The vhost_user_read_cb() and rte_vhost_driver_unregister()
can be called at the same time by 2 threads. Eg thread1
calls vhost_user_read_cb() and removes the vsocket from
conn_list, then thread2 calls rte_vhost_driver_unregister()
and frees the vsocket since it is NOT in the conn_list.
So thread1 will access invalid memory when trying to
reconnect.
The fix is to move the "removing of vsocket from conn_list"
to end of the vhost_user_read_cb(), then avoid the race
condition.
The core trace is:
Program terminated with signal 11, Segmentation fault.
Fixes:
af1475918124 ("vhost: introduce API to start a specific driver")
Cc: stable@dpdk.org
Signed-off-by: Zhike Wang <wangzhike@jd.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
vhost_destroy_device(conn->vid);
+ if (vsocket->reconnect) {
+ create_unix_socket(vsocket);
+ vhost_user_start_client(vsocket);
+ }
+
pthread_mutex_lock(&vsocket->conn_mutex);
TAILQ_REMOVE(&vsocket->conn_list, conn, next);
pthread_mutex_unlock(&vsocket->conn_mutex);
free(conn);
-
- if (vsocket->reconnect) {
- create_unix_socket(vsocket);
- vhost_user_start_client(vsocket);
- }
}
}