[PATCH] iscsi: conn error (1020) each time iscsi session logout

From: Vaughan Cao
Date: Wed Dec 18 2013 - 01:45:05 EST


We do a normal login/logout process to iscsi server. iscsiadm report success,
but we always see the following error just before conn shutdown in dmesg.

Oct 15 05:30:09 vmhodtest019 iscsid: Connection1:0 to [target:
iqn.1986-03.com.sun:02:7b863a18-045a-cb04-c686-841f17df2f9c, portal:
10.182.32.162,3260] through [iface: default] is operational now
Oct 15 05:30:42 vmhodtest019 kernel: connection1:0: detected conn error
(1020)
Oct 15 05:30:42 vmhodtest019 iscsid: Connection1:0 to [target:
iqn.1986-03.com.sun:02:7b863a18-045a-cb04-c686-841f17df2f9c, portal:
10.182.32.162,3260] through [iface: default] is shutdown.

It's because iscsi_tcp module evaluates socket state in data_ready() callback,
and that detect the socket close. However, this socket close on target peer
is in response to the logout request from initiator. So this is not an error
that should be reported out. I quiesce it by checking session state and err
value accordingly.

Signed-off-by: Vaughan Cao <vaughan.cao@xxxxxxxxxx>
---
drivers/scsi/libiscsi.c | 6 ++++++
1 file changed, 6 insertions(+)

diff --git a/drivers/scsi/libiscsi.c b/drivers/scsi/libiscsi.c
index 415f2c0..84171ef 100644
--- a/drivers/scsi/libiscsi.c
+++ b/drivers/scsi/libiscsi.c
@@ -1360,6 +1360,12 @@ void iscsi_conn_failure(struct iscsi_conn *conn, enum iscsi_err err)
spin_unlock_bh(&session->lock);
return;
}
+ /* Target closed the connection in response to logout */
+ if (session->state == ISCSI_STATE_LOGGING_OUT &&
+ err == ISCSI_ERR_TCP_CONN_CLOSE) {
+ spin_unlock_bh(&session->lock);
+ return;
+ }

if (conn->stop_stage == 0)
session->state = ISCSI_STATE_FAILED;
--
1.8.3.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/