diff options
| author |   Boris Burkov <boris@bur.io> | 2025-03-05 15:16:57 -0800 |
|---|---|---|
| committer |   David Sterba <dsterba@suse.com> | 2025-03-18 20:35:51 +0100 |
| commit | 7511e29cf1355b2c47d0effb39e463119913e2f6 (patch) | |
| tree | 19bdcb729fd136f9f1cbe1c27cd93f53d4602ba7 | |
| parent | btrfs: fix block group refcount race in btrfs_create_pending_block_groups() (diff) | |
| download | wireguard-linux-7511e29cf1355b2c47d0effb39e463119913e2f6.tar.xz wireguard-linux-7511e29cf1355b2c47d0effb39e463119913e2f6.zip | |
btrfs: harden block_group::bg_list against list_del() races
As far as I can tell, these calls of list_del_init() on bg_list cannot
run concurrently with btrfs_mark_bg_unused() or btrfs_mark_bg_to_reclaim(),
as they are in transaction error paths and situations where the block
group is readonly.
However, if there is any chance at all of racing with mark_bg_unused(),
or a different future user of bg_list, better to be safe than sorry.
Otherwise we risk the following interleaving (bg_list refcount in parens)
T1 (some random op) T2 (btrfs_mark_bg_unused)
!list_empty(&bg->bg_list); (1)
list_del_init(&bg->bg_list); (1)
list_move_tail (1)
btrfs_put_block_group (0)
btrfs_delete_unused_bgs
bg = list_first_entry
list_del_init(&bg->bg_list);
btrfs_put_block_group(bg); (-1)
Ultimately, this results in a broken ref count that hits zero one deref
early and the real final deref underflows the refcount, resulting in a WARNING.
Reviewed-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Boris Burkov <boris@bur.io>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Diffstat (limited to '')
| -rw-r--r-- | fs/btrfs/extent-tree.c | 8 | ||||
| -rw-r--r-- | fs/btrfs/transaction.c | 12 |
2 files changed, 20 insertions, 0 deletions
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 5de1a1293c93..957230abd827 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -2868,7 +2868,15 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans) block_group->length, &trimmed); + /* + * Not strictly necessary to lock, as the block_group should be + * read-only from btrfs_delete_unused_bgs(). + */ + ASSERT(block_group->ro); + spin_lock(&fs_info->unused_bgs_lock); list_del_init(&block_group->bg_list); + spin_unlock(&fs_info->unused_bgs_lock); + btrfs_unfreeze_block_group(block_group); btrfs_put_block_group(block_group); diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index db8fe291d010..470dfc3a1a5c 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -160,7 +160,13 @@ void btrfs_put_transaction(struct btrfs_transaction *transaction) cache = list_first_entry(&transaction->deleted_bgs, struct btrfs_block_group, bg_list); + /* + * Not strictly necessary to lock, as no other task will be using a + * block_group on the deleted_bgs list during a transaction abort. + */ + spin_lock(&transaction->fs_info->unused_bgs_lock); list_del_init(&cache->bg_list); + spin_unlock(&transaction->fs_info->unused_bgs_lock); btrfs_unfreeze_block_group(cache); btrfs_put_block_group(cache); } @@ -2096,7 +2102,13 @@ static void btrfs_cleanup_pending_block_groups(struct btrfs_trans_handle *trans) list_for_each_entry_safe(block_group, tmp, &trans->new_bgs, bg_list) { btrfs_dec_delayed_refs_rsv_bg_inserts(fs_info); + /* + * Not strictly necessary to lock, as no other task will be using a + * block_group on the new_bgs list during a transaction abort. + */ + spin_lock(&fs_info->unused_bgs_lock); list_del_init(&block_group->bg_list); + spin_unlock(&fs_info->unused_bgs_lock); } } |