linux-block.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Mike Snitzer <snitzer@redhat.com>
To: Jens Axboe <axboe@kernel.dk>
Cc: linux-block@vger.kernel.org, dm-devel@redhat.com,
	Mikulas Patocka <mpatocka@redhat.com>
Subject: [PATCH v4 5/7] block: switch to per-cpu in-flight counters
Date: Thu,  6 Dec 2018 11:41:20 -0500	[thread overview]
Message-ID: <20181206164122.2166-6-snitzer@redhat.com> (raw)
In-Reply-To: <20181206164122.2166-1-snitzer@redhat.com>

From: Mikulas Patocka <mpatocka@redhat.com>

Now when part_round_stats is gone, we can switch to per-cpu in-flight
counters.

We use the local-atomic type local_t, so that if part_inc_in_flight or
part_dec_in_flight is reentrantly called from an interrupt, the value will
be correct.

The other counters could be corrupted due to reentrant interrupt, but the
corruption only results in slight counter skew - the in_flight counter
must be exact, so it needs local_t.

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
---
 block/genhd.c         | 43 +++++++++++++++++++++++++++++++++----------
 include/linux/genhd.h | 29 ++++++++++++++++++++++-------
 2 files changed, 55 insertions(+), 17 deletions(-)

diff --git a/block/genhd.c b/block/genhd.c
index cdf174d7d329..9827a2c05db7 100644
--- a/block/genhd.c
+++ b/block/genhd.c
@@ -50,9 +50,9 @@ void part_inc_in_flight(struct request_queue *q, struct hd_struct *part, int rw)
 	if (queue_is_mq(q))
 		return;
 
-	atomic_inc(&part->in_flight[rw]);
+	part_stat_local_inc(part, in_flight[rw]);
 	if (part->partno)
-		atomic_inc(&part_to_disk(part)->part0.in_flight[rw]);
+		part_stat_local_inc(&part_to_disk(part)->part0, in_flight[rw]);
 }
 
 void part_dec_in_flight(struct request_queue *q, struct hd_struct *part, int rw)
@@ -60,38 +60,61 @@ void part_dec_in_flight(struct request_queue *q, struct hd_struct *part, int rw)
 	if (queue_is_mq(q))
 		return;
 
-	atomic_dec(&part->in_flight[rw]);
+	part_stat_local_dec(part, in_flight[rw]);
 	if (part->partno)
-		atomic_dec(&part_to_disk(part)->part0.in_flight[rw]);
+		part_stat_local_dec(&part_to_disk(part)->part0, in_flight[rw]);
 }
 
 void part_in_flight(struct request_queue *q, struct hd_struct *part,
 		    unsigned int inflight[2])
 {
+	int cpu;
+
 	if (queue_is_mq(q)) {
 		blk_mq_in_flight(q, part, inflight);
 		return;
 	}
 
-	inflight[0] = atomic_read(&part->in_flight[0]) +
-			atomic_read(&part->in_flight[1]);
+	inflight[0] = 0;
+	for_each_possible_cpu(cpu) {
+		inflight[0] += part_stat_local_read_cpu(part, in_flight[0], cpu) +
+			       part_stat_local_read_cpu(part, in_flight[1], cpu);
+	}
+	if ((int)inflight[0] < 0)
+		inflight[0] = 0;
+
 	if (part->partno) {
 		part = &part_to_disk(part)->part0;
-		inflight[1] = atomic_read(&part->in_flight[0]) +
-				atomic_read(&part->in_flight[1]);
+		inflight[1] = 0;
+		for_each_possible_cpu(cpu) {
+			inflight[1] += part_stat_local_read_cpu(part, in_flight[0], cpu) +
+				       part_stat_local_read_cpu(part, in_flight[1], cpu);
+		}
+		if ((int)inflight[1] < 0)
+			inflight[1] = 0;
 	}
 }
 
 void part_in_flight_rw(struct request_queue *q, struct hd_struct *part,
 		       unsigned int inflight[2])
 {
+	int cpu;
+
 	if (queue_is_mq(q)) {
 		blk_mq_in_flight_rw(q, part, inflight);
 		return;
 	}
 
-	inflight[0] = atomic_read(&part->in_flight[0]);
-	inflight[1] = atomic_read(&part->in_flight[1]);
+	inflight[0] = 0;
+	inflight[1] = 0;
+	for_each_possible_cpu(cpu) {
+		inflight[0] += part_stat_local_read_cpu(part, in_flight[0], cpu);
+		inflight[1] += part_stat_local_read_cpu(part, in_flight[1], cpu);
+	}
+	if ((int)inflight[0] < 0)
+		inflight[0] = 0;
+	if ((int)inflight[1] < 0)
+		inflight[1] = 0;
 }
 
 struct hd_struct *__disk_get_part(struct gendisk *disk, int partno)
diff --git a/include/linux/genhd.h b/include/linux/genhd.h
index 838c2a7a40c5..636b4f687e35 100644
--- a/include/linux/genhd.h
+++ b/include/linux/genhd.h
@@ -17,6 +17,7 @@
 #include <linux/percpu-refcount.h>
 #include <linux/uuid.h>
 #include <linux/blk_types.h>
+#include <asm/local.h>
 
 #ifdef CONFIG_BLOCK
 
@@ -89,6 +90,7 @@ struct disk_stats {
 	unsigned long merges[NR_STAT_GROUPS];
 	unsigned long io_ticks;
 	unsigned long time_in_queue;
+	local_t in_flight[2];
 };
 
 #define PARTITION_META_INFO_VOLNAMELTH	64
@@ -122,7 +124,6 @@ struct hd_struct {
 	int make_it_fail;
 #endif
 	unsigned long stamp;
-	atomic_t in_flight[2];
 #ifdef	CONFIG_SMP
 	struct disk_stats __percpu *dkstats;
 #else
@@ -295,8 +296,11 @@ extern struct hd_struct *disk_map_sector_rcu(struct gendisk *disk,
 #define part_stat_lock()	({ rcu_read_lock(); get_cpu(); })
 #define part_stat_unlock()	do { put_cpu(); rcu_read_unlock(); } while (0)
 
-#define __part_stat_add(part, field, addnd)				\
-	(per_cpu_ptr((part)->dkstats, smp_processor_id())->field += (addnd))
+#define part_stat_get_cpu(part, field, cpu)					\
+	(per_cpu_ptr((part)->dkstats, (cpu))->field)
+
+#define part_stat_get(part, field)					\
+	part_stat_get_cpu(part, field, smp_processor_id())
 
 #define part_stat_read(part, field)					\
 ({									\
@@ -333,10 +337,9 @@ static inline void free_part_stats(struct hd_struct *part)
 #define part_stat_lock()	({ rcu_read_lock(); 0; })
 #define part_stat_unlock()	rcu_read_unlock()
 
-#define __part_stat_add(part, field, addnd)				\
-	((part)->dkstats.field += addnd)
-
-#define part_stat_read(part, field)	((part)->dkstats.field)
+#define part_stat_get(part, field)		((part)->dkstats.field)
+#define part_stat_get_cpu(part, field, cpu)	part_stat_get(part, field)
+#define part_stat_read(part, field)		part_stat_get(part, field)
 
 static inline void part_stat_set_all(struct hd_struct *part, int value)
 {
@@ -362,6 +365,9 @@ static inline void free_part_stats(struct hd_struct *part)
 	 part_stat_read(part, field[STAT_WRITE]) +			\
 	 part_stat_read(part, field[STAT_DISCARD]))
 
+#define __part_stat_add(part, field, addnd)				\
+	(part_stat_get(part, field) += (addnd))
+
 #define part_stat_add(part, field, addnd)	do {			\
 	__part_stat_add((part), field, addnd);				\
 	if ((part)->partno)						\
@@ -376,6 +382,15 @@ static inline void free_part_stats(struct hd_struct *part)
 #define part_stat_sub(gendiskp, field, subnd)				\
 	part_stat_add(gendiskp, field, -subnd)
 
+#define part_stat_local_dec(gendiskp, field)				\
+	local_dec(&(part_stat_get(gendiskp, field)))
+#define part_stat_local_inc(gendiskp, field)				\
+	local_inc(&(part_stat_get(gendiskp, field)))
+#define part_stat_local_read(gendiskp, field)				\
+	local_read(&(part_stat_get(gendiskp, field)))
+#define part_stat_local_read_cpu(gendiskp, field, cpu)			\
+	local_read(&(part_stat_get_cpu(gendiskp, field, cpu)))
+
 void part_in_flight(struct request_queue *q, struct hd_struct *part,
 		    unsigned int inflight[2]);
 void part_in_flight_rw(struct request_queue *q, struct hd_struct *part,
-- 
2.15.0


  parent reply	other threads:[~2018-12-06 16:41 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-12-06 16:41 [PATCH v4 0/7] per-cpu in_flight counters for bio-based drivers Mike Snitzer
2018-12-06 16:41 ` [PATCH v4 1/7] dm: dont rewrite dm_disk(md)->part0.in_flight Mike Snitzer
2018-12-06 16:41 ` [PATCH v4 2/7] dm rq: leverage blk_mq_queue_busy() to check for outstanding IO Mike Snitzer
2018-12-06 16:41 ` [PATCH v4 3/7] block: stop passing 'cpu' to all percpu stats methods Mike Snitzer
2018-12-06 16:41 ` [PATCH v4 4/7] block: delete part_round_stats and switch to less precise counting Mike Snitzer
2018-12-06 16:41 ` Mike Snitzer [this message]
2018-12-06 16:41 ` [PATCH v4 6/7] block: return just one value from part_in_flight Mike Snitzer
2018-12-06 16:41 ` [PATCH v4 7/7] dm: remove the pending IO accounting Mike Snitzer
2018-12-06 18:00 ` [PATCH v4 0/7] per-cpu in_flight counters for bio-based drivers Mike Snitzer
2018-12-06 19:04   ` Mike Snitzer
2018-12-10 15:35 ` Jens Axboe

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20181206164122.2166-6-snitzer@redhat.com \
    --to=snitzer@redhat.com \
    --cc=axboe@kernel.dk \
    --cc=dm-devel@redhat.com \
    --cc=linux-block@vger.kernel.org \
    --cc=mpatocka@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).