On Thu, Feb 24, 2022 at 05:06:40PM +0800, Wang Jianchao (Kuaishou) wrote:

This patch makes blk-rq-qos policies pluggable as following,
(1) Add code to maintain the rq_qos_ops. A rq-qos policy need to
    register itself with rq_qos_register(). The original enum
    rq_qos_id will be removed in following patch. They will use
    a dynamic id maintained by rq_qos_ida.
(2) Add .init callback into rq_qos_ops. We use it to initialize the
    resource.
(3) Add /sys/block/x/queue/qos
    We can use '+name' or "-name" to open or close the blk-rq-qos
    policy.

This sunds like most of these should be separate patches.

 {
 	struct request_queue *q = rqos->q;
-	const char *dir_name = rq_qos_id_to_name(rqos->id);
+	const char *dir_name;
+
+	if (rqos->ops->name)
+		dir_name = rqos->ops->name;
+	else
+		dir_name = rq_qos_id_to_name(rqos->id);

Plase split out a patch to add the name to all the ops and remove
rq_qos_id_to_name, which can be at the beginning of the series.

+	spin_lock_irq(&q->rqos_lock);
+	pos = q->rq_qos;
+	if (pos) {
+		while (pos->next)
+			pos = pos->next;
+		pos->next = rqos;
+	} else {
+		q->rq_qos = rqos;
+	}

I think another really useful prep patch would be to switch the queue
linkage to use a hlist_head instead of all this open coded list
manipulation.

+	rqos = rq_qos_by_name(q, qosname);
+	if ((buf[0] == '+' && rqos)) {
+		ret = -EEXIST;
+		goto out;
+	}
+
+	if ((buf[0] == '-' && !rqos)) {

These two conditionals have pretty odd extra braces.