btrfs: fix race between quota enable and quota rescan ioctl
authorFilipe Manana <fdmanana@suse.com>
Tue, 23 Aug 2022 11:45:42 +0000 (12:45 +0100)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 26 Oct 2022 10:34:27 +0000 (12:34 +0200)
commit4b996a3014ef014af8f97b60c35f5289210a4720
tree7c6c195a271c80fd0c1720461e1a1946b01515e6
parent0d942303430824f785ac7ed8434a05d87bf4bb13
btrfs: fix race between quota enable and quota rescan ioctl

commit 331cd9461412e103d07595a10289de90004ac890 upstream.

When enabling quotas, at btrfs_quota_enable(), after committing the
transaction, we change fs_info->quota_root to point to the quota root we
created and set BTRFS_FS_QUOTA_ENABLED at fs_info->flags. Then we try
to start the qgroup rescan worker, first by initializing it with a call
to qgroup_rescan_init() - however if that fails we end up freeing the
quota root but we leave fs_info->quota_root still pointing to it, this
can later result in a use-after-free somewhere else.

We have previously set the flags BTRFS_FS_QUOTA_ENABLED and
BTRFS_QGROUP_STATUS_FLAG_ON, so we can only fail with -EINPROGRESS at
btrfs_quota_enable(), which is possible if someone already called the
quota rescan ioctl, and therefore started the rescan worker.

So fix this by ignoring an -EINPROGRESS and asserting we can't get any
other error.

Reported-by: Ye Bin <yebin10@huawei.com>
Link: https://lore.kernel.org/linux-btrfs/20220823015931.421355-1-yebin10@huawei.com/
CC: stable@vger.kernel.org # 4.19+
Reviewed-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
fs/btrfs/qgroup.c