ batadv_neigh_node_new(struct batadv_hard_iface *hard_iface,
neigh_node->if_incoming = hard_iface;
neigh_node->orig_node = orig_node;
- INIT_LIST_HEAD(&neigh_node->bonding_list);
-
/* extra reference for return */
atomic_set(&neigh_node->refcount, 2);
@@ -208,20 +206,13 @@ out:
static void batadv_orig_node_free_rcu(struct rcu_head *rcu)
{
struct hlist_node *node_tmp;
- struct batadv_neigh_node *neigh_node, *tmp_neigh_node;
+ struct batadv_neigh_node *neigh_node;
struct batadv_orig_node *orig_node;
orig_node = container_of(rcu, struct batadv_orig_node, rcu);
spin_lock_bh(&orig_node->neigh_list_lock);
- /* for all bonding members ... */
- list_for_each_entry_safe(neigh_node, tmp_neigh_node,
- &orig_node->bond_list, bonding_list) {
- list_del_rcu(&neigh_node->bonding_list);
- batadv_neigh_node_free_ref(neigh_node);
- }
-
/* for all neighbors towards this originator ... */
hlist_for_each_entry_safe(neigh_node, node_tmp,
&orig_node->neigh_list, list) {
@@ -325,7 +316,6 @@ struct batadv_orig_node *batadv_orig_node_new(struct batadv_priv *bat_priv,
return NULL;
INIT_HLIST_HEAD(&orig_node->neigh_list);
- INIT_LIST_HEAD(&orig_node->bond_list);
INIT_LIST_HEAD(&orig_node->vlan_list);
spin_lock_init(&orig_node->bcast_seqno_lock);
spin_lock_init(&orig_node->neigh_list_lock);
@@ -350,8 +340,6 @@ struct batadv_orig_node *batadv_orig_node_new(struct batadv_priv *bat_priv,
orig_node->bcast_seqno_reset = reset_time;
orig_node->batman_seqno_reset = reset_time;
- atomic_set(&orig_node->bond_candidates, 0);
-
/* create a vlan object for the "untagged" LAN */
vlan = batadv_orig_node_vlan_new(orig_node, BATADV_NO_FLAGS);
if (!vlan)
@@ -416,7 +404,6 @@ batadv_purge_orig_neighbors(struct batadv_priv *bat_priv,
neigh_purged = true;
hlist_del_rcu(&neigh_node->list);
- batadv_bonding_candidate_del(orig_node, neigh_node);
batadv_neigh_node_free_ref(neigh_node);
} else {
/* store the best_neighbour if this is the first
diff --git a/net/batman-adv/routing.c b/net/batman-adv/routing.c
index 7e8aa17..4fd2687 100644
--- a/net/batman-adv/routing.c
+++ b/net/batman-adv/routing.c
@@ -96,115 +96,6 @@ out:
batadv_neigh_node_free_ref(router);
}
-/* caller must hold the neigh_list_lock */
-void batadv_bonding_candidate_del(struct batadv_orig_node *orig_node,
- struct batadv_neigh_node *neigh_node)
-{
- /* this neighbor is not part of our candidate list */
- if (list_empty(&neigh_node->bonding_list))
- goto out;
-
- list_del_rcu(&neigh_node->bonding_list);
- INIT_LIST_HEAD(&neigh_node->bonding_list);
- batadv_neigh_node_free_ref(neigh_node);
- atomic_dec(&orig_node->bond_candidates);
-
-out:
- return;
-}
-
-/**
- * batadv_bonding_candidate_add - consider a new link for bonding mode towards
- * the given originator
- * @bat_priv: the bat priv with all the soft interface information
- * @orig_node: the target node
- * @neigh_node: the neighbor representing the new link to consider for bonding
- * mode
- */
-void batadv_bonding_candidate_add(struct batadv_priv *bat_priv,
- struct batadv_orig_node *orig_node,
- struct batadv_neigh_node *neigh_node)
-{
- struct batadv_algo_ops *bao = bat_priv->bat_algo_ops;
- struct batadv_neigh_node *tmp_neigh_node, *router = NULL;
- uint8_t interference_candidate = 0;
-
- spin_lock_bh(&orig_node->neigh_list_lock);
-
- /* only consider if it has the same primary address ... */
- if (!batadv_compare_eth(orig_node->orig,
- neigh_node->orig_node->primary_addr))
- goto candidate_del;
-
- router = batadv_orig_node_get_router(orig_node);
- if (!router)
- goto candidate_del;
-
-
- /* ... and is good enough to be considered */
- if (bao->bat_neigh_is_equiv_or_better(neigh_node, router))
- goto candidate_del;
-
- /* check if we have another candidate with the same mac address or
- * interface. If we do, we won't select this candidate because of
- * possible interference.
- */
- hlist_for_each_entry_rcu(tmp_neigh_node,
- &orig_node->neigh_list, list) {
- if (tmp_neigh_node == neigh_node)
- continue;
-
- /* we only care if the other candidate is even
- * considered as candidate.
- */
- if (list_empty(&tmp_neigh_node->bonding_list))
- continue;
-
- if ((neigh_node->if_incoming == tmp_neigh_node->if_incoming) ||
- (batadv_compare_eth(neigh_node->addr,
- tmp_neigh_node->addr))) {
- interference_candidate = 1;
- break;
- }
- }
-
- /* don't care further if it is an interference candidate */
- if (interference_candidate)
- goto candidate_del;
-
- /* this neighbor already is part of our candidate list */
- if (!list_empty(&neigh_node->bonding_list))
- goto out;
-
- if (!atomic_inc_not_zero(&neigh_node->refcount))
- goto out;
-
- list_add_rcu(&neigh_node->bonding_list, &orig_node->bond_list);
- atomic_inc(&orig_node->bond_candidates);
- goto out;
-
-candidate_del:
- batadv_bonding_candidate_del(orig_node, neigh_node);
-
-out:
- spin_unlock_bh(&orig_node->neigh_list_lock);
-
- if (router)
- batadv_neigh_node_free_ref(router);
-}
-
-/* copy primary address for bonding */
-void
-batadv_bonding_save_primary(const struct batadv_orig_node *orig_node,
- struct batadv_orig_node *orig_neigh_node,
- const struct batadv_ogm_packet *batman_ogm_packet)
-{
- if (!(batman_ogm_packet->flags & BATADV_PRIMARIES_FIRST_HOP))
- return;
-
- memcpy(orig_neigh_node->primary_addr, orig_node->orig, ETH_ALEN);
-}
-
/* checks whether the host restarted and is in the protection time.
* returns:
* 0 if the packet is to be accepted
@@ -459,114 +350,6 @@ out:
return ret;
}
-/* In the bonding case, send the packets in a round
- * robin fashion over the remaining interfaces.
- *
- * This method rotates the bonding list and increases the
- * returned router's refcount.
- */
-static struct batadv_neigh_node *
-batadv_find_bond_router(struct batadv_orig_node *primary_orig,
- const struct batadv_hard_iface *recv_if)
-{
- struct batadv_neigh_node *tmp_neigh_node;
- struct batadv_neigh_node *router = NULL, *first_candidate = NULL;
-
- rcu_read_lock();
- list_for_each_entry_rcu(tmp_neigh_node, &primary_orig->bond_list,
- bonding_list) {
- if (!first_candidate)
- first_candidate = tmp_neigh_node;
-
- /* recv_if == NULL on the first node. */
- if (tmp_neigh_node->if_incoming == recv_if)
- continue;
-
- if (!atomic_inc_not_zero(&tmp_neigh_node->refcount))
- continue;
-
- router = tmp_neigh_node;
- break;
- }
-
- /* use the first candidate if nothing was found. */
- if (!router && first_candidate &&
- atomic_inc_not_zero(&first_candidate->refcount))
- router = first_candidate;
-
- if (!router)
- goto out;
-
- /* selected should point to the next element
- * after the current router
- */
- spin_lock_bh(&primary_orig->neigh_list_lock);
- /* this is a list_move(), which unfortunately
- * does not exist as rcu version
- */
- list_del_rcu(&primary_orig->bond_list);
- list_add_rcu(&primary_orig->bond_list,
- &router->bonding_list);
- spin_unlock_bh(&primary_orig->neigh_list_lock);
-
-out:
- rcu_read_unlock();
- return router;
-}
-
-/**
- * batadv_find_ifalter_router - find the best of the remaining candidates which
- * are not using this interface
- * @bat_priv: the bat priv with all the soft interface information
- * @primary_orig: the destination
- * @recv_if: the interface that the router returned by this function has to not
- * use
- *
- * Returns the best candidate towards primary_orig that is not using recv_if.
- * Increases the returned neighbor's refcount
- */
-static struct batadv_neigh_node *
-batadv_find_ifalter_router(struct batadv_priv *bat_priv,
- struct batadv_orig_node *primary_orig,
- const struct batadv_hard_iface *recv_if)
-{
- struct batadv_neigh_node *router = NULL, *first_candidate = NULL;
- struct batadv_algo_ops *bao = bat_priv->bat_algo_ops;
- struct batadv_neigh_node *tmp_neigh_node;
-
- rcu_read_lock();
- list_for_each_entry_rcu(tmp_neigh_node, &primary_orig->bond_list,
- bonding_list) {
- if (!first_candidate)
- first_candidate = tmp_neigh_node;
-
- /* recv_if == NULL on the first node. */
- if (tmp_neigh_node->if_incoming == recv_if)
- continue;
-
- if (router && bao->bat_neigh_cmp(tmp_neigh_node, router))
- continue;
-
- if (!atomic_inc_not_zero(&tmp_neigh_node->refcount))
- continue;
-
- /* decrement refcount of previously selected router */
- if (router)
- batadv_neigh_node_free_ref(router);
-
- /* we found a better router (or at least one valid router) */
- router = tmp_neigh_node;
- }
-
- /* use the first candidate if nothing was found. */
- if (!router && first_candidate &&
- atomic_inc_not_zero(&first_candidate->refcount))
- router = first_candidate;
-
- rcu_read_unlock();
- return router;
-}
-
/**
* batadv_check_unicast_packet - Check for malformed unicast packets
* @bat_priv: the bat priv with all the soft interface information
@@ -604,95 +387,30 @@ static int batadv_check_unicast_packet(struct batadv_priv *bat_priv,
return 0;
}
-/* find a suitable router for this originator, and use
- * bonding if possible. increases the found neighbors
- * refcount.
+/**
+ * batadv_find_router - find a suitable router for this originator
+ * @bat_priv: the bat priv with all the soft interface information
+ * @orig_node: the destination node
+ * @recv_if: pointer to interface this packet was received on
+ *
+ * Returns the router which should be used for this orig_node on
+ * this interface, or NULL if not available.
*/
struct batadv_neigh_node *
batadv_find_router(struct batadv_priv *bat_priv,
struct batadv_orig_node *orig_node,
const struct batadv_hard_iface *recv_if)
{
- struct batadv_orig_node *primary_orig_node;
- struct batadv_orig_node *router_orig;
struct batadv_neigh_node *router;
- static uint8_t zero_mac[ETH_ALEN] = {0, 0, 0, 0, 0, 0};
- int bonding_enabled;
- uint8_t *primary_addr;
if (!orig_node)
return NULL;
router = batadv_orig_node_get_router(orig_node);
- if (!router)
- goto err;
- /* without bonding, the first node should
- * always choose the default router.
- */
- bonding_enabled = atomic_read(&bat_priv->bonding);
+ /* TODO: fill this later with new bonding mechanism */
- rcu_read_lock();
- /* select default router to output */
- router_orig = router->orig_node;
- if (!router_orig)
- goto err_unlock;
-
- if ((!recv_if) && (!bonding_enabled))
- goto return_router;
-
- primary_addr = router_orig->primary_addr;
-
- /* if we have something in the primary_addr, we can search
- * for a potential bonding candidate.
- */
- if (batadv_compare_eth(primary_addr, zero_mac))
- goto return_router;
-
- /* find the orig_node which has the primary interface. might
- * even be the same as our router_orig in many cases
- */
- if (batadv_compare_eth(primary_addr, router_orig->orig)) {
- primary_orig_node = router_orig;
- } else {
- primary_orig_node = batadv_orig_hash_find(bat_priv,
- primary_addr);
- if (!primary_orig_node)
- goto return_router;
-
- batadv_orig_node_free_ref(primary_orig_node);
- }
-
- /* with less than 2 candidates, we can't do any
- * bonding and prefer the original router.
- */
- if (atomic_read(&primary_orig_node->bond_candidates) < 2)
- goto return_router;
-
- /* all nodes between should choose a candidate which
- * is is not on the interface where the packet came
- * in.
- */
- batadv_neigh_node_free_ref(router);
-
- if (bonding_enabled)
- router = batadv_find_bond_router(primary_orig_node, recv_if);
- else
- router = batadv_find_ifalter_router(bat_priv, primary_orig_node,
- recv_if);
-
-return_router:
- if (router && router->if_incoming->if_status != BATADV_IF_ACTIVE)
- goto err_unlock;
-
- rcu_read_unlock();
return router;
-err_unlock:
- rcu_read_unlock();
-err:
- if (router)
- batadv_neigh_node_free_ref(router);
- return NULL;
}
static int batadv_route_unicast_packet(struct sk_buff *skb,
diff --git a/net/batman-adv/routing.h b/net/batman-adv/routing.h
index c1eadfa..192f0aa 100644
--- a/net/batman-adv/routing.h
+++ b/net/batman-adv/routing.h
@@ -44,15 +44,6 @@ struct batadv_neigh_node *
batadv_find_router(struct batadv_priv *bat_priv,
struct batadv_orig_node *orig_node,
const struct batadv_hard_iface *recv_if);
-void batadv_bonding_candidate_del(struct batadv_orig_node *orig_node,
- struct batadv_neigh_node *neigh_node);
-void batadv_bonding_candidate_add(struct batadv_priv *bat_priv,
- struct batadv_orig_node *orig_node,
- struct batadv_neigh_node *neigh_node);
-void batadv_bonding_save_primary(const struct batadv_orig_node *orig_node,
- struct batadv_orig_node *orig_neigh_node,
- const struct batadv_ogm_packet
- *batman_ogm_packet);
int batadv_window_protected(struct batadv_priv *bat_priv, int32_t seq_num_diff,
unsigned long *last_reset);
diff --git a/net/batman-adv/types.h b/net/batman-adv/types.h
index 0430a04..0f68845 100644
--- a/net/batman-adv/types.h
+++ b/net/batman-adv/types.h
@@ -199,12 +199,10 @@ struct batadv_orig_bat_iv {
* last_bcast_seqno)
* @last_bcast_seqno: last broadcast sequence number received by this host
* @neigh_list: list of potential next hop neighbor towards this orig node
- * @neigh_list_lock: lock protecting neigh_list, router and bonding_list
+ * @neigh_list_lock: lock protecting neigh_list and router
* @hash_entry: hlist node for batadv_priv::orig_hash
* @bat_priv: pointer to soft_iface this orig node belongs to
* @bcast_seqno_lock: lock protecting bcast_bits & last_bcast_seqno
- * @bond_candidates: how many candidates are available
- * @bond_list: list of bonding candidates
* @refcount: number of contexts the object is used
* @rcu: struct used for freeing in an RCU-safe manner
* @in_coding_list: list of nodes this orig can hear
@@ -240,14 +238,12 @@ struct batadv_orig_node {
DECLARE_BITMAP(bcast_bits, BATADV_TQ_LOCAL_WINDOW_SIZE);
uint32_t last_bcast_seqno;
struct hlist_head neigh_list;
- /* neigh_list_lock protects: neigh_list, router & bonding_list */
+ /* neigh_list_lock protects: neigh_list and router */
spinlock_t neigh_list_lock;
struct hlist_node hash_entry;
struct batadv_priv *bat_priv;
/* bcast_seqno_lock protects: bcast_bits & last_bcast_seqno */
spinlock_t bcast_seqno_lock;
- atomic_t bond_candidates;
- struct list_head bond_list;
atomic_t refcount;
struct rcu_head rcu;
#ifdef CONFIG_BATMAN_ADV_NC
@@ -320,7 +316,6 @@ struct batadv_neigh_bat_iv {
* @if_incoming: pointer to incoming hard interface
* @last_seen: when last packet via this neighbor was received
* @last_ttl: last received ttl from this neigh node
- * @bonding_list: list node for batadv_orig_node::bond_list
* @refcount: number of contexts the object is used
* @rcu: struct used for freeing in an RCU-safe manner
* @bat_iv: B.A.T.M.A.N. IV private structure
@@ -332,7 +327,6 @@ struct batadv_neigh_node {
struct batadv_hard_iface *if_incoming;
unsigned long last_seen;
uint8_t last_ttl;
- struct list_head bonding_list;
atomic_t refcount;
struct rcu_head rcu;
struct batadv_neigh_bat_iv bat_iv;
--
1.8.5.2
^ permalink raw reply related
* pull request: batman-adv 2014-01-13
From: Antonio Quartulli @ 2014-01-13 22:31 UTC (permalink / raw)
To: davem; +Cc: netdev, b.a.t.m.a.n
Hello David,
here you have another pull request intended for net-next/linux-3.14.
Patches from 1 to 8 implement an optimization for multi-interface setups.
It allows batman-adv nodes to take advantage of all the interfaces they
have in a smarter way than what the protocol was doing before. In
a few words this feature allows a single node to make routing decision by
using a sort of "network-wide knowledge". For more details please read [1].
Along with this feature you also have some specific debug facilities that
should make the netadmin life easier in case of troubles.
Patch 11 is adding some compile time checks against the size of the
packets we send over the wire. This should ensure us that no future member
reordering messes up with the padding of those structures.
The remaining patches are minor changes and cleanups. In particular:
- drop the CRC16 dependency from Kbuild
- remove "returns" at the end of void functions
- update release version
- update copyright years to 2014
- add missing sysfs attributes in batman-adv.txt document
Please pull or let me know of any problem!
Thanks a lot,
Antonio
[1] http://www.open-mesh.org/projects/batman-adv/wiki/Network-wide-multi-link-optimization
The following changes since commit f9577a376e28d7b8367319e103d6511d41c1c5fa:
Merge branch 'alx_stats' (2014-01-11 20:53:03 -0800)
are available in the git repository at:
git://git.open-mesh.org/linux-merge.git tags/batman-adv-for-davem
for you to fetch changes up to 12afc36e38b3b6a0ec9bda71632c2285e7fdbab2:
batman-adv: drop dependency against CRC16 (2014-01-12 14:41:21 +0100)
----------------------------------------------------------------
Included changes:
- drop dependency against CRC16
- move to new release version
- add size check at compile time for packet structs
- update copyright years in every file
- implement new bonding/interface alternation feature
----------------------------------------------------------------
Antonio Quartulli (3):
batman-adv: remove returns at the end of void functions
batman-adv: add missing sysfs attributes to README
batman-adv: drop dependency against CRC16
Simon Wunderlich (11):
batman-adv: remove bonding and interface alternating
batman-adv: split tq information in neigh_node struct
batman-adv: split out router from orig_node
batman-adv: add WiFi penalty
batman-adv: consider outgoing interface in OGM sending
batman-adv: add bonding again
batman-adv: add debugfs structure for information per interface
batman-adv: add debugfs support to view multiif tables
batman-adv: add build checks for packet sizes
batman-adv: update copyright years for 2014
batman-adv: Start new development cycle
Documentation/networking/batman-adv.txt | 9 +-
net/batman-adv/Kconfig | 1 -
net/batman-adv/Makefile | 2 +-
net/batman-adv/bat_algo.h | 2 +-
net/batman-adv/bat_iv_ogm.c | 882 +++++++++++++++++++++-----------
net/batman-adv/bitarray.c | 2 +-
net/batman-adv/bitarray.h | 2 +-
net/batman-adv/bridge_loop_avoidance.c | 2 +-
net/batman-adv/bridge_loop_avoidance.h | 2 +-
net/batman-adv/debugfs.c | 91 +++-
net/batman-adv/debugfs.h | 4 +-
net/batman-adv/distributed-arp-table.c | 5 +-
net/batman-adv/distributed-arp-table.h | 2 +-
net/batman-adv/fragmentation.c | 2 +-
net/batman-adv/fragmentation.h | 2 +-
net/batman-adv/gateway_client.c | 88 +++-
net/batman-adv/gateway_client.h | 2 +-
net/batman-adv/gateway_common.c | 2 +-
net/batman-adv/gateway_common.h | 2 +-
net/batman-adv/hard-interface.c | 11 +-
net/batman-adv/hard-interface.h | 20 +-
net/batman-adv/hash.c | 2 +-
net/batman-adv/hash.h | 2 +-
net/batman-adv/icmp_socket.c | 5 +-
net/batman-adv/icmp_socket.h | 2 +-
net/batman-adv/main.c | 26 +-
net/batman-adv/main.h | 10 +-
net/batman-adv/network-coding.c | 51 +-
net/batman-adv/network-coding.h | 9 +-
net/batman-adv/originator.c | 536 +++++++++++++++++--
net/batman-adv/originator.h | 22 +-
net/batman-adv/packet.h | 2 +-
net/batman-adv/routing.c | 429 ++++++----------
net/batman-adv/routing.h | 14 +-
net/batman-adv/send.c | 21 +-
net/batman-adv/send.h | 2 +-
net/batman-adv/soft-interface.c | 4 +-
net/batman-adv/soft-interface.h | 2 +-
net/batman-adv/sysfs.c | 2 +-
net/batman-adv/sysfs.h | 2 +-
net/batman-adv/translation-table.c | 8 +-
net/batman-adv/translation-table.h | 2 +-
net/batman-adv/types.h | 132 +++--
43 files changed, 1652 insertions(+), 768 deletions(-)
^ permalink raw reply
* Re: [PATCH 1/6] [v6] phylib: Add Clause 45 read/write functions
From: David Miller @ 2014-01-13 22:30 UTC (permalink / raw)
To: shh.xie
Cc: jg1.han, mugunthanvnm, f.fainelli, netdev, linux-kernel,
Shaohui.Xie, afleming
In-Reply-To: <1389335109-28179-1-git-send-email-shh.xie@gmail.com>
Series applied, thanks.
^ permalink raw reply
* Re: [PATCH V2 0/4] misc: xgene: Add support for APM X-Gene SoC Queue Manager/Traffic Manager
From: Ravi Patel @ 2014-01-13 22:18 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Greg KH, Loc Ho, davem, netdev, linux-kernel,
devicetree@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
Jon Masters, patches@apm.com, Keyur Chudgar
In-Reply-To: <201401122219.11593.arnd@arndb.de>
On Sun, Jan 12, 2014 at 1:19 PM, Arnd Bergmann <arnd@arndb.de> wrote:
> On Friday 10 January 2014, Ravi Patel wrote:
>
>> Do you want any further clarification or document related to QMTM.
>> We want to make sure everyone is on same page, understand and
>> conclude upon that QMTM is a device and and not a bus or a dma
>> engine.
>
> I have a much better understanding now, but there are still a few open
> questions from my side. Let me try to explain in my own words what I
> think is the relevant information (part of this is still guessing).
> It took me a while to figure out what it does from your description,
> and then some more time to see what it's actually good for (as
> opposed to adding complexity).
>
> Please confirm or correct the individual statements in this
> description:
>
> The QMTM serves as a relay for short (a few bytes) messages between
> the OS software and various slave hardware blocks on the SoC.
> The messages are typically but not always DMA descriptors used by the
> slave device used for starting bus master transactions by the slave,
> or for notifying sofware about the competion of a DMA transaction.
>
> The message format is specific to the slave device and the QMTM
> only understands the common header of the message.
>
> OS software sees the messages in cache-coherent memory and does
> not require any cache flushes or MMIO access for inbound messages
> and only a single posted MMIO write for outbound messages.
>
> The queues are likely designed to be per-thread and don't require
> software-side locking.
>
> For outbound messages, the QMTM is the bus master of a device-to-device
> DMA transaction that gets started once a message is queued and the
> device has signaled that it is ready for receiving it. The QMTM needs
> to know the bus address of the device as well as a slave ID for
> the signal pin.
> For inbound messages, the QMTM slave initiates a busmaster transaction
> and needs to know the bus address of its QMTM port, while the QMTM
> needs to know only the slave ID that is associated with the queue.
>
> In addition to those hardware properties, the QMTM driver needs to
> set up a memory buffer for the message queue as seen by the CPU,
> and needs tell the QMTM the location as well as some other
> properties such as the message length.
Your description is correct upto this point.
> For inbound messages, the QMTM serves a similar purpose as an MSI
> controller, ensuring that inbound DMA data has arrived in RAM
> before an interrupt is delivered to the CPU and thereby avoiding
> the need for an expensive MMIO read to serialize the DMA.
For inbound messages, slave device generates message on a completion
of a inbound DMA operation or any relevant operation targeted to the
CPU. The QMTM's role is to just trigger an interrupt to CPU when there
is a new message arrived from a slave device. QMTM doesn't know what
the message was for. It is upto the upper layer drivers to decide how
to process this message.
> The resources managed by the QMTM are both SoC-global (e.g. bus
> bandwidth) and slave specific (e.g. ethernet bandwith or buffer space).
> Global resource management is performed to prevent one slave
> device from monopolizing the system or preventing other slaves
> from making forward progress.
> Examples for local resource management (I had to think about this
> a long time, but probably some of these are wrong) would be
> * balancing between multiple non-busmaster devices connected to
> a dma-engine
> * distributing incoming ethernet data to the available CPUs based on
> a flow classifier in the MAC, e.g. by IOV MAC address, VLAN tag
> or even individual TCP connection depending on the NIC's capabilities.
> * 802.1p flow control for incoming ethernet data based on the amount
> of data queued up between the MAC and the driver
> * interrupt mitigation for both inbound data and outbound completion,
> by delaying the IRQ to the OS until multiple messages have arrived
> or a queue specific amount of time has passed.
> * controlling the amount of outbound buffer space per flow to minimize
> buffer-bloat between an ethernet driver and the NIC hardware.
> * reordering data from outbound flows based on priority.
>
> This is basically my current interpretation, I hope I got at least
> some of it right this time ;-)
You have got them right. Although we have taken Ethernet examples here,
most of the local resource management apply to other slave devices also.
^ permalink raw reply
* Re: [PATCH net-next v4 0/3] path mtu hardening patches
From: Hannes Frederic Sowa @ 2014-01-13 22:15 UTC (permalink / raw)
To: John Heffner, David Miller, Netdev, Eric Dumazet,
steffen.klassert, fweimer
In-Reply-To: <20140113220356.GL6586@order.stressinduktion.org>
On Mon, Jan 13, 2014 at 11:03:56PM +0100, Hannes Frederic Sowa wrote:
> I really don't like to depend on firewalling to do that. Especially on
> big routers one can use the routing table to protect interfaces for
> management and thus don't need to introduce stateful firewalling to
> realize a secure router setup which could cause performance degradation,
> especially with lots of small and shortlived flows (e.g. UDP/DNS).
This may get better if maybe some work is put into bringing this patch
forward: http://comments.gmane.org/gmane.linux.network/268758
Greetings,
Hannes
^ permalink raw reply
* Re: [PATCH net-next v4 0/3] path mtu hardening patches
From: David Miller @ 2014-01-13 22:12 UTC (permalink / raw)
To: johnwheffner; +Cc: netdev, eric.dumazet, steffen.klassert, fweimer
In-Reply-To: <CABrhC0=CfPxB7rNrA0ukfZyNLuTNSBB2srohr-Yt2b4k6oHWxg@mail.gmail.com>
From: John Heffner <johnwheffner@gmail.com>
Date: Mon, 13 Jan 2014 16:50:38 -0500
> Perhaps I misunderstood your description of FreeBSD then. It seems
> hard for me to believe that MTU discovery for UDP is broken by default
> in FreeBSD. It was not as of a couple years ago...
FreeBSD never implemented PMTU for anything other than TCP.
In fact, outside of Linux and perhaps Solaris, very few networking
stacks have implemented UDP PMTU.
^ permalink raw reply
* Re: [PATCH net-next v4 0/3] path mtu hardening patches
From: Hannes Frederic Sowa @ 2014-01-13 22:03 UTC (permalink / raw)
To: John Heffner
Cc: David Miller, Netdev, Eric Dumazet, steffen.klassert, fweimer
In-Reply-To: <CABrhC0=CfPxB7rNrA0ukfZyNLuTNSBB2srohr-Yt2b4k6oHWxg@mail.gmail.com>
On Mon, Jan 13, 2014 at 04:50:38PM -0500, John Heffner wrote:
> On Mon, Jan 13, 2014 at 4:28 PM, Hannes Frederic Sowa
> <hannes@stressinduktion.org> wrote:
> > On Mon, Jan 13, 2014 at 04:08:22PM -0500, John Heffner wrote:
> >> Would it be sufficient to allow Linux to be configured in a way that
> >> matches FreeBSD's behavior? (I believe you can do this easily with
> >> stateful firewall rules now, or possibly in the ICMP processing code
> >> with a sysctl switch.) I feel this would be a much cleaner approach.
> >
> > Actually, this is part of this series. The hardened path mtu mode provides
> > exactly that (Patch 3).
> >
> > But because we cannot switch this on by default, I also protected the
> > forwarding path. UDP path mtu discovery has been too long available on
> > Linux and, I guess, a lot of applications, especially running on routers,
> > depend on that.
>
> Perhaps I misunderstood your description of FreeBSD then. It seems
> hard for me to believe that MTU discovery for UDP is broken by default
> in FreeBSD. It was not as of a couple years ago...
I have to admit, I have not tested that. But I could not find an icmp handler
dealing with pmtu updates in the UDP protocol path. Neither did the icmp
layer.
TCP and sctp do handle the PRC_MSGSIZE callbacks and update the path
mtu on the route.
Maybe I overlooked something and I also did not check the history just
their current subversion checkout.
> The nice thing about stateful firewall rules is that they give you
> fine-grained policies over which ICMP messages you want to trust, and
> can filter out messages that don't match "connections" with existing
> state across a wide variety of protocols (including TCP, UDP and
> ICMP).
I really don't like to depend on firewalling to do that. Especially on
big routers one can use the routing table to protect interfaces for
management and thus don't need to introduce stateful firewalling to
realize a secure router setup which could cause performance degradation,
especially with lots of small and shortlived flows (e.g. UDP/DNS).
Greetings,
Hannes
^ permalink raw reply
* Re: [PATCH net-next 0/2] vxlan updates
From: Cong Wang @ 2014-01-13 21:59 UTC (permalink / raw)
To: Daniel Borkmann; +Cc: David Miller, netdev
In-Reply-To: <1389634880-4138-1-git-send-email-dborkman@redhat.com>
On Mon, Jan 13, 2014 at 9:41 AM, Daniel Borkmann <dborkman@redhat.com> wrote:
> Did the split into two patches upon request from Cong Wang.
>
> Changelog:
>
> v1->v2:
> - Removed BUG_ON as it's not needed.
> v2->v3:
> - Removed dev->reg_state check for netns.
> v3->v4:
> - Removed list_del(), we seem to do it in some places and
> in some others not; we agreed it's not really necessary.
> - Split patch into 2 patches, notifier part and module
> unload cleanup part.
Looks good to my eyes, so
Reviewed-by: Cong Wang <cwang@twopensource.com>
^ permalink raw reply
* RE: [PATCH 07/15] net: i40e calls skb_set_hash
From: Brown, Aaron F @ 2014-01-13 21:53 UTC (permalink / raw)
To: Kirsher, Jeffrey T, Tom Herbert; +Cc: David Miller, netdev
In-Reply-To: <CAL3LdT4M0gHP5bW3A1GDJGy2zrt4+YQpK3i8WhAp31N2yHsQkg@mail.gmail.com>
> From: netdev-owner@vger.kernel.org [mailto:netdev-owner@vger.kernel.org]
> On Behalf Of Jeff Kirsher
> Sent: Wednesday, December 18, 2013 12:40 AM
> To: Tom Herbert
> Cc: David Miller; netdev
> Subject: Re: [PATCH 07/15] net: i40e calls skb_set_hash
>
> On Tue, Dec 17, 2013 at 11:27 PM, Tom Herbert <therbert@google.com> wrote:
> > Drivers should call skb_set_hash to set the hash and its type in an
> > skbuff.
> >
> > Signed-off-by: Tom Herbert <therbert@google.com>
> > ---
> > drivers/net/ethernet/intel/i40e/i40e_txrx.c | 3 ++-
> > 1 file changed, 2 insertions(+), 1 deletion(-)
> >
>
> I have added this patch to my queue, thanks Tom.
And unfortunately it fails to compile giving the following error:
------------------------------------------------------------------------
drivers/net/ethernet/intel/i40e/i40e_txrx.c: In function 'i40e_clean_rx_irq':
drivers/net/ethernet/intel/i40e/i40e_txrx.c:1088: error: too few arguments to function 'i40e_rx_checksum'
^ permalink raw reply
* Re: [PATCH net-next v4 0/3] path mtu hardening patches
From: John Heffner @ 2014-01-13 21:50 UTC (permalink / raw)
To: John Heffner, David Miller, Netdev, Eric Dumazet,
steffen.klassert, fweimer
In-Reply-To: <20140113212808.GJ6586@order.stressinduktion.org>
On Mon, Jan 13, 2014 at 4:28 PM, Hannes Frederic Sowa
<hannes@stressinduktion.org> wrote:
> On Mon, Jan 13, 2014 at 04:08:22PM -0500, John Heffner wrote:
>> Would it be sufficient to allow Linux to be configured in a way that
>> matches FreeBSD's behavior? (I believe you can do this easily with
>> stateful firewall rules now, or possibly in the ICMP processing code
>> with a sysctl switch.) I feel this would be a much cleaner approach.
>
> Actually, this is part of this series. The hardened path mtu mode provides
> exactly that (Patch 3).
>
> But because we cannot switch this on by default, I also protected the
> forwarding path. UDP path mtu discovery has been too long available on
> Linux and, I guess, a lot of applications, especially running on routers,
> depend on that.
Perhaps I misunderstood your description of FreeBSD then. It seems
hard for me to believe that MTU discovery for UDP is broken by default
in FreeBSD. It was not as of a couple years ago...
The nice thing about stateful firewall rules is that they give you
fine-grained policies over which ICMP messages you want to trust, and
can filter out messages that don't match "connections" with existing
state across a wide variety of protocols (including TCP, UDP and
ICMP).
-John
^ permalink raw reply
* [PATCH RFC] reciprocal_divide: correction/update of the algorithm
From: Hannes Frederic Sowa @ 2014-01-13 21:42 UTC (permalink / raw)
To: netdev; +Cc: dborkman, eric.dumazet, linux-kernel, darkjames-ws
This patch is a RFC and part of a series Daniel Borkmann and me want to
do when introducing prandom_u32_range{,_ro} and prandom_u32_max{,_ro}
helpers later this week.
At first Jakub Zawadzki noticed that some divisions by reciprocal_divide
were not correct:
http://www.wireshark.org/~darkjames/reciprocal-buggy.c
He could also show this with BPF:
http://www.wireshark.org/~darkjames/set-and-dump-filter-k-bug.c
Also: reciprocal_value and reciprocal_divide always return 0 for
divisions by 1. This is a bit worrisome as those functions also get
used in mm/slab.c and lib/flex_array.c. Bonding already seems to check
for the 1-divisor case and handles that correctly. We don't know about
other problems, yet.
I propose an correction/update of the algorithm
based on the paper "T. Granlund and P. L. Montgomery:
Division by Invariant Integers Using Multiplication"
<http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.1.2556>
The assembler implementation from Agner Fog, found here
<http://www.agner.org/optimize/asmlib.zip>, helped a lot while
implementing.
I would like to have feedback if people see problems with this patch or
have concerns about performance. I did some testing on x86-64 and found
no problems so far but did no performance evaluation, yet.
The current code does break the call-sides of reciprocal_divide. The necessary
changes will be part of the full series, then.
Thanks!
---
include/linux/reciprocal_div.h | 12 +++++++++---
lib/reciprocal_div.c | 22 ++++++++++++++++++----
2 files changed, 27 insertions(+), 7 deletions(-)
diff --git a/include/linux/reciprocal_div.h b/include/linux/reciprocal_div.h
index f9c90b3..6f17a87 100644
--- a/include/linux/reciprocal_div.h
+++ b/include/linux/reciprocal_div.h
@@ -22,11 +22,17 @@
* Should not be called before each reciprocal_divide(),
* or else the performance is slower than a normal divide.
*/
-extern u32 reciprocal_value(u32 B);
+struct reciprocal_value {
+ u32 m;
+ u8 sh1, sh2;
+};
-static inline u32 reciprocal_divide(u32 A, u32 R)
+struct reciprocal_value reciprocal_value(u32 d);
+
+static inline u32 reciprocal_divide(u32 a, struct reciprocal_value R)
{
- return (u32)(((u64)A * R) >> 32);
+ u32 t = (u32)(((u64)a * R.m) >> 32);
+ return (t + ((a - t) >> R.sh1)) >> R.sh2;
}
#endif
diff --git a/lib/reciprocal_div.c b/lib/reciprocal_div.c
index 75510e9..b741b30 100644
--- a/lib/reciprocal_div.c
+++ b/lib/reciprocal_div.c
@@ -1,11 +1,25 @@
+#include <linux/kernel.h>
#include <asm/div64.h>
#include <linux/reciprocal_div.h>
#include <linux/export.h>
-u32 reciprocal_value(u32 k)
+/* For a description of the algorithmus please look at
+ * linux/reciprocal_div.h
+ */
+
+struct reciprocal_value reciprocal_value(u32 d)
{
- u64 val = (1LL << 32) + (k - 1);
- do_div(val, k);
- return (u32)val;
+ struct reciprocal_value R;
+ u64 m;
+ int l;
+
+ l = fls(d - 1);
+ m = ((1ULL << 32) * ((1ULL << l) - d));
+ do_div(m, d);
+ ++m;
+ R.m = (u32)m;
+ R.sh1 = min(l, 1);
+ R.sh2 = max(l-1, 0);
+ return R;
}
EXPORT_SYMBOL(reciprocal_value);
--
1.8.4.2
^ permalink raw reply related
* Re: [PATCH] net: Check skb->rxhash in gro_receive
From: Eric Dumazet @ 2014-01-13 21:37 UTC (permalink / raw)
To: Ben Hutchings; +Cc: David Miller, therbert, netdev
In-Reply-To: <1389646213.2025.159.camel@bwh-desktop.uk.level5networks.com>
On Mon, 2014-01-13 at 20:50 +0000, Ben Hutchings wrote:
> According to the original report of that vulnerability:
>
> > skb_flow_dissect() were used by several places:
> > - packet scheduler that want classify flows
> > - skb_get_rxhash() that will be used by RPS, vxlan, multiqueue
> > tap,macvtap packet fanout
> > - skb_probe_transport_header() which was used for probing transport
> > header for DODGY packets
> > - __skb_get_poff() which will be used by socket filter
>
> So flow dissector is already part of the attack surface for both local
> and remote users in common configurations.
Take a debian or Android distro, and this bug is not hit on 'common
configurations'. Send them a 'packet of death', they will not hang,
unless some admin set up RPS/RFS on the incoming device.
Anyway, I see no point pushing this schem (flow dissect all incoming
packets). A router has no gain going to L4 header, but still might
need GRO.
^ permalink raw reply
* Re: [PATCH] net: Check skb->rxhash in gro_receive
From: Eric Dumazet @ 2014-01-13 21:36 UTC (permalink / raw)
To: Tom Herbert; +Cc: David Miller, Linux Netdev List
In-Reply-To: <CA+mtBx_a0cnqXHFk3zRuQQnRNUqMet=N3h10HPgTGghyGoh=xA@mail.gmail.com>
On Mon, 2014-01-13 at 13:24 -0800, Tom Herbert wrote:
>
> Or maybe nobody uses IPIP. ;-)
The bug happened even without IPIP being used on the node.
Changing one byte from 0x45 to 0x40 was enough to trigger it.
Since we do not perform header checksum in flow dissector, normal
corrupted traffic on the Internet would have a chance to trigger the
infinite loop.
^ permalink raw reply
* Re: [PATCH net-next v4 0/3] path mtu hardening patches
From: Hannes Frederic Sowa @ 2014-01-13 21:28 UTC (permalink / raw)
To: John Heffner
Cc: David Miller, Netdev, Eric Dumazet, steffen.klassert, fweimer
In-Reply-To: <CABrhC0=9ff8aLG4y4gxiGrGBBHvdcWD2pCC3wyLTveKHRySg4A@mail.gmail.com>
On Mon, Jan 13, 2014 at 04:08:22PM -0500, John Heffner wrote:
> On Mon, Jan 13, 2014 at 3:42 PM, Hannes Frederic Sowa
> <hannes@stressinduktion.org> wrote:
> > On Mon, Jan 13, 2014 at 02:35:31PM -0500, John Heffner wrote:
> >> My only comment would be not to look to me as the only source of
> >> reason not to include this change. I've been largely disconnected
> >> from Linux development for several years and don't have time to get
> >> into a protracted discussion on this topic.
> >>
> >> FWIW, I still have doubts as to whether this is the best approach to
> >> solving the underlying problem. I still haven't heard any reason why
> >> firewall rules and other administrative best practices, such as using
> >
> > Because we currently cannot easily filter icmp payloads and check whether
> > it is in a response for a local socket or a malicious one.
> >
> >> separate management and forwarding interfaces on a router, don't
> >> practically solve this problem.
> >
> > I don't think this is practiable, especially in times of small devices
> > doing routing (e.g. smartphones).
> >
> >> I'd also be curious to hear what
> >> dedicated routing operating systems do, and why I haven't heard about
> >> widespread fragmentation DoS attacks.
> >
> > My old Cisco didn't honour those pmtu packets (at least in default
> > configuration) and FreeBSD only accepts pmtu information for TCP sockets
> > where it also verifies the sequence number. It does not react to pmtu
> > notifications in response to icmp or udp payloads.
> >
> > Routing path does use the pmtu values on FreeBSD, though. But it is much
> > harder to inject path mtu packets there because, as said, they are only
> > accepted for tcp.
>
> Would it be sufficient to allow Linux to be configured in a way that
> matches FreeBSD's behavior? (I believe you can do this easily with
> stateful firewall rules now, or possibly in the ICMP processing code
> with a sysctl switch.) I feel this would be a much cleaner approach.
Actually, this is part of this series. The hardened path mtu mode provides
exactly that (Patch 3).
But because we cannot switch this on by default, I also protected the
forwarding path. UDP path mtu discovery has been too long available on
Linux and, I guess, a lot of applications, especially running on routers,
depend on that.
Greetings,
Hannes
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox