On Fri, Nov 30 2018 at  4:52pm -0500,
Jens Axboe [off-list ref] wrote:

On 11/30/18 2:44 PM, Mike Snitzer wrote:

quoted

On Tue, Nov 27 2018 at  7:42pm -0500,
Mikulas Patocka [off-list ref] wrote:

quoted

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>

---
 block/bio.c           |    4 ++--
 block/blk-core.c      |    4 ++--
 block/blk-merge.c     |    2 +-
 block/genhd.c         |   47 +++++++++++++++++++++++++++++++++++------------
 drivers/md/dm.c       |    4 +---
 include/linux/genhd.h |    7 ++++---
 6 files changed, 45 insertions(+), 23 deletions(-)

Index: linux-block/block/genhd.c
===================================================================

--- linux-block.orig/block/genhd.c	2018-11-28 00:09:59.000000000 +0100
+++ linux-block/block/genhd.c	2018-11-28 00:09:59.000000000 +0100

@@ -45,53 +45,76 @@ static void disk_add_events(struct gendi
 static void disk_del_events(struct gendisk *disk);
 static void disk_release_events(struct gendisk *disk);
 
-void part_inc_in_flight(struct request_queue *q, struct hd_struct *part, int rw)
+void part_inc_in_flight(struct request_queue *q, int cpu, struct hd_struct *part, int rw)
 {
 	if (queue_is_mq(q))
 		return;
 
-	atomic_inc(&part->in_flight[rw]);
+	local_inc(per_cpu_ptr(part->dkstats, cpu)->in_flight);
 	if (part->partno)
-		atomic_inc(&part_to_disk(part)->part0.in_flight[rw]);
+		local_inc(per_cpu_ptr(part_to_disk(part)->part0.dkstats, cpu)->in_flight);
 }
 
-void part_dec_in_flight(struct request_queue *q, struct hd_struct *part, int rw)
+void part_dec_in_flight(struct request_queue *q, int cpu, struct hd_struct *part, int rw)
 {
 	if (queue_is_mq(q))
 		return;
 
-	atomic_dec(&part->in_flight[rw]);
+	local_dec(per_cpu_ptr(part->dkstats, cpu)->in_flight);
 	if (part->partno)
-		atomic_dec(&part_to_disk(part)->part0.in_flight[rw]);
+		local_dec(per_cpu_ptr(part_to_disk(part)->part0.dkstats, cpu)->in_flight);
 }

The above is wrong, needs this incremental fix:

Should also be using this_cpu_inc/dec instead.

How so?  Mikulas is using local_t (original header speaks to why)