nvmet-tcp: fix potential race of tcp socket closing accept_work

[ Upstream commit 0fbcfb089a3f2f2a731d01f0aec8f7697a849c28 ]

When we accept a TCP connection and allocate an nvmet-tcp queue we should
make sure not to fully establish it or reference it as the connection may
be already closing, which triggers queue release work, which does not
fence against queue establishment.

In order to address such a race, we make sure to check the sk_state and
contain the queue reference to be done underneath the sk_callback_lock
such that the queue release work correctly fences against it.

Fixes: 872d26a391da ("nvmet-tcp: add NVMe over TCP target driver")
Reported-by: Elad Grupi <elad.grupi@dell.com>
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

diff --git a/drivers/nvme/target/tcp.c b/drivers/nvme/target/tcp.c
index 577ce7d403ae92c41bccac8b489d1ca67c154162..8b0485ada315b9f139e3e42b8131d70ea466b92f 100644 (file)
--- a/drivers/nvme/target/tcp.c
+++ b/drivers/nvme/target/tcp.c
@@ -1485,17 +1485,27 @@ static int nvmet_tcp_set_queue_sock(struct nvmet_tcp_queue *queue)
        if (inet->rcv_tos > 0)
                ip_sock_set_tos(sock->sk, inet->rcv_tos);
 
+       ret = 0;
        write_lock_bh(&sock->sk->sk_callback_lock);
-       sock->sk->sk_user_data = queue;
-       queue->data_ready = sock->sk->sk_data_ready;
-       sock->sk->sk_data_ready = nvmet_tcp_data_ready;
-       queue->state_change = sock->sk->sk_state_change;
-       sock->sk->sk_state_change = nvmet_tcp_state_change;
-       queue->write_space = sock->sk->sk_write_space;
-       sock->sk->sk_write_space = nvmet_tcp_write_space;
+       if (sock->sk->sk_state != TCP_ESTABLISHED) {
+               /*
+                * If the socket is already closing, don't even start
+                * consuming it
+                */
+               ret = -ENOTCONN;
+       } else {
+               sock->sk->sk_user_data = queue;
+               queue->data_ready = sock->sk->sk_data_ready;
+               sock->sk->sk_data_ready = nvmet_tcp_data_ready;
+               queue->state_change = sock->sk->sk_state_change;
+               sock->sk->sk_state_change = nvmet_tcp_state_change;
+               queue->write_space = sock->sk->sk_write_space;
+               sock->sk->sk_write_space = nvmet_tcp_write_space;
+               queue_work_on(queue_cpu(queue), nvmet_tcp_wq, &queue->io_work);
+       }
        write_unlock_bh(&sock->sk->sk_callback_lock);
 
-       return 0;
+       return ret;
 }
 
 static int nvmet_tcp_alloc_queue(struct nvmet_tcp_port *port,
@@ -1543,8 +1553,6 @@ static int nvmet_tcp_alloc_queue(struct nvmet_tcp_port *port,
        if (ret)
                goto out_destroy_sq;
 
-       queue_work_on(queue_cpu(queue), nvmet_tcp_wq, &queue->io_work);
-
        return 0;
 out_destroy_sq:
        mutex_lock(&nvmet_tcp_queue_mutex);