linux/net/dsa/switch.c
Vladimir Oltean 3f9e79f31e net: dsa: avoid suspicious RCU usage for synced VLAN-aware MAC addresses
[ Upstream commit d06f925f13 ]

When using the felix driver (the only one which supports UC filtering
and MC filtering) as a DSA master for a random other DSA switch, one can
see the following stack trace when the downstream switch ports join a
VLAN-aware bridge:

=============================
WARNING: suspicious RCU usage
-----------------------------
net/8021q/vlan_core.c:238 suspicious rcu_dereference_protected() usage!

stack backtrace:
Workqueue: dsa_ordered dsa_slave_switchdev_event_work
Call trace:
 lockdep_rcu_suspicious+0x170/0x210
 vlan_for_each+0x8c/0x188
 dsa_slave_sync_uc+0x128/0x178
 __hw_addr_sync_dev+0x138/0x158
 dsa_slave_set_rx_mode+0x58/0x70
 __dev_set_rx_mode+0x88/0xa8
 dev_uc_add+0x74/0xa0
 dsa_port_bridge_host_fdb_add+0xec/0x180
 dsa_slave_switchdev_event_work+0x7c/0x1c8
 process_one_work+0x290/0x568

What it's saying is that vlan_for_each() expects rtnl_lock() context and
it's not getting it, when it's called from the DSA master's ndo_set_rx_mode().

The caller of that - dsa_slave_set_rx_mode() - is the slave DSA
interface's dsa_port_bridge_host_fdb_add() which comes from the deferred
dsa_slave_switchdev_event_work().

We went to great lengths to avoid the rtnl_lock() context in that call
path in commit 0faf890fc5 ("net: dsa: drop rtnl_lock from
dsa_slave_switchdev_event_work"), and calling rtnl_lock() is simply not
an option due to the possibility of deadlocking when calling
dsa_flush_workqueue() from the call paths that do hold rtnl_lock() -
basically all of them.

So, when the DSA master calls vlan_for_each() from its ndo_set_rx_mode(),
the state of the 8021q driver on this device is really not protected
from concurrent access by anything.

Looking at net/8021q/, I don't think that vlan_info->vid_list was
particularly designed with RCU traversal in mind, so introducing an RCU
read-side form of vlan_for_each() - vlan_for_each_rcu() - won't be so
easy, and it also wouldn't be exactly what we need anyway.

In general I believe that the solution isn't in net/8021q/ anyway;
vlan_for_each() is not cut out for this task. DSA doesn't need rtnl_lock()
to be held per se - since it's not a netdev state change that we're
blocking, but rather, just concurrent additions/removals to a VLAN list.
We don't even need sleepable context - the callback of vlan_for_each()
just schedules deferred work.

The proposed escape is to remove the dependency on vlan_for_each() and
to open-code a non-sleepable, rtnl-free alternative to that, based on
copies of the VLAN list modified from .ndo_vlan_rx_add_vid() and
.ndo_vlan_rx_kill_vid().

Fixes: 64fdc5f341 ("net: dsa: sync unicast and multicast addresses for VLAN filters too")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Link: https://lore.kernel.org/r/20230626154402.3154454-1-vladimir.oltean@nxp.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
2023-07-19 16:35:37 +02:00

1135 lines
26 KiB
C

// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Handling of a single switch chip, part of a switch fabric
*
* Copyright (c) 2017 Savoir-faire Linux Inc.
* Vivien Didelot <vivien.didelot@savoirfairelinux.com>
*/
#include <linux/if_bridge.h>
#include <linux/netdevice.h>
#include <linux/notifier.h>
#include <linux/if_vlan.h>
#include <net/switchdev.h>
#include "dsa.h"
#include "netlink.h"
#include "port.h"
#include "slave.h"
#include "switch.h"
#include "tag_8021q.h"
#include "trace.h"
static unsigned int dsa_switch_fastest_ageing_time(struct dsa_switch *ds,
unsigned int ageing_time)
{
struct dsa_port *dp;
dsa_switch_for_each_port(dp, ds)
if (dp->ageing_time && dp->ageing_time < ageing_time)
ageing_time = dp->ageing_time;
return ageing_time;
}
static int dsa_switch_ageing_time(struct dsa_switch *ds,
struct dsa_notifier_ageing_time_info *info)
{
unsigned int ageing_time = info->ageing_time;
if (ds->ageing_time_min && ageing_time < ds->ageing_time_min)
return -ERANGE;
if (ds->ageing_time_max && ageing_time > ds->ageing_time_max)
return -ERANGE;
/* Program the fastest ageing time in case of multiple bridges */
ageing_time = dsa_switch_fastest_ageing_time(ds, ageing_time);
if (ds->ops->set_ageing_time)
return ds->ops->set_ageing_time(ds, ageing_time);
return 0;
}
static bool dsa_port_mtu_match(struct dsa_port *dp,
struct dsa_notifier_mtu_info *info)
{
return dp == info->dp || dsa_port_is_dsa(dp) || dsa_port_is_cpu(dp);
}
static int dsa_switch_mtu(struct dsa_switch *ds,
struct dsa_notifier_mtu_info *info)
{
struct dsa_port *dp;
int ret;
if (!ds->ops->port_change_mtu)
return -EOPNOTSUPP;
dsa_switch_for_each_port(dp, ds) {
if (dsa_port_mtu_match(dp, info)) {
ret = ds->ops->port_change_mtu(ds, dp->index,
info->mtu);
if (ret)
return ret;
}
}
return 0;
}
static int dsa_switch_bridge_join(struct dsa_switch *ds,
struct dsa_notifier_bridge_info *info)
{
int err;
if (info->dp->ds == ds) {
if (!ds->ops->port_bridge_join)
return -EOPNOTSUPP;
err = ds->ops->port_bridge_join(ds, info->dp->index,
info->bridge,
&info->tx_fwd_offload,
info->extack);
if (err)
return err;
}
if (info->dp->ds != ds && ds->ops->crosschip_bridge_join) {
err = ds->ops->crosschip_bridge_join(ds,
info->dp->ds->dst->index,
info->dp->ds->index,
info->dp->index,
info->bridge,
info->extack);
if (err)
return err;
}
return 0;
}
static int dsa_switch_bridge_leave(struct dsa_switch *ds,
struct dsa_notifier_bridge_info *info)
{
if (info->dp->ds == ds && ds->ops->port_bridge_leave)
ds->ops->port_bridge_leave(ds, info->dp->index, info->bridge);
if (info->dp->ds != ds && ds->ops->crosschip_bridge_leave)
ds->ops->crosschip_bridge_leave(ds, info->dp->ds->dst->index,
info->dp->ds->index,
info->dp->index,
info->bridge);
return 0;
}
/* Matches for all upstream-facing ports (the CPU port and all upstream-facing
* DSA links) that sit between the targeted port on which the notifier was
* emitted and its dedicated CPU port.
*/
static bool dsa_port_host_address_match(struct dsa_port *dp,
const struct dsa_port *targeted_dp)
{
struct dsa_port *cpu_dp = targeted_dp->cpu_dp;
if (dsa_switch_is_upstream_of(dp->ds, targeted_dp->ds))
return dp->index == dsa_towards_port(dp->ds, cpu_dp->ds->index,
cpu_dp->index);
return false;
}
static struct dsa_mac_addr *dsa_mac_addr_find(struct list_head *addr_list,
const unsigned char *addr, u16 vid,
struct dsa_db db)
{
struct dsa_mac_addr *a;
list_for_each_entry(a, addr_list, list)
if (ether_addr_equal(a->addr, addr) && a->vid == vid &&
dsa_db_equal(&a->db, &db))
return a;
return NULL;
}
static int dsa_port_do_mdb_add(struct dsa_port *dp,
const struct switchdev_obj_port_mdb *mdb,
struct dsa_db db)
{
struct dsa_switch *ds = dp->ds;
struct dsa_mac_addr *a;
int port = dp->index;
int err = 0;
/* No need to bother with refcounting for user ports */
if (!(dsa_port_is_cpu(dp) || dsa_port_is_dsa(dp))) {
err = ds->ops->port_mdb_add(ds, port, mdb, db);
trace_dsa_mdb_add_hw(dp, mdb->addr, mdb->vid, &db, err);
return err;
}
mutex_lock(&dp->addr_lists_lock);
a = dsa_mac_addr_find(&dp->mdbs, mdb->addr, mdb->vid, db);
if (a) {
refcount_inc(&a->refcount);
trace_dsa_mdb_add_bump(dp, mdb->addr, mdb->vid, &db,
&a->refcount);
goto out;
}
a = kzalloc(sizeof(*a), GFP_KERNEL);
if (!a) {
err = -ENOMEM;
goto out;
}
err = ds->ops->port_mdb_add(ds, port, mdb, db);
trace_dsa_mdb_add_hw(dp, mdb->addr, mdb->vid, &db, err);
if (err) {
kfree(a);
goto out;
}
ether_addr_copy(a->addr, mdb->addr);
a->vid = mdb->vid;
a->db = db;
refcount_set(&a->refcount, 1);
list_add_tail(&a->list, &dp->mdbs);
out:
mutex_unlock(&dp->addr_lists_lock);
return err;
}
static int dsa_port_do_mdb_del(struct dsa_port *dp,
const struct switchdev_obj_port_mdb *mdb,
struct dsa_db db)
{
struct dsa_switch *ds = dp->ds;
struct dsa_mac_addr *a;
int port = dp->index;
int err = 0;
/* No need to bother with refcounting for user ports */
if (!(dsa_port_is_cpu(dp) || dsa_port_is_dsa(dp))) {
err = ds->ops->port_mdb_del(ds, port, mdb, db);
trace_dsa_mdb_del_hw(dp, mdb->addr, mdb->vid, &db, err);
return err;
}
mutex_lock(&dp->addr_lists_lock);
a = dsa_mac_addr_find(&dp->mdbs, mdb->addr, mdb->vid, db);
if (!a) {
trace_dsa_mdb_del_not_found(dp, mdb->addr, mdb->vid, &db);
err = -ENOENT;
goto out;
}
if (!refcount_dec_and_test(&a->refcount)) {
trace_dsa_mdb_del_drop(dp, mdb->addr, mdb->vid, &db,
&a->refcount);
goto out;
}
err = ds->ops->port_mdb_del(ds, port, mdb, db);
trace_dsa_mdb_del_hw(dp, mdb->addr, mdb->vid, &db, err);
if (err) {
refcount_set(&a->refcount, 1);
goto out;
}
list_del(&a->list);
kfree(a);
out:
mutex_unlock(&dp->addr_lists_lock);
return err;
}
static int dsa_port_do_fdb_add(struct dsa_port *dp, const unsigned char *addr,
u16 vid, struct dsa_db db)
{
struct dsa_switch *ds = dp->ds;
struct dsa_mac_addr *a;
int port = dp->index;
int err = 0;
/* No need to bother with refcounting for user ports */
if (!(dsa_port_is_cpu(dp) || dsa_port_is_dsa(dp))) {
err = ds->ops->port_fdb_add(ds, port, addr, vid, db);
trace_dsa_fdb_add_hw(dp, addr, vid, &db, err);
return err;
}
mutex_lock(&dp->addr_lists_lock);
a = dsa_mac_addr_find(&dp->fdbs, addr, vid, db);
if (a) {
refcount_inc(&a->refcount);
trace_dsa_fdb_add_bump(dp, addr, vid, &db, &a->refcount);
goto out;
}
a = kzalloc(sizeof(*a), GFP_KERNEL);
if (!a) {
err = -ENOMEM;
goto out;
}
err = ds->ops->port_fdb_add(ds, port, addr, vid, db);
trace_dsa_fdb_add_hw(dp, addr, vid, &db, err);
if (err) {
kfree(a);
goto out;
}
ether_addr_copy(a->addr, addr);
a->vid = vid;
a->db = db;
refcount_set(&a->refcount, 1);
list_add_tail(&a->list, &dp->fdbs);
out:
mutex_unlock(&dp->addr_lists_lock);
return err;
}
static int dsa_port_do_fdb_del(struct dsa_port *dp, const unsigned char *addr,
u16 vid, struct dsa_db db)
{
struct dsa_switch *ds = dp->ds;
struct dsa_mac_addr *a;
int port = dp->index;
int err = 0;
/* No need to bother with refcounting for user ports */
if (!(dsa_port_is_cpu(dp) || dsa_port_is_dsa(dp))) {
err = ds->ops->port_fdb_del(ds, port, addr, vid, db);
trace_dsa_fdb_del_hw(dp, addr, vid, &db, err);
return err;
}
mutex_lock(&dp->addr_lists_lock);
a = dsa_mac_addr_find(&dp->fdbs, addr, vid, db);
if (!a) {
trace_dsa_fdb_del_not_found(dp, addr, vid, &db);
err = -ENOENT;
goto out;
}
if (!refcount_dec_and_test(&a->refcount)) {
trace_dsa_fdb_del_drop(dp, addr, vid, &db, &a->refcount);
goto out;
}
err = ds->ops->port_fdb_del(ds, port, addr, vid, db);
trace_dsa_fdb_del_hw(dp, addr, vid, &db, err);
if (err) {
refcount_set(&a->refcount, 1);
goto out;
}
list_del(&a->list);
kfree(a);
out:
mutex_unlock(&dp->addr_lists_lock);
return err;
}
static int dsa_switch_do_lag_fdb_add(struct dsa_switch *ds, struct dsa_lag *lag,
const unsigned char *addr, u16 vid,
struct dsa_db db)
{
struct dsa_mac_addr *a;
int err = 0;
mutex_lock(&lag->fdb_lock);
a = dsa_mac_addr_find(&lag->fdbs, addr, vid, db);
if (a) {
refcount_inc(&a->refcount);
trace_dsa_lag_fdb_add_bump(lag->dev, addr, vid, &db,
&a->refcount);
goto out;
}
a = kzalloc(sizeof(*a), GFP_KERNEL);
if (!a) {
err = -ENOMEM;
goto out;
}
err = ds->ops->lag_fdb_add(ds, *lag, addr, vid, db);
trace_dsa_lag_fdb_add_hw(lag->dev, addr, vid, &db, err);
if (err) {
kfree(a);
goto out;
}
ether_addr_copy(a->addr, addr);
a->vid = vid;
a->db = db;
refcount_set(&a->refcount, 1);
list_add_tail(&a->list, &lag->fdbs);
out:
mutex_unlock(&lag->fdb_lock);
return err;
}
static int dsa_switch_do_lag_fdb_del(struct dsa_switch *ds, struct dsa_lag *lag,
const unsigned char *addr, u16 vid,
struct dsa_db db)
{
struct dsa_mac_addr *a;
int err = 0;
mutex_lock(&lag->fdb_lock);
a = dsa_mac_addr_find(&lag->fdbs, addr, vid, db);
if (!a) {
trace_dsa_lag_fdb_del_not_found(lag->dev, addr, vid, &db);
err = -ENOENT;
goto out;
}
if (!refcount_dec_and_test(&a->refcount)) {
trace_dsa_lag_fdb_del_drop(lag->dev, addr, vid, &db,
&a->refcount);
goto out;
}
err = ds->ops->lag_fdb_del(ds, *lag, addr, vid, db);
trace_dsa_lag_fdb_del_hw(lag->dev, addr, vid, &db, err);
if (err) {
refcount_set(&a->refcount, 1);
goto out;
}
list_del(&a->list);
kfree(a);
out:
mutex_unlock(&lag->fdb_lock);
return err;
}
static int dsa_switch_host_fdb_add(struct dsa_switch *ds,
struct dsa_notifier_fdb_info *info)
{
struct dsa_port *dp;
int err = 0;
if (!ds->ops->port_fdb_add)
return -EOPNOTSUPP;
dsa_switch_for_each_port(dp, ds) {
if (dsa_port_host_address_match(dp, info->dp)) {
if (dsa_port_is_cpu(dp) && info->dp->cpu_port_in_lag) {
err = dsa_switch_do_lag_fdb_add(ds, dp->lag,
info->addr,
info->vid,
info->db);
} else {
err = dsa_port_do_fdb_add(dp, info->addr,
info->vid, info->db);
}
if (err)
break;
}
}
return err;
}
static int dsa_switch_host_fdb_del(struct dsa_switch *ds,
struct dsa_notifier_fdb_info *info)
{
struct dsa_port *dp;
int err = 0;
if (!ds->ops->port_fdb_del)
return -EOPNOTSUPP;
dsa_switch_for_each_port(dp, ds) {
if (dsa_port_host_address_match(dp, info->dp)) {
if (dsa_port_is_cpu(dp) && info->dp->cpu_port_in_lag) {
err = dsa_switch_do_lag_fdb_del(ds, dp->lag,
info->addr,
info->vid,
info->db);
} else {
err = dsa_port_do_fdb_del(dp, info->addr,
info->vid, info->db);
}
if (err)
break;
}
}
return err;
}
static int dsa_switch_fdb_add(struct dsa_switch *ds,
struct dsa_notifier_fdb_info *info)
{
int port = dsa_towards_port(ds, info->dp->ds->index, info->dp->index);
struct dsa_port *dp = dsa_to_port(ds, port);
if (!ds->ops->port_fdb_add)
return -EOPNOTSUPP;
return dsa_port_do_fdb_add(dp, info->addr, info->vid, info->db);
}
static int dsa_switch_fdb_del(struct dsa_switch *ds,
struct dsa_notifier_fdb_info *info)
{
int port = dsa_towards_port(ds, info->dp->ds->index, info->dp->index);
struct dsa_port *dp = dsa_to_port(ds, port);
if (!ds->ops->port_fdb_del)
return -EOPNOTSUPP;
return dsa_port_do_fdb_del(dp, info->addr, info->vid, info->db);
}
static int dsa_switch_lag_fdb_add(struct dsa_switch *ds,
struct dsa_notifier_lag_fdb_info *info)
{
struct dsa_port *dp;
if (!ds->ops->lag_fdb_add)
return -EOPNOTSUPP;
/* Notify switch only if it has a port in this LAG */
dsa_switch_for_each_port(dp, ds)
if (dsa_port_offloads_lag(dp, info->lag))
return dsa_switch_do_lag_fdb_add(ds, info->lag,
info->addr, info->vid,
info->db);
return 0;
}
static int dsa_switch_lag_fdb_del(struct dsa_switch *ds,
struct dsa_notifier_lag_fdb_info *info)
{
struct dsa_port *dp;
if (!ds->ops->lag_fdb_del)
return -EOPNOTSUPP;
/* Notify switch only if it has a port in this LAG */
dsa_switch_for_each_port(dp, ds)
if (dsa_port_offloads_lag(dp, info->lag))
return dsa_switch_do_lag_fdb_del(ds, info->lag,
info->addr, info->vid,
info->db);
return 0;
}
static int dsa_switch_lag_change(struct dsa_switch *ds,
struct dsa_notifier_lag_info *info)
{
if (info->dp->ds == ds && ds->ops->port_lag_change)
return ds->ops->port_lag_change(ds, info->dp->index);
if (info->dp->ds != ds && ds->ops->crosschip_lag_change)
return ds->ops->crosschip_lag_change(ds, info->dp->ds->index,
info->dp->index);
return 0;
}
static int dsa_switch_lag_join(struct dsa_switch *ds,
struct dsa_notifier_lag_info *info)
{
if (info->dp->ds == ds && ds->ops->port_lag_join)
return ds->ops->port_lag_join(ds, info->dp->index, info->lag,
info->info, info->extack);
if (info->dp->ds != ds && ds->ops->crosschip_lag_join)
return ds->ops->crosschip_lag_join(ds, info->dp->ds->index,
info->dp->index, info->lag,
info->info, info->extack);
return -EOPNOTSUPP;
}
static int dsa_switch_lag_leave(struct dsa_switch *ds,
struct dsa_notifier_lag_info *info)
{
if (info->dp->ds == ds && ds->ops->port_lag_leave)
return ds->ops->port_lag_leave(ds, info->dp->index, info->lag);
if (info->dp->ds != ds && ds->ops->crosschip_lag_leave)
return ds->ops->crosschip_lag_leave(ds, info->dp->ds->index,
info->dp->index, info->lag);
return -EOPNOTSUPP;
}
static int dsa_switch_mdb_add(struct dsa_switch *ds,
struct dsa_notifier_mdb_info *info)
{
int port = dsa_towards_port(ds, info->dp->ds->index, info->dp->index);
struct dsa_port *dp = dsa_to_port(ds, port);
if (!ds->ops->port_mdb_add)
return -EOPNOTSUPP;
return dsa_port_do_mdb_add(dp, info->mdb, info->db);
}
static int dsa_switch_mdb_del(struct dsa_switch *ds,
struct dsa_notifier_mdb_info *info)
{
int port = dsa_towards_port(ds, info->dp->ds->index, info->dp->index);
struct dsa_port *dp = dsa_to_port(ds, port);
if (!ds->ops->port_mdb_del)
return -EOPNOTSUPP;
return dsa_port_do_mdb_del(dp, info->mdb, info->db);
}
static int dsa_switch_host_mdb_add(struct dsa_switch *ds,
struct dsa_notifier_mdb_info *info)
{
struct dsa_port *dp;
int err = 0;
if (!ds->ops->port_mdb_add)
return -EOPNOTSUPP;
dsa_switch_for_each_port(dp, ds) {
if (dsa_port_host_address_match(dp, info->dp)) {
err = dsa_port_do_mdb_add(dp, info->mdb, info->db);
if (err)
break;
}
}
return err;
}
static int dsa_switch_host_mdb_del(struct dsa_switch *ds,
struct dsa_notifier_mdb_info *info)
{
struct dsa_port *dp;
int err = 0;
if (!ds->ops->port_mdb_del)
return -EOPNOTSUPP;
dsa_switch_for_each_port(dp, ds) {
if (dsa_port_host_address_match(dp, info->dp)) {
err = dsa_port_do_mdb_del(dp, info->mdb, info->db);
if (err)
break;
}
}
return err;
}
/* Port VLANs match on the targeted port and on all DSA ports */
static bool dsa_port_vlan_match(struct dsa_port *dp,
struct dsa_notifier_vlan_info *info)
{
return dsa_port_is_dsa(dp) || dp == info->dp;
}
/* Host VLANs match on the targeted port's CPU port, and on all DSA ports
* (upstream and downstream) of that switch and its upstream switches.
*/
static bool dsa_port_host_vlan_match(struct dsa_port *dp,
const struct dsa_port *targeted_dp)
{
struct dsa_port *cpu_dp = targeted_dp->cpu_dp;
if (dsa_switch_is_upstream_of(dp->ds, targeted_dp->ds))
return dsa_port_is_dsa(dp) || dp == cpu_dp;
return false;
}
struct dsa_vlan *dsa_vlan_find(struct list_head *vlan_list,
const struct switchdev_obj_port_vlan *vlan)
{
struct dsa_vlan *v;
list_for_each_entry(v, vlan_list, list)
if (v->vid == vlan->vid)
return v;
return NULL;
}
static int dsa_port_do_vlan_add(struct dsa_port *dp,
const struct switchdev_obj_port_vlan *vlan,
struct netlink_ext_ack *extack)
{
struct dsa_switch *ds = dp->ds;
int port = dp->index;
struct dsa_vlan *v;
int err = 0;
/* No need to bother with refcounting for user ports. */
if (!(dsa_port_is_cpu(dp) || dsa_port_is_dsa(dp))) {
err = ds->ops->port_vlan_add(ds, port, vlan, extack);
trace_dsa_vlan_add_hw(dp, vlan, err);
return err;
}
/* No need to propagate on shared ports the existing VLANs that were
* re-notified after just the flags have changed. This would cause a
* refcount bump which we need to avoid, since it unbalances the
* additions with the deletions.
*/
if (vlan->changed)
return 0;
mutex_lock(&dp->vlans_lock);
v = dsa_vlan_find(&dp->vlans, vlan);
if (v) {
refcount_inc(&v->refcount);
trace_dsa_vlan_add_bump(dp, vlan, &v->refcount);
goto out;
}
v = kzalloc(sizeof(*v), GFP_KERNEL);
if (!v) {
err = -ENOMEM;
goto out;
}
err = ds->ops->port_vlan_add(ds, port, vlan, extack);
trace_dsa_vlan_add_hw(dp, vlan, err);
if (err) {
kfree(v);
goto out;
}
v->vid = vlan->vid;
refcount_set(&v->refcount, 1);
list_add_tail(&v->list, &dp->vlans);
out:
mutex_unlock(&dp->vlans_lock);
return err;
}
static int dsa_port_do_vlan_del(struct dsa_port *dp,
const struct switchdev_obj_port_vlan *vlan)
{
struct dsa_switch *ds = dp->ds;
int port = dp->index;
struct dsa_vlan *v;
int err = 0;
/* No need to bother with refcounting for user ports */
if (!(dsa_port_is_cpu(dp) || dsa_port_is_dsa(dp))) {
err = ds->ops->port_vlan_del(ds, port, vlan);
trace_dsa_vlan_del_hw(dp, vlan, err);
return err;
}
mutex_lock(&dp->vlans_lock);
v = dsa_vlan_find(&dp->vlans, vlan);
if (!v) {
trace_dsa_vlan_del_not_found(dp, vlan);
err = -ENOENT;
goto out;
}
if (!refcount_dec_and_test(&v->refcount)) {
trace_dsa_vlan_del_drop(dp, vlan, &v->refcount);
goto out;
}
err = ds->ops->port_vlan_del(ds, port, vlan);
trace_dsa_vlan_del_hw(dp, vlan, err);
if (err) {
refcount_set(&v->refcount, 1);
goto out;
}
list_del(&v->list);
kfree(v);
out:
mutex_unlock(&dp->vlans_lock);
return err;
}
static int dsa_switch_vlan_add(struct dsa_switch *ds,
struct dsa_notifier_vlan_info *info)
{
struct dsa_port *dp;
int err;
if (!ds->ops->port_vlan_add)
return -EOPNOTSUPP;
dsa_switch_for_each_port(dp, ds) {
if (dsa_port_vlan_match(dp, info)) {
err = dsa_port_do_vlan_add(dp, info->vlan,
info->extack);
if (err)
return err;
}
}
return 0;
}
static int dsa_switch_vlan_del(struct dsa_switch *ds,
struct dsa_notifier_vlan_info *info)
{
struct dsa_port *dp;
int err;
if (!ds->ops->port_vlan_del)
return -EOPNOTSUPP;
dsa_switch_for_each_port(dp, ds) {
if (dsa_port_vlan_match(dp, info)) {
err = dsa_port_do_vlan_del(dp, info->vlan);
if (err)
return err;
}
}
return 0;
}
static int dsa_switch_host_vlan_add(struct dsa_switch *ds,
struct dsa_notifier_vlan_info *info)
{
struct dsa_port *dp;
int err;
if (!ds->ops->port_vlan_add)
return -EOPNOTSUPP;
dsa_switch_for_each_port(dp, ds) {
if (dsa_port_host_vlan_match(dp, info->dp)) {
err = dsa_port_do_vlan_add(dp, info->vlan,
info->extack);
if (err)
return err;
}
}
return 0;
}
static int dsa_switch_host_vlan_del(struct dsa_switch *ds,
struct dsa_notifier_vlan_info *info)
{
struct dsa_port *dp;
int err;
if (!ds->ops->port_vlan_del)
return -EOPNOTSUPP;
dsa_switch_for_each_port(dp, ds) {
if (dsa_port_host_vlan_match(dp, info->dp)) {
err = dsa_port_do_vlan_del(dp, info->vlan);
if (err)
return err;
}
}
return 0;
}
static int dsa_switch_change_tag_proto(struct dsa_switch *ds,
struct dsa_notifier_tag_proto_info *info)
{
const struct dsa_device_ops *tag_ops = info->tag_ops;
struct dsa_port *dp, *cpu_dp;
int err;
if (!ds->ops->change_tag_protocol)
return -EOPNOTSUPP;
ASSERT_RTNL();
err = ds->ops->change_tag_protocol(ds, tag_ops->proto);
if (err)
return err;
dsa_switch_for_each_cpu_port(cpu_dp, ds)
dsa_port_set_tag_protocol(cpu_dp, tag_ops);
/* Now that changing the tag protocol can no longer fail, let's update
* the remaining bits which are "duplicated for faster access", and the
* bits that depend on the tagger, such as the MTU.
*/
dsa_switch_for_each_user_port(dp, ds) {
struct net_device *slave = dp->slave;
dsa_slave_setup_tagger(slave);
/* rtnl_mutex is held in dsa_tree_change_tag_proto */
dsa_slave_change_mtu(slave, slave->mtu);
}
return 0;
}
/* We use the same cross-chip notifiers to inform both the tagger side, as well
* as the switch side, of connection and disconnection events.
* Since ds->tagger_data is owned by the tagger, it isn't a hard error if the
* switch side doesn't support connecting to this tagger, and therefore, the
* fact that we don't disconnect the tagger side doesn't constitute a memory
* leak: the tagger will still operate with persistent per-switch memory, just
* with the switch side unconnected to it. What does constitute a hard error is
* when the switch side supports connecting but fails.
*/
static int
dsa_switch_connect_tag_proto(struct dsa_switch *ds,
struct dsa_notifier_tag_proto_info *info)
{
const struct dsa_device_ops *tag_ops = info->tag_ops;
int err;
/* Notify the new tagger about the connection to this switch */
if (tag_ops->connect) {
err = tag_ops->connect(ds);
if (err)
return err;
}
if (!ds->ops->connect_tag_protocol)
return -EOPNOTSUPP;
/* Notify the switch about the connection to the new tagger */
err = ds->ops->connect_tag_protocol(ds, tag_ops->proto);
if (err) {
/* Revert the new tagger's connection to this tree */
if (tag_ops->disconnect)
tag_ops->disconnect(ds);
return err;
}
return 0;
}
static int
dsa_switch_disconnect_tag_proto(struct dsa_switch *ds,
struct dsa_notifier_tag_proto_info *info)
{
const struct dsa_device_ops *tag_ops = info->tag_ops;
/* Notify the tagger about the disconnection from this switch */
if (tag_ops->disconnect && ds->tagger_data)
tag_ops->disconnect(ds);
/* No need to notify the switch, since it shouldn't have any
* resources to tear down
*/
return 0;
}
static int
dsa_switch_master_state_change(struct dsa_switch *ds,
struct dsa_notifier_master_state_info *info)
{
if (!ds->ops->master_state_change)
return 0;
ds->ops->master_state_change(ds, info->master, info->operational);
return 0;
}
static int dsa_switch_event(struct notifier_block *nb,
unsigned long event, void *info)
{
struct dsa_switch *ds = container_of(nb, struct dsa_switch, nb);
int err;
switch (event) {
case DSA_NOTIFIER_AGEING_TIME:
err = dsa_switch_ageing_time(ds, info);
break;
case DSA_NOTIFIER_BRIDGE_JOIN:
err = dsa_switch_bridge_join(ds, info);
break;
case DSA_NOTIFIER_BRIDGE_LEAVE:
err = dsa_switch_bridge_leave(ds, info);
break;
case DSA_NOTIFIER_FDB_ADD:
err = dsa_switch_fdb_add(ds, info);
break;
case DSA_NOTIFIER_FDB_DEL:
err = dsa_switch_fdb_del(ds, info);
break;
case DSA_NOTIFIER_HOST_FDB_ADD:
err = dsa_switch_host_fdb_add(ds, info);
break;
case DSA_NOTIFIER_HOST_FDB_DEL:
err = dsa_switch_host_fdb_del(ds, info);
break;
case DSA_NOTIFIER_LAG_FDB_ADD:
err = dsa_switch_lag_fdb_add(ds, info);
break;
case DSA_NOTIFIER_LAG_FDB_DEL:
err = dsa_switch_lag_fdb_del(ds, info);
break;
case DSA_NOTIFIER_LAG_CHANGE:
err = dsa_switch_lag_change(ds, info);
break;
case DSA_NOTIFIER_LAG_JOIN:
err = dsa_switch_lag_join(ds, info);
break;
case DSA_NOTIFIER_LAG_LEAVE:
err = dsa_switch_lag_leave(ds, info);
break;
case DSA_NOTIFIER_MDB_ADD:
err = dsa_switch_mdb_add(ds, info);
break;
case DSA_NOTIFIER_MDB_DEL:
err = dsa_switch_mdb_del(ds, info);
break;
case DSA_NOTIFIER_HOST_MDB_ADD:
err = dsa_switch_host_mdb_add(ds, info);
break;
case DSA_NOTIFIER_HOST_MDB_DEL:
err = dsa_switch_host_mdb_del(ds, info);
break;
case DSA_NOTIFIER_VLAN_ADD:
err = dsa_switch_vlan_add(ds, info);
break;
case DSA_NOTIFIER_VLAN_DEL:
err = dsa_switch_vlan_del(ds, info);
break;
case DSA_NOTIFIER_HOST_VLAN_ADD:
err = dsa_switch_host_vlan_add(ds, info);
break;
case DSA_NOTIFIER_HOST_VLAN_DEL:
err = dsa_switch_host_vlan_del(ds, info);
break;
case DSA_NOTIFIER_MTU:
err = dsa_switch_mtu(ds, info);
break;
case DSA_NOTIFIER_TAG_PROTO:
err = dsa_switch_change_tag_proto(ds, info);
break;
case DSA_NOTIFIER_TAG_PROTO_CONNECT:
err = dsa_switch_connect_tag_proto(ds, info);
break;
case DSA_NOTIFIER_TAG_PROTO_DISCONNECT:
err = dsa_switch_disconnect_tag_proto(ds, info);
break;
case DSA_NOTIFIER_TAG_8021Q_VLAN_ADD:
err = dsa_switch_tag_8021q_vlan_add(ds, info);
break;
case DSA_NOTIFIER_TAG_8021Q_VLAN_DEL:
err = dsa_switch_tag_8021q_vlan_del(ds, info);
break;
case DSA_NOTIFIER_MASTER_STATE_CHANGE:
err = dsa_switch_master_state_change(ds, info);
break;
default:
err = -EOPNOTSUPP;
break;
}
if (err)
dev_dbg(ds->dev, "breaking chain for DSA event %lu (%d)\n",
event, err);
return notifier_from_errno(err);
}
/**
* dsa_tree_notify - Execute code for all switches in a DSA switch tree.
* @dst: collection of struct dsa_switch devices to notify.
* @e: event, must be of type DSA_NOTIFIER_*
* @v: event-specific value.
*
* Given a struct dsa_switch_tree, this can be used to run a function once for
* each member DSA switch. The other alternative of traversing the tree is only
* through its ports list, which does not uniquely list the switches.
*/
int dsa_tree_notify(struct dsa_switch_tree *dst, unsigned long e, void *v)
{
struct raw_notifier_head *nh = &dst->nh;
int err;
err = raw_notifier_call_chain(nh, e, v);
return notifier_to_errno(err);
}
/**
* dsa_broadcast - Notify all DSA trees in the system.
* @e: event, must be of type DSA_NOTIFIER_*
* @v: event-specific value.
*
* Can be used to notify the switching fabric of events such as cross-chip
* bridging between disjoint trees (such as islands of tagger-compatible
* switches bridged by an incompatible middle switch).
*
* WARNING: this function is not reliable during probe time, because probing
* between trees is asynchronous and not all DSA trees might have probed.
*/
int dsa_broadcast(unsigned long e, void *v)
{
struct dsa_switch_tree *dst;
int err = 0;
list_for_each_entry(dst, &dsa_tree_list, list) {
err = dsa_tree_notify(dst, e, v);
if (err)
break;
}
return err;
}
int dsa_switch_register_notifier(struct dsa_switch *ds)
{
ds->nb.notifier_call = dsa_switch_event;
return raw_notifier_chain_register(&ds->dst->nh, &ds->nb);
}
void dsa_switch_unregister_notifier(struct dsa_switch *ds)
{
int err;
err = raw_notifier_chain_unregister(&ds->dst->nh, &ds->nb);
if (err)
dev_err(ds->dev, "failed to unregister notifier (%d)\n", err);
}