Hi,
在 2025/09/01 14:43, Damien Le Moal 写道:
On 9/1/25 12:32 PM, Yu Kuai wrote:
From: Yu Kuai <yukuai3@xxxxxxxxxx>
Unify bio split code, and prepare to fix disordered split IO.
Signed-off-by: Yu Kuai <yukuai3@xxxxxxxxxx>
[...]
@@ -1586,18 +1577,13 @@ static void raid1_write_request(struct mddev *mddev, struct bio *bio,
max_sectors = min_t(int, max_sectors,
BIO_MAX_VECS * (PAGE_SIZE >> 9));
if (max_sectors < bio_sectors(bio)) {
- struct bio *split = bio_split(bio, max_sectors,
- GFP_NOIO, &conf->bio_split);
-
- if (IS_ERR(split)) {
- error = PTR_ERR(split);
+ bio = bio_submit_split_bioset(bio, max_sectors,
+ &conf->bio_split);
+ if (!bio) {
+ set_bit(R1BIO_Returned, &r1_bio->state);
Before it was "set_bit(R1BIO_Uptodate, &r1_bio->state);" that was done. Now it
is R1BIO_Returned that is set. The commit message does not mention this change
at all. Is this a bug fix ? If yes, that should be in a pre patch before the
conversion to using bio_submit_split_bioset().
There should be no functional changes, before the change we:
1) set bio->bi_status to split error value;
2) set R1BIO_Uptodate;
3) raid_end_bio_io() check R1BIO_Returned is not set, and call
call_bio_endio();
4) call_bio_endio() check R1BIO_Uptodate is already set, keep the
bio->bi_status that is by split error;
With this change:
1) bio_submit_split_bioset() already fail the bio will split error;
2) set R1BIO_Returned;
3) raid_end_bio_io() check R1BIO_Returned is set and do nothing with the
bio;
And the same with raid10 in patch 8,9;
Perhaps I'll emphasize there is no function changes and explain a bit.
Thanks,
Kuai
goto err_handle;
}
- bio_chain(split, bio);
- trace_block_split(split, bio->bi_iter.bi_sector);
- submit_bio_noacct(bio);
- bio = split;
r1_bio->master_bio = bio;
r1_bio->sectors = max_sectors;
}
@@ -1687,8 +1673,6 @@ static void raid1_write_request(struct mddev *mddev, struct bio *bio,
}
}
- bio->bi_status = errno_to_blk_status(error);
- set_bit(R1BIO_Uptodate, &r1_bio->state);
raid_end_bio_io(r1_bio);
}
