[PATCH 16/17] nvme: implement multipath access to nvme subsystems

[Sagi Grimberg]

>   static inline bool nvme_req_needs_retry(struct request *req)
>   {
>   	if (blk_noretry_request(req))
> @@ -143,6 +204,11 @@ static inline bool nvme_req_needs_retry(struct request *req)
>   void nvme_complete_rq(struct request *req)
>   {
>   	if (unlikely(nvme_req(req)->status && nvme_req_needs_retry(req))) {
> +		if (nvme_req_needs_failover(req)) {
> +			nvme_failover_req(req);
> +			return;
> +		}
> +
>   		nvme_req(req)->retries++;
>   		blk_mq_requeue_request(req, true);
>   		return;

Nit, consider having the !nvme_req_needs_failover() case in an else
clause, just a suggestion though.



> @@ -171,6 +237,18 @@ void nvme_cancel_request(struct request *req, void *data, bool reserved)
>   }
>   EXPORT_SYMBOL_GPL(nvme_cancel_request);

Question,

Does the statement in nvme_cancel_request:

         if (blk_queue_dying(req->q))
                 status |= NVME_SC_DNR;

still holds?