[SCSI] zfcp: fix device registration issues

The patch fixes following issues:

(1) Replace scsi_add_device with scsi_scan_target.
(Thus the rport instead of the scsi_host becomes parent of a
scsi_target again.)

(2) Avoid scsi_device allocation during registration of an remote port.
(Would be done during fc_scsi_scan_rport.)

(3) Fix queuecommand behaviour when an zfcp unit is blocked.
(Call scsi_done with DID_NO_CONNECT instead of returning
SCSI_MLQUEUE_DEVICE_BUSY otherwise we might end up waiting
for completion in blk_execute_rq for ever.)

Signed-off-by: Andreas Herrmann <aherrman@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>

1 files changed, 7 insertions, 4 deletions

diff --git a/drivers/s390/scsi/zfcp_erp.c b/drivers/s390/scsi/zfcp_erp.c
index e3c4bdd29a60..57cb628a05aa 100644
--- a/drivers/s390/scsi/zfcp_erp.c
+++ b/drivers/s390/scsi/zfcp_erp.c
@@ -3391,10 +3391,13 @@ zfcp_erp_action_cleanup(int action, struct zfcp_adapter *adapter,
                     && (!atomic_test_mask(ZFCP_STATUS_UNIT_TEMPORARY,
                                           &unit->status))
                     && !unit->device
-                    && port->rport)
-                        scsi_add_device(port->adapter->scsi_host, 0,
-                                        port->rport->scsi_target_id,
-                                        unit->scsi_lun);
+                    && port->rport) {
+                        atomic_set_mask(ZFCP_STATUS_UNIT_REGISTERED,
+                                        &unit->status);
+                        scsi_scan_target(&port->rport->dev, 0,
+                                         port->rport->scsi_target_id,
+                                         unit->scsi_lun, 0);
+                }
                 zfcp_unit_put(unit);
                 break;
         case ZFCP_ERP_ACTION_REOPEN_PORT_FORCED: