Skip to contentSkip to conversation

nbd: clear DISCONNECT_REQUESTED flag once disconnection occurs.

PATCH Applied [email protected] 3 participants 3 messages May 30, 2018

Subsystems (2)
Block layer Network block device (NBD)
lore
Share & tools
lore

[email protected]

author
#0 lore raw

PATCH 1/1 nbd: clear DISCONNECT_REQUESTED flag once disconnection occurs.

From: Kevin Vigor <[email protected]>

When a userspace client requests a NBD device be disconnected, the DISCONNECT_REQUESTED flag is set. While this flag is set, the driver will not inform userspace when a connection is closed.

Unfortunately the flag was never cleared, so once a disconnect was requested the driver would thereafter never tell userspace about a closed connection. Thus when connections failed due to timeout, no attempt to reconnect was made and eventually the device would fail.

Fix by clearing the DISCONNECT_REQUESTED flag (and setting the DISCONNECTED flag) once all connections are closed.

Changes relative to v1 (https://marc.info/?l=linux-block&m=152763540418902):

* remove pointless wake_up() -> wake_up_all() change.

Signed-off-by Kevin Vigor <[email protected]>
---
 drivers/block/nbd.c | 21 +++++++++++++++------
 1 file changed, 15 insertions(+), 6 deletions(-)
drivers/block/nbd.c +15 -6
index afbc202..1cd041b 100644
@@ -213,7 +213,15 @@ static void nbd_mark_nsock_dead(struct nbd_device *nbd, struct nbd_sock *nsock,
 	}
 	if (!nsock->dead) {
 		kernel_sock_shutdown(nsock->sock, SHUT_RDWR);
-		atomic_dec(&nbd->config->live_connections);
+		if (atomic_dec_return(&nbd->config->live_connections) == 0) {
+			if (test_and_clear_bit(NBD_DISCONNECT_REQUESTED,
+					       &nbd->config->runtime_flags)) {
+				set_bit(NBD_DISCONNECTED,
+					&nbd->config->runtime_flags);
+				dev_info(nbd_to_dev(nbd),
+					"Disconnected due to user request.\n");
+			}
+		}
 	}
 	nsock->dead = true;
 	nsock->pending = NULL;

@@ -292,7 +300,9 @@ static enum blk_eh_timer_return nbd_xmit_timeout(struct request *req,
 
 	if (config->num_connections > 1) {
 		dev_err_ratelimited(nbd_to_dev(nbd),
-				    "Connection timed out, retrying\n");
+				    "Connection timed out, retrying (%d/%d alive)\n",
+				    atomic_read(&config->live_connections),
+				    config->num_connections);
 		/*
 		 * Hooray we have more connections, requeue this IO, the submit
 		 * path will put it on a real connection.

@@ -714,10 +724,9 @@ static int wait_for_reconnect(struct nbd_device *nbd)
 		return 0;
 	if (test_bit(NBD_DISCONNECTED, &config->runtime_flags))
 		return 0;
-	wait_event_timeout(config->conn_wait,
-			   atomic_read(&config->live_connections),
-			   config->dead_conn_timeout);
-	return atomic_read(&config->live_connections);
+	return wait_event_timeout(config->conn_wait,
+				  atomic_read(&config->live_connections) > 0,
+				  config->dead_conn_timeout) > 0;
 }
 
 static int nbd_handle_cmd(struct nbd_cmd *cmd, int index)
Signature · 1 line
2.7.4
maintainer Reviewed-by on v1 1/1
#1 lore raw
Quoted 11 lines · On Wed, May 30, 2018 at 10:45:11AM -0600, [email protected] wrote:

From: Kevin Vigor <[email protected]>

When a userspace client requests a NBD device be disconnected, the DISCONNECT_REQUESTED flag is set. While this flag is set, the driver will not inform userspace when a connection is closed.

Unfortunately the flag was never cleared, so once a disconnect was requested the driver would thereafter never tell userspace about a closed connection. Thus when connections failed due to timeout, no attempt to reconnect was made and eventually the device would fail.

Quoted 7 lines

Fix by clearing the DISCONNECT_REQUESTED flag (and setting the DISCONNECTED flag) once all connections are closed.

Changes relative to v1 (https://marc.info/?l=linux-block&m=152763540418902):

* remove pointless wake_up() -> wake_up_all() change.

Usually you want to put the changelog after the --- bit below so we can all see it but it doesn't end up in the git changelog.

Reviewed-by Josef Bacik <[email protected]>

Thanks,

Josef

maintainer Applied
#2 lore raw
In reply to [email protected]
Quoted 6 lines · On 5/30/18 10:45 AM, [email protected] wrote:

From: Kevin Vigor <[email protected]>

When a userspace client requests a NBD device be disconnected, the DISCONNECT_REQUESTED flag is set. While this flag is set, the driver will not inform userspace when a connection is closed.

Quoted 7 lines

Unfortunately the flag was never cleared, so once a disconnect was requested the driver would thereafter never tell userspace about a closed connection. Thus when connections failed due to timeout, no attempt to reconnect was made and eventually the device would fail.

Fix by clearing the DISCONNECT_REQUESTED flag (and setting the DISCONNECTED flag) once all connections are closed.

Applied, thanks Kevin.

Signature · 1 line
Jens Axboe