thr3ads.net - Linux Ethernet Bridging - [Bridge] [PATCH net-next 00/16] mlxsw: Add VxLAN learning support [Nov 2018]

If this information is useful, please help other people find it:
Share via:

Ido Schimmel

2018-Nov-21 08:02 UTC

[Bridge] [PATCH net-next 00/16] mlxsw: Add VxLAN learning support

This patchset adds VxLAN learning support in the mlxsw driver.

The first five patches from Petr add the required switchdev APIs which
allow device drivers to notify the VxLAN driver about learned / aged-out
FDB entries.

First in patch #1, an unnecessary argument is dropped from
__vxlan_fdb_delete().

In patches #2-#4, the VxLAN FDB handling code is extended to make
sending the switchdev events configurable; to mark user-added entries as
such; and to make sure HW-learned FDB entries do not take over
user-added ones.

Finally in patch #5, the necessary switchdev notifications are added and
handled by VxLAN, similarly to how this is handled in the bridge driver.

Patch #6 allows changing of the VxLAN's device ageing time since it is
useful for the selftest in the last patch.

Patch #7 adds support for querying bridge port flags of a given
netdevice, as a new entry should not be learned and notified to the
bridge driver in case learning is disabled on the bridge port.

Next patches gradually add learning support in mlxsw.

The last patch adds a new test case for VxLAN learning.

Ido Schimmel (11):
  vxlan: Allow changing ageing time
  bridge: Allow querying bridge port flags
  mlxsw: reg: Add definition of unicast tunnel record for SFN register
  mlxsw: spectrum_fid: Store ifindex of NVE device in FID
  mlxsw: spectrum_fid: Allow FID lookup by its index
  mlxsw: spectrum_nve: Add API to resolve learned IP addresses
  mlxsw: spectrum_switchdev: Process learned VxLAN FDB entries
  mlxsw: spectrum_switchdev: Allow deletion of learned FDB entries
  mlxsw: spectrum_nve: Allow VxLAN learning
  selftests: mlxsw: Consider VxLAN learning enabled as valid
  selftests: forwarding: vxlan_bridge_1d: Add learning test

Petr Machata (5):
  vxlan: __vxlan_fdb_delete(): Drop unused argument vid
  vxlan: vxlan_fdb_notify(): Make switchdev notification configurable
  vxlan: Mark user-added FDB entries
  vxlan: Don't override user-added entries with ext-learned ones
  vxlan: Add hardware FDB learning

 drivers/net/ethernet/mellanox/mlxsw/reg.h     |  64 ++++++
 .../net/ethernet/mellanox/mlxsw/spectrum.h    |   8 +-
 .../ethernet/mellanox/mlxsw/spectrum_fid.c    |  56 ++++-
 .../ethernet/mellanox/mlxsw/spectrum_nve.c    |  16 +-
 .../mellanox/mlxsw/spectrum_nve_vxlan.c       |   8 +-
 .../mellanox/mlxsw/spectrum_switchdev.c       | 206 ++++++++++++++++--
 drivers/net/vxlan.c                           | 192 ++++++++++++----
 include/linux/if_bridge.h                     |   6 +
 include/net/switchdev.h                       |   2 +
 include/net/vxlan.h                           |   1 +
 net/bridge/br_if.c                            |  12 +
 .../selftests/drivers/net/mlxsw/vxlan.sh      |   2 +-
 .../net/forwarding/vxlan_bridge_1d.sh         | 108 +++++++++
 13 files changed, 606 insertions(+), 75 deletions(-)

-- 
2.19.1

Ido Schimmel

2018-Nov-21 08:02 UTC

head link

[Bridge] [PATCH net-next 01/16] vxlan: __vxlan_fdb_delete(): Drop unused argument vid

From: Petr Machata <petrm at mellanox.com>

This argument is necessary for vxlan_fdb_delete(), the API of which is
prescribed by ndo_fdb_del, but __vxlan_fdb_delete() doesn't need it.
Therefore drop it.

Signed-off-by: Petr Machata <petrm at mellanox.com>
Signed-off-by: Ido Schimmel <idosch at mellanox.com>
---
 drivers/net/vxlan.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index c3e65e78f015..99ab7844476f 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -977,7 +977,7 @@ static int vxlan_fdb_add(struct ndmsg *ndm, struct nlattr
*tb[],
 static int __vxlan_fdb_delete(struct vxlan_dev *vxlan,
 			      const unsigned char *addr, union vxlan_addr ip,
 			      __be16 port, __be32 src_vni, __be32 vni,
-			      u32 ifindex, u16 vid)
+			      u32 ifindex)
 {
 	struct vxlan_fdb *f;
 	struct vxlan_rdst *rd = NULL;
@@ -1024,8 +1024,7 @@ static int vxlan_fdb_delete(struct ndmsg *ndm, struct
nlattr *tb[],
 		return err;
 
 	spin_lock_bh(&vxlan->hash_lock);
-	err = __vxlan_fdb_delete(vxlan, addr, ip, port, src_vni, vni, ifindex,
-				 vid);
+	err = __vxlan_fdb_delete(vxlan, addr, ip, port, src_vni, vni, ifindex);
 	spin_unlock_bh(&vxlan->hash_lock);
 
 	return err;
@@ -3611,7 +3610,7 @@ static int vxlan_changelink(struct net_device *dev, struct
nlattr *tb[],
 					   vxlan->cfg.dst_port,
 					   old_dst.remote_vni,
 					   old_dst.remote_vni,
-					   old_dst.remote_ifindex, 0);
+					   old_dst.remote_ifindex);
 
 		if (!vxlan_addr_any(&dst->remote_ip)) {
 			err = vxlan_fdb_create(vxlan, all_zeros_mac,
-- 
2.19.1

Ido Schimmel

2018-Nov-21 08:02 UTC

head link

[Bridge] [PATCH net-next 02/16] vxlan: vxlan_fdb_notify(): Make switchdev notification configurable

From: Petr Machata <petrm at mellanox.com>

In a following patch, vxlan is extended to allow hardware FDB learning.
For FDB entries learned this way, switchdev notifications should not be
sent again, because the driver already knows about these entries.

To that end, add an argument vxlan_fdb_notify() to determine whether
the switchdev notifications should be sent. Propagate the argument to
all call sites transitively, eventually passing true in all root calls.

Signed-off-by: Petr Machata <petrm at mellanox.com>
Signed-off-by: Ido Schimmel <idosch at mellanox.com>
---
 drivers/net/vxlan.c | 71 ++++++++++++++++++++++++++-------------------
 1 file changed, 41 insertions(+), 30 deletions(-)

diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index 99ab7844476f..441290464912 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -384,15 +384,19 @@ static void vxlan_fdb_switchdev_call_notifiers(struct
vxlan_dev *vxlan,
 }
 
 static void vxlan_fdb_notify(struct vxlan_dev *vxlan, struct vxlan_fdb *fdb,
-			     struct vxlan_rdst *rd, int type)
+			     struct vxlan_rdst *rd, int type, bool swdev_notify)
 {
-	switch (type) {
-	case RTM_NEWNEIGH:
-		vxlan_fdb_switchdev_call_notifiers(vxlan, fdb, rd, true);
-		break;
-	case RTM_DELNEIGH:
-		vxlan_fdb_switchdev_call_notifiers(vxlan, fdb, rd, false);
-		break;
+	if (swdev_notify) {
+		switch (type) {
+		case RTM_NEWNEIGH:
+			vxlan_fdb_switchdev_call_notifiers(vxlan, fdb, rd,
+							   true);
+			break;
+		case RTM_DELNEIGH:
+			vxlan_fdb_switchdev_call_notifiers(vxlan, fdb, rd,
+							   false);
+			break;
+		}
 	}
 
 	__vxlan_fdb_notify(vxlan, fdb, rd, type);
@@ -409,7 +413,7 @@ static void vxlan_ip_miss(struct net_device *dev, union
vxlan_addr *ipa)
 		.remote_vni = cpu_to_be32(VXLAN_N_VID),
 	};
 
-	vxlan_fdb_notify(vxlan, &f, &remote, RTM_GETNEIGH);
+	vxlan_fdb_notify(vxlan, &f, &remote, RTM_GETNEIGH, true);
 }
 
 static void vxlan_fdb_miss(struct vxlan_dev *vxlan, const u8
eth_addr[ETH_ALEN])
@@ -421,7 +425,7 @@ static void vxlan_fdb_miss(struct vxlan_dev *vxlan, const u8
eth_addr[ETH_ALEN])
 
 	memcpy(f.eth_addr, eth_addr, ETH_ALEN);
 
-	vxlan_fdb_notify(vxlan, &f, &remote, RTM_GETNEIGH);
+	vxlan_fdb_notify(vxlan, &f, &remote, RTM_GETNEIGH, true);
 }
 
 /* Hash Ethernet address */
@@ -756,7 +760,8 @@ static int vxlan_fdb_update(struct vxlan_dev *vxlan,
 			    const u8 *mac, union vxlan_addr *ip,
 			    __u16 state, __u16 flags,
 			    __be16 port, __be32 src_vni, __be32 vni,
-			    __u32 ifindex, __u8 ndm_flags)
+			    __u32 ifindex, __u8 ndm_flags,
+			    bool swdev_notify)
 {
 	__u8 fdb_flags = (ndm_flags & ~NTF_USE);
 	struct vxlan_rdst *rd = NULL;
@@ -822,7 +827,7 @@ static int vxlan_fdb_update(struct vxlan_dev *vxlan,
 	if (notify) {
 		if (rd == NULL)
 			rd = first_remote_rtnl(f);
-		vxlan_fdb_notify(vxlan, f, rd, RTM_NEWNEIGH);
+		vxlan_fdb_notify(vxlan, f, rd, RTM_NEWNEIGH, swdev_notify);
 	}
 
 	return 0;
@@ -841,7 +846,7 @@ static void vxlan_fdb_free(struct rcu_head *head)
 }
 
 static void vxlan_fdb_destroy(struct vxlan_dev *vxlan, struct vxlan_fdb *f,
-			      bool do_notify)
+			      bool do_notify, bool swdev_notify)
 {
 	struct vxlan_rdst *rd;
 
@@ -851,7 +856,8 @@ static void vxlan_fdb_destroy(struct vxlan_dev *vxlan,
struct vxlan_fdb *f,
 	--vxlan->addrcnt;
 	if (do_notify)
 		list_for_each_entry(rd, &f->remotes, list)
-			vxlan_fdb_notify(vxlan, f, rd, RTM_DELNEIGH);
+			vxlan_fdb_notify(vxlan, f, rd, RTM_DELNEIGH,
+					 swdev_notify);
 
 	hlist_del_rcu(&f->hlist);
 	call_rcu(&f->rcu, vxlan_fdb_free);
@@ -866,10 +872,10 @@ static void vxlan_dst_free(struct rcu_head *head)
 }
 
 static void vxlan_fdb_dst_destroy(struct vxlan_dev *vxlan, struct vxlan_fdb *f,
-				  struct vxlan_rdst *rd)
+				  struct vxlan_rdst *rd, bool swdev_notify)
 {
 	list_del_rcu(&rd->list);
-	vxlan_fdb_notify(vxlan, f, rd, RTM_DELNEIGH);
+	vxlan_fdb_notify(vxlan, f, rd, RTM_DELNEIGH, swdev_notify);
 	call_rcu(&rd->rcu, vxlan_dst_free);
 }
 
@@ -968,7 +974,8 @@ static int vxlan_fdb_add(struct ndmsg *ndm, struct nlattr
*tb[],
 
 	spin_lock_bh(&vxlan->hash_lock);
 	err = vxlan_fdb_update(vxlan, addr, &ip, ndm->ndm_state, flags,
-			       port, src_vni, vni, ifindex, ndm->ndm_flags);
+			       port, src_vni, vni, ifindex, ndm->ndm_flags,
+			       true);
 	spin_unlock_bh(&vxlan->hash_lock);
 
 	return err;
@@ -977,7 +984,7 @@ static int vxlan_fdb_add(struct ndmsg *ndm, struct nlattr
*tb[],
 static int __vxlan_fdb_delete(struct vxlan_dev *vxlan,
 			      const unsigned char *addr, union vxlan_addr ip,
 			      __be16 port, __be32 src_vni, __be32 vni,
-			      u32 ifindex)
+			      u32 ifindex, bool swdev_notify)
 {
 	struct vxlan_fdb *f;
 	struct vxlan_rdst *rd = NULL;
@@ -997,11 +1004,11 @@ static int __vxlan_fdb_delete(struct vxlan_dev *vxlan,
 	 * otherwise destroy the fdb entry
 	 */
 	if (rd && !list_is_singular(&f->remotes)) {
-		vxlan_fdb_dst_destroy(vxlan, f, rd);
+		vxlan_fdb_dst_destroy(vxlan, f, rd, swdev_notify);
 		goto out;
 	}
 
-	vxlan_fdb_destroy(vxlan, f, true);
+	vxlan_fdb_destroy(vxlan, f, true, swdev_notify);
 
 out:
 	return 0;
@@ -1024,7 +1031,8 @@ static int vxlan_fdb_delete(struct ndmsg *ndm, struct
nlattr *tb[],
 		return err;
 
 	spin_lock_bh(&vxlan->hash_lock);
-	err = __vxlan_fdb_delete(vxlan, addr, ip, port, src_vni, vni, ifindex);
+	err = __vxlan_fdb_delete(vxlan, addr, ip, port, src_vni, vni, ifindex,
+				 true);
 	spin_unlock_bh(&vxlan->hash_lock);
 
 	return err;
@@ -1102,7 +1110,7 @@ static bool vxlan_snoop(struct net_device *dev,
 
 		rdst->remote_ip = *src_ip;
 		f->updated = jiffies;
-		vxlan_fdb_notify(vxlan, f, rdst, RTM_NEWNEIGH);
+		vxlan_fdb_notify(vxlan, f, rdst, RTM_NEWNEIGH, true);
 	} else {
 		/* learned new entry */
 		spin_lock(&vxlan->hash_lock);
@@ -1115,7 +1123,7 @@ static bool vxlan_snoop(struct net_device *dev,
 					 vxlan->cfg.dst_port,
 					 vni,
 					 vxlan->default_dst.remote_vni,
-					 ifindex, NTF_SELF);
+					 ifindex, NTF_SELF, true);
 		spin_unlock(&vxlan->hash_lock);
 	}
 
@@ -2499,7 +2507,7 @@ static void vxlan_cleanup(struct timer_list *t)
 					   "garbage collect %pM\n",
 					   f->eth_addr);
 				f->state = NUD_STALE;
-				vxlan_fdb_destroy(vxlan, f, true);
+				vxlan_fdb_destroy(vxlan, f, true, true);
 			} else if (time_before(timeout, next_timer))
 				next_timer = timeout;
 		}
@@ -2550,7 +2558,7 @@ static void vxlan_fdb_delete_default(struct vxlan_dev
*vxlan, __be32 vni)
 	spin_lock_bh(&vxlan->hash_lock);
 	f = __vxlan_find_mac(vxlan, all_zeros_mac, vni);
 	if (f)
-		vxlan_fdb_destroy(vxlan, f, true);
+		vxlan_fdb_destroy(vxlan, f, true, true);
 	spin_unlock_bh(&vxlan->hash_lock);
 }
 
@@ -2604,7 +2612,7 @@ static void vxlan_flush(struct vxlan_dev *vxlan, bool
do_all)
 				continue;
 			/* the all_zeros_mac entry is deleted at vxlan_uninit */
 			if (!is_zero_ether_addr(f->eth_addr))
-				vxlan_fdb_destroy(vxlan, f, true);
+				vxlan_fdb_destroy(vxlan, f, true, true);
 		}
 	}
 	spin_unlock_bh(&vxlan->hash_lock);
@@ -3342,13 +3350,14 @@ static int __vxlan_dev_create(struct net *net, struct
net_device *dev,
 
 	/* notify default fdb entry */
 	if (f)
-		vxlan_fdb_notify(vxlan, f, first_remote_rtnl(f), RTM_NEWNEIGH);
+		vxlan_fdb_notify(vxlan, f, first_remote_rtnl(f), RTM_NEWNEIGH,
+				 true);
 
 	list_add(&vxlan->next, &vn->vxlan_list);
 	return 0;
 errout:
 	if (f)
-		vxlan_fdb_destroy(vxlan, f, false);
+		vxlan_fdb_destroy(vxlan, f, false, false);
 	return err;
 }
 
@@ -3610,7 +3619,8 @@ static int vxlan_changelink(struct net_device *dev, struct
nlattr *tb[],
 					   vxlan->cfg.dst_port,
 					   old_dst.remote_vni,
 					   old_dst.remote_vni,
-					   old_dst.remote_ifindex);
+					   old_dst.remote_ifindex,
+					   true);
 
 		if (!vxlan_addr_any(&dst->remote_ip)) {
 			err = vxlan_fdb_create(vxlan, all_zeros_mac,
@@ -3625,7 +3635,8 @@ static int vxlan_changelink(struct net_device *dev, struct
nlattr *tb[],
 				spin_unlock_bh(&vxlan->hash_lock);
 				return err;
 			}
-			vxlan_fdb_notify(vxlan, f, first_remote_rtnl(f), RTM_NEWNEIGH);
+			vxlan_fdb_notify(vxlan, f, first_remote_rtnl(f),
+					 RTM_NEWNEIGH, true);
 		}
 		spin_unlock_bh(&vxlan->hash_lock);
 	}
-- 
2.19.1

Ido Schimmel

2018-Nov-21 08:02 UTC

head link

[Bridge] [PATCH net-next 03/16] vxlan: Mark user-added FDB entries

From: Petr Machata <petrm at mellanox.com>

The VXLAN driver needs to differentiate between FDB entries learned by
the VXLAN driver, and those added by the user. The latter ones shouldn't
be taken over by external learning events. This is in accordance with
bridge behavior.

Therefore, extend the flags bitfield to 16 bits and add a new private
NTF flag to mark the user-added entries.

This seems preferable to adding a dedicated boolean, because passing the
flag, unlike passing e.g. a true, makes it clear what the meaning of the
bit is.

Signed-off-by: Petr Machata <petrm at mellanox.com>
Signed-off-by: Ido Schimmel <idosch at mellanox.com>
---
 drivers/net/vxlan.c | 17 +++++++++++------
 include/net/vxlan.h |  1 +
 2 files changed, 12 insertions(+), 6 deletions(-)

diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index 441290464912..9171c1f42fe9 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -79,9 +79,11 @@ struct vxlan_fdb {
 	u8		  eth_addr[ETH_ALEN];
 	u16		  state;	/* see ndm_state */
 	__be32		  vni;
-	u8		  flags;	/* see ndm_flags */
+	u16		  flags;	/* see ndm_flags and below */
 };
 
+#define NTF_VXLAN_ADDED_BY_USER 0x100
+
 /* salt for hash table */
 static u32 vxlan_salt __read_mostly;
 
@@ -376,6 +378,7 @@ static void vxlan_fdb_switchdev_call_notifiers(struct
vxlan_dev *vxlan,
 		.remote_ifindex = rd->remote_ifindex,
 		.vni = fdb->vni,
 		.offloaded = rd->offloaded,
+		.added_by_user = fdb->flags & NTF_VXLAN_ADDED_BY_USER,
 	};
 	memcpy(info.eth_addr, fdb->eth_addr, ETH_ALEN);
 
@@ -544,6 +547,7 @@ int vxlan_fdb_find_uc(struct net_device *dev, const u8 *mac,
__be32 vni,
 	fdb_info->remote_ifindex = rdst->remote_ifindex;
 	fdb_info->vni = vni;
 	fdb_info->offloaded = rdst->offloaded;
+	fdb_info->added_by_user = f->flags & NTF_VXLAN_ADDED_BY_USER;
 	ether_addr_copy(fdb_info->eth_addr, mac);
 
 out:
@@ -704,7 +708,7 @@ static int vxlan_gro_complete(struct sock *sk, struct
sk_buff *skb, int nhoff)
 
 static struct vxlan_fdb *vxlan_fdb_alloc(struct vxlan_dev *vxlan,
 					 const u8 *mac, __u16 state,
-					 __be32 src_vni, __u8 ndm_flags)
+					 __be32 src_vni, __u16 ndm_flags)
 {
 	struct vxlan_fdb *f;
 
@@ -724,7 +728,7 @@ static struct vxlan_fdb *vxlan_fdb_alloc(struct vxlan_dev
*vxlan,
 static int vxlan_fdb_create(struct vxlan_dev *vxlan,
 			    const u8 *mac, union vxlan_addr *ip,
 			    __u16 state, __be16 port, __be32 src_vni,
-			    __be32 vni, __u32 ifindex, __u8 ndm_flags,
+			    __be32 vni, __u32 ifindex, __u16 ndm_flags,
 			    struct vxlan_fdb **fdb)
 {
 	struct vxlan_rdst *rd = NULL;
@@ -760,10 +764,10 @@ static int vxlan_fdb_update(struct vxlan_dev *vxlan,
 			    const u8 *mac, union vxlan_addr *ip,
 			    __u16 state, __u16 flags,
 			    __be16 port, __be32 src_vni, __be32 vni,
-			    __u32 ifindex, __u8 ndm_flags,
+			    __u32 ifindex, __u16 ndm_flags,
 			    bool swdev_notify)
 {
-	__u8 fdb_flags = (ndm_flags & ~NTF_USE);
+	__u16 fdb_flags = (ndm_flags & ~NTF_USE);
 	struct vxlan_rdst *rd = NULL;
 	struct vxlan_fdb *f;
 	int notify = 0;
@@ -974,7 +978,8 @@ static int vxlan_fdb_add(struct ndmsg *ndm, struct nlattr
*tb[],
 
 	spin_lock_bh(&vxlan->hash_lock);
 	err = vxlan_fdb_update(vxlan, addr, &ip, ndm->ndm_state, flags,
-			       port, src_vni, vni, ifindex, ndm->ndm_flags,
+			       port, src_vni, vni, ifindex,
+			       ndm->ndm_flags | NTF_VXLAN_ADDED_BY_USER,
 			       true);
 	spin_unlock_bh(&vxlan->hash_lock);
 
diff --git a/include/net/vxlan.h b/include/net/vxlan.h
index ec999c49df1f..b73c670df184 100644
--- a/include/net/vxlan.h
+++ b/include/net/vxlan.h
@@ -421,6 +421,7 @@ struct switchdev_notifier_vxlan_fdb_info {
 	u8 eth_addr[ETH_ALEN];
 	__be32 vni;
 	bool offloaded;
+	bool added_by_user;
 };
 
 #if IS_ENABLED(CONFIG_VXLAN)
-- 
2.19.1

Ido Schimmel

2018-Nov-21 08:02 UTC

head link

[Bridge] [PATCH net-next 04/16] vxlan: Don't override user-added entries with ext-learned ones

From: Petr Machata <petrm at mellanox.com>

When an external learning event collides with an user-added entry, the
user-added entry shouldn't be taken over. Otherwise on an unlearn event
the entry would be completely lost, even though the user added it by
hand.

Therefore skip update of FDB flags and state for these cases. This is in
accordance with the bridge behavior.

Signed-off-by: Petr Machata <petrm at mellanox.com>
Signed-off-by: Ido Schimmel <idosch at mellanox.com>
---
 drivers/net/vxlan.c | 26 +++++++++++++++++---------
 1 file changed, 17 insertions(+), 9 deletions(-)

diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index 9171c1f42fe9..b50705a50686 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -780,16 +780,24 @@ static int vxlan_fdb_update(struct vxlan_dev *vxlan,
 				   "lost race to create %pM\n", mac);
 			return -EEXIST;
 		}
-		if (f->state != state) {
-			f->state = state;
-			f->updated = jiffies;
-			notify = 1;
-		}
-		if (f->flags != fdb_flags) {
-			f->flags = fdb_flags;
-			f->updated = jiffies;
-			notify = 1;
+
+		/* Do not allow an externally learned entry to take over an
+		 * entry added by the user.
+		 */
+		if (!(fdb_flags & NTF_EXT_LEARNED) ||
+		    !(f->flags & NTF_VXLAN_ADDED_BY_USER)) {
+			if (f->state != state) {
+				f->state = state;
+				f->updated = jiffies;
+				notify = 1;
+			}
+			if (f->flags != fdb_flags) {
+				f->flags = fdb_flags;
+				f->updated = jiffies;
+				notify = 1;
+			}
 		}
+
 		if ((flags & NLM_F_REPLACE)) {
 			/* Only change unicasts */
 			if (!(is_multicast_ether_addr(f->eth_addr) ||
-- 
2.19.1

Ido Schimmel

2018-Nov-21 08:02 UTC

head link

[Bridge] [PATCH net-next 05/16] vxlan: Add hardware FDB learning

From: Petr Machata <petrm at mellanox.com>

In order to allow devices to signal learning events to VXLAN, introduce
two new switchdev messages: SWITCHDEV_VXLAN_FDB_ADD_TO_BRIDGE and
SWITCHDEV_VXLAN_FDB_DEL_TO_BRIDGE.

Listen to these notifications in the vxlan driver. The FDB entries
learned this way have an NTF_EXT_LEARNED flag, and only entries marked
as such can be unlearned by the _DEL_ event. They are also immediately
marked as offloaded. This is the same behavior that the bridge driver
observes.

Signed-off-by: Petr Machata <petrm at mellanox.com>
Signed-off-by: Ido Schimmel <idosch at mellanox.com>
---
 drivers/net/vxlan.c     | 73 ++++++++++++++++++++++++++++++++++++++++-
 include/net/switchdev.h |  2 ++
 2 files changed, 74 insertions(+), 1 deletion(-)

diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index b50705a50686..03ba1b56ba6d 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -3923,18 +3923,89 @@ vxlan_fdb_offloaded_set(struct net_device *dev,
 	spin_unlock_bh(&vxlan->hash_lock);
 }
 
+static int
+vxlan_fdb_external_learn_add(struct net_device *dev,
+			     struct switchdev_notifier_vxlan_fdb_info *fdb_info)
+{
+	struct vxlan_dev *vxlan = netdev_priv(dev);
+	int err;
+
+	spin_lock_bh(&vxlan->hash_lock);
+	err = vxlan_fdb_update(vxlan, fdb_info->eth_addr,
&fdb_info->remote_ip,
+			       NUD_REACHABLE,
+			       NLM_F_CREATE | NLM_F_REPLACE,
+			       fdb_info->remote_port,
+			       fdb_info->vni,
+			       fdb_info->remote_vni,
+			       fdb_info->remote_ifindex,
+			       NTF_USE | NTF_SELF | NTF_EXT_LEARNED,
+			       false);
+	spin_unlock_bh(&vxlan->hash_lock);
+
+	return err;
+}
+
+static int
+vxlan_fdb_external_learn_del(struct net_device *dev,
+			     struct switchdev_notifier_vxlan_fdb_info *fdb_info)
+{
+	struct vxlan_dev *vxlan = netdev_priv(dev);
+	struct vxlan_fdb *f;
+	int err = 0;
+
+	spin_lock_bh(&vxlan->hash_lock);
+
+	f = vxlan_find_mac(vxlan, fdb_info->eth_addr, fdb_info->vni);
+	if (!f)
+		err = -ENOENT;
+	else if (f->flags & NTF_EXT_LEARNED)
+		err = __vxlan_fdb_delete(vxlan, fdb_info->eth_addr,
+					 fdb_info->remote_ip,
+					 fdb_info->remote_port,
+					 fdb_info->vni,
+					 fdb_info->remote_vni,
+					 fdb_info->remote_ifindex,
+					 false);
+
+	spin_unlock_bh(&vxlan->hash_lock);
+
+	return err;
+}
+
 static int vxlan_switchdev_event(struct notifier_block *unused,
 				 unsigned long event, void *ptr)
 {
 	struct net_device *dev = switchdev_notifier_info_to_dev(ptr);
+	struct switchdev_notifier_vxlan_fdb_info *fdb_info;
+	int err = 0;
 
 	switch (event) {
 	case SWITCHDEV_VXLAN_FDB_OFFLOADED:
 		vxlan_fdb_offloaded_set(dev, ptr);
 		break;
+	case SWITCHDEV_VXLAN_FDB_ADD_TO_BRIDGE:
+		fdb_info = ptr;
+		err = vxlan_fdb_external_learn_add(dev, fdb_info);
+		if (err) {
+			err = notifier_from_errno(err);
+			break;
+		}
+		fdb_info->offloaded = true;
+		vxlan_fdb_offloaded_set(dev, fdb_info);
+		break;
+	case SWITCHDEV_VXLAN_FDB_DEL_TO_BRIDGE:
+		fdb_info = ptr;
+		err = vxlan_fdb_external_learn_del(dev, fdb_info);
+		if (err) {
+			err = notifier_from_errno(err);
+			break;
+		}
+		fdb_info->offloaded = false;
+		vxlan_fdb_offloaded_set(dev, fdb_info);
+		break;
 	}
 
-	return 0;
+	return err;
 }
 
 static struct notifier_block vxlan_switchdev_notifier_block __read_mostly = {
diff --git a/include/net/switchdev.h b/include/net/switchdev.h
index 881ecb1555bf..7b371e7c4bc6 100644
--- a/include/net/switchdev.h
+++ b/include/net/switchdev.h
@@ -146,6 +146,8 @@ enum switchdev_notifier_type {
 	SWITCHDEV_FDB_DEL_TO_DEVICE,
 	SWITCHDEV_FDB_OFFLOADED,
 
+	SWITCHDEV_VXLAN_FDB_ADD_TO_BRIDGE,
+	SWITCHDEV_VXLAN_FDB_DEL_TO_BRIDGE,
 	SWITCHDEV_VXLAN_FDB_ADD_TO_DEVICE,
 	SWITCHDEV_VXLAN_FDB_DEL_TO_DEVICE,
 	SWITCHDEV_VXLAN_FDB_OFFLOADED,
-- 
2.19.1

Ido Schimmel

2018-Nov-21 08:02 UTC

head link

[Bridge] [PATCH net-next 06/16] vxlan: Allow changing ageing time

In a similar fashion to the bridge device, allow changing the ageing
time of the VxLAN device by scheduling its timer to fire if the ageing
time changed.

One use case is selftests where learning / ageing of VxLAN FDB entries
is tested. The default ageing time is 5 minutes, which is too long for a
simple selftest.

Signed-off-by: Ido Schimmel <idosch at mellanox.com>
Reviewed-by: Petr Machata <petrm at mellanox.com>
---
 drivers/net/vxlan.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index 03ba1b56ba6d..911066299a83 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -3458,11 +3458,8 @@ static int vxlan_nl2conf(struct nlattr *tb[], struct
nlattr *data[],
 		conf->flags |= VXLAN_F_LEARN;
 	}
 
-	if (data[IFLA_VXLAN_AGEING]) {
-		if (changelink)
-			return -EOPNOTSUPP;
+	if (data[IFLA_VXLAN_AGEING])
 		conf->age_interval = nla_get_u32(data[IFLA_VXLAN_AGEING]);
-	}
 
 	if (data[IFLA_VXLAN_PROXY]) {
 		if (changelink)
@@ -3607,6 +3604,7 @@ static int vxlan_changelink(struct net_device *dev, struct
nlattr *tb[],
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct vxlan_rdst *dst = &vxlan->default_dst;
+	unsigned long old_age_interval;
 	struct vxlan_rdst old_dst;
 	struct vxlan_config conf;
 	struct vxlan_fdb *f = NULL;
@@ -3617,12 +3615,16 @@ static int vxlan_changelink(struct net_device *dev,
struct nlattr *tb[],
 	if (err)
 		return err;
 
+	old_age_interval = vxlan->cfg.age_interval;
 	memcpy(&old_dst, dst, sizeof(struct vxlan_rdst));
 
 	err = vxlan_dev_configure(vxlan->net, dev, &conf, true, extack);
 	if (err)
 		return err;
 
+	if (old_age_interval != vxlan->cfg.age_interval)
+		mod_timer(&vxlan->age_timer, jiffies);
+
 	/* handle default dst entry */
 	if (!vxlan_addr_equal(&dst->remote_ip, &old_dst.remote_ip)) {
 		spin_lock_bh(&vxlan->hash_lock);
-- 
2.19.1

Ido Schimmel

2018-Nov-21 08:02 UTC

head link

[Bridge] [PATCH net-next 07/16] bridge: Allow querying bridge port flags

Allow querying bridge port flags so that drivers capable of performing
VxLAN learning will update the bridge driver only if learning is enabled
on its bridge port corresponding to the VxLAN device.

Signed-off-by: Ido Schimmel <idosch at mellanox.com>
Reviewed-by: Petr Machata <petrm at mellanox.com>
---
 include/linux/if_bridge.h |  6 ++++++
 net/bridge/br_if.c        | 12 ++++++++++++
 2 files changed, 18 insertions(+)

diff --git a/include/linux/if_bridge.h b/include/linux/if_bridge.h
index c20c7e197d07..ef7c3d376b21 100644
--- a/include/linux/if_bridge.h
+++ b/include/linux/if_bridge.h
@@ -119,6 +119,7 @@ static inline int br_vlan_get_info(const struct net_device
*dev, u16 vid,
 struct net_device *br_fdb_find_port(const struct net_device *br_dev,
 				    const unsigned char *addr,
 				    __u16 vid);
+bool br_port_flag_is_set(const struct net_device *dev, unsigned long flag);
 #else
 static inline struct net_device *
 br_fdb_find_port(const struct net_device *br_dev,
@@ -127,6 +128,11 @@ br_fdb_find_port(const struct net_device *br_dev,
 {
 	return NULL;
 }
+static inline bool
+br_port_flag_is_set(const struct net_device *dev, unsigned long flag)
+{
+	return false;
+}
 #endif
 
 #endif
diff --git a/net/bridge/br_if.c b/net/bridge/br_if.c
index 9b46d2dc4c22..d4863f5679ac 100644
--- a/net/bridge/br_if.c
+++ b/net/bridge/br_if.c
@@ -741,3 +741,15 @@ void br_port_flags_change(struct net_bridge_port *p,
unsigned long mask)
 	if (mask & BR_NEIGH_SUPPRESS)
 		br_recalculate_neigh_suppress_enabled(br);
 }
+
+bool br_port_flag_is_set(const struct net_device *dev, unsigned long flag)
+{
+	struct net_bridge_port *p;
+
+	p = br_port_get_rtnl_rcu(dev);
+	if (!p)
+		return false;
+
+	return p->flags & flag;
+}
+EXPORT_SYMBOL_GPL(br_port_flag_is_set);
-- 
2.19.1

Ido Schimmel

2018-Nov-21 08:02 UTC

head link

[Bridge] [PATCH net-next 08/16] mlxsw: reg: Add definition of unicast tunnel record for SFN register

Will be used to process learned FDB records from an NVE tunnel.

Signed-off-by: Ido Schimmel <idosch at mellanox.com>
Reviewed-by: Petr Machata <petrm at mellanox.com>
---
 drivers/net/ethernet/mellanox/mlxsw/reg.h | 64 +++++++++++++++++++++++
 1 file changed, 64 insertions(+)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/reg.h
b/drivers/net/ethernet/mellanox/mlxsw/reg.h
index be2ffbd19e3a..5c8a38f61a93 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/reg.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/reg.h
@@ -641,6 +641,10 @@ enum mlxsw_reg_sfn_rec_type {
 	MLXSW_REG_SFN_REC_TYPE_AGED_OUT_MAC = 0x7,
 	/* Aged-out MAC address on a LAG port. */
 	MLXSW_REG_SFN_REC_TYPE_AGED_OUT_MAC_LAG = 0x8,
+	/* Learned unicast tunnel record. */
+	MLXSW_REG_SFN_REC_TYPE_LEARNED_UNICAST_TUNNEL = 0xD,
+	/* Aged-out unicast tunnel record. */
+	MLXSW_REG_SFN_REC_TYPE_AGED_OUT_UNICAST_TUNNEL = 0xE,
 };
 
 /* reg_sfn_rec_type
@@ -704,6 +708,66 @@ static inline void mlxsw_reg_sfn_mac_lag_unpack(char
*payload, int rec_index,
 	*p_lag_id = mlxsw_reg_sfn_mac_lag_lag_id_get(payload, rec_index);
 }
 
+/* reg_sfn_uc_tunnel_uip_msb
+ * When protocol is IPv4, the most significant byte of the underlay IPv4
+ * address of the remote VTEP.
+ * When protocol is IPv6, reserved.
+ * Access: RO
+ */
+MLXSW_ITEM32_INDEXED(reg, sfn, uc_tunnel_uip_msb, MLXSW_REG_SFN_BASE_LEN, 24,
+		     8, MLXSW_REG_SFN_REC_LEN, 0x08, false);
+
+enum mlxsw_reg_sfn_uc_tunnel_protocol {
+	MLXSW_REG_SFN_UC_TUNNEL_PROTOCOL_IPV4,
+	MLXSW_REG_SFN_UC_TUNNEL_PROTOCOL_IPV6,
+};
+
+/* reg_sfn_uc_tunnel_protocol
+ * IP protocol.
+ * Access: RO
+ */
+MLXSW_ITEM32_INDEXED(reg, sfn, uc_tunnel_protocol, MLXSW_REG_SFN_BASE_LEN, 27,
+		     1, MLXSW_REG_SFN_REC_LEN, 0x0C, false);
+
+/* reg_sfn_uc_tunnel_uip_lsb
+ * When protocol is IPv4, the least significant bytes of the underlay
+ * IPv4 address of the remote VTEP.
+ * When protocol is IPv6, ipv6_id to be queried from TNIPSD.
+ * Access: RO
+ */
+MLXSW_ITEM32_INDEXED(reg, sfn, uc_tunnel_uip_lsb, MLXSW_REG_SFN_BASE_LEN, 0,
+		     24, MLXSW_REG_SFN_REC_LEN, 0x0C, false);
+
+enum mlxsw_reg_sfn_tunnel_port {
+	MLXSW_REG_SFN_TUNNEL_PORT_NVE,
+	MLXSW_REG_SFN_TUNNEL_PORT_VPLS,
+	MLXSW_REG_SFN_TUNNEL_FLEX_TUNNEL0,
+	MLXSW_REG_SFN_TUNNEL_FLEX_TUNNEL1,
+};
+
+/* reg_sfn_uc_tunnel_port
+ * Tunnel port.
+ * Reserved on Spectrum.
+ * Access: RO
+ */
+MLXSW_ITEM32_INDEXED(reg, sfn, tunnel_port, MLXSW_REG_SFN_BASE_LEN, 0, 4,
+		     MLXSW_REG_SFN_REC_LEN, 0x10, false);
+
+static inline void
+mlxsw_reg_sfn_uc_tunnel_unpack(char *payload, int rec_index, char *mac,
+			       u16 *p_fid, u32 *p_uip,
+			       enum mlxsw_reg_sfn_uc_tunnel_protocol *p_proto)
+{
+	u32 uip_msb, uip_lsb;
+
+	mlxsw_reg_sfn_rec_mac_memcpy_from(payload, rec_index, mac);
+	*p_fid = mlxsw_reg_sfn_mac_fid_get(payload, rec_index);
+	uip_msb = mlxsw_reg_sfn_uc_tunnel_uip_msb_get(payload, rec_index);
+	uip_lsb = mlxsw_reg_sfn_uc_tunnel_uip_lsb_get(payload, rec_index);
+	*p_uip = uip_msb << 24 | uip_lsb;
+	*p_proto = mlxsw_reg_sfn_uc_tunnel_protocol_get(payload, rec_index);
+}
+
 /* SPMS - Switch Port MSTP/RSTP State Register
  * -------------------------------------------
  * Configures the spanning tree state of a physical port.
-- 
2.19.1

Ido Schimmel

2018-Nov-21 08:02 UTC

head link

[Bridge] [PATCH net-next 09/16] mlxsw: spectrum_fid: Store ifindex of NVE device in FID

The driver periodically polls for new FDB entries learned by the device.
In the case of an FDB entry learned from a VxLAN tunnel, the
notification includes the IP of the remote VTEP, the filtering
identifier (FID) and the source MAC address of the overlay packet.

Assuming learning is enabled in the VxLAN and bridge drivers, the driver
needs to generate a notification and update them about the new FDB
entry.

Store the ifindex of the NVE device in the FID so that the driver will
be able to update the VxLAN and bridge drivers using it.

Signed-off-by: Ido Schimmel <idosch at mellanox.com>
Reviewed-by: Petr Machata <petrm at mellanox.com>
---
 drivers/net/ethernet/mellanox/mlxsw/spectrum.h     |  3 ++-
 drivers/net/ethernet/mellanox/mlxsw/spectrum_fid.c | 14 +++++++++++++-
 drivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c |  2 +-
 3 files changed, 16 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
index 0875a79cbe7b..b8f00863a0c6 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
@@ -721,6 +721,7 @@ int mlxsw_sp_setup_tc_prio(struct mlxsw_sp_port
*mlxsw_sp_port,
 			   struct tc_prio_qopt_offload *p);
 
 /* spectrum_fid.c */
+int mlxsw_sp_fid_nve_ifindex(const struct mlxsw_sp_fid *fid, int *nve_ifindex);
 struct mlxsw_sp_fid *mlxsw_sp_fid_lookup_by_vni(struct mlxsw_sp *mlxsw_sp,
 						__be32 vni);
 int mlxsw_sp_fid_vni(const struct mlxsw_sp_fid *fid, __be32 *vni);
@@ -728,7 +729,7 @@ int mlxsw_sp_fid_nve_flood_index_set(struct mlxsw_sp_fid
*fid,
 				     u32 nve_flood_index);
 void mlxsw_sp_fid_nve_flood_index_clear(struct mlxsw_sp_fid *fid);
 bool mlxsw_sp_fid_nve_flood_index_is_set(const struct mlxsw_sp_fid *fid);
-int mlxsw_sp_fid_vni_set(struct mlxsw_sp_fid *fid, __be32 vni);
+int mlxsw_sp_fid_vni_set(struct mlxsw_sp_fid *fid, __be32 vni, int
nve_ifindex);
 void mlxsw_sp_fid_vni_clear(struct mlxsw_sp_fid *fid);
 bool mlxsw_sp_fid_vni_is_set(const struct mlxsw_sp_fid *fid);
 int mlxsw_sp_fid_flood_set(struct mlxsw_sp_fid *fid,
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_fid.c
b/drivers/net/ethernet/mellanox/mlxsw/spectrum_fid.c
index a3db033d7399..ad2bc8e3b2ed 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_fid.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_fid.c
@@ -30,6 +30,7 @@ struct mlxsw_sp_fid {
 	struct rhash_head vni_ht_node;
 	__be32 vni;
 	u32 nve_flood_index;
+	int nve_ifindex;
 	u8 vni_valid:1,
 	   nve_flood_index_valid:1;
 };
@@ -113,6 +114,16 @@ static const int *mlxsw_sp_packet_type_sfgc_types[] = {
 	[MLXSW_SP_FLOOD_TYPE_MC]	= mlxsw_sp_sfgc_mc_packet_types,
 };
 
+int mlxsw_sp_fid_nve_ifindex(const struct mlxsw_sp_fid *fid, int *nve_ifindex)
+{
+	if (!fid->vni_valid)
+		return -EINVAL;
+
+	*nve_ifindex = fid->nve_ifindex;
+
+	return 0;
+}
+
 struct mlxsw_sp_fid *mlxsw_sp_fid_lookup_by_vni(struct mlxsw_sp *mlxsw_sp,
 						__be32 vni)
 {
@@ -173,7 +184,7 @@ bool mlxsw_sp_fid_nve_flood_index_is_set(const struct
mlxsw_sp_fid *fid)
 	return fid->nve_flood_index_valid;
 }
 
-int mlxsw_sp_fid_vni_set(struct mlxsw_sp_fid *fid, __be32 vni)
+int mlxsw_sp_fid_vni_set(struct mlxsw_sp_fid *fid, __be32 vni, int nve_ifindex)
 {
 	struct mlxsw_sp_fid_family *fid_family = fid->fid_family;
 	const struct mlxsw_sp_fid_ops *ops = fid_family->ops;
@@ -183,6 +194,7 @@ int mlxsw_sp_fid_vni_set(struct mlxsw_sp_fid *fid, __be32
vni)
 	if (WARN_ON(!ops->vni_set || fid->vni_valid))
 		return -EINVAL;
 
+	fid->nve_ifindex = nve_ifindex;
 	fid->vni = vni;
 	err = rhashtable_lookup_insert_fast(&mlxsw_sp->fid_core->vni_ht,
 					    &fid->vni_ht_node,
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c
b/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c
index ad06d9969bc1..4498e3b3c3f0 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c
@@ -803,7 +803,7 @@ int mlxsw_sp_nve_fid_enable(struct mlxsw_sp *mlxsw_sp,
struct mlxsw_sp_fid *fid,
 		return err;
 	}
 
-	err = mlxsw_sp_fid_vni_set(fid, params->vni);
+	err = mlxsw_sp_fid_vni_set(fid, params->vni, params->dev->ifindex);
 	if (err) {
 		NL_SET_ERR_MSG_MOD(extack, "Failed to set VNI on FID");
 		goto err_fid_vni_set;
-- 
2.19.1

Ido Schimmel

2018-Nov-21 08:02 UTC

head link

[Bridge] [PATCH net-next 10/16] mlxsw: spectrum_fid: Allow FID lookup by its index

When processing a notification about a new FDB entry learned from a
VxLAN tunnel, the driver is provided with the FID index among other
parameters.

The driver potentially needs to update the bridge and VxLAN drivers
about the new entry using a pointer to the VxLAN device and the
corresponding VNI.

These two parameters are stored in the FID, so add a new function that
allows looking up a FID based on its index.

Signed-off-by: Ido Schimmel <idosch at mellanox.com>
Reviewed-by: Petr Machata <petrm at mellanox.com>
---
 .../net/ethernet/mellanox/mlxsw/spectrum.h    |  2 +
 .../ethernet/mellanox/mlxsw/spectrum_fid.c    | 42 ++++++++++++++++++-
 2 files changed, 42 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
index b8f00863a0c6..879ac65886af 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
@@ -721,6 +721,8 @@ int mlxsw_sp_setup_tc_prio(struct mlxsw_sp_port
*mlxsw_sp_port,
 			   struct tc_prio_qopt_offload *p);
 
 /* spectrum_fid.c */
+struct mlxsw_sp_fid *mlxsw_sp_fid_lookup_by_index(struct mlxsw_sp *mlxsw_sp,
+						  u16 fid_index);
 int mlxsw_sp_fid_nve_ifindex(const struct mlxsw_sp_fid *fid, int *nve_ifindex);
 struct mlxsw_sp_fid *mlxsw_sp_fid_lookup_by_vni(struct mlxsw_sp *mlxsw_sp,
 						__be32 vni);
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_fid.c
b/drivers/net/ethernet/mellanox/mlxsw/spectrum_fid.c
index ad2bc8e3b2ed..71b2d20afcc2 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_fid.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_fid.c
@@ -15,6 +15,7 @@
 struct mlxsw_sp_fid_family;
 
 struct mlxsw_sp_fid_core {
+	struct rhashtable fid_ht;
 	struct rhashtable vni_ht;
 	struct mlxsw_sp_fid_family *fid_family_arr[MLXSW_SP_FID_TYPE_MAX];
 	unsigned int *port_fid_mappings;
@@ -26,6 +27,7 @@ struct mlxsw_sp_fid {
 	unsigned int ref_count;
 	u16 fid_index;
 	struct mlxsw_sp_fid_family *fid_family;
+	struct rhash_head ht_node;
 
 	struct rhash_head vni_ht_node;
 	__be32 vni;
@@ -45,6 +47,12 @@ struct mlxsw_sp_fid_8021d {
 	int br_ifindex;
 };
 
+static const struct rhashtable_params mlxsw_sp_fid_ht_params = {
+	.key_len = sizeof_field(struct mlxsw_sp_fid, fid_index),
+	.key_offset = offsetof(struct mlxsw_sp_fid, fid_index),
+	.head_offset = offsetof(struct mlxsw_sp_fid, ht_node),
+};
+
 static const struct rhashtable_params mlxsw_sp_fid_vni_ht_params = {
 	.key_len = sizeof_field(struct mlxsw_sp_fid, vni),
 	.key_offset = offsetof(struct mlxsw_sp_fid, vni),
@@ -114,6 +122,19 @@ static const int *mlxsw_sp_packet_type_sfgc_types[] = {
 	[MLXSW_SP_FLOOD_TYPE_MC]	= mlxsw_sp_sfgc_mc_packet_types,
 };
 
+struct mlxsw_sp_fid *mlxsw_sp_fid_lookup_by_index(struct mlxsw_sp *mlxsw_sp,
+						  u16 fid_index)
+{
+	struct mlxsw_sp_fid *fid;
+
+	fid = rhashtable_lookup_fast(&mlxsw_sp->fid_core->fid_ht,
&fid_index,
+				     mlxsw_sp_fid_ht_params);
+	if (fid)
+		fid->ref_count++;
+
+	return fid;
+}
+
 int mlxsw_sp_fid_nve_ifindex(const struct mlxsw_sp_fid *fid, int *nve_ifindex)
 {
 	if (!fid->vni_valid)
@@ -956,10 +977,17 @@ static struct mlxsw_sp_fid *mlxsw_sp_fid_get(struct
mlxsw_sp *mlxsw_sp,
 	if (err)
 		goto err_configure;
 
+	err = rhashtable_insert_fast(&mlxsw_sp->fid_core->fid_ht,
&fid->ht_node,
+				     mlxsw_sp_fid_ht_params);
+	if (err)
+		goto err_rhashtable_insert;
+
 	list_add(&fid->list, &fid_family->fids_list);
 	fid->ref_count++;
 	return fid;
 
+err_rhashtable_insert:
+	fid->fid_family->ops->deconfigure(fid);
 err_configure:
 	__clear_bit(fid_index - fid_family->start_index,
 		    fid_family->fids_bitmap);
@@ -971,6 +999,7 @@ static struct mlxsw_sp_fid *mlxsw_sp_fid_get(struct mlxsw_sp
*mlxsw_sp,
 void mlxsw_sp_fid_put(struct mlxsw_sp_fid *fid)
 {
 	struct mlxsw_sp_fid_family *fid_family = fid->fid_family;
+	struct mlxsw_sp *mlxsw_sp = fid_family->mlxsw_sp;
 
 	if (--fid->ref_count == 1 && fid->rif) {
 		/* Destroy the associated RIF and let it drop the last
@@ -979,6 +1008,8 @@ void mlxsw_sp_fid_put(struct mlxsw_sp_fid *fid)
 		return mlxsw_sp_rif_destroy(fid->rif);
 	} else if (fid->ref_count == 0) {
 		list_del(&fid->list);
+		rhashtable_remove_fast(&mlxsw_sp->fid_core->fid_ht,
+				       &fid->ht_node, mlxsw_sp_fid_ht_params);
 		fid->fid_family->ops->deconfigure(fid);
 		__clear_bit(fid->fid_index - fid_family->start_index,
 			    fid_family->fids_bitmap);
@@ -1138,9 +1169,13 @@ int mlxsw_sp_fids_init(struct mlxsw_sp *mlxsw_sp)
 		return -ENOMEM;
 	mlxsw_sp->fid_core = fid_core;
 
+	err = rhashtable_init(&fid_core->fid_ht, &mlxsw_sp_fid_ht_params);
+	if (err)
+		goto err_rhashtable_fid_init;
+
 	err = rhashtable_init(&fid_core->vni_ht,
&mlxsw_sp_fid_vni_ht_params);
 	if (err)
-		goto err_rhashtable_init;
+		goto err_rhashtable_vni_init;
 
 	fid_core->port_fid_mappings = kcalloc(max_ports, sizeof(unsigned int),
 					      GFP_KERNEL);
@@ -1169,7 +1204,9 @@ int mlxsw_sp_fids_init(struct mlxsw_sp *mlxsw_sp)
 	kfree(fid_core->port_fid_mappings);
 err_alloc_port_fid_mappings:
 	rhashtable_destroy(&fid_core->vni_ht);
-err_rhashtable_init:
+err_rhashtable_vni_init:
+	rhashtable_destroy(&fid_core->fid_ht);
+err_rhashtable_fid_init:
 	kfree(fid_core);
 	return err;
 }
@@ -1184,5 +1221,6 @@ void mlxsw_sp_fids_fini(struct mlxsw_sp *mlxsw_sp)
 					       fid_core->fid_family_arr[i]);
 	kfree(fid_core->port_fid_mappings);
 	rhashtable_destroy(&fid_core->vni_ht);
+	rhashtable_destroy(&fid_core->fid_ht);
 	kfree(fid_core);
 }
-- 
2.19.1

Ido Schimmel

2018-Nov-21 08:02 UTC

head link

[Bridge] [PATCH net-next 11/16] mlxsw: spectrum_nve: Add API to resolve learned IP addresses

FDB notifications for entries learned from an NVE tunnel contain the IP
address of the remote VTEP. In the case of IPv4 underlay, the IP address
is specified as-is. IPv6 addresses on the other hand, are specified as
handles which then need to be used to query the actual address from the
device.

Only IPv4 underlay is currently supported, so we cannot receive
notifications for IPv6 addresses and therefore an error is returned when
one tries to resolve such an address.

Signed-off-by: Ido Schimmel <idosch at mellanox.com>
Reviewed-by: Petr Machata <petrm at mellanox.com>
---
 drivers/net/ethernet/mellanox/mlxsw/spectrum.h     |  3 +++
 drivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c | 14 ++++++++++++++
 2 files changed, 17 insertions(+)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
index 879ac65886af..973a9d2901f7 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
@@ -813,6 +813,9 @@ struct mlxsw_sp_nve_params {
 extern const struct mlxsw_sp_nve_ops *mlxsw_sp1_nve_ops_arr[];
 extern const struct mlxsw_sp_nve_ops *mlxsw_sp2_nve_ops_arr[];
 
+int mlxsw_sp_nve_learned_ip_resolve(struct mlxsw_sp *mlxsw_sp, u32 uip,
+				    enum mlxsw_sp_l3proto proto,
+				    union mlxsw_sp_l3addr *addr);
 int mlxsw_sp_nve_flood_ip_add(struct mlxsw_sp *mlxsw_sp,
 			      struct mlxsw_sp_fid *fid,
 			      enum mlxsw_sp_l3proto proto,
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c
b/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c
index 4498e3b3c3f0..c4d5a0865c8f 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c
@@ -174,6 +174,20 @@ mlxsw_sp_nve_mc_record_ops_arr[] = {
 	[MLXSW_SP_L3_PROTO_IPV6] = &mlxsw_sp_nve_mc_record_ipv6_ops,
 };
 
+int mlxsw_sp_nve_learned_ip_resolve(struct mlxsw_sp *mlxsw_sp, u32 uip,
+				    enum mlxsw_sp_l3proto proto,
+				    union mlxsw_sp_l3addr *addr)
+{
+	switch (proto) {
+	case MLXSW_SP_L3_PROTO_IPV4:
+		addr->addr4 = cpu_to_be32(uip);
+		return 0;
+	default:
+		WARN_ON(1);
+		return -EINVAL;
+	}
+}
+
 static struct mlxsw_sp_nve_mc_list *
 mlxsw_sp_nve_mc_list_find(struct mlxsw_sp *mlxsw_sp,
 			  const struct mlxsw_sp_nve_mc_list_key *key)
-- 
2.19.1

Ido Schimmel

2018-Nov-21 08:02 UTC

head link

[Bridge] [PATCH net-next 12/16] mlxsw: spectrum_switchdev: Process learned VxLAN FDB entries

Start processing two new entry types in addition to current ones:
* Learned unicast tunnel entry
* Aged-out unicast tunnel entry

In both cases the device reports on a new {MAC, FID, IP address} tuple
that was learned / aged-out. Based on this notification, the driver
instructs the device to add / delete the entry to / from its database.

The driver also makes sure to notify the bridge and VxLAN drivers about
the new entry.

Signed-off-by: Ido Schimmel <idosch at mellanox.com>
Reviewed-by: Petr Machata <petrm at mellanox.com>
---
 .../mellanox/mlxsw/spectrum_switchdev.c       | 203 ++++++++++++++++--
 1 file changed, 189 insertions(+), 14 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
index 739a51f0a366..d3d1c1d00d98 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
@@ -2311,6 +2311,71 @@ void mlxsw_sp_bridge_vxlan_leave(struct mlxsw_sp
*mlxsw_sp,
 	bridge_device->ops->vxlan_leave(bridge_device, vxlan_dev);
 }
 
+static void
+mlxsw_sp_switchdev_vxlan_addr_convert(const union vxlan_addr *vxlan_addr,
+				      enum mlxsw_sp_l3proto *proto,
+				      union mlxsw_sp_l3addr *addr)
+{
+	if (vxlan_addr->sa.sa_family == AF_INET) {
+		addr->addr4 = vxlan_addr->sin.sin_addr.s_addr;
+		*proto = MLXSW_SP_L3_PROTO_IPV4;
+	} else {
+		addr->addr6 = vxlan_addr->sin6.sin6_addr;
+		*proto = MLXSW_SP_L3_PROTO_IPV6;
+	}
+}
+
+static void
+mlxsw_sp_switchdev_addr_vxlan_convert(enum mlxsw_sp_l3proto proto,
+				      const union mlxsw_sp_l3addr *addr,
+				      union vxlan_addr *vxlan_addr)
+{
+	switch (proto) {
+	case MLXSW_SP_L3_PROTO_IPV4:
+		vxlan_addr->sa.sa_family = AF_INET;
+		vxlan_addr->sin.sin_addr.s_addr = addr->addr4;
+		break;
+	case MLXSW_SP_L3_PROTO_IPV6:
+		vxlan_addr->sa.sa_family = AF_INET6;
+		vxlan_addr->sin6.sin6_addr = addr->addr6;
+		break;
+	}
+}
+
+static void mlxsw_sp_fdb_vxlan_call_notifiers(struct net_device *dev,
+					      const char *mac,
+					      enum mlxsw_sp_l3proto proto,
+					      union mlxsw_sp_l3addr *addr,
+					      __be32 vni, bool adding)
+{
+	struct switchdev_notifier_vxlan_fdb_info info;
+	struct vxlan_dev *vxlan = netdev_priv(dev);
+	enum switchdev_notifier_type type;
+
+	type = adding ? SWITCHDEV_VXLAN_FDB_ADD_TO_BRIDGE :
+			SWITCHDEV_VXLAN_FDB_DEL_TO_BRIDGE;
+	mlxsw_sp_switchdev_addr_vxlan_convert(proto, addr, &info.remote_ip);
+	info.remote_port = vxlan->cfg.dst_port;
+	info.remote_vni = vni;
+	info.remote_ifindex = 0;
+	ether_addr_copy(info.eth_addr, mac);
+	info.vni = vni;
+	info.offloaded = adding;
+	call_switchdev_notifiers(type, dev, &info.info);
+}
+
+static void mlxsw_sp_fdb_nve_call_notifiers(struct net_device *dev,
+					    const char *mac,
+					    enum mlxsw_sp_l3proto proto,
+					    union mlxsw_sp_l3addr *addr,
+					    __be32 vni,
+					    bool adding)
+{
+	if (netif_is_vxlan(dev))
+		mlxsw_sp_fdb_vxlan_call_notifiers(dev, mac, proto, addr, vni,
+						  adding);
+}
+
 static void
 mlxsw_sp_fdb_call_notifiers(enum switchdev_notifier_type type,
 			    const char *mac, u16 vid,
@@ -2442,6 +2507,122 @@ static void mlxsw_sp_fdb_notify_mac_lag_process(struct
mlxsw_sp *mlxsw_sp,
 	goto do_fdb_op;
 }
 
+static int
+__mlxsw_sp_fdb_notify_mac_uc_tunnel_process(struct mlxsw_sp *mlxsw_sp,
+					    const struct mlxsw_sp_fid *fid,
+					    bool adding,
+					    struct net_device **nve_dev,
+					    u16 *p_vid, __be32 *p_vni)
+{
+	struct mlxsw_sp_bridge_device *bridge_device;
+	struct net_device *br_dev, *dev;
+	int nve_ifindex;
+	int err;
+
+	err = mlxsw_sp_fid_nve_ifindex(fid, &nve_ifindex);
+	if (err)
+		return err;
+
+	err = mlxsw_sp_fid_vni(fid, p_vni);
+	if (err)
+		return err;
+
+	dev = __dev_get_by_index(&init_net, nve_ifindex);
+	if (!dev)
+		return -EINVAL;
+	*nve_dev = dev;
+
+	if (!netif_running(dev))
+		return -EINVAL;
+
+	if (adding && !br_port_flag_is_set(dev, BR_LEARNING))
+		return -EINVAL;
+
+	if (adding && netif_is_vxlan(dev)) {
+		struct vxlan_dev *vxlan = netdev_priv(dev);
+
+		if (!(vxlan->cfg.flags & VXLAN_F_LEARN))
+			return -EINVAL;
+	}
+
+	br_dev = netdev_master_upper_dev_get(dev);
+	if (!br_dev)
+		return -EINVAL;
+
+	bridge_device = mlxsw_sp_bridge_device_find(mlxsw_sp->bridge, br_dev);
+	if (!bridge_device)
+		return -EINVAL;
+
+	*p_vid = bridge_device->ops->fid_vid(bridge_device, fid);
+
+	return 0;
+}
+
+static void mlxsw_sp_fdb_notify_mac_uc_tunnel_process(struct mlxsw_sp
*mlxsw_sp,
+						      char *sfn_pl,
+						      int rec_index,
+						      bool adding)
+{
+	enum mlxsw_reg_sfn_uc_tunnel_protocol sfn_proto;
+	enum switchdev_notifier_type type;
+	struct net_device *nve_dev;
+	union mlxsw_sp_l3addr addr;
+	struct mlxsw_sp_fid *fid;
+	char mac[ETH_ALEN];
+	u16 fid_index, vid;
+	__be32 vni;
+	u32 uip;
+	int err;
+
+	mlxsw_reg_sfn_uc_tunnel_unpack(sfn_pl, rec_index, mac, &fid_index,
+				       &uip, &sfn_proto);
+
+	fid = mlxsw_sp_fid_lookup_by_index(mlxsw_sp, fid_index);
+	if (!fid)
+		goto err_fid_lookup;
+
+	err = mlxsw_sp_nve_learned_ip_resolve(mlxsw_sp, uip,
+					      (enum mlxsw_sp_l3proto) sfn_proto,
+					      &addr);
+	if (err)
+		goto err_ip_resolve;
+
+	err = __mlxsw_sp_fdb_notify_mac_uc_tunnel_process(mlxsw_sp, fid, adding,
+							  &nve_dev, &vid, &vni);
+	if (err)
+		goto err_fdb_process;
+
+	err = mlxsw_sp_port_fdb_tunnel_uc_op(mlxsw_sp, mac, fid_index,
+					     (enum mlxsw_sp_l3proto) sfn_proto,
+					     &addr, adding, true);
+	if (err)
+		goto err_fdb_op;
+
+	mlxsw_sp_fdb_nve_call_notifiers(nve_dev, mac,
+					(enum mlxsw_sp_l3proto) sfn_proto,
+					&addr, vni, adding);
+
+	type = adding ? SWITCHDEV_FDB_ADD_TO_BRIDGE :
+			SWITCHDEV_FDB_DEL_TO_BRIDGE;
+	mlxsw_sp_fdb_call_notifiers(type, mac, vid, nve_dev, adding);
+
+	mlxsw_sp_fid_put(fid);
+
+	return;
+
+err_fdb_op:
+err_fdb_process:
+err_ip_resolve:
+	mlxsw_sp_fid_put(fid);
+err_fid_lookup:
+	/* Remove an FDB entry in case we cannot process it. Otherwise the
+	 * device will keep sending the same notification over and over again.
+	 */
+	mlxsw_sp_port_fdb_tunnel_uc_op(mlxsw_sp, mac, fid_index,
+				       (enum mlxsw_sp_l3proto) sfn_proto, &addr,
+				       false, true);
+}
+
 static void mlxsw_sp_fdb_notify_rec_process(struct mlxsw_sp *mlxsw_sp,
 					    char *sfn_pl, int rec_index)
 {
@@ -2462,6 +2643,14 @@ static void mlxsw_sp_fdb_notify_rec_process(struct
mlxsw_sp *mlxsw_sp,
 		mlxsw_sp_fdb_notify_mac_lag_process(mlxsw_sp, sfn_pl,
 						    rec_index, false);
 		break;
+	case MLXSW_REG_SFN_REC_TYPE_LEARNED_UNICAST_TUNNEL:
+		mlxsw_sp_fdb_notify_mac_uc_tunnel_process(mlxsw_sp, sfn_pl,
+							  rec_index, true);
+		break;
+	case MLXSW_REG_SFN_REC_TYPE_AGED_OUT_UNICAST_TUNNEL:
+		mlxsw_sp_fdb_notify_mac_uc_tunnel_process(mlxsw_sp, sfn_pl,
+							  rec_index, false);
+		break;
 	}
 }
 
@@ -2516,20 +2705,6 @@ struct mlxsw_sp_switchdev_event_work {
 	unsigned long event;
 };
 
-static void
-mlxsw_sp_switchdev_vxlan_addr_convert(const union vxlan_addr *vxlan_addr,
-				      enum mlxsw_sp_l3proto *proto,
-				      union mlxsw_sp_l3addr *addr)
-{
-	if (vxlan_addr->sa.sa_family == AF_INET) {
-		addr->addr4 = vxlan_addr->sin.sin_addr.s_addr;
-		*proto = MLXSW_SP_L3_PROTO_IPV4;
-	} else {
-		addr->addr6 = vxlan_addr->sin6.sin6_addr;
-		*proto = MLXSW_SP_L3_PROTO_IPV6;
-	}
-}
-
 static void
 mlxsw_sp_switchdev_bridge_vxlan_fdb_event(struct mlxsw_sp *mlxsw_sp,
 					  struct mlxsw_sp_switchdev_event_work *
-- 
2.19.1

Ido Schimmel

2018-Nov-21 08:02 UTC

head link

[Bridge] [PATCH net-next 13/16] mlxsw: spectrum_switchdev: Allow deletion of learned FDB entries

Allow users to delete learned FDB entries from the bridge's FDB before
enabling VxLAN learning.

Signed-off-by: Ido Schimmel <idosch at mellanox.com>
Reviewed-by: Petr Machata <petrm at mellanox.com>
---
 drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
index d3d1c1d00d98..b32a5ee57fb9 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
@@ -2770,7 +2770,8 @@ mlxsw_sp_switchdev_bridge_nve_fdb_event(struct
mlxsw_sp_switchdev_event_work *
 	    switchdev_work->event != SWITCHDEV_FDB_DEL_TO_DEVICE)
 		return;
 
-	if (!switchdev_work->fdb_info.added_by_user)
+	if (switchdev_work->event == SWITCHDEV_FDB_ADD_TO_DEVICE &&
+	    !switchdev_work->fdb_info.added_by_user)
 		return;
 
 	if (!netif_running(dev))
-- 
2.19.1

Ido Schimmel

2018-Nov-21 08:02 UTC

head link

[Bridge] [PATCH net-next 14/16] mlxsw: spectrum_nve: Allow VxLAN learning

Up until now the driver returned an error when learning was enabled on a
VxLAN device enslaved to an offloaded bridge.

Previous patches added VxLAN learning support, so remove the check.

Signed-off-by: Ido Schimmel <idosch at mellanox.com>
Reviewed-by: Petr Machata <petrm at mellanox.com>
---
 drivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c | 8 ++------
 1 file changed, 2 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c
b/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c
index d21c7be5b1c9..4e9cc00a88fd 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c
@@ -17,7 +17,8 @@
 #define MLXSW_SP_NVE_VXLAN_PARSING_DEPTH 128
 #define MLXSW_SP_NVE_DEFAULT_PARSING_DEPTH 96
 
-#define MLXSW_SP_NVE_VXLAN_SUPPORTED_FLAGS	VXLAN_F_UDP_ZERO_CSUM_TX
+#define MLXSW_SP_NVE_VXLAN_SUPPORTED_FLAGS	(VXLAN_F_UDP_ZERO_CSUM_TX | \
+						 VXLAN_F_LEARN)
 
 static bool mlxsw_sp1_nve_vxlan_can_offload(const struct mlxsw_sp_nve *nve,
 					    const struct net_device *dev,
@@ -61,11 +62,6 @@ static bool mlxsw_sp1_nve_vxlan_can_offload(const struct
mlxsw_sp_nve *nve,
 		return false;
 	}
 
-	if (cfg->flags & VXLAN_F_LEARN) {
-		NL_SET_ERR_MSG_MOD(extack, "VxLAN: Learning is not supported");
-		return false;
-	}
-
 	if (!(cfg->flags & VXLAN_F_UDP_ZERO_CSUM_TX)) {
 		NL_SET_ERR_MSG_MOD(extack, "VxLAN: UDP checksum is not supported");
 		return false;
-- 
2.19.1

Ido Schimmel

2018-Nov-21 08:02 UTC

head link

[Bridge] [PATCH net-next 15/16] selftests: mlxsw: Consider VxLAN learning enabled as valid

The test currently expects that a configuration which includes a VxLAN
device with learning enabled to fail.

Previous patches enabled VxLAN learning in mlxsw, so change the test
accordingly.

Signed-off-by: Ido Schimmel <idosch at mellanox.com>
Reviewed-by: Petr Machata <petrm at mellanox.com>
---
 tools/testing/selftests/drivers/net/mlxsw/vxlan.sh | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/testing/selftests/drivers/net/mlxsw/vxlan.sh
b/tools/testing/selftests/drivers/net/mlxsw/vxlan.sh
index 52e78adfe081..ffa79b738f2a 100755
--- a/tools/testing/selftests/drivers/net/mlxsw/vxlan.sh
+++ b/tools/testing/selftests/drivers/net/mlxsw/vxlan.sh
@@ -167,7 +167,7 @@ sanitization_single_dev_learning_enabled_test()
 	ip link add name vxlan0 up type vxlan id 10 learning noudpcsum \
 		ttl 20 tos inherit local 198.51.100.1 dstport 4789
 
-	sanitization_single_dev_test_fail
+	sanitization_single_dev_test_pass
 
 	ip link del dev vxlan0
 	ip link del dev br0
-- 
2.19.1

Ido Schimmel

2018-Nov-21 08:02 UTC

head link

[Bridge] [PATCH net-next 16/16] selftests: forwarding: vxlan_bridge_1d: Add learning test

Add a test which checks that the VxLAN driver can learn FDB entries and
that these entries are correctly deleted and aged-out.

Signed-off-by: Ido Schimmel <idosch at mellanox.com>
Reviewed-by: Petr Machata <petrm at mellanox.com>
---
 .../net/forwarding/vxlan_bridge_1d.sh         | 108 ++++++++++++++++++
 1 file changed, 108 insertions(+)

diff --git a/tools/testing/selftests/net/forwarding/vxlan_bridge_1d.sh
b/tools/testing/selftests/net/forwarding/vxlan_bridge_1d.sh
index 5cc6ac74eb74..56cef3b1c194 100755
--- a/tools/testing/selftests/net/forwarding/vxlan_bridge_1d.sh
+++ b/tools/testing/selftests/net/forwarding/vxlan_bridge_1d.sh
@@ -76,6 +76,7 @@ export VXPORT
 	ping_ipv4
 	test_flood
 	test_unicast
+	test_learning
     "}
 
 NUM_NETIFS=6
@@ -663,6 +664,113 @@ test_ecn_decap()
 	test_ecn_decap_error
 }
 
+test_learning()
+{
+	local mac=de:ad:be:ef:13:37
+	local dst=192.0.2.100
+
+	# Enable learning on the VxLAN device and set ageing time to 10 seconds
+	ip link set dev br1 type bridge ageing_time 1000
+	ip link set dev vx1 type vxlan ageing 10
+	ip link set dev vx1 type vxlan learning
+	reapply_config
+
+	# Check that flooding works
+	RET=0
+
+	vxlan_flood_test $mac $dst 10 10 10
+
+	log_test "VXLAN: flood before learning"
+
+	# Send a packet with source mac set to $mac from host w2 and check that
+	# a corresponding entry is created in VxLAN device vx1
+	RET=0
+
+	in_ns ns1 $MZ w2 -c 1 -p 64 -a $mac -b ff:ff:ff:ff:ff:ff -B $dst \
+		-t icmp -q
+	sleep 1
+
+	bridge fdb show brport vx1 | grep $mac | grep -q self
+	check_err $?
+	bridge fdb show brport vx1 | grep $mac | grep -q -v self
+	check_err $?
+
+	log_test "VXLAN: show learned FDB entry"
+
+	# Repeat first test and check that packets only reach host w2 in ns1
+	RET=0
+
+	vxlan_flood_test $mac $dst 0 10 0
+
+	log_test "VXLAN: learned FDB entry"
+
+	# Delete the learned FDB entry from the VxLAN and bridge devices and
+	# check that packets are flooded
+	RET=0
+
+	bridge fdb del dev vx1 $mac master self
+	sleep 1
+
+	vxlan_flood_test $mac $dst 10 10 10
+
+	log_test "VXLAN: deletion of learned FDB entry"
+
+	# Re-learn the first FDB entry and check that it is correctly aged-out
+	RET=0
+
+	in_ns ns1 $MZ w2 -c 1 -p 64 -a $mac -b ff:ff:ff:ff:ff:ff -B $dst \
+		-t icmp -q
+	sleep 1
+
+	bridge fdb show brport vx1 | grep $mac | grep -q self
+	check_err $?
+	bridge fdb show brport vx1 | grep $mac | grep -q -v self
+	check_err $?
+
+	vxlan_flood_test $mac $dst 0 10 0
+
+	sleep 20
+
+	bridge fdb show brport vx1 | grep $mac | grep -q self
+	check_fail $?
+	bridge fdb show brport vx1 | grep $mac | grep -q -v self
+	check_fail $?
+
+	vxlan_flood_test $mac $dst 10 10 10
+
+	log_test "VXLAN: Ageing of learned FDB entry"
+
+	# Toggle learning on the bridge port and check that the bridge's FDB
+	# is populated only when it should
+	RET=0
+
+	ip link set dev vx1 type bridge_slave learning off
+
+	in_ns ns1 $MZ w2 -c 1 -p 64 -a $mac -b ff:ff:ff:ff:ff:ff -B $dst \
+		-t icmp -q
+	sleep 1
+
+	bridge fdb show brport vx1 | grep $mac | grep -q -v self
+	check_fail $?
+
+	ip link set dev vx1 type bridge_slave learning on
+
+	in_ns ns1 $MZ w2 -c 1 -p 64 -a $mac -b ff:ff:ff:ff:ff:ff -B $dst \
+		-t icmp -q
+	sleep 1
+
+	bridge fdb show brport vx1 | grep $mac | grep -q -v self
+	check_err $?
+
+	log_test "VXLAN: learning toggling on bridge port"
+
+	# Restore previous settings
+	ip link set dev vx1 type vxlan nolearning
+	ip link set dev vx1 type vxlan ageing 300
+	ip link set dev br1 type bridge ageing_time 30000
+	reapply_config
+}
+
 test_all()
 {
 	echo "Running tests with UDP port $VXPORT"
-- 
2.19.1

David Miller

2018-Nov-22 01:10 UTC

head link

[Bridge] [PATCH net-next 00/16] mlxsw: Add VxLAN learning support

From: Ido Schimmel <idosch at mellanox.com>
Date: Wed, 21 Nov 2018 08:02:32 +0000
> This patchset adds VxLAN learning support in the mlxsw driver.
Looks great, series applied, thanks Ido.

Linux Ethernet Bridging - Nov 2018 - [Bridge] [PATCH net-next 00/16] mlxsw: Add VxLAN learning support

[Bridge] [PATCH net-next 00/16] mlxsw: Add VxLAN learning support

[Bridge] [PATCH net-next 01/16] vxlan: __vxlan_fdb_delete(): Drop unused argument vid

[Bridge] [PATCH net-next 02/16] vxlan: vxlan_fdb_notify(): Make switchdev notification configurable

[Bridge] [PATCH net-next 03/16] vxlan: Mark user-added FDB entries

[Bridge] [PATCH net-next 04/16] vxlan: Don't override user-added entries with ext-learned ones

[Bridge] [PATCH net-next 05/16] vxlan: Add hardware FDB learning

[Bridge] [PATCH net-next 06/16] vxlan: Allow changing ageing time

[Bridge] [PATCH net-next 07/16] bridge: Allow querying bridge port flags

[Bridge] [PATCH net-next 08/16] mlxsw: reg: Add definition of unicast tunnel record for SFN register

[Bridge] [PATCH net-next 09/16] mlxsw: spectrum_fid: Store ifindex of NVE device in FID

[Bridge] [PATCH net-next 10/16] mlxsw: spectrum_fid: Allow FID lookup by its index

[Bridge] [PATCH net-next 11/16] mlxsw: spectrum_nve: Add API to resolve learned IP addresses

[Bridge] [PATCH net-next 12/16] mlxsw: spectrum_switchdev: Process learned VxLAN FDB entries

[Bridge] [PATCH net-next 13/16] mlxsw: spectrum_switchdev: Allow deletion of learned FDB entries

[Bridge] [PATCH net-next 14/16] mlxsw: spectrum_nve: Allow VxLAN learning

[Bridge] [PATCH net-next 15/16] selftests: mlxsw: Consider VxLAN learning enabled as valid

[Bridge] [PATCH net-next 16/16] selftests: forwarding: vxlan_bridge_1d: Add learning test

[Bridge] [PATCH net-next 00/16] mlxsw: Add VxLAN learning support