[PATCH 2/2] nvme-multipath: add 'latency' iopolicy

[Hannes Reinecke]

Add an I/O policy for selecting paths based upon the least latency.

Signed-off-by: Hannes Reinecke <hare at suse.de>
---
 drivers/nvme/host/core.c      |  1 +
 drivers/nvme/host/multipath.c | 19 +++++++++++++++++--
 drivers/nvme/host/nvme.h      |  5 +++++
 3 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index b52e9c9bffd6..bd506284af1f 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -397,6 +397,7 @@ void nvme_complete_rq(struct request *req)
 
 	trace_nvme_complete_rq(req);
 	nvme_cleanup_cmd(req);
+	nvme_mpath_complete_request(req);
 
 	/*
 	 * Completions of long-running commands should not be able to
diff --git a/drivers/nvme/host/multipath.c b/drivers/nvme/host/multipath.c
index 7813608038bc..8052ba8085f0 100644
--- a/drivers/nvme/host/multipath.c
+++ b/drivers/nvme/host/multipath.c
@@ -18,6 +18,7 @@ static const char *nvme_iopolicy_names[] = {
 	[NVME_IOPOLICY_NUMA]	= "numa",
 	[NVME_IOPOLICY_RR]	= "round-robin",
 	[NVME_IOPOLICY_BW]	= "bandwidth",
+	[NVME_IOPOLICY_LAT]	= "latency",
 	[NVME_IOPOLICY_EWMA]	= "ewma",
 };
 
@@ -33,6 +34,8 @@ static int nvme_set_iopolicy(const char *val, const struct kernel_param *kp)
 		iopolicy = NVME_IOPOLICY_RR;
 	else if (!strncmp(val, "bandwidth", 9))
 		iopolicy = NVME_IOPOLICY_BW;
+	else if (!strncmp(val, "latency", 7))
+		iopolicy = NVME_IOPOLICY_LAT;
 	else if (!strncmp(val, "ewma", 4))
 		iopolicy = NVME_IOPOLICY_EWMA;
 	else
@@ -49,7 +52,7 @@ static int nvme_get_iopolicy(char *buf, const struct kernel_param *kp)
 module_param_call(iopolicy, nvme_set_iopolicy, nvme_get_iopolicy,
 	&iopolicy, 0644);
 MODULE_PARM_DESC(iopolicy,
-	"Default multipath I/O policy; 'numa' (default), 'round-robin', 'bandwidth', or 'ewma'");
+	"Default multipath I/O policy; 'numa' (default), 'round-robin', 'bandwidth', 'latency', or 'ewma'");
 
 void nvme_mpath_default_iopolicy(struct nvme_subsystem *subsys)
 {
@@ -153,6 +156,17 @@ void nvme_mpath_end_request(struct request *rq)
 			 nvme_req(rq)->start_time);
 }
 
+void nvme_mpath_complete_request(struct request *req)
+{
+	struct nvme_ns *ns = req->q->queuedata;
+
+	if (iopolicy == NVME_IOPOLICY_LAT) {
+		u64 lat = ktime_get_ns() - req->start_time_ns;
+
+		ns->path_weight += lat >> 10;
+	}
+}
+
 void nvme_kick_requeue_lists(struct nvme_ctrl *ctrl)
 {
 	struct nvme_ns *ns;
@@ -267,7 +281,8 @@ static struct nvme_ns *__nvme_find_path(struct nvme_ns_head *head, int node)
 
 		if (iopolicy == NVME_IOPOLICY_NUMA)
 			distance = node_distance(node, ns->ctrl->numa_node);
-		else if (iopolicy == NVME_IOPOLICY_BW)
+		else if (iopolicy == NVME_IOPOLICY_LAT ||
+			 iopolicy == NVME_IOPOLICY_BW)
 			distance = ns->path_weight;
 		else if (iopolicy == NVME_IOPOLICY_EWMA)
 			distance = nvme_path_ewma(ns);
diff --git a/drivers/nvme/host/nvme.h b/drivers/nvme/host/nvme.h
index 4eabb6ee563a..d9e9c1b5236a 100644
--- a/drivers/nvme/host/nvme.h
+++ b/drivers/nvme/host/nvme.h
@@ -391,6 +391,7 @@ enum nvme_iopolicy {
 	NVME_IOPOLICY_NUMA,
 	NVME_IOPOLICY_RR,
 	NVME_IOPOLICY_BW,
+	NVME_IOPOLICY_LAT,
 	NVME_IOPOLICY_EWMA,
 };
 
@@ -904,6 +905,7 @@ void nvme_mpath_clear_ctrl_paths(struct nvme_ctrl *ctrl);
 void nvme_mpath_shutdown_disk(struct nvme_ns_head *head);
 void nvme_mpath_start_request(struct request *rq);
 void nvme_mpath_end_request(struct request *rq);
+void nvme_mpath_complete_request(struct request *rq);
 
 static inline void nvme_trace_bio_complete(struct request *req)
 {
@@ -995,6 +997,9 @@ static inline void nvme_mpath_start_request(struct request *rq)
 static inline void nvme_mpath_end_request(struct request *rq)
 {
 }
+static inline void nvme_mpath_complete_request(struct request *rq)
+{
+}
 #endif /* CONFIG_NVME_MULTIPATH */
 
 int nvme_revalidate_zones(struct nvme_ns *ns);
-- 
2.35.3