[SCSI] ibmvfc: Fix command completion handling

Commands which are completed by the VIOS are placed on a CRQ
in kernel memory for the ibmvfc driver to process. Each CRQ
entry is 16 bytes. The ibmvfc driver reads the first 8 bytes
to check if the entry is valid, then reads the next 8 bytes to get
the handle, which is a pointer the completed command. This fixes
an issue seen on Power 7 where the processor reordered the
loads from memory, resulting in processing command completion
with a stale handle. This could result in command timeouts,
and also early completion of commands.

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

drivers/scsi/ibmvscsi/ibmvfc.c

@@ -3013,6 +3013,7 @@ static struct ibmvfc_async_crq *ibmvfc_next_async_crq(struct ibmvfc_host *vhost)
 	if (crq->valid & 0x80) {
 		if (++async_crq->cur == async_crq->size)
 			async_crq->cur = 0;
+		rmb();
 	} else
 		crq = NULL;
 
@@ -3035,6 +3036,7 @@ static struct ibmvfc_crq *ibmvfc_next_crq(struct ibmvfc_host *vhost)
 	if (crq->valid & 0x80) {
 		if (++queue->cur == queue->size)
 			queue->cur = 0;
+		rmb();
 	} else
 		crq = NULL;
 
@@ -3083,12 +3085,14 @@ static void ibmvfc_tasklet(void *data)
 		while ((async = ibmvfc_next_async_crq(vhost)) != NULL) {
 			ibmvfc_handle_async(async, vhost);
 			async->valid = 0;
+			wmb();
 		}
 
 		/* Pull all the valid messages off the CRQ */
 		while ((crq = ibmvfc_next_crq(vhost)) != NULL) {
 			ibmvfc_handle_crq(crq, vhost);
 			crq->valid = 0;
+			wmb();
 		}
 
 		vio_enable_interrupts(vdev);
@@ -3096,10 +3100,12 @@ static void ibmvfc_tasklet(void *data)
 			vio_disable_interrupts(vdev);
 			ibmvfc_handle_async(async, vhost);
 			async->valid = 0;
+			wmb();
 		} else if ((crq = ibmvfc_next_crq(vhost)) != NULL) {
 			vio_disable_interrupts(vdev);
 			ibmvfc_handle_crq(crq, vhost);
 			crq->valid = 0;
+			wmb();
 		} else
 			done = 1;
 	}