2023-03-10 07:40:20

by Yu Kuai

Subject: [PATCH v2 1/6] md/raid10: don't call bio_start_io_acct twice for bio which experienced read error

From: Yu Kuai <[email protected]>

handle_read_error() will resumit r10_bio by raid10_read_request(), which
will call bio_start_io_acct() again, while bio_end_io_acct() will only
be called once.

Fix the problem by don't account io again from handle_read_error().

Fixes: 528bc2cf2fcc ("md/raid10: enable io accounting")
Signed-off-by: Yu Kuai <[email protected]>
Acked-by: Guoqing Jiang <[email protected]>
---
drivers/md/raid10.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
index 6c66357f92f5..4f8edb6ea3e2 100644
--- a/drivers/md/raid10.c
+++ b/drivers/md/raid10.c
@@ -1173,7 +1173,7 @@ 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 handle_error)
{
struct r10conf *conf = mddev->private;
struct bio *read_bio;
@@ -1244,7 +1244,7 @@ static void raid10_read_request(struct mddev *mddev, struct bio *bio,
}
slot = r10_bio->read_slot;

- if (blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
+ if (!handle_error && blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
r10_bio->start_time = bio_start_io_acct(bio);
read_bio = bio_alloc_clone(rdev->bdev, bio, gfp, &mddev->bio_set);

@@ -1578,7 +1578,7 @@ static void __make_request(struct mddev *mddev, struct bio *bio, int sectors)
conf->geo.raid_disks);

if (bio_data_dir(bio) == READ)
- raid10_read_request(mddev, bio, r10_bio);
+ raid10_read_request(mddev, bio, r10_bio, false);
else
raid10_write_request(mddev, bio, r10_bio);
}
@@ -2980,7 +2980,7 @@ static void handle_read_error(struct mddev *mddev, struct r10bio *r10_bio)
rdev_dec_pending(rdev, mddev);
allow_barrier(conf);
r10_bio->state = 0;
- raid10_read_request(mddev, r10_bio->master_bio, r10_bio);
+ raid10_read_request(mddev, r10_bio->master_bio, r10_bio, true);
}

static void handle_write_completed(struct r10conf *conf, struct r10bio *r10_bio)
--
2.31.1

2023-03-13 22:09:00

by Song Liu

Subject: Re: [PATCH v2 1/6] md/raid10: don't call bio_start_io_acct twice for bio which experienced read error

On Thu, Mar 9, 2023 at 11:39 PM Yu Kuai <[email protected]> wrote:
>
> From: Yu Kuai <[email protected]>
>
> handle_read_error() will resumit r10_bio by raid10_read_request(), which
> will call bio_start_io_acct() again, while bio_end_io_acct() will only
> be called once.
>
> Fix the problem by don't account io again from handle_read_error().
>
> Fixes: 528bc2cf2fcc ("md/raid10: enable io accounting")
> Signed-off-by: Yu Kuai <[email protected]>
> Acked-by: Guoqing Jiang <[email protected]>

I would rather keep same argument lists for raid10_read_request
and raid10_write_request. How about we do something like this
instead?

diff --git i/drivers/md/raid10.c w/drivers/md/raid10.c
index 6b39e6c7ada3..13f33a8a8fe8 100644
--- i/drivers/md/raid10.c
+++ w/drivers/md/raid10.c
@@ -1248,7 +1248,8 @@ static void raid10_read_request(struct mddev
*mddev, struct bio *bio,
}
slot = r10_bio->read_slot;

- if (blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
+ if (!r10_bio->start_time &&
+ blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
r10_bio->start_time = bio_start_io_acct(bio);
read_bio = bio_alloc_clone(rdev->bdev, bio, gfp, &mddev->bio_set);

@@ -1578,6 +1579,7 @@ 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);

Thanks,
Song

> ---
> drivers/md/raid10.c | 8 ++++----
> 1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
> index 6c66357f92f5..4f8edb6ea3e2 100644
> --- a/drivers/md/raid10.c
> +++ b/drivers/md/raid10.c
> @@ -1173,7 +1173,7 @@ 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 handle_error)
> {
> struct r10conf *conf = mddev->private;
> struct bio *read_bio;
> @@ -1244,7 +1244,7 @@ static void raid10_read_request(struct mddev *mddev, struct bio *bio,
> }
> slot = r10_bio->read_slot;
>
> - if (blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
> + if (!handle_error && blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
> r10_bio->start_time = bio_start_io_acct(bio);
> read_bio = bio_alloc_clone(rdev->bdev, bio, gfp, &mddev->bio_set);
>
> @@ -1578,7 +1578,7 @@ static void __make_request(struct mddev *mddev, struct bio *bio, int sectors)
> conf->geo.raid_disks);
>
> if (bio_data_dir(bio) == READ)
> - raid10_read_request(mddev, bio, r10_bio);
> + raid10_read_request(mddev, bio, r10_bio, false);
> else
> raid10_write_request(mddev, bio, r10_bio);
> }
> @@ -2980,7 +2980,7 @@ static void handle_read_error(struct mddev *mddev, struct r10bio *r10_bio)
> rdev_dec_pending(rdev, mddev);
> allow_barrier(conf);
> r10_bio->state = 0;
> - raid10_read_request(mddev, r10_bio->master_bio, r10_bio);
> + raid10_read_request(mddev, r10_bio->master_bio, r10_bio, true);
> }
>
> static void handle_write_completed(struct r10conf *conf, struct r10bio *r10_bio)
> --
> 2.31.1
>

2023-03-14 01:17:56

by Yu Kuai

Subject: Re: [PATCH v2 1/6] md/raid10: don't call bio_start_io_acct twice for bio which experienced read error

Hi,

在 2023/03/14 6:08, Song Liu 写道:
> On Thu, Mar 9, 2023 at 11:39 PM Yu Kuai <[email protected]> wrote:
>>
>> From: Yu Kuai <[email protected]>
>>
>> handle_read_error() will resumit r10_bio by raid10_read_request(), which
>> will call bio_start_io_acct() again, while bio_end_io_acct() will only
>> be called once.
>>
>> Fix the problem by don't account io again from handle_read_error().
>>
>> Fixes: 528bc2cf2fcc ("md/raid10: enable io accounting")
>> Signed-off-by: Yu Kuai <[email protected]>
>> Acked-by: Guoqing Jiang <[email protected]>
>
> I would rather keep same argument lists for raid10_read_request
> and raid10_write_request. How about we do something like this
> instead?
>
> diff --git i/drivers/md/raid10.c w/drivers/md/raid10.c
> index 6b39e6c7ada3..13f33a8a8fe8 100644
> --- i/drivers/md/raid10.c
> +++ w/drivers/md/raid10.c
> @@ -1248,7 +1248,8 @@ static void raid10_read_request(struct mddev
> *mddev, struct bio *bio,
> }
> slot = r10_bio->read_slot;
>
> - if (blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
> + if (!r10_bio->start_time &&
> + blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
> r10_bio->start_time = bio_start_io_acct(bio);
> read_bio = bio_alloc_clone(rdev->bdev, bio, gfp, &mddev->bio_set);
>
> @@ -1578,6 +1579,7 @@ 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);
>
> Thanks,
> Song

Of course, this looks better. I'll send a new verison for this patch.

Thanks,
Kuai

2023-03-14 01:24:56

by Yu Kuai

Subject: [PATCH v3] md/raid10: don't call bio_start_io_acct twice for bio which experienced read error

From: Yu Kuai <[email protected]>

handle_read_error() will resumit r10_bio by raid10_read_request(), which
will call bio_start_io_acct() again, while bio_end_io_acct() will only
be called once.

Fix the problem by don't account io again from handle_read_error().

Fixes: 528bc2cf2fcc ("md/raid10: enable io accounting")
Suggested-by: Song Liu <[email protected]>
Signed-off-by: Yu Kuai <[email protected]>
---
Changes in v3:
- use r10_bo->start_time instead of adding a new paramater.
Changes in v3:
- Change the patch title

drivers/md/raid10.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
index 6c66357f92f5..3483fdf796ec 100644
--- a/drivers/md/raid10.c
+++ b/drivers/md/raid10.c
@@ -1244,7 +1244,8 @@ static void raid10_read_request(struct mddev *mddev, struct bio *bio,
}
slot = r10_bio->read_slot;

- if (blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
+ if (!r10_bio->start_time &&
+ blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
r10_bio->start_time = bio_start_io_acct(bio);
read_bio = bio_alloc_clone(rdev->bdev, bio, gfp, &mddev->bio_set);

@@ -1574,6 +1575,7 @@ 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);

--
2.31.1

2023-03-14 20:43:46

by Song Liu

Subject: Re: [PATCH v3] md/raid10: don't call bio_start_io_acct twice for bio which experienced read error

On Mon, Mar 13, 2023 at 6:23 PM Yu Kuai <[email protected]> wrote:
>
> From: Yu Kuai <[email protected]>
>
> handle_read_error() will resumit r10_bio by raid10_read_request(), which
> will call bio_start_io_acct() again, while bio_end_io_acct() will only
> be called once.
>
> Fix the problem by don't account io again from handle_read_error().
>
> Fixes: 528bc2cf2fcc ("md/raid10: enable io accounting")
> Suggested-by: Song Liu <[email protected]>
> Signed-off-by: Yu Kuai <[email protected]>

Applied to md-next.

Thanks!
Song

> ---
> Changes in v3:
> - use r10_bo->start_time instead of adding a new paramater.
> Changes in v3:
> - Change the patch title
>
> drivers/md/raid10.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
> index 6c66357f92f5..3483fdf796ec 100644
> --- a/drivers/md/raid10.c
> +++ b/drivers/md/raid10.c
> @@ -1244,7 +1244,8 @@ static void raid10_read_request(struct mddev *mddev, struct bio *bio,
> }
> slot = r10_bio->read_slot;
>
> - if (blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
> + if (!r10_bio->start_time &&
> + blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
> r10_bio->start_time = bio_start_io_acct(bio);
> read_bio = bio_alloc_clone(rdev->bdev, bio, gfp, &mddev->bio_set);
>
> @@ -1574,6 +1575,7 @@ 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);
>
> --
> 2.31.1
>