@@ -785,6 +785,7 @@ static int bcm_delete_rx_op(struct list_head *ops, struct bcm_msg_head *mh,
bcm_rx_handler, op);
list_del(&op->list);
+ synchronize_rcu();
bcm_remove_op(op);
return 1; /* done */
}
@@ -1533,9 +1534,13 @@ static int bcm_release(struct socket *sock)
REGMASK(op->can_id),
bcm_rx_handler, op);
- bcm_remove_op(op);
}
+ synchronize_rcu();
+
+ list_for_each_entry_safe(op, next, &bo->rx_ops, list)
+ bcm_remove_op(op);
+
#if IS_ENABLED(CONFIG_PROC_FS)
/* remove procfs entry */
if (net->can.bcmproc_dir && bo->bcm_proc_read)
can_rx_register() callbacks may be called concurrently to the call to can_rx_unregister(). The callbacks and callback data, though, are protected by RCU and the struct sock reference count. So the callback data is really attached to the life of sk, meaning that it should be released on sk_destruct. However, bcm_remove_op() calls tasklet_kill(), and RCU callbacks may be called under RCU softirq, so that cannot be used on kernels before the introduction of HRTIMER_MODE_SOFT. However, bcm_rx_handler() is called under RCU protection, so after calling can_rx_unregister(), we may call synchronize_rcu() in order to wait for any RCU read-side critical sections to finish. That is, bcm_rx_handler() won't be called anymore for those ops. So, we only free them, after we do that synchronize_rcu(). Fixes: ffd980f976e7 ("[CAN]: Add broadcast manager (bcm) protocol") Link: https://lore.kernel.org/r/20210619161813.2098382-1-cascardo@canonical.com Cc: linux-stable <stable@vger.kernel.org> Reported-by: syzbot+0f7e7e5e2f4f40fa89c0@syzkaller.appspotmail.com Reported-by: Norbert Slusarek <nslusarek@gmx.net> Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@canonical.com> Acked-by: Oliver Hartkopp <socketcan@hartkopp.net> Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> (cherry picked from commit d5f9023fa61ee8b94f37a93f08e94b136cf1e463) CVE-2021-3609 Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@canonical.com> --- net/can/bcm.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)