blkio: Implement dynamic io controlling policy registration

o One of the goals of block IO controller is that it should be able to
  support mulitple io control policies, some of which be operational at
  higher level in storage hierarchy.

o To begin with, we had one io controlling policy implemented by CFQ, and
  I hard coded the CFQ functions called by blkio. This created issues when
  CFQ is compiled as module.

o This patch implements a basic dynamic io controlling policy registration
  functionality in blkio. This is similar to elevator functionality where
  ioschedulers register the functions dynamically.

o Now in future, when more IO controlling policies are implemented, these
  can dynakically register with block IO controller.

Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>

4 files changed, 69 insertions, 12 deletions

diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index 4d4a277..3ad497f 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -15,7 +15,9 @@
 #include <linux/kdev_t.h>
 #include <linux/module.h>
 #include "blk-cgroup.h"
-#include "cfq-iosched.h"
+
+static DEFINE_SPINLOCK(blkio_list_lock);
+static LIST_HEAD(blkio_list);
 
 struct blkio_cgroup blkio_root_cgroup = { .weight = 2*BLKIO_WEIGHT_DEFAULT };
 EXPORT_SYMBOL_GPL(blkio_root_cgroup);
@@ -138,6 +140,7 @@ blkiocg_weight_write(struct cgroup *cgroup, struct cftype *cftype, u64 val)
 	struct blkio_cgroup *blkcg;
 	struct blkio_group *blkg;
 	struct hlist_node *n;
+	struct blkio_policy_type *blkiop;
 
 	if (val < BLKIO_WEIGHT_MIN || val > BLKIO_WEIGHT_MAX)
 		return -EINVAL;
@@ -145,8 +148,13 @@ blkiocg_weight_write(struct cgroup *cgroup, struct cftype *cftype, u64 val)
 	blkcg = cgroup_to_blkio_cgroup(cgroup);
 	spin_lock_irq(&blkcg->lock);
 	blkcg->weight = (unsigned int)val;
-	hlist_for_each_entry(blkg, n, &blkcg->blkg_list, blkcg_node)
-		cfq_update_blkio_group_weight(blkg, blkcg->weight);
+	hlist_for_each_entry(blkg, n, &blkcg->blkg_list, blkcg_node) {
+		spin_lock(&blkio_list_lock);
+		list_for_each_entry(blkiop, &blkio_list, list)
+			blkiop->ops.blkio_update_group_weight_fn(blkg,
+					blkcg->weight);
+		spin_unlock(&blkio_list_lock);
+	}
 	spin_unlock_irq(&blkcg->lock);
 	return 0;
 }
@@ -224,6 +232,7 @@ static void blkiocg_destroy(struct cgroup_subsys *subsys, struct cgroup *cgroup)
 	unsigned long flags;
 	struct blkio_group *blkg;
 	void *key;
+	struct blkio_policy_type *blkiop;
 
 	rcu_read_lock();
 remove_entry:
@@ -249,7 +258,10 @@ remove_entry:
 	 * we have more policies in place, we need some dynamic registration
 	 * of callback function.
 	 */
-	cfq_unlink_blkio_group(key, blkg);
+	spin_lock(&blkio_list_lock);
+	list_for_each_entry(blkiop, &blkio_list, list)
+		blkiop->ops.blkio_unlink_group_fn(key, blkg);
+	spin_unlock(&blkio_list_lock);
 	goto remove_entry;
 done:
 	free_css_id(&blkio_subsys, &blkcg->css);
@@ -330,3 +342,19 @@ struct cgroup_subsys blkio_subsys = {
 	.subsys_id = blkio_subsys_id,
 	.use_id = 1,
 };
+
+void blkio_policy_register(struct blkio_policy_type *blkiop)
+{
+	spin_lock(&blkio_list_lock);
+	list_add_tail(&blkiop->list, &blkio_list);
+	spin_unlock(&blkio_list_lock);
+}
+EXPORT_SYMBOL_GPL(blkio_policy_register);
+
+void blkio_policy_unregister(struct blkio_policy_type *blkiop)
+{
+	spin_lock(&blkio_list_lock);
+	list_del_init(&blkiop->list);
+	spin_unlock(&blkio_list_lock);
+}
+EXPORT_SYMBOL_GPL(blkio_policy_unregister);
diff --git a/block/blk-cgroup.h b/block/blk-cgroup.h
index 4f89b96..4d316df 100644
--- a/block/blk-cgroup.h
+++ b/block/blk-cgroup.h
@@ -46,11 +46,35 @@ struct blkio_group {
 extern bool blkiocg_css_tryget(struct blkio_cgroup *blkcg);
 extern void blkiocg_css_put(struct blkio_cgroup *blkcg);
 
+typedef void (blkio_unlink_group_fn) (void *key, struct blkio_group *blkg);
+typedef void (blkio_update_group_weight_fn) (struct blkio_group *blkg,
+						unsigned int weight);
+
+struct blkio_policy_ops {
+	blkio_unlink_group_fn *blkio_unlink_group_fn;
+	blkio_update_group_weight_fn *blkio_update_group_weight_fn;
+};
+
+struct blkio_policy_type {
+	struct list_head list;
+	struct blkio_policy_ops ops;
+};
+
+/* Blkio controller policy registration */
+extern void blkio_policy_register(struct blkio_policy_type *);
+extern void blkio_policy_unregister(struct blkio_policy_type *);
+
 #else
 
 struct blkio_group {
 };
 
+struct blkio_policy_type {
+};
+
+static inline void blkio_policy_register(struct blkio_policy_type *blkiop) { }
+static inline void blkio_policy_unregister(struct blkio_policy_type *blkiop) { }
+
 #endif
 
 #define BLKIO_WEIGHT_MIN	100
diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c
index 7f3f343..78f4829 100644
--- a/block/cfq-iosched.c
+++ b/block/cfq-iosched.c
@@ -14,7 +14,6 @@
 #include <linux/ioprio.h>
 #include <linux/blktrace_api.h>
 #include "blk-cgroup.h"
-#include "cfq-iosched.h"
 
 /*
  * tunables
@@ -3855,6 +3854,17 @@ static struct elevator_type iosched_cfq = {
 	.elevator_owner =	THIS_MODULE,
 };
 
+#ifdef CONFIG_CFQ_GROUP_IOSCHED
+static struct blkio_policy_type blkio_policy_cfq = {
+	.ops = {
+		.blkio_unlink_group_fn =	cfq_unlink_blkio_group,
+		.blkio_update_group_weight_fn =	cfq_update_blkio_group_weight,
+	},
+};
+#else
+static struct blkio_policy_type blkio_policy_cfq;
+#endif
+
 static int __init cfq_init(void)
 {
 	/*
@@ -3869,6 +3879,7 @@ static int __init cfq_init(void)
 		return -ENOMEM;
 
 	elv_register(&iosched_cfq);
+	blkio_policy_register(&blkio_policy_cfq);
 
 	return 0;
 }
@@ -3876,6 +3887,7 @@ static int __init cfq_init(void)
 static void __exit cfq_exit(void)
 {
 	DECLARE_COMPLETION_ONSTACK(all_gone);
+	blkio_policy_unregister(&blkio_policy_cfq);
 	elv_unregister(&iosched_cfq);
 	ioc_gone = &all_gone;
 	/* ioc_gone's update must be visible before reading ioc_count */
diff --git a/block/cfq-iosched.h b/block/cfq-iosched.h
deleted file mode 100644
index ef7b479..0000000
--- a/block/cfq-iosched.h
+++ /dev/null
@@ -1,7 +0,0 @@
-#ifndef CFQ_IOSCHED_H
-#define CFQ_IOSCHED_H
-
-void cfq_unlink_blkio_group(void *, struct blkio_group *);
-void cfq_update_blkio_group_weight(struct blkio_group *, unsigned int);
-
-#endif