4.12-stable review patch. If anyone has any objections, please let me know. ------------------ From: Shaohua Li <shli@xxxxxx> commit 16d56e2fcc1fc15b981369653c3b41d7ff0b443d upstream. After bio is submitted, we should not clone it as its bi_iter might be invalid by driver. This is the case of behind_master_bio. In certain situration, we could dispatch behind_master_bio immediately for the first disk and then clone it for other disks. https://bugzilla.kernel.org/show_bug.cgi?id=196383 Reported-and-tested-by: Markus <m4rkusxxl@xxxxxx> Reviewed-by: Ming Lei <ming.lei@xxxxxxxxxx> Fix: 841c1316c7da(md: raid1: improve write behind) Signed-off-by: Shaohua Li <shli@xxxxxx> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- drivers/md/raid1.c | 34 +++++++++++++--------------------- 1 file changed, 13 insertions(+), 21 deletions(-) --- a/drivers/md/raid1.c +++ b/drivers/md/raid1.c @@ -491,10 +491,6 @@ static void raid1_end_write_request(stru } if (behind) { - /* we release behind master bio when all write are done */ - if (r1_bio->behind_master_bio == bio) - to_put = NULL; - if (test_bit(WriteMostly, &rdev->flags)) atomic_dec(&r1_bio->behind_remaining); @@ -1087,7 +1083,7 @@ static void unfreeze_array(struct r1conf wake_up(&conf->wait_barrier); } -static struct bio *alloc_behind_master_bio(struct r1bio *r1_bio, +static void alloc_behind_master_bio(struct r1bio *r1_bio, struct bio *bio) { int size = bio->bi_iter.bi_size; @@ -1097,11 +1093,13 @@ static struct bio *alloc_behind_master_b behind_bio = bio_alloc_mddev(GFP_NOIO, vcnt, r1_bio->mddev); if (!behind_bio) - goto fail; + return; /* discard op, we don't support writezero/writesame yet */ - if (!bio_has_data(bio)) + if (!bio_has_data(bio)) { + behind_bio->bi_iter.bi_size = size; goto skip_copy; + } while (i < vcnt && size) { struct page *page; @@ -1122,14 +1120,13 @@ skip_copy: r1_bio->behind_master_bio = behind_bio;; set_bit(R1BIO_BehindIO, &r1_bio->state); - return behind_bio; + return; free_pages: pr_debug("%dB behind alloc failed, doing sync I/O\n", bio->bi_iter.bi_size); bio_free_pages(behind_bio); -fail: - return behind_bio; + bio_put(behind_bio); } struct raid1_plug_cb { @@ -1482,7 +1479,7 @@ static void raid1_write_request(struct m (atomic_read(&bitmap->behind_writes) < mddev->bitmap_info.max_write_behind) && !waitqueue_active(&bitmap->behind_wait)) { - mbio = alloc_behind_master_bio(r1_bio, bio); + alloc_behind_master_bio(r1_bio, bio); } bitmap_startwrite(bitmap, r1_bio->sector, @@ -1492,14 +1489,11 @@ static void raid1_write_request(struct m first_clone = 0; } - if (!mbio) { - if (r1_bio->behind_master_bio) - mbio = bio_clone_fast(r1_bio->behind_master_bio, - GFP_NOIO, - mddev->bio_set); - else - mbio = bio_clone_fast(bio, GFP_NOIO, mddev->bio_set); - } + if (r1_bio->behind_master_bio) + mbio = bio_clone_fast(r1_bio->behind_master_bio, + GFP_NOIO, mddev->bio_set); + else + mbio = bio_clone_fast(bio, GFP_NOIO, mddev->bio_set); if (r1_bio->behind_master_bio) { if (test_bit(WriteMostly, &conf->mirrors[i].rdev->flags)) @@ -2367,8 +2361,6 @@ static int narrow_write_error(struct r1b wbio = bio_clone_fast(r1_bio->behind_master_bio, GFP_NOIO, mddev->bio_set); - /* We really need a _all clone */ - wbio->bi_iter = (struct bvec_iter){ 0 }; } else { wbio = bio_clone_fast(r1_bio->master_bio, GFP_NOIO, mddev->bio_set);