Commit 2221d954 authored by Vlad Buslov's avatar Vlad Buslov Committed by Saeed Mahameed
Browse files

net/mlx5e: Refactor neigh update infrastructure



Following patches in series implements route update which can cause encap
entries to migrate between routing devices. Consecutively, their parent
nhe's need to be also transferable between devices instead of having neigh
device as a part of their immutable key. Move neigh device from struct
mlx5_neigh to struct mlx5e_neigh_hash_entry and check that nhe and neigh
devices are the same in workqueue neigh update handler.

Save neigh net_device that can change dynamically in dedicated nhe->dev
field. With FIB event handler that is implemented in following patches
changing nhe->dev, NETEVENT_DELAY_PROBE_TIME_UPDATE handler can
concurrently access the nhe entry when traversing neigh list under rcu read
lock. Processing stale values in that handler doesn't change the handler
logic, so just wrap all accesses to the dev pointer in {WRITE|READ}_ONCE()
helpers.

Signed-off-by: default avatarVlad Buslov <vladbu@nvidia.com>
Signed-off-by: default avatarDmytro Linkin <dlinkin@nvidia.com>
Reviewed-by: default avatarRoi Dayan <roid@nvidia.com>
Signed-off-by: default avatarSaeed Mahameed <saeedm@nvidia.com>
parent 777bb800
Loading
Loading
Loading
Loading
+2 −2
Original line number Diff line number Diff line
@@ -15,7 +15,7 @@ TRACE_EVENT(mlx5e_rep_neigh_update,
	    TP_PROTO(const struct mlx5e_neigh_hash_entry *nhe, const u8 *ha,
		     bool neigh_connected),
	    TP_ARGS(nhe, ha, neigh_connected),
	    TP_STRUCT__entry(__string(devname, nhe->m_neigh.dev->name)
	    TP_STRUCT__entry(__string(devname, nhe->neigh_dev->name)
			     __array(u8, ha, ETH_ALEN)
			     __array(u8, v4, 4)
			     __array(u8, v6, 16)
@@ -25,7 +25,7 @@ TRACE_EVENT(mlx5e_rep_neigh_update,
			struct in6_addr *pin6;
			__be32 *p32;

			__assign_str(devname, mn->dev->name);
			__assign_str(devname, nhe->neigh_dev->name);
			__entry->neigh_connected = neigh_connected;
			memcpy(__entry->ha, ha, ETH_ALEN);

+2 −2
Original line number Diff line number Diff line
@@ -77,7 +77,7 @@ TRACE_EVENT(mlx5e_stats_flower,
TRACE_EVENT(mlx5e_tc_update_neigh_used_value,
	    TP_PROTO(const struct mlx5e_neigh_hash_entry *nhe, bool neigh_used),
	    TP_ARGS(nhe, neigh_used),
	    TP_STRUCT__entry(__string(devname, nhe->m_neigh.dev->name)
	    TP_STRUCT__entry(__string(devname, nhe->neigh_dev->name)
			     __array(u8, v4, 4)
			     __array(u8, v6, 16)
			     __field(bool, neigh_used)
@@ -86,7 +86,7 @@ TRACE_EVENT(mlx5e_tc_update_neigh_used_value,
			struct in6_addr *pin6;
			__be32 *p32;

			__assign_str(devname, mn->dev->name);
			__assign_str(devname, nhe->neigh_dev->name);
			__entry->neigh_used = neigh_used;

			p32 = (__be32 *)__entry->v4;
+11 −5
Original line number Diff line number Diff line
@@ -129,10 +129,10 @@ static void mlx5e_rep_neigh_update(struct work_struct *work)
							     work);
	struct mlx5e_neigh_hash_entry *nhe = update_work->nhe;
	struct neighbour *n = update_work->n;
	bool neigh_connected, same_dev;
	struct mlx5e_encap_entry *e;
	unsigned char ha[ETH_ALEN];
	struct mlx5e_priv *priv;
	bool neigh_connected;
	u8 nud_state, dead;

	rtnl_lock();
@@ -146,12 +146,16 @@ static void mlx5e_rep_neigh_update(struct work_struct *work)
	memcpy(ha, n->ha, ETH_ALEN);
	nud_state = n->nud_state;
	dead = n->dead;
	same_dev = READ_ONCE(nhe->neigh_dev) == n->dev;
	read_unlock_bh(&n->lock);

	neigh_connected = (nud_state & NUD_VALID) && !dead;

	trace_mlx5e_rep_neigh_update(nhe, ha, neigh_connected);

	if (!same_dev)
		goto out;

	list_for_each_entry(e, &nhe->encap_list, encap_list) {
		if (!mlx5e_encap_take(e))
			continue;
@@ -160,6 +164,7 @@ static void mlx5e_rep_neigh_update(struct work_struct *work)
		mlx5e_rep_update_flows(priv, e, neigh_connected, ha);
		mlx5e_encap_put(priv, e);
	}
out:
	rtnl_unlock();
	mlx5e_release_neigh_update_work(update_work);
}
@@ -175,7 +180,6 @@ static struct neigh_update_work *mlx5e_alloc_neigh_update_work(struct mlx5e_priv
	if (WARN_ON(!update_work))
		return NULL;

	m_neigh.dev = n->dev;
	m_neigh.family = n->ops->family;
	memcpy(&m_neigh.dst_ip, n->primary_key, n->tbl->key_len);

@@ -246,7 +250,7 @@ static int mlx5e_rep_netevent_event(struct notifier_block *nb,
		rcu_read_lock();
		list_for_each_entry_rcu(nhe, &neigh_update->neigh_list,
					neigh_list) {
			if (p->dev == nhe->m_neigh.dev) {
			if (p->dev == READ_ONCE(nhe->neigh_dev)) {
				found = true;
				break;
			}
@@ -369,7 +373,8 @@ mlx5e_rep_neigh_entry_lookup(struct mlx5e_priv *priv,
}

int mlx5e_rep_neigh_entry_create(struct mlx5e_priv *priv,
				 struct mlx5e_encap_entry *e,
				 struct mlx5e_neigh *m_neigh,
				 struct net_device *neigh_dev,
				 struct mlx5e_neigh_hash_entry **nhe)
{
	int err;
@@ -379,10 +384,11 @@ int mlx5e_rep_neigh_entry_create(struct mlx5e_priv *priv,
		return -ENOMEM;

	(*nhe)->priv = priv;
	memcpy(&(*nhe)->m_neigh, &e->m_neigh, sizeof(e->m_neigh));
	memcpy(&(*nhe)->m_neigh, m_neigh, sizeof(*m_neigh));
	spin_lock_init(&(*nhe)->encap_list_lock);
	INIT_LIST_HEAD(&(*nhe)->encap_list);
	refcount_set(&(*nhe)->refcnt, 1);
	WRITE_ONCE((*nhe)->neigh_dev, neigh_dev);

	err = mlx5e_rep_neigh_entry_insert(priv, *nhe);
	if (err)
+2 −1
Original line number Diff line number Diff line
@@ -16,7 +16,8 @@ struct mlx5e_neigh_hash_entry *
mlx5e_rep_neigh_entry_lookup(struct mlx5e_priv *priv,
			     struct mlx5e_neigh *m_neigh);
int mlx5e_rep_neigh_entry_create(struct mlx5e_priv *priv,
				 struct mlx5e_encap_entry *e,
				 struct mlx5e_neigh *m_neigh,
				 struct net_device *neigh_dev,
				 struct mlx5e_neigh_hash_entry **nhe);
void mlx5e_rep_neigh_entry_release(struct mlx5e_neigh_hash_entry *nhe);

+5 −3
Original line number Diff line number Diff line
@@ -26,7 +26,9 @@ struct mlx5e_rep_indr_block_priv {
};

int mlx5e_rep_encap_entry_attach(struct mlx5e_priv *priv,
				 struct mlx5e_encap_entry *e)
				 struct mlx5e_encap_entry *e,
				 struct mlx5e_neigh *m_neigh,
				 struct net_device *neigh_dev)
{
	struct mlx5e_rep_priv *rpriv = priv->ppriv;
	struct mlx5_rep_uplink_priv *uplink_priv = &rpriv->uplink_priv;
@@ -39,9 +41,9 @@ int mlx5e_rep_encap_entry_attach(struct mlx5e_priv *priv,
		return err;

	mutex_lock(&rpriv->neigh_update.encap_lock);
	nhe = mlx5e_rep_neigh_entry_lookup(priv, &e->m_neigh);
	nhe = mlx5e_rep_neigh_entry_lookup(priv, m_neigh);
	if (!nhe) {
		err = mlx5e_rep_neigh_entry_create(priv, e, &nhe);
		err = mlx5e_rep_neigh_entry_create(priv, m_neigh, neigh_dev, &nhe);
		if (err) {
			mutex_unlock(&rpriv->neigh_update.encap_lock);
			mlx5_tun_entropy_refcount_dec(tun_entropy,
Loading