]> git.proxmox.com Git - mirror_ubuntu-focal-kernel.git/commitdiff
mlxsw: spectrum: Drop refcounting of IPIP entries
authorPetr Machata <petrm@mellanox.com>
Mon, 16 Oct 2017 14:26:39 +0000 (16:26 +0200)
committerDavid S. Miller <davem@davemloft.net>
Mon, 16 Oct 2017 20:30:33 +0000 (21:30 +0100)
Formerly, IPIP entries were created lazily by next hops that referenced
an offloadable IP-in-IP netdevice. However now that they are created
eagerly as a reaction to events on such netdevices, the reference
counting is useless. Hence drop it.

The routes whose next hops reference an offloaded IP-in-IP netdevice
actually linger around a bit after their device is unregistered.
However, mlxsw_sp_ipip_entry_destroy() also destroys the backing
loopback, and mlxsw_sp_rif_destroy() transitively (via
mlxsw_sp_nexthop_rif_gone_sync()) calls mlxsw_sp_nexthop_ipip_fini(),
which unlinks the IPIP entry from a next hop. Thus no dangling pointers
are left behind for the brief window after netdevice is gone, but routes
not yet.

Signed-off-by: Petr Machata <petrm@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/net/ethernet/mellanox/mlxsw/spectrum_ipip.h
drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c

index 1c2db831d83b1c4f05a0d469e08a89e2ccf86c8f..6fb49129ce8783e07746512c28575b1cdc61401a 100644 (file)
@@ -47,7 +47,6 @@ struct mlxsw_sp_ipip_entry {
        enum mlxsw_sp_ipip_type ipipt;
        struct net_device *ol_dev; /* Overlay. */
        struct mlxsw_sp_rif_ipip_lb *ol_lb;
-       unsigned int ref_count; /* Number of next hops using the tunnel. */
        struct mlxsw_sp_fib_entry *decap_fib_entry;
        struct list_head ipip_list_node;
 };
index 082cf00eaadb8b41ffed0c659da6283b8102584c..3330120f2f8e29fdbdd391a18ce24699e51e5777 100644 (file)
@@ -1002,9 +1002,8 @@ err_ol_ipip_lb_create:
 }
 
 static void
-mlxsw_sp_ipip_entry_destroy(struct mlxsw_sp_ipip_entry *ipip_entry)
+mlxsw_sp_ipip_entry_dealloc(struct mlxsw_sp_ipip_entry *ipip_entry)
 {
-       WARN_ON(ipip_entry->ref_count > 0);
        mlxsw_sp_rif_destroy(&ipip_entry->ol_lb->common);
        kfree(ipip_entry);
 }
@@ -1200,9 +1199,9 @@ mlxsw_sp_ipip_entry_find_decap(struct mlxsw_sp *mlxsw_sp,
 }
 
 static struct mlxsw_sp_ipip_entry *
-mlxsw_sp_ipip_entry_get(struct mlxsw_sp *mlxsw_sp,
-                       enum mlxsw_sp_ipip_type ipipt,
-                       struct net_device *ol_dev)
+mlxsw_sp_ipip_entry_create(struct mlxsw_sp *mlxsw_sp,
+                          enum mlxsw_sp_ipip_type ipipt,
+                          struct net_device *ol_dev)
 {
        u32 ul_tb_id = mlxsw_sp_ipip_dev_ul_tb_id(ol_dev);
        struct mlxsw_sp_router *router = mlxsw_sp->router;
@@ -1210,15 +1209,12 @@ mlxsw_sp_ipip_entry_get(struct mlxsw_sp *mlxsw_sp,
        enum mlxsw_sp_l3proto ul_proto;
        union mlxsw_sp_l3addr saddr;
 
+       /* The configuration where several tunnels have the same local address
+        * in the same underlay table needs special treatment in the HW. That is
+        * currently not implemented in the driver.
+        */
        list_for_each_entry(ipip_entry, &mlxsw_sp->router->ipip_list,
                            ipip_list_node) {
-               if (ipip_entry->ol_dev == ol_dev)
-                       goto inc_ref_count;
-
-               /* The configuration where several tunnels have the same local
-                * address in the same underlay table needs special treatment in
-                * the HW. That is currently not implemented in the driver.
-                */
                ul_proto = router->ipip_ops_arr[ipip_entry->ipipt]->ul_proto;
                saddr = mlxsw_sp_ipip_netdev_saddr(ul_proto, ol_dev);
                if (mlxsw_sp_ipip_entry_saddr_matches(mlxsw_sp, ul_proto, saddr,
@@ -1233,19 +1229,15 @@ mlxsw_sp_ipip_entry_get(struct mlxsw_sp *mlxsw_sp,
        list_add_tail(&ipip_entry->ipip_list_node,
                      &mlxsw_sp->router->ipip_list);
 
-inc_ref_count:
-       ++ipip_entry->ref_count;
        return ipip_entry;
 }
 
 static void
-mlxsw_sp_ipip_entry_put(struct mlxsw_sp *mlxsw_sp,
-                       struct mlxsw_sp_ipip_entry *ipip_entry)
+mlxsw_sp_ipip_entry_destroy(struct mlxsw_sp *mlxsw_sp,
+                           struct mlxsw_sp_ipip_entry *ipip_entry)
 {
-       if (--ipip_entry->ref_count == 0) {
-               list_del(&ipip_entry->ipip_list_node);
-               mlxsw_sp_ipip_entry_destroy(ipip_entry);
-       }
+       list_del(&ipip_entry->ipip_list_node);
+       mlxsw_sp_ipip_entry_dealloc(ipip_entry);
 }
 
 static bool
@@ -1338,7 +1330,8 @@ static int mlxsw_sp_netdevice_ipip_reg_event(struct mlxsw_sp *mlxsw_sp,
                                                     MLXSW_SP_L3_PROTO_IPV4) ||
            router->ipip_ops_arr[ipipt]->can_offload(mlxsw_sp, ol_dev,
                                                     MLXSW_SP_L3_PROTO_IPV6)) {
-               ipip_entry = mlxsw_sp_ipip_entry_get(mlxsw_sp, ipipt, ol_dev);
+               ipip_entry = mlxsw_sp_ipip_entry_create(mlxsw_sp, ipipt,
+                                                       ol_dev);
                if (IS_ERR(ipip_entry))
                        return PTR_ERR(ipip_entry);
        }
@@ -1353,7 +1346,7 @@ static void mlxsw_sp_netdevice_ipip_unreg_event(struct mlxsw_sp *mlxsw_sp,
 
        ipip_entry = mlxsw_sp_ipip_entry_find_by_ol_dev(mlxsw_sp, ol_dev);
        if (ipip_entry)
-               mlxsw_sp_ipip_entry_put(mlxsw_sp, ipip_entry);
+               mlxsw_sp_ipip_entry_destroy(mlxsw_sp, ipip_entry);
 }
 
 static int mlxsw_sp_netdevice_ipip_up_event(struct mlxsw_sp *mlxsw_sp,
@@ -2939,16 +2932,15 @@ static void mlxsw_sp_nexthop_neigh_fini(struct mlxsw_sp *mlxsw_sp,
 }
 
 static int mlxsw_sp_nexthop_ipip_init(struct mlxsw_sp *mlxsw_sp,
-                                     enum mlxsw_sp_ipip_type ipipt,
                                      struct mlxsw_sp_nexthop *nh,
                                      struct net_device *ol_dev)
 {
        if (!nh->nh_grp->gateway || nh->ipip_entry)
                return 0;
 
-       nh->ipip_entry = mlxsw_sp_ipip_entry_get(mlxsw_sp, ipipt, ol_dev);
-       if (IS_ERR(nh->ipip_entry))
-               return PTR_ERR(nh->ipip_entry);
+       nh->ipip_entry = mlxsw_sp_ipip_entry_find_by_ol_dev(mlxsw_sp, ol_dev);
+       if (!nh->ipip_entry)
+               return -ENOENT;
 
        __mlxsw_sp_nexthop_neigh_update(nh, false);
        return 0;
@@ -2963,7 +2955,6 @@ static void mlxsw_sp_nexthop_ipip_fini(struct mlxsw_sp *mlxsw_sp,
                return;
 
        __mlxsw_sp_nexthop_neigh_update(nh, true);
-       mlxsw_sp_ipip_entry_put(mlxsw_sp, ipip_entry);
        nh->ipip_entry = NULL;
 }
 
@@ -3007,7 +2998,7 @@ static int mlxsw_sp_nexthop4_type_init(struct mlxsw_sp *mlxsw_sp,
            router->ipip_ops_arr[ipipt]->can_offload(mlxsw_sp, dev,
                                                     MLXSW_SP_L3_PROTO_IPV4)) {
                nh->type = MLXSW_SP_NEXTHOP_TYPE_IPIP;
-               err = mlxsw_sp_nexthop_ipip_init(mlxsw_sp, ipipt, nh, dev);
+               err = mlxsw_sp_nexthop_ipip_init(mlxsw_sp, nh, dev);
                if (err)
                        return err;
                mlxsw_sp_nexthop_rif_init(nh, &nh->ipip_entry->ol_lb->common);
@@ -4269,7 +4260,7 @@ static int mlxsw_sp_nexthop6_type_init(struct mlxsw_sp *mlxsw_sp,
            router->ipip_ops_arr[ipipt]->can_offload(mlxsw_sp, dev,
                                                     MLXSW_SP_L3_PROTO_IPV6)) {
                nh->type = MLXSW_SP_NEXTHOP_TYPE_IPIP;
-               err = mlxsw_sp_nexthop_ipip_init(mlxsw_sp, ipipt, nh, dev);
+               err = mlxsw_sp_nexthop_ipip_init(mlxsw_sp, nh, dev);
                if (err)
                        return err;
                mlxsw_sp_nexthop_rif_init(nh, &nh->ipip_entry->ol_lb->common);