diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index ad6843f..9af7257 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -15,6 +15,7 @@
 #include <linux/kdev_t.h>
 #include <linux/module.h>
 #include <linux/err.h>
+#include <linux/blkdev.h>
 #include "blk-cgroup.h"
 
 static DEFINE_SPINLOCK(blkio_list_lock);
@@ -55,6 +56,26 @@
 }
 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;
@@ -65,6 +86,41 @@
 }
 EXPORT_SYMBOL_GPL(blkiocg_update_timeslice_used);
 
+void blkiocg_update_request_dispatch_stats(struct blkio_group *blkg,
+						struct request *rq)
+{
+	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)
 {
@@ -325,12 +381,12 @@
 #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->stats.dequeue += dequeue;
 }
-EXPORT_SYMBOL_GPL(blkiocg_update_blkio_group_dequeue_stats);
+EXPORT_SYMBOL_GPL(blkiocg_update_dequeue_stats);
 #endif
 
 struct cftype blkio_files[] = {
diff --git a/block/blk-cgroup.h b/block/blk-cgroup.h
index 5c5e529..80010ef 100644
--- a/block/blk-cgroup.h
+++ b/block/blk-cgroup.h
@@ -112,12 +112,12 @@
 {
 	return blkg->path;
 }
-void blkiocg_update_blkio_group_dequeue_stats(struct blkio_group *blkg,
+void blkiocg_update_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) {}
+static inline void blkiocg_update_dequeue_stats(struct blkio_group *blkg,
+						unsigned long dequeue) {}
 #endif
 
 #if defined(CONFIG_BLK_CGROUP) || defined(CONFIG_BLK_CGROUP_MODULE)
@@ -130,6 +130,10 @@
 						void *key);
 void blkiocg_update_timeslice_used(struct blkio_group *blkg,
 					unsigned long time);
+void blkiocg_update_request_dispatch_stats(struct blkio_group *blkg,
+						struct request *rq);
+void blkiocg_update_request_completion_stats(struct blkio_group *blkg,
+						struct request *rq);
 #else
 struct cgroup;
 static inline struct blkio_cgroup *
@@ -147,5 +151,9 @@
 blkiocg_lookup_group(struct blkio_cgroup *blkcg, void *key) { return NULL; }
 static inline void blkiocg_update_timeslice_used(struct blkio_group *blkg,
 						unsigned long time) {}
+static inline void blkiocg_update_request_dispatch_stats(
+		struct blkio_group *blkg, struct request *rq) {}
+static inline void blkiocg_update_request_completion_stats(
+		struct blkio_group *blkg, struct request *rq) {}
 #endif
 #endif /* _BLK_CGROUP_H */
diff --git a/block/blk-core.c b/block/blk-core.c
index 9fe174d..1d94f15 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -127,6 +127,7 @@
 	rq->tag = -1;
 	rq->ref_count = 1;
 	rq->start_time = jiffies;
+	set_start_time_ns(rq);
 }
 EXPORT_SYMBOL(blk_rq_init);
 
@@ -1855,8 +1856,10 @@
 	 * and to it is freed is accounted as io that is in progress at
 	 * the driver side.
 	 */
-	if (blk_account_rq(rq))
+	if (blk_account_rq(rq)) {
 		q->in_flight[rq_is_sync(rq)]++;
+		set_io_start_time_ns(rq);
+	}
 }
 
 /**
@@ -2517,4 +2520,3 @@
 
 	return 0;
 }
-
diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c
index c5161bb..42028e7 100644
--- a/block/cfq-iosched.c
+++ b/block/cfq-iosched.c
@@ -855,7 +855,7 @@
 	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);
+	blkiocg_update_dequeue_stats(&cfqg->blkg, 1);
 }
 
 static inline unsigned int cfq_cfqq_slice_usage(struct cfq_queue *cfqq)
@@ -1865,6 +1865,7 @@
 	elv_dispatch_sort(q, rq);
 
 	cfqd->rq_in_flight[cfq_cfqq_sync(cfqq)]++;
+	blkiocg_update_request_dispatch_stats(&cfqq->cfqg->blkg, rq);
 }
 
 /*
@@ -3285,6 +3286,7 @@
 	WARN_ON(!cfqq->dispatched);
 	cfqd->rq_in_driver--;
 	cfqq->dispatched--;
+	blkiocg_update_request_completion_stats(&cfqq->cfqg->blkg, rq);
 
 	cfqd->rq_in_flight[cfq_cfqq_sync(cfqq)]--;
 
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 6690e8b..f3fff8b 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -194,7 +194,10 @@
 
 	struct gendisk *rq_disk;
 	unsigned long start_time;
-
+#ifdef CONFIG_BLK_CGROUP
+	unsigned long long start_time_ns;
+	unsigned long long io_start_time_ns;    /* when passed to hardware */
+#endif
 	/* Number of scatter-gather DMA addr+len pairs after
 	 * physical address coalescing is performed.
 	 */
@@ -1196,6 +1199,21 @@
 struct work_struct;
 int kblockd_schedule_work(struct request_queue *q, struct work_struct *work);
 
+#ifdef CONFIG_BLK_CGROUP
+static inline void set_start_time_ns(struct request *req)
+{
+	req->start_time_ns = sched_clock();
+}
+
+static inline void set_io_start_time_ns(struct request *req)
+{
+	req->io_start_time_ns = sched_clock();
+}
+#else
+static inline void set_start_time_ns(struct request *req) {}
+static inline void set_io_start_time_ns(struct request *req) {}
+#endif
+
 #define MODULE_ALIAS_BLOCKDEV(major,minor) \
 	MODULE_ALIAS("block-major-" __stringify(major) "-" __stringify(minor))
 #define MODULE_ALIAS_BLOCKDEV_MAJOR(major) \