[patch net-next 0/8] mlxsw: cleanup neigh handling

[patch net-next 0/8] mlxsw: cleanup neigh handling

[Jiri Pirko]

From: Jiri Pirko <jiri@mellanox.com>

Ido says:

This series addresses long standing issues in the mlxsw driver
concerning neighbour reflection. It also prepares the code for follow-up
changes dealing with proper resource cleanup and nexthop reflection.

The first two patches convert the neighbour reflection code to use an
ordered workqueue, to prevent re-ordering of NEIGH_UPDATE events that
may happen following subsequent patches.

The third to fifth patches remove the ndo_neigh_{construct,destroy}
entry points from the driver, thereby relying only on NEIGH_UPDATE
events for neighbour reflection. This simplifies the code considerably.

Last patches are fallout and adjust nits in the code I noticed while
going over it.

Ido Schimmel (8):
  mlxsw: core: Queue work immediately instead of delaying it
  mlxsw: spectrum_router: Use ordered workqueue for neigh updates
  mlxsw: spectrum_router: Remove unused variable
  mlxsw: spectrum_router: Simplify neighbour reflection
  net: remove ndo_neigh_{construct,destroy} from stacked devices
  mlxsw: spectrum_router: Remove redundant check
  mlxsw: spectrum_router: Don't read 'nud_state' without lock
  mlxsw: spectrum_router: Fix typo in comment

 drivers/net/bonding/bond_main.c                    |   2 -
 drivers/net/ethernet/mellanox/mlxsw/core.c         |   6 +-
 drivers/net/ethernet/mellanox/mlxsw/core.h         |   2 +-
 drivers/net/ethernet/mellanox/mlxsw/spectrum.c     |   2 -
 drivers/net/ethernet/mellanox/mlxsw/spectrum.h     |   4 -
 .../net/ethernet/mellanox/mlxsw/spectrum_router.c  | 287 +++++++++++----------
 drivers/net/team/team.c                            |   2 -
 include/linux/netdevice.h                          |   4 -
 net/8021q/vlan_dev.c                               |   2 -
 net/bridge/br_device.c                             |   2 -
 net/core/dev.c                                     |  44 ----
 11 files changed, 148 insertions(+), 209 deletions(-)

-- 
2.7.4

[patch net-next 1/8] mlxsw: core: Queue work immediately instead of delaying it

[Jiri Pirko]

From: Ido Schimmel <idosch@mellanox.com>

We always use zero delay before queueing a work on the ordered workqueue
('mlxsw_owq'), so use work_struct directly instead of delayable work.

Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlxsw/core.c            | 6 +++---
 drivers/net/ethernet/mellanox/mlxsw/core.h            | 2 +-
 drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c | 8 ++++----
 3 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/core.c b/drivers/net/ethernet/mellanox/mlxsw/core.c
index 57a9884..a4c0784 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/core.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/core.c
@@ -1901,11 +1901,11 @@ int mlxsw_core_schedule_dw(struct delayed_work *dwork, unsigned long delay)
 }
 EXPORT_SYMBOL(mlxsw_core_schedule_dw);
 
-int mlxsw_core_schedule_odw(struct delayed_work *dwork, unsigned long delay)
+bool mlxsw_core_schedule_work(struct work_struct *work)
 {
-	return queue_delayed_work(mlxsw_owq, dwork, delay);
+	return queue_work(mlxsw_owq, work);
 }
-EXPORT_SYMBOL(mlxsw_core_schedule_odw);
+EXPORT_SYMBOL(mlxsw_core_schedule_work);
 
 void mlxsw_core_flush_owq(void)
 {
diff --git a/drivers/net/ethernet/mellanox/mlxsw/core.h b/drivers/net/ethernet/mellanox/mlxsw/core.h
index a7f94fb..cf38cf9 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/core.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/core.h
@@ -207,7 +207,7 @@ enum devlink_port_type mlxsw_core_port_type_get(struct mlxsw_core *mlxsw_core,
 						u8 local_port);
 
 int mlxsw_core_schedule_dw(struct delayed_work *dwork, unsigned long delay);
-int mlxsw_core_schedule_odw(struct delayed_work *dwork, unsigned long delay);
+bool mlxsw_core_schedule_work(struct work_struct *work);
 void mlxsw_core_flush_owq(void);
 
 #define MLXSW_CONFIG_PROFILE_SWID_COUNT 8
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 9e494a4..01e86e7 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -1964,7 +1964,7 @@ static void __mlxsw_sp_router_fini(struct mlxsw_sp *mlxsw_sp)
 }
 
 struct mlxsw_sp_fib_event_work {
-	struct delayed_work dw;
+	struct work_struct work;
 	struct fib_entry_notifier_info fen_info;
 	struct mlxsw_sp *mlxsw_sp;
 	unsigned long event;
@@ -1973,7 +1973,7 @@ struct mlxsw_sp_fib_event_work {
 static void mlxsw_sp_router_fib_event_work(struct work_struct *work)
 {
 	struct mlxsw_sp_fib_event_work *fib_work =
-		container_of(work, struct mlxsw_sp_fib_event_work, dw.work);
+		container_of(work, struct mlxsw_sp_fib_event_work, work);
 	struct mlxsw_sp *mlxsw_sp = fib_work->mlxsw_sp;
 	int err;
 
@@ -2014,7 +2014,7 @@ static int mlxsw_sp_router_fib_event(struct notifier_block *nb,
 	if (WARN_ON(!fib_work))
 		return NOTIFY_BAD;
 
-	INIT_DELAYED_WORK(&fib_work->dw, mlxsw_sp_router_fib_event_work);
+	INIT_WORK(&fib_work->work, mlxsw_sp_router_fib_event_work);
 	fib_work->mlxsw_sp = mlxsw_sp;
 	fib_work->event = event;
 
@@ -2029,7 +2029,7 @@ static int mlxsw_sp_router_fib_event(struct notifier_block *nb,
 		break;
 	}
 
-	mlxsw_core_schedule_odw(&fib_work->dw, 0);
+	mlxsw_core_schedule_work(&fib_work->work);
 
 	return NOTIFY_DONE;
 }
-- 
2.7.4

[patch net-next 2/8] mlxsw: spectrum_router: Use ordered workqueue for neigh updates

[Jiri Pirko]

From: Ido Schimmel <idosch@mellanox.com>

We currently associate each neighbour entry with a work item, so it's
not possible to have multiple events queued for the same neighbour
entry. However, this is about to be changed so that the neighbour entry
is only resolved when the work item is scheduled.

The above can result in a mismatch between the kernel's and the device's
neighbour table, unless the associated work items are processed in the
order in which they were submitted.

Do that by migrating the NEIGH_UPDATE work items to be processed in the
ordered workqueue which was recently introduced in mlxsw in commit
a3832b31898f ("mlxsw: core: Create an ordered workqueue for FIB
offload").

Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 01e86e7..0d8201e 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -614,7 +614,7 @@ struct mlxsw_sp_neigh_entry {
 	struct mlxsw_sp_neigh_key key;
 	u16 rif;
 	bool offloaded;
-	struct delayed_work dw;
+	struct work_struct work;
 	struct mlxsw_sp_port *mlxsw_sp_port;
 	unsigned char ha[ETH_ALEN];
 	struct list_head nexthop_list; /* list of nexthops using
@@ -659,7 +659,7 @@ mlxsw_sp_neigh_entry_create(struct neighbour *n, u16 rif)
 		return NULL;
 	neigh_entry->key.n = n;
 	neigh_entry->rif = rif;
-	INIT_DELAYED_WORK(&neigh_entry->dw, mlxsw_sp_router_neigh_update_hw);
+	INIT_WORK(&neigh_entry->work, mlxsw_sp_router_neigh_update_hw);
 	INIT_LIST_HEAD(&neigh_entry->nexthop_list);
 	return neigh_entry;
 }
@@ -935,7 +935,7 @@ mlxsw_sp_nexthop_neigh_update(struct mlxsw_sp *mlxsw_sp,
 static void mlxsw_sp_router_neigh_update_hw(struct work_struct *work)
 {
 	struct mlxsw_sp_neigh_entry *neigh_entry =
-		container_of(work, struct mlxsw_sp_neigh_entry, dw.work);
+		container_of(work, struct mlxsw_sp_neigh_entry, work);
 	struct neighbour *n = neigh_entry->key.n;
 	struct mlxsw_sp_port *mlxsw_sp_port = neigh_entry->mlxsw_sp_port;
 	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
@@ -1052,7 +1052,7 @@ int mlxsw_sp_router_netevent_event(struct notifier_block *unused,
 		 * work.
 		 */
 		neigh_clone(n);
-		if (!mlxsw_core_schedule_dw(&neigh_entry->dw, 0)) {
+		if (!mlxsw_core_schedule_work(&neigh_entry->work)) {
 			neigh_release(n);
 			mlxsw_sp_port_dev_put(mlxsw_sp_port);
 		}
-- 
2.7.4

[patch net-next 3/8] mlxsw: spectrum_router: Remove unused variable

[Jiri Pirko]

From: Ido Schimmel <idosch@mellanox.com>

Since commit 33b1341cd1bf ("mlxsw: spectrum_router: Fix handling of
neighbour structure") we no longer use destination IP for neighbour
lookup, so remove it.

Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 0d8201e..851799d 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -1004,7 +1004,6 @@ int mlxsw_sp_router_netevent_event(struct notifier_block *unused,
 	struct net_device *dev;
 	struct neigh_parms *p;
 	struct neighbour *n;
-	u32 dip;
 
 	switch (event) {
 	case NETEVENT_DELAY_PROBE_TIME_UPDATE:
@@ -1039,7 +1038,6 @@ int mlxsw_sp_router_netevent_event(struct notifier_block *unused,
 			return NOTIFY_DONE;
 
 		mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
-		dip = ntohl(*((__be32 *) n->primary_key));
 		neigh_entry = mlxsw_sp_neigh_entry_lookup(mlxsw_sp, n);
 		if (WARN_ON(!neigh_entry)) {
 			mlxsw_sp_port_dev_put(mlxsw_sp_port);
-- 
2.7.4

[patch net-next 4/8] mlxsw: spectrum_router: Simplify neighbour reflection

[Jiri Pirko]

From: Ido Schimmel <idosch@mellanox.com>

Up until now we had two interfaces for neighbour related configuration:
ndo_neigh_{construct,destroy} and NEIGH_UPDATE netevents. The ndos were
used to add and remove neighbours from the driver's cache, whereas the
netevent was used to reflect the neighbours into the device's tables.

However, if the NUD state of a neighbour isn't NUD_VALID or if the
neighbour is dead, then there's really no reason for us to keep it
inside our cache. The only exception to this rule are neighbours that
are also used for nexthops, which we periodically refresh to get them
resolved.

We can therefore eliminate the ndo entry point into the driver and
simplify the code, making it similar to the FIB reflection, which is
based solely on events. This also helps us avoid a locking issue, in
which the RIF cache was traversed without proper locking during
insertion into the neigh entry cache.

Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlxsw/spectrum.c     |   2 -
 drivers/net/ethernet/mellanox/mlxsw/spectrum.h     |   4 -
 .../net/ethernet/mellanox/mlxsw/spectrum_router.c  | 263 +++++++++++----------
 3 files changed, 135 insertions(+), 134 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
index 8a52c86..c2a6751 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
@@ -1400,8 +1400,6 @@ static const struct net_device_ops mlxsw_sp_port_netdev_ops = {
 	.ndo_get_offload_stats	= mlxsw_sp_port_get_offload_stats,
 	.ndo_vlan_rx_add_vid	= mlxsw_sp_port_add_vid,
 	.ndo_vlan_rx_kill_vid	= mlxsw_sp_port_kill_vid,
-	.ndo_neigh_construct	= mlxsw_sp_router_neigh_construct,
-	.ndo_neigh_destroy	= mlxsw_sp_router_neigh_destroy,
 	.ndo_fdb_add		= switchdev_port_fdb_add,
 	.ndo_fdb_del		= switchdev_port_fdb_del,
 	.ndo_fdb_dump		= switchdev_port_fdb_dump,
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
index 4d251e0..dcd641a 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
@@ -599,10 +599,6 @@ static inline void mlxsw_sp_port_dcb_fini(struct mlxsw_sp_port *mlxsw_sp_port)
 
 int mlxsw_sp_router_init(struct mlxsw_sp *mlxsw_sp);
 void mlxsw_sp_router_fini(struct mlxsw_sp *mlxsw_sp);
-int mlxsw_sp_router_neigh_construct(struct net_device *dev,
-				    struct neighbour *n);
-void mlxsw_sp_router_neigh_destroy(struct net_device *dev,
-				   struct neighbour *n);
 int mlxsw_sp_router_netevent_event(struct notifier_block *unused,
 				   unsigned long event, void *ptr);
 
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 851799d..c338b08 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -613,9 +613,7 @@ struct mlxsw_sp_neigh_entry {
 	struct rhash_head ht_node;
 	struct mlxsw_sp_neigh_key key;
 	u16 rif;
-	bool offloaded;
-	struct work_struct work;
-	struct mlxsw_sp_port *mlxsw_sp_port;
+	bool connected;
 	unsigned char ha[ETH_ALEN];
 	struct list_head nexthop_list; /* list of nexthops using
 					* this neigh entry
@@ -629,105 +627,88 @@ static const struct rhashtable_params mlxsw_sp_neigh_ht_params = {
 	.key_len = sizeof(struct mlxsw_sp_neigh_key),
 };
 
-static int
-mlxsw_sp_neigh_entry_insert(struct mlxsw_sp *mlxsw_sp,
-			    struct mlxsw_sp_neigh_entry *neigh_entry)
-{
-	return rhashtable_insert_fast(&mlxsw_sp->router.neigh_ht,
-				      &neigh_entry->ht_node,
-				      mlxsw_sp_neigh_ht_params);
-}
-
-static void
-mlxsw_sp_neigh_entry_remove(struct mlxsw_sp *mlxsw_sp,
-			    struct mlxsw_sp_neigh_entry *neigh_entry)
-{
-	rhashtable_remove_fast(&mlxsw_sp->router.neigh_ht,
-			       &neigh_entry->ht_node,
-			       mlxsw_sp_neigh_ht_params);
-}
-
-static void mlxsw_sp_router_neigh_update_hw(struct work_struct *work);
-
 static struct mlxsw_sp_neigh_entry *
-mlxsw_sp_neigh_entry_create(struct neighbour *n, u16 rif)
+mlxsw_sp_neigh_entry_alloc(struct mlxsw_sp *mlxsw_sp, struct neighbour *n,
+			   u16 rif)
 {
 	struct mlxsw_sp_neigh_entry *neigh_entry;
 
-	neigh_entry = kzalloc(sizeof(*neigh_entry), GFP_ATOMIC);
+	neigh_entry = kzalloc(sizeof(*neigh_entry), GFP_KERNEL);
 	if (!neigh_entry)
 		return NULL;
+
 	neigh_entry->key.n = n;
 	neigh_entry->rif = rif;
-	INIT_WORK(&neigh_entry->work, mlxsw_sp_router_neigh_update_hw);
 	INIT_LIST_HEAD(&neigh_entry->nexthop_list);
+
 	return neigh_entry;
 }
 
-static void
-mlxsw_sp_neigh_entry_destroy(struct mlxsw_sp_neigh_entry *neigh_entry)
+static void mlxsw_sp_neigh_entry_free(struct mlxsw_sp_neigh_entry *neigh_entry)
 {
 	kfree(neigh_entry);
 }
 
-static struct mlxsw_sp_neigh_entry *
-mlxsw_sp_neigh_entry_lookup(struct mlxsw_sp *mlxsw_sp, struct neighbour *n)
+static int
+mlxsw_sp_neigh_entry_insert(struct mlxsw_sp *mlxsw_sp,
+			    struct mlxsw_sp_neigh_entry *neigh_entry)
 {
-	struct mlxsw_sp_neigh_key key;
+	return rhashtable_insert_fast(&mlxsw_sp->router.neigh_ht,
+				      &neigh_entry->ht_node,
+				      mlxsw_sp_neigh_ht_params);
+}
 
-	key.n = n;
-	return rhashtable_lookup_fast(&mlxsw_sp->router.neigh_ht,
-				      &key, mlxsw_sp_neigh_ht_params);
+static void
+mlxsw_sp_neigh_entry_remove(struct mlxsw_sp *mlxsw_sp,
+			    struct mlxsw_sp_neigh_entry *neigh_entry)
+{
+	rhashtable_remove_fast(&mlxsw_sp->router.neigh_ht,
+			       &neigh_entry->ht_node,
+			       mlxsw_sp_neigh_ht_params);
 }
 
-int mlxsw_sp_router_neigh_construct(struct net_device *dev,
-				    struct neighbour *n)
+static struct mlxsw_sp_neigh_entry *
+mlxsw_sp_neigh_entry_create(struct mlxsw_sp *mlxsw_sp, struct neighbour *n)
 {
-	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
-	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
 	struct mlxsw_sp_neigh_entry *neigh_entry;
 	struct mlxsw_sp_rif *r;
 	int err;
 
-	if (n->tbl != &arp_tbl)
-		return 0;
-
-	neigh_entry = mlxsw_sp_neigh_entry_lookup(mlxsw_sp, n);
-	if (neigh_entry)
-		return 0;
-
 	r = mlxsw_sp_rif_find_by_dev(mlxsw_sp, n->dev);
-	if (WARN_ON(!r))
-		return -EINVAL;
+	if (!r)
+		return ERR_PTR(-EINVAL);
 
-	neigh_entry = mlxsw_sp_neigh_entry_create(n, r->rif);
+	neigh_entry = mlxsw_sp_neigh_entry_alloc(mlxsw_sp, n, r->rif);
 	if (!neigh_entry)
-		return -ENOMEM;
+		return ERR_PTR(-ENOMEM);
+
 	err = mlxsw_sp_neigh_entry_insert(mlxsw_sp, neigh_entry);
 	if (err)
 		goto err_neigh_entry_insert;
-	return 0;
+
+	return neigh_entry;
 
 err_neigh_entry_insert:
-	mlxsw_sp_neigh_entry_destroy(neigh_entry);
-	return err;
+	mlxsw_sp_neigh_entry_free(neigh_entry);
+	return ERR_PTR(err);
 }
 
-void mlxsw_sp_router_neigh_destroy(struct net_device *dev,
-				   struct neighbour *n)
+static void
+mlxsw_sp_neigh_entry_destroy(struct mlxsw_sp *mlxsw_sp,
+			     struct mlxsw_sp_neigh_entry *neigh_entry)
 {
-	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
-	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
-	struct mlxsw_sp_neigh_entry *neigh_entry;
+	mlxsw_sp_neigh_entry_remove(mlxsw_sp, neigh_entry);
+	mlxsw_sp_neigh_entry_free(neigh_entry);
+}
 
-	if (n->tbl != &arp_tbl)
-		return;
+static struct mlxsw_sp_neigh_entry *
+mlxsw_sp_neigh_entry_lookup(struct mlxsw_sp *mlxsw_sp, struct neighbour *n)
+{
+	struct mlxsw_sp_neigh_key key;
 
-	neigh_entry = mlxsw_sp_neigh_entry_lookup(mlxsw_sp, n);
-	if (!neigh_entry)
-		return;
-	mlxsw_sp_neigh_entry_remove(mlxsw_sp, neigh_entry);
-	mlxsw_sp_neigh_entry_destroy(neigh_entry);
+	key.n = n;
+	return rhashtable_lookup_fast(&mlxsw_sp->router.neigh_ht,
+				      &key, mlxsw_sp_neigh_ht_params);
 }
 
 static void
@@ -932,76 +913,99 @@ mlxsw_sp_nexthop_neigh_update(struct mlxsw_sp *mlxsw_sp,
 			      struct mlxsw_sp_neigh_entry *neigh_entry,
 			      bool removing);
 
-static void mlxsw_sp_router_neigh_update_hw(struct work_struct *work)
+static enum mlxsw_reg_rauht_op mlxsw_sp_rauht_op(bool adding)
+{
+	return adding ? MLXSW_REG_RAUHT_OP_WRITE_ADD :
+			MLXSW_REG_RAUHT_OP_WRITE_DELETE;
+}
+
+static void
+mlxsw_sp_router_neigh_entry_op4(struct mlxsw_sp *mlxsw_sp,
+				struct mlxsw_sp_neigh_entry *neigh_entry,
+				enum mlxsw_reg_rauht_op op)
 {
-	struct mlxsw_sp_neigh_entry *neigh_entry =
-		container_of(work, struct mlxsw_sp_neigh_entry, work);
 	struct neighbour *n = neigh_entry->key.n;
-	struct mlxsw_sp_port *mlxsw_sp_port = neigh_entry->mlxsw_sp_port;
-	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	u32 dip = ntohl(*((__be32 *) n->primary_key));
 	char rauht_pl[MLXSW_REG_RAUHT_LEN];
-	struct net_device *dev;
+
+	mlxsw_reg_rauht_pack4(rauht_pl, op, neigh_entry->rif, neigh_entry->ha,
+			      dip);
+	mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(rauht), rauht_pl);
+}
+
+static void
+mlxsw_sp_neigh_entry_update(struct mlxsw_sp *mlxsw_sp,
+			    struct mlxsw_sp_neigh_entry *neigh_entry,
+			    bool adding)
+{
+	if (!adding && !neigh_entry->connected)
+		return;
+	neigh_entry->connected = adding;
+	if (neigh_entry->key.n->tbl == &arp_tbl)
+		mlxsw_sp_router_neigh_entry_op4(mlxsw_sp, neigh_entry,
+						mlxsw_sp_rauht_op(adding));
+	else
+		WARN_ON_ONCE(1);
+}
+
+struct mlxsw_sp_neigh_event_work {
+	struct work_struct work;
+	struct mlxsw_sp *mlxsw_sp;
+	struct neighbour *n;
+};
+
+static void mlxsw_sp_router_neigh_event_work(struct work_struct *work)
+{
+	struct mlxsw_sp_neigh_event_work *neigh_work =
+		container_of(work, struct mlxsw_sp_neigh_event_work, work);
+	struct mlxsw_sp *mlxsw_sp = neigh_work->mlxsw_sp;
+	struct mlxsw_sp_neigh_entry *neigh_entry;
+	struct neighbour *n = neigh_work->n;
+	unsigned char ha[ETH_ALEN];
 	bool entry_connected;
 	u8 nud_state, dead;
-	bool updating;
-	bool removing;
-	bool adding;
-	u32 dip;
-	int err;
 
+	/* If these parameters are changed after we release the lock,
+	 * then we are guaranteed to receive another event letting us
+	 * know about it.
+	 */
 	read_lock_bh(&n->lock);
-	dip = ntohl(*((__be32 *) n->primary_key));
-	memcpy(neigh_entry->ha, n->ha, sizeof(neigh_entry->ha));
+	memcpy(ha, n->ha, ETH_ALEN);
 	nud_state = n->nud_state;
 	dead = n->dead;
-	dev = n->dev;
 	read_unlock_bh(&n->lock);
 
+	rtnl_lock();
 	entry_connected = nud_state & NUD_VALID && !dead;
-	adding = (!neigh_entry->offloaded) && entry_connected;
-	updating = neigh_entry->offloaded && entry_connected;
-	removing = neigh_entry->offloaded && !entry_connected;
-
-	if (adding || updating) {
-		mlxsw_reg_rauht_pack4(rauht_pl, MLXSW_REG_RAUHT_OP_WRITE_ADD,
-				      neigh_entry->rif,
-				      neigh_entry->ha, dip);
-		err = mlxsw_reg_write(mlxsw_sp->core,
-				      MLXSW_REG(rauht), rauht_pl);
-		if (err) {
-			netdev_err(dev, "Could not add neigh %pI4h\n", &dip);
-			neigh_entry->offloaded = false;
-		} else {
-			neigh_entry->offloaded = true;
-		}
-		mlxsw_sp_nexthop_neigh_update(mlxsw_sp, neigh_entry, false);
-	} else if (removing) {
-		mlxsw_reg_rauht_pack4(rauht_pl, MLXSW_REG_RAUHT_OP_WRITE_DELETE,
-				      neigh_entry->rif,
-				      neigh_entry->ha, dip);
-		err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(rauht),
-				      rauht_pl);
-		if (err) {
-			netdev_err(dev, "Could not delete neigh %pI4h\n", &dip);
-			neigh_entry->offloaded = true;
-		} else {
-			neigh_entry->offloaded = false;
-		}
-		mlxsw_sp_nexthop_neigh_update(mlxsw_sp, neigh_entry, true);
+	neigh_entry = mlxsw_sp_neigh_entry_lookup(mlxsw_sp, n);
+	if (!entry_connected && !neigh_entry)
+		goto out;
+	if (!neigh_entry) {
+		neigh_entry = mlxsw_sp_neigh_entry_create(mlxsw_sp, n);
+		if (IS_ERR(neigh_entry))
+			goto out;
 	}
 
+	memcpy(neigh_entry->ha, ha, ETH_ALEN);
+	mlxsw_sp_neigh_entry_update(mlxsw_sp, neigh_entry, entry_connected);
+	mlxsw_sp_nexthop_neigh_update(mlxsw_sp, neigh_entry, !entry_connected);
+
+	if (!neigh_entry->connected && list_empty(&neigh_entry->nexthop_list))
+		mlxsw_sp_neigh_entry_destroy(mlxsw_sp, neigh_entry);
+
+out:
+	rtnl_unlock();
 	neigh_release(n);
-	mlxsw_sp_port_dev_put(mlxsw_sp_port);
+	kfree(neigh_work);
 }
 
 int mlxsw_sp_router_netevent_event(struct notifier_block *unused,
 				   unsigned long event, void *ptr)
 {
-	struct mlxsw_sp_neigh_entry *neigh_entry;
+	struct mlxsw_sp_neigh_event_work *neigh_work;
 	struct mlxsw_sp_port *mlxsw_sp_port;
 	struct mlxsw_sp *mlxsw_sp;
 	unsigned long interval;
-	struct net_device *dev;
 	struct neigh_parms *p;
 	struct neighbour *n;
 
@@ -1028,32 +1032,31 @@ int mlxsw_sp_router_netevent_event(struct notifier_block *unused,
 		break;
 	case NETEVENT_NEIGH_UPDATE:
 		n = ptr;
-		dev = n->dev;
 
 		if (n->tbl != &arp_tbl)
 			return NOTIFY_DONE;
 
-		mlxsw_sp_port = mlxsw_sp_port_lower_dev_hold(dev);
+		mlxsw_sp_port = mlxsw_sp_port_lower_dev_hold(n->dev);
 		if (!mlxsw_sp_port)
 			return NOTIFY_DONE;
 
-		mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
-		neigh_entry = mlxsw_sp_neigh_entry_lookup(mlxsw_sp, n);
-		if (WARN_ON(!neigh_entry)) {
+		neigh_work = kzalloc(sizeof(*neigh_work), GFP_ATOMIC);
+		if (!neigh_work) {
 			mlxsw_sp_port_dev_put(mlxsw_sp_port);
-			return NOTIFY_DONE;
+			return NOTIFY_BAD;
 		}
-		neigh_entry->mlxsw_sp_port = mlxsw_sp_port;
+
+		INIT_WORK(&neigh_work->work, mlxsw_sp_router_neigh_event_work);
+		neigh_work->mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+		neigh_work->n = n;
 
 		/* Take a reference to ensure the neighbour won't be
 		 * destructed until we drop the reference in delayed
 		 * work.
 		 */
 		neigh_clone(n);
-		if (!mlxsw_core_schedule_work(&neigh_entry->work)) {
-			neigh_release(n);
-			mlxsw_sp_port_dev_put(mlxsw_sp_port);
-		}
+		mlxsw_core_schedule_work(&neigh_work->work);
+		mlxsw_sp_port_dev_put(mlxsw_sp_port);
 		break;
 	}
 
@@ -1334,14 +1337,11 @@ mlxsw_sp_nexthop_neigh_update(struct mlxsw_sp *mlxsw_sp,
 {
 	struct mlxsw_sp_nexthop *nh;
 
-	/* Take RTNL mutex here to prevent lists from changes */
-	rtnl_lock();
 	list_for_each_entry(nh, &neigh_entry->nexthop_list,
 			    neigh_list_node) {
 		__mlxsw_sp_nexthop_neigh_update(nh, removing);
 		mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh->nh_grp);
 	}
-	rtnl_unlock();
 }
 
 static int mlxsw_sp_nexthop_init(struct mlxsw_sp *mlxsw_sp,
@@ -1368,8 +1368,11 @@ static int mlxsw_sp_nexthop_init(struct mlxsw_sp *mlxsw_sp,
 	}
 	neigh_entry = mlxsw_sp_neigh_entry_lookup(mlxsw_sp, n);
 	if (!neigh_entry) {
-		neigh_release(n);
-		return -EINVAL;
+		neigh_entry = mlxsw_sp_neigh_entry_create(mlxsw_sp, n);
+		if (IS_ERR(neigh_entry)) {
+			neigh_release(n);
+			return -EINVAL;
+		}
 	}
 
 	/* If that is the first nexthop connected to that neigh, add to
@@ -1395,6 +1398,7 @@ static void mlxsw_sp_nexthop_fini(struct mlxsw_sp *mlxsw_sp,
 				  struct mlxsw_sp_nexthop *nh)
 {
 	struct mlxsw_sp_neigh_entry *neigh_entry = nh->neigh_entry;
+	struct neighbour *n = neigh_entry->key.n;
 
 	__mlxsw_sp_nexthop_neigh_update(nh, true);
 	list_del(&nh->neigh_list_node);
@@ -1405,7 +1409,10 @@ static void mlxsw_sp_nexthop_fini(struct mlxsw_sp *mlxsw_sp,
 	if (list_empty(&nh->neigh_entry->nexthop_list))
 		list_del(&nh->neigh_entry->nexthop_neighs_list_node);
 
-	neigh_release(neigh_entry->key.n);
+	if (!neigh_entry->connected && list_empty(&neigh_entry->nexthop_list))
+		mlxsw_sp_neigh_entry_destroy(mlxsw_sp, neigh_entry);
+
+	neigh_release(n);
 }
 
 static struct mlxsw_sp_nexthop_group *
-- 
2.7.4

[patch net-next 5/8] net: remove ndo_neigh_{construct,destroy} from stacked devices

[Jiri Pirko]

From: Ido Schimmel <idosch@mellanox.com>

In commit 18bfb924f000 ("net: introduce default neigh_construct/destroy
ndo calls for L2 upper devices") we added these ndos to stacked devices
such as team and bond, so that calls will be propagated to mlxsw.

However, previous commit removed the reliance on these ndos and no new
users of these ndos have appeared since above mentioned commit. We can
therefore safely remove this dead code.

Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 drivers/net/bonding/bond_main.c |  2 --
 drivers/net/team/team.c         |  2 --
 include/linux/netdevice.h       |  4 ----
 net/8021q/vlan_dev.c            |  2 --
 net/bridge/br_device.c          |  2 --
 net/core/dev.c                  | 44 -----------------------------------------
 6 files changed, 56 deletions(-)

diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
index e6af047..6321f12 100644
--- a/drivers/net/bonding/bond_main.c
+++ b/drivers/net/bonding/bond_main.c
@@ -4145,8 +4145,6 @@ static const struct net_device_ops bond_netdev_ops = {
 	.ndo_add_slave		= bond_enslave,
 	.ndo_del_slave		= bond_release,
 	.ndo_fix_features	= bond_fix_features,
-	.ndo_neigh_construct	= netdev_default_l2upper_neigh_construct,
-	.ndo_neigh_destroy	= netdev_default_l2upper_neigh_destroy,
 	.ndo_bridge_setlink	= switchdev_port_bridge_setlink,
 	.ndo_bridge_getlink	= switchdev_port_bridge_getlink,
 	.ndo_bridge_dellink	= switchdev_port_bridge_dellink,
diff --git a/drivers/net/team/team.c b/drivers/net/team/team.c
index a371176..4a24b5d 100644
--- a/drivers/net/team/team.c
+++ b/drivers/net/team/team.c
@@ -2001,8 +2001,6 @@ static const struct net_device_ops team_netdev_ops = {
 	.ndo_add_slave		= team_add_slave,
 	.ndo_del_slave		= team_del_slave,
 	.ndo_fix_features	= team_fix_features,
-	.ndo_neigh_construct	= netdev_default_l2upper_neigh_construct,
-	.ndo_neigh_destroy	= netdev_default_l2upper_neigh_destroy,
 	.ndo_change_carrier     = team_change_carrier,
 	.ndo_bridge_setlink	= switchdev_port_bridge_setlink,
 	.ndo_bridge_getlink	= switchdev_port_bridge_getlink,
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 014fbe2..58afbd1 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -3888,10 +3888,6 @@ void *netdev_lower_dev_get_private(struct net_device *dev,
 				   struct net_device *lower_dev);
 void netdev_lower_state_changed(struct net_device *lower_dev,
 				void *lower_state_info);
-int netdev_default_l2upper_neigh_construct(struct net_device *dev,
-					   struct neighbour *n);
-void netdev_default_l2upper_neigh_destroy(struct net_device *dev,
-					  struct neighbour *n);
 
 /* RSS keys are 40 or 52 bytes long */
 #define NETDEV_RSS_KEY_LEN 52
diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c
index 116455a..e97ab82 100644
--- a/net/8021q/vlan_dev.c
+++ b/net/8021q/vlan_dev.c
@@ -791,8 +791,6 @@ static const struct net_device_ops vlan_netdev_ops = {
 	.ndo_netpoll_cleanup	= vlan_dev_netpoll_cleanup,
 #endif
 	.ndo_fix_features	= vlan_dev_fix_features,
-	.ndo_neigh_construct	= netdev_default_l2upper_neigh_construct,
-	.ndo_neigh_destroy	= netdev_default_l2upper_neigh_destroy,
 	.ndo_fdb_add		= switchdev_port_fdb_add,
 	.ndo_fdb_del		= switchdev_port_fdb_del,
 	.ndo_fdb_dump		= switchdev_port_fdb_dump,
diff --git a/net/bridge/br_device.c b/net/bridge/br_device.c
index 6c46d1b..5ba0b55 100644
--- a/net/bridge/br_device.c
+++ b/net/bridge/br_device.c
@@ -347,8 +347,6 @@ static const struct net_device_ops br_netdev_ops = {
 	.ndo_add_slave		 = br_add_slave,
 	.ndo_del_slave		 = br_del_slave,
 	.ndo_fix_features        = br_fix_features,
-	.ndo_neigh_construct	 = netdev_default_l2upper_neigh_construct,
-	.ndo_neigh_destroy	 = netdev_default_l2upper_neigh_destroy,
 	.ndo_fdb_add		 = br_fdb_add,
 	.ndo_fdb_del		 = br_fdb_delete,
 	.ndo_fdb_dump		 = br_fdb_dump,
diff --git a/net/core/dev.c b/net/core/dev.c
index 404d2e6..3e1a601 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -6111,50 +6111,6 @@ void netdev_lower_state_changed(struct net_device *lower_dev,
 }
 EXPORT_SYMBOL(netdev_lower_state_changed);
 
-int netdev_default_l2upper_neigh_construct(struct net_device *dev,
-					   struct neighbour *n)
-{
-	struct net_device *lower_dev, *stop_dev;
-	struct list_head *iter;
-	int err;
-
-	netdev_for_each_lower_dev(dev, lower_dev, iter) {
-		if (!lower_dev->netdev_ops->ndo_neigh_construct)
-			continue;
-		err = lower_dev->netdev_ops->ndo_neigh_construct(lower_dev, n);
-		if (err) {
-			stop_dev = lower_dev;
-			goto rollback;
-		}
-	}
-	return 0;
-
-rollback:
-	netdev_for_each_lower_dev(dev, lower_dev, iter) {
-		if (lower_dev == stop_dev)
-			break;
-		if (!lower_dev->netdev_ops->ndo_neigh_destroy)
-			continue;
-		lower_dev->netdev_ops->ndo_neigh_destroy(lower_dev, n);
-	}
-	return err;
-}
-EXPORT_SYMBOL_GPL(netdev_default_l2upper_neigh_construct);
-
-void netdev_default_l2upper_neigh_destroy(struct net_device *dev,
-					  struct neighbour *n)
-{
-	struct net_device *lower_dev;
-	struct list_head *iter;
-
-	netdev_for_each_lower_dev(dev, lower_dev, iter) {
-		if (!lower_dev->netdev_ops->ndo_neigh_destroy)
-			continue;
-		lower_dev->netdev_ops->ndo_neigh_destroy(lower_dev, n);
-	}
-}
-EXPORT_SYMBOL_GPL(netdev_default_l2upper_neigh_destroy);
-
 static void dev_change_rx_flags(struct net_device *dev, int flags)
 {
 	const struct net_device_ops *ops = dev->netdev_ops;
-- 
2.7.4

[patch net-next 6/8] mlxsw: spectrum_router: Remove redundant check

[Jiri Pirko]

From: Ido Schimmel <idosch@mellanox.com>

We only add neighbour entries that are also used for nexthops to
'nexthop_neighs_list', so when iterating over this list there's no need
to check that the entry is indeed used for nexthops.

Remove the redundant check.

Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c | 12 ++++--------
 1 file changed, 4 insertions(+), 8 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index c338b08..98a6d03 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -847,13 +847,11 @@ static void mlxsw_sp_router_neighs_update_nh(struct mlxsw_sp *mlxsw_sp)
 	/* Take RTNL mutex here to prevent lists from changes */
 	rtnl_lock();
 	list_for_each_entry(neigh_entry, &mlxsw_sp->router.nexthop_neighs_list,
-			    nexthop_neighs_list_node) {
+			    nexthop_neighs_list_node)
 		/* If this neigh have nexthops, make the kernel think this neigh
 		 * is active regardless of the traffic.
 		 */
-		if (!list_empty(&neigh_entry->nexthop_list))
-			neigh_event_send(neigh_entry->key.n, NULL);
-	}
+		neigh_event_send(neigh_entry->key.n, NULL);
 	rtnl_unlock();
 }
 
@@ -897,11 +895,9 @@ static void mlxsw_sp_router_probe_unresolved_nexthops(struct work_struct *work)
 	 */
 	rtnl_lock();
 	list_for_each_entry(neigh_entry, &mlxsw_sp->router.nexthop_neighs_list,
-			    nexthop_neighs_list_node) {
-		if (!(neigh_entry->key.n->nud_state & NUD_VALID) &&
-		    !list_empty(&neigh_entry->nexthop_list))
+			    nexthop_neighs_list_node)
+		if (!(neigh_entry->key.n->nud_state & NUD_VALID))
 			neigh_event_send(neigh_entry->key.n, NULL);
-	}
 	rtnl_unlock();
 
 	mlxsw_core_schedule_dw(&mlxsw_sp->router.nexthop_probe_dw,
-- 
2.7.4

[patch net-next 7/8] mlxsw: spectrum_router: Don't read 'nud_state' without lock

[Jiri Pirko]

From: Ido Schimmel <idosch@mellanox.com>

We periodically ask the neighbouring system to try and resolve
neighbours that are used for nexthops, but aren't currently resolved.

However, 'nud_state' is protected by the neighbour lock, so we shouldn't
access it without taking it. Instead, we can simply check the
'connected' field of the neighbour entry, which we update upon
NEIGH_UPDATE events.

Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 98a6d03..ec022f9 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -896,7 +896,7 @@ static void mlxsw_sp_router_probe_unresolved_nexthops(struct work_struct *work)
 	rtnl_lock();
 	list_for_each_entry(neigh_entry, &mlxsw_sp->router.nexthop_neighs_list,
 			    nexthop_neighs_list_node)
-		if (!(neigh_entry->key.n->nud_state & NUD_VALID))
+		if (!neigh_entry->connected)
 			neigh_event_send(neigh_entry->key.n, NULL);
 	rtnl_unlock();
 
-- 
2.7.4

[patch net-next 8/8] mlxsw: spectrum_router: Fix typo in comment

[Jiri Pirko]

From: Ido Schimmel <idosch@mellanox.com>

Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index ec022f9..b19f69f 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -1353,7 +1353,7 @@ static int mlxsw_sp_nexthop_init(struct mlxsw_sp *mlxsw_sp,
 	/* Take a reference of neigh here ensuring that neigh would
 	 * not be detructed before the nexthop entry is finished.
 	 * The reference is taken either in neigh_lookup() or
-	 * in neith_create() in case n is not found.
+	 * in neigh_create() in case n is not found.
 	 */
 	n = neigh_lookup(&arp_tbl, &fib_nh->nh_gw, dev);
 	if (!n) {
-- 
2.7.4

Re: [patch net-next 0/8] mlxsw: cleanup neigh handling

[David Miller]

From: Jiri Pirko <jiri@resnulli.us>
Date: Mon,  6 Feb 2017 16:20:09 +0100

> From: Jiri Pirko <jiri@mellanox.com>
> 
> Ido says:
> 
> This series addresses long standing issues in the mlxsw driver
> concerning neighbour reflection. It also prepares the code for follow-up
> changes dealing with proper resource cleanup and nexthop reflection.
> 
> The first two patches convert the neighbour reflection code to use an
> ordered workqueue, to prevent re-ordering of NEIGH_UPDATE events that
> may happen following subsequent patches.
> 
> The third to fifth patches remove the ndo_neigh_{construct,destroy}
> entry points from the driver, thereby relying only on NEIGH_UPDATE
> events for neighbour reflection. This simplifies the code considerably.
> 
> Last patches are fallout and adjust nits in the code I noticed while
> going over it.

Series applied, thanks.

Looks like 6lowpan is the only remaining user of the ndo_neigh_*() ops
now.

Re: [patch net-next 0/8] mlxsw: cleanup neigh handling

[Jiri Pirko]

Mon, Feb 06, 2017 at 05:28:44PM CET, davem@davemloft.net wrote:
>From: Jiri Pirko <jiri@resnulli.us>
>Date: Mon,  6 Feb 2017 16:20:09 +0100
>
>> From: Jiri Pirko <jiri@mellanox.com>
>> 
>> Ido says:
>> 
>> This series addresses long standing issues in the mlxsw driver
>> concerning neighbour reflection. It also prepares the code for follow-up
>> changes dealing with proper resource cleanup and nexthop reflection.
>> 
>> The first two patches convert the neighbour reflection code to use an
>> ordered workqueue, to prevent re-ordering of NEIGH_UPDATE events that
>> may happen following subsequent patches.
>> 
>> The third to fifth patches remove the ndo_neigh_{construct,destroy}
>> entry points from the driver, thereby relying only on NEIGH_UPDATE
>> events for neighbour reflection. This simplifies the code considerably.
>> 
>> Last patches are fallout and adjust nits in the code I noticed while
>> going over it.
>
>Series applied, thanks.
>
>Looks like 6lowpan is the only remaining user of the ndo_neigh_*() ops
>now.

net/atm/clip.c