linux/net/bridge/br_multicast.c

4679 lines
123 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Bridge multicast support.
*
* Copyright (c) 2010 Herbert Xu <herbert@gondor.apana.org.au>
*/
#include <linux/err.h>
#include <linux/export.h>
#include <linux/if_ether.h>
#include <linux/igmp.h>
#include <linux/in.h>
#include <linux/jhash.h>
#include <linux/kernel.h>
#include <linux/log2.h>
#include <linux/netdevice.h>
#include <linux/netfilter_bridge.h>
#include <linux/random.h>
#include <linux/rculist.h>
#include <linux/skbuff.h>
#include <linux/slab.h>
#include <linux/timer.h>
#include <linux/inetdevice.h>
#include <linux/mroute.h>
#include <net/ip.h>
#include <net/switchdev.h>
#if IS_ENABLED(CONFIG_IPV6)
#include <linux/icmpv6.h>
#include <net/ipv6.h>
#include <net/mld.h>
#include <net/ip6_checksum.h>
#include <net/addrconf.h>
#endif
#include "br_private.h"
#include "br_private_mcast_eht.h"
static const struct rhashtable_params br_mdb_rht_params = {
.head_offset = offsetof(struct net_bridge_mdb_entry, rhnode),
.key_offset = offsetof(struct net_bridge_mdb_entry, addr),
.key_len = sizeof(struct br_ip),
.automatic_shrinking = true,
};
static const struct rhashtable_params br_sg_port_rht_params = {
.head_offset = offsetof(struct net_bridge_port_group, rhnode),
.key_offset = offsetof(struct net_bridge_port_group, key),
.key_len = sizeof(struct net_bridge_port_group_sg_key),
.automatic_shrinking = true,
};
static void br_multicast_start_querier(struct net_bridge_mcast *brmctx,
struct bridge_mcast_own_query *query);
static void br_ip4_multicast_add_router(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx);
static void br_ip4_multicast_leave_group(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
__be32 group,
__u16 vid,
const unsigned char *src);
static void br_multicast_port_group_rexmit(struct timer_list *t);
static void
br_multicast_rport_del_notify(struct net_bridge_mcast_port *pmctx, bool deleted);
static void br_ip6_multicast_add_router(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx);
#if IS_ENABLED(CONFIG_IPV6)
static void br_ip6_multicast_leave_group(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
const struct in6_addr *group,
__u16 vid, const unsigned char *src);
#endif
static struct net_bridge_port_group *
__br_multicast_add_group(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct br_ip *group,
const unsigned char *src,
u8 filter_mode,
bool igmpv2_mldv1,
bool blocked);
static void br_multicast_find_del_pg(struct net_bridge *br,
struct net_bridge_port_group *pg);
static void __br_multicast_stop(struct net_bridge_mcast *brmctx);
static struct net_bridge_port_group *
br_sg_port_find(struct net_bridge *br,
struct net_bridge_port_group_sg_key *sg_p)
{
lockdep_assert_held_once(&br->multicast_lock);
return rhashtable_lookup_fast(&br->sg_port_tbl, sg_p,
br_sg_port_rht_params);
}
static struct net_bridge_mdb_entry *br_mdb_ip_get_rcu(struct net_bridge *br,
struct br_ip *dst)
{
return rhashtable_lookup(&br->mdb_hash_tbl, dst, br_mdb_rht_params);
}
struct net_bridge_mdb_entry *br_mdb_ip_get(struct net_bridge *br,
struct br_ip *dst)
{
struct net_bridge_mdb_entry *ent;
lockdep_assert_held_once(&br->multicast_lock);
rcu_read_lock();
ent = rhashtable_lookup(&br->mdb_hash_tbl, dst, br_mdb_rht_params);
rcu_read_unlock();
return ent;
}
static struct net_bridge_mdb_entry *br_mdb_ip4_get(struct net_bridge *br,
__be32 dst, __u16 vid)
{
struct br_ip br_dst;
memset(&br_dst, 0, sizeof(br_dst));
br_dst.dst.ip4 = dst;
br_dst.proto = htons(ETH_P_IP);
br_dst.vid = vid;
return br_mdb_ip_get(br, &br_dst);
}
#if IS_ENABLED(CONFIG_IPV6)
static struct net_bridge_mdb_entry *br_mdb_ip6_get(struct net_bridge *br,
const struct in6_addr *dst,
__u16 vid)
{
struct br_ip br_dst;
memset(&br_dst, 0, sizeof(br_dst));
br_dst.dst.ip6 = *dst;
br_dst.proto = htons(ETH_P_IPV6);
br_dst.vid = vid;
return br_mdb_ip_get(br, &br_dst);
}
#endif
struct net_bridge_mdb_entry *br_mdb_get(struct net_bridge_mcast *brmctx,
struct sk_buff *skb, u16 vid)
{
struct net_bridge *br = brmctx->br;
struct br_ip ip;
if (!br_opt_get(br, BROPT_MULTICAST_ENABLED) ||
br_multicast_ctx_vlan_global_disabled(brmctx))
return NULL;
if (BR_INPUT_SKB_CB(skb)->igmp)
return NULL;
memset(&ip, 0, sizeof(ip));
ip.proto = skb->protocol;
ip.vid = vid;
switch (skb->protocol) {
case htons(ETH_P_IP):
ip.dst.ip4 = ip_hdr(skb)->daddr;
if (brmctx->multicast_igmp_version == 3) {
struct net_bridge_mdb_entry *mdb;
ip.src.ip4 = ip_hdr(skb)->saddr;
mdb = br_mdb_ip_get_rcu(br, &ip);
if (mdb)
return mdb;
ip.src.ip4 = 0;
}
break;
#if IS_ENABLED(CONFIG_IPV6)
case htons(ETH_P_IPV6):
ip.dst.ip6 = ipv6_hdr(skb)->daddr;
if (brmctx->multicast_mld_version == 2) {
struct net_bridge_mdb_entry *mdb;
ip.src.ip6 = ipv6_hdr(skb)->saddr;
mdb = br_mdb_ip_get_rcu(br, &ip);
if (mdb)
return mdb;
memset(&ip.src.ip6, 0, sizeof(ip.src.ip6));
}
break;
#endif
default:
ip.proto = 0;
ether_addr_copy(ip.dst.mac_addr, eth_hdr(skb)->h_dest);
}
return br_mdb_ip_get_rcu(br, &ip);
}
/* IMPORTANT: this function must be used only when the contexts cannot be
* passed down (e.g. timer) and must be used for read-only purposes because
* the vlan snooping option can change, so it can return any context
* (non-vlan or vlan). Its initial intended purpose is to read timer values
* from the *current* context based on the option. At worst that could lead
* to inconsistent timers when the contexts are changed, i.e. src timer
* which needs to re-arm with a specific delay taken from the old context
*/
static struct net_bridge_mcast_port *
br_multicast_pg_to_port_ctx(const struct net_bridge_port_group *pg)
{
struct net_bridge_mcast_port *pmctx = &pg->key.port->multicast_ctx;
struct net_bridge_vlan *vlan;
lockdep_assert_held_once(&pg->key.port->br->multicast_lock);
/* if vlan snooping is disabled use the port's multicast context */
if (!pg->key.addr.vid ||
!br_opt_get(pg->key.port->br, BROPT_MCAST_VLAN_SNOOPING_ENABLED))
goto out;
/* locking is tricky here, due to different rules for multicast and
* vlans we need to take rcu to find the vlan and make sure it has
* the BR_VLFLAG_MCAST_ENABLED flag set, it can only change under
* multicast_lock which must be already held here, so the vlan's pmctx
* can safely be used on return
*/
rcu_read_lock();
vlan = br_vlan_find(nbp_vlan_group(pg->key.port), pg->key.addr.vid);
if (vlan && !br_multicast_port_ctx_vlan_disabled(&vlan->port_mcast_ctx))
pmctx = &vlan->port_mcast_ctx;
else
pmctx = NULL;
rcu_read_unlock();
out:
return pmctx;
}
/* when snooping we need to check if the contexts should be used
* in the following order:
* - if pmctx is non-NULL (port), check if it should be used
* - if pmctx is NULL (bridge), check if brmctx should be used
*/
static bool
br_multicast_ctx_should_use(const struct net_bridge_mcast *brmctx,
const struct net_bridge_mcast_port *pmctx)
{
if (!netif_running(brmctx->br->dev))
return false;
if (pmctx)
return !br_multicast_port_ctx_state_disabled(pmctx);
else
return !br_multicast_ctx_vlan_disabled(brmctx);
}
static bool br_port_group_equal(struct net_bridge_port_group *p,
struct net_bridge_port *port,
const unsigned char *src)
{
if (p->key.port != port)
return false;
if (!(port->flags & BR_MULTICAST_TO_UNICAST))
return true;
return ether_addr_equal(src, p->eth_addr);
}
static void __fwd_add_star_excl(struct net_bridge_mcast_port *pmctx,
struct net_bridge_port_group *pg,
struct br_ip *sg_ip)
{
struct net_bridge_port_group_sg_key sg_key;
struct net_bridge_port_group *src_pg;
struct net_bridge_mcast *brmctx;
memset(&sg_key, 0, sizeof(sg_key));
brmctx = br_multicast_port_ctx_get_global(pmctx);
sg_key.port = pg->key.port;
sg_key.addr = *sg_ip;
if (br_sg_port_find(brmctx->br, &sg_key))
return;
src_pg = __br_multicast_add_group(brmctx, pmctx,
sg_ip, pg->eth_addr,
MCAST_INCLUDE, false, false);
if (IS_ERR_OR_NULL(src_pg) ||
src_pg->rt_protocol != RTPROT_KERNEL)
return;
src_pg->flags |= MDB_PG_FLAGS_STAR_EXCL;
}
static void __fwd_del_star_excl(struct net_bridge_port_group *pg,
struct br_ip *sg_ip)
{
struct net_bridge_port_group_sg_key sg_key;
struct net_bridge *br = pg->key.port->br;
struct net_bridge_port_group *src_pg;
memset(&sg_key, 0, sizeof(sg_key));
sg_key.port = pg->key.port;
sg_key.addr = *sg_ip;
src_pg = br_sg_port_find(br, &sg_key);
if (!src_pg || !(src_pg->flags & MDB_PG_FLAGS_STAR_EXCL) ||
src_pg->rt_protocol != RTPROT_KERNEL)
return;
br_multicast_find_del_pg(br, src_pg);
}
/* When a port group transitions to (or is added as) EXCLUDE we need to add it
* to all other ports' S,G entries which are not blocked by the current group
* for proper replication, the assumption is that any S,G blocked entries
* are already added so the S,G,port lookup should skip them.
* When a port group transitions from EXCLUDE -> INCLUDE mode or is being
* deleted we need to remove it from all ports' S,G entries where it was
* automatically installed before (i.e. where it's MDB_PG_FLAGS_STAR_EXCL).
*/
void br_multicast_star_g_handle_mode(struct net_bridge_port_group *pg,
u8 filter_mode)
{
struct net_bridge *br = pg->key.port->br;
struct net_bridge_port_group *pg_lst;
struct net_bridge_mcast_port *pmctx;
struct net_bridge_mdb_entry *mp;
struct br_ip sg_ip;
if (WARN_ON(!br_multicast_is_star_g(&pg->key.addr)))
return;
mp = br_mdb_ip_get(br, &pg->key.addr);
if (!mp)
return;
pmctx = br_multicast_pg_to_port_ctx(pg);
if (!pmctx)
return;
memset(&sg_ip, 0, sizeof(sg_ip));
sg_ip = pg->key.addr;
for (pg_lst = mlock_dereference(mp->ports, br);
pg_lst;
pg_lst = mlock_dereference(pg_lst->next, br)) {
struct net_bridge_group_src *src_ent;
if (pg_lst == pg)
continue;
hlist_for_each_entry(src_ent, &pg_lst->src_list, node) {
if (!(src_ent->flags & BR_SGRP_F_INSTALLED))
continue;
sg_ip.src = src_ent->addr.src;
switch (filter_mode) {
case MCAST_INCLUDE:
__fwd_del_star_excl(pg, &sg_ip);
break;
case MCAST_EXCLUDE:
__fwd_add_star_excl(pmctx, pg, &sg_ip);
break;
}
}
}
}
/* called when adding a new S,G with host_joined == false by default */
static void br_multicast_sg_host_state(struct net_bridge_mdb_entry *star_mp,
struct net_bridge_port_group *sg)
{
struct net_bridge_mdb_entry *sg_mp;
if (WARN_ON(!br_multicast_is_star_g(&star_mp->addr)))
return;
if (!star_mp->host_joined)
return;
sg_mp = br_mdb_ip_get(star_mp->br, &sg->key.addr);
if (!sg_mp)
return;
sg_mp->host_joined = true;
}
/* set the host_joined state of all of *,G's S,G entries */
static void br_multicast_star_g_host_state(struct net_bridge_mdb_entry *star_mp)
{
struct net_bridge *br = star_mp->br;
struct net_bridge_mdb_entry *sg_mp;
struct net_bridge_port_group *pg;
struct br_ip sg_ip;
if (WARN_ON(!br_multicast_is_star_g(&star_mp->addr)))
return;
memset(&sg_ip, 0, sizeof(sg_ip));
sg_ip = star_mp->addr;
for (pg = mlock_dereference(star_mp->ports, br);
pg;
pg = mlock_dereference(pg->next, br)) {
struct net_bridge_group_src *src_ent;
hlist_for_each_entry(src_ent, &pg->src_list, node) {
if (!(src_ent->flags & BR_SGRP_F_INSTALLED))
continue;
sg_ip.src = src_ent->addr.src;
sg_mp = br_mdb_ip_get(br, &sg_ip);
if (!sg_mp)
continue;
sg_mp->host_joined = star_mp->host_joined;
}
}
}
static void br_multicast_sg_del_exclude_ports(struct net_bridge_mdb_entry *sgmp)
{
struct net_bridge_port_group __rcu **pp;
struct net_bridge_port_group *p;
/* *,G exclude ports are only added to S,G entries */
if (WARN_ON(br_multicast_is_star_g(&sgmp->addr)))
return;
/* we need the STAR_EXCLUDE ports if there are non-STAR_EXCLUDE ports
* we should ignore perm entries since they're managed by user-space
*/
for (pp = &sgmp->ports;
(p = mlock_dereference(*pp, sgmp->br)) != NULL;
pp = &p->next)
if (!(p->flags & (MDB_PG_FLAGS_STAR_EXCL |
MDB_PG_FLAGS_PERMANENT)))
return;
/* currently the host can only have joined the *,G which means
* we treat it as EXCLUDE {}, so for an S,G it's considered a
* STAR_EXCLUDE entry and we can safely leave it
*/
sgmp->host_joined = false;
for (pp = &sgmp->ports;
(p = mlock_dereference(*pp, sgmp->br)) != NULL;) {
if (!(p->flags & MDB_PG_FLAGS_PERMANENT))
br_multicast_del_pg(sgmp, p, pp);
else
pp = &p->next;
}
}
void br_multicast_sg_add_exclude_ports(struct net_bridge_mdb_entry *star_mp,
struct net_bridge_port_group *sg)
{
struct net_bridge_port_group_sg_key sg_key;
struct net_bridge *br = star_mp->br;
struct net_bridge_mcast_port *pmctx;
struct net_bridge_port_group *pg;
struct net_bridge_mcast *brmctx;
if (WARN_ON(br_multicast_is_star_g(&sg->key.addr)))
return;
if (WARN_ON(!br_multicast_is_star_g(&star_mp->addr)))
return;
br_multicast_sg_host_state(star_mp, sg);
memset(&sg_key, 0, sizeof(sg_key));
sg_key.addr = sg->key.addr;
/* we need to add all exclude ports to the S,G */
for (pg = mlock_dereference(star_mp->ports, br);
pg;
pg = mlock_dereference(pg->next, br)) {
struct net_bridge_port_group *src_pg;
if (pg == sg || pg->filter_mode == MCAST_INCLUDE)
continue;
sg_key.port = pg->key.port;
if (br_sg_port_find(br, &sg_key))
continue;
pmctx = br_multicast_pg_to_port_ctx(pg);
if (!pmctx)
continue;
brmctx = br_multicast_port_ctx_get_global(pmctx);
src_pg = __br_multicast_add_group(brmctx, pmctx,
&sg->key.addr,
sg->eth_addr,
MCAST_INCLUDE, false, false);
if (IS_ERR_OR_NULL(src_pg) ||
src_pg->rt_protocol != RTPROT_KERNEL)
continue;
src_pg->flags |= MDB_PG_FLAGS_STAR_EXCL;
}
}
static void br_multicast_fwd_src_add(struct net_bridge_group_src *src)
{
struct net_bridge_mdb_entry *star_mp;
struct net_bridge_mcast_port *pmctx;
struct net_bridge_port_group *sg;
struct net_bridge_mcast *brmctx;
struct br_ip sg_ip;
if (src->flags & BR_SGRP_F_INSTALLED)
return;
memset(&sg_ip, 0, sizeof(sg_ip));
pmctx = br_multicast_pg_to_port_ctx(src->pg);
if (!pmctx)
return;
brmctx = br_multicast_port_ctx_get_global(pmctx);
sg_ip = src->pg->key.addr;
sg_ip.src = src->addr.src;
sg = __br_multicast_add_group(brmctx, pmctx, &sg_ip,
src->pg->eth_addr, MCAST_INCLUDE, false,
!timer_pending(&src->timer));
if (IS_ERR_OR_NULL(sg))
return;
src->flags |= BR_SGRP_F_INSTALLED;
sg->flags &= ~MDB_PG_FLAGS_STAR_EXCL;
/* if it was added by user-space as perm we can skip next steps */
if (sg->rt_protocol != RTPROT_KERNEL &&
(sg->flags & MDB_PG_FLAGS_PERMANENT))
return;
/* the kernel is now responsible for removing this S,G */
del_timer(&sg->timer);
star_mp = br_mdb_ip_get(src->br, &src->pg->key.addr);
if (!star_mp)
return;
br_multicast_sg_add_exclude_ports(star_mp, sg);
}
static void br_multicast_fwd_src_remove(struct net_bridge_group_src *src,
bool fastleave)
{
struct net_bridge_port_group *p, *pg = src->pg;
struct net_bridge_port_group __rcu **pp;
struct net_bridge_mdb_entry *mp;
struct br_ip sg_ip;
memset(&sg_ip, 0, sizeof(sg_ip));
sg_ip = pg->key.addr;
sg_ip.src = src->addr.src;
mp = br_mdb_ip_get(src->br, &sg_ip);
if (!mp)
return;
for (pp = &mp->ports;
(p = mlock_dereference(*pp, src->br)) != NULL;
pp = &p->next) {
if (!br_port_group_equal(p, pg->key.port, pg->eth_addr))
continue;
if (p->rt_protocol != RTPROT_KERNEL &&
(p->flags & MDB_PG_FLAGS_PERMANENT))
break;
if (fastleave)
p->flags |= MDB_PG_FLAGS_FAST_LEAVE;
br_multicast_del_pg(mp, p, pp);
break;
}
src->flags &= ~BR_SGRP_F_INSTALLED;
}
/* install S,G and based on src's timer enable or disable forwarding */
static void br_multicast_fwd_src_handle(struct net_bridge_group_src *src)
{
struct net_bridge_port_group_sg_key sg_key;
struct net_bridge_port_group *sg;
u8 old_flags;
br_multicast_fwd_src_add(src);
memset(&sg_key, 0, sizeof(sg_key));
sg_key.addr = src->pg->key.addr;
sg_key.addr.src = src->addr.src;
sg_key.port = src->pg->key.port;
sg = br_sg_port_find(src->br, &sg_key);
if (!sg || (sg->flags & MDB_PG_FLAGS_PERMANENT))
return;
old_flags = sg->flags;
if (timer_pending(&src->timer))
sg->flags &= ~MDB_PG_FLAGS_BLOCKED;
else
sg->flags |= MDB_PG_FLAGS_BLOCKED;
if (old_flags != sg->flags) {
struct net_bridge_mdb_entry *sg_mp;
sg_mp = br_mdb_ip_get(src->br, &sg_key.addr);
if (!sg_mp)
return;
br_mdb_notify(src->br->dev, sg_mp, sg, RTM_NEWMDB);
}
}
static void br_multicast_destroy_mdb_entry(struct net_bridge_mcast_gc *gc)
{
struct net_bridge_mdb_entry *mp;
mp = container_of(gc, struct net_bridge_mdb_entry, mcast_gc);
WARN_ON(!hlist_unhashed(&mp->mdb_node));
WARN_ON(mp->ports);
del_timer_sync(&mp->timer);
kfree_rcu(mp, rcu);
}
static void br_multicast_del_mdb_entry(struct net_bridge_mdb_entry *mp)
{
struct net_bridge *br = mp->br;
rhashtable_remove_fast(&br->mdb_hash_tbl, &mp->rhnode,
br_mdb_rht_params);
hlist_del_init_rcu(&mp->mdb_node);
hlist_add_head(&mp->mcast_gc.gc_node, &br->mcast_gc_list);
queue_work(system_long_wq, &br->mcast_gc_work);
}
static void br_multicast_group_expired(struct timer_list *t)
{
struct net_bridge_mdb_entry *mp = from_timer(mp, t, timer);
struct net_bridge *br = mp->br;
spin_lock(&br->multicast_lock);
if (hlist_unhashed(&mp->mdb_node) || !netif_running(br->dev) ||
timer_pending(&mp->timer))
goto out;
br_multicast_host_leave(mp, true);
if (mp->ports)
goto out;
br_multicast_del_mdb_entry(mp);
out:
spin_unlock(&br->multicast_lock);
}
static void br_multicast_destroy_group_src(struct net_bridge_mcast_gc *gc)
{
struct net_bridge_group_src *src;
src = container_of(gc, struct net_bridge_group_src, mcast_gc);
WARN_ON(!hlist_unhashed(&src->node));
del_timer_sync(&src->timer);
kfree_rcu(src, rcu);
}
void br_multicast_del_group_src(struct net_bridge_group_src *src,
bool fastleave)
{
struct net_bridge *br = src->pg->key.port->br;
br_multicast_fwd_src_remove(src, fastleave);
hlist_del_init_rcu(&src->node);
src->pg->src_ents--;
hlist_add_head(&src->mcast_gc.gc_node, &br->mcast_gc_list);
queue_work(system_long_wq, &br->mcast_gc_work);
}
static void br_multicast_destroy_port_group(struct net_bridge_mcast_gc *gc)
{
struct net_bridge_port_group *pg;
pg = container_of(gc, struct net_bridge_port_group, mcast_gc);
WARN_ON(!hlist_unhashed(&pg->mglist));
WARN_ON(!hlist_empty(&pg->src_list));
del_timer_sync(&pg->rexmit_timer);
del_timer_sync(&pg->timer);
kfree_rcu(pg, rcu);
}
void br_multicast_del_pg(struct net_bridge_mdb_entry *mp,
struct net_bridge_port_group *pg,
struct net_bridge_port_group __rcu **pp)
{
struct net_bridge *br = pg->key.port->br;
struct net_bridge_group_src *ent;
struct hlist_node *tmp;
rcu_assign_pointer(*pp, pg->next);
hlist_del_init(&pg->mglist);
br_multicast_eht_clean_sets(pg);
hlist_for_each_entry_safe(ent, tmp, &pg->src_list, node)
br_multicast_del_group_src(ent, false);
br_mdb_notify(br->dev, mp, pg, RTM_DELMDB);
if (!br_multicast_is_star_g(&mp->addr)) {
rhashtable_remove_fast(&br->sg_port_tbl, &pg->rhnode,
br_sg_port_rht_params);
br_multicast_sg_del_exclude_ports(mp);
} else {
br_multicast_star_g_handle_mode(pg, MCAST_INCLUDE);
}
hlist_add_head(&pg->mcast_gc.gc_node, &br->mcast_gc_list);
queue_work(system_long_wq, &br->mcast_gc_work);
if (!mp->ports && !mp->host_joined && netif_running(br->dev))
mod_timer(&mp->timer, jiffies);
}
static void br_multicast_find_del_pg(struct net_bridge *br,
struct net_bridge_port_group *pg)
{
struct net_bridge_port_group __rcu **pp;
struct net_bridge_mdb_entry *mp;
struct net_bridge_port_group *p;
mp = br_mdb_ip_get(br, &pg->key.addr);
if (WARN_ON(!mp))
return;
for (pp = &mp->ports;
(p = mlock_dereference(*pp, br)) != NULL;
pp = &p->next) {
if (p != pg)
continue;
br_multicast_del_pg(mp, pg, pp);
return;
}
WARN_ON(1);
}
static void br_multicast_port_group_expired(struct timer_list *t)
{
struct net_bridge_port_group *pg = from_timer(pg, t, timer);
struct net_bridge_group_src *src_ent;
struct net_bridge *br = pg->key.port->br;
struct hlist_node *tmp;
bool changed;
spin_lock(&br->multicast_lock);
if (!netif_running(br->dev) || timer_pending(&pg->timer) ||
hlist_unhashed(&pg->mglist) || pg->flags & MDB_PG_FLAGS_PERMANENT)
goto out;
changed = !!(pg->filter_mode == MCAST_EXCLUDE);
pg->filter_mode = MCAST_INCLUDE;
hlist_for_each_entry_safe(src_ent, tmp, &pg->src_list, node) {
if (!timer_pending(&src_ent->timer)) {
br_multicast_del_group_src(src_ent, false);
changed = true;
}
}
if (hlist_empty(&pg->src_list)) {
br_multicast_find_del_pg(br, pg);
} else if (changed) {
struct net_bridge_mdb_entry *mp = br_mdb_ip_get(br, &pg->key.addr);
if (changed && br_multicast_is_star_g(&pg->key.addr))
br_multicast_star_g_handle_mode(pg, MCAST_INCLUDE);
if (WARN_ON(!mp))
goto out;
br_mdb_notify(br->dev, mp, pg, RTM_NEWMDB);
}
out:
spin_unlock(&br->multicast_lock);
}
static void br_multicast_gc(struct hlist_head *head)
{
struct net_bridge_mcast_gc *gcent;
struct hlist_node *tmp;
hlist_for_each_entry_safe(gcent, tmp, head, gc_node) {
hlist_del_init(&gcent->gc_node);
gcent->destroy(gcent);
}
}
static struct sk_buff *br_ip4_multicast_alloc_query(struct net_bridge_mcast *brmctx,
struct net_bridge_port_group *pg,
__be32 ip_dst, __be32 group,
bool with_srcs, bool over_lmqt,
u8 sflag, u8 *igmp_type,
bool *need_rexmit)
{
struct net_bridge_port *p = pg ? pg->key.port : NULL;
struct net_bridge_group_src *ent;
size_t pkt_size, igmp_hdr_size;
unsigned long now = jiffies;
struct igmpv3_query *ihv3;
void *csum_start = NULL;
__sum16 *csum = NULL;
struct sk_buff *skb;
struct igmphdr *ih;
struct ethhdr *eth;
unsigned long lmqt;
struct iphdr *iph;
u16 lmqt_srcs = 0;
igmp_hdr_size = sizeof(*ih);
if (brmctx->multicast_igmp_version == 3) {
igmp_hdr_size = sizeof(*ihv3);
if (pg && with_srcs) {
lmqt = now + (brmctx->multicast_last_member_interval *
brmctx->multicast_last_member_count);
hlist_for_each_entry(ent, &pg->src_list, node) {
if (over_lmqt == time_after(ent->timer.expires,
lmqt) &&
ent->src_query_rexmit_cnt > 0)
lmqt_srcs++;
}
if (!lmqt_srcs)
return NULL;
igmp_hdr_size += lmqt_srcs * sizeof(__be32);
}
}
pkt_size = sizeof(*eth) + sizeof(*iph) + 4 + igmp_hdr_size;
if ((p && pkt_size > p->dev->mtu) ||
pkt_size > brmctx->br->dev->mtu)
return NULL;
skb = netdev_alloc_skb_ip_align(brmctx->br->dev, pkt_size);
if (!skb)
goto out;
skb->protocol = htons(ETH_P_IP);
skb_reset_mac_header(skb);
eth = eth_hdr(skb);
ether_addr_copy(eth->h_source, brmctx->br->dev->dev_addr);
ip_eth_mc_map(ip_dst, eth->h_dest);
eth->h_proto = htons(ETH_P_IP);
skb_put(skb, sizeof(*eth));
skb_set_network_header(skb, skb->len);
iph = ip_hdr(skb);
iph->tot_len = htons(pkt_size - sizeof(*eth));
iph->version = 4;
iph->ihl = 6;
iph->tos = 0xc0;
iph->id = 0;
iph->frag_off = htons(IP_DF);
iph->ttl = 1;
iph->protocol = IPPROTO_IGMP;
iph->saddr = br_opt_get(brmctx->br, BROPT_MULTICAST_QUERY_USE_IFADDR) ?
inet_select_addr(brmctx->br->dev, 0, RT_SCOPE_LINK) : 0;
iph->daddr = ip_dst;
((u8 *)&iph[1])[0] = IPOPT_RA;
((u8 *)&iph[1])[1] = 4;
((u8 *)&iph[1])[2] = 0;
((u8 *)&iph[1])[3] = 0;
ip_send_check(iph);
skb_put(skb, 24);
skb_set_transport_header(skb, skb->len);
*igmp_type = IGMP_HOST_MEMBERSHIP_QUERY;
switch (brmctx->multicast_igmp_version) {
case 2:
ih = igmp_hdr(skb);
ih->type = IGMP_HOST_MEMBERSHIP_QUERY;
ih->code = (group ? brmctx->multicast_last_member_interval :
brmctx->multicast_query_response_interval) /
(HZ / IGMP_TIMER_SCALE);
ih->group = group;
ih->csum = 0;
csum = &ih->csum;
csum_start = (void *)ih;
break;
case 3:
ihv3 = igmpv3_query_hdr(skb);
ihv3->type = IGMP_HOST_MEMBERSHIP_QUERY;
ihv3->code = (group ? brmctx->multicast_last_member_interval :
brmctx->multicast_query_response_interval) /
(HZ / IGMP_TIMER_SCALE);
ihv3->group = group;
ihv3->qqic = brmctx->multicast_query_interval / HZ;
ihv3->nsrcs = htons(lmqt_srcs);
ihv3->resv = 0;
ihv3->suppress = sflag;
ihv3->qrv = 2;
ihv3->csum = 0;
csum = &ihv3->csum;
csum_start = (void *)ihv3;
if (!pg || !with_srcs)
break;
lmqt_srcs = 0;
hlist_for_each_entry(ent, &pg->src_list, node) {
if (over_lmqt == time_after(ent->timer.expires,
lmqt) &&
ent->src_query_rexmit_cnt > 0) {
ihv3->srcs[lmqt_srcs++] = ent->addr.src.ip4;
ent->src_query_rexmit_cnt--;
if (need_rexmit && ent->src_query_rexmit_cnt)
*need_rexmit = true;
}
}
if (WARN_ON(lmqt_srcs != ntohs(ihv3->nsrcs))) {
kfree_skb(skb);
return NULL;
}
break;
}
if (WARN_ON(!csum || !csum_start)) {
kfree_skb(skb);
return NULL;
}
*csum = ip_compute_csum(csum_start, igmp_hdr_size);
skb_put(skb, igmp_hdr_size);
__skb_pull(skb, sizeof(*eth));
out:
return skb;
}
#if IS_ENABLED(CONFIG_IPV6)
static struct sk_buff *br_ip6_multicast_alloc_query(struct net_bridge_mcast *brmctx,
struct net_bridge_port_group *pg,
const struct in6_addr *ip6_dst,
const struct in6_addr *group,
bool with_srcs, bool over_llqt,
u8 sflag, u8 *igmp_type,
bool *need_rexmit)
{
struct net_bridge_port *p = pg ? pg->key.port : NULL;
struct net_bridge_group_src *ent;
size_t pkt_size, mld_hdr_size;
unsigned long now = jiffies;
struct mld2_query *mld2q;
void *csum_start = NULL;
unsigned long interval;
__sum16 *csum = NULL;
struct ipv6hdr *ip6h;
struct mld_msg *mldq;
struct sk_buff *skb;
unsigned long llqt;
struct ethhdr *eth;
u16 llqt_srcs = 0;
u8 *hopopt;
mld_hdr_size = sizeof(*mldq);
if (brmctx->multicast_mld_version == 2) {
mld_hdr_size = sizeof(*mld2q);
if (pg && with_srcs) {
llqt = now + (brmctx->multicast_last_member_interval *
brmctx->multicast_last_member_count);
hlist_for_each_entry(ent, &pg->src_list, node) {
if (over_llqt == time_after(ent->timer.expires,
llqt) &&
ent->src_query_rexmit_cnt > 0)
llqt_srcs++;
}
if (!llqt_srcs)
return NULL;
mld_hdr_size += llqt_srcs * sizeof(struct in6_addr);
}
}
pkt_size = sizeof(*eth) + sizeof(*ip6h) + 8 + mld_hdr_size;
if ((p && pkt_size > p->dev->mtu) ||
pkt_size > brmctx->br->dev->mtu)
return NULL;
skb = netdev_alloc_skb_ip_align(brmctx->br->dev, pkt_size);
if (!skb)
goto out;
skb->protocol = htons(ETH_P_IPV6);
/* Ethernet header */
skb_reset_mac_header(skb);
eth = eth_hdr(skb);
ether_addr_copy(eth->h_source, brmctx->br->dev->dev_addr);
eth->h_proto = htons(ETH_P_IPV6);
skb_put(skb, sizeof(*eth));
/* IPv6 header + HbH option */
skb_set_network_header(skb, skb->len);
ip6h = ipv6_hdr(skb);
*(__force __be32 *)ip6h = htonl(0x60000000);
ip6h->payload_len = htons(8 + mld_hdr_size);
ip6h->nexthdr = IPPROTO_HOPOPTS;
ip6h->hop_limit = 1;
ip6h->daddr = *ip6_dst;
if (ipv6_dev_get_saddr(dev_net(brmctx->br->dev), brmctx->br->dev,
&ip6h->daddr, 0, &ip6h->saddr)) {
kfree_skb(skb);
br_opt_toggle(brmctx->br, BROPT_HAS_IPV6_ADDR, false);
return NULL;
}
br_opt_toggle(brmctx->br, BROPT_HAS_IPV6_ADDR, true);
ipv6_eth_mc_map(&ip6h->daddr, eth->h_dest);
hopopt = (u8 *)(ip6h + 1);
hopopt[0] = IPPROTO_ICMPV6; /* next hdr */
hopopt[1] = 0; /* length of HbH */
hopopt[2] = IPV6_TLV_ROUTERALERT; /* Router Alert */
hopopt[3] = 2; /* Length of RA Option */
hopopt[4] = 0; /* Type = 0x0000 (MLD) */
hopopt[5] = 0;
hopopt[6] = IPV6_TLV_PAD1; /* Pad1 */
hopopt[7] = IPV6_TLV_PAD1; /* Pad1 */
skb_put(skb, sizeof(*ip6h) + 8);
/* ICMPv6 */
skb_set_transport_header(skb, skb->len);
interval = ipv6_addr_any(group) ?
brmctx->multicast_query_response_interval :
brmctx->multicast_last_member_interval;
*igmp_type = ICMPV6_MGM_QUERY;
switch (brmctx->multicast_mld_version) {
case 1:
mldq = (struct mld_msg *)icmp6_hdr(skb);
mldq->mld_type = ICMPV6_MGM_QUERY;
mldq->mld_code = 0;
mldq->mld_cksum = 0;
mldq->mld_maxdelay = htons((u16)jiffies_to_msecs(interval));
mldq->mld_reserved = 0;
mldq->mld_mca = *group;
csum = &mldq->mld_cksum;
csum_start = (void *)mldq;
break;
case 2:
mld2q = (struct mld2_query *)icmp6_hdr(skb);
mld2q->mld2q_mrc = htons((u16)jiffies_to_msecs(interval));
mld2q->mld2q_type = ICMPV6_MGM_QUERY;
mld2q->mld2q_code = 0;
mld2q->mld2q_cksum = 0;
mld2q->mld2q_resv1 = 0;
mld2q->mld2q_resv2 = 0;
mld2q->mld2q_suppress = sflag;
mld2q->mld2q_qrv = 2;
mld2q->mld2q_nsrcs = htons(llqt_srcs);
mld2q->mld2q_qqic = brmctx->multicast_query_interval / HZ;
mld2q->mld2q_mca = *group;
csum = &mld2q->mld2q_cksum;
csum_start = (void *)mld2q;
if (!pg || !with_srcs)
break;
llqt_srcs = 0;
hlist_for_each_entry(ent, &pg->src_list, node) {
if (over_llqt == time_after(ent->timer.expires,
llqt) &&
ent->src_query_rexmit_cnt > 0) {
mld2q->mld2q_srcs[llqt_srcs++] = ent->addr.src.ip6;
ent->src_query_rexmit_cnt--;
if (need_rexmit && ent->src_query_rexmit_cnt)
*need_rexmit = true;
}
}
if (WARN_ON(llqt_srcs != ntohs(mld2q->mld2q_nsrcs))) {
kfree_skb(skb);
return NULL;
}
break;
}
if (WARN_ON(!csum || !csum_start)) {
kfree_skb(skb);
return NULL;
}
*csum = csum_ipv6_magic(&ip6h->saddr, &ip6h->daddr, mld_hdr_size,
IPPROTO_ICMPV6,
csum_partial(csum_start, mld_hdr_size, 0));
skb_put(skb, mld_hdr_size);
__skb_pull(skb, sizeof(*eth));
out:
return skb;
}
#endif
static struct sk_buff *br_multicast_alloc_query(struct net_bridge_mcast *brmctx,
struct net_bridge_port_group *pg,
struct br_ip *ip_dst,
struct br_ip *group,
bool with_srcs, bool over_lmqt,
u8 sflag, u8 *igmp_type,
bool *need_rexmit)
{
__be32 ip4_dst;
switch (group->proto) {
case htons(ETH_P_IP):
ip4_dst = ip_dst ? ip_dst->dst.ip4 : htonl(INADDR_ALLHOSTS_GROUP);
return br_ip4_multicast_alloc_query(brmctx, pg,
ip4_dst, group->dst.ip4,
with_srcs, over_lmqt,
sflag, igmp_type,
need_rexmit);
#if IS_ENABLED(CONFIG_IPV6)
case htons(ETH_P_IPV6): {
struct in6_addr ip6_dst;
if (ip_dst)
ip6_dst = ip_dst->dst.ip6;
else
ipv6_addr_set(&ip6_dst, htonl(0xff020000), 0, 0,
htonl(1));
return br_ip6_multicast_alloc_query(brmctx, pg,
&ip6_dst, &group->dst.ip6,
with_srcs, over_lmqt,
sflag, igmp_type,
need_rexmit);
}
#endif
}
return NULL;
}
struct net_bridge_mdb_entry *br_multicast_new_group(struct net_bridge *br,
struct br_ip *group)
{
struct net_bridge_mdb_entry *mp;
int err;
mp = br_mdb_ip_get(br, group);
if (mp)
return mp;
if (atomic_read(&br->mdb_hash_tbl.nelems) >= br->hash_max) {
br_opt_toggle(br, BROPT_MULTICAST_ENABLED, false);
return ERR_PTR(-E2BIG);
}
mp = kzalloc(sizeof(*mp), GFP_ATOMIC);
if (unlikely(!mp))
return ERR_PTR(-ENOMEM);
mp->br = br;
mp->addr = *group;
mp->mcast_gc.destroy = br_multicast_destroy_mdb_entry;
timer_setup(&mp->timer, br_multicast_group_expired, 0);
err = rhashtable_lookup_insert_fast(&br->mdb_hash_tbl, &mp->rhnode,
br_mdb_rht_params);
if (err) {
kfree(mp);
mp = ERR_PTR(err);
} else {
hlist_add_head_rcu(&mp->mdb_node, &br->mdb_list);
}
return mp;
}
static void br_multicast_group_src_expired(struct timer_list *t)
{
struct net_bridge_group_src *src = from_timer(src, t, timer);
struct net_bridge_port_group *pg;
struct net_bridge *br = src->br;
spin_lock(&br->multicast_lock);
if (hlist_unhashed(&src->node) || !netif_running(br->dev) ||
timer_pending(&src->timer))
goto out;
pg = src->pg;
if (pg->filter_mode == MCAST_INCLUDE) {
br_multicast_del_group_src(src, false);
if (!hlist_empty(&pg->src_list))
goto out;
br_multicast_find_del_pg(br, pg);
} else {
br_multicast_fwd_src_handle(src);
}
out:
spin_unlock(&br->multicast_lock);
}
struct net_bridge_group_src *
br_multicast_find_group_src(struct net_bridge_port_group *pg, struct br_ip *ip)
{
struct net_bridge_group_src *ent;
switch (ip->proto) {
case htons(ETH_P_IP):
hlist_for_each_entry(ent, &pg->src_list, node)
if (ip->src.ip4 == ent->addr.src.ip4)
return ent;
break;
#if IS_ENABLED(CONFIG_IPV6)
case htons(ETH_P_IPV6):
hlist_for_each_entry(ent, &pg->src_list, node)
if (!ipv6_addr_cmp(&ent->addr.src.ip6, &ip->src.ip6))
return ent;
break;
#endif
}
return NULL;
}
static struct net_bridge_group_src *
br_multicast_new_group_src(struct net_bridge_port_group *pg, struct br_ip *src_ip)
{
struct net_bridge_group_src *grp_src;
if (unlikely(pg->src_ents >= PG_SRC_ENT_LIMIT))
return NULL;
switch (src_ip->proto) {
case htons(ETH_P_IP):
if (ipv4_is_zeronet(src_ip->src.ip4) ||
ipv4_is_multicast(src_ip->src.ip4))
return NULL;
break;
#if IS_ENABLED(CONFIG_IPV6)
case htons(ETH_P_IPV6):
if (ipv6_addr_any(&src_ip->src.ip6) ||
ipv6_addr_is_multicast(&src_ip->src.ip6))
return NULL;
break;
#endif
}
grp_src = kzalloc(sizeof(*grp_src), GFP_ATOMIC);
if (unlikely(!grp_src))
return NULL;
grp_src->pg = pg;
grp_src->br = pg->key.port->br;
grp_src->addr = *src_ip;
grp_src->mcast_gc.destroy = br_multicast_destroy_group_src;
timer_setup(&grp_src->timer, br_multicast_group_src_expired, 0);
hlist_add_head_rcu(&grp_src->node, &pg->src_list);
pg->src_ents++;
return grp_src;
}
struct net_bridge_port_group *br_multicast_new_port_group(
struct net_bridge_port *port,
struct br_ip *group,
struct net_bridge_port_group __rcu *next,
unsigned char flags,
const unsigned char *src,
u8 filter_mode,
u8 rt_protocol)
{
struct net_bridge_port_group *p;
p = kzalloc(sizeof(*p), GFP_ATOMIC);
if (unlikely(!p))
return NULL;
p->key.addr = *group;
p->key.port = port;
p->flags = flags;
p->filter_mode = filter_mode;
p->rt_protocol = rt_protocol;
p->eht_host_tree = RB_ROOT;
p->eht_set_tree = RB_ROOT;
p->mcast_gc.destroy = br_multicast_destroy_port_group;
INIT_HLIST_HEAD(&p->src_list);
if (!br_multicast_is_star_g(group) &&
rhashtable_lookup_insert_fast(&port->br->sg_port_tbl, &p->rhnode,
br_sg_port_rht_params)) {
kfree(p);
return NULL;
}
rcu_assign_pointer(p->next, next);
timer_setup(&p->timer, br_multicast_port_group_expired, 0);
timer_setup(&p->rexmit_timer, br_multicast_port_group_rexmit, 0);
hlist_add_head(&p->mglist, &port->mglist);
if (src)
memcpy(p->eth_addr, src, ETH_ALEN);
else
eth_broadcast_addr(p->eth_addr);
return p;
}
void br_multicast_host_join(struct net_bridge_mdb_entry *mp, bool notify)
{
if (!mp->host_joined) {
mp->host_joined = true;
if (br_multicast_is_star_g(&mp->addr))
br_multicast_star_g_host_state(mp);
if (notify)
br_mdb_notify(mp->br->dev, mp, NULL, RTM_NEWMDB);
}
if (br_group_is_l2(&mp->addr))
return;
mod_timer(&mp->timer,
jiffies + mp->br->multicast_ctx.multicast_membership_interval);
}
void br_multicast_host_leave(struct net_bridge_mdb_entry *mp, bool notify)
{
if (!mp->host_joined)
return;
mp->host_joined = false;
if (br_multicast_is_star_g(&mp->addr))
br_multicast_star_g_host_state(mp);
if (notify)
br_mdb_notify(mp->br->dev, mp, NULL, RTM_DELMDB);
}
static struct net_bridge_port_group *
__br_multicast_add_group(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct br_ip *group,
const unsigned char *src,
u8 filter_mode,
bool igmpv2_mldv1,
bool blocked)
{
struct net_bridge_port_group __rcu **pp;
struct net_bridge_port_group *p = NULL;
struct net_bridge_mdb_entry *mp;
Revert "bridge: only expire the mdb entry when query is received" While this commit was a good attempt to fix issues occuring when no multicast querier is present, this commit still has two more issues: 1) There are cases where mdb entries do not expire even if there is a querier present. The bridge will unnecessarily continue flooding multicast packets on the according ports. 2) Never removing an mdb entry could be exploited for a Denial of Service by an attacker on the local link, slowly, but steadily eating up all memory. Actually, this commit became obsolete with "bridge: disable snooping if there is no querier" (b00589af3b) which included fixes for a few more cases. Therefore reverting the following commits (the commit stated in the commit message plus three of its follow up fixes): ==================== Revert "bridge: update mdb expiration timer upon reports." This reverts commit f144febd93d5ee534fdf23505ab091b2b9088edc. Revert "bridge: do not call setup_timer() multiple times" This reverts commit 1faabf2aab1fdaa1ace4e8c829d1b9cf7bfec2f1. Revert "bridge: fix some kernel warning in multicast timer" This reverts commit c7e8e8a8f7a70b343ca1e0f90a31e35ab2d16de1. Revert "bridge: only expire the mdb entry when query is received" This reverts commit 9f00b2e7cf241fa389733d41b615efdaa2cb0f5b. ==================== CC: Cong Wang <amwang@redhat.com> Signed-off-by: Linus Lüssing <linus.luessing@web.de> Reviewed-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-10-20 02:58:57 +04:00
unsigned long now = jiffies;
if (!br_multicast_ctx_should_use(brmctx, pmctx))
goto out;
mp = br_multicast_new_group(brmctx->br, group);
if (IS_ERR(mp))
return ERR_CAST(mp);
if (!pmctx) {
br_multicast_host_join(mp, true);
goto out;
}
for (pp = &mp->ports;
(p = mlock_dereference(*pp, brmctx->br)) != NULL;
pp = &p->next) {
if (br_port_group_equal(p, pmctx->port, src))
Revert "bridge: only expire the mdb entry when query is received" While this commit was a good attempt to fix issues occuring when no multicast querier is present, this commit still has two more issues: 1) There are cases where mdb entries do not expire even if there is a querier present. The bridge will unnecessarily continue flooding multicast packets on the according ports. 2) Never removing an mdb entry could be exploited for a Denial of Service by an attacker on the local link, slowly, but steadily eating up all memory. Actually, this commit became obsolete with "bridge: disable snooping if there is no querier" (b00589af3b) which included fixes for a few more cases. Therefore reverting the following commits (the commit stated in the commit message plus three of its follow up fixes): ==================== Revert "bridge: update mdb expiration timer upon reports." This reverts commit f144febd93d5ee534fdf23505ab091b2b9088edc. Revert "bridge: do not call setup_timer() multiple times" This reverts commit 1faabf2aab1fdaa1ace4e8c829d1b9cf7bfec2f1. Revert "bridge: fix some kernel warning in multicast timer" This reverts commit c7e8e8a8f7a70b343ca1e0f90a31e35ab2d16de1. Revert "bridge: only expire the mdb entry when query is received" This reverts commit 9f00b2e7cf241fa389733d41b615efdaa2cb0f5b. ==================== CC: Cong Wang <amwang@redhat.com> Signed-off-by: Linus Lüssing <linus.luessing@web.de> Reviewed-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-10-20 02:58:57 +04:00
goto found;
if ((unsigned long)p->key.port < (unsigned long)pmctx->port)
break;
}
p = br_multicast_new_port_group(pmctx->port, group, *pp, 0, src,
filter_mode, RTPROT_KERNEL);
if (unlikely(!p)) {
p = ERR_PTR(-ENOMEM);
goto out;
}
rcu_assign_pointer(*pp, p);
if (blocked)
p->flags |= MDB_PG_FLAGS_BLOCKED;
br_mdb_notify(brmctx->br->dev, mp, p, RTM_NEWMDB);
Revert "bridge: only expire the mdb entry when query is received" While this commit was a good attempt to fix issues occuring when no multicast querier is present, this commit still has two more issues: 1) There are cases where mdb entries do not expire even if there is a querier present. The bridge will unnecessarily continue flooding multicast packets on the according ports. 2) Never removing an mdb entry could be exploited for a Denial of Service by an attacker on the local link, slowly, but steadily eating up all memory. Actually, this commit became obsolete with "bridge: disable snooping if there is no querier" (b00589af3b) which included fixes for a few more cases. Therefore reverting the following commits (the commit stated in the commit message plus three of its follow up fixes): ==================== Revert "bridge: update mdb expiration timer upon reports." This reverts commit f144febd93d5ee534fdf23505ab091b2b9088edc. Revert "bridge: do not call setup_timer() multiple times" This reverts commit 1faabf2aab1fdaa1ace4e8c829d1b9cf7bfec2f1. Revert "bridge: fix some kernel warning in multicast timer" This reverts commit c7e8e8a8f7a70b343ca1e0f90a31e35ab2d16de1. Revert "bridge: only expire the mdb entry when query is received" This reverts commit 9f00b2e7cf241fa389733d41b615efdaa2cb0f5b. ==================== CC: Cong Wang <amwang@redhat.com> Signed-off-by: Linus Lüssing <linus.luessing@web.de> Reviewed-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-10-20 02:58:57 +04:00
found:
if (igmpv2_mldv1)
mod_timer(&p->timer,
now + brmctx->multicast_membership_interval);
out:
return p;
}
static int br_multicast_add_group(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct br_ip *group,
const unsigned char *src,
u8 filter_mode,
bool igmpv2_mldv1)
{
struct net_bridge_port_group *pg;
int err;
spin_lock(&brmctx->br->multicast_lock);
pg = __br_multicast_add_group(brmctx, pmctx, group, src, filter_mode,
igmpv2_mldv1, false);
/* NULL is considered valid for host joined groups */
err = PTR_ERR_OR_ZERO(pg);
spin_unlock(&brmctx->br->multicast_lock);
return err;
}
static int br_ip4_multicast_add_group(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
__be32 group,
__u16 vid,
const unsigned char *src,
bool igmpv2)
{
struct br_ip br_group;
u8 filter_mode;
if (ipv4_is_local_multicast(group))
return 0;
memset(&br_group, 0, sizeof(br_group));
br_group.dst.ip4 = group;
br_group.proto = htons(ETH_P_IP);
br_group.vid = vid;
filter_mode = igmpv2 ? MCAST_EXCLUDE : MCAST_INCLUDE;
return br_multicast_add_group(brmctx, pmctx, &br_group, src,
filter_mode, igmpv2);
}
#if IS_ENABLED(CONFIG_IPV6)
static int br_ip6_multicast_add_group(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
const struct in6_addr *group,
__u16 vid,
const unsigned char *src,
bool mldv1)
{
struct br_ip br_group;
u8 filter_mode;
if (ipv6_addr_is_ll_all_nodes(group))
return 0;
memset(&br_group, 0, sizeof(br_group));
br_group.dst.ip6 = *group;
br_group.proto = htons(ETH_P_IPV6);
br_group.vid = vid;
filter_mode = mldv1 ? MCAST_EXCLUDE : MCAST_INCLUDE;
return br_multicast_add_group(brmctx, pmctx, &br_group, src,
filter_mode, mldv1);
}
#endif
static bool br_multicast_rport_del(struct hlist_node *rlist)
{
if (hlist_unhashed(rlist))
return false;
hlist_del_init_rcu(rlist);
return true;
}
static bool br_ip4_multicast_rport_del(struct net_bridge_mcast_port *pmctx)
{
return br_multicast_rport_del(&pmctx->ip4_rlist);
}
static bool br_ip6_multicast_rport_del(struct net_bridge_mcast_port *pmctx)
{
#if IS_ENABLED(CONFIG_IPV6)
return br_multicast_rport_del(&pmctx->ip6_rlist);
#else
return false;
#endif
}
static void br_multicast_router_expired(struct net_bridge_mcast_port *pmctx,
struct timer_list *t,
struct hlist_node *rlist)
{
struct net_bridge *br = pmctx->port->br;
bool del;
spin_lock(&br->multicast_lock);
if (pmctx->multicast_router == MDB_RTR_TYPE_DISABLED ||
pmctx->multicast_router == MDB_RTR_TYPE_PERM ||
timer_pending(t))
goto out;
del = br_multicast_rport_del(rlist);
br_multicast_rport_del_notify(pmctx, del);
out:
spin_unlock(&br->multicast_lock);
}
static void br_ip4_multicast_router_expired(struct timer_list *t)
{
struct net_bridge_mcast_port *pmctx = from_timer(pmctx, t,
ip4_mc_router_timer);
br_multicast_router_expired(pmctx, t, &pmctx->ip4_rlist);
}
#if IS_ENABLED(CONFIG_IPV6)
static void br_ip6_multicast_router_expired(struct timer_list *t)
{
struct net_bridge_mcast_port *pmctx = from_timer(pmctx, t,
ip6_mc_router_timer);
br_multicast_router_expired(pmctx, t, &pmctx->ip6_rlist);
}
#endif
static void br_mc_router_state_change(struct net_bridge *p,
bool is_mc_router)
{
struct switchdev_attr attr = {
.orig_dev = p->dev,
.id = SWITCHDEV_ATTR_ID_BRIDGE_MROUTER,
.flags = SWITCHDEV_F_DEFER,
.u.mrouter = is_mc_router,
};
switchdev_port_attr_set(p->dev, &attr, NULL);
}
static void br_multicast_local_router_expired(struct net_bridge_mcast *brmctx,
struct timer_list *timer)
{
spin_lock(&brmctx->br->multicast_lock);
if (brmctx->multicast_router == MDB_RTR_TYPE_DISABLED ||
brmctx->multicast_router == MDB_RTR_TYPE_PERM ||
br_ip4_multicast_is_router(brmctx) ||
br_ip6_multicast_is_router(brmctx))
goto out;
br_mc_router_state_change(brmctx->br, false);
out:
spin_unlock(&brmctx->br->multicast_lock);
}
static void br_ip4_multicast_local_router_expired(struct timer_list *t)
{
struct net_bridge_mcast *brmctx = from_timer(brmctx, t,
ip4_mc_router_timer);
br_multicast_local_router_expired(brmctx, t);
}
#if IS_ENABLED(CONFIG_IPV6)
static void br_ip6_multicast_local_router_expired(struct timer_list *t)
{
struct net_bridge_mcast *brmctx = from_timer(brmctx, t,
ip6_mc_router_timer);
br_multicast_local_router_expired(brmctx, t);
}
#endif
static void br_multicast_querier_expired(struct net_bridge_mcast *brmctx,
struct bridge_mcast_own_query *query)
{
spin_lock(&brmctx->br->multicast_lock);
if (!netif_running(brmctx->br->dev) ||
br_multicast_ctx_vlan_global_disabled(brmctx) ||
!br_opt_get(brmctx->br, BROPT_MULTICAST_ENABLED))
goto out;
br_multicast_start_querier(brmctx, query);
out:
spin_unlock(&brmctx->br->multicast_lock);
}
static void br_ip4_multicast_querier_expired(struct timer_list *t)
{
struct net_bridge_mcast *brmctx = from_timer(brmctx, t,
ip4_other_query.timer);
br_multicast_querier_expired(brmctx, &brmctx->ip4_own_query);
}
#if IS_ENABLED(CONFIG_IPV6)
static void br_ip6_multicast_querier_expired(struct timer_list *t)
{
struct net_bridge_mcast *brmctx = from_timer(brmctx, t,
ip6_other_query.timer);
br_multicast_querier_expired(brmctx, &brmctx->ip6_own_query);
}
#endif
static void br_multicast_select_own_querier(struct net_bridge_mcast *brmctx,
struct br_ip *ip,
struct sk_buff *skb)
{
if (ip->proto == htons(ETH_P_IP))
brmctx->ip4_querier.addr.src.ip4 = ip_hdr(skb)->saddr;
#if IS_ENABLED(CONFIG_IPV6)
else
brmctx->ip6_querier.addr.src.ip6 = ipv6_hdr(skb)->saddr;
#endif
}
static void __br_multicast_send_query(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct net_bridge_port_group *pg,
struct br_ip *ip_dst,
struct br_ip *group,
bool with_srcs,
u8 sflag,
bool *need_rexmit)
{
bool over_lmqt = !!sflag;
struct sk_buff *skb;
u8 igmp_type;
again_under_lmqt:
skb = br_multicast_alloc_query(brmctx, pg, ip_dst, group, with_srcs,
over_lmqt, sflag, &igmp_type,
need_rexmit);
if (!skb)
return;
if (pmctx) {
skb->dev = pmctx->port->dev;
br_multicast_count(brmctx->br, pmctx->port, skb, igmp_type,
BR_MCAST_DIR_TX);
2015-09-16 04:04:16 +03:00
NF_HOOK(NFPROTO_BRIDGE, NF_BR_LOCAL_OUT,
dev_net(pmctx->port->dev), NULL, skb, NULL, skb->dev,
br_dev_queue_push_xmit);
if (over_lmqt && with_srcs && sflag) {
over_lmqt = false;
goto again_under_lmqt;
}
} else {
br_multicast_select_own_querier(brmctx, group, skb);
br_multicast_count(brmctx->br, NULL, skb, igmp_type,
BR_MCAST_DIR_RX);
netif_rx(skb);
}
}
static void br_multicast_send_query(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct bridge_mcast_own_query *own_query)
{
struct bridge_mcast_other_query *other_query = NULL;
struct br_ip br_group;
unsigned long time;
if (!br_multicast_ctx_should_use(brmctx, pmctx) ||
!br_opt_get(brmctx->br, BROPT_MULTICAST_ENABLED) ||
!br_opt_get(brmctx->br, BROPT_MULTICAST_QUERIER))
return;
memset(&br_group.dst, 0, sizeof(br_group.dst));
if (pmctx ? (own_query == &pmctx->ip4_own_query) :
(own_query == &brmctx->ip4_own_query)) {
other_query = &brmctx->ip4_other_query;
br_group.proto = htons(ETH_P_IP);
#if IS_ENABLED(CONFIG_IPV6)
} else {
other_query = &brmctx->ip6_other_query;
br_group.proto = htons(ETH_P_IPV6);
#endif
}
if (!other_query || timer_pending(&other_query->timer))
return;
__br_multicast_send_query(brmctx, pmctx, NULL, NULL, &br_group, false,
0, NULL);
time = jiffies;
time += own_query->startup_sent < brmctx->multicast_startup_query_count ?
brmctx->multicast_startup_query_interval :
brmctx->multicast_query_interval;
mod_timer(&own_query->timer, time);
}
static void
br_multicast_port_query_expired(struct net_bridge_mcast_port *pmctx,
struct bridge_mcast_own_query *query)
{
struct net_bridge *br = pmctx->port->br;
struct net_bridge_mcast *brmctx;
spin_lock(&br->multicast_lock);
if (br_multicast_port_ctx_state_stopped(pmctx))
goto out;
brmctx = br_multicast_port_ctx_get_global(pmctx);
if (query->startup_sent < brmctx->multicast_startup_query_count)
query->startup_sent++;
br_multicast_send_query(brmctx, pmctx, query);
out:
spin_unlock(&br->multicast_lock);
}
static void br_ip4_multicast_port_query_expired(struct timer_list *t)
{
struct net_bridge_mcast_port *pmctx = from_timer(pmctx, t,
ip4_own_query.timer);
br_multicast_port_query_expired(pmctx, &pmctx->ip4_own_query);
}
#if IS_ENABLED(CONFIG_IPV6)
static void br_ip6_multicast_port_query_expired(struct timer_list *t)
{
struct net_bridge_mcast_port *pmctx = from_timer(pmctx, t,
ip6_own_query.timer);
br_multicast_port_query_expired(pmctx, &pmctx->ip6_own_query);
}
#endif
static void br_multicast_port_group_rexmit(struct timer_list *t)
{
struct net_bridge_port_group *pg = from_timer(pg, t, rexmit_timer);
struct bridge_mcast_other_query *other_query = NULL;
struct net_bridge *br = pg->key.port->br;
struct net_bridge_mcast_port *pmctx;
struct net_bridge_mcast *brmctx;
bool need_rexmit = false;
spin_lock(&br->multicast_lock);
if (!netif_running(br->dev) || hlist_unhashed(&pg->mglist) ||
!br_opt_get(br, BROPT_MULTICAST_ENABLED) ||
!br_opt_get(br, BROPT_MULTICAST_QUERIER))
goto out;
pmctx = br_multicast_pg_to_port_ctx(pg);
if (!pmctx)
goto out;
brmctx = br_multicast_port_ctx_get_global(pmctx);
if (pg->key.addr.proto == htons(ETH_P_IP))
other_query = &brmctx->ip4_other_query;
#if IS_ENABLED(CONFIG_IPV6)
else
other_query = &brmctx->ip6_other_query;
#endif
if (!other_query || timer_pending(&other_query->timer))
goto out;
if (pg->grp_query_rexmit_cnt) {
pg->grp_query_rexmit_cnt--;
__br_multicast_send_query(brmctx, pmctx, pg, &pg->key.addr,
&pg->key.addr, false, 1, NULL);
}
__br_multicast_send_query(brmctx, pmctx, pg, &pg->key.addr,
&pg->key.addr, true, 0, &need_rexmit);
if (pg->grp_query_rexmit_cnt || need_rexmit)
mod_timer(&pg->rexmit_timer, jiffies +
brmctx->multicast_last_member_interval);
out:
spin_unlock(&br->multicast_lock);
}
static int br_mc_disabled_update(struct net_device *dev, bool value,
struct netlink_ext_ack *extack)
{
struct switchdev_attr attr = {
.orig_dev = dev,
.id = SWITCHDEV_ATTR_ID_BRIDGE_MC_DISABLED,
.flags = SWITCHDEV_F_DEFER,
.u.mc_disabled = !value,
};
return switchdev_port_attr_set(dev, &attr, extack);
}
void br_multicast_port_ctx_init(struct net_bridge_port *port,
struct net_bridge_vlan *vlan,
struct net_bridge_mcast_port *pmctx)
{
pmctx->port = port;
pmctx->vlan = vlan;
pmctx->multicast_router = MDB_RTR_TYPE_TEMP_QUERY;
timer_setup(&pmctx->ip4_mc_router_timer,
br_ip4_multicast_router_expired, 0);
timer_setup(&pmctx->ip4_own_query.timer,
br_ip4_multicast_port_query_expired, 0);
#if IS_ENABLED(CONFIG_IPV6)
timer_setup(&pmctx->ip6_mc_router_timer,
br_ip6_multicast_router_expired, 0);
timer_setup(&pmctx->ip6_own_query.timer,
br_ip6_multicast_port_query_expired, 0);
#endif
}
void br_multicast_port_ctx_deinit(struct net_bridge_mcast_port *pmctx)
{
#if IS_ENABLED(CONFIG_IPV6)
del_timer_sync(&pmctx->ip6_mc_router_timer);
#endif
del_timer_sync(&pmctx->ip4_mc_router_timer);
}
int br_multicast_add_port(struct net_bridge_port *port)
{
int err;
port->multicast_eht_hosts_limit = BR_MCAST_DEFAULT_EHT_HOSTS_LIMIT;
br_multicast_port_ctx_init(port, NULL, &port->multicast_ctx);
err = br_mc_disabled_update(port->dev,
br_opt_get(port->br,
BROPT_MULTICAST_ENABLED),
NULL);
if (err && err != -EOPNOTSUPP)
return err;
port->mcast_stats = netdev_alloc_pcpu_stats(struct bridge_mcast_stats);
if (!port->mcast_stats)
return -ENOMEM;
return 0;
}
void br_multicast_del_port(struct net_bridge_port *port)
{
struct net_bridge *br = port->br;
struct net_bridge_port_group *pg;
HLIST_HEAD(deleted_head);
struct hlist_node *n;
/* Take care of the remaining groups, only perm ones should be left */
spin_lock_bh(&br->multicast_lock);
hlist_for_each_entry_safe(pg, n, &port->mglist, mglist)
br_multicast_find_del_pg(br, pg);
hlist_move_list(&br->mcast_gc_list, &deleted_head);
spin_unlock_bh(&br->multicast_lock);
br_multicast_gc(&deleted_head);
br_multicast_port_ctx_deinit(&port->multicast_ctx);
free_percpu(port->mcast_stats);
}
static void br_multicast_enable(struct bridge_mcast_own_query *query)
{
query->startup_sent = 0;
if (try_to_del_timer_sync(&query->timer) >= 0 ||
del_timer(&query->timer))
mod_timer(&query->timer, jiffies);
}
static void __br_multicast_enable_port_ctx(struct net_bridge_mcast_port *pmctx)
{
struct net_bridge *br = pmctx->port->br;
struct net_bridge_mcast *brmctx;
brmctx = br_multicast_port_ctx_get_global(pmctx);
if (!br_opt_get(br, BROPT_MULTICAST_ENABLED) ||
!netif_running(br->dev))
bridge: multicast: restore perm router ports on multicast enable Satish reported a problem with the perm multicast router ports not getting reenabled after some series of events, in particular if it happens that the multicast snooping has been disabled and the port goes to disabled state then it will be deleted from the router port list, but if it moves into non-disabled state it will not be re-added because the mcast snooping is still disabled, and enabling snooping later does nothing. Here are the steps to reproduce, setup br0 with snooping enabled and eth1 added as a perm router (multicast_router = 2): 1. $ echo 0 > /sys/class/net/br0/bridge/multicast_snooping 2. $ ip l set eth1 down ^ This step deletes the interface from the router list 3. $ ip l set eth1 up ^ This step does not add it again because mcast snooping is disabled 4. $ echo 1 > /sys/class/net/br0/bridge/multicast_snooping 5. $ bridge -d -s mdb show <empty> At this point we have mcast enabled and eth1 as a perm router (value = 2) but it is not in the router list which is incorrect. After this change: 1. $ echo 0 > /sys/class/net/br0/bridge/multicast_snooping 2. $ ip l set eth1 down ^ This step deletes the interface from the router list 3. $ ip l set eth1 up ^ This step does not add it again because mcast snooping is disabled 4. $ echo 1 > /sys/class/net/br0/bridge/multicast_snooping 5. $ bridge -d -s mdb show router ports on br0: eth1 Note: we can directly do br_multicast_enable_port for all because the querier timer already has checks for the port state and will simply expire if it's in blocking/disabled. See the comment added by commit 9aa66382163e7 ("bridge: multicast: add a comment to br_port_state_selection about blocking state") Fixes: 561f1103a2b7 ("bridge: Add multicast_snooping sysfs toggle") Reported-by: Satish Ashok <sashok@cumulusnetworks.com> Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-10-18 19:09:48 +03:00
return;
br_multicast_enable(&pmctx->ip4_own_query);
#if IS_ENABLED(CONFIG_IPV6)
br_multicast_enable(&pmctx->ip6_own_query);
#endif
if (pmctx->multicast_router == MDB_RTR_TYPE_PERM) {
br_ip4_multicast_add_router(brmctx, pmctx);
br_ip6_multicast_add_router(brmctx, pmctx);
}
bridge: multicast: restore perm router ports on multicast enable Satish reported a problem with the perm multicast router ports not getting reenabled after some series of events, in particular if it happens that the multicast snooping has been disabled and the port goes to disabled state then it will be deleted from the router port list, but if it moves into non-disabled state it will not be re-added because the mcast snooping is still disabled, and enabling snooping later does nothing. Here are the steps to reproduce, setup br0 with snooping enabled and eth1 added as a perm router (multicast_router = 2): 1. $ echo 0 > /sys/class/net/br0/bridge/multicast_snooping 2. $ ip l set eth1 down ^ This step deletes the interface from the router list 3. $ ip l set eth1 up ^ This step does not add it again because mcast snooping is disabled 4. $ echo 1 > /sys/class/net/br0/bridge/multicast_snooping 5. $ bridge -d -s mdb show <empty> At this point we have mcast enabled and eth1 as a perm router (value = 2) but it is not in the router list which is incorrect. After this change: 1. $ echo 0 > /sys/class/net/br0/bridge/multicast_snooping 2. $ ip l set eth1 down ^ This step deletes the interface from the router list 3. $ ip l set eth1 up ^ This step does not add it again because mcast snooping is disabled 4. $ echo 1 > /sys/class/net/br0/bridge/multicast_snooping 5. $ bridge -d -s mdb show router ports on br0: eth1 Note: we can directly do br_multicast_enable_port for all because the querier timer already has checks for the port state and will simply expire if it's in blocking/disabled. See the comment added by commit 9aa66382163e7 ("bridge: multicast: add a comment to br_port_state_selection about blocking state") Fixes: 561f1103a2b7 ("bridge: Add multicast_snooping sysfs toggle") Reported-by: Satish Ashok <sashok@cumulusnetworks.com> Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-10-18 19:09:48 +03:00
}
bridge: multicast: restore perm router ports on multicast enable Satish reported a problem with the perm multicast router ports not getting reenabled after some series of events, in particular if it happens that the multicast snooping has been disabled and the port goes to disabled state then it will be deleted from the router port list, but if it moves into non-disabled state it will not be re-added because the mcast snooping is still disabled, and enabling snooping later does nothing. Here are the steps to reproduce, setup br0 with snooping enabled and eth1 added as a perm router (multicast_router = 2): 1. $ echo 0 > /sys/class/net/br0/bridge/multicast_snooping 2. $ ip l set eth1 down ^ This step deletes the interface from the router list 3. $ ip l set eth1 up ^ This step does not add it again because mcast snooping is disabled 4. $ echo 1 > /sys/class/net/br0/bridge/multicast_snooping 5. $ bridge -d -s mdb show <empty> At this point we have mcast enabled and eth1 as a perm router (value = 2) but it is not in the router list which is incorrect. After this change: 1. $ echo 0 > /sys/class/net/br0/bridge/multicast_snooping 2. $ ip l set eth1 down ^ This step deletes the interface from the router list 3. $ ip l set eth1 up ^ This step does not add it again because mcast snooping is disabled 4. $ echo 1 > /sys/class/net/br0/bridge/multicast_snooping 5. $ bridge -d -s mdb show router ports on br0: eth1 Note: we can directly do br_multicast_enable_port for all because the querier timer already has checks for the port state and will simply expire if it's in blocking/disabled. See the comment added by commit 9aa66382163e7 ("bridge: multicast: add a comment to br_port_state_selection about blocking state") Fixes: 561f1103a2b7 ("bridge: Add multicast_snooping sysfs toggle") Reported-by: Satish Ashok <sashok@cumulusnetworks.com> Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-10-18 19:09:48 +03:00
void br_multicast_enable_port(struct net_bridge_port *port)
{
struct net_bridge *br = port->br;
spin_lock_bh(&br->multicast_lock);
__br_multicast_enable_port_ctx(&port->multicast_ctx);
spin_unlock_bh(&br->multicast_lock);
}
static void __br_multicast_disable_port_ctx(struct net_bridge_mcast_port *pmctx)
{
struct net_bridge_port_group *pg;
hlist: drop the node parameter from iterators I'm not sure why, but the hlist for each entry iterators were conceived list_for_each_entry(pos, head, member) The hlist ones were greedy and wanted an extra parameter: hlist_for_each_entry(tpos, pos, head, member) Why did they need an extra pos parameter? I'm not quite sure. Not only they don't really need it, it also prevents the iterator from looking exactly like the list iterator, which is unfortunate. Besides the semantic patch, there was some manual work required: - Fix up the actual hlist iterators in linux/list.h - Fix up the declaration of other iterators based on the hlist ones. - A very small amount of places were using the 'node' parameter, this was modified to use 'obj->member' instead. - Coccinelle didn't handle the hlist_for_each_entry_safe iterator properly, so those had to be fixed up manually. The semantic patch which is mostly the work of Peter Senna Tschudin is here: @@ iterator name hlist_for_each_entry, hlist_for_each_entry_continue, hlist_for_each_entry_from, hlist_for_each_entry_rcu, hlist_for_each_entry_rcu_bh, hlist_for_each_entry_continue_rcu_bh, for_each_busy_worker, ax25_uid_for_each, ax25_for_each, inet_bind_bucket_for_each, sctp_for_each_hentry, sk_for_each, sk_for_each_rcu, sk_for_each_from, sk_for_each_safe, sk_for_each_bound, hlist_for_each_entry_safe, hlist_for_each_entry_continue_rcu, nr_neigh_for_each, nr_neigh_for_each_safe, nr_node_for_each, nr_node_for_each_safe, for_each_gfn_indirect_valid_sp, for_each_gfn_sp, for_each_host; type T; expression a,c,d,e; identifier b; statement S; @@ -T b; <+... when != b ( hlist_for_each_entry(a, - b, c, d) S | hlist_for_each_entry_continue(a, - b, c) S | hlist_for_each_entry_from(a, - b, c) S | hlist_for_each_entry_rcu(a, - b, c, d) S | hlist_for_each_entry_rcu_bh(a, - b, c, d) S | hlist_for_each_entry_continue_rcu_bh(a, - b, c) S | for_each_busy_worker(a, c, - b, d) S | ax25_uid_for_each(a, - b, c) S | ax25_for_each(a, - b, c) S | inet_bind_bucket_for_each(a, - b, c) S | sctp_for_each_hentry(a, - b, c) S | sk_for_each(a, - b, c) S | sk_for_each_rcu(a, - b, c) S | sk_for_each_from -(a, b) +(a) S + sk_for_each_from(a) S | sk_for_each_safe(a, - b, c, d) S | sk_for_each_bound(a, - b, c) S | hlist_for_each_entry_safe(a, - b, c, d, e) S | hlist_for_each_entry_continue_rcu(a, - b, c) S | nr_neigh_for_each(a, - b, c) S | nr_neigh_for_each_safe(a, - b, c, d) S | nr_node_for_each(a, - b, c) S | nr_node_for_each_safe(a, - b, c, d) S | - for_each_gfn_sp(a, c, d, b) S + for_each_gfn_sp(a, c, d) S | - for_each_gfn_indirect_valid_sp(a, c, d, b) S + for_each_gfn_indirect_valid_sp(a, c, d) S | for_each_host(a, - b, c) S | for_each_host_safe(a, - b, c, d) S | for_each_mesh_entry(a, - b, c, d) S ) ...+> [akpm@linux-foundation.org: drop bogus change from net/ipv4/raw.c] [akpm@linux-foundation.org: drop bogus hunk from net/ipv6/raw.c] [akpm@linux-foundation.org: checkpatch fixes] [akpm@linux-foundation.org: fix warnings] [akpm@linux-foudnation.org: redo intrusive kvm changes] Tested-by: Peter Senna Tschudin <peter.senna@gmail.com> Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Signed-off-by: Sasha Levin <sasha.levin@oracle.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Gleb Natapov <gleb@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-28 05:06:00 +04:00
struct hlist_node *n;
bool del = false;
hlist_for_each_entry_safe(pg, n, &pmctx->port->mglist, mglist)
if (!(pg->flags & MDB_PG_FLAGS_PERMANENT) &&
(!br_multicast_port_ctx_is_vlan(pmctx) ||
pg->key.addr.vid == pmctx->vlan->vid))
br_multicast_find_del_pg(pmctx->port->br, pg);
del |= br_ip4_multicast_rport_del(pmctx);
del_timer(&pmctx->ip4_mc_router_timer);
del_timer(&pmctx->ip4_own_query.timer);
del |= br_ip6_multicast_rport_del(pmctx);
#if IS_ENABLED(CONFIG_IPV6)
del_timer(&pmctx->ip6_mc_router_timer);
del_timer(&pmctx->ip6_own_query.timer);
#endif
br_multicast_rport_del_notify(pmctx, del);
}
void br_multicast_disable_port(struct net_bridge_port *port)
{
spin_lock_bh(&port->br->multicast_lock);
__br_multicast_disable_port_ctx(&port->multicast_ctx);
spin_unlock_bh(&port->br->multicast_lock);
}
static int __grp_src_delete_marked(struct net_bridge_port_group *pg)
{
struct net_bridge_group_src *ent;
struct hlist_node *tmp;
int deleted = 0;
hlist_for_each_entry_safe(ent, tmp, &pg->src_list, node)
if (ent->flags & BR_SGRP_F_DELETE) {
br_multicast_del_group_src(ent, false);
deleted++;
}
return deleted;
}
static void __grp_src_mod_timer(struct net_bridge_group_src *src,
unsigned long expires)
{
mod_timer(&src->timer, expires);
br_multicast_fwd_src_handle(src);
}
static void __grp_src_query_marked_and_rexmit(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct net_bridge_port_group *pg)
{
struct bridge_mcast_other_query *other_query = NULL;
u32 lmqc = brmctx->multicast_last_member_count;
unsigned long lmqt, lmi, now = jiffies;
struct net_bridge_group_src *ent;
if (!netif_running(brmctx->br->dev) ||
!br_opt_get(brmctx->br, BROPT_MULTICAST_ENABLED))
return;
if (pg->key.addr.proto == htons(ETH_P_IP))
other_query = &brmctx->ip4_other_query;
#if IS_ENABLED(CONFIG_IPV6)
else
other_query = &brmctx->ip6_other_query;
#endif
lmqt = now + br_multicast_lmqt(brmctx);
hlist_for_each_entry(ent, &pg->src_list, node) {
if (ent->flags & BR_SGRP_F_SEND) {
ent->flags &= ~BR_SGRP_F_SEND;
if (ent->timer.expires > lmqt) {
if (br_opt_get(brmctx->br,
BROPT_MULTICAST_QUERIER) &&
other_query &&
!timer_pending(&other_query->timer))
ent->src_query_rexmit_cnt = lmqc;
__grp_src_mod_timer(ent, lmqt);
}
}
}
if (!br_opt_get(brmctx->br, BROPT_MULTICAST_QUERIER) ||
!other_query || timer_pending(&other_query->timer))
return;
__br_multicast_send_query(brmctx, pmctx, pg, &pg->key.addr,
&pg->key.addr, true, 1, NULL);
lmi = now + brmctx->multicast_last_member_interval;
if (!timer_pending(&pg->rexmit_timer) ||
time_after(pg->rexmit_timer.expires, lmi))
mod_timer(&pg->rexmit_timer, lmi);
}
static void __grp_send_query_and_rexmit(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct net_bridge_port_group *pg)
{
struct bridge_mcast_other_query *other_query = NULL;
unsigned long now = jiffies, lmi;
if (!netif_running(brmctx->br->dev) ||
!br_opt_get(brmctx->br, BROPT_MULTICAST_ENABLED))
return;
if (pg->key.addr.proto == htons(ETH_P_IP))
other_query = &brmctx->ip4_other_query;
#if IS_ENABLED(CONFIG_IPV6)
else
other_query = &brmctx->ip6_other_query;
#endif
if (br_opt_get(brmctx->br, BROPT_MULTICAST_QUERIER) &&
other_query && !timer_pending(&other_query->timer)) {
lmi = now + brmctx->multicast_last_member_interval;
pg->grp_query_rexmit_cnt = brmctx->multicast_last_member_count - 1;
__br_multicast_send_query(brmctx, pmctx, pg, &pg->key.addr,
&pg->key.addr, false, 0, NULL);
if (!timer_pending(&pg->rexmit_timer) ||
time_after(pg->rexmit_timer.expires, lmi))
mod_timer(&pg->rexmit_timer, lmi);
}
if (pg->filter_mode == MCAST_EXCLUDE &&
(!timer_pending(&pg->timer) ||
time_after(pg->timer.expires, now + br_multicast_lmqt(brmctx))))
mod_timer(&pg->timer, now + br_multicast_lmqt(brmctx));
}
/* State Msg type New state Actions
* INCLUDE (A) IS_IN (B) INCLUDE (A+B) (B)=GMI
* INCLUDE (A) ALLOW (B) INCLUDE (A+B) (B)=GMI
* EXCLUDE (X,Y) ALLOW (A) EXCLUDE (X+A,Y-A) (A)=GMI
*/
static bool br_multicast_isinc_allow(const struct net_bridge_mcast *brmctx,
struct net_bridge_port_group *pg, void *h_addr,
void *srcs, u32 nsrcs, size_t addr_size,
int grec_type)
{
struct net_bridge_group_src *ent;
unsigned long now = jiffies;
bool changed = false;
struct br_ip src_ip;
u32 src_idx;
memset(&src_ip, 0, sizeof(src_ip));
src_ip.proto = pg->key.addr.proto;
for (src_idx = 0; src_idx < nsrcs; src_idx++) {
memcpy(&src_ip.src, srcs + (src_idx * addr_size), addr_size);
ent = br_multicast_find_group_src(pg, &src_ip);
if (!ent) {
ent = br_multicast_new_group_src(pg, &src_ip);
if (ent)
changed = true;
}
if (ent)
__grp_src_mod_timer(ent, now + br_multicast_gmi(brmctx));
}
if (br_multicast_eht_handle(brmctx, pg, h_addr, srcs, nsrcs, addr_size,
grec_type))
changed = true;
return changed;
}
/* State Msg type New state Actions
* INCLUDE (A) IS_EX (B) EXCLUDE (A*B,B-A) (B-A)=0
* Delete (A-B)
* Group Timer=GMI
*/
static void __grp_src_isexc_incl(const struct net_bridge_mcast *brmctx,
struct net_bridge_port_group *pg, void *h_addr,
void *srcs, u32 nsrcs, size_t addr_size,
int grec_type)
{
struct net_bridge_group_src *ent;
struct br_ip src_ip;
u32 src_idx;
hlist_for_each_entry(ent, &pg->src_list, node)
ent->flags |= BR_SGRP_F_DELETE;
memset(&src_ip, 0, sizeof(src_ip));
src_ip.proto = pg->key.addr.proto;
for (src_idx = 0; src_idx < nsrcs; src_idx++) {
memcpy(&src_ip.src, srcs + (src_idx * addr_size), addr_size);
ent = br_multicast_find_group_src(pg, &src_ip);
if (ent)
ent->flags &= ~BR_SGRP_F_DELETE;
else
ent = br_multicast_new_group_src(pg, &src_ip);
if (ent)
br_multicast_fwd_src_handle(ent);
}
br_multicast_eht_handle(brmctx, pg, h_addr, srcs, nsrcs, addr_size,
grec_type);
__grp_src_delete_marked(pg);
}
/* State Msg type New state Actions
* EXCLUDE (X,Y) IS_EX (A) EXCLUDE (A-Y,Y*A) (A-X-Y)=GMI
* Delete (X-A)
* Delete (Y-A)
* Group Timer=GMI
*/
static bool __grp_src_isexc_excl(const struct net_bridge_mcast *brmctx,
struct net_bridge_port_group *pg, void *h_addr,
void *srcs, u32 nsrcs, size_t addr_size,
int grec_type)
{
struct net_bridge_group_src *ent;
unsigned long now = jiffies;
bool changed = false;
struct br_ip src_ip;
u32 src_idx;
hlist_for_each_entry(ent, &pg->src_list, node)
ent->flags |= BR_SGRP_F_DELETE;
memset(&src_ip, 0, sizeof(src_ip));
src_ip.proto = pg->key.addr.proto;
for (src_idx = 0; src_idx < nsrcs; src_idx++) {
memcpy(&src_ip.src, srcs + (src_idx * addr_size), addr_size);
ent = br_multicast_find_group_src(pg, &src_ip);
if (ent) {
ent->flags &= ~BR_SGRP_F_DELETE;
} else {
ent = br_multicast_new_group_src(pg, &src_ip);
if (ent) {
__grp_src_mod_timer(ent,
now + br_multicast_gmi(brmctx));
changed = true;
}
}
}
if (br_multicast_eht_handle(brmctx, pg, h_addr, srcs, nsrcs, addr_size,
grec_type))
changed = true;
if (__grp_src_delete_marked(pg))
changed = true;
return changed;
}
static bool br_multicast_isexc(const struct net_bridge_mcast *brmctx,
struct net_bridge_port_group *pg, void *h_addr,
void *srcs, u32 nsrcs, size_t addr_size,
int grec_type)
{
bool changed = false;
switch (pg->filter_mode) {
case MCAST_INCLUDE:
__grp_src_isexc_incl(brmctx, pg, h_addr, srcs, nsrcs, addr_size,
grec_type);
br_multicast_star_g_handle_mode(pg, MCAST_EXCLUDE);
changed = true;
break;
case MCAST_EXCLUDE:
changed = __grp_src_isexc_excl(brmctx, pg, h_addr, srcs, nsrcs,
addr_size, grec_type);
break;
}
pg->filter_mode = MCAST_EXCLUDE;
mod_timer(&pg->timer, jiffies + br_multicast_gmi(brmctx));
return changed;
}
/* State Msg type New state Actions
* INCLUDE (A) TO_IN (B) INCLUDE (A+B) (B)=GMI
* Send Q(G,A-B)
*/
static bool __grp_src_toin_incl(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct net_bridge_port_group *pg, void *h_addr,
void *srcs, u32 nsrcs, size_t addr_size,
int grec_type)
{
u32 src_idx, to_send = pg->src_ents;
struct net_bridge_group_src *ent;
unsigned long now = jiffies;
bool changed = false;
struct br_ip src_ip;
hlist_for_each_entry(ent, &pg->src_list, node)
ent->flags |= BR_SGRP_F_SEND;
memset(&src_ip, 0, sizeof(src_ip));
src_ip.proto = pg->key.addr.proto;
for (src_idx = 0; src_idx < nsrcs; src_idx++) {
memcpy(&src_ip.src, srcs + (src_idx * addr_size), addr_size);
ent = br_multicast_find_group_src(pg, &src_ip);
if (ent) {
ent->flags &= ~BR_SGRP_F_SEND;
to_send--;
} else {
ent = br_multicast_new_group_src(pg, &src_ip);
if (ent)
changed = true;
}
if (ent)
__grp_src_mod_timer(ent, now + br_multicast_gmi(brmctx));
}
if (br_multicast_eht_handle(brmctx, pg, h_addr, srcs, nsrcs, addr_size,
grec_type))
changed = true;
if (to_send)
__grp_src_query_marked_and_rexmit(brmctx, pmctx, pg);
return changed;
}
/* State Msg type New state Actions
* EXCLUDE (X,Y) TO_IN (A) EXCLUDE (X+A,Y-A) (A)=GMI
* Send Q(G,X-A)
* Send Q(G)
*/
static bool __grp_src_toin_excl(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct net_bridge_port_group *pg, void *h_addr,
void *srcs, u32 nsrcs, size_t addr_size,
int grec_type)
{
u32 src_idx, to_send = pg->src_ents;
struct net_bridge_group_src *ent;
unsigned long now = jiffies;
bool changed = false;
struct br_ip src_ip;
hlist_for_each_entry(ent, &pg->src_list, node)
if (timer_pending(&ent->timer))
ent->flags |= BR_SGRP_F_SEND;
memset(&src_ip, 0, sizeof(src_ip));
src_ip.proto = pg->key.addr.proto;
for (src_idx = 0; src_idx < nsrcs; src_idx++) {
memcpy(&src_ip.src, srcs + (src_idx * addr_size), addr_size);
ent = br_multicast_find_group_src(pg, &src_ip);
if (ent) {
if (timer_pending(&ent->timer)) {
ent->flags &= ~BR_SGRP_F_SEND;
to_send--;
}
} else {
ent = br_multicast_new_group_src(pg, &src_ip);
if (ent)
changed = true;
}
if (ent)
__grp_src_mod_timer(ent, now + br_multicast_gmi(brmctx));
}
if (br_multicast_eht_handle(brmctx, pg, h_addr, srcs, nsrcs, addr_size,
grec_type))
changed = true;
if (to_send)
__grp_src_query_marked_and_rexmit(brmctx, pmctx, pg);
__grp_send_query_and_rexmit(brmctx, pmctx, pg);
return changed;
}
static bool br_multicast_toin(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct net_bridge_port_group *pg, void *h_addr,
void *srcs, u32 nsrcs, size_t addr_size,
int grec_type)
{
bool changed = false;
switch (pg->filter_mode) {
case MCAST_INCLUDE:
changed = __grp_src_toin_incl(brmctx, pmctx, pg, h_addr, srcs,
nsrcs, addr_size, grec_type);
break;
case MCAST_EXCLUDE:
changed = __grp_src_toin_excl(brmctx, pmctx, pg, h_addr, srcs,
nsrcs, addr_size, grec_type);
break;
}
if (br_multicast_eht_should_del_pg(pg)) {
pg->flags |= MDB_PG_FLAGS_FAST_LEAVE;
br_multicast_find_del_pg(pg->key.port->br, pg);
/* a notification has already been sent and we shouldn't
* access pg after the delete so we have to return false
*/
changed = false;
}
return changed;
}
/* State Msg type New state Actions
* INCLUDE (A) TO_EX (B) EXCLUDE (A*B,B-A) (B-A)=0
* Delete (A-B)
* Send Q(G,A*B)
* Group Timer=GMI
*/
static void __grp_src_toex_incl(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct net_bridge_port_group *pg, void *h_addr,
void *srcs, u32 nsrcs, size_t addr_size,
int grec_type)
{
struct net_bridge_group_src *ent;
u32 src_idx, to_send = 0;
struct br_ip src_ip;
hlist_for_each_entry(ent, &pg->src_list, node)
ent->flags = (ent->flags & ~BR_SGRP_F_SEND) | BR_SGRP_F_DELETE;
memset(&src_ip, 0, sizeof(src_ip));
src_ip.proto = pg->key.addr.proto;
for (src_idx = 0; src_idx < nsrcs; src_idx++) {
memcpy(&src_ip.src, srcs + (src_idx * addr_size), addr_size);
ent = br_multicast_find_group_src(pg, &src_ip);
if (ent) {
ent->flags = (ent->flags & ~BR_SGRP_F_DELETE) |
BR_SGRP_F_SEND;
to_send++;
} else {
ent = br_multicast_new_group_src(pg, &src_ip);
}
if (ent)
br_multicast_fwd_src_handle(ent);
}
br_multicast_eht_handle(brmctx, pg, h_addr, srcs, nsrcs, addr_size,
grec_type);
__grp_src_delete_marked(pg);
if (to_send)
__grp_src_query_marked_and_rexmit(brmctx, pmctx, pg);
}
/* State Msg type New state Actions
* EXCLUDE (X,Y) TO_EX (A) EXCLUDE (A-Y,Y*A) (A-X-Y)=Group Timer
* Delete (X-A)
* Delete (Y-A)
* Send Q(G,A-Y)
* Group Timer=GMI
*/
static bool __grp_src_toex_excl(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct net_bridge_port_group *pg, void *h_addr,
void *srcs, u32 nsrcs, size_t addr_size,
int grec_type)
{
struct net_bridge_group_src *ent;
u32 src_idx, to_send = 0;
bool changed = false;
struct br_ip src_ip;
hlist_for_each_entry(ent, &pg->src_list, node)
ent->flags = (ent->flags & ~BR_SGRP_F_SEND) | BR_SGRP_F_DELETE;
memset(&src_ip, 0, sizeof(src_ip));
src_ip.proto = pg->key.addr.proto;
for (src_idx = 0; src_idx < nsrcs; src_idx++) {
memcpy(&src_ip.src, srcs + (src_idx * addr_size), addr_size);
ent = br_multicast_find_group_src(pg, &src_ip);
if (ent) {
ent->flags &= ~BR_SGRP_F_DELETE;
} else {
ent = br_multicast_new_group_src(pg, &src_ip);
if (ent) {
__grp_src_mod_timer(ent, pg->timer.expires);
changed = true;
}
}
if (ent && timer_pending(&ent->timer)) {
ent->flags |= BR_SGRP_F_SEND;
to_send++;
}
}
if (br_multicast_eht_handle(brmctx, pg, h_addr, srcs, nsrcs, addr_size,
grec_type))
changed = true;
if (__grp_src_delete_marked(pg))
changed = true;
if (to_send)
__grp_src_query_marked_and_rexmit(brmctx, pmctx, pg);
return changed;
}
static bool br_multicast_toex(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct net_bridge_port_group *pg, void *h_addr,
void *srcs, u32 nsrcs, size_t addr_size,
int grec_type)
{
bool changed = false;
switch (pg->filter_mode) {
case MCAST_INCLUDE:
__grp_src_toex_incl(brmctx, pmctx, pg, h_addr, srcs, nsrcs,
addr_size, grec_type);
br_multicast_star_g_handle_mode(pg, MCAST_EXCLUDE);
changed = true;
break;
case MCAST_EXCLUDE:
changed = __grp_src_toex_excl(brmctx, pmctx, pg, h_addr, srcs,
nsrcs, addr_size, grec_type);
break;
}
pg->filter_mode = MCAST_EXCLUDE;
mod_timer(&pg->timer, jiffies + br_multicast_gmi(brmctx));
return changed;
}
/* State Msg type New state Actions
* INCLUDE (A) BLOCK (B) INCLUDE (A) Send Q(G,A*B)
*/
static bool __grp_src_block_incl(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct net_bridge_port_group *pg, void *h_addr,
void *srcs, u32 nsrcs, size_t addr_size, int grec_type)
{
struct net_bridge_group_src *ent;
u32 src_idx, to_send = 0;
bool changed = false;
struct br_ip src_ip;
hlist_for_each_entry(ent, &pg->src_list, node)
ent->flags &= ~BR_SGRP_F_SEND;
memset(&src_ip, 0, sizeof(src_ip));
src_ip.proto = pg->key.addr.proto;
for (src_idx = 0; src_idx < nsrcs; src_idx++) {
memcpy(&src_ip.src, srcs + (src_idx * addr_size), addr_size);
ent = br_multicast_find_group_src(pg, &src_ip);
if (ent) {
ent->flags |= BR_SGRP_F_SEND;
to_send++;
}
}
if (br_multicast_eht_handle(brmctx, pg, h_addr, srcs, nsrcs, addr_size,
grec_type))
changed = true;
if (to_send)
__grp_src_query_marked_and_rexmit(brmctx, pmctx, pg);
return changed;
}
/* State Msg type New state Actions
* EXCLUDE (X,Y) BLOCK (A) EXCLUDE (X+(A-Y),Y) (A-X-Y)=Group Timer
* Send Q(G,A-Y)
*/
static bool __grp_src_block_excl(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct net_bridge_port_group *pg, void *h_addr,
void *srcs, u32 nsrcs, size_t addr_size, int grec_type)
{
struct net_bridge_group_src *ent;
u32 src_idx, to_send = 0;
bool changed = false;
struct br_ip src_ip;
hlist_for_each_entry(ent, &pg->src_list, node)
ent->flags &= ~BR_SGRP_F_SEND;
memset(&src_ip, 0, sizeof(src_ip));
src_ip.proto = pg->key.addr.proto;
for (src_idx = 0; src_idx < nsrcs; src_idx++) {
memcpy(&src_ip.src, srcs + (src_idx * addr_size), addr_size);
ent = br_multicast_find_group_src(pg, &src_ip);
if (!ent) {
ent = br_multicast_new_group_src(pg, &src_ip);
if (ent) {
__grp_src_mod_timer(ent, pg->timer.expires);
changed = true;
}
}
if (ent && timer_pending(&ent->timer)) {
ent->flags |= BR_SGRP_F_SEND;
to_send++;
}
}
if (br_multicast_eht_handle(brmctx, pg, h_addr, srcs, nsrcs, addr_size,
grec_type))
changed = true;
if (to_send)
__grp_src_query_marked_and_rexmit(brmctx, pmctx, pg);
return changed;
}
static bool br_multicast_block(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct net_bridge_port_group *pg, void *h_addr,
void *srcs, u32 nsrcs, size_t addr_size, int grec_type)
{
bool changed = false;
switch (pg->filter_mode) {
case MCAST_INCLUDE:
changed = __grp_src_block_incl(brmctx, pmctx, pg, h_addr, srcs,
nsrcs, addr_size, grec_type);
break;
case MCAST_EXCLUDE:
changed = __grp_src_block_excl(brmctx, pmctx, pg, h_addr, srcs,
nsrcs, addr_size, grec_type);
break;
}
if ((pg->filter_mode == MCAST_INCLUDE && hlist_empty(&pg->src_list)) ||
br_multicast_eht_should_del_pg(pg)) {
if (br_multicast_eht_should_del_pg(pg))
pg->flags |= MDB_PG_FLAGS_FAST_LEAVE;
br_multicast_find_del_pg(pg->key.port->br, pg);
/* a notification has already been sent and we shouldn't
* access pg after the delete so we have to return false
*/
changed = false;
}
return changed;
}
static struct net_bridge_port_group *
br_multicast_find_port(struct net_bridge_mdb_entry *mp,
struct net_bridge_port *p,
const unsigned char *src)
{
struct net_bridge *br __maybe_unused = mp->br;
struct net_bridge_port_group *pg;
for (pg = mlock_dereference(mp->ports, br);
pg;
pg = mlock_dereference(pg->next, br))
if (br_port_group_equal(pg, p, src))
return pg;
return NULL;
}
static int br_ip4_multicast_igmp3_report(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct sk_buff *skb,
u16 vid)
{
bool igmpv2 = brmctx->multicast_igmp_version == 2;
struct net_bridge_mdb_entry *mdst;
struct net_bridge_port_group *pg;
const unsigned char *src;
struct igmpv3_report *ih;
struct igmpv3_grec *grec;
int i, len, num, type;
__be32 group, *h_addr;
bool changed = false;
int err = 0;
net: bridge: mcast: fix stale nsrcs pointer in igmp3/mld2 report handling We take a pointer to grec prior to calling pskb_may_pull and use it afterwards to get nsrcs so record nsrcs before the pull when handling igmp3 and we get a pointer to nsrcs and call pskb_may_pull when handling mld2 which again could lead to reading 2 bytes out-of-bounds. ================================================================== BUG: KASAN: use-after-free in br_multicast_rcv+0x480c/0x4ad0 [bridge] Read of size 2 at addr ffff8880421302b4 by task ksoftirqd/1/16 CPU: 1 PID: 16 Comm: ksoftirqd/1 Tainted: G OE 5.2.0-rc6+ #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1 04/01/2014 Call Trace: dump_stack+0x71/0xab print_address_description+0x6a/0x280 ? br_multicast_rcv+0x480c/0x4ad0 [bridge] __kasan_report+0x152/0x1aa ? br_multicast_rcv+0x480c/0x4ad0 [bridge] ? br_multicast_rcv+0x480c/0x4ad0 [bridge] kasan_report+0xe/0x20 br_multicast_rcv+0x480c/0x4ad0 [bridge] ? br_multicast_disable_port+0x150/0x150 [bridge] ? ktime_get_with_offset+0xb4/0x150 ? __kasan_kmalloc.constprop.6+0xa6/0xf0 ? __netif_receive_skb+0x1b0/0x1b0 ? br_fdb_update+0x10e/0x6e0 [bridge] ? br_handle_frame_finish+0x3c6/0x11d0 [bridge] br_handle_frame_finish+0x3c6/0x11d0 [bridge] ? br_pass_frame_up+0x3a0/0x3a0 [bridge] ? virtnet_probe+0x1c80/0x1c80 [virtio_net] br_handle_frame+0x731/0xd90 [bridge] ? select_idle_sibling+0x25/0x7d0 ? br_handle_frame_finish+0x11d0/0x11d0 [bridge] __netif_receive_skb_core+0xced/0x2d70 ? virtqueue_get_buf_ctx+0x230/0x1130 [virtio_ring] ? do_xdp_generic+0x20/0x20 ? virtqueue_napi_complete+0x39/0x70 [virtio_net] ? virtnet_poll+0x94d/0xc78 [virtio_net] ? receive_buf+0x5120/0x5120 [virtio_net] ? __netif_receive_skb_one_core+0x97/0x1d0 __netif_receive_skb_one_core+0x97/0x1d0 ? __netif_receive_skb_core+0x2d70/0x2d70 ? _raw_write_trylock+0x100/0x100 ? __queue_work+0x41e/0xbe0 process_backlog+0x19c/0x650 ? _raw_read_lock_irq+0x40/0x40 net_rx_action+0x71e/0xbc0 ? __switch_to_asm+0x40/0x70 ? napi_complete_done+0x360/0x360 ? __switch_to_asm+0x34/0x70 ? __switch_to_asm+0x40/0x70 ? __schedule+0x85e/0x14d0 __do_softirq+0x1db/0x5f9 ? takeover_tasklets+0x5f0/0x5f0 run_ksoftirqd+0x26/0x40 smpboot_thread_fn+0x443/0x680 ? sort_range+0x20/0x20 ? schedule+0x94/0x210 ? __kthread_parkme+0x78/0xf0 ? sort_range+0x20/0x20 kthread+0x2ae/0x3a0 ? kthread_create_worker_on_cpu+0xc0/0xc0 ret_from_fork+0x35/0x40 The buggy address belongs to the page: page:ffffea0001084c00 refcount:0 mapcount:-128 mapping:0000000000000000 index:0x0 flags: 0xffffc000000000() raw: 00ffffc000000000 ffffea0000cfca08 ffffea0001098608 0000000000000000 raw: 0000000000000000 0000000000000003 00000000ffffff7f 0000000000000000 page dumped because: kasan: bad access detected Memory state around the buggy address: ffff888042130180: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ffff888042130200: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff > ffff888042130280: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ^ ffff888042130300: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ffff888042130380: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ================================================================== Disabling lock debugging due to kernel taint Fixes: bc8c20acaea1 ("bridge: multicast: treat igmpv3 report with INCLUDE and no sources as a leave") Reported-by: Martin Weinelt <martin@linuxlounge.net> Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com> Tested-by: Martin Weinelt <martin@linuxlounge.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-07-02 15:00:18 +03:00
u16 nsrcs;
ih = igmpv3_report_hdr(skb);
num = ntohs(ih->ngrec);
len = skb_transport_offset(skb) + sizeof(*ih);
for (i = 0; i < num; i++) {
len += sizeof(*grec);
if (!ip_mc_may_pull(skb, len))
return -EINVAL;
grec = (void *)(skb->data + len - sizeof(*grec));
group = grec->grec_mca;
type = grec->grec_type;
net: bridge: mcast: fix stale nsrcs pointer in igmp3/mld2 report handling We take a pointer to grec prior to calling pskb_may_pull and use it afterwards to get nsrcs so record nsrcs before the pull when handling igmp3 and we get a pointer to nsrcs and call pskb_may_pull when handling mld2 which again could lead to reading 2 bytes out-of-bounds. ================================================================== BUG: KASAN: use-after-free in br_multicast_rcv+0x480c/0x4ad0 [bridge] Read of size 2 at addr ffff8880421302b4 by task ksoftirqd/1/16 CPU: 1 PID: 16 Comm: ksoftirqd/1 Tainted: G OE 5.2.0-rc6+ #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1 04/01/2014 Call Trace: dump_stack+0x71/0xab print_address_description+0x6a/0x280 ? br_multicast_rcv+0x480c/0x4ad0 [bridge] __kasan_report+0x152/0x1aa ? br_multicast_rcv+0x480c/0x4ad0 [bridge] ? br_multicast_rcv+0x480c/0x4ad0 [bridge] kasan_report+0xe/0x20 br_multicast_rcv+0x480c/0x4ad0 [bridge] ? br_multicast_disable_port+0x150/0x150 [bridge] ? ktime_get_with_offset+0xb4/0x150 ? __kasan_kmalloc.constprop.6+0xa6/0xf0 ? __netif_receive_skb+0x1b0/0x1b0 ? br_fdb_update+0x10e/0x6e0 [bridge] ? br_handle_frame_finish+0x3c6/0x11d0 [bridge] br_handle_frame_finish+0x3c6/0x11d0 [bridge] ? br_pass_frame_up+0x3a0/0x3a0 [bridge] ? virtnet_probe+0x1c80/0x1c80 [virtio_net] br_handle_frame+0x731/0xd90 [bridge] ? select_idle_sibling+0x25/0x7d0 ? br_handle_frame_finish+0x11d0/0x11d0 [bridge] __netif_receive_skb_core+0xced/0x2d70 ? virtqueue_get_buf_ctx+0x230/0x1130 [virtio_ring] ? do_xdp_generic+0x20/0x20 ? virtqueue_napi_complete+0x39/0x70 [virtio_net] ? virtnet_poll+0x94d/0xc78 [virtio_net] ? receive_buf+0x5120/0x5120 [virtio_net] ? __netif_receive_skb_one_core+0x97/0x1d0 __netif_receive_skb_one_core+0x97/0x1d0 ? __netif_receive_skb_core+0x2d70/0x2d70 ? _raw_write_trylock+0x100/0x100 ? __queue_work+0x41e/0xbe0 process_backlog+0x19c/0x650 ? _raw_read_lock_irq+0x40/0x40 net_rx_action+0x71e/0xbc0 ? __switch_to_asm+0x40/0x70 ? napi_complete_done+0x360/0x360 ? __switch_to_asm+0x34/0x70 ? __switch_to_asm+0x40/0x70 ? __schedule+0x85e/0x14d0 __do_softirq+0x1db/0x5f9 ? takeover_tasklets+0x5f0/0x5f0 run_ksoftirqd+0x26/0x40 smpboot_thread_fn+0x443/0x680 ? sort_range+0x20/0x20 ? schedule+0x94/0x210 ? __kthread_parkme+0x78/0xf0 ? sort_range+0x20/0x20 kthread+0x2ae/0x3a0 ? kthread_create_worker_on_cpu+0xc0/0xc0 ret_from_fork+0x35/0x40 The buggy address belongs to the page: page:ffffea0001084c00 refcount:0 mapcount:-128 mapping:0000000000000000 index:0x0 flags: 0xffffc000000000() raw: 00ffffc000000000 ffffea0000cfca08 ffffea0001098608 0000000000000000 raw: 0000000000000000 0000000000000003 00000000ffffff7f 0000000000000000 page dumped because: kasan: bad access detected Memory state around the buggy address: ffff888042130180: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ffff888042130200: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff > ffff888042130280: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ^ ffff888042130300: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ffff888042130380: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ================================================================== Disabling lock debugging due to kernel taint Fixes: bc8c20acaea1 ("bridge: multicast: treat igmpv3 report with INCLUDE and no sources as a leave") Reported-by: Martin Weinelt <martin@linuxlounge.net> Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com> Tested-by: Martin Weinelt <martin@linuxlounge.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-07-02 15:00:18 +03:00
nsrcs = ntohs(grec->grec_nsrcs);
net: bridge: mcast: fix stale nsrcs pointer in igmp3/mld2 report handling We take a pointer to grec prior to calling pskb_may_pull and use it afterwards to get nsrcs so record nsrcs before the pull when handling igmp3 and we get a pointer to nsrcs and call pskb_may_pull when handling mld2 which again could lead to reading 2 bytes out-of-bounds. ================================================================== BUG: KASAN: use-after-free in br_multicast_rcv+0x480c/0x4ad0 [bridge] Read of size 2 at addr ffff8880421302b4 by task ksoftirqd/1/16 CPU: 1 PID: 16 Comm: ksoftirqd/1 Tainted: G OE 5.2.0-rc6+ #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1 04/01/2014 Call Trace: dump_stack+0x71/0xab print_address_description+0x6a/0x280 ? br_multicast_rcv+0x480c/0x4ad0 [bridge] __kasan_report+0x152/0x1aa ? br_multicast_rcv+0x480c/0x4ad0 [bridge] ? br_multicast_rcv+0x480c/0x4ad0 [bridge] kasan_report+0xe/0x20 br_multicast_rcv+0x480c/0x4ad0 [bridge] ? br_multicast_disable_port+0x150/0x150 [bridge] ? ktime_get_with_offset+0xb4/0x150 ? __kasan_kmalloc.constprop.6+0xa6/0xf0 ? __netif_receive_skb+0x1b0/0x1b0 ? br_fdb_update+0x10e/0x6e0 [bridge] ? br_handle_frame_finish+0x3c6/0x11d0 [bridge] br_handle_frame_finish+0x3c6/0x11d0 [bridge] ? br_pass_frame_up+0x3a0/0x3a0 [bridge] ? virtnet_probe+0x1c80/0x1c80 [virtio_net] br_handle_frame+0x731/0xd90 [bridge] ? select_idle_sibling+0x25/0x7d0 ? br_handle_frame_finish+0x11d0/0x11d0 [bridge] __netif_receive_skb_core+0xced/0x2d70 ? virtqueue_get_buf_ctx+0x230/0x1130 [virtio_ring] ? do_xdp_generic+0x20/0x20 ? virtqueue_napi_complete+0x39/0x70 [virtio_net] ? virtnet_poll+0x94d/0xc78 [virtio_net] ? receive_buf+0x5120/0x5120 [virtio_net] ? __netif_receive_skb_one_core+0x97/0x1d0 __netif_receive_skb_one_core+0x97/0x1d0 ? __netif_receive_skb_core+0x2d70/0x2d70 ? _raw_write_trylock+0x100/0x100 ? __queue_work+0x41e/0xbe0 process_backlog+0x19c/0x650 ? _raw_read_lock_irq+0x40/0x40 net_rx_action+0x71e/0xbc0 ? __switch_to_asm+0x40/0x70 ? napi_complete_done+0x360/0x360 ? __switch_to_asm+0x34/0x70 ? __switch_to_asm+0x40/0x70 ? __schedule+0x85e/0x14d0 __do_softirq+0x1db/0x5f9 ? takeover_tasklets+0x5f0/0x5f0 run_ksoftirqd+0x26/0x40 smpboot_thread_fn+0x443/0x680 ? sort_range+0x20/0x20 ? schedule+0x94/0x210 ? __kthread_parkme+0x78/0xf0 ? sort_range+0x20/0x20 kthread+0x2ae/0x3a0 ? kthread_create_worker_on_cpu+0xc0/0xc0 ret_from_fork+0x35/0x40 The buggy address belongs to the page: page:ffffea0001084c00 refcount:0 mapcount:-128 mapping:0000000000000000 index:0x0 flags: 0xffffc000000000() raw: 00ffffc000000000 ffffea0000cfca08 ffffea0001098608 0000000000000000 raw: 0000000000000000 0000000000000003 00000000ffffff7f 0000000000000000 page dumped because: kasan: bad access detected Memory state around the buggy address: ffff888042130180: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ffff888042130200: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff > ffff888042130280: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ^ ffff888042130300: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ffff888042130380: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ================================================================== Disabling lock debugging due to kernel taint Fixes: bc8c20acaea1 ("bridge: multicast: treat igmpv3 report with INCLUDE and no sources as a leave") Reported-by: Martin Weinelt <martin@linuxlounge.net> Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com> Tested-by: Martin Weinelt <martin@linuxlounge.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-07-02 15:00:18 +03:00
len += nsrcs * 4;
if (!ip_mc_may_pull(skb, len))
return -EINVAL;
switch (type) {
case IGMPV3_MODE_IS_INCLUDE:
case IGMPV3_MODE_IS_EXCLUDE:
case IGMPV3_CHANGE_TO_INCLUDE:
case IGMPV3_CHANGE_TO_EXCLUDE:
case IGMPV3_ALLOW_NEW_SOURCES:
case IGMPV3_BLOCK_OLD_SOURCES:
break;
default:
continue;
}
src = eth_hdr(skb)->h_source;
if (nsrcs == 0 &&
(type == IGMPV3_CHANGE_TO_INCLUDE ||
type == IGMPV3_MODE_IS_INCLUDE)) {
if (!pmctx || igmpv2) {
br_ip4_multicast_leave_group(brmctx, pmctx,
group, vid, src);
continue;
}
} else {
err = br_ip4_multicast_add_group(brmctx, pmctx, group,
vid, src, igmpv2);
if (err)
break;
}
if (!pmctx || igmpv2)
continue;
spin_lock_bh(&brmctx->br->multicast_lock);
if (!br_multicast_ctx_should_use(brmctx, pmctx))
goto unlock_continue;
mdst = br_mdb_ip4_get(brmctx->br, group, vid);
if (!mdst)
goto unlock_continue;
pg = br_multicast_find_port(mdst, pmctx->port, src);
if (!pg || (pg->flags & MDB_PG_FLAGS_PERMANENT))
goto unlock_continue;
/* reload grec and host addr */
grec = (void *)(skb->data + len - sizeof(*grec) - (nsrcs * 4));
h_addr = &ip_hdr(skb)->saddr;
switch (type) {
case IGMPV3_ALLOW_NEW_SOURCES:
changed = br_multicast_isinc_allow(brmctx, pg, h_addr,
grec->grec_src,
nsrcs, sizeof(__be32), type);
break;
case IGMPV3_MODE_IS_INCLUDE:
changed = br_multicast_isinc_allow(brmctx, pg, h_addr,
grec->grec_src,
nsrcs, sizeof(__be32), type);
break;
case IGMPV3_MODE_IS_EXCLUDE:
changed = br_multicast_isexc(brmctx, pg, h_addr,
grec->grec_src,
nsrcs, sizeof(__be32), type);
break;
case IGMPV3_CHANGE_TO_INCLUDE:
changed = br_multicast_toin(brmctx, pmctx, pg, h_addr,
grec->grec_src,
nsrcs, sizeof(__be32), type);
break;
case IGMPV3_CHANGE_TO_EXCLUDE:
changed = br_multicast_toex(brmctx, pmctx, pg, h_addr,
grec->grec_src,
nsrcs, sizeof(__be32), type);
break;
case IGMPV3_BLOCK_OLD_SOURCES:
changed = br_multicast_block(brmctx, pmctx, pg, h_addr,
grec->grec_src,
nsrcs, sizeof(__be32), type);
break;
}
if (changed)
br_mdb_notify(brmctx->br->dev, mdst, pg, RTM_NEWMDB);
unlock_continue:
spin_unlock_bh(&brmctx->br->multicast_lock);
}
return err;
}
#if IS_ENABLED(CONFIG_IPV6)
static int br_ip6_multicast_mld2_report(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct sk_buff *skb,
u16 vid)
{
bool mldv1 = brmctx->multicast_mld_version == 1;
struct net_bridge_mdb_entry *mdst;
struct net_bridge_port_group *pg;
unsigned int nsrcs_offset;
const unsigned char *src;
struct icmp6hdr *icmp6h;
struct in6_addr *h_addr;
struct mld2_grec *grec;
unsigned int grec_len;
bool changed = false;
int i, len, num;
int err = 0;
if (!ipv6_mc_may_pull(skb, sizeof(*icmp6h)))
return -EINVAL;
icmp6h = icmp6_hdr(skb);
num = ntohs(icmp6h->icmp6_dataun.un_data16[1]);
len = skb_transport_offset(skb) + sizeof(*icmp6h);
for (i = 0; i < num; i++) {
net: bridge: mcast: fix stale nsrcs pointer in igmp3/mld2 report handling We take a pointer to grec prior to calling pskb_may_pull and use it afterwards to get nsrcs so record nsrcs before the pull when handling igmp3 and we get a pointer to nsrcs and call pskb_may_pull when handling mld2 which again could lead to reading 2 bytes out-of-bounds. ================================================================== BUG: KASAN: use-after-free in br_multicast_rcv+0x480c/0x4ad0 [bridge] Read of size 2 at addr ffff8880421302b4 by task ksoftirqd/1/16 CPU: 1 PID: 16 Comm: ksoftirqd/1 Tainted: G OE 5.2.0-rc6+ #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1 04/01/2014 Call Trace: dump_stack+0x71/0xab print_address_description+0x6a/0x280 ? br_multicast_rcv+0x480c/0x4ad0 [bridge] __kasan_report+0x152/0x1aa ? br_multicast_rcv+0x480c/0x4ad0 [bridge] ? br_multicast_rcv+0x480c/0x4ad0 [bridge] kasan_report+0xe/0x20 br_multicast_rcv+0x480c/0x4ad0 [bridge] ? br_multicast_disable_port+0x150/0x150 [bridge] ? ktime_get_with_offset+0xb4/0x150 ? __kasan_kmalloc.constprop.6+0xa6/0xf0 ? __netif_receive_skb+0x1b0/0x1b0 ? br_fdb_update+0x10e/0x6e0 [bridge] ? br_handle_frame_finish+0x3c6/0x11d0 [bridge] br_handle_frame_finish+0x3c6/0x11d0 [bridge] ? br_pass_frame_up+0x3a0/0x3a0 [bridge] ? virtnet_probe+0x1c80/0x1c80 [virtio_net] br_handle_frame+0x731/0xd90 [bridge] ? select_idle_sibling+0x25/0x7d0 ? br_handle_frame_finish+0x11d0/0x11d0 [bridge] __netif_receive_skb_core+0xced/0x2d70 ? virtqueue_get_buf_ctx+0x230/0x1130 [virtio_ring] ? do_xdp_generic+0x20/0x20 ? virtqueue_napi_complete+0x39/0x70 [virtio_net] ? virtnet_poll+0x94d/0xc78 [virtio_net] ? receive_buf+0x5120/0x5120 [virtio_net] ? __netif_receive_skb_one_core+0x97/0x1d0 __netif_receive_skb_one_core+0x97/0x1d0 ? __netif_receive_skb_core+0x2d70/0x2d70 ? _raw_write_trylock+0x100/0x100 ? __queue_work+0x41e/0xbe0 process_backlog+0x19c/0x650 ? _raw_read_lock_irq+0x40/0x40 net_rx_action+0x71e/0xbc0 ? __switch_to_asm+0x40/0x70 ? napi_complete_done+0x360/0x360 ? __switch_to_asm+0x34/0x70 ? __switch_to_asm+0x40/0x70 ? __schedule+0x85e/0x14d0 __do_softirq+0x1db/0x5f9 ? takeover_tasklets+0x5f0/0x5f0 run_ksoftirqd+0x26/0x40 smpboot_thread_fn+0x443/0x680 ? sort_range+0x20/0x20 ? schedule+0x94/0x210 ? __kthread_parkme+0x78/0xf0 ? sort_range+0x20/0x20 kthread+0x2ae/0x3a0 ? kthread_create_worker_on_cpu+0xc0/0xc0 ret_from_fork+0x35/0x40 The buggy address belongs to the page: page:ffffea0001084c00 refcount:0 mapcount:-128 mapping:0000000000000000 index:0x0 flags: 0xffffc000000000() raw: 00ffffc000000000 ffffea0000cfca08 ffffea0001098608 0000000000000000 raw: 0000000000000000 0000000000000003 00000000ffffff7f 0000000000000000 page dumped because: kasan: bad access detected Memory state around the buggy address: ffff888042130180: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ffff888042130200: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff > ffff888042130280: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ^ ffff888042130300: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ffff888042130380: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ================================================================== Disabling lock debugging due to kernel taint Fixes: bc8c20acaea1 ("bridge: multicast: treat igmpv3 report with INCLUDE and no sources as a leave") Reported-by: Martin Weinelt <martin@linuxlounge.net> Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com> Tested-by: Martin Weinelt <martin@linuxlounge.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-07-02 15:00:18 +03:00
__be16 *_nsrcs, __nsrcs;
u16 nsrcs;
nsrcs_offset = len + offsetof(struct mld2_grec, grec_nsrcs);
if (skb_transport_offset(skb) + ipv6_transport_len(skb) <
nsrcs_offset + sizeof(__nsrcs))
return -EINVAL;
net: bridge: mcast: fix stale nsrcs pointer in igmp3/mld2 report handling We take a pointer to grec prior to calling pskb_may_pull and use it afterwards to get nsrcs so record nsrcs before the pull when handling igmp3 and we get a pointer to nsrcs and call pskb_may_pull when handling mld2 which again could lead to reading 2 bytes out-of-bounds. ================================================================== BUG: KASAN: use-after-free in br_multicast_rcv+0x480c/0x4ad0 [bridge] Read of size 2 at addr ffff8880421302b4 by task ksoftirqd/1/16 CPU: 1 PID: 16 Comm: ksoftirqd/1 Tainted: G OE 5.2.0-rc6+ #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1 04/01/2014 Call Trace: dump_stack+0x71/0xab print_address_description+0x6a/0x280 ? br_multicast_rcv+0x480c/0x4ad0 [bridge] __kasan_report+0x152/0x1aa ? br_multicast_rcv+0x480c/0x4ad0 [bridge] ? br_multicast_rcv+0x480c/0x4ad0 [bridge] kasan_report+0xe/0x20 br_multicast_rcv+0x480c/0x4ad0 [bridge] ? br_multicast_disable_port+0x150/0x150 [bridge] ? ktime_get_with_offset+0xb4/0x150 ? __kasan_kmalloc.constprop.6+0xa6/0xf0 ? __netif_receive_skb+0x1b0/0x1b0 ? br_fdb_update+0x10e/0x6e0 [bridge] ? br_handle_frame_finish+0x3c6/0x11d0 [bridge] br_handle_frame_finish+0x3c6/0x11d0 [bridge] ? br_pass_frame_up+0x3a0/0x3a0 [bridge] ? virtnet_probe+0x1c80/0x1c80 [virtio_net] br_handle_frame+0x731/0xd90 [bridge] ? select_idle_sibling+0x25/0x7d0 ? br_handle_frame_finish+0x11d0/0x11d0 [bridge] __netif_receive_skb_core+0xced/0x2d70 ? virtqueue_get_buf_ctx+0x230/0x1130 [virtio_ring] ? do_xdp_generic+0x20/0x20 ? virtqueue_napi_complete+0x39/0x70 [virtio_net] ? virtnet_poll+0x94d/0xc78 [virtio_net] ? receive_buf+0x5120/0x5120 [virtio_net] ? __netif_receive_skb_one_core+0x97/0x1d0 __netif_receive_skb_one_core+0x97/0x1d0 ? __netif_receive_skb_core+0x2d70/0x2d70 ? _raw_write_trylock+0x100/0x100 ? __queue_work+0x41e/0xbe0 process_backlog+0x19c/0x650 ? _raw_read_lock_irq+0x40/0x40 net_rx_action+0x71e/0xbc0 ? __switch_to_asm+0x40/0x70 ? napi_complete_done+0x360/0x360 ? __switch_to_asm+0x34/0x70 ? __switch_to_asm+0x40/0x70 ? __schedule+0x85e/0x14d0 __do_softirq+0x1db/0x5f9 ? takeover_tasklets+0x5f0/0x5f0 run_ksoftirqd+0x26/0x40 smpboot_thread_fn+0x443/0x680 ? sort_range+0x20/0x20 ? schedule+0x94/0x210 ? __kthread_parkme+0x78/0xf0 ? sort_range+0x20/0x20 kthread+0x2ae/0x3a0 ? kthread_create_worker_on_cpu+0xc0/0xc0 ret_from_fork+0x35/0x40 The buggy address belongs to the page: page:ffffea0001084c00 refcount:0 mapcount:-128 mapping:0000000000000000 index:0x0 flags: 0xffffc000000000() raw: 00ffffc000000000 ffffea0000cfca08 ffffea0001098608 0000000000000000 raw: 0000000000000000 0000000000000003 00000000ffffff7f 0000000000000000 page dumped because: kasan: bad access detected Memory state around the buggy address: ffff888042130180: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ffff888042130200: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff > ffff888042130280: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ^ ffff888042130300: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ffff888042130380: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ================================================================== Disabling lock debugging due to kernel taint Fixes: bc8c20acaea1 ("bridge: multicast: treat igmpv3 report with INCLUDE and no sources as a leave") Reported-by: Martin Weinelt <martin@linuxlounge.net> Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com> Tested-by: Martin Weinelt <martin@linuxlounge.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-07-02 15:00:18 +03:00
_nsrcs = skb_header_pointer(skb, nsrcs_offset,
sizeof(__nsrcs), &__nsrcs);
if (!_nsrcs)
return -EINVAL;
net: bridge: mcast: fix stale nsrcs pointer in igmp3/mld2 report handling We take a pointer to grec prior to calling pskb_may_pull and use it afterwards to get nsrcs so record nsrcs before the pull when handling igmp3 and we get a pointer to nsrcs and call pskb_may_pull when handling mld2 which again could lead to reading 2 bytes out-of-bounds. ================================================================== BUG: KASAN: use-after-free in br_multicast_rcv+0x480c/0x4ad0 [bridge] Read of size 2 at addr ffff8880421302b4 by task ksoftirqd/1/16 CPU: 1 PID: 16 Comm: ksoftirqd/1 Tainted: G OE 5.2.0-rc6+ #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1 04/01/2014 Call Trace: dump_stack+0x71/0xab print_address_description+0x6a/0x280 ? br_multicast_rcv+0x480c/0x4ad0 [bridge] __kasan_report+0x152/0x1aa ? br_multicast_rcv+0x480c/0x4ad0 [bridge] ? br_multicast_rcv+0x480c/0x4ad0 [bridge] kasan_report+0xe/0x20 br_multicast_rcv+0x480c/0x4ad0 [bridge] ? br_multicast_disable_port+0x150/0x150 [bridge] ? ktime_get_with_offset+0xb4/0x150 ? __kasan_kmalloc.constprop.6+0xa6/0xf0 ? __netif_receive_skb+0x1b0/0x1b0 ? br_fdb_update+0x10e/0x6e0 [bridge] ? br_handle_frame_finish+0x3c6/0x11d0 [bridge] br_handle_frame_finish+0x3c6/0x11d0 [bridge] ? br_pass_frame_up+0x3a0/0x3a0 [bridge] ? virtnet_probe+0x1c80/0x1c80 [virtio_net] br_handle_frame+0x731/0xd90 [bridge] ? select_idle_sibling+0x25/0x7d0 ? br_handle_frame_finish+0x11d0/0x11d0 [bridge] __netif_receive_skb_core+0xced/0x2d70 ? virtqueue_get_buf_ctx+0x230/0x1130 [virtio_ring] ? do_xdp_generic+0x20/0x20 ? virtqueue_napi_complete+0x39/0x70 [virtio_net] ? virtnet_poll+0x94d/0xc78 [virtio_net] ? receive_buf+0x5120/0x5120 [virtio_net] ? __netif_receive_skb_one_core+0x97/0x1d0 __netif_receive_skb_one_core+0x97/0x1d0 ? __netif_receive_skb_core+0x2d70/0x2d70 ? _raw_write_trylock+0x100/0x100 ? __queue_work+0x41e/0xbe0 process_backlog+0x19c/0x650 ? _raw_read_lock_irq+0x40/0x40 net_rx_action+0x71e/0xbc0 ? __switch_to_asm+0x40/0x70 ? napi_complete_done+0x360/0x360 ? __switch_to_asm+0x34/0x70 ? __switch_to_asm+0x40/0x70 ? __schedule+0x85e/0x14d0 __do_softirq+0x1db/0x5f9 ? takeover_tasklets+0x5f0/0x5f0 run_ksoftirqd+0x26/0x40 smpboot_thread_fn+0x443/0x680 ? sort_range+0x20/0x20 ? schedule+0x94/0x210 ? __kthread_parkme+0x78/0xf0 ? sort_range+0x20/0x20 kthread+0x2ae/0x3a0 ? kthread_create_worker_on_cpu+0xc0/0xc0 ret_from_fork+0x35/0x40 The buggy address belongs to the page: page:ffffea0001084c00 refcount:0 mapcount:-128 mapping:0000000000000000 index:0x0 flags: 0xffffc000000000() raw: 00ffffc000000000 ffffea0000cfca08 ffffea0001098608 0000000000000000 raw: 0000000000000000 0000000000000003 00000000ffffff7f 0000000000000000 page dumped because: kasan: bad access detected Memory state around the buggy address: ffff888042130180: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ffff888042130200: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff > ffff888042130280: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ^ ffff888042130300: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ffff888042130380: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ================================================================== Disabling lock debugging due to kernel taint Fixes: bc8c20acaea1 ("bridge: multicast: treat igmpv3 report with INCLUDE and no sources as a leave") Reported-by: Martin Weinelt <martin@linuxlounge.net> Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com> Tested-by: Martin Weinelt <martin@linuxlounge.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-07-02 15:00:18 +03:00
nsrcs = ntohs(*_nsrcs);
grec_len = struct_size(grec, grec_src, nsrcs);
if (!ipv6_mc_may_pull(skb, len + grec_len))
return -EINVAL;
grec = (struct mld2_grec *)(skb->data + len);
len += grec_len;
switch (grec->grec_type) {
case MLD2_MODE_IS_INCLUDE:
case MLD2_MODE_IS_EXCLUDE:
case MLD2_CHANGE_TO_INCLUDE:
case MLD2_CHANGE_TO_EXCLUDE:
case MLD2_ALLOW_NEW_SOURCES:
case MLD2_BLOCK_OLD_SOURCES:
break;
default:
continue;
}
src = eth_hdr(skb)->h_source;
if ((grec->grec_type == MLD2_CHANGE_TO_INCLUDE ||
grec->grec_type == MLD2_MODE_IS_INCLUDE) &&
net: bridge: mcast: fix stale nsrcs pointer in igmp3/mld2 report handling We take a pointer to grec prior to calling pskb_may_pull and use it afterwards to get nsrcs so record nsrcs before the pull when handling igmp3 and we get a pointer to nsrcs and call pskb_may_pull when handling mld2 which again could lead to reading 2 bytes out-of-bounds. ================================================================== BUG: KASAN: use-after-free in br_multicast_rcv+0x480c/0x4ad0 [bridge] Read of size 2 at addr ffff8880421302b4 by task ksoftirqd/1/16 CPU: 1 PID: 16 Comm: ksoftirqd/1 Tainted: G OE 5.2.0-rc6+ #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1 04/01/2014 Call Trace: dump_stack+0x71/0xab print_address_description+0x6a/0x280 ? br_multicast_rcv+0x480c/0x4ad0 [bridge] __kasan_report+0x152/0x1aa ? br_multicast_rcv+0x480c/0x4ad0 [bridge] ? br_multicast_rcv+0x480c/0x4ad0 [bridge] kasan_report+0xe/0x20 br_multicast_rcv+0x480c/0x4ad0 [bridge] ? br_multicast_disable_port+0x150/0x150 [bridge] ? ktime_get_with_offset+0xb4/0x150 ? __kasan_kmalloc.constprop.6+0xa6/0xf0 ? __netif_receive_skb+0x1b0/0x1b0 ? br_fdb_update+0x10e/0x6e0 [bridge] ? br_handle_frame_finish+0x3c6/0x11d0 [bridge] br_handle_frame_finish+0x3c6/0x11d0 [bridge] ? br_pass_frame_up+0x3a0/0x3a0 [bridge] ? virtnet_probe+0x1c80/0x1c80 [virtio_net] br_handle_frame+0x731/0xd90 [bridge] ? select_idle_sibling+0x25/0x7d0 ? br_handle_frame_finish+0x11d0/0x11d0 [bridge] __netif_receive_skb_core+0xced/0x2d70 ? virtqueue_get_buf_ctx+0x230/0x1130 [virtio_ring] ? do_xdp_generic+0x20/0x20 ? virtqueue_napi_complete+0x39/0x70 [virtio_net] ? virtnet_poll+0x94d/0xc78 [virtio_net] ? receive_buf+0x5120/0x5120 [virtio_net] ? __netif_receive_skb_one_core+0x97/0x1d0 __netif_receive_skb_one_core+0x97/0x1d0 ? __netif_receive_skb_core+0x2d70/0x2d70 ? _raw_write_trylock+0x100/0x100 ? __queue_work+0x41e/0xbe0 process_backlog+0x19c/0x650 ? _raw_read_lock_irq+0x40/0x40 net_rx_action+0x71e/0xbc0 ? __switch_to_asm+0x40/0x70 ? napi_complete_done+0x360/0x360 ? __switch_to_asm+0x34/0x70 ? __switch_to_asm+0x40/0x70 ? __schedule+0x85e/0x14d0 __do_softirq+0x1db/0x5f9 ? takeover_tasklets+0x5f0/0x5f0 run_ksoftirqd+0x26/0x40 smpboot_thread_fn+0x443/0x680 ? sort_range+0x20/0x20 ? schedule+0x94/0x210 ? __kthread_parkme+0x78/0xf0 ? sort_range+0x20/0x20 kthread+0x2ae/0x3a0 ? kthread_create_worker_on_cpu+0xc0/0xc0 ret_from_fork+0x35/0x40 The buggy address belongs to the page: page:ffffea0001084c00 refcount:0 mapcount:-128 mapping:0000000000000000 index:0x0 flags: 0xffffc000000000() raw: 00ffffc000000000 ffffea0000cfca08 ffffea0001098608 0000000000000000 raw: 0000000000000000 0000000000000003 00000000ffffff7f 0000000000000000 page dumped because: kasan: bad access detected Memory state around the buggy address: ffff888042130180: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ffff888042130200: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff > ffff888042130280: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ^ ffff888042130300: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ffff888042130380: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ================================================================== Disabling lock debugging due to kernel taint Fixes: bc8c20acaea1 ("bridge: multicast: treat igmpv3 report with INCLUDE and no sources as a leave") Reported-by: Martin Weinelt <martin@linuxlounge.net> Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com> Tested-by: Martin Weinelt <martin@linuxlounge.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-07-02 15:00:18 +03:00
nsrcs == 0) {
if (!pmctx || mldv1) {
br_ip6_multicast_leave_group(brmctx, pmctx,
&grec->grec_mca,
vid, src);
continue;
}
} else {
err = br_ip6_multicast_add_group(brmctx, pmctx,
&grec->grec_mca, vid,
src, mldv1);
if (err)
break;
}
if (!pmctx || mldv1)
continue;
spin_lock_bh(&brmctx->br->multicast_lock);
if (!br_multicast_ctx_should_use(brmctx, pmctx))
goto unlock_continue;
mdst = br_mdb_ip6_get(brmctx->br, &grec->grec_mca, vid);
if (!mdst)
goto unlock_continue;
pg = br_multicast_find_port(mdst, pmctx->port, src);
if (!pg || (pg->flags & MDB_PG_FLAGS_PERMANENT))
goto unlock_continue;
h_addr = &ipv6_hdr(skb)->saddr;
switch (grec->grec_type) {
case MLD2_ALLOW_NEW_SOURCES:
changed = br_multicast_isinc_allow(brmctx, pg, h_addr,
grec->grec_src, nsrcs,
sizeof(struct in6_addr),
grec->grec_type);
break;
case MLD2_MODE_IS_INCLUDE:
changed = br_multicast_isinc_allow(brmctx, pg, h_addr,
grec->grec_src, nsrcs,
sizeof(struct in6_addr),
grec->grec_type);
break;
case MLD2_MODE_IS_EXCLUDE:
changed = br_multicast_isexc(brmctx, pg, h_addr,
grec->grec_src, nsrcs,
sizeof(struct in6_addr),
grec->grec_type);
break;
case MLD2_CHANGE_TO_INCLUDE:
changed = br_multicast_toin(brmctx, pmctx, pg, h_addr,
grec->grec_src, nsrcs,
sizeof(struct in6_addr),
grec->grec_type);
break;
case MLD2_CHANGE_TO_EXCLUDE:
changed = br_multicast_toex(brmctx, pmctx, pg, h_addr,
grec->grec_src, nsrcs,
sizeof(struct in6_addr),
grec->grec_type);
break;
case MLD2_BLOCK_OLD_SOURCES:
changed = br_multicast_block(brmctx, pmctx, pg, h_addr,
grec->grec_src, nsrcs,
sizeof(struct in6_addr),
grec->grec_type);
break;
}
if (changed)
br_mdb_notify(brmctx->br->dev, mdst, pg, RTM_NEWMDB);
unlock_continue:
spin_unlock_bh(&brmctx->br->multicast_lock);
}
return err;
}
#endif
static bool br_ip4_multicast_select_querier(struct net_bridge_mcast *brmctx,
struct net_bridge_port *port,
__be32 saddr)
{
if (!timer_pending(&brmctx->ip4_own_query.timer) &&
!timer_pending(&brmctx->ip4_other_query.timer))
goto update;
if (!brmctx->ip4_querier.addr.src.ip4)
goto update;
if (ntohl(saddr) <= ntohl(brmctx->ip4_querier.addr.src.ip4))
goto update;
return false;
update:
brmctx->ip4_querier.addr.src.ip4 = saddr;
/* update protected by general multicast_lock by caller */
rcu_assign_pointer(brmctx->ip4_querier.port, port);
return true;
}
#if IS_ENABLED(CONFIG_IPV6)
static bool br_ip6_multicast_select_querier(struct net_bridge_mcast *brmctx,
struct net_bridge_port *port,
struct in6_addr *saddr)
{
if (!timer_pending(&brmctx->ip6_own_query.timer) &&
!timer_pending(&brmctx->ip6_other_query.timer))
goto update;
if (ipv6_addr_cmp(saddr, &brmctx->ip6_querier.addr.src.ip6) <= 0)
goto update;
return false;
update:
brmctx->ip6_querier.addr.src.ip6 = *saddr;
/* update protected by general multicast_lock by caller */
rcu_assign_pointer(brmctx->ip6_querier.port, port);
return true;
}
#endif
static void
br_multicast_update_query_timer(struct net_bridge_mcast *brmctx,
struct bridge_mcast_other_query *query,
unsigned long max_delay)
{
if (!timer_pending(&query->timer))
query->delay_time = jiffies + max_delay;
mod_timer(&query->timer, jiffies + brmctx->multicast_querier_interval);
}
static void br_port_mc_router_state_change(struct net_bridge_port *p,
bool is_mc_router)
{
struct switchdev_attr attr = {
.orig_dev = p->dev,
.id = SWITCHDEV_ATTR_ID_PORT_MROUTER,
.flags = SWITCHDEV_F_DEFER,
.u.mrouter = is_mc_router,
};
switchdev_port_attr_set(p->dev, &attr, NULL);
}
static struct net_bridge_port *
br_multicast_rport_from_node(struct net_bridge_mcast *brmctx,
struct hlist_head *mc_router_list,
struct hlist_node *rlist)
{
struct net_bridge_mcast_port *pmctx;
#if IS_ENABLED(CONFIG_IPV6)
if (mc_router_list == &brmctx->ip6_mc_router_list)
pmctx = hlist_entry(rlist, struct net_bridge_mcast_port,
ip6_rlist);
else
#endif
pmctx = hlist_entry(rlist, struct net_bridge_mcast_port,
ip4_rlist);
return pmctx->port;
}
static struct hlist_node *
br_multicast_get_rport_slot(struct net_bridge_mcast *brmctx,
struct net_bridge_port *port,
struct hlist_head *mc_router_list)
{
struct hlist_node *slot = NULL;
struct net_bridge_port *p;
struct hlist_node *rlist;
hlist_for_each(rlist, mc_router_list) {
p = br_multicast_rport_from_node(brmctx, mc_router_list, rlist);
if ((unsigned long)port >= (unsigned long)p)
break;
slot = rlist;
}
return slot;
}
static bool br_multicast_no_router_otherpf(struct net_bridge_mcast_port *pmctx,
struct hlist_node *rnode)
{
#if IS_ENABLED(CONFIG_IPV6)
if (rnode != &pmctx->ip6_rlist)
return hlist_unhashed(&pmctx->ip6_rlist);
else
return hlist_unhashed(&pmctx->ip4_rlist);
#else
return true;
#endif
}
/* Add port to router_list
* list is maintained ordered by pointer value
* and locked by br->multicast_lock and RCU
*/
static void br_multicast_add_router(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct hlist_node *rlist,
struct hlist_head *mc_router_list)
{
struct hlist_node *slot;
if (!hlist_unhashed(rlist))
return;
slot = br_multicast_get_rport_slot(brmctx, pmctx->port, mc_router_list);
if (slot)
hlist_add_behind_rcu(rlist, slot);
else
hlist_add_head_rcu(rlist, mc_router_list);
/* For backwards compatibility for now, only notify if we
* switched from no IPv4/IPv6 multicast router to a new
* IPv4 or IPv6 multicast router.
*/
if (br_multicast_no_router_otherpf(pmctx, rlist)) {
br_rtr_notify(pmctx->port->br->dev, pmctx->port, RTM_NEWMDB);
br_port_mc_router_state_change(pmctx->port, true);
}
}
/* Add port to router_list
* list is maintained ordered by pointer value
* and locked by br->multicast_lock and RCU
*/
static void br_ip4_multicast_add_router(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx)
{
br_multicast_add_router(brmctx, pmctx, &pmctx->ip4_rlist,
&brmctx->ip4_mc_router_list);
}
/* Add port to router_list
* list is maintained ordered by pointer value
* and locked by br->multicast_lock and RCU
*/
static void br_ip6_multicast_add_router(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx)
{
#if IS_ENABLED(CONFIG_IPV6)
br_multicast_add_router(brmctx, pmctx, &pmctx->ip6_rlist,
&brmctx->ip6_mc_router_list);
#endif
}
static void br_multicast_mark_router(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct timer_list *timer,
struct hlist_node *rlist,
struct hlist_head *mc_router_list)
{
unsigned long now = jiffies;
if (!br_multicast_ctx_should_use(brmctx, pmctx))
return;
if (!pmctx) {
if (brmctx->multicast_router == MDB_RTR_TYPE_TEMP_QUERY) {
if (!br_ip4_multicast_is_router(brmctx) &&
!br_ip6_multicast_is_router(brmctx))
br_mc_router_state_change(brmctx->br, true);
mod_timer(timer, now + brmctx->multicast_querier_interval);
}
return;
}
if (pmctx->multicast_router == MDB_RTR_TYPE_DISABLED ||
pmctx->multicast_router == MDB_RTR_TYPE_PERM)
return;
br_multicast_add_router(brmctx, pmctx, rlist, mc_router_list);
mod_timer(timer, now + brmctx->multicast_querier_interval);
}
static void br_ip4_multicast_mark_router(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx)
{
struct timer_list *timer = &brmctx->ip4_mc_router_timer;
struct hlist_node *rlist = NULL;
if (pmctx) {
timer = &pmctx->ip4_mc_router_timer;
rlist = &pmctx->ip4_rlist;
}
br_multicast_mark_router(brmctx, pmctx, timer, rlist,
&brmctx->ip4_mc_router_list);
}
static void br_ip6_multicast_mark_router(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx)
{
#if IS_ENABLED(CONFIG_IPV6)
struct timer_list *timer = &brmctx->ip6_mc_router_timer;
struct hlist_node *rlist = NULL;
if (pmctx) {
timer = &pmctx->ip6_mc_router_timer;
rlist = &pmctx->ip6_rlist;
}
br_multicast_mark_router(brmctx, pmctx, timer, rlist,
&brmctx->ip6_mc_router_list);
#endif
}
static void
br_ip4_multicast_query_received(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct bridge_mcast_other_query *query,
struct br_ip *saddr,
unsigned long max_delay)
{
if (!br_ip4_multicast_select_querier(brmctx, pmctx->port, saddr->src.ip4))
return;
br_multicast_update_query_timer(brmctx, query, max_delay);
br_ip4_multicast_mark_router(brmctx, pmctx);
}
#if IS_ENABLED(CONFIG_IPV6)
static void
br_ip6_multicast_query_received(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct bridge_mcast_other_query *query,
struct br_ip *saddr,
unsigned long max_delay)
{
if (!br_ip6_multicast_select_querier(brmctx, pmctx->port, &saddr->src.ip6))
return;
br_multicast_update_query_timer(brmctx, query, max_delay);
br_ip6_multicast_mark_router(brmctx, pmctx);
}
#endif
static void br_ip4_multicast_query(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct sk_buff *skb,
u16 vid)
{
unsigned int transport_len = ip_transport_len(skb);
const struct iphdr *iph = ip_hdr(skb);
struct igmphdr *ih = igmp_hdr(skb);
struct net_bridge_mdb_entry *mp;
struct igmpv3_query *ih3;
struct net_bridge_port_group *p;
struct net_bridge_port_group __rcu **pp;
struct br_ip saddr;
unsigned long max_delay;
unsigned long now = jiffies;
__be32 group;
spin_lock(&brmctx->br->multicast_lock);
if (!br_multicast_ctx_should_use(brmctx, pmctx))
goto out;
group = ih->group;
if (transport_len == sizeof(*ih)) {
max_delay = ih->code * (HZ / IGMP_TIMER_SCALE);
if (!max_delay) {
max_delay = 10 * HZ;
group = 0;
}
} else if (transport_len >= sizeof(*ih3)) {
ih3 = igmpv3_query_hdr(skb);
if (ih3->nsrcs ||
(brmctx->multicast_igmp_version == 3 && group &&
ih3->suppress))
goto out;
max_delay = ih3->code ?
IGMPV3_MRC(ih3->code) * (HZ / IGMP_TIMER_SCALE) : 1;
} else {
goto out;
}
if (!group) {
saddr.proto = htons(ETH_P_IP);
saddr.src.ip4 = iph->saddr;
br_ip4_multicast_query_received(brmctx, pmctx,
&brmctx->ip4_other_query,
&saddr, max_delay);
goto out;
}
mp = br_mdb_ip4_get(brmctx->br, group, vid);
if (!mp)
goto out;
max_delay *= brmctx->multicast_last_member_count;
if (mp->host_joined &&
(timer_pending(&mp->timer) ?
time_after(mp->timer.expires, now + max_delay) :
try_to_del_timer_sync(&mp->timer) >= 0))
mod_timer(&mp->timer, now + max_delay);
for (pp = &mp->ports;
(p = mlock_dereference(*pp, brmctx->br)) != NULL;
pp = &p->next) {
if (timer_pending(&p->timer) ?
time_after(p->timer.expires, now + max_delay) :
try_to_del_timer_sync(&p->timer) >= 0 &&
(brmctx->multicast_igmp_version == 2 ||
p->filter_mode == MCAST_EXCLUDE))
mod_timer(&p->timer, now + max_delay);
}
out:
spin_unlock(&brmctx->br->multicast_lock);
}
#if IS_ENABLED(CONFIG_IPV6)
static int br_ip6_multicast_query(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct sk_buff *skb,
u16 vid)
{
unsigned int transport_len = ipv6_transport_len(skb);
struct mld_msg *mld;
struct net_bridge_mdb_entry *mp;
struct mld2_query *mld2q;
struct net_bridge_port_group *p;
struct net_bridge_port_group __rcu **pp;
struct br_ip saddr;
unsigned long max_delay;
unsigned long now = jiffies;
unsigned int offset = skb_transport_offset(skb);
const struct in6_addr *group = NULL;
bool is_general_query;
int err = 0;
spin_lock(&brmctx->br->multicast_lock);
if (!br_multicast_ctx_should_use(brmctx, pmctx))
goto out;
if (transport_len == sizeof(*mld)) {
if (!pskb_may_pull(skb, offset + sizeof(*mld))) {
err = -EINVAL;
goto out;
}
mld = (struct mld_msg *) icmp6_hdr(skb);
max_delay = msecs_to_jiffies(ntohs(mld->mld_maxdelay));
if (max_delay)
group = &mld->mld_mca;
} else {
if (!pskb_may_pull(skb, offset + sizeof(*mld2q))) {
err = -EINVAL;
goto out;
}
mld2q = (struct mld2_query *)icmp6_hdr(skb);
if (!mld2q->mld2q_nsrcs)
group = &mld2q->mld2q_mca;
if (brmctx->multicast_mld_version == 2 &&
!ipv6_addr_any(&mld2q->mld2q_mca) &&
mld2q->mld2q_suppress)
goto out;
max_delay = max(msecs_to_jiffies(mldv2_mrc(mld2q)), 1UL);
}
is_general_query = group && ipv6_addr_any(group);
if (is_general_query) {
saddr.proto = htons(ETH_P_IPV6);
saddr.src.ip6 = ipv6_hdr(skb)->saddr;
br_ip6_multicast_query_received(brmctx, pmctx,
&brmctx->ip6_other_query,
&saddr, max_delay);
goto out;
} else if (!group) {
goto out;
}
mp = br_mdb_ip6_get(brmctx->br, group, vid);
if (!mp)
goto out;
max_delay *= brmctx->multicast_last_member_count;
if (mp->host_joined &&
(timer_pending(&mp->timer) ?
time_after(mp->timer.expires, now + max_delay) :
try_to_del_timer_sync(&mp->timer) >= 0))
mod_timer(&mp->timer, now + max_delay);
for (pp = &mp->ports;
(p = mlock_dereference(*pp, brmctx->br)) != NULL;
pp = &p->next) {
if (timer_pending(&p->timer) ?
time_after(p->timer.expires, now + max_delay) :
try_to_del_timer_sync(&p->timer) >= 0 &&
(brmctx->multicast_mld_version == 1 ||
p->filter_mode == MCAST_EXCLUDE))
mod_timer(&p->timer, now + max_delay);
}
out:
spin_unlock(&brmctx->br->multicast_lock);
return err;
}
#endif
static void
br_multicast_leave_group(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct br_ip *group,
struct bridge_mcast_other_query *other_query,
struct bridge_mcast_own_query *own_query,
const unsigned char *src)
{
struct net_bridge_mdb_entry *mp;
struct net_bridge_port_group *p;
unsigned long now;
unsigned long time;
spin_lock(&brmctx->br->multicast_lock);
if (!br_multicast_ctx_should_use(brmctx, pmctx))
goto out;
mp = br_mdb_ip_get(brmctx->br, group);
if (!mp)
goto out;
if (pmctx && (pmctx->port->flags & BR_MULTICAST_FAST_LEAVE)) {
struct net_bridge_port_group __rcu **pp;
for (pp = &mp->ports;
(p = mlock_dereference(*pp, brmctx->br)) != NULL;
pp = &p->next) {
if (!br_port_group_equal(p, pmctx->port, src))
continue;
if (p->flags & MDB_PG_FLAGS_PERMANENT)
break;
p->flags |= MDB_PG_FLAGS_FAST_LEAVE;
br_multicast_del_pg(mp, p, pp);
}
goto out;
}
if (timer_pending(&other_query->timer))
goto out;
if (br_opt_get(brmctx->br, BROPT_MULTICAST_QUERIER)) {
__br_multicast_send_query(brmctx, pmctx, NULL, NULL, &mp->addr,
false, 0, NULL);
time = jiffies + brmctx->multicast_last_member_count *
brmctx->multicast_last_member_interval;
mod_timer(&own_query->timer, time);
for (p = mlock_dereference(mp->ports, brmctx->br);
p != NULL;
p = mlock_dereference(p->next, brmctx->br)) {
if (!br_port_group_equal(p, pmctx->port, src))
continue;
if (!hlist_unhashed(&p->mglist) &&
(timer_pending(&p->timer) ?
time_after(p->timer.expires, time) :
try_to_del_timer_sync(&p->timer) >= 0)) {
mod_timer(&p->timer, time);
}
break;
}
}
now = jiffies;
time = now + brmctx->multicast_last_member_count *
brmctx->multicast_last_member_interval;
if (!pmctx) {
if (mp->host_joined &&
(timer_pending(&mp->timer) ?
time_after(mp->timer.expires, time) :
try_to_del_timer_sync(&mp->timer) >= 0)) {
mod_timer(&mp->timer, time);
}
Revert "bridge: only expire the mdb entry when query is received" While this commit was a good attempt to fix issues occuring when no multicast querier is present, this commit still has two more issues: 1) There are cases where mdb entries do not expire even if there is a querier present. The bridge will unnecessarily continue flooding multicast packets on the according ports. 2) Never removing an mdb entry could be exploited for a Denial of Service by an attacker on the local link, slowly, but steadily eating up all memory. Actually, this commit became obsolete with "bridge: disable snooping if there is no querier" (b00589af3b) which included fixes for a few more cases. Therefore reverting the following commits (the commit stated in the commit message plus three of its follow up fixes): ==================== Revert "bridge: update mdb expiration timer upon reports." This reverts commit f144febd93d5ee534fdf23505ab091b2b9088edc. Revert "bridge: do not call setup_timer() multiple times" This reverts commit 1faabf2aab1fdaa1ace4e8c829d1b9cf7bfec2f1. Revert "bridge: fix some kernel warning in multicast timer" This reverts commit c7e8e8a8f7a70b343ca1e0f90a31e35ab2d16de1. Revert "bridge: only expire the mdb entry when query is received" This reverts commit 9f00b2e7cf241fa389733d41b615efdaa2cb0f5b. ==================== CC: Cong Wang <amwang@redhat.com> Signed-off-by: Linus Lüssing <linus.luessing@web.de> Reviewed-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-10-20 02:58:57 +04:00
goto out;
}
for (p = mlock_dereference(mp->ports, brmctx->br);
Revert "bridge: only expire the mdb entry when query is received" While this commit was a good attempt to fix issues occuring when no multicast querier is present, this commit still has two more issues: 1) There are cases where mdb entries do not expire even if there is a querier present. The bridge will unnecessarily continue flooding multicast packets on the according ports. 2) Never removing an mdb entry could be exploited for a Denial of Service by an attacker on the local link, slowly, but steadily eating up all memory. Actually, this commit became obsolete with "bridge: disable snooping if there is no querier" (b00589af3b) which included fixes for a few more cases. Therefore reverting the following commits (the commit stated in the commit message plus three of its follow up fixes): ==================== Revert "bridge: update mdb expiration timer upon reports." This reverts commit f144febd93d5ee534fdf23505ab091b2b9088edc. Revert "bridge: do not call setup_timer() multiple times" This reverts commit 1faabf2aab1fdaa1ace4e8c829d1b9cf7bfec2f1. Revert "bridge: fix some kernel warning in multicast timer" This reverts commit c7e8e8a8f7a70b343ca1e0f90a31e35ab2d16de1. Revert "bridge: only expire the mdb entry when query is received" This reverts commit 9f00b2e7cf241fa389733d41b615efdaa2cb0f5b. ==================== CC: Cong Wang <amwang@redhat.com> Signed-off-by: Linus Lüssing <linus.luessing@web.de> Reviewed-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-10-20 02:58:57 +04:00
p != NULL;
p = mlock_dereference(p->next, brmctx->br)) {
if (p->key.port != pmctx->port)
Revert "bridge: only expire the mdb entry when query is received" While this commit was a good attempt to fix issues occuring when no multicast querier is present, this commit still has two more issues: 1) There are cases where mdb entries do not expire even if there is a querier present. The bridge will unnecessarily continue flooding multicast packets on the according ports. 2) Never removing an mdb entry could be exploited for a Denial of Service by an attacker on the local link, slowly, but steadily eating up all memory. Actually, this commit became obsolete with "bridge: disable snooping if there is no querier" (b00589af3b) which included fixes for a few more cases. Therefore reverting the following commits (the commit stated in the commit message plus three of its follow up fixes): ==================== Revert "bridge: update mdb expiration timer upon reports." This reverts commit f144febd93d5ee534fdf23505ab091b2b9088edc. Revert "bridge: do not call setup_timer() multiple times" This reverts commit 1faabf2aab1fdaa1ace4e8c829d1b9cf7bfec2f1. Revert "bridge: fix some kernel warning in multicast timer" This reverts commit c7e8e8a8f7a70b343ca1e0f90a31e35ab2d16de1. Revert "bridge: only expire the mdb entry when query is received" This reverts commit 9f00b2e7cf241fa389733d41b615efdaa2cb0f5b. ==================== CC: Cong Wang <amwang@redhat.com> Signed-off-by: Linus Lüssing <linus.luessing@web.de> Reviewed-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-10-20 02:58:57 +04:00
continue;
if (!hlist_unhashed(&p->mglist) &&
(timer_pending(&p->timer) ?
time_after(p->timer.expires, time) :
try_to_del_timer_sync(&p->timer) >= 0)) {
mod_timer(&p->timer, time);
}
break;
}
out:
spin_unlock(&brmctx->br->multicast_lock);
}
static void br_ip4_multicast_leave_group(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
__be32 group,
__u16 vid,
const unsigned char *src)
{
struct br_ip br_group;
struct bridge_mcast_own_query *own_query;
if (ipv4_is_local_multicast(group))
return;
own_query = pmctx ? &pmctx->ip4_own_query : &brmctx->ip4_own_query;
memset(&br_group, 0, sizeof(br_group));
br_group.dst.ip4 = group;
br_group.proto = htons(ETH_P_IP);
br_group.vid = vid;
br_multicast_leave_group(brmctx, pmctx, &br_group,
&brmctx->ip4_other_query,
own_query, src);
}
#if IS_ENABLED(CONFIG_IPV6)
static void br_ip6_multicast_leave_group(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
const struct in6_addr *group,
__u16 vid,
const unsigned char *src)
{
struct br_ip br_group;
struct bridge_mcast_own_query *own_query;
if (ipv6_addr_is_ll_all_nodes(group))
return;
own_query = pmctx ? &pmctx->ip6_own_query : &brmctx->ip6_own_query;
memset(&br_group, 0, sizeof(br_group));
br_group.dst.ip6 = *group;
br_group.proto = htons(ETH_P_IPV6);
br_group.vid = vid;
br_multicast_leave_group(brmctx, pmctx, &br_group,
&brmctx->ip6_other_query,
own_query, src);
}
#endif
static void br_multicast_err_count(const struct net_bridge *br,
const struct net_bridge_port *p,
__be16 proto)
{
struct bridge_mcast_stats __percpu *stats;
struct bridge_mcast_stats *pstats;
if (!br_opt_get(br, BROPT_MULTICAST_STATS_ENABLED))
return;
if (p)
stats = p->mcast_stats;
else
stats = br->mcast_stats;
if (WARN_ON(!stats))
return;
pstats = this_cpu_ptr(stats);
u64_stats_update_begin(&pstats->syncp);
switch (proto) {
case htons(ETH_P_IP):
pstats->mstats.igmp_parse_errors++;
break;
#if IS_ENABLED(CONFIG_IPV6)
case htons(ETH_P_IPV6):
pstats->mstats.mld_parse_errors++;
break;
#endif
}
u64_stats_update_end(&pstats->syncp);
}
static void br_multicast_pim(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
const struct sk_buff *skb)
{
unsigned int offset = skb_transport_offset(skb);
struct pimhdr *pimhdr, _pimhdr;
pimhdr = skb_header_pointer(skb, offset, sizeof(_pimhdr), &_pimhdr);
if (!pimhdr || pim_hdr_version(pimhdr) != PIM_VERSION ||
pim_hdr_type(pimhdr) != PIM_TYPE_HELLO)
return;
spin_lock(&brmctx->br->multicast_lock);
br_ip4_multicast_mark_router(brmctx, pmctx);
spin_unlock(&brmctx->br->multicast_lock);
}
static int br_ip4_multicast_mrd_rcv(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct sk_buff *skb)
{
if (ip_hdr(skb)->protocol != IPPROTO_IGMP ||
igmp_hdr(skb)->type != IGMP_MRDISC_ADV)
return -ENOMSG;
spin_lock(&brmctx->br->multicast_lock);
br_ip4_multicast_mark_router(brmctx, pmctx);
spin_unlock(&brmctx->br->multicast_lock);
return 0;
}
static int br_multicast_ipv4_rcv(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct sk_buff *skb,
u16 vid)
{
struct net_bridge_port *p = pmctx ? pmctx->port : NULL;
const unsigned char *src;
struct igmphdr *ih;
int err;
err = ip_mc_check_igmp(skb);
if (err == -ENOMSG) {
if (!ipv4_is_local_multicast(ip_hdr(skb)->daddr)) {
BR_INPUT_SKB_CB(skb)->mrouters_only = 1;
} else if (pim_ipv4_all_pim_routers(ip_hdr(skb)->daddr)) {
if (ip_hdr(skb)->protocol == IPPROTO_PIM)
br_multicast_pim(brmctx, pmctx, skb);
} else if (ipv4_is_all_snoopers(ip_hdr(skb)->daddr)) {
br_ip4_multicast_mrd_rcv(brmctx, pmctx, skb);
}
return 0;
} else if (err < 0) {
br_multicast_err_count(brmctx->br, p, skb->protocol);
return err;
}
ih = igmp_hdr(skb);
src = eth_hdr(skb)->h_source;
BR_INPUT_SKB_CB(skb)->igmp = ih->type;
switch (ih->type) {
case IGMP_HOST_MEMBERSHIP_REPORT:
case IGMPV2_HOST_MEMBERSHIP_REPORT:
BR_INPUT_SKB_CB(skb)->mrouters_only = 1;
err = br_ip4_multicast_add_group(brmctx, pmctx, ih->group, vid,
src, true);
break;
case IGMPV3_HOST_MEMBERSHIP_REPORT:
err = br_ip4_multicast_igmp3_report(brmctx, pmctx, skb, vid);
break;
case IGMP_HOST_MEMBERSHIP_QUERY:
br_ip4_multicast_query(brmctx, pmctx, skb, vid);
break;
case IGMP_HOST_LEAVE_MESSAGE:
br_ip4_multicast_leave_group(brmctx, pmctx, ih->group, vid, src);
break;
}
br_multicast_count(brmctx->br, p, skb, BR_INPUT_SKB_CB(skb)->igmp,
BR_MCAST_DIR_RX);
return err;
}
#if IS_ENABLED(CONFIG_IPV6)
static void br_ip6_multicast_mrd_rcv(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
net: bridge: mcast: fix broken length + header check for MRDv6 Adv. The IPv6 Multicast Router Advertisements parsing has the following two issues: For one thing, ICMPv6 MRD Advertisements are smaller than ICMPv6 MLD messages (ICMPv6 MRD Adv.: 8 bytes vs. ICMPv6 MLDv1/2: >= 24 bytes, assuming MLDv2 Reports with at least one multicast address entry). When ipv6_mc_check_mld_msg() tries to parse an Multicast Router Advertisement its MLD length check will fail - and it will wrongly return -EINVAL, even if we have a valid MRD Advertisement. With the returned -EINVAL the bridge code will assume a broken packet and will wrongly discard it, potentially leading to multicast packet loss towards multicast routers. The second issue is the MRD header parsing in br_ip6_multicast_mrd_rcv(): It wrongly checks for an ICMPv6 header immediately after the IPv6 header (IPv6 next header type). However according to RFC4286, section 2 all MRD messages contain a Router Alert option (just like MLD). So instead there is an IPv6 Hop-by-Hop option for the Router Alert between the IPv6 and ICMPv6 header, again leading to the bridge wrongly discarding Multicast Router Advertisements. To fix these two issues, introduce a new return value -ENODATA to ipv6_mc_check_mld() to indicate a valid ICMPv6 packet with a hop-by-hop option which is not an MLD but potentially an MRD packet. This also simplifies further parsing in the bridge code, as ipv6_mc_check_mld() already fully checks the ICMPv6 header and hop-by-hop option. These issues were found and fixed with the help of the mrdisc tool (https://github.com/troglobit/mrdisc). Fixes: 4b3087c7e37f ("bridge: Snoop Multicast Router Advertisements") Signed-off-by: Linus Lüssing <linus.luessing@c0d3.blue> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-04-25 18:27:35 +03:00
struct sk_buff *skb)
{
if (icmp6_hdr(skb)->icmp6_type != ICMPV6_MRDISC_ADV)
net: bridge: mcast: fix broken length + header check for MRDv6 Adv. The IPv6 Multicast Router Advertisements parsing has the following two issues: For one thing, ICMPv6 MRD Advertisements are smaller than ICMPv6 MLD messages (ICMPv6 MRD Adv.: 8 bytes vs. ICMPv6 MLDv1/2: >= 24 bytes, assuming MLDv2 Reports with at least one multicast address entry). When ipv6_mc_check_mld_msg() tries to parse an Multicast Router Advertisement its MLD length check will fail - and it will wrongly return -EINVAL, even if we have a valid MRD Advertisement. With the returned -EINVAL the bridge code will assume a broken packet and will wrongly discard it, potentially leading to multicast packet loss towards multicast routers. The second issue is the MRD header parsing in br_ip6_multicast_mrd_rcv(): It wrongly checks for an ICMPv6 header immediately after the IPv6 header (IPv6 next header type). However according to RFC4286, section 2 all MRD messages contain a Router Alert option (just like MLD). So instead there is an IPv6 Hop-by-Hop option for the Router Alert between the IPv6 and ICMPv6 header, again leading to the bridge wrongly discarding Multicast Router Advertisements. To fix these two issues, introduce a new return value -ENODATA to ipv6_mc_check_mld() to indicate a valid ICMPv6 packet with a hop-by-hop option which is not an MLD but potentially an MRD packet. This also simplifies further parsing in the bridge code, as ipv6_mc_check_mld() already fully checks the ICMPv6 header and hop-by-hop option. These issues were found and fixed with the help of the mrdisc tool (https://github.com/troglobit/mrdisc). Fixes: 4b3087c7e37f ("bridge: Snoop Multicast Router Advertisements") Signed-off-by: Linus Lüssing <linus.luessing@c0d3.blue> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-04-25 18:27:35 +03:00
return;
spin_lock(&brmctx->br->multicast_lock);
br_ip6_multicast_mark_router(brmctx, pmctx);
spin_unlock(&brmctx->br->multicast_lock);
}
static int br_multicast_ipv6_rcv(struct net_bridge_mcast *brmctx,
struct net_bridge_mcast_port *pmctx,
struct sk_buff *skb,
u16 vid)
{
struct net_bridge_port *p = pmctx ? pmctx->port : NULL;
const unsigned char *src;
struct mld_msg *mld;
int err;
err = ipv6_mc_check_mld(skb);
net: bridge: mcast: fix broken length + header check for MRDv6 Adv. The IPv6 Multicast Router Advertisements parsing has the following two issues: For one thing, ICMPv6 MRD Advertisements are smaller than ICMPv6 MLD messages (ICMPv6 MRD Adv.: 8 bytes vs. ICMPv6 MLDv1/2: >= 24 bytes, assuming MLDv2 Reports with at least one multicast address entry). When ipv6_mc_check_mld_msg() tries to parse an Multicast Router Advertisement its MLD length check will fail - and it will wrongly return -EINVAL, even if we have a valid MRD Advertisement. With the returned -EINVAL the bridge code will assume a broken packet and will wrongly discard it, potentially leading to multicast packet loss towards multicast routers. The second issue is the MRD header parsing in br_ip6_multicast_mrd_rcv(): It wrongly checks for an ICMPv6 header immediately after the IPv6 header (IPv6 next header type). However according to RFC4286, section 2 all MRD messages contain a Router Alert option (just like MLD). So instead there is an IPv6 Hop-by-Hop option for the Router Alert between the IPv6 and ICMPv6 header, again leading to the bridge wrongly discarding Multicast Router Advertisements. To fix these two issues, introduce a new return value -ENODATA to ipv6_mc_check_mld() to indicate a valid ICMPv6 packet with a hop-by-hop option which is not an MLD but potentially an MRD packet. This also simplifies further parsing in the bridge code, as ipv6_mc_check_mld() already fully checks the ICMPv6 header and hop-by-hop option. These issues were found and fixed with the help of the mrdisc tool (https://github.com/troglobit/mrdisc). Fixes: 4b3087c7e37f ("bridge: Snoop Multicast Router Advertisements") Signed-off-by: Linus Lüssing <linus.luessing@c0d3.blue> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-04-25 18:27:35 +03:00
if (err == -ENOMSG || err == -ENODATA) {
if (!ipv6_addr_is_ll_all_nodes(&ipv6_hdr(skb)->daddr))
BR_INPUT_SKB_CB(skb)->mrouters_only = 1;
net: bridge: mcast: fix broken length + header check for MRDv6 Adv. The IPv6 Multicast Router Advertisements parsing has the following two issues: For one thing, ICMPv6 MRD Advertisements are smaller than ICMPv6 MLD messages (ICMPv6 MRD Adv.: 8 bytes vs. ICMPv6 MLDv1/2: >= 24 bytes, assuming MLDv2 Reports with at least one multicast address entry). When ipv6_mc_check_mld_msg() tries to parse an Multicast Router Advertisement its MLD length check will fail - and it will wrongly return -EINVAL, even if we have a valid MRD Advertisement. With the returned -EINVAL the bridge code will assume a broken packet and will wrongly discard it, potentially leading to multicast packet loss towards multicast routers. The second issue is the MRD header parsing in br_ip6_multicast_mrd_rcv(): It wrongly checks for an ICMPv6 header immediately after the IPv6 header (IPv6 next header type). However according to RFC4286, section 2 all MRD messages contain a Router Alert option (just like MLD). So instead there is an IPv6 Hop-by-Hop option for the Router Alert between the IPv6 and ICMPv6 header, again leading to the bridge wrongly discarding Multicast Router Advertisements. To fix these two issues, introduce a new return value -ENODATA to ipv6_mc_check_mld() to indicate a valid ICMPv6 packet with a hop-by-hop option which is not an MLD but potentially an MRD packet. This also simplifies further parsing in the bridge code, as ipv6_mc_check_mld() already fully checks the ICMPv6 header and hop-by-hop option. These issues were found and fixed with the help of the mrdisc tool (https://github.com/troglobit/mrdisc). Fixes: 4b3087c7e37f ("bridge: Snoop Multicast Router Advertisements") Signed-off-by: Linus Lüssing <linus.luessing@c0d3.blue> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-04-25 18:27:35 +03:00
if (err == -ENODATA &&
ipv6_addr_is_all_snoopers(&ipv6_hdr(skb)->daddr))
br_ip6_multicast_mrd_rcv(brmctx, pmctx, skb);
return 0;
} else if (err < 0) {
br_multicast_err_count(brmctx->br, p, skb->protocol);
return err;
}
mld = (struct mld_msg *)skb_transport_header(skb);
BR_INPUT_SKB_CB(skb)->igmp = mld->mld_type;
switch (mld->mld_type) {
case ICMPV6_MGM_REPORT:
src = eth_hdr(skb)->h_source;
BR_INPUT_SKB_CB(skb)->mrouters_only = 1;
err = br_ip6_multicast_add_group(brmctx, pmctx, &mld->mld_mca,
vid, src, true);
break;
case ICMPV6_MLD2_REPORT:
err = br_ip6_multicast_mld2_report(brmctx, pmctx, skb, vid);
break;
case ICMPV6_MGM_QUERY:
err = br_ip6_multicast_query(brmctx, pmctx, skb, vid);
break;
case ICMPV6_MGM_REDUCTION:
src = eth_hdr(skb)->h_source;
br_ip6_multicast_leave_group(brmctx, pmctx, &mld->mld_mca, vid,
src);
break;
}
br_multicast_count(brmctx->br, p, skb, BR_INPUT_SKB_CB(skb)->igmp,
BR_MCAST_DIR_RX);
return err;
}
#endif
int br_multicast_rcv(struct net_bridge_mcast **brmctx,
struct net_bridge_mcast_port **pmctx,
struct net_bridge_vlan *vlan,
struct sk_buff *skb, u16 vid)
{
int ret = 0;
BR_INPUT_SKB_CB(skb)->igmp = 0;
BR_INPUT_SKB_CB(skb)->mrouters_only = 0;
if (!br_opt_get((*brmctx)->br, BROPT_MULTICAST_ENABLED))
return 0;
if (br_opt_get((*brmctx)->br, BROPT_MCAST_VLAN_SNOOPING_ENABLED) && vlan) {
const struct net_bridge_vlan *masterv;
/* the vlan has the master flag set only when transmitting
* through the bridge device
*/
if (br_vlan_is_master(vlan)) {
masterv = vlan;
*brmctx = &vlan->br_mcast_ctx;
*pmctx = NULL;
} else {
masterv = vlan->brvlan;
*brmctx = &vlan->brvlan->br_mcast_ctx;
*pmctx = &vlan->port_mcast_ctx;
}
if (!(masterv->priv_flags & BR_VLFLAG_GLOBAL_MCAST_ENABLED))
return 0;
}
switch (skb->protocol) {
case htons(ETH_P_IP):
ret = br_multicast_ipv4_rcv(*brmctx, *pmctx, skb, vid);
break;
#if IS_ENABLED(CONFIG_IPV6)
case htons(ETH_P_IPV6):
ret = br_multicast_ipv6_rcv(*brmctx, *pmctx, skb, vid);
break;
#endif
}
return ret;
}
static void br_multicast_query_expired(struct net_bridge_mcast *brmctx,
struct bridge_mcast_own_query *query,
struct bridge_mcast_querier *querier)
{
spin_lock(&brmctx->br->multicast_lock);
if (br_multicast_ctx_vlan_disabled(brmctx))
goto out;
if (query->startup_sent < brmctx->multicast_startup_query_count)
query->startup_sent++;
RCU_INIT_POINTER(querier->port, NULL);
br_multicast_send_query(brmctx, NULL, query);
out:
spin_unlock(&brmctx->br->multicast_lock);
}
static void br_ip4_multicast_query_expired(struct timer_list *t)
{
struct net_bridge_mcast *brmctx = from_timer(brmctx, t,
ip4_own_query.timer);
br_multicast_query_expired(brmctx, &brmctx->ip4_own_query,
&brmctx->ip4_querier);
}
#if IS_ENABLED(CONFIG_IPV6)
static void br_ip6_multicast_query_expired(struct timer_list *t)
{
struct net_bridge_mcast *brmctx = from_timer(brmctx, t,
ip6_own_query.timer);
br_multicast_query_expired(brmctx, &brmctx->ip6_own_query,
&brmctx->ip6_querier);
}
#endif
static void br_multicast_gc_work(struct work_struct *work)
{
struct net_bridge *br = container_of(work, struct net_bridge,
mcast_gc_work);
HLIST_HEAD(deleted_head);
spin_lock_bh(&br->multicast_lock);
hlist_move_list(&br->mcast_gc_list, &deleted_head);
spin_unlock_bh(&br->multicast_lock);
br_multicast_gc(&deleted_head);
}
void br_multicast_ctx_init(struct net_bridge *br,
struct net_bridge_vlan *vlan,
struct net_bridge_mcast *brmctx)
{
brmctx->br = br;
brmctx->vlan = vlan;
brmctx->multicast_router = MDB_RTR_TYPE_TEMP_QUERY;
brmctx->multicast_last_member_count = 2;
brmctx->multicast_startup_query_count = 2;
brmctx->multicast_last_member_interval = HZ;
brmctx->multicast_query_response_interval = 10 * HZ;
brmctx->multicast_startup_query_interval = 125 * HZ / 4;
brmctx->multicast_query_interval = 125 * HZ;
brmctx->multicast_querier_interval = 255 * HZ;
brmctx->multicast_membership_interval = 260 * HZ;
brmctx->ip4_other_query.delay_time = 0;
brmctx->ip4_querier.port = NULL;
brmctx->multicast_igmp_version = 2;
#if IS_ENABLED(CONFIG_IPV6)
brmctx->multicast_mld_version = 1;
brmctx->ip6_other_query.delay_time = 0;
brmctx->ip6_querier.port = NULL;
#endif
timer_setup(&brmctx->ip4_mc_router_timer,
br_ip4_multicast_local_router_expired, 0);
timer_setup(&brmctx->ip4_other_query.timer,
br_ip4_multicast_querier_expired, 0);
timer_setup(&brmctx->ip4_own_query.timer,
br_ip4_multicast_query_expired, 0);
#if IS_ENABLED(CONFIG_IPV6)
timer_setup(&brmctx->ip6_mc_router_timer,
br_ip6_multicast_local_router_expired, 0);
timer_setup(&brmctx->ip6_other_query.timer,
br_ip6_multicast_querier_expired, 0);
timer_setup(&brmctx->ip6_own_query.timer,
br_ip6_multicast_query_expired, 0);
#endif
}
void br_multicast_ctx_deinit(struct net_bridge_mcast *brmctx)
{
__br_multicast_stop(brmctx);
}
void br_multicast_init(struct net_bridge *br)
{
br->hash_max = BR_MULTICAST_DEFAULT_HASH_MAX;
br_multicast_ctx_init(br, NULL, &br->multicast_ctx);
br_opt_toggle(br, BROPT_MULTICAST_ENABLED, true);
br_opt_toggle(br, BROPT_HAS_IPV6_ADDR, true);
spin_lock_init(&br->multicast_lock);
INIT_HLIST_HEAD(&br->mdb_list);
INIT_HLIST_HEAD(&br->mcast_gc_list);
INIT_WORK(&br->mcast_gc_work, br_multicast_gc_work);
}
static void br_ip4_multicast_join_snoopers(struct net_bridge *br)
{
struct in_device *in_dev = in_dev_get(br->dev);
if (!in_dev)
return;
net: Fix ip_mc_{dec,inc}_group allocation context After 4effd28c1245 ("bridge: join all-snoopers multicast address"), I started seeing the following sleep in atomic warnings: [ 26.763893] BUG: sleeping function called from invalid context at mm/slab.h:421 [ 26.771425] in_atomic(): 1, irqs_disabled(): 0, pid: 1658, name: sh [ 26.777855] INFO: lockdep is turned off. [ 26.781916] CPU: 0 PID: 1658 Comm: sh Not tainted 5.0.0-rc4 #20 [ 26.787943] Hardware name: BCM97278SV (DT) [ 26.792118] Call trace: [ 26.794645] dump_backtrace+0x0/0x170 [ 26.798391] show_stack+0x24/0x30 [ 26.801787] dump_stack+0xa4/0xe4 [ 26.805182] ___might_sleep+0x208/0x218 [ 26.809102] __might_sleep+0x78/0x88 [ 26.812762] kmem_cache_alloc_trace+0x64/0x28c [ 26.817301] igmp_group_dropped+0x150/0x230 [ 26.821573] ip_mc_dec_group+0x1b0/0x1f8 [ 26.825585] br_ip4_multicast_leave_snoopers.isra.11+0x174/0x190 [ 26.831704] br_multicast_toggle+0x78/0xcc [ 26.835887] store_bridge_parm+0xc4/0xfc [ 26.839894] multicast_snooping_store+0x3c/0x4c [ 26.844517] dev_attr_store+0x44/0x5c [ 26.848262] sysfs_kf_write+0x50/0x68 [ 26.852006] kernfs_fop_write+0x14c/0x1b4 [ 26.856102] __vfs_write+0x60/0x190 [ 26.859668] vfs_write+0xc8/0x168 [ 26.863059] ksys_write+0x70/0xc8 [ 26.866449] __arm64_sys_write+0x24/0x30 [ 26.870458] el0_svc_common+0xa0/0x11c [ 26.874291] el0_svc_handler+0x38/0x70 [ 26.878120] el0_svc+0x8/0xc while toggling the bridge's multicast_snooping attribute dynamically. Pass a gfp_t down to igmpv3_add_delrec(), introduce __igmp_group_dropped() and introduce __ip_mc_dec_group() to take a gfp_t argument. Similarly introduce ____ip_mc_inc_group() and __ip_mc_inc_group() to allow caller to specify gfp_t. IPv6 part of the patch appears fine. Fixes: 4effd28c1245 ("bridge: join all-snoopers multicast address") Signed-off-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-02 07:20:52 +03:00
__ip_mc_inc_group(in_dev, htonl(INADDR_ALLSNOOPERS_GROUP), GFP_ATOMIC);
in_dev_put(in_dev);
}
#if IS_ENABLED(CONFIG_IPV6)
static void br_ip6_multicast_join_snoopers(struct net_bridge *br)
{
struct in6_addr addr;
ipv6_addr_set(&addr, htonl(0xff020000), 0, 0, htonl(0x6a));
ipv6_dev_mc_inc(br->dev, &addr);
}
#else
static inline void br_ip6_multicast_join_snoopers(struct net_bridge *br)
{
}
#endif
bridge: Fix a deadlock when enabling multicast snooping When enabling multicast snooping, bridge module deadlocks on multicast_lock if 1) IPv6 is enabled, and 2) there is an existing querier on the same L2 network. The deadlock was caused by the following sequence: While holding the lock, br_multicast_open calls br_multicast_join_snoopers, which eventually causes IP stack to (attempt to) send out a Listener Report (in igmp6_join_group). Since the destination Ethernet address is a multicast address, br_dev_xmit feeds the packet back to the bridge via br_multicast_rcv, which in turn calls br_multicast_add_group, which then deadlocks on multicast_lock. The fix is to move the call br_multicast_join_snoopers outside of the critical section. This works since br_multicast_join_snoopers only deals with IP and does not modify any multicast data structures of the bridge, so there's no need to hold the lock. Steps to reproduce: 1. sysctl net.ipv6.conf.all.force_mld_version=1 2. have another querier 3. ip link set dev bridge type bridge mcast_snooping 0 && \ ip link set dev bridge type bridge mcast_snooping 1 < deadlock > A typical call trace looks like the following: [ 936.251495] _raw_spin_lock+0x5c/0x68 [ 936.255221] br_multicast_add_group+0x40/0x170 [bridge] [ 936.260491] br_multicast_rcv+0x7ac/0xe30 [bridge] [ 936.265322] br_dev_xmit+0x140/0x368 [bridge] [ 936.269689] dev_hard_start_xmit+0x94/0x158 [ 936.273876] __dev_queue_xmit+0x5ac/0x7f8 [ 936.277890] dev_queue_xmit+0x10/0x18 [ 936.281563] neigh_resolve_output+0xec/0x198 [ 936.285845] ip6_finish_output2+0x240/0x710 [ 936.290039] __ip6_finish_output+0x130/0x170 [ 936.294318] ip6_output+0x6c/0x1c8 [ 936.297731] NF_HOOK.constprop.0+0xd8/0xe8 [ 936.301834] igmp6_send+0x358/0x558 [ 936.305326] igmp6_join_group.part.0+0x30/0xf0 [ 936.309774] igmp6_group_added+0xfc/0x110 [ 936.313787] __ipv6_dev_mc_inc+0x1a4/0x290 [ 936.317885] ipv6_dev_mc_inc+0x10/0x18 [ 936.321677] br_multicast_open+0xbc/0x110 [bridge] [ 936.326506] br_multicast_toggle+0xec/0x140 [bridge] Fixes: 4effd28c1245 ("bridge: join all-snoopers multicast address") Signed-off-by: Joseph Huang <Joseph.Huang@garmin.com> Acked-by: Nikolay Aleksandrov <nikolay@nvidia.com> Link: https://lore.kernel.org/r/20201204235628.50653-1-Joseph.Huang@garmin.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2020-12-05 02:56:28 +03:00
void br_multicast_join_snoopers(struct net_bridge *br)
{
br_ip4_multicast_join_snoopers(br);
br_ip6_multicast_join_snoopers(br);
}
static void br_ip4_multicast_leave_snoopers(struct net_bridge *br)
{
struct in_device *in_dev = in_dev_get(br->dev);
if (WARN_ON(!in_dev))
return;
net: Fix ip_mc_{dec,inc}_group allocation context After 4effd28c1245 ("bridge: join all-snoopers multicast address"), I started seeing the following sleep in atomic warnings: [ 26.763893] BUG: sleeping function called from invalid context at mm/slab.h:421 [ 26.771425] in_atomic(): 1, irqs_disabled(): 0, pid: 1658, name: sh [ 26.777855] INFO: lockdep is turned off. [ 26.781916] CPU: 0 PID: 1658 Comm: sh Not tainted 5.0.0-rc4 #20 [ 26.787943] Hardware name: BCM97278SV (DT) [ 26.792118] Call trace: [ 26.794645] dump_backtrace+0x0/0x170 [ 26.798391] show_stack+0x24/0x30 [ 26.801787] dump_stack+0xa4/0xe4 [ 26.805182] ___might_sleep+0x208/0x218 [ 26.809102] __might_sleep+0x78/0x88 [ 26.812762] kmem_cache_alloc_trace+0x64/0x28c [ 26.817301] igmp_group_dropped+0x150/0x230 [ 26.821573] ip_mc_dec_group+0x1b0/0x1f8 [ 26.825585] br_ip4_multicast_leave_snoopers.isra.11+0x174/0x190 [ 26.831704] br_multicast_toggle+0x78/0xcc [ 26.835887] store_bridge_parm+0xc4/0xfc [ 26.839894] multicast_snooping_store+0x3c/0x4c [ 26.844517] dev_attr_store+0x44/0x5c [ 26.848262] sysfs_kf_write+0x50/0x68 [ 26.852006] kernfs_fop_write+0x14c/0x1b4 [ 26.856102] __vfs_write+0x60/0x190 [ 26.859668] vfs_write+0xc8/0x168 [ 26.863059] ksys_write+0x70/0xc8 [ 26.866449] __arm64_sys_write+0x24/0x30 [ 26.870458] el0_svc_common+0xa0/0x11c [ 26.874291] el0_svc_handler+0x38/0x70 [ 26.878120] el0_svc+0x8/0xc while toggling the bridge's multicast_snooping attribute dynamically. Pass a gfp_t down to igmpv3_add_delrec(), introduce __igmp_group_dropped() and introduce __ip_mc_dec_group() to take a gfp_t argument. Similarly introduce ____ip_mc_inc_group() and __ip_mc_inc_group() to allow caller to specify gfp_t. IPv6 part of the patch appears fine. Fixes: 4effd28c1245 ("bridge: join all-snoopers multicast address") Signed-off-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-02 07:20:52 +03:00
__ip_mc_dec_group(in_dev, htonl(INADDR_ALLSNOOPERS_GROUP), GFP_ATOMIC);
in_dev_put(in_dev);
}
#if IS_ENABLED(CONFIG_IPV6)
static void br_ip6_multicast_leave_snoopers(struct net_bridge *br)
{
struct in6_addr addr;
ipv6_addr_set(&addr, htonl(0xff020000), 0, 0, htonl(0x6a));
ipv6_dev_mc_dec(br->dev, &addr);
}
#else
static inline void br_ip6_multicast_leave_snoopers(struct net_bridge *br)
{
}
#endif
bridge: Fix a deadlock when enabling multicast snooping When enabling multicast snooping, bridge module deadlocks on multicast_lock if 1) IPv6 is enabled, and 2) there is an existing querier on the same L2 network. The deadlock was caused by the following sequence: While holding the lock, br_multicast_open calls br_multicast_join_snoopers, which eventually causes IP stack to (attempt to) send out a Listener Report (in igmp6_join_group). Since the destination Ethernet address is a multicast address, br_dev_xmit feeds the packet back to the bridge via br_multicast_rcv, which in turn calls br_multicast_add_group, which then deadlocks on multicast_lock. The fix is to move the call br_multicast_join_snoopers outside of the critical section. This works since br_multicast_join_snoopers only deals with IP and does not modify any multicast data structures of the bridge, so there's no need to hold the lock. Steps to reproduce: 1. sysctl net.ipv6.conf.all.force_mld_version=1 2. have another querier 3. ip link set dev bridge type bridge mcast_snooping 0 && \ ip link set dev bridge type bridge mcast_snooping 1 < deadlock > A typical call trace looks like the following: [ 936.251495] _raw_spin_lock+0x5c/0x68 [ 936.255221] br_multicast_add_group+0x40/0x170 [bridge] [ 936.260491] br_multicast_rcv+0x7ac/0xe30 [bridge] [ 936.265322] br_dev_xmit+0x140/0x368 [bridge] [ 936.269689] dev_hard_start_xmit+0x94/0x158 [ 936.273876] __dev_queue_xmit+0x5ac/0x7f8 [ 936.277890] dev_queue_xmit+0x10/0x18 [ 936.281563] neigh_resolve_output+0xec/0x198 [ 936.285845] ip6_finish_output2+0x240/0x710 [ 936.290039] __ip6_finish_output+0x130/0x170 [ 936.294318] ip6_output+0x6c/0x1c8 [ 936.297731] NF_HOOK.constprop.0+0xd8/0xe8 [ 936.301834] igmp6_send+0x358/0x558 [ 936.305326] igmp6_join_group.part.0+0x30/0xf0 [ 936.309774] igmp6_group_added+0xfc/0x110 [ 936.313787] __ipv6_dev_mc_inc+0x1a4/0x290 [ 936.317885] ipv6_dev_mc_inc+0x10/0x18 [ 936.321677] br_multicast_open+0xbc/0x110 [bridge] [ 936.326506] br_multicast_toggle+0xec/0x140 [bridge] Fixes: 4effd28c1245 ("bridge: join all-snoopers multicast address") Signed-off-by: Joseph Huang <Joseph.Huang@garmin.com> Acked-by: Nikolay Aleksandrov <nikolay@nvidia.com> Link: https://lore.kernel.org/r/20201204235628.50653-1-Joseph.Huang@garmin.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2020-12-05 02:56:28 +03:00
void br_multicast_leave_snoopers(struct net_bridge *br)
{
br_ip4_multicast_leave_snoopers(br);
br_ip6_multicast_leave_snoopers(br);
}
static void __br_multicast_open_query(struct net_bridge *br,
struct bridge_mcast_own_query *query)
{
query->startup_sent = 0;
if (!br_opt_get(br, BROPT_MULTICAST_ENABLED))
return;
mod_timer(&query->timer, jiffies);
}
static void __br_multicast_open(struct net_bridge_mcast *brmctx)
{
__br_multicast_open_query(brmctx->br, &brmctx->ip4_own_query);
#if IS_ENABLED(CONFIG_IPV6)
__br_multicast_open_query(brmctx->br, &brmctx->ip6_own_query);
#endif
}
void br_multicast_open(struct net_bridge *br)
{
ASSERT_RTNL();
if (br_opt_get(br, BROPT_MCAST_VLAN_SNOOPING_ENABLED)) {
struct net_bridge_vlan_group *vg;
struct net_bridge_vlan *vlan;
vg = br_vlan_group(br);
if (vg) {
list_for_each_entry(vlan, &vg->vlan_list, vlist) {
struct net_bridge_mcast *brmctx;
brmctx = &vlan->br_mcast_ctx;
if (br_vlan_is_brentry(vlan) &&
!br_multicast_ctx_vlan_disabled(brmctx))
__br_multicast_open(&vlan->br_mcast_ctx);
}
}
}
__br_multicast_open(&br->multicast_ctx);
}
static void __br_multicast_stop(struct net_bridge_mcast *brmctx)
{
del_timer_sync(&brmctx->ip4_mc_router_timer);
del_timer_sync(&brmctx->ip4_other_query.timer);
del_timer_sync(&brmctx->ip4_own_query.timer);
#if IS_ENABLED(CONFIG_IPV6)
del_timer_sync(&brmctx->ip6_mc_router_timer);
del_timer_sync(&brmctx->ip6_other_query.timer);
del_timer_sync(&brmctx->ip6_own_query.timer);
#endif
}
void br_multicast_toggle_one_vlan(struct net_bridge_vlan *vlan, bool on)
{
struct net_bridge *br;
/* it's okay to check for the flag without the multicast lock because it
* can only change under RTNL -> multicast_lock, we need the latter to
* sync with timers and packets
*/
if (on == !!(vlan->priv_flags & BR_VLFLAG_MCAST_ENABLED))
return;
if (br_vlan_is_master(vlan)) {
br = vlan->br;
if (!br_vlan_is_brentry(vlan) ||
(on &&
br_multicast_ctx_vlan_global_disabled(&vlan->br_mcast_ctx)))
return;
spin_lock_bh(&br->multicast_lock);
vlan->priv_flags ^= BR_VLFLAG_MCAST_ENABLED;
spin_unlock_bh(&br->multicast_lock);
if (on)
__br_multicast_open(&vlan->br_mcast_ctx);
else
__br_multicast_stop(&vlan->br_mcast_ctx);
} else {
struct net_bridge_mcast *brmctx;
brmctx = br_multicast_port_ctx_get_global(&vlan->port_mcast_ctx);
if (on && br_multicast_ctx_vlan_global_disabled(brmctx))
return;
br = vlan->port->br;
spin_lock_bh(&br->multicast_lock);
vlan->priv_flags ^= BR_VLFLAG_MCAST_ENABLED;
if (on)
__br_multicast_enable_port_ctx(&vlan->port_mcast_ctx);
else
__br_multicast_disable_port_ctx(&vlan->port_mcast_ctx);
spin_unlock_bh(&br->multicast_lock);
}
}
void br_multicast_toggle_vlan(struct net_bridge_vlan *vlan, bool on)
{
struct net_bridge_port *p;
if (WARN_ON_ONCE(!br_vlan_is_master(vlan)))
return;
list_for_each_entry(p, &vlan->br->port_list, list) {
struct net_bridge_vlan *vport;
vport = br_vlan_find(nbp_vlan_group(p), vlan->vid);
if (!vport)
continue;
br_multicast_toggle_one_vlan(vport, on);
}
}
int br_multicast_toggle_vlan_snooping(struct net_bridge *br, bool on,
struct netlink_ext_ack *extack)
{
struct net_bridge_vlan_group *vg;
struct net_bridge_vlan *vlan;
struct net_bridge_port *p;
if (br_opt_get(br, BROPT_MCAST_VLAN_SNOOPING_ENABLED) == on)
return 0;
if (on && !br_opt_get(br, BROPT_VLAN_ENABLED)) {
NL_SET_ERR_MSG_MOD(extack, "Cannot enable multicast vlan snooping with vlan filtering disabled");
return -EINVAL;
}
vg = br_vlan_group(br);
if (!vg)
return 0;
br_opt_toggle(br, BROPT_MCAST_VLAN_SNOOPING_ENABLED, on);
/* disable/enable non-vlan mcast contexts based on vlan snooping */
if (on)
__br_multicast_stop(&br->multicast_ctx);
else
__br_multicast_open(&br->multicast_ctx);
list_for_each_entry(p, &br->port_list, list) {
if (on)
br_multicast_disable_port(p);
else
br_multicast_enable_port(p);
}
list_for_each_entry(vlan, &vg->vlan_list, vlist)
br_multicast_toggle_vlan(vlan, on);
return 0;
}
void br_multicast_stop(struct net_bridge *br)
{
ASSERT_RTNL();
if (br_opt_get(br, BROPT_MCAST_VLAN_SNOOPING_ENABLED)) {
struct net_bridge_vlan_group *vg;
struct net_bridge_vlan *vlan;
vg = br_vlan_group(br);
if (vg) {
list_for_each_entry(vlan, &vg->vlan_list, vlist) {
struct net_bridge_mcast *brmctx;
brmctx = &vlan->br_mcast_ctx;
if (br_vlan_is_brentry(vlan) &&
!br_multicast_ctx_vlan_disabled(brmctx))
__br_multicast_stop(&vlan->br_mcast_ctx);
}
}
}
__br_multicast_stop(&br->multicast_ctx);
}
void br_multicast_dev_del(struct net_bridge *br)
{
struct net_bridge_mdb_entry *mp;
HLIST_HEAD(deleted_head);
struct hlist_node *tmp;
spin_lock_bh(&br->multicast_lock);
hlist_for_each_entry_safe(mp, tmp, &br->mdb_list, mdb_node)
br_multicast_del_mdb_entry(mp);
hlist_move_list(&br->mcast_gc_list, &deleted_head);
spin_unlock_bh(&br->multicast_lock);
br_multicast_ctx_deinit(&br->multicast_ctx);
br_multicast_gc(&deleted_head);
cancel_work_sync(&br->mcast_gc_work);
rcu_barrier();
}
int br_multicast_set_router(struct net_bridge *br, unsigned long val)
{
struct net_bridge_mcast *brmctx = &br->multicast_ctx;
int err = -EINVAL;
spin_lock_bh(&br->multicast_lock);
switch (val) {
case MDB_RTR_TYPE_DISABLED:
case MDB_RTR_TYPE_PERM:
br_mc_router_state_change(br, val == MDB_RTR_TYPE_PERM);
del_timer(&brmctx->ip4_mc_router_timer);
#if IS_ENABLED(CONFIG_IPV6)
del_timer(&brmctx->ip6_mc_router_timer);
#endif
brmctx->multicast_router = val;
err = 0;
break;
case MDB_RTR_TYPE_TEMP_QUERY:
if (brmctx->multicast_router != MDB_RTR_TYPE_TEMP_QUERY)
br_mc_router_state_change(br, false);
brmctx->multicast_router = val;
err = 0;
break;
}
spin_unlock_bh(&br->multicast_lock);
return err;
}
static void
br_multicast_rport_del_notify(struct net_bridge_mcast_port *pmctx, bool deleted)
{
if (!deleted)
return;
/* For backwards compatibility for now, only notify if there is
* no multicast router anymore for both IPv4 and IPv6.
*/
if (!hlist_unhashed(&pmctx->ip4_rlist))
return;
#if IS_ENABLED(CONFIG_IPV6)
if (!hlist_unhashed(&pmctx->ip6_rlist))
return;
#endif
br_rtr_notify(pmctx->port->br->dev, pmctx->port, RTM_DELMDB);
br_port_mc_router_state_change(pmctx->port, false);
/* don't allow timer refresh */
if (pmctx->multicast_router == MDB_RTR_TYPE_TEMP)
pmctx->multicast_router = MDB_RTR_TYPE_TEMP_QUERY;
}
int br_multicast_set_port_router(struct net_bridge_port *p, unsigned long val)
{
struct net_bridge_mcast *brmctx = &p->br->multicast_ctx;
struct net_bridge_mcast_port *pmctx = &p->multicast_ctx;
unsigned long now = jiffies;
int err = -EINVAL;
bool del = false;
spin_lock(&p->br->multicast_lock);
if (pmctx->multicast_router == val) {
/* Refresh the temp router port timer */
if (pmctx->multicast_router == MDB_RTR_TYPE_TEMP) {
mod_timer(&pmctx->ip4_mc_router_timer,
now + brmctx->multicast_querier_interval);
#if IS_ENABLED(CONFIG_IPV6)
mod_timer(&pmctx->ip6_mc_router_timer,
now + brmctx->multicast_querier_interval);
#endif
}
err = 0;
goto unlock;
}
switch (val) {
case MDB_RTR_TYPE_DISABLED:
pmctx->multicast_router = MDB_RTR_TYPE_DISABLED;
del |= br_ip4_multicast_rport_del(pmctx);
del_timer(&pmctx->ip4_mc_router_timer);
del |= br_ip6_multicast_rport_del(pmctx);
#if IS_ENABLED(CONFIG_IPV6)
del_timer(&pmctx->ip6_mc_router_timer);
#endif
br_multicast_rport_del_notify(pmctx, del);
break;
case MDB_RTR_TYPE_TEMP_QUERY:
pmctx->multicast_router = MDB_RTR_TYPE_TEMP_QUERY;
del |= br_ip4_multicast_rport_del(pmctx);
del |= br_ip6_multicast_rport_del(pmctx);
br_multicast_rport_del_notify(pmctx, del);
break;
case MDB_RTR_TYPE_PERM:
pmctx->multicast_router = MDB_RTR_TYPE_PERM;
del_timer(&pmctx->ip4_mc_router_timer);
br_ip4_multicast_add_router(brmctx, pmctx);
#if IS_ENABLED(CONFIG_IPV6)
del_timer(&pmctx->ip6_mc_router_timer);
#endif
br_ip6_multicast_add_router(brmctx, pmctx);
break;
case MDB_RTR_TYPE_TEMP:
pmctx->multicast_router = MDB_RTR_TYPE_TEMP;
br_ip4_multicast_mark_router(brmctx, pmctx);
br_ip6_multicast_mark_router(brmctx, pmctx);
break;
default:
goto unlock;
}
err = 0;
unlock:
spin_unlock(&p->br->multicast_lock);
return err;
}
static void br_multicast_start_querier(struct net_bridge_mcast *brmctx,
struct bridge_mcast_own_query *query)
{
struct net_bridge_port *port;
__br_multicast_open_query(brmctx->br, query);
rcu_read_lock();
list_for_each_entry_rcu(port, &brmctx->br->port_list, list) {
if (port->state == BR_STATE_DISABLED ||
port->state == BR_STATE_BLOCKING)
continue;
if (query == &brmctx->ip4_own_query)
br_multicast_enable(&port->multicast_ctx.ip4_own_query);
#if IS_ENABLED(CONFIG_IPV6)
else
br_multicast_enable(&port->multicast_ctx.ip6_own_query);
#endif
}
rcu_read_unlock();
}
int br_multicast_toggle(struct net_bridge *br, unsigned long val,
struct netlink_ext_ack *extack)
{
bridge: multicast: restore perm router ports on multicast enable Satish reported a problem with the perm multicast router ports not getting reenabled after some series of events, in particular if it happens that the multicast snooping has been disabled and the port goes to disabled state then it will be deleted from the router port list, but if it moves into non-disabled state it will not be re-added because the mcast snooping is still disabled, and enabling snooping later does nothing. Here are the steps to reproduce, setup br0 with snooping enabled and eth1 added as a perm router (multicast_router = 2): 1. $ echo 0 > /sys/class/net/br0/bridge/multicast_snooping 2. $ ip l set eth1 down ^ This step deletes the interface from the router list 3. $ ip l set eth1 up ^ This step does not add it again because mcast snooping is disabled 4. $ echo 1 > /sys/class/net/br0/bridge/multicast_snooping 5. $ bridge -d -s mdb show <empty> At this point we have mcast enabled and eth1 as a perm router (value = 2) but it is not in the router list which is incorrect. After this change: 1. $ echo 0 > /sys/class/net/br0/bridge/multicast_snooping 2. $ ip l set eth1 down ^ This step deletes the interface from the router list 3. $ ip l set eth1 up ^ This step does not add it again because mcast snooping is disabled 4. $ echo 1 > /sys/class/net/br0/bridge/multicast_snooping 5. $ bridge -d -s mdb show router ports on br0: eth1 Note: we can directly do br_multicast_enable_port for all because the querier timer already has checks for the port state and will simply expire if it's in blocking/disabled. See the comment added by commit 9aa66382163e7 ("bridge: multicast: add a comment to br_port_state_selection about blocking state") Fixes: 561f1103a2b7 ("bridge: Add multicast_snooping sysfs toggle") Reported-by: Satish Ashok <sashok@cumulusnetworks.com> Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-10-18 19:09:48 +03:00
struct net_bridge_port *port;
bridge: Fix a deadlock when enabling multicast snooping When enabling multicast snooping, bridge module deadlocks on multicast_lock if 1) IPv6 is enabled, and 2) there is an existing querier on the same L2 network. The deadlock was caused by the following sequence: While holding the lock, br_multicast_open calls br_multicast_join_snoopers, which eventually causes IP stack to (attempt to) send out a Listener Report (in igmp6_join_group). Since the destination Ethernet address is a multicast address, br_dev_xmit feeds the packet back to the bridge via br_multicast_rcv, which in turn calls br_multicast_add_group, which then deadlocks on multicast_lock. The fix is to move the call br_multicast_join_snoopers outside of the critical section. This works since br_multicast_join_snoopers only deals with IP and does not modify any multicast data structures of the bridge, so there's no need to hold the lock. Steps to reproduce: 1. sysctl net.ipv6.conf.all.force_mld_version=1 2. have another querier 3. ip link set dev bridge type bridge mcast_snooping 0 && \ ip link set dev bridge type bridge mcast_snooping 1 < deadlock > A typical call trace looks like the following: [ 936.251495] _raw_spin_lock+0x5c/0x68 [ 936.255221] br_multicast_add_group+0x40/0x170 [bridge] [ 936.260491] br_multicast_rcv+0x7ac/0xe30 [bridge] [ 936.265322] br_dev_xmit+0x140/0x368 [bridge] [ 936.269689] dev_hard_start_xmit+0x94/0x158 [ 936.273876] __dev_queue_xmit+0x5ac/0x7f8 [ 936.277890] dev_queue_xmit+0x10/0x18 [ 936.281563] neigh_resolve_output+0xec/0x198 [ 936.285845] ip6_finish_output2+0x240/0x710 [ 936.290039] __ip6_finish_output+0x130/0x170 [ 936.294318] ip6_output+0x6c/0x1c8 [ 936.297731] NF_HOOK.constprop.0+0xd8/0xe8 [ 936.301834] igmp6_send+0x358/0x558 [ 936.305326] igmp6_join_group.part.0+0x30/0xf0 [ 936.309774] igmp6_group_added+0xfc/0x110 [ 936.313787] __ipv6_dev_mc_inc+0x1a4/0x290 [ 936.317885] ipv6_dev_mc_inc+0x10/0x18 [ 936.321677] br_multicast_open+0xbc/0x110 [bridge] [ 936.326506] br_multicast_toggle+0xec/0x140 [bridge] Fixes: 4effd28c1245 ("bridge: join all-snoopers multicast address") Signed-off-by: Joseph Huang <Joseph.Huang@garmin.com> Acked-by: Nikolay Aleksandrov <nikolay@nvidia.com> Link: https://lore.kernel.org/r/20201204235628.50653-1-Joseph.Huang@garmin.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2020-12-05 02:56:28 +03:00
bool change_snoopers = false;
int err = 0;
spin_lock_bh(&br->multicast_lock);
if (!!br_opt_get(br, BROPT_MULTICAST_ENABLED) == !!val)
goto unlock;
err = br_mc_disabled_update(br->dev, val, extack);
if (err == -EOPNOTSUPP)
err = 0;
if (err)
goto unlock;
br_opt_toggle(br, BROPT_MULTICAST_ENABLED, !!val);
if (!br_opt_get(br, BROPT_MULTICAST_ENABLED)) {
bridge: Fix a deadlock when enabling multicast snooping When enabling multicast snooping, bridge module deadlocks on multicast_lock if 1) IPv6 is enabled, and 2) there is an existing querier on the same L2 network. The deadlock was caused by the following sequence: While holding the lock, br_multicast_open calls br_multicast_join_snoopers, which eventually causes IP stack to (attempt to) send out a Listener Report (in igmp6_join_group). Since the destination Ethernet address is a multicast address, br_dev_xmit feeds the packet back to the bridge via br_multicast_rcv, which in turn calls br_multicast_add_group, which then deadlocks on multicast_lock. The fix is to move the call br_multicast_join_snoopers outside of the critical section. This works since br_multicast_join_snoopers only deals with IP and does not modify any multicast data structures of the bridge, so there's no need to hold the lock. Steps to reproduce: 1. sysctl net.ipv6.conf.all.force_mld_version=1 2. have another querier 3. ip link set dev bridge type bridge mcast_snooping 0 && \ ip link set dev bridge type bridge mcast_snooping 1 < deadlock > A typical call trace looks like the following: [ 936.251495] _raw_spin_lock+0x5c/0x68 [ 936.255221] br_multicast_add_group+0x40/0x170 [bridge] [ 936.260491] br_multicast_rcv+0x7ac/0xe30 [bridge] [ 936.265322] br_dev_xmit+0x140/0x368 [bridge] [ 936.269689] dev_hard_start_xmit+0x94/0x158 [ 936.273876] __dev_queue_xmit+0x5ac/0x7f8 [ 936.277890] dev_queue_xmit+0x10/0x18 [ 936.281563] neigh_resolve_output+0xec/0x198 [ 936.285845] ip6_finish_output2+0x240/0x710 [ 936.290039] __ip6_finish_output+0x130/0x170 [ 936.294318] ip6_output+0x6c/0x1c8 [ 936.297731] NF_HOOK.constprop.0+0xd8/0xe8 [ 936.301834] igmp6_send+0x358/0x558 [ 936.305326] igmp6_join_group.part.0+0x30/0xf0 [ 936.309774] igmp6_group_added+0xfc/0x110 [ 936.313787] __ipv6_dev_mc_inc+0x1a4/0x290 [ 936.317885] ipv6_dev_mc_inc+0x10/0x18 [ 936.321677] br_multicast_open+0xbc/0x110 [bridge] [ 936.326506] br_multicast_toggle+0xec/0x140 [bridge] Fixes: 4effd28c1245 ("bridge: join all-snoopers multicast address") Signed-off-by: Joseph Huang <Joseph.Huang@garmin.com> Acked-by: Nikolay Aleksandrov <nikolay@nvidia.com> Link: https://lore.kernel.org/r/20201204235628.50653-1-Joseph.Huang@garmin.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2020-12-05 02:56:28 +03:00
change_snoopers = true;
goto unlock;
}
if (!netif_running(br->dev))
goto unlock;
bridge: multicast: restore perm router ports on multicast enable Satish reported a problem with the perm multicast router ports not getting reenabled after some series of events, in particular if it happens that the multicast snooping has been disabled and the port goes to disabled state then it will be deleted from the router port list, but if it moves into non-disabled state it will not be re-added because the mcast snooping is still disabled, and enabling snooping later does nothing. Here are the steps to reproduce, setup br0 with snooping enabled and eth1 added as a perm router (multicast_router = 2): 1. $ echo 0 > /sys/class/net/br0/bridge/multicast_snooping 2. $ ip l set eth1 down ^ This step deletes the interface from the router list 3. $ ip l set eth1 up ^ This step does not add it again because mcast snooping is disabled 4. $ echo 1 > /sys/class/net/br0/bridge/multicast_snooping 5. $ bridge -d -s mdb show <empty> At this point we have mcast enabled and eth1 as a perm router (value = 2) but it is not in the router list which is incorrect. After this change: 1. $ echo 0 > /sys/class/net/br0/bridge/multicast_snooping 2. $ ip l set eth1 down ^ This step deletes the interface from the router list 3. $ ip l set eth1 up ^ This step does not add it again because mcast snooping is disabled 4. $ echo 1 > /sys/class/net/br0/bridge/multicast_snooping 5. $ bridge -d -s mdb show router ports on br0: eth1 Note: we can directly do br_multicast_enable_port for all because the querier timer already has checks for the port state and will simply expire if it's in blocking/disabled. See the comment added by commit 9aa66382163e7 ("bridge: multicast: add a comment to br_port_state_selection about blocking state") Fixes: 561f1103a2b7 ("bridge: Add multicast_snooping sysfs toggle") Reported-by: Satish Ashok <sashok@cumulusnetworks.com> Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-10-18 19:09:48 +03:00
br_multicast_open(br);
list_for_each_entry(port, &br->port_list, list)
__br_multicast_enable_port_ctx(&port->multicast_ctx);
bridge: Fix a deadlock when enabling multicast snooping When enabling multicast snooping, bridge module deadlocks on multicast_lock if 1) IPv6 is enabled, and 2) there is an existing querier on the same L2 network. The deadlock was caused by the following sequence: While holding the lock, br_multicast_open calls br_multicast_join_snoopers, which eventually causes IP stack to (attempt to) send out a Listener Report (in igmp6_join_group). Since the destination Ethernet address is a multicast address, br_dev_xmit feeds the packet back to the bridge via br_multicast_rcv, which in turn calls br_multicast_add_group, which then deadlocks on multicast_lock. The fix is to move the call br_multicast_join_snoopers outside of the critical section. This works since br_multicast_join_snoopers only deals with IP and does not modify any multicast data structures of the bridge, so there's no need to hold the lock. Steps to reproduce: 1. sysctl net.ipv6.conf.all.force_mld_version=1 2. have another querier 3. ip link set dev bridge type bridge mcast_snooping 0 && \ ip link set dev bridge type bridge mcast_snooping 1 < deadlock > A typical call trace looks like the following: [ 936.251495] _raw_spin_lock+0x5c/0x68 [ 936.255221] br_multicast_add_group+0x40/0x170 [bridge] [ 936.260491] br_multicast_rcv+0x7ac/0xe30 [bridge] [ 936.265322] br_dev_xmit+0x140/0x368 [bridge] [ 936.269689] dev_hard_start_xmit+0x94/0x158 [ 936.273876] __dev_queue_xmit+0x5ac/0x7f8 [ 936.277890] dev_queue_xmit+0x10/0x18 [ 936.281563] neigh_resolve_output+0xec/0x198 [ 936.285845] ip6_finish_output2+0x240/0x710 [ 936.290039] __ip6_finish_output+0x130/0x170 [ 936.294318] ip6_output+0x6c/0x1c8 [ 936.297731] NF_HOOK.constprop.0+0xd8/0xe8 [ 936.301834] igmp6_send+0x358/0x558 [ 936.305326] igmp6_join_group.part.0+0x30/0xf0 [ 936.309774] igmp6_group_added+0xfc/0x110 [ 936.313787] __ipv6_dev_mc_inc+0x1a4/0x290 [ 936.317885] ipv6_dev_mc_inc+0x10/0x18 [ 936.321677] br_multicast_open+0xbc/0x110 [bridge] [ 936.326506] br_multicast_toggle+0xec/0x140 [bridge] Fixes: 4effd28c1245 ("bridge: join all-snoopers multicast address") Signed-off-by: Joseph Huang <Joseph.Huang@garmin.com> Acked-by: Nikolay Aleksandrov <nikolay@nvidia.com> Link: https://lore.kernel.org/r/20201204235628.50653-1-Joseph.Huang@garmin.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2020-12-05 02:56:28 +03:00
change_snoopers = true;
unlock:
spin_unlock_bh(&br->multicast_lock);
bridge: Fix a deadlock when enabling multicast snooping When enabling multicast snooping, bridge module deadlocks on multicast_lock if 1) IPv6 is enabled, and 2) there is an existing querier on the same L2 network. The deadlock was caused by the following sequence: While holding the lock, br_multicast_open calls br_multicast_join_snoopers, which eventually causes IP stack to (attempt to) send out a Listener Report (in igmp6_join_group). Since the destination Ethernet address is a multicast address, br_dev_xmit feeds the packet back to the bridge via br_multicast_rcv, which in turn calls br_multicast_add_group, which then deadlocks on multicast_lock. The fix is to move the call br_multicast_join_snoopers outside of the critical section. This works since br_multicast_join_snoopers only deals with IP and does not modify any multicast data structures of the bridge, so there's no need to hold the lock. Steps to reproduce: 1. sysctl net.ipv6.conf.all.force_mld_version=1 2. have another querier 3. ip link set dev bridge type bridge mcast_snooping 0 && \ ip link set dev bridge type bridge mcast_snooping 1 < deadlock > A typical call trace looks like the following: [ 936.251495] _raw_spin_lock+0x5c/0x68 [ 936.255221] br_multicast_add_group+0x40/0x170 [bridge] [ 936.260491] br_multicast_rcv+0x7ac/0xe30 [bridge] [ 936.265322] br_dev_xmit+0x140/0x368 [bridge] [ 936.269689] dev_hard_start_xmit+0x94/0x158 [ 936.273876] __dev_queue_xmit+0x5ac/0x7f8 [ 936.277890] dev_queue_xmit+0x10/0x18 [ 936.281563] neigh_resolve_output+0xec/0x198 [ 936.285845] ip6_finish_output2+0x240/0x710 [ 936.290039] __ip6_finish_output+0x130/0x170 [ 936.294318] ip6_output+0x6c/0x1c8 [ 936.297731] NF_HOOK.constprop.0+0xd8/0xe8 [ 936.301834] igmp6_send+0x358/0x558 [ 936.305326] igmp6_join_group.part.0+0x30/0xf0 [ 936.309774] igmp6_group_added+0xfc/0x110 [ 936.313787] __ipv6_dev_mc_inc+0x1a4/0x290 [ 936.317885] ipv6_dev_mc_inc+0x10/0x18 [ 936.321677] br_multicast_open+0xbc/0x110 [bridge] [ 936.326506] br_multicast_toggle+0xec/0x140 [bridge] Fixes: 4effd28c1245 ("bridge: join all-snoopers multicast address") Signed-off-by: Joseph Huang <Joseph.Huang@garmin.com> Acked-by: Nikolay Aleksandrov <nikolay@nvidia.com> Link: https://lore.kernel.org/r/20201204235628.50653-1-Joseph.Huang@garmin.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2020-12-05 02:56:28 +03:00
/* br_multicast_join_snoopers has the potential to cause
* an MLD Report/Leave to be delivered to br_multicast_rcv,
* which would in turn call br_multicast_add_group, which would
* attempt to acquire multicast_lock. This function should be
* called after the lock has been released to avoid deadlocks on
* multicast_lock.
*
* br_multicast_leave_snoopers does not have the problem since
* br_multicast_rcv first checks BROPT_MULTICAST_ENABLED, and
* returns without calling br_multicast_ipv4/6_rcv if it's not
* enabled. Moved both functions out just for symmetry.
*/
if (change_snoopers) {
if (br_opt_get(br, BROPT_MULTICAST_ENABLED))
br_multicast_join_snoopers(br);
else
br_multicast_leave_snoopers(br);
}
return err;
}
bool br_multicast_enabled(const struct net_device *dev)
{
struct net_bridge *br = netdev_priv(dev);
return !!br_opt_get(br, BROPT_MULTICAST_ENABLED);
}
EXPORT_SYMBOL_GPL(br_multicast_enabled);
bool br_multicast_router(const struct net_device *dev)
{
struct net_bridge *br = netdev_priv(dev);
bool is_router;
spin_lock_bh(&br->multicast_lock);
is_router = br_multicast_is_router(&br->multicast_ctx, NULL);
spin_unlock_bh(&br->multicast_lock);
return is_router;
}
EXPORT_SYMBOL_GPL(br_multicast_router);
int br_multicast_set_querier(struct net_bridge *br, unsigned long val)
{
struct net_bridge_mcast *brmctx = &br->multicast_ctx;
unsigned long max_delay;
val = !!val;
spin_lock_bh(&br->multicast_lock);
if (br_opt_get(br, BROPT_MULTICAST_QUERIER) == val)
goto unlock;
br_opt_toggle(br, BROPT_MULTICAST_QUERIER, !!val);
if (!val)
goto unlock;
max_delay = brmctx->multicast_query_response_interval;
if (!timer_pending(&brmctx->ip4_other_query.timer))
brmctx->ip4_other_query.delay_time = jiffies + max_delay;
br_multicast_start_querier(brmctx, &brmctx->ip4_own_query);
#if IS_ENABLED(CONFIG_IPV6)
if (!timer_pending(&brmctx->ip6_other_query.timer))
brmctx->ip6_other_query.delay_time = jiffies + max_delay;
br_multicast_start_querier(brmctx, &brmctx->ip6_own_query);
#endif
unlock:
spin_unlock_bh(&br->multicast_lock);
return 0;
}
int br_multicast_set_igmp_version(struct net_bridge *br, unsigned long val)
{
/* Currently we support only version 2 and 3 */
switch (val) {
case 2:
case 3:
break;
default:
return -EINVAL;
}
spin_lock_bh(&br->multicast_lock);
br->multicast_ctx.multicast_igmp_version = val;
spin_unlock_bh(&br->multicast_lock);
return 0;
}
#if IS_ENABLED(CONFIG_IPV6)
int br_multicast_set_mld_version(struct net_bridge *br, unsigned long val)
{
/* Currently we support version 1 and 2 */
switch (val) {
case 1:
case 2:
break;
default:
return -EINVAL;
}
spin_lock_bh(&br->multicast_lock);
br->multicast_ctx.multicast_mld_version = val;
spin_unlock_bh(&br->multicast_lock);
return 0;
}
#endif
/**
* br_multicast_list_adjacent - Returns snooped multicast addresses
* @dev: The bridge port adjacent to which to retrieve addresses
* @br_ip_list: The list to store found, snooped multicast IP addresses in
*
* Creates a list of IP addresses (struct br_ip_list) sensed by the multicast
* snooping feature on all bridge ports of dev's bridge device, excluding
* the addresses from dev itself.
*
* Returns the number of items added to br_ip_list.
*
* Notes:
* - br_ip_list needs to be initialized by caller
* - br_ip_list might contain duplicates in the end
* (needs to be taken care of by caller)
* - br_ip_list needs to be freed by caller
*/
int br_multicast_list_adjacent(struct net_device *dev,
struct list_head *br_ip_list)
{
struct net_bridge *br;
struct net_bridge_port *port;
struct net_bridge_port_group *group;
struct br_ip_list *entry;
int count = 0;
rcu_read_lock();
if (!br_ip_list || !netif_is_bridge_port(dev))
goto unlock;
port = br_port_get_rcu(dev);
if (!port || !port->br)
goto unlock;
br = port->br;
list_for_each_entry_rcu(port, &br->port_list, list) {
if (!port->dev || port->dev == dev)
continue;
hlist_for_each_entry_rcu(group, &port->mglist, mglist) {
entry = kmalloc(sizeof(*entry), GFP_ATOMIC);
if (!entry)
goto unlock;
entry->addr = group->key.addr;
list_add(&entry->list, br_ip_list);
count++;
}
}
unlock:
rcu_read_unlock();
return count;
}
EXPORT_SYMBOL_GPL(br_multicast_list_adjacent);
/**
* br_multicast_has_querier_anywhere - Checks for a querier on a bridge
* @dev: The bridge port providing the bridge on which to check for a querier
* @proto: The protocol family to check for: IGMP -> ETH_P_IP, MLD -> ETH_P_IPV6
*
* Checks whether the given interface has a bridge on top and if so returns
* true if a valid querier exists anywhere on the bridged link layer.
* Otherwise returns false.
*/
bool br_multicast_has_querier_anywhere(struct net_device *dev, int proto)
{
struct net_bridge *br;
struct net_bridge_port *port;
struct ethhdr eth;
bool ret = false;
rcu_read_lock();
if (!netif_is_bridge_port(dev))
goto unlock;
port = br_port_get_rcu(dev);
if (!port || !port->br)
goto unlock;
br = port->br;
memset(&eth, 0, sizeof(eth));
eth.h_proto = htons(proto);
ret = br_multicast_querier_exists(&br->multicast_ctx, &eth, NULL);
unlock:
rcu_read_unlock();
return ret;
}
EXPORT_SYMBOL_GPL(br_multicast_has_querier_anywhere);
/**
* br_multicast_has_querier_adjacent - Checks for a querier behind a bridge port
* @dev: The bridge port adjacent to which to check for a querier
* @proto: The protocol family to check for: IGMP -> ETH_P_IP, MLD -> ETH_P_IPV6
*
* Checks whether the given interface has a bridge on top and if so returns
* true if a selected querier is behind one of the other ports of this
* bridge. Otherwise returns false.
*/
bool br_multicast_has_querier_adjacent(struct net_device *dev, int proto)
{
struct net_bridge_mcast *brmctx;
struct net_bridge *br;
struct net_bridge_port *port;
bool ret = false;
rcu_read_lock();
if (!netif_is_bridge_port(dev))
goto unlock;
port = br_port_get_rcu(dev);
if (!port || !port->br)
goto unlock;
br = port->br;
brmctx = &br->multicast_ctx;
switch (proto) {
case ETH_P_IP:
if (!timer_pending(&brmctx->ip4_other_query.timer) ||
rcu_dereference(brmctx->ip4_querier.port) == port)
goto unlock;
break;
#if IS_ENABLED(CONFIG_IPV6)
case ETH_P_IPV6:
if (!timer_pending(&brmctx->ip6_other_query.timer) ||
rcu_dereference(brmctx->ip6_querier.port) == port)
goto unlock;
break;
#endif
default:
goto unlock;
}
ret = true;
unlock:
rcu_read_unlock();
return ret;
}
EXPORT_SYMBOL_GPL(br_multicast_has_querier_adjacent);
/**
* br_multicast_has_router_adjacent - Checks for a router behind a bridge port
* @dev: The bridge port adjacent to which to check for a multicast router
* @proto: The protocol family to check for: IGMP -> ETH_P_IP, MLD -> ETH_P_IPV6
*
* Checks whether the given interface has a bridge on top and if so returns
* true if a multicast router is behind one of the other ports of this
* bridge. Otherwise returns false.
*/
bool br_multicast_has_router_adjacent(struct net_device *dev, int proto)
{
struct net_bridge_mcast_port *pmctx;
struct net_bridge_mcast *brmctx;
struct net_bridge_port *port;
bool ret = false;
rcu_read_lock();
port = br_port_get_check_rcu(dev);
if (!port)
goto unlock;
brmctx = &port->br->multicast_ctx;
switch (proto) {
case ETH_P_IP:
hlist_for_each_entry_rcu(pmctx, &brmctx->ip4_mc_router_list,
ip4_rlist) {
if (pmctx->port == port)
continue;
ret = true;
goto unlock;
}
break;
#if IS_ENABLED(CONFIG_IPV6)
case ETH_P_IPV6:
hlist_for_each_entry_rcu(pmctx, &brmctx->ip6_mc_router_list,
ip6_rlist) {
if (pmctx->port == port)
continue;
ret = true;
goto unlock;
}
break;
#endif
default:
/* when compiled without IPv6 support, be conservative and
* always assume presence of an IPv6 multicast router
*/
ret = true;
}
unlock:
rcu_read_unlock();
return ret;
}
EXPORT_SYMBOL_GPL(br_multicast_has_router_adjacent);
static void br_mcast_stats_add(struct bridge_mcast_stats __percpu *stats,
const struct sk_buff *skb, u8 type, u8 dir)
{
struct bridge_mcast_stats *pstats = this_cpu_ptr(stats);
__be16 proto = skb->protocol;
unsigned int t_len;
u64_stats_update_begin(&pstats->syncp);
switch (proto) {
case htons(ETH_P_IP):
t_len = ntohs(ip_hdr(skb)->tot_len) - ip_hdrlen(skb);
switch (type) {
case IGMP_HOST_MEMBERSHIP_REPORT:
pstats->mstats.igmp_v1reports[dir]++;
break;
case IGMPV2_HOST_MEMBERSHIP_REPORT:
pstats->mstats.igmp_v2reports[dir]++;
break;
case IGMPV3_HOST_MEMBERSHIP_REPORT:
pstats->mstats.igmp_v3reports[dir]++;
break;
case IGMP_HOST_MEMBERSHIP_QUERY:
if (t_len != sizeof(struct igmphdr)) {
pstats->mstats.igmp_v3queries[dir]++;
} else {
unsigned int offset = skb_transport_offset(skb);
struct igmphdr *ih, _ihdr;
ih = skb_header_pointer(skb, offset,
sizeof(_ihdr), &_ihdr);
if (!ih)
break;
if (!ih->code)
pstats->mstats.igmp_v1queries[dir]++;
else
pstats->mstats.igmp_v2queries[dir]++;
}
break;
case IGMP_HOST_LEAVE_MESSAGE:
pstats->mstats.igmp_leaves[dir]++;
break;
}
break;
#if IS_ENABLED(CONFIG_IPV6)
case htons(ETH_P_IPV6):
t_len = ntohs(ipv6_hdr(skb)->payload_len) +
sizeof(struct ipv6hdr);
t_len -= skb_network_header_len(skb);
switch (type) {
case ICMPV6_MGM_REPORT:
pstats->mstats.mld_v1reports[dir]++;
break;
case ICMPV6_MLD2_REPORT:
pstats->mstats.mld_v2reports[dir]++;
break;
case ICMPV6_MGM_QUERY:
if (t_len != sizeof(struct mld_msg))
pstats->mstats.mld_v2queries[dir]++;
else
pstats->mstats.mld_v1queries[dir]++;
break;
case ICMPV6_MGM_REDUCTION:
pstats->mstats.mld_leaves[dir]++;
break;
}
break;
#endif /* CONFIG_IPV6 */
}
u64_stats_update_end(&pstats->syncp);
}
void br_multicast_count(struct net_bridge *br,
const struct net_bridge_port *p,
const struct sk_buff *skb, u8 type, u8 dir)
{
struct bridge_mcast_stats __percpu *stats;
/* if multicast_disabled is true then igmp type can't be set */
if (!type || !br_opt_get(br, BROPT_MULTICAST_STATS_ENABLED))
return;
if (p)
stats = p->mcast_stats;
else
stats = br->mcast_stats;
if (WARN_ON(!stats))
return;
br_mcast_stats_add(stats, skb, type, dir);
}
int br_multicast_init_stats(struct net_bridge *br)
{
br->mcast_stats = netdev_alloc_pcpu_stats(struct bridge_mcast_stats);
if (!br->mcast_stats)
return -ENOMEM;
return 0;
}
void br_multicast_uninit_stats(struct net_bridge *br)
{
free_percpu(br->mcast_stats);
}
/* noinline for https://bugs.llvm.org/show_bug.cgi?id=45802#c9 */
static noinline_for_stack void mcast_stats_add_dir(u64 *dst, u64 *src)
{
dst[BR_MCAST_DIR_RX] += src[BR_MCAST_DIR_RX];
dst[BR_MCAST_DIR_TX] += src[BR_MCAST_DIR_TX];
}
void br_multicast_get_stats(const struct net_bridge *br,
const struct net_bridge_port *p,
struct br_mcast_stats *dest)
{
struct bridge_mcast_stats __percpu *stats;
struct br_mcast_stats tdst;
int i;
memset(dest, 0, sizeof(*dest));
if (p)
stats = p->mcast_stats;
else
stats = br->mcast_stats;
if (WARN_ON(!stats))
return;
memset(&tdst, 0, sizeof(tdst));
for_each_possible_cpu(i) {
struct bridge_mcast_stats *cpu_stats = per_cpu_ptr(stats, i);
struct br_mcast_stats temp;
unsigned int start;
do {
start = u64_stats_fetch_begin_irq(&cpu_stats->syncp);
memcpy(&temp, &cpu_stats->mstats, sizeof(temp));
} while (u64_stats_fetch_retry_irq(&cpu_stats->syncp, start));
mcast_stats_add_dir(tdst.igmp_v1queries, temp.igmp_v1queries);
mcast_stats_add_dir(tdst.igmp_v2queries, temp.igmp_v2queries);
mcast_stats_add_dir(tdst.igmp_v3queries, temp.igmp_v3queries);
mcast_stats_add_dir(tdst.igmp_leaves, temp.igmp_leaves);
mcast_stats_add_dir(tdst.igmp_v1reports, temp.igmp_v1reports);
mcast_stats_add_dir(tdst.igmp_v2reports, temp.igmp_v2reports);
mcast_stats_add_dir(tdst.igmp_v3reports, temp.igmp_v3reports);
tdst.igmp_parse_errors += temp.igmp_parse_errors;
mcast_stats_add_dir(tdst.mld_v1queries, temp.mld_v1queries);
mcast_stats_add_dir(tdst.mld_v2queries, temp.mld_v2queries);
mcast_stats_add_dir(tdst.mld_leaves, temp.mld_leaves);
mcast_stats_add_dir(tdst.mld_v1reports, temp.mld_v1reports);
mcast_stats_add_dir(tdst.mld_v2reports, temp.mld_v2reports);
tdst.mld_parse_errors += temp.mld_parse_errors;
}
memcpy(dest, &tdst, sizeof(*dest));
}
int br_mdb_hash_init(struct net_bridge *br)
{
int err;
err = rhashtable_init(&br->sg_port_tbl, &br_sg_port_rht_params);
if (err)
return err;
err = rhashtable_init(&br->mdb_hash_tbl, &br_mdb_rht_params);
if (err) {
rhashtable_destroy(&br->sg_port_tbl);
return err;
}
return 0;
}
void br_mdb_hash_fini(struct net_bridge *br)
{
rhashtable_destroy(&br->sg_port_tbl);
rhashtable_destroy(&br->mdb_hash_tbl);
}