Re: [PATCH net v4] tipc: serialize udp bearer replicast list updates
From: Weiming Shi
Date: Tue Jul 14 2026 - 15:23:11 EST
Tung Quang Nguyen <tung.quang.nguyen@xxxxxxxx> 于2026年7月13日周一 20:06写道:
>
> >Subject: [PATCH net v4] tipc: serialize udp bearer replicast list updates
> >
> >tipc_udp_rcast_add() and cleanup_bearer() both update ub->rcast.list with
> >list_add_rcu() / list_del_rcu(), but nothing serializes them. The add runs from
> >the encap receive softirq (via tipc_udp_rcast_disc()) without rtnl_lock(), so it
> >can race the cleanup delete and corrupt the list:
> >
> > list_del corruption. prev->next should be ffff8880298d7ab8,
> > but was ffff88802449ad38. (prev=ffff888027e3ec98)
> > kernel BUG at lib/list_debug.c:62!
> > RIP: __list_del_entry_valid_or_report+0x17a/0x200
> > Workqueue: events cleanup_bearer
> > Call Trace:
> > cleanup_bearer (net/tipc/udp_media.c:811)
> > process_one_work (kernel/workqueue.c:3302)
> > worker_thread (kernel/workqueue.c:3466)
> >
> >The bearer can be enabled from an unprivileged user namespace, as the
> >TIPCv2 generic-netlink ops carry no GENL_ADMIN_PERM.
> >
> >Add a spinlock to struct udp_bearer and take it around the list_add_rcu() in
> >tipc_udp_rcast_add() and the list_del_rcu() loop in cleanup_bearer() so the
> >two writers can no longer corrupt the list.
> >
> >Reject a duplicate peer under the same lock before allocating, and remove
> >tipc_udp_is_known_peer(). The old lockless pre-check in
> >tipc_udp_rcast_disc() was racy: two softirqs discovering the same peer could
> >both find it absent and add it twice. Doing the check under rcast_lock in
> >tipc_udp_rcast_add(), before the allocation, makes it the single point of truth
> >for both the discovery and the netlink add paths and keeps a flood of the same
> >address from churning the allocator.
> >
> >Fixes: ef20cd4dd163 ("tipc: introduce UDP replicast")
> >Reported-by: Xiang Mei <xmei5@xxxxxxx>
> >Suggested-by: Tung Nguyen <tung.quang.nguyen@xxxxxxxx>
> >Signed-off-by: Weiming Shi <bestswngs@xxxxxxxxx>
> >---
> >v4: (per Tung's review)
> > - Reject a duplicate under rcast_lock before allocating rcast, so a
> > flood of the same address can't churn the allocator.
> >v3:
> > - Check for a duplicate peer in tipc_udp_rcast_add() under rcast_lock
> > and remove tipc_udp_is_known_peer().
> >v2:
> > - Narrow the lock to the list mutation.
> >
> > net/tipc/udp_media.c | 53 ++++++++++++++++++++------------------------
> > 1 file changed, 24 insertions(+), 29 deletions(-)
> >
> >diff --git a/net/tipc/udp_media.c b/net/tipc/udp_media.c index
> >62ae7f5b5..3f0d75d72 100644
> >--- a/net/tipc/udp_media.c
> >+++ b/net/tipc/udp_media.c
> >@@ -94,6 +94,7 @@ struct udp_replicast {
> > * @ifindex: local address scope
> > * @work: used to schedule deferred work on a bearer
> > * @rcast: associated udp_replicast container
> >+ * @rcast_lock: serialize updates to @rcast.list against concurrent
> >updaters
> > */
> > struct udp_bearer {
> > struct tipc_bearer __rcu *bearer;
> >@@ -101,6 +102,7 @@ struct udp_bearer {
> > u32 ifindex;
> > struct work_struct work;
> > struct udp_replicast rcast;
> >+ spinlock_t rcast_lock; /* protects rcast.list */
> > };
> >
> > static int tipc_udp_is_mcast_addr(struct udp_media_addr *addr) @@ -278,26
> >+280,6 @@ static int tipc_udp_send_msg(struct net *net, struct sk_buff *skb,
> > return err;
> > }
> >
> >-static bool tipc_udp_is_known_peer(struct tipc_bearer *b,
> >- struct udp_media_addr *addr)
> >-{
> >- struct udp_replicast *rcast, *tmp;
> >- struct udp_bearer *ub;
> >-
> >- ub = rcu_dereference_rtnl(b->media_ptr);
> >- if (!ub) {
> >- pr_err_ratelimited("UDP bearer instance not found\n");
> >- return false;
> >- }
> >-
> >- list_for_each_entry_safe(rcast, tmp, &ub->rcast.list, list) {
> >- if (!memcmp(&rcast->addr, addr, sizeof(struct
> >udp_media_addr)))
> >- return true;
> >- }
> >-
> >- return false;
> >-}
> >-
> > static int tipc_udp_rcast_add(struct tipc_bearer *b,
> > struct udp_media_addr *addr)
> > {
> >@@ -308,16 +290,34 @@ static int tipc_udp_rcast_add(struct tipc_bearer *b,
> > if (!ub)
> > return -ENODEV;
> >
> >+ /* Serialize with other updaters and reject a duplicate before
> >+ * allocating, so a flood of the same address can't churn the
> >+ * allocator.
> >+ */
> >+ spin_lock_bh(&ub->rcast_lock);
> >+ list_for_each_entry(rcast, &ub->rcast.list, list) {
> >+ if (!memcmp(&rcast->addr, addr, sizeof(*addr))) {
> >+ spin_unlock_bh(&ub->rcast_lock);
> >+ return 0;
> >+ }
> >+ }
> >+
> > rcast = kmalloc_obj(*rcast, GFP_ATOMIC);
> >- if (!rcast)
> >+ if (!rcast) {
> >+ spin_unlock_bh(&ub->rcast_lock);
> > return -ENOMEM;
> >+ }
> >
> > if (dst_cache_init(&rcast->dst_cache, GFP_ATOMIC)) {
> >+ spin_unlock_bh(&ub->rcast_lock);
> > kfree(rcast);
> > return -ENOMEM;
> > }
> >
> > memcpy(&rcast->addr, addr, sizeof(struct udp_media_addr));
> >+ list_add_rcu(&rcast->list, &ub->rcast.list);
> >+ b->bcast_addr.broadcast = TIPC_REPLICAST_SUPPORT;
> >+ spin_unlock_bh(&ub->rcast_lock);
>
> sashiko reports potential memory leak when tipc_udp_nl_bearer_add() adds an element to the list right after cleanup_bearer() frees all element in the list.
> I think we need to add more code to handle this case as below:
>
> --- a/net/tipc/udp_media.c
> +++ b/net/tipc/udp_media.c
> @@ -95,6 +95,7 @@ struct udp_replicast {
> * @work: used to schedule deferred work on a bearer
> * @rcast: associated udp_replicast container
> * @rcast_lock: serialize updates to @rcast.list against concurrent updaters
> + * @bearer_disabled: flag to not add rcast to rcast.list if bearer was disabled
> */
> struct udp_bearer {
> struct tipc_bearer __rcu *bearer;
> @@ -103,6 +104,7 @@ struct udp_bearer {
> struct work_struct work;
> struct udp_replicast rcast;
> spinlock_t rcast_lock; /* protects rcast.list */
> + bool bearer_disabled;
> };
>
> static int tipc_udp_is_mcast_addr(struct udp_media_addr *addr)
> @@ -295,6 +297,10 @@ static int tipc_udp_rcast_add(struct tipc_bearer *b,
> * allocator.
> */
> spin_lock_bh(&ub->rcast_lock);
> + if (ub->bearer_disabled) {
> + spin_unlock_bh(&ub->rcast_lock);
> + return 0;
> + }
> list_for_each_entry(rcast, &ub->rcast.list, list) {
> if (!memcmp(&rcast->addr, addr, sizeof(*addr))) {
> spin_unlock_bh(&ub->rcast_lock);
> @@ -817,6 +823,7 @@ static void cleanup_bearer(struct work_struct *work)
> list_del_rcu(&rcast->list);
> call_rcu_hurry(&rcast->rcu, rcast_free_rcu);
> }
> + ub->bearer_disabled = true;
> spin_unlock_bh(&ub->rcast_lock);
>
> tn = tipc_net(sock_net(ub->sk));
>
> >
> > if (ntohs(addr->proto) == ETH_P_IP)
> > pr_info("New replicast peer: %pI4\n", &rcast->addr.ipv4); @@
> >-325,8 +325,6 @@ static int tipc_udp_rcast_add(struct tipc_bearer *b,
> > else if (ntohs(addr->proto) == ETH_P_IPV6)
> > pr_info("New replicast peer: %pI6\n", &rcast->addr.ipv6);
> >#endif
> >- b->bcast_addr.broadcast = TIPC_REPLICAST_SUPPORT;
> >- list_add_rcu(&rcast->list, &ub->rcast.list);
> > return 0;
> > }
> >
> >@@ -361,9 +359,6 @@ static int tipc_udp_rcast_disc(struct tipc_bearer *b,
> >struct sk_buff *skb)
> > return 0;
> > }
> >
> >- if (likely(tipc_udp_is_known_peer(b, &src)))
> >- return 0;
> >-
> > return tipc_udp_rcast_add(b, &src);
> > }
> >
> >@@ -644,9 +639,6 @@ int tipc_udp_nl_bearer_add(struct tipc_bearer *b,
> >struct nlattr *attr)
> > return -EINVAL;
> > }
> >
> >- if (tipc_udp_is_known_peer(b, &addr))
> >- return 0;
> >-
> > return tipc_udp_rcast_add(b, &addr);
> > }
> >
> >@@ -679,6 +671,7 @@ static int tipc_udp_enable(struct net *net, struct
> >tipc_bearer *b,
> > return -ENOMEM;
> >
> > INIT_LIST_HEAD(&ub->rcast.list);
> >+ spin_lock_init(&ub->rcast_lock);
> >
> > if (!attrs[TIPC_NLA_BEARER_UDP_OPTS])
> > goto err;
> >@@ -819,10 +812,12 @@ static void cleanup_bearer(struct work_struct
> >*work)
> > struct udp_replicast *rcast, *tmp;
> > struct tipc_net *tn;
> >
> >+ spin_lock_bh(&ub->rcast_lock);
> > list_for_each_entry_safe(rcast, tmp, &ub->rcast.list, list) {
> > list_del_rcu(&rcast->list);
> > call_rcu_hurry(&rcast->rcu, rcast_free_rcu);
> > }
> >+ spin_unlock_bh(&ub->rcast_lock);
> >
> > tn = tipc_net(sock_net(ub->sk));
> >
> >--
> >2.43.0
>
v5 sent. Thanks.