[SCSI] ibmvfc: Fix Virtual I/O failover hang

If a Virtual I/O server fails in a dual virtual I/O server multipath
configuration, ensure we delete all remote ports so that path failover
can occur. For a single path configuration, the remote ports will
go into devloss state.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>

diff --git a/drivers/scsi/ibmvscsi/ibmvfc.c b/drivers/scsi/ibmvscsi/ibmvfc.c
index b7650613b8c2606caef87d83a6b0e2a74cc7ca09..bdfa223a7dbb9e047c278b6414c5135e49a47001 100644 (file)
--- a/drivers/scsi/ibmvscsi/ibmvfc.c
+++ b/drivers/scsi/ibmvscsi/ibmvfc.c
@@ -4306,8 +4306,8 @@ static void ibmvfc_do_work(struct ibmvfc_host *vhost)
                spin_lock_irqsave(vhost->host->host_lock, flags);
                if (rc == H_CLOSED)
                        vio_enable_interrupts(to_vio_dev(vhost->dev));
-               else if (rc || (rc = ibmvfc_send_crq_init(vhost)) ||
-                        (rc = vio_enable_interrupts(to_vio_dev(vhost->dev)))) {
+               if (rc || (rc = ibmvfc_send_crq_init(vhost)) ||
+                   (rc = vio_enable_interrupts(to_vio_dev(vhost->dev)))) {
                        ibmvfc_link_down(vhost, IBMVFC_LINK_DEAD);
                        dev_err(vhost->dev, "Error after reset (rc=%d)\n", rc);
                }