[PATCH AUTOSEL 4.19 64/73] vhost/vsock: fix reset orphans race with close timeout

From: Sasha Levin
Date: Wed Dec 12 2018 - 23:46:33 EST


From: Stefan Hajnoczi <stefanha@xxxxxxxxxx>

[ Upstream commit c38f57da428b033f2721b611d84b1f40bde674a8 ]

If a local process has closed a connected socket and hasn't received a
RST packet yet, then the socket remains in the table until a timeout
expires.

When a vhost_vsock instance is released with the timeout still pending,
the socket is never freed because vhost_vsock has already set the
SOCK_DONE flag.

Check if the close timer is pending and let it close the socket. This
prevents the race which can leak sockets.

Reported-by: Maximilian Riemensberger <riemensberger@xxxxxxxxxx>
Cc: Graham Whaley <graham.whaley@xxxxxxxxx>
Signed-off-by: Stefan Hajnoczi <stefanha@xxxxxxxxxx>
Signed-off-by: Michael S. Tsirkin <mst@xxxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
drivers/vhost/vsock.c | 22 +++++++++++++++-------
1 file changed, 15 insertions(+), 7 deletions(-)

diff --git a/drivers/vhost/vsock.c b/drivers/vhost/vsock.c
index 34bc3ab40c6d..731e2ea2aeca 100644
--- a/drivers/vhost/vsock.c
+++ b/drivers/vhost/vsock.c
@@ -563,13 +563,21 @@ static void vhost_vsock_reset_orphans(struct sock *sk)
* executing.
*/

- if (!vhost_vsock_get(vsk->remote_addr.svm_cid)) {
- sock_set_flag(sk, SOCK_DONE);
- vsk->peer_shutdown = SHUTDOWN_MASK;
- sk->sk_state = SS_UNCONNECTED;
- sk->sk_err = ECONNRESET;
- sk->sk_error_report(sk);
- }
+ /* If the peer is still valid, no need to reset connection */
+ if (vhost_vsock_get(vsk->remote_addr.svm_cid))
+ return;
+
+ /* If the close timeout is pending, let it expire. This avoids races
+ * with the timeout callback.
+ */
+ if (vsk->close_work_scheduled)
+ return;
+
+ sock_set_flag(sk, SOCK_DONE);
+ vsk->peer_shutdown = SHUTDOWN_MASK;
+ sk->sk_state = SS_UNCONNECTED;
+ sk->sk_err = ECONNRESET;
+ sk->sk_error_report(sk);
}

static int vhost_vsock_dev_release(struct inode *inode, struct file *file)
--
2.19.1