[PATCH] raid1: prevent unnecessary call to wake_up() in fast path

[Jack Wang <jinpu.wang@xxxxxxxxx>]

wake_up is called unconditionally in fast path such as make_request(),
which cause lock contention under high concurrency
    raid1_end_write_request
     wake_up
      __wake_up_common_lock
       spin_lock_irqsave

Improve performance by only call wake_up() if waitqueue is not empty

Fio test script:

[global]
name=random reads and writes
ioengine=libaio
direct=1
readwrite=randrw
rwmixread=70
iodepth=64
buffered=0
filename=/dev/md0
size=1G
runtime=30
time_based
randrepeat=0
norandommap
refill_buffers
ramp_time=10
bs=4k
numjobs=400
group_reporting=1
[job1]

Test result with ramdisk raid1 on a EPYC:

	Before this patch       With this patch
	READ	BW=4621MB/s	BW=7337MB/s
	WRITE	BW=1980MB/s	BW=1675MB/s

The patch is inspired by Yu Kuai's change for raid10:
https://lore.kernel.org/r/20230621105728.1268542-1-yukuai1@xxxxxxxxxxxxxxx

Cc: Yu Kuai <yukuai3@xxxxxxxxxx>
Signed-off-by: Jack Wang <jinpu.wang@xxxxxxxxx>
---
 drivers/md/raid1.c | 17 ++++++++++++-----
 1 file changed, 12 insertions(+), 5 deletions(-)

diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c
index f834d99a36f6..808c91f338e6 100644
--- a/drivers/md/raid1.c
+++ b/drivers/md/raid1.c
@@ -789,11 +789,17 @@ static int read_balance(struct r1conf *conf, struct r1bio *r1_bio, int *max_sect
 	return best_disk;
 }
 
+static void wake_up_barrier(struct r1conf *conf)
+{
+	if (wq_has_sleeper(&conf->wait_barrier))
+		wake_up(&conf->wait_barrier);
+}
+
 static void flush_bio_list(struct r1conf *conf, struct bio *bio)
 {
 	/* flush any pending bitmap writes to disk before proceeding w/ I/O */
 	raid1_prepare_flush_writes(conf->mddev->bitmap);
-	wake_up(&conf->wait_barrier);
+	wake_up_barrier(conf);
 
 	while (bio) { /* submit pending writes */
 		struct bio *next = bio->bi_next;
@@ -835,6 +841,7 @@ static void flush_pending_writes(struct r1conf *conf)
 		spin_unlock_irq(&conf->device_lock);
 }
 
+
 /* Barriers....
  * Sometimes we need to suspend IO while we do something else,
  * either some resync/recovery, or reconfigure the array.
@@ -970,7 +977,7 @@ static bool _wait_barrier(struct r1conf *conf, int idx, bool nowait)
 	 * In case freeze_array() is waiting for
 	 * get_unqueued_pending() == extra
 	 */
-	wake_up(&conf->wait_barrier);
+	wake_up_barrier(conf);
 	/* Wait for the barrier in same barrier unit bucket to drop. */
 
 	/* Return false when nowait flag is set */
@@ -1013,7 +1020,7 @@ static bool wait_read_barrier(struct r1conf *conf, sector_t sector_nr, bool nowa
 	 * In case freeze_array() is waiting for
 	 * get_unqueued_pending() == extra
 	 */
-	wake_up(&conf->wait_barrier);
+	wake_up_barrier(conf);
 	/* Wait for array to be unfrozen */
 
 	/* Return false when nowait flag is set */
@@ -1042,7 +1049,7 @@ static bool wait_barrier(struct r1conf *conf, sector_t sector_nr, bool nowait)
 static void _allow_barrier(struct r1conf *conf, int idx)
 {
 	atomic_dec(&conf->nr_pending[idx]);
-	wake_up(&conf->wait_barrier);
+	wake_up_barrier(conf);
 }
 
 static void allow_barrier(struct r1conf *conf, sector_t sector_nr)
@@ -1171,7 +1178,7 @@ static void raid1_unplug(struct blk_plug_cb *cb, bool from_schedule)
 		spin_lock_irq(&conf->device_lock);
 		bio_list_merge(&conf->pending_bio_list, &plug->pending);
 		spin_unlock_irq(&conf->device_lock);
-		wake_up(&conf->wait_barrier);
+		wake_up_barrier(conf);
 		md_wakeup_thread(mddev->thread);
 		kfree(plug);
 		return;
-- 
2.34.1