From db29d79b34d9593179de5f868be45c650923e7b4 Mon Sep 17 00:00:00 2001 From: Yu Kuai Date: Fri, 24 Nov 2023 15:59:53 +0800 Subject: dm-raid: delay flushing event_work() after reconfig_mutex is released After commit db5e653d7c9f ("md: delay choosing sync action to md_start_sync()"), md_start_sync() will hold 'reconfig_mutex', however, in order to make sure event_work is done, __md_stop() will flush workqueue with reconfig_mutex grabbed, hence if sync_work is still pending, deadlock will be triggered. Fortunately, former pacthes to fix stopping sync_thread already make sure all sync_work is done already, hence such deadlock is not possible anymore. However, in order not to cause confusions for people by this implicit dependency, delay flushing event_work to dm-raid where 'reconfig_mutex' is not held, and add some comments to emphasize that the workqueue can't be flushed with 'reconfig_mutex'. Fixes: db5e653d7c9f ("md: delay choosing sync action to md_start_sync()") Depends-on: f52f5c71f3d4 ("md: fix stopping sync thread") Signed-off-by: Yu Kuai Acked-by: Xiao Ni Signed-off-by: Mike Snitzer --- drivers/md/dm-raid.c | 3 +++ drivers/md/md.c | 11 ++++++++--- 2 files changed, 11 insertions(+), 3 deletions(-) (limited to 'drivers') diff --git a/drivers/md/dm-raid.c b/drivers/md/dm-raid.c index 91ebdcc6e9a8..eb009d6bb03a 100644 --- a/drivers/md/dm-raid.c +++ b/drivers/md/dm-raid.c @@ -3317,6 +3317,9 @@ static void raid_dtr(struct dm_target *ti) mddev_lock_nointr(&rs->md); md_stop(&rs->md); mddev_unlock(&rs->md); + + if (work_pending(&rs->md.event_work)) + flush_work(&rs->md.event_work); raid_set_free(rs); } diff --git a/drivers/md/md.c b/drivers/md/md.c index b066abbffd10..9bdd57324c37 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -82,6 +82,14 @@ static struct module *md_cluster_mod; static DECLARE_WAIT_QUEUE_HEAD(resync_wait); static struct workqueue_struct *md_wq; + +/* + * This workqueue is used for sync_work to register new sync_thread, and for + * del_work to remove rdev, and for event_work that is only set by dm-raid. + * + * Noted that sync_work will grab reconfig_mutex, hence never flush this + * workqueue whith reconfig_mutex grabbed. + */ static struct workqueue_struct *md_misc_wq; struct workqueue_struct *md_bitmap_wq; @@ -6330,9 +6338,6 @@ static void __md_stop(struct mddev *mddev) struct md_personality *pers = mddev->pers; md_bitmap_destroy(mddev); mddev_detach(mddev); - /* Ensure ->event_work is done */ - if (mddev->event_work.func) - flush_workqueue(md_misc_wq); spin_lock(&mddev->lock); mddev->pers = NULL; spin_unlock(&mddev->lock); -- cgit From b86f4b790c998afdbc88fe1aa55cfe89c4068726 Mon Sep 17 00:00:00 2001 From: Mikulas Patocka Date: Tue, 5 Dec 2023 16:39:16 +0100 Subject: dm-integrity: don't modify bio's immutable bio_vec in integrity_metadata() __bio_for_each_segment assumes that the first struct bio_vec argument doesn't change - it calls "bio_advance_iter_single((bio), &(iter), (bvl).bv_len)" to advance the iterator. Unfortunately, the dm-integrity code changes the bio_vec with "bv.bv_len -= pos". When this code path is taken, the iterator would be out of sync and dm-integrity would report errors. This happens if the machine is out of memory and "kmalloc" fails. Fix this bug by making a copy of "bv" and changing the copy instead. Fixes: 7eada909bfd7 ("dm: add integrity target") Cc: stable@vger.kernel.org # v4.12+ Signed-off-by: Mikulas Patocka Signed-off-by: Mike Snitzer --- drivers/md/dm-integrity.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) (limited to 'drivers') diff --git a/drivers/md/dm-integrity.c b/drivers/md/dm-integrity.c index e85c688fd91e..c5f03aab4552 100644 --- a/drivers/md/dm-integrity.c +++ b/drivers/md/dm-integrity.c @@ -1755,11 +1755,12 @@ static void integrity_metadata(struct work_struct *w) sectors_to_process = dio->range.n_sectors; __bio_for_each_segment(bv, bio, iter, dio->bio_details.bi_iter) { + struct bio_vec bv_copy = bv; unsigned int pos; char *mem, *checksums_ptr; again: - mem = bvec_kmap_local(&bv); + mem = bvec_kmap_local(&bv_copy); pos = 0; checksums_ptr = checksums; do { @@ -1768,7 +1769,7 @@ again: sectors_to_process -= ic->sectors_per_block; pos += ic->sectors_per_block << SECTOR_SHIFT; sector += ic->sectors_per_block; - } while (pos < bv.bv_len && sectors_to_process && checksums != checksums_onstack); + } while (pos < bv_copy.bv_len && sectors_to_process && checksums != checksums_onstack); kunmap_local(mem); r = dm_integrity_rw_tag(ic, checksums, &dio->metadata_block, &dio->metadata_offset, @@ -1793,9 +1794,9 @@ again: if (!sectors_to_process) break; - if (unlikely(pos < bv.bv_len)) { - bv.bv_offset += pos; - bv.bv_len -= pos; + if (unlikely(pos < bv_copy.bv_len)) { + bv_copy.bv_offset += pos; + bv_copy.bv_len -= pos; goto again; } } -- cgit From 6849302fdff126997765d16df355b73231f130d4 Mon Sep 17 00:00:00 2001 From: Mike Snitzer Date: Wed, 13 Dec 2023 14:46:19 -0500 Subject: dm audit: fix Kconfig so DM_AUDIT depends on BLK_DEV_DM Signed-off-by: Mike Snitzer --- drivers/md/Kconfig | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers') diff --git a/drivers/md/Kconfig b/drivers/md/Kconfig index 2a8b081bce7d..3ff87cb4dc49 100644 --- a/drivers/md/Kconfig +++ b/drivers/md/Kconfig @@ -660,6 +660,7 @@ config DM_ZONED config DM_AUDIT bool "DM audit events" + depends on BLK_DEV_DM depends on AUDIT help Generate audit events for device-mapper. -- cgit