From: "Jérémy Jean" <Jeremy.Jean@oss.cyber.gouv.fr>
To: Tung Quang Nguyen <tung.quang.nguyen@est.tech>,
Jon Maloy <jmaloy@redhat.com>
Cc: netdev@vger.kernel.org, tipc-discussion@lists.sourceforge.net,
linux-kernel@vger.kernel.org,
"Jérémy Jean" <Jeremy.Jean@oss.cyber.gouv.fr>,
stable@vger.kernel.org
Subject: [PATCH] tipc: fix use-after-free in netns teardown
Date: Sat, 3 Oct 2026 19:25:31 +0000 [thread overview]
Message-ID: <20261003192530.2414682-2-Jeremy.Jean@oss.cyber.gouv.fr> (raw)
conn_put() decreases the reference count of a connection, and when it
reaches zero, tipc_conn_kref_release() removes the connection from
conn_idr and frees it. Removing the entry is protected by idr_lock,
but decreasing the count is not.
During netns dismantle, it may happen that a connection gets its count
dropped to zero by conn_put() while tipc_topsrv_stop() still holds the
lock idr_lock. tipc_topsrv_stop() calls conn_get() on that connection
whose count is already zero, and then releases idr_lock.
tipc_conn_kref_release() can then free the connection before
tipc_conn_close() uses it. KASAN reports the UAF as:
BUG: KASAN: slab-use-after-free in tipc_conn_close (net/tipc/topsrv.c:158)
Read of size 8 at addr ff1100000cf4f408 by task kworker/u16:3/70
Stop accepting connections before walking conn_idr. Use
kref_get_unless_zero() to avoid taking references when the count is zero.
Retry from the first entry, releasing idr_lock and calling cond_resched()
between iterations so pending work can finish.
Fixes: 667eeab4999e ("tipc: Fix use-after-free in tipc_conn_close().")
Cc: stable@vger.kernel.org
Assisted-by: LLM
Signed-off-by: Jérémy Jean <Jeremy.Jean@oss.cyber.gouv.fr>
---
net/tipc/topsrv.c | 17 ++++++++++++-----
1 file changed, 12 insertions(+), 5 deletions(-)
diff --git a/net/tipc/topsrv.c b/net/tipc/topsrv.c
index af530c9ed8402..9ab2349f71190 100644
--- a/net/tipc/topsrv.c
+++ b/net/tipc/topsrv.c
@@ -700,16 +700,23 @@ static void tipc_topsrv_stop(struct net *net)
struct tipc_conn *con;
int id;
+ write_lock_bh(&lsock->sk->sk_callback_lock);
+ lsock->sk->sk_user_data = NULL;
+ write_unlock_bh(&lsock->sk->sk_callback_lock);
+ cancel_work_sync(&srv->awork);
+
spin_lock_bh(&srv->idr_lock);
- for (id = 0; srv->idr_in_use; id++) {
- con = idr_find(&srv->conn_idr, id);
+ for (id = 0; srv->idr_in_use; id = 0) {
+ con = idr_get_next(&srv->conn_idr, &id);
+ if (con && !kref_get_unless_zero(&con->kref))
+ con = NULL;
+ spin_unlock_bh(&srv->idr_lock);
if (con) {
- conn_get(con);
- spin_unlock_bh(&srv->idr_lock);
tipc_conn_close(con);
conn_put(con);
- spin_lock_bh(&srv->idr_lock);
}
+ cond_resched();
+ spin_lock_bh(&srv->idr_lock);
}
__module_get(lsock->ops->owner);
__module_get(lsock->sk->sk_prot_creator->owner);
--
2.47.3
next reply other threads:[~2026-10-03 19:26 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-10-03 19:25 Jérémy Jean [this message]
2026-10-03 19:29 ` [PATCH] tipc: fix use-after-free in netns teardown netdev-bot+sinfo
2026-10-04 19:27 ` netdev-bot+sashiko
2026-10-04 20:45 ` Jérémy Jean
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20261003192530.2414682-2-Jeremy.Jean@oss.cyber.gouv.fr \
--to=jeremy.jean@oss.cyber.gouv.fr \
--cc=jmaloy@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=stable@vger.kernel.org \
--cc=tipc-discussion@lists.sourceforge.net \
--cc=tung.quang.nguyen@est.tech \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox