--- /dev/null
+From: Gerald Schaefer <geraldsc@de.ibm.com>
+Subject: zfcp: deadlock between scheduled task, ERP and scan_work
+References: bnc#487755,LTC#52845
+
+Symptom: I/O stalls due to a pending ERP action.
+Problem: Under certain timing constraints a deadlock can prevent
+ the ERP to finish. The ERP is trying to force the wka
+ port offline waiting for the scheduled work to finish.
+ The scheduled work is waiting for the refcount to
+ turn zero and another task is waiting for the ERP to
+ finish before descrementing the refcounter. ->deadlock
+Solution: Don't wait for the refcount to turn zero. If required
+ the wka port offline function is called again later.
+
+Acked-by: John Jolly <jjolly@suse.de>
+
+---
+ drivers/s390/scsi/zfcp_fc.c | 3 ---
+ 1 file changed, 3 deletions(-)
+
+Index: linux-sles11/drivers/s390/scsi/zfcp_fc.c
+===================================================================
+--- linux-sles11.orig/drivers/s390/scsi/zfcp_fc.c
++++ linux-sles11/drivers/s390/scsi/zfcp_fc.c
+@@ -75,9 +75,6 @@ static void zfcp_wka_port_offline(struct
+ struct zfcp_wka_port *wka_port =
+ container_of(dw, struct zfcp_wka_port, work);
+
+- wait_event(wka_port->completion_wq,
+- atomic_read(&wka_port->refcount) == 0);
+-
+ mutex_lock(&wka_port->mutex);
+ if ((atomic_read(&wka_port->refcount) != 0) ||
+ (wka_port->status != ZFCP_WKA_PORT_ONLINE))