nvme-pci: remove volatile cqes

The completion queue entry is not volatile once the phase is confirmed.
Remove the volatile keywords and check the phase using the appropriate
READ_ONCE() accessor, allowing the compiler to optimize the remaining
completion path.
Signed-off-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>

drivers/nvme/host/pci.c

@@ -166,7 +166,7 @@ struct nvme_queue {
 	void *sq_cmds;
 	 /* only used for poll queues: */
 	spinlock_t cq_poll_lock ____cacheline_aligned_in_smp;
-	volatile struct nvme_completion *cqes;
+	struct nvme_completion *cqes;
 	dma_addr_t sq_dma_addr;
 	dma_addr_t cq_dma_addr;
 	u32 __iomem *q_db;
@@ -922,8 +922,9 @@ static void nvme_pci_complete_rq(struct request *req)
 /* We read the CQE phase first to check if the rest of the entry is valid */
 static inline bool nvme_cqe_pending(struct nvme_queue *nvmeq)
 {
-	return (le16_to_cpu(nvmeq->cqes[nvmeq->cq_head].status) & 1) ==
-			nvmeq->cq_phase;
+	struct nvme_completion *hcqe = &nvmeq->cqes[nvmeq->cq_head];
+
+	return (le16_to_cpu(READ_ONCE(hcqe->status)) & 1) == nvmeq->cq_phase;
 }
 
 static inline void nvme_ring_cq_doorbell(struct nvme_queue *nvmeq)
@@ -944,7 +945,7 @@ static inline struct blk_mq_tags *nvme_queue_tagset(struct nvme_queue *nvmeq)
 
 static inline void nvme_handle_cqe(struct nvme_queue *nvmeq, u16 idx)
 {
-	volatile struct nvme_completion *cqe = &nvmeq->cqes[idx];
+	struct nvme_completion *cqe = &nvmeq->cqes[idx];
 	struct request *req;
 
 	if (unlikely(cqe->command_id >= nvmeq->q_depth)) {