[PATCH v2 1/1] nvme-tcp: fix wrong status on deferred digest error

[Xixin Liu]

A C2HData digest error stores a host status code in req->status. Without
DATA_SUCCESS completion goes through the rsp path, which passed that host
value straight into complete as CQE Status wire encoding.

On the rsp path, shift left when a deferred host error is stored,
otherwise use cqe->status.

Fixes: 1ba2e507f55c ("nvme-tcp: Do not reset transport on data digest errors")
Signed-off-by: Xixin Liu <liuxixin at kylinos.cn>
---
 drivers/nvme/host/tcp.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

--- a/drivers/nvme/host/tcp.c
+++ b/drivers/nvme/host/tcp.c
@@ -617,6 +617,7 @@ static int nvme_tcp_process_nvme_cqe(struct nvme_tcp_queue *queue,
 {
 	struct nvme_tcp_request *req;
 	struct request *rq;
+	__le16 status;
 
 	rq = nvme_find_rq(nvme_tcp_tagset(queue), cqe->command_id);
 	if (!rq) {
@@ -628,10 +629,12 @@ static int nvme_tcp_process_nvme_cqe(struct nvme_tcp_queue *queue,
 	}
 
 	req = blk_mq_rq_to_pdu(rq);
 	if (req->status == cpu_to_le16(NVME_SC_SUCCESS))
-		req->status = cqe->status;
+		status = cqe->status;
+	else
+		status = cpu_to_le16(le16_to_cpu(req->status) << 1);
 
-	if (!nvme_try_complete_req(rq, req->status, cqe->result))
+	if (!nvme_try_complete_req(rq, status, cqe->result))
 		nvme_complete_rq(rq);
 	queue->nr_cqe++;
 
-- 
2.53.0