Re: [PATCH v2 3/3] md/raid1: keep holding the barrier in handle_read_error()
From: Yu Kuai <hidden>
Date: 2023-07-11 12:38:54
Hi, 在 2023/07/09 18:29, Xueshi Hu 写道:
quoted hunk ↗ jump to hunk
From: Xueshi Hu <redacted> Raid reshape changes the r1conf::raid_disks and mempool. Keep holding the barrier in handle_read_error() to avoid raid reshape. Signed-off-by: Xueshi Hu <redacted> --- drivers/md/raid1.c | 25 ++++++++++++------------- 1 file changed, 12 insertions(+), 13 deletions(-)diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c index 6c54786067b4..43a9c095f94d 100644 --- a/drivers/md/raid1.c +++ b/drivers/md/raid1.c@@ -1240,20 +1240,20 @@ static void raid1_read_request(struct mddev *mddev, struct bio *bio, rcu_read_unlock(); } - /* - * Still need barrier for READ in case that whole - * array is frozen. - */ - if (!wait_read_barrier(conf, bio->bi_iter.bi_sector, - bio->bi_opf & REQ_NOWAIT)) { - bio_wouldblock_error(bio); - return; - } - - if (!r1_bio) + if (!r1_bio) { + /* + * Still need barrier for READ in case that whole + * array is frozen. + */ + if (!wait_read_barrier(conf, bio->bi_iter.bi_sector, + bio->bi_opf & REQ_NOWAIT)) { + bio_wouldblock_error(bio); + return; + } r1_bio = alloc_r1bio(mddev, bio); - else + } else init_r1bio(r1_bio, mddev, bio); + r1_bio->sectors = max_read_sectors; /*@@ -2527,7 +2527,6 @@ static void handle_read_error(struct r1conf *conf, struct r1bio *r1_bio) } rdev_dec_pending(rdev, conf->mddev); - allow_barrier(conf, r1_bio->sector);
I think it's better to just move allow_barrier() to after raid1_read_request(). Thanks, Kuai
bio = r1_bio->master_bio; /* Reuse the old r1_bio so that the IO_BLOCKED settings are preserved */