block: use ktime_get_ns() instead of sched_clock() for cfq and bfq

cfq and bfq have some internal fields that use sched_clock() which can
trivially use ktime_get_ns() instead. Their timestamp fields in struct
request can also use ktime_get_ns(), which resolves the 8 year old
comment added by commit 28f4197e5d47 ("block: disable preemption before
using sched_clock()").

Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>

block/bfq-cgroup.c

@@ -55,13 +55,13 @@ BFQG_FLAG_FNS(empty)
 /* This should be called with the scheduler lock held. */
 static void bfqg_stats_update_group_wait_time(struct bfqg_stats *stats)
 {
-	unsigned long long now;
+	u64 now;
 
 	if (!bfqg_stats_waiting(stats))
 		return;
 
-	now = sched_clock();
-	if (time_after64(now, stats->start_group_wait_time))
+	now = ktime_get_ns();
+	if (now > stats->start_group_wait_time)
 		blkg_stat_add(&stats->group_wait_time,
 			      now - stats->start_group_wait_time);
 	bfqg_stats_clear_waiting(stats);
@@ -77,20 +77,20 @@ static void bfqg_stats_set_start_group_wait_time(struct bfq_group *bfqg,
 		return;
 	if (bfqg == curr_bfqg)
 		return;
-	stats->start_group_wait_time = sched_clock();
+	stats->start_group_wait_time = ktime_get_ns();
 	bfqg_stats_mark_waiting(stats);
 }
 
 /* This should be called with the scheduler lock held. */
 static void bfqg_stats_end_empty_time(struct bfqg_stats *stats)
 {
-	unsigned long long now;
+	u64 now;
 
 	if (!bfqg_stats_empty(stats))
 		return;
 
-	now = sched_clock();
-	if (time_after64(now, stats->start_empty_time))
+	now = ktime_get_ns();
+	if (now > stats->start_empty_time)
 		blkg_stat_add(&stats->empty_time,
 			      now - stats->start_empty_time);
 	bfqg_stats_clear_empty(stats);
@@ -116,7 +116,7 @@ void bfqg_stats_set_start_empty_time(struct bfq_group *bfqg)
 	if (bfqg_stats_empty(stats))
 		return;
 
-	stats->start_empty_time = sched_clock();
+	stats->start_empty_time = ktime_get_ns();
 	bfqg_stats_mark_empty(stats);
 }
 
@@ -125,9 +125,9 @@ void bfqg_stats_update_idle_time(struct bfq_group *bfqg)
 	struct bfqg_stats *stats = &bfqg->stats;
 
 	if (bfqg_stats_idling(stats)) {
-		unsigned long long now = sched_clock();
+		u64 now = ktime_get_ns();
 
-		if (time_after64(now, stats->start_idle_time))
+		if (now > stats->start_idle_time)
 			blkg_stat_add(&stats->idle_time,
 				      now - stats->start_idle_time);
 		bfqg_stats_clear_idling(stats);
@@ -138,7 +138,7 @@ void bfqg_stats_set_start_idle_time(struct bfq_group *bfqg)
 {
 	struct bfqg_stats *stats = &bfqg->stats;
 
-	stats->start_idle_time = sched_clock();
+	stats->start_idle_time = ktime_get_ns();
 	bfqg_stats_mark_idling(stats);
 }
 
@@ -171,18 +171,18 @@ void bfqg_stats_update_io_merged(struct bfq_group *bfqg, unsigned int op)
 	blkg_rwstat_add(&bfqg->stats.merged, op, 1);
 }
 
-void bfqg_stats_update_completion(struct bfq_group *bfqg, uint64_t start_time,
-				  uint64_t io_start_time, unsigned int op)
+void bfqg_stats_update_completion(struct bfq_group *bfqg, u64 start_time_ns,
+				  u64 io_start_time_ns, unsigned int op)
 {
 	struct bfqg_stats *stats = &bfqg->stats;
-	unsigned long long now = sched_clock();
+	u64 now = ktime_get_ns();
 
-	if (time_after64(now, io_start_time))
+	if (now > io_start_time_ns)
 		blkg_rwstat_add(&stats->service_time, op,
-				now - io_start_time);
-	if (time_after64(io_start_time, start_time))
+				now - io_start_time_ns);
+	if (io_start_time_ns > start_time_ns)
 		blkg_rwstat_add(&stats->wait_time, op,
-				io_start_time - start_time);
+				io_start_time_ns - start_time_ns);
 }
 
 #else /* CONFIG_BFQ_GROUP_IOSCHED && CONFIG_DEBUG_BLK_CGROUP */
@@ -191,8 +191,8 @@ void bfqg_stats_update_io_add(struct bfq_group *bfqg, struct bfq_queue *bfqq,
 			      unsigned int op) { }
 void bfqg_stats_update_io_remove(struct bfq_group *bfqg, unsigned int op) { }
 void bfqg_stats_update_io_merged(struct bfq_group *bfqg, unsigned int op) { }
-void bfqg_stats_update_completion(struct bfq_group *bfqg, uint64_t start_time,
-				  uint64_t io_start_time, unsigned int op) { }
+void bfqg_stats_update_completion(struct bfq_group *bfqg, u64 start_time_ns,
+				  u64 io_start_time_ns, unsigned int op) { }
 void bfqg_stats_update_dequeue(struct bfq_group *bfqg) { }
 void bfqg_stats_set_start_empty_time(struct bfq_group *bfqg) { }
 void bfqg_stats_update_idle_time(struct bfq_group *bfqg) { }

block/bfq-iosched.h

@@ -732,9 +732,9 @@ struct bfqg_stats {
 	/* total time with empty current active q with other requests queued */
 	struct blkg_stat		empty_time;
 	/* fields after this shouldn't be cleared on stat reset */
-	uint64_t			start_group_wait_time;
-	uint64_t			start_idle_time;
-	uint64_t			start_empty_time;
+	u64				start_group_wait_time;
+	u64				start_idle_time;
+	u64				start_empty_time;
 	uint16_t			flags;
 #endif	/* CONFIG_BFQ_GROUP_IOSCHED && CONFIG_DEBUG_BLK_CGROUP */
 };
@@ -856,8 +856,8 @@ void bfqg_stats_update_io_add(struct bfq_group *bfqg, struct bfq_queue *bfqq,
 			      unsigned int op);
 void bfqg_stats_update_io_remove(struct bfq_group *bfqg, unsigned int op);
 void bfqg_stats_update_io_merged(struct bfq_group *bfqg, unsigned int op);
-void bfqg_stats_update_completion(struct bfq_group *bfqg, uint64_t start_time,
-				  uint64_t io_start_time, unsigned int op);
+void bfqg_stats_update_completion(struct bfq_group *bfqg, u64 start_time_ns,
+				  u64 io_start_time_ns, unsigned int op);
 void bfqg_stats_update_dequeue(struct bfq_group *bfqg);
 void bfqg_stats_set_start_empty_time(struct bfq_group *bfqg);
 void bfqg_stats_update_idle_time(struct bfq_group *bfqg);

block/cfq-iosched.c

@@ -210,9 +210,9 @@ struct cfqg_stats {
 	/* total time with empty current active q with other requests queued */
 	struct blkg_stat		empty_time;
 	/* fields after this shouldn't be cleared on stat reset */
-	uint64_t			start_group_wait_time;
-	uint64_t			start_idle_time;
-	uint64_t			start_empty_time;
+	u64				start_group_wait_time;
+	u64				start_idle_time;
+	u64				start_empty_time;
 	uint16_t			flags;
 #endif	/* CONFIG_DEBUG_BLK_CGROUP */
 #endif	/* CONFIG_CFQ_GROUP_IOSCHED */
@@ -491,13 +491,13 @@ CFQG_FLAG_FNS(empty)
 /* This should be called with the queue_lock held. */
 static void cfqg_stats_update_group_wait_time(struct cfqg_stats *stats)
 {
-	unsigned long long now;
+	u64 now;
 
 	if (!cfqg_stats_waiting(stats))
 		return;
 
-	now = sched_clock();
-	if (time_after64(now, stats->start_group_wait_time))
+	now = ktime_get_ns();
+	if (now > stats->start_group_wait_time)
 		blkg_stat_add(&stats->group_wait_time,
 			      now - stats->start_group_wait_time);
 	cfqg_stats_clear_waiting(stats);
@@ -513,20 +513,20 @@ static void cfqg_stats_set_start_group_wait_time(struct cfq_group *cfqg,
 		return;
 	if (cfqg == curr_cfqg)
 		return;
-	stats->start_group_wait_time = sched_clock();
+	stats->start_group_wait_time = ktime_get_ns();
 	cfqg_stats_mark_waiting(stats);
 }
 
 /* This should be called with the queue_lock held. */
 static void cfqg_stats_end_empty_time(struct cfqg_stats *stats)
 {
-	unsigned long long now;
+	u64 now;
 
 	if (!cfqg_stats_empty(stats))
 		return;
 
-	now = sched_clock();
-	if (time_after64(now, stats->start_empty_time))
+	now = ktime_get_ns();
+	if (now > stats->start_empty_time)
 		blkg_stat_add(&stats->empty_time,
 			      now - stats->start_empty_time);
 	cfqg_stats_clear_empty(stats);
@@ -552,7 +552,7 @@ static void cfqg_stats_set_start_empty_time(struct cfq_group *cfqg)
 	if (cfqg_stats_empty(stats))
 		return;
 
-	stats->start_empty_time = sched_clock();
+	stats->start_empty_time = ktime_get_ns();
 	cfqg_stats_mark_empty(stats);
 }
 
@@ -561,9 +561,9 @@ static void cfqg_stats_update_idle_time(struct cfq_group *cfqg)
 	struct cfqg_stats *stats = &cfqg->stats;
 
 	if (cfqg_stats_idling(stats)) {
-		unsigned long long now = sched_clock();
+		u64 now = ktime_get_ns();
 
-		if (time_after64(now, stats->start_idle_time))
+		if (now > stats->start_idle_time)
 			blkg_stat_add(&stats->idle_time,
 				      now - stats->start_idle_time);
 		cfqg_stats_clear_idling(stats);
@@ -576,7 +576,7 @@ static void cfqg_stats_set_start_idle_time(struct cfq_group *cfqg)
 
 	BUG_ON(cfqg_stats_idling(stats));
 
-	stats->start_idle_time = sched_clock();
+	stats->start_idle_time = ktime_get_ns();
 	cfqg_stats_mark_idling(stats);
 }
 
@@ -701,17 +701,19 @@ static inline void cfqg_stats_update_io_merged(struct cfq_group *cfqg,
 }
 
 static inline void cfqg_stats_update_completion(struct cfq_group *cfqg,
-			uint64_t start_time, uint64_t io_start_time,
-			unsigned int op)
+						u64 start_time_ns,
+						u64 io_start_time_ns,
+						unsigned int op)
 {
 	struct cfqg_stats *stats = &cfqg->stats;
-	unsigned long long now = sched_clock();
+	u64 now = ktime_get_ns();
 
-	if (time_after64(now, io_start_time))
-		blkg_rwstat_add(&stats->service_time, op, now - io_start_time);
-	if (time_after64(io_start_time, start_time))
+	if (now > io_start_time_ns)
+		blkg_rwstat_add(&stats->service_time, op,
+				now - io_start_time_ns);
+	if (io_start_time_ns > start_time_ns)
 		blkg_rwstat_add(&stats->wait_time, op,
-				io_start_time - start_time);
+				io_start_time_ns - start_time_ns);
 }
 
 /* @stats = 0 */
@@ -797,8 +799,9 @@ static inline void cfqg_stats_update_io_remove(struct cfq_group *cfqg,
 static inline void cfqg_stats_update_io_merged(struct cfq_group *cfqg,
 			unsigned int op) { }
 static inline void cfqg_stats_update_completion(struct cfq_group *cfqg,
-			uint64_t start_time, uint64_t io_start_time,
-			unsigned int op) { }
+						u64 start_time_ns,
+						u64 io_start_time_ns,
+						unsigned int op) { }
 
 #endif	/* CONFIG_CFQ_GROUP_IOSCHED */
 

include/linux/blkdev.h

@@ -1799,42 +1799,33 @@ int kblockd_schedule_work_on(int cpu, struct work_struct *work);
 int kblockd_mod_delayed_work_on(int cpu, struct delayed_work *dwork, unsigned long delay);
 
 #ifdef CONFIG_BLK_CGROUP
-/*
- * This should not be using sched_clock(). A real patch is in progress
- * to fix this up, until that is in place we need to disable preemption
- * around sched_clock() in this function and set_io_start_time_ns().
- */
 static inline void set_start_time_ns(struct request *req)
 {
-	preempt_disable();
-	req->cgroup_start_time_ns = sched_clock();
-	preempt_enable();
+	req->cgroup_start_time_ns = ktime_get_ns();
 }
 
 static inline void set_io_start_time_ns(struct request *req)
 {
-	preempt_disable();
-	req->cgroup_io_start_time_ns = sched_clock();
-	preempt_enable();
+	req->cgroup_io_start_time_ns = ktime_get_ns();
 }
 
-static inline uint64_t rq_start_time_ns(struct request *req)
+static inline u64 rq_start_time_ns(struct request *req)
 {
 	return req->cgroup_start_time_ns;
 }
 
-static inline uint64_t rq_io_start_time_ns(struct request *req)
+static inline u64 rq_io_start_time_ns(struct request *req)
 {
 	return req->cgroup_io_start_time_ns;
 }
 #else
 static inline void set_start_time_ns(struct request *req) {}
 static inline void set_io_start_time_ns(struct request *req) {}
-static inline uint64_t rq_start_time_ns(struct request *req)
+static inline u64 rq_start_time_ns(struct request *req)
 {
 	return 0;
 }
-static inline uint64_t rq_io_start_time_ns(struct request *req)
+static inline u64 rq_io_start_time_ns(struct request *req)
 {
 	return 0;
 }