blkio: Export disk time and sectors used by a group to user space

o Export disk time and sector used by a group to user space through cgroup interface. o Also export a "dequeue" interface to cgroup which keeps track of how many a times a group was deleted from service tree. Helps in debugging. Signed-off-by: Vivek Goyal <vgoyal@redhat.com> Signed-off-by: Jens Axboe <jens.axboe@oracle.com>

blkio: Export disk time and sectors used by a group to user space
o Export disk time and sector used by a group to user space through cgroup interface. o Also export a "dequeue" interface to cgroup which keeps track of how many a times a group was deleted from service tree. Helps in debugging. Signed-off-by: Vivek Goyal <vgoyal@redhat.com> Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Vivek Goyal · Jens Axboe
1 parent 2868ef7b39
Showing 3 changed files with 99 additions and 6 deletions Side-by-side Diff
block/blk-cgroup.c
block/blk-cgroup.h
block/cfq-iosched.c
@@ -11,6 +11,8 @@
  * 	              Nauman Rafique <nauman@google.com>
  */
 #include <linux/ioprio.h>
+#include <linux/seq_file.h>
+#include <linux/kdev_t.h>
 #include "blk-cgroup.h"
  
 extern void cfq_unlink_blkio_group(void *, struct blkio_group *);
  
@@ -23,8 +25,15 @@
 			    struct blkio_cgroup, css);
 }
  
+void blkiocg_update_blkio_group_stats(struct blkio_group *blkg,
+			unsigned long time, unsigned long sectors)
+{
+	blkg->time += time;
+	blkg->sectors += sectors;
+}
+
 void blkiocg_add_blkio_group(struct blkio_cgroup *blkcg,
-				struct blkio_group *blkg, void *key)
+			struct blkio_group *blkg, void *key, dev_t dev)
 {
 	unsigned long flags;
  
@@ -37,6 +46,7 @@
 	/* Need to take css reference ? */
 	cgroup_path(blkcg->css.cgroup, blkg->path, sizeof(blkg->path));
 #endif
+	blkg->dev = dev;
 }
  
 static void __blkiocg_del_blkio_group(struct blkio_group *blkg)
  
@@ -115,12 +125,64 @@
 	return 0;
 }
  
+#define SHOW_FUNCTION_PER_GROUP(__VAR)					\
+static int blkiocg_##__VAR##_read(struct cgroup *cgroup,		\
+			struct cftype *cftype, struct seq_file *m)	\
+{									\
+	struct blkio_cgroup *blkcg;					\
+	struct blkio_group *blkg;					\
+	struct hlist_node *n;						\
+									\
+	if (!cgroup_lock_live_group(cgroup))				\
+		return -ENODEV;						\
+									\
+	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);	\
+	}								\
+	rcu_read_unlock();						\
+	cgroup_unlock();						\
+	return 0;							\
+}
+
+SHOW_FUNCTION_PER_GROUP(time);
+SHOW_FUNCTION_PER_GROUP(sectors);
+#ifdef CONFIG_DEBUG_BLK_CGROUP
+SHOW_FUNCTION_PER_GROUP(dequeue);
+#endif
+#undef SHOW_FUNCTION_PER_GROUP
+
+#ifdef CONFIG_DEBUG_BLK_CGROUP
+void blkiocg_update_blkio_group_dequeue_stats(struct blkio_group *blkg,
+			unsigned long dequeue)
+{
+	blkg->dequeue += dequeue;
+}
+#endif
+
 struct cftype blkio_files[] = {
 	{
 		.name = "weight",
 		.read_u64 = blkiocg_weight_read,
 		.write_u64 = blkiocg_weight_write,
 	},
+	{
+		.name = "time",
+		.read_seq_string = blkiocg_time_read,
+	},
+	{
+		.name = "sectors",
+		.read_seq_string = blkiocg_sectors_read,
+	},
+#ifdef CONFIG_DEBUG_BLK_CGROUP
+       {
+		.name = "dequeue",
+		.read_seq_string = blkiocg_dequeue_read,
+       },
+#endif
 };
  
 static int blkiocg_populate(struct cgroup_subsys *subsys, struct cgroup *cgroup)
@@ -30,7 +30,15 @@
 #ifdef CONFIG_DEBUG_BLK_CGROUP
 	/* Store cgroup path */
 	char path[128];
+	/* How many times this group has been removed from service tree */
+	unsigned long dequeue;
 #endif
+	/* The device MKDEV(major, minor), this group has been created for */
+	dev_t   dev;
+
+	/* total disk time and nr sectors dispatched by this group */
+	unsigned long time;
+	unsigned long sectors;
 };
  
 #define BLKIO_WEIGHT_MIN	100
  
  
  
  
@@ -42,24 +50,30 @@
 {
 	return blkg->path;
 }
+void blkiocg_update_blkio_group_dequeue_stats(struct blkio_group *blkg,
+				unsigned long dequeue);
 #else
 static inline char *blkg_path(struct blkio_group *blkg) { return NULL; }
+static inline void blkiocg_update_blkio_group_dequeue_stats(
+			struct blkio_group *blkg, unsigned long dequeue) {}
 #endif
  
 #ifdef CONFIG_BLK_CGROUP
 extern struct blkio_cgroup blkio_root_cgroup;
 extern struct blkio_cgroup *cgroup_to_blkio_cgroup(struct cgroup *cgroup);
 extern void blkiocg_add_blkio_group(struct blkio_cgroup *blkcg,
-				struct blkio_group *blkg, void *key);
+			struct blkio_group *blkg, void *key, dev_t dev);
 extern int blkiocg_del_blkio_group(struct blkio_group *blkg);
 extern struct blkio_group *blkiocg_lookup_group(struct blkio_cgroup *blkcg,
 						void *key);
+void blkiocg_update_blkio_group_stats(struct blkio_group *blkg,
+			unsigned long time, unsigned long sectors);
 #else
 static inline struct blkio_cgroup *
 cgroup_to_blkio_cgroup(struct cgroup *cgroup) { return NULL; }
  
 static inline void blkiocg_add_blkio_group(struct blkio_cgroup *blkcg,
-			struct blkio_group *blkg, void *key)
+			struct blkio_group *blkg, void *key, dev_t dev)
 {
 }
  
@@ -68,6 +82,10 @@
  
 static inline struct blkio_group *
 blkiocg_lookup_group(struct blkio_cgroup *blkcg, void *key) { return NULL; }
+static inline void blkiocg_update_blkio_group_stats(struct blkio_group *blkg,
+			unsigned long time, unsigned long sectors)
+{
+}
 #endif
 #endif /* _BLK_CGROUP_H */
@@ -143,6 +143,8 @@
 	struct cfq_rb_root *service_tree;
 	struct cfq_queue *new_cfqq;
 	struct cfq_group *cfqg;
+	/* Sectors dispatched in current dispatch round */
+	unsigned long nr_sectors;
 };
  
 /*
@@ -852,6 +854,7 @@
 	if (!RB_EMPTY_NODE(&cfqg->rb_node))
 		cfq_rb_erase(&cfqg->rb_node, st);
 	cfqg->saved_workload_slice = 0;
+	blkiocg_update_blkio_group_dequeue_stats(&cfqg->blkg, 1);
 }
  
 static inline unsigned int cfq_cfqq_slice_usage(struct cfq_queue *cfqq)
@@ -878,7 +881,8 @@
 			slice_used = allocated_slice;
 	}
  
-	cfq_log_cfqq(cfqq->cfqd, cfqq, "sl_used=%u", slice_used);
+	cfq_log_cfqq(cfqq->cfqd, cfqq, "sl_used=%u sect=%lu", slice_used,
+				cfqq->nr_sectors);
 	return slice_used;
 }
  
@@ -906,6 +910,8 @@
  
 	cfq_log_cfqg(cfqd, cfqg, "served: vt=%llu min_vt=%llu", cfqg->vdisktime,
 					st->min_vdisktime);
+	blkiocg_update_blkio_group_stats(&cfqg->blkg, used_sl,
+						cfqq->nr_sectors);
 }
  
 #ifdef CONFIG_CFQ_GROUP_IOSCHED
@@ -924,6 +930,8 @@
 	void *key = cfqd;
 	int i, j;
 	struct cfq_rb_root *st;
+	struct backing_dev_info *bdi = &cfqd->queue->backing_dev_info;
+	unsigned int major, minor;
  
 	/* Do we need to take this reference */
 	if (!css_tryget(&blkcg->css))
@@ -951,7 +959,9 @@
 	atomic_set(&cfqg->ref, 1);
  
 	/* Add group onto cgroup list */
-	blkiocg_add_blkio_group(blkcg, &cfqg->blkg, (void *)cfqd);
+	sscanf(dev_name(bdi->dev), "%u:%u", &major, &minor);
+	blkiocg_add_blkio_group(blkcg, &cfqg->blkg, (void *)cfqd,
+					MKDEV(major, minor));
  
 	/* Add group on cfqd list */
 	hlist_add_head(&cfqg->cfqd_node, &cfqd->cfqg_list);
@@ -1478,6 +1488,7 @@
 		cfqq->dispatch_start = jiffies;
 		cfqq->slice_end = 0;
 		cfqq->slice_dispatch = 0;
+		cfqq->nr_sectors = 0;
  
 		cfq_clear_cfqq_wait_request(cfqq);
 		cfq_clear_cfqq_must_dispatch(cfqq);
@@ -1801,6 +1812,7 @@
  
 	if (cfq_cfqq_sync(cfqq))
 		cfqd->sync_flight++;
+	cfqq->nr_sectors += blk_rq_sectors(rq);
 }
  
 /*
@@ -3513,7 +3525,8 @@
 	 * to make sure that cfq_put_cfqg() does not try to kfree root group
 	 */
 	atomic_set(&cfqg->ref, 1);
-	blkiocg_add_blkio_group(&blkio_root_cgroup, &cfqg->blkg, (void *)cfqd);
+	blkiocg_add_blkio_group(&blkio_root_cgroup, &cfqg->blkg, (void *)cfqd,
+					0);
 #endif
 	/*
 	 * Not strictly needed (since RB_ROOT just clears the node and we
...	...	@@ -11,6 +11,8 @@
11	11	* Nauman Rafique <nauman@google.com>
12	12	*/
13	13	#include <linux/ioprio.h>
	14	+#include <linux/seq_file.h>
	15	+#include <linux/kdev_t.h>
14	16	#include "blk-cgroup.h"
15	17
16	18	extern void cfq_unlink_blkio_group(void , struct blkio_group );
17	19
...	...	@@ -23,8 +25,15 @@
23	25	struct blkio_cgroup, css);
24	26	}
25	27
	28	+void blkiocg_update_blkio_group_stats(struct blkio_group *blkg,
	29	+ unsigned long time, unsigned long sectors)
	30	+{
	31	+ blkg->time += time;
	32	+ blkg->sectors += sectors;
	33	+}
	34	+
26	35	void blkiocg_add_blkio_group(struct blkio_cgroup *blkcg,
27		- struct blkio_group blkg, void key)
	36	+ struct blkio_group blkg, void key, dev_t dev)
28	37	{
29	38	unsigned long flags;
30	39
...	...	@@ -37,6 +46,7 @@
37	46	/* Need to take css reference ? */
38	47	cgroup_path(blkcg->css.cgroup, blkg->path, sizeof(blkg->path));
39	48	#endif
	49	+ blkg->dev = dev;
40	50	}
41	51
42	52	static void __blkiocg_del_blkio_group(struct blkio_group *blkg)
43	53
...	...	@@ -115,12 +125,64 @@
115	125	return 0;
116	126	}
117	127
	128	+#define SHOW_FUNCTION_PER_GROUP(__VAR) \
	129	+static int blkiocg_##__VAR##_read(struct cgroup *cgroup, \
	130	+ struct cftype cftype, struct seq_file m) \
	131	+{ \
	132	+ struct blkio_cgroup *blkcg; \
	133	+ struct blkio_group *blkg; \
	134	+ struct hlist_node *n; \
	135	+ \
	136	+ if (!cgroup_lock_live_group(cgroup)) \
	137	+ return -ENODEV; \
	138	+ \
	139	+ blkcg = cgroup_to_blkio_cgroup(cgroup); \
	140	+ rcu_read_lock(); \
	141	+ hlist_for_each_entry_rcu(blkg, n, &blkcg->blkg_list, blkcg_node) {\
	142	+ if (blkg->dev) \
	143	+ seq_printf(m, "%u:%u %lu\n", MAJOR(blkg->dev), \
	144	+ MINOR(blkg->dev), blkg->__VAR); \
	145	+ } \
	146	+ rcu_read_unlock(); \
	147	+ cgroup_unlock(); \
	148	+ return 0; \
	149	+}
	150	+
	151	+SHOW_FUNCTION_PER_GROUP(time);
	152	+SHOW_FUNCTION_PER_GROUP(sectors);
	153	+#ifdef CONFIG_DEBUG_BLK_CGROUP
	154	+SHOW_FUNCTION_PER_GROUP(dequeue);
	155	+#endif
	156	+#undef SHOW_FUNCTION_PER_GROUP
	157	+
	158	+#ifdef CONFIG_DEBUG_BLK_CGROUP
	159	+void blkiocg_update_blkio_group_dequeue_stats(struct blkio_group *blkg,
	160	+ unsigned long dequeue)
	161	+{
	162	+ blkg->dequeue += dequeue;
	163	+}
	164	+#endif
	165	+
118	166	struct cftype blkio_files[] = {
119	167	{
120	168	.name = "weight",
121	169	.read_u64 = blkiocg_weight_read,
122	170	.write_u64 = blkiocg_weight_write,
123	171	},
	172	+ {
	173	+ .name = "time",
	174	+ .read_seq_string = blkiocg_time_read,
	175	+ },
	176	+ {
	177	+ .name = "sectors",
	178	+ .read_seq_string = blkiocg_sectors_read,
	179	+ },
	180	+#ifdef CONFIG_DEBUG_BLK_CGROUP
	181	+ {
	182	+ .name = "dequeue",
	183	+ .read_seq_string = blkiocg_dequeue_read,
	184	+ },
	185	+#endif
124	186	};
125	187
126	188	static int blkiocg_populate(struct cgroup_subsys subsys, struct cgroup cgroup)
...	...	@@ -30,7 +30,15 @@
30	30	#ifdef CONFIG_DEBUG_BLK_CGROUP
31	31	/* Store cgroup path */
32	32	char path[128];
	33	+ /* How many times this group has been removed from service tree */
	34	+ unsigned long dequeue;
33	35	#endif
	36	+ /* The device MKDEV(major, minor), this group has been created for */
	37	+ dev_t dev;
	38	+
	39	+ /* total disk time and nr sectors dispatched by this group */
	40	+ unsigned long time;
	41	+ unsigned long sectors;
34	42	};
35	43
36	44	#define BLKIO_WEIGHT_MIN 100
37	45
38	46
39	47
40	48
...	...	@@ -42,24 +50,30 @@
42	50	{
43	51	return blkg->path;
44	52	}
	53	+void blkiocg_update_blkio_group_dequeue_stats(struct blkio_group *blkg,
	54	+ unsigned long dequeue);
45	55	#else
46	56	static inline char blkg_path(struct blkio_group blkg) { return NULL; }
	57	+static inline void blkiocg_update_blkio_group_dequeue_stats(
	58	+ struct blkio_group *blkg, unsigned long dequeue) {}
47	59	#endif
48	60
49	61	#ifdef CONFIG_BLK_CGROUP
50	62	extern struct blkio_cgroup blkio_root_cgroup;
51	63	extern struct blkio_cgroup cgroup_to_blkio_cgroup(struct cgroup cgroup);
52	64	extern void blkiocg_add_blkio_group(struct blkio_cgroup *blkcg,
53		- struct blkio_group blkg, void key);
	65	+ struct blkio_group blkg, void key, dev_t dev);
54	66	extern int blkiocg_del_blkio_group(struct blkio_group *blkg);
55	67	extern struct blkio_group blkiocg_lookup_group(struct blkio_cgroup blkcg,
56	68	void *key);
	69	+void blkiocg_update_blkio_group_stats(struct blkio_group *blkg,
	70	+ unsigned long time, unsigned long sectors);
57	71	#else
58	72	static inline struct blkio_cgroup *
59	73	cgroup_to_blkio_cgroup(struct cgroup *cgroup) { return NULL; }
60	74
61	75	static inline void blkiocg_add_blkio_group(struct blkio_cgroup *blkcg,
62		- struct blkio_group blkg, void key)
	76	+ struct blkio_group blkg, void key, dev_t dev)
63	77	{
64	78	}
65	79
...	...	@@ -68,6 +82,10 @@
68	82
69	83	static inline struct blkio_group *
70	84	blkiocg_lookup_group(struct blkio_cgroup blkcg, void key) { return NULL; }
	85	+static inline void blkiocg_update_blkio_group_stats(struct blkio_group *blkg,
	86	+ unsigned long time, unsigned long sectors)
	87	+{
	88	+}
71	89	#endif
72	90	#endif /* _BLK_CGROUP_H */
...	...	@@ -143,6 +143,8 @@
143	143	struct cfq_rb_root *service_tree;
144	144	struct cfq_queue *new_cfqq;
145	145	struct cfq_group *cfqg;
	146	+ /* Sectors dispatched in current dispatch round */
	147	+ unsigned long nr_sectors;
146	148	};
147	149
148	150	/*
...	...	@@ -852,6 +854,7 @@
852	854	if (!RB_EMPTY_NODE(&cfqg->rb_node))
853	855	cfq_rb_erase(&cfqg->rb_node, st);
854	856	cfqg->saved_workload_slice = 0;
	857	+ blkiocg_update_blkio_group_dequeue_stats(&cfqg->blkg, 1);
855	858	}
856	859
857	860	static inline unsigned int cfq_cfqq_slice_usage(struct cfq_queue *cfqq)
...	...	@@ -878,7 +881,8 @@
878	881	slice_used = allocated_slice;
879	882	}
880	883
881		- cfq_log_cfqq(cfqq->cfqd, cfqq, "sl_used=%u", slice_used);
	884	+ cfq_log_cfqq(cfqq->cfqd, cfqq, "sl_used=%u sect=%lu", slice_used,
	885	+ cfqq->nr_sectors);
882	886	return slice_used;
883	887	}
884	888
...	...	@@ -906,6 +910,8 @@
906	910
907	911	cfq_log_cfqg(cfqd, cfqg, "served: vt=%llu min_vt=%llu", cfqg->vdisktime,
908	912	st->min_vdisktime);
	913	+ blkiocg_update_blkio_group_stats(&cfqg->blkg, used_sl,
	914	+ cfqq->nr_sectors);
909	915	}
910	916
911	917	#ifdef CONFIG_CFQ_GROUP_IOSCHED
...	...	@@ -924,6 +930,8 @@
924	930	void *key = cfqd;
925	931	int i, j;
926	932	struct cfq_rb_root *st;
	933	+ struct backing_dev_info *bdi = &cfqd->queue->backing_dev_info;
	934	+ unsigned int major, minor;
927	935
928	936	/* Do we need to take this reference */
929	937	if (!css_tryget(&blkcg->css))
...	...	@@ -951,7 +959,9 @@
951	959	atomic_set(&cfqg->ref, 1);
952	960
953	961	/* Add group onto cgroup list */
954		- blkiocg_add_blkio_group(blkcg, &cfqg->blkg, (void *)cfqd);
	962	+ sscanf(dev_name(bdi->dev), "%u:%u", &major, &minor);
	963	+ blkiocg_add_blkio_group(blkcg, &cfqg->blkg, (void *)cfqd,
	964	+ MKDEV(major, minor));
955	965
956	966	/* Add group on cfqd list */
957	967	hlist_add_head(&cfqg->cfqd_node, &cfqd->cfqg_list);
...	...	@@ -1478,6 +1488,7 @@
1478	1488	cfqq->dispatch_start = jiffies;
1479	1489	cfqq->slice_end = 0;
1480	1490	cfqq->slice_dispatch = 0;
	1491	+ cfqq->nr_sectors = 0;
1481	1492
1482	1493	cfq_clear_cfqq_wait_request(cfqq);
1483	1494	cfq_clear_cfqq_must_dispatch(cfqq);
...	...	@@ -1801,6 +1812,7 @@
1801	1812
1802	1813	if (cfq_cfqq_sync(cfqq))
1803	1814	cfqd->sync_flight++;
	1815	+ cfqq->nr_sectors += blk_rq_sectors(rq);
1804	1816	}
1805	1817
1806	1818	/*
...	...	@@ -3513,7 +3525,8 @@
3513	3525	* to make sure that cfq_put_cfqg() does not try to kfree root group
3514	3526	*/
3515	3527	atomic_set(&cfqg->ref, 1);
3516		- blkiocg_add_blkio_group(&blkio_root_cgroup, &cfqg->blkg, (void *)cfqd);
	3528	+ blkiocg_add_blkio_group(&blkio_root_cgroup, &cfqg->blkg, (void *)cfqd,
	3529	+ 0);
3517	3530	#endif
3518	3531	/*
3519	3532	* Not strictly needed (since RB_ROOT just clears the node and we