md/raid10: switch to use md_account_bio() for io accounting

Make sure that 'active_io' will represent inflight io instead of io that
is dispatching, and io accounting from all levels will be consistent.

Signed-off-by: Yu Kuai <yukuai3@huawei.com>
Reviewed-by: Xiao Ni <xni@redhat.com>
Signed-off-by: Song Liu <song@kernel.org>
Link: https://lore.kernel.org/r/20230621165110.1498313-6-yukuai1@huaweicloud.com

diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
index 5051149..d42e9b7 100644 (file)
--- a/drivers/md/raid10.c
+++ b/drivers/md/raid10.c
@@ -325,8 +325,6 @@ static void raid_end_bio_io(struct r10bio *r10_bio)
        if (!test_bit(R10BIO_Uptodate, &r10_bio->state))
                bio->bi_status = BLK_STS_IOERR;
 
-       if (r10_bio->start_time)
-               bio_end_io_acct(bio, r10_bio->start_time);
        bio_endio(bio);
        /*
         * Wake up any possible resync thread that waits for the device
@@ -1172,7 +1170,7 @@ static bool regular_request_wait(struct mddev *mddev, struct r10conf *conf,
 }
 
 static void raid10_read_request(struct mddev *mddev, struct bio *bio,
-                               struct r10bio *r10_bio)
+                               struct r10bio *r10_bio, bool io_accounting)
 {
        struct r10conf *conf = mddev->private;
        struct bio *read_bio;
@@ -1243,9 +1241,10 @@ static void raid10_read_request(struct mddev *mddev, struct bio *bio,
        }
        slot = r10_bio->read_slot;
 
-       if (!r10_bio->start_time &&
-           blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
-               r10_bio->start_time = bio_start_io_acct(bio);
+       if (io_accounting) {
+               md_account_bio(mddev, &bio);
+               r10_bio->master_bio = bio;
+       }
        read_bio = bio_alloc_clone(rdev->bdev, bio, gfp, &mddev->bio_set);
 
        r10_bio->devs[slot].bio = read_bio;
@@ -1543,8 +1542,8 @@ static void raid10_write_request(struct mddev *mddev, struct bio *bio,
                r10_bio->master_bio = bio;
        }
 
-       if (blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
-               r10_bio->start_time = bio_start_io_acct(bio);
+       md_account_bio(mddev, &bio);
+       r10_bio->master_bio = bio;
        atomic_set(&r10_bio->remaining, 1);
        md_bitmap_startwrite(mddev->bitmap, r10_bio->sector, r10_bio->sectors, 0);
 
@@ -1571,12 +1570,11 @@ static void __make_request(struct mddev *mddev, struct bio *bio, int sectors)
        r10_bio->sector = bio->bi_iter.bi_sector;
        r10_bio->state = 0;
        r10_bio->read_slot = -1;
-       r10_bio->start_time = 0;
        memset(r10_bio->devs, 0, sizeof(r10_bio->devs[0]) *
                        conf->geo.raid_disks);
 
        if (bio_data_dir(bio) == READ)
-               raid10_read_request(mddev, bio, r10_bio);
+               raid10_read_request(mddev, bio, r10_bio, true);
        else
                raid10_write_request(mddev, bio, r10_bio);
 }
@@ -2985,7 +2983,7 @@ static void handle_read_error(struct mddev *mddev, struct r10bio *r10_bio)
 
        rdev_dec_pending(rdev, mddev);
        r10_bio->state = 0;
-       raid10_read_request(mddev, r10_bio->master_bio, r10_bio);
+       raid10_read_request(mddev, r10_bio->master_bio, r10_bio, false);
        /*
         * allow_barrier after re-submit to ensure no sync io
         * can be issued while regular io pending.

diff --git a/drivers/md/raid10.h b/drivers/md/raid10.h
index 63e48b1..2e75e88 100644 (file)
--- a/drivers/md/raid10.h
+++ b/drivers/md/raid10.h
@@ -123,7 +123,6 @@ struct r10bio {
        sector_t                sector; /* virtual sector number */
        int                     sectors;
        unsigned long           state;
-       unsigned long           start_time;
        struct mddev            *mddev;
        /*
         * original bio going to /dev/mdx