[PATCH 5.10 084/452] nbd: Fix hung on disconnect request if socket is closed before

From: Greg Kroah-Hartman
Date: Tue Jun 07 2022 - 13:28:09 EST


From: Xie Yongji <xieyongji@xxxxxxxxxxxxx>

[ Upstream commit 491bf8f236fdeec698fa6744993f1ecf3fafd1a5 ]

When userspace closes the socket before sending a disconnect
request, the following I/O requests will be blocked in
wait_for_reconnect() until dead timeout. This will cause the
following disconnect request also hung on blk_mq_quiesce_queue().
That means we have no way to disconnect a nbd device if there
are some I/O requests waiting for reconnecting until dead timeout.
It's not expected. So let's wake up the thread waiting for
reconnecting directly when a disconnect request is sent.

Reported-by: Xu Jianhai <zero.xu@xxxxxxxxxxxxx>
Signed-off-by: Xie Yongji <xieyongji@xxxxxxxxxxxxx>
Reviewed-by: Josef Bacik <josef@xxxxxxxxxxxxxx>
Link: https://lore.kernel.org/r/20220322080639.142-1-xieyongji@xxxxxxxxxxxxx
Signed-off-by: Jens Axboe <axboe@xxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
drivers/block/nbd.c | 13 +++++++++----
1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c
index 59c452fff835..ecde800ba210 100644
--- a/drivers/block/nbd.c
+++ b/drivers/block/nbd.c
@@ -880,11 +880,15 @@ static int wait_for_reconnect(struct nbd_device *nbd)
struct nbd_config *config = nbd->config;
if (!config->dead_conn_timeout)
return 0;
- if (test_bit(NBD_RT_DISCONNECTED, &config->runtime_flags))
+
+ if (!wait_event_timeout(config->conn_wait,
+ test_bit(NBD_RT_DISCONNECTED,
+ &config->runtime_flags) ||
+ atomic_read(&config->live_connections) > 0,
+ config->dead_conn_timeout))
return 0;
- return wait_event_timeout(config->conn_wait,
- atomic_read(&config->live_connections) > 0,
- config->dead_conn_timeout) > 0;
+
+ return !test_bit(NBD_RT_DISCONNECTED, &config->runtime_flags);
}

static int nbd_handle_cmd(struct nbd_cmd *cmd, int index)
@@ -2029,6 +2033,7 @@ static void nbd_disconnect_and_put(struct nbd_device *nbd)
mutex_lock(&nbd->config_lock);
nbd_disconnect(nbd);
sock_shutdown(nbd);
+ wake_up(&nbd->config->conn_wait);
/*
* Make sure recv thread has finished, so it does not drop the last
* config ref and try to destroy the workqueue from inside the work
--
2.35.1