Fix scsi device starvation handling

Doing some debugging after the SDET failure, I found that
if we reject a command at zero depth, we can get into the
infinite loop where we refuse a command to a LUN because it
is starved, but there's no returning command to run the SCSI
queues and fix the starvation.

The solution is to accept the command and remove the LUN from
the starved list.

drivers/scsi/scsi_lib.c

@@ -1095,14 +1095,17 @@ static inline int scsi_host_queue_ready(struct request_queue *q,
 			return 0;
 		}
 	}
-	if (!list_empty(&sdev->starved_entry))
-		return 0;
 	if ((shost->can_queue > 0 && shost->host_busy >= shost->can_queue) ||
 	    shost->host_blocked || shost->host_self_blocked) {
+		if (list_empty(&sdev->starved_entry))
 			list_add_tail(&sdev->starved_entry, &shost->starved_list);
 		return 0;
 	}
 
+	/* We're OK to process the command, so we can't be starved */
+	if (!list_empty(&sdev->starved_entry))
+		list_del_init(&sdev->starved_entry);
+
 	return 1;
 }
 
@@ -1129,6 +1132,7 @@ static void scsi_request_fn(struct request_queue *q)
 	 * the host is no longer able to accept any more requests.
 	 */
 	while (!blk_queue_plugged(q)) {
+		int rtn;
 		/*
 		 * get next queueable request.  We do this early to make sure
 		 * that the request is fully prepared even if we cannot 
@@ -1182,8 +1186,17 @@ static void scsi_request_fn(struct request_queue *q)
 		/*
 		 * Dispatch the command to the low-level driver.
 		 */
-		scsi_dispatch_cmd(cmd);
+		rtn = scsi_dispatch_cmd(cmd);
 		spin_lock_irq(q->queue_lock);
+		if(rtn) {
+			/* we're refusing the command; because of
+			 * the way locks get dropped, we need to 
+			 * check here if plugging is required */
+			if(sdev->device_busy == 0)
+				blk_plug_device(q);
+
+			break;
+		}
 	}
 
 	return;
@@ -1204,6 +1217,8 @@ static void scsi_request_fn(struct request_queue *q)
 		blk_queue_end_tag(q, req);
 	__elv_add_request(q, req, 0, 0);
 	sdev->device_busy--;
+	if(sdev->device_busy == 0)
+		blk_plug_device(q);
 }
 
 u64 scsi_calculate_bounce_limit(struct Scsi_Host *shost)