NVMe: reap completion entries when deleting queue

Make sure that there are no unprocesssed entries on a completion
queue before deleting it, and check for validity of the CQ
door bell before writing completions to it.

This fixes problems with doing a sysfs reset of the device while
it's handling IO.

Tested-by: Jon Derrick <jonathan.derrick@intel.com>

Signed-off-by: Jens Axboe <axboe@fb.com>

drivers/nvme/host/pci.c

@@ -968,7 +968,8 @@ static void __nvme_process_cq(struct nvme_queue *nvmeq, unsigned int *tag)
 	if (head == nvmeq->cq_head && phase == nvmeq->cq_phase)
 		return;
 
-	writel(head, nvmeq->q_db + nvmeq->dev->db_stride);
+	if (likely(nvmeq->cq_vector >= 0))
+		writel(head, nvmeq->q_db + nvmeq->dev->db_stride);
 	nvmeq->cq_head = head;
 	nvmeq->cq_phase = phase;
 
@@ -2787,6 +2788,10 @@ static void nvme_del_queue_end(struct nvme_queue *nvmeq)
 {
 	struct nvme_delq_ctx *dq = nvmeq->cmdinfo.ctx;
 	nvme_put_dq(dq);
+
+	spin_lock_irq(&nvmeq->q_lock);
+	nvme_process_cq(nvmeq);
+	spin_unlock_irq(&nvmeq->q_lock);
 }
 
 static int adapter_async_del_queue(struct nvme_queue *nvmeq, u8 opcode,