X-Git-Url: http://ftp.safe.ca/?a=blobdiff_plain;f=block%2Fblk-cgroup.c;h=9af7257f429c34dfa444f54d5e4907d7f06bd10a;hb=31373d09da5b7fe21fe6f781e92bd534a3495f00;hp=73a5525cc0a2440e2fa4ddffe3972c3543b56189;hpb=f2eecb91522686edf8199947b77f435a4031d92f;p=safe%2Fjmp%2Flinux-2.6

diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index 73a5525..9af7257 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -13,23 +13,113 @@
 #include <linux/ioprio.h>
 #include <linux/seq_file.h>
 #include <linux/kdev_t.h>
+#include <linux/module.h>
+#include <linux/err.h>
+#include <linux/blkdev.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);
+
+static struct cgroup_subsys_state *blkiocg_create(struct cgroup_subsys *,
+						  struct cgroup *);
+static int blkiocg_can_attach(struct cgroup_subsys *, struct cgroup *,
+			      struct task_struct *, bool);
+static void blkiocg_attach(struct cgroup_subsys *, struct cgroup *,
+			   struct cgroup *, struct task_struct *, bool);
+static void blkiocg_destroy(struct cgroup_subsys *, struct cgroup *);
+static int blkiocg_populate(struct cgroup_subsys *, struct cgroup *);
+
+struct cgroup_subsys blkio_subsys = {
+	.name = "blkio",
+	.create = blkiocg_create,
+	.can_attach = blkiocg_can_attach,
+	.attach = blkiocg_attach,
+	.destroy = blkiocg_destroy,
+	.populate = blkiocg_populate,
+#ifdef CONFIG_BLK_CGROUP
+	/* note: blkio_subsys_id is otherwise defined in blk-cgroup.h */
+	.subsys_id = blkio_subsys_id,
+#endif
+	.use_id = 1,
+	.module = THIS_MODULE,
+};
+EXPORT_SYMBOL_GPL(blkio_subsys);
 
 struct blkio_cgroup *cgroup_to_blkio_cgroup(struct cgroup *cgroup)
 {
 	return container_of(cgroup_subsys_state(cgroup, blkio_subsys_id),
 			    struct blkio_cgroup, css);
 }
+EXPORT_SYMBOL_GPL(cgroup_to_blkio_cgroup);
+
+/*
+ * Add to the appropriate stat variable depending on the request type.
+ * This should be called with the blkg->stats_lock held.
+ */
+void io_add_stat(uint64_t *stat, uint64_t add, unsigned int flags)
+{
+	if (flags & REQ_RW)
+		stat[IO_WRITE] += add;
+	else
+		stat[IO_READ] += add;
+	/*
+	 * Everywhere in the block layer, an IO is treated as sync if it is a
+	 * read or a SYNC write. We follow the same norm.
+	 */
+	if (!(flags & REQ_RW) || flags & REQ_RW_SYNC)
+		stat[IO_SYNC] += add;
+	else
+		stat[IO_ASYNC] += add;
+}
+
+void blkiocg_update_timeslice_used(struct blkio_group *blkg, unsigned long time)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&blkg->stats_lock, flags);
+	blkg->stats.time += time;
+	spin_unlock_irqrestore(&blkg->stats_lock, flags);
+}
+EXPORT_SYMBOL_GPL(blkiocg_update_timeslice_used);
 
-void blkiocg_update_blkio_group_stats(struct blkio_group *blkg,
-			unsigned long time, unsigned long sectors)
+void blkiocg_update_request_dispatch_stats(struct blkio_group *blkg,
+						struct request *rq)
 {
-	blkg->time += time;
-	blkg->sectors += sectors;
+	struct blkio_group_stats *stats;
+	unsigned long flags;
+
+	spin_lock_irqsave(&blkg->stats_lock, flags);
+	stats = &blkg->stats;
+	stats->sectors += blk_rq_sectors(rq);
+	io_add_stat(stats->io_serviced, 1, rq->cmd_flags);
+	io_add_stat(stats->io_service_bytes, blk_rq_sectors(rq) << 9,
+			rq->cmd_flags);
+	spin_unlock_irqrestore(&blkg->stats_lock, flags);
+}
+
+void blkiocg_update_request_completion_stats(struct blkio_group *blkg,
+						struct request *rq)
+{
+	struct blkio_group_stats *stats;
+	unsigned long flags;
+	unsigned long long now = sched_clock();
+
+	spin_lock_irqsave(&blkg->stats_lock, flags);
+	stats = &blkg->stats;
+	if (time_after64(now, rq->io_start_time_ns))
+		io_add_stat(stats->io_service_time, now - rq->io_start_time_ns,
+				rq->cmd_flags);
+	if (time_after64(rq->io_start_time_ns, rq->start_time_ns))
+		io_add_stat(stats->io_wait_time,
+				rq->io_start_time_ns - rq->start_time_ns,
+				rq->cmd_flags);
+	spin_unlock_irqrestore(&blkg->stats_lock, flags);
 }
+EXPORT_SYMBOL_GPL(blkiocg_update_request_completion_stats);
 
 void blkiocg_add_blkio_group(struct blkio_cgroup *blkcg,
 			struct blkio_group *blkg, void *key, dev_t dev)
@@ -47,6 +137,7 @@ void blkiocg_add_blkio_group(struct blkio_cgroup *blkcg,
 #endif
 	blkg->dev = dev;
 }
+EXPORT_SYMBOL_GPL(blkiocg_add_blkio_group);
 
 static void __blkiocg_del_blkio_group(struct blkio_group *blkg)
 {
@@ -81,6 +172,7 @@ out:
 	rcu_read_unlock();
 	return ret;
 }
+EXPORT_SYMBOL_GPL(blkiocg_del_blkio_group);
 
 /* called under rcu_read_lock(). */
 struct blkio_group *blkiocg_lookup_group(struct blkio_cgroup *blkcg, void *key)
@@ -97,6 +189,7 @@ struct blkio_group *blkiocg_lookup_group(struct blkio_cgroup *blkcg, void *key)
 
 	return NULL;
 }
+EXPORT_SYMBOL_GPL(blkiocg_lookup_group);
 
 #define SHOW_FUNCTION(__VAR)						\
 static u64 blkiocg_##__VAR##_read(struct cgroup *cgroup,		\
@@ -117,26 +210,140 @@ 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;
 
 	blkcg = cgroup_to_blkio_cgroup(cgroup);
+	spin_lock(&blkio_list_lock);
 	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) {
+		list_for_each_entry(blkiop, &blkio_list, list)
+			blkiop->ops.blkio_update_group_weight_fn(blkg,
+					blkcg->weight);
+	}
 	spin_unlock_irq(&blkcg->lock);
+	spin_unlock(&blkio_list_lock);
 	return 0;
 }
 
-#define SHOW_FUNCTION_PER_GROUP(__VAR)					\
+static int
+blkiocg_reset_write(struct cgroup *cgroup, struct cftype *cftype, u64 val)
+{
+	struct blkio_cgroup *blkcg;
+	struct blkio_group *blkg;
+	struct hlist_node *n;
+	struct blkio_group_stats *stats;
+
+	blkcg = cgroup_to_blkio_cgroup(cgroup);
+	spin_lock_irq(&blkcg->lock);
+	hlist_for_each_entry(blkg, n, &blkcg->blkg_list, blkcg_node) {
+		spin_lock(&blkg->stats_lock);
+		stats = &blkg->stats;
+		memset(stats, 0, sizeof(struct blkio_group_stats));
+		spin_unlock(&blkg->stats_lock);
+	}
+	spin_unlock_irq(&blkcg->lock);
+	return 0;
+}
+
+void get_key_name(int type, char *disk_id, char *str, int chars_left)
+{
+	strlcpy(str, disk_id, chars_left);
+	chars_left -= strlen(str);
+	if (chars_left <= 0) {
+		printk(KERN_WARNING
+			"Possibly incorrect cgroup stat display format");
+		return;
+	}
+	switch (type) {
+	case IO_READ:
+		strlcat(str, " Read", chars_left);
+		break;
+	case IO_WRITE:
+		strlcat(str, " Write", chars_left);
+		break;
+	case IO_SYNC:
+		strlcat(str, " Sync", chars_left);
+		break;
+	case IO_ASYNC:
+		strlcat(str, " Async", chars_left);
+		break;
+	case IO_TYPE_MAX:
+		strlcat(str, " Total", chars_left);
+		break;
+	default:
+		strlcat(str, " Invalid", chars_left);
+	}
+}
+
+typedef uint64_t (get_var) (struct blkio_group *, int);
+
+#define MAX_KEY_LEN 100
+uint64_t get_typed_stat(struct blkio_group *blkg, struct cgroup_map_cb *cb,
+		get_var *getvar, char *disk_id)
+{
+	uint64_t disk_total;
+	char key_str[MAX_KEY_LEN];
+	int type;
+
+	for (type = 0; type < IO_TYPE_MAX; type++) {
+		get_key_name(type, disk_id, key_str, MAX_KEY_LEN);
+		cb->fill(cb, key_str, getvar(blkg, type));
+	}
+	disk_total = getvar(blkg, IO_READ) + getvar(blkg, IO_WRITE);
+	get_key_name(IO_TYPE_MAX, disk_id, key_str, MAX_KEY_LEN);
+	cb->fill(cb, key_str, disk_total);
+	return disk_total;
+}
+
+uint64_t get_stat(struct blkio_group *blkg, struct cgroup_map_cb *cb,
+		get_var *getvar, char *disk_id)
+{
+	uint64_t var = getvar(blkg, 0);
+	cb->fill(cb, disk_id, var);
+	return var;
+}
+
+#define GET_STAT_INDEXED(__VAR)						\
+uint64_t get_##__VAR##_stat(struct blkio_group *blkg, int type)		\
+{									\
+	return blkg->stats.__VAR[type];					\
+}									\
+
+GET_STAT_INDEXED(io_service_bytes);
+GET_STAT_INDEXED(io_serviced);
+GET_STAT_INDEXED(io_service_time);
+GET_STAT_INDEXED(io_wait_time);
+#undef GET_STAT_INDEXED
+
+#define GET_STAT(__VAR, __CONV)						\
+uint64_t get_##__VAR##_stat(struct blkio_group *blkg, int dummy)	\
+{									\
+	uint64_t data = blkg->stats.__VAR;				\
+	if (__CONV)							\
+		data = (uint64_t)jiffies_to_msecs(data) * NSEC_PER_MSEC;\
+	return data;							\
+}
+
+GET_STAT(time, 1);
+GET_STAT(sectors, 0);
+#ifdef CONFIG_DEBUG_BLK_CGROUP
+GET_STAT(dequeue, 0);
+#endif
+#undef GET_STAT
+
+#define SHOW_FUNCTION_PER_GROUP(__VAR, get_stats, getvar, show_total)	\
 static int blkiocg_##__VAR##_read(struct cgroup *cgroup,		\
-			struct cftype *cftype, struct seq_file *m)	\
+		struct cftype *cftype, struct cgroup_map_cb *cb)	\
 {									\
 	struct blkio_cgroup *blkcg;					\
 	struct blkio_group *blkg;					\
 	struct hlist_node *n;						\
+	uint64_t cgroup_total = 0;					\
+	char disk_id[10];						\
 									\
 	if (!cgroup_lock_live_group(cgroup))				\
 		return -ENODEV;						\
@@ -144,28 +351,42 @@ static int blkiocg_##__VAR##_read(struct cgroup *cgroup,		\
 	blkcg = cgroup_to_blkio_cgroup(cgroup);				\
 	rcu_read_lock();						\
 	hlist_for_each_entry_rcu(blkg, n, &blkcg->blkg_list, blkcg_node) {\
-		if (blkg->dev)						\
-			seq_printf(m, "%u:%u %lu\n", MAJOR(blkg->dev),	\
-				 MINOR(blkg->dev), blkg->__VAR);	\
+		if (blkg->dev) {					\
+			spin_lock_irq(&blkg->stats_lock);		\
+			snprintf(disk_id, 10, "%u:%u", MAJOR(blkg->dev),\
+					MINOR(blkg->dev));		\
+			cgroup_total += get_stats(blkg, cb, getvar,	\
+						disk_id);		\
+			spin_unlock_irq(&blkg->stats_lock);		\
+		}							\
 	}								\
+	if (show_total)							\
+		cb->fill(cb, "Total", cgroup_total);			\
 	rcu_read_unlock();						\
 	cgroup_unlock();						\
 	return 0;							\
 }
 
-SHOW_FUNCTION_PER_GROUP(time);
-SHOW_FUNCTION_PER_GROUP(sectors);
+SHOW_FUNCTION_PER_GROUP(time, get_stat, get_time_stat, 0);
+SHOW_FUNCTION_PER_GROUP(sectors, get_stat, get_sectors_stat, 0);
+SHOW_FUNCTION_PER_GROUP(io_service_bytes, get_typed_stat,
+			get_io_service_bytes_stat, 1);
+SHOW_FUNCTION_PER_GROUP(io_serviced, get_typed_stat, get_io_serviced_stat, 1);
+SHOW_FUNCTION_PER_GROUP(io_service_time, get_typed_stat,
+			get_io_service_time_stat, 1);
+SHOW_FUNCTION_PER_GROUP(io_wait_time, get_typed_stat, get_io_wait_time_stat, 1);
 #ifdef CONFIG_DEBUG_BLK_CGROUP
-SHOW_FUNCTION_PER_GROUP(dequeue);
+SHOW_FUNCTION_PER_GROUP(dequeue, get_stat, get_dequeue_stat, 0);
 #endif
 #undef SHOW_FUNCTION_PER_GROUP
 
 #ifdef CONFIG_DEBUG_BLK_CGROUP
-void blkiocg_update_blkio_group_dequeue_stats(struct blkio_group *blkg,
+void blkiocg_update_dequeue_stats(struct blkio_group *blkg,
 			unsigned long dequeue)
 {
-	blkg->dequeue += dequeue;
+	blkg->stats.dequeue += dequeue;
 }
+EXPORT_SYMBOL_GPL(blkiocg_update_dequeue_stats);
 #endif
 
 struct cftype blkio_files[] = {
@@ -176,16 +397,38 @@ struct cftype blkio_files[] = {
 	},
 	{
 		.name = "time",
-		.read_seq_string = blkiocg_time_read,
+		.read_map = blkiocg_time_read,
+		.write_u64 = blkiocg_reset_write,
 	},
 	{
 		.name = "sectors",
-		.read_seq_string = blkiocg_sectors_read,
+		.read_map = blkiocg_sectors_read,
+		.write_u64 = blkiocg_reset_write,
+	},
+	{
+		.name = "io_service_bytes",
+		.read_map = blkiocg_io_service_bytes_read,
+		.write_u64 = blkiocg_reset_write,
+	},
+	{
+		.name = "io_serviced",
+		.read_map = blkiocg_io_serviced_read,
+		.write_u64 = blkiocg_reset_write,
+	},
+	{
+		.name = "io_service_time",
+		.read_map = blkiocg_io_service_time_read,
+		.write_u64 = blkiocg_reset_write,
+	},
+	{
+		.name = "io_wait_time",
+		.read_map = blkiocg_io_wait_time_read,
+		.write_u64 = blkiocg_reset_write,
 	},
 #ifdef CONFIG_DEBUG_BLK_CGROUP
        {
 		.name = "dequeue",
-		.read_seq_string = blkiocg_dequeue_read,
+		.read_map = blkiocg_dequeue_read,
        },
 #endif
 };
@@ -202,6 +445,7 @@ 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:
@@ -227,12 +471,16 @@ 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);
 	rcu_read_unlock();
-	kfree(blkcg);
+	if (blkcg != &blkio_root_cgroup)
+		kfree(blkcg);
 }
 
 static struct cgroup_subsys_state *
@@ -298,13 +546,32 @@ static void blkiocg_attach(struct cgroup_subsys *subsys, struct cgroup *cgroup,
 	task_unlock(tsk);
 }
 
-struct cgroup_subsys blkio_subsys = {
-	.name = "blkio",
-	.create = blkiocg_create,
-	.can_attach = blkiocg_can_attach,
-	.attach = blkiocg_attach,
-	.destroy = blkiocg_destroy,
-	.populate = blkiocg_populate,
-	.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);
+
+static int __init init_cgroup_blkio(void)
+{
+	return cgroup_load_subsys(&blkio_subsys);
+}
+
+static void __exit exit_cgroup_blkio(void)
+{
+	cgroup_unload_subsys(&blkio_subsys);
+}
+
+module_init(init_cgroup_blkio);
+module_exit(exit_cgroup_blkio);
+MODULE_LICENSE("GPL");