Allowing switchdev to prevent vlan deletion and error out in __vlan_del could cause multiple different issues - inconsistent state, memory leaks when flushing, NULL pointer dereference on bridge error when flushing. It doesn't make sense to allow it to stop __vlan_del, so log the error and continue with software vlan deletion. This is also consistent with 8021q behaviour.
Suggested-by: Ido Schimmel <[email protected]> Fixes: bf361ad38165 ("net: bridge: check __vlan_vid_del for error") Fixes: 5454f5c28eca ("net: bridge: vlan: check for errors from __vlan_del in __vlan_flush") Fixes: 2594e9064a57 ("bridge: vlan: add per-vlan struct and move to rhashtables") Fixes: 9c86ce2c1ae3 ("net: bridge: Notify about bridge VLANs") Signed-off-by: Nikolay Aleksandrov <[email protected]> --- v2: this is a new patch, don't fail vlan deletion as suggested by Ido v1: https://lore.kernel.org/netdev/[email protected]/ to Sashiko (from local clone check): [Severity: Critical] Could this introduce a regression where net/bridge/br_vlan.c:__vlan_del() leaves stale VLAN membership in hardware after tearing down the software VLAN? For the port branch, __vlan_vid_del() has already converted -EOPNOTSUPP to success. An error reaching this warning is therefore a substantive offload deletion failure. Nik: No, a driver should not block VLAN deletion and this behaviour is in line with 8021q. Also blocking it could leave incosistent state. It is best to print a warning and allow software vlan deletion to proceed. net/bridge/br_vlan.c | 31 ++++++++++++++----------------- 1 file changed, 14 insertions(+), 17 deletions(-) diff --git a/net/bridge/br_vlan.c b/net/bridge/br_vlan.c index 1e0e436629ec..92b3cb621a26 100644 --- a/net/bridge/br_vlan.c +++ b/net/bridge/br_vlan.c @@ -387,12 +387,12 @@ static int __vlan_add(struct net_bridge_vlan *v, u16 flags, goto out; } -static int __vlan_del(struct net_bridge_vlan *v) +static void __vlan_del(struct net_bridge_vlan *v) { struct net_bridge_vlan *masterv = v; struct net_bridge_vlan_group *vg; struct net_bridge_port *p = NULL; - int err = 0; + int err; if (br_vlan_is_master(v)) { vg = br_vlan_group(v->br); @@ -406,12 +406,16 @@ static int __vlan_del(struct net_bridge_vlan *v) if (p) { err = __vlan_vid_del(p->dev, p->br, v); if (err) - goto out; + br_warn(p->br, + "port %u(%s) failed to delete vlan %u from switchdev: %pe\n", + (unsigned int)p->port_no, p->dev->name, + v->vid, ERR_PTR(err)); } else { err = br_switchdev_port_vlan_del(v->br->dev, v->vid); if (err && err != -EOPNOTSUPP) - goto out; - err = 0; + br_warn(v->br, + "failed to delete bridge vlan %u from switchdev: %pe\n", + v->vid, ERR_PTR(err)); } if (br_vlan_should_use(v)) { @@ -431,8 +435,6 @@ static int __vlan_del(struct net_bridge_vlan *v) } br_vlan_put_master(masterv); -out: - return err; } static void __vlan_group_free(struct net_bridge_vlan_group *vg) @@ -449,7 +451,6 @@ static void __vlan_flush(const struct net_bridge *br, { struct net_bridge_vlan *vlan, *tmp; u16 v_start = 0, v_end = 0; - int err; __vlan_delete_pvid(vg, vg->pvid); list_for_each_entry_safe(vlan, tmp, &vg->vlan_list, vlist) { @@ -463,13 +464,7 @@ static void __vlan_flush(const struct net_bridge *br, } v_end = vlan->vid; - err = __vlan_del(vlan); - if (err) { - br_err(br, - "port %u(%s) failed to delete vlan %d: %pe\n", - (unsigned int) p->port_no, p->dev->name, - vlan->vid, ERR_PTR(err)); - } + __vlan_del(vlan); } /* notify about the last/whole vlan range */ @@ -837,8 +832,9 @@ int br_vlan_delete(struct net_bridge *br, u16 vid) br_fdb_delete_by_port(br, NULL, vid, 0); vlan_tunnel_info_del(vg, v); + __vlan_del(v); - return __vlan_del(v); + return 0; } void br_vlan_flush(struct net_bridge *br) @@ -1368,8 +1364,9 @@ int nbp_vlan_delete(struct net_bridge_port *port, u16 vid) return -ENOENT; br_fdb_find_delete_local(port->br, port, port->dev->dev_addr, vid); br_fdb_delete_by_port(port->br, port, vid, 0); + __vlan_del(v); - return __vlan_del(v); + return 0; } void nbp_vlan_flush(struct net_bridge_port *port) -- 2.47.3
