tipc: fix NULL deref in tipc_lxc_xmit() on node up
Subsystems (4)
PATCH v3 1/1 tipc: fix NULL deref in tipc_lxc_xmit() on node up
tipc_named_node_up() builds a bulk of this node's cluster-scope service bindings for a peer that just came up and sends it with tipc_node_xmit(). Neither step tolerates an empty skb chain: named_distribute() finishes with buf_msg(skb_peek_tail(list)) to tag the last message, and for a same-host peer tipc_node_xmit() routes into tipc_lxc_xmit(), which reads buf_msg(skb_peek(list)). skb_peek*() returns NULL on an empty chain, so buf_msg(NULL) faults.
The chain is empty in two cases: when cluster_scope itself is empty, and when named_distribute() bails out on an allocation failure without queueing any buffer. cluster_scope is legitimately empty during the window in tipc_net_finalize() between setting the node address, after which peers can link up and trigger tipc_named_node_up(), and tipc_nametbl_publish() inserting the first self-binding. A peer linking in that window crashes the node. It is reachable by an unprivileged user, who can gain CAP_NET_ADMIN in a private net namespace and drive TIPC there.
Oops: general protection fault, probably for non-canonical address KASAN: null-ptr-deref in range [0x00000000000000d8-0x00000000000000df] RIP: 0010:tipc_lxc_xmit (net/tipc/node.c:1629 net/tipc/msg.h:202) tipc_node_xmit (net/tipc/node.c:1718) tipc_named_node_up (net/tipc/name_distr.c:222) tipc_node_write_unlock (net/tipc/node.c:428) tipc_rcv (net/tipc/node.c:2185) tipc_l2_rcv_msg (net/tipc/bearer.c:669)
Distribute only when cluster_scope is non-empty, and send only when named_distribute() actually produced a buffer. An empty bulk carries no bindings, so not sending it changes nothing. tipc_node_xmit() does not touch cluster_scope, so move it out of cluster_scope_lock.
---
v3:
- Guard tipc_node_xmit() with skb_queue_empty(&head) so an empty chain from a named_distribute() allocation failure is not sent either, and move it out of cluster_scope_lock since it does not use cluster_scope (Tung Quang Nguyen).
v2:
- Guard in tipc_named_node_up() instead of inside named_distribute().
net/tipc/name_distr.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)
net/tipc/name_distr.c +5 -2
@@ -218,9 +218,12 @@ void tipc_named_node_up(struct net *net, u32 dnode, u16 capabilities)spin_unlock_bh (&tn->nametbl_lock);read_lock_bh (&nt->cluster_scope_lock); -named_distribute (net, &head, dnode, &nt->cluster_scope, seqno); -tipc_node_xmit (net, &head, dnode,0 ); +if (!list_empty (&nt->cluster_scope)) +named_distribute (net, &head, dnode, &nt->cluster_scope, seqno);read_unlock_bh (&nt->cluster_scope_lock); + +if (!skb_queue_empty (&head)) +tipc_node_xmit (net, &head, dnode,0 ); }/**
Signature · 1 line
2.43.0
Jakub Kicinski
maintainerQuoted 7 lines · On Thu, 9 Jul 2026 07:47:19 -0700 Weiming Shi wrote:
tipc_named_node_up()builds a bulk of this node's cluster-scope service bindings for a peer that just came up and sends it withtipc_node_xmit(). Neither step tolerates an empty skb chain:named_distribute()finishes with buf_msg(skb_peek_tail(list)) to tag the last message, and for a same-host peertipc_node_xmit()routes intotipc_lxc_xmit(), which reads buf_msg(skb_peek(list)). skb_peek*() returns NULL on an empty chain, sobuf_msg(NULL)faults.
This patch has been sitting in PW for 2 weeks, due to maintainers being on vacation. Please repost if it's still relevant.
Signature · 1 line
pw-bot: defer
Quoted 10 lines · Jakub Kicinski <[email protected]> 于2026年7月22日周三 02:59写道:
Quoted 7 lines · On Thu, 9 Jul 2026 07:47:19 -0700 Weiming Shi wrote:
tipc_named_node_up()builds a bulk of this node's cluster-scope service bindings for a peer that just came up and sends it withtipc_node_xmit(). Neither step tolerates an empty skb chain:named_distribute()finishes with buf_msg(skb_peek_tail(list)) to tag the last message, and for a same-host peertipc_node_xmit()routes intotipc_lxc_xmit(), which reads buf_msg(skb_peek(list)). skb_peek*() returns NULL on an empty chain, sobuf_msg(NULL)faults.
Quoted 4 lines
This patch has been sitting in PW for 2 weeks, due to maintainers being on vacation. Please repost if it's still relevant.
--
pw-bot: defer
Hi Jakub,
Thanks for the ping. This is already being actively worked in another thread:
No need to repost this v3 separatel.
Thanks,
Weiming