[PATCH v5 5/6] nvme-fc: use nvme_ctrl_reconnect to decide reconnect attempts

[Daniel Wagner]

The fc transport handles the DNR for reconnect correctly, though it
ignores all the negative error code returned by the LLDD. Thus, use the
nvme_ctrl_reconnect helper to ensure to have consistent behavior for all
transports.

Signed-off-by: Daniel Wagner <dwagner at suse.de>
---
 drivers/nvme/host/fc.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/drivers/nvme/host/fc.c b/drivers/nvme/host/fc.c
index a5b29e9ad342..a58f08304459 100644
--- a/drivers/nvme/host/fc.c
+++ b/drivers/nvme/host/fc.c
@@ -3301,7 +3301,7 @@ nvme_fc_reconnect_or_delete(struct nvme_fc_ctrl *ctrl, int status)
 	struct nvme_fc_rport *rport = ctrl->rport;
 	struct nvme_fc_remote_port *portptr = &rport->remoteport;
 	unsigned long recon_delay = ctrl->ctrl.opts->reconnect_delay * HZ;
-	bool recon = true;
+	bool recon = nvme_ctrl_reconnect(status);
 
 	if (nvme_ctrl_state(&ctrl->ctrl) != NVME_CTRL_CONNECTING)
 		return;
@@ -3310,8 +3310,6 @@ nvme_fc_reconnect_or_delete(struct nvme_fc_ctrl *ctrl, int status)
 		dev_info(ctrl->ctrl.device,
 			"NVME-FC{%d}: reset: Reconnect attempt failed (%d)\n",
 			ctrl->cnum, status);
-		if (status > 0 && (status & NVME_SC_DNR))
-			recon = false;
 	} else if (time_after_eq(jiffies, rport->dev_loss_end))
 		recon = false;
 
-- 
2.44.0