linux-btrfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Qu Wenruo <wqu@suse.com>
To: linux-btrfs@vger.kernel.org
Subject: [PATCH 3/3] btrfs: allow BTRFS_IOC_SNAP_DESTROY_V2 to remove ghost subvolume
Date: Mon, 28 Jun 2021 18:16:37 +0800	[thread overview]
Message-ID: <20210628101637.349718-4-wqu@suse.com> (raw)
In-Reply-To: <20210628101637.349718-1-wqu@suse.com>

There is a report from the mail list that some subvolumes don't have any
ROOT_REF/BACKREF and has 0 ref.
But without an ORPHAN item.

Such ghost subvolumes can't be deleted by any ioctl but only rely on
btrfs-progs to add ORPHAN item.

Normally kernel only needs to gracefully abort/reject such corrupted
structure, but in this case we have all the needed infrastructures and
interface to allow BTRFS_IOC_SNAP_DESTROY_V2 to delete it.

So add the ability to delete such ghost subvolumes to
BTRFS_IOC_SNAP_DESTROY_V2.

Signed-off-by: Qu Wenruo <wqu@suse.com>
---
 fs/btrfs/ioctl.c | 78 ++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 78 insertions(+)

diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 889e27c24e3a..06d3c293cffd 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -2892,6 +2892,81 @@ static int btrfs_ioctl_get_subvol_rootref(struct file *file, void __user *argp)
 	return ret;
 }
 
+/*
+ * Special case that some subvolume has missing ORPHAN_ITEM, but its refs is
+ * already 0 (without any ROOT_REF/BACKREF).
+ * In that case such subvolume is only taking space while unable to be deleted.
+ *
+ * No reproducer to reproduce such corruption, but it won't hurt to cleanup them
+ * as we can reuse existing code since we only need to insert an orphan item and
+ * queue them to be deleted.
+ */
+static int __cold remove_ghost_subvol(struct btrfs_fs_info *fs_info,
+				      u64 rootid)
+{
+	struct btrfs_trans_handle *trans;
+	struct btrfs_root *root;
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	int ret;
+
+	root = btrfs_get_fs_root(fs_info, rootid, false);
+	if (IS_ERR(root)) {
+		ret = PTR_ERR(root);
+		return ret;
+	}
+
+	/* A ghost subvolume is already a problem, better to output a warning */
+	btrfs_warn(fs_info, "root %llu has no refs nor orphan item", rootid); 
+	if (btrfs_root_refs(&root->root_item) != 0) {
+		/* We get some strange root */
+		btrfs_warn(fs_info,
+			"root %llu has %u refs, but no proper root backref",
+			rootid, btrfs_root_refs(&root->root_item));
+		ret = -EUCLEAN;
+		goto out;
+	}
+
+	/* Already has orphan inserted */
+	if (test_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, &root->state))
+		goto out;
+
+	path = btrfs_alloc_path();
+	if (!path) {
+		ret = -ENOMEM;
+		goto out;
+	}
+	key.objectid = BTRFS_ORPHAN_OBJECTID;
+	key.type = BTRFS_ORPHAN_ITEM_KEY;
+	key.offset = rootid;
+
+	ret = btrfs_search_slot(NULL, fs_info->tree_root, &key, path, 0, 0);
+	btrfs_free_path(path);
+	/* Either error or there is already an orphan item */
+	if (ret <= 0)
+		goto out;
+
+	trans = btrfs_start_transaction(fs_info->tree_root, 1);
+	if (IS_ERR(trans)) {
+		ret = PTR_ERR(trans);
+		goto out;
+	}
+
+	ret = btrfs_insert_orphan_item(trans, fs_info->tree_root, rootid);
+	if (ret < 0 && ret != -EEXIST) {
+		btrfs_abort_transaction(trans, ret);
+		goto end_trans;
+	}
+	ret = 0;
+	btrfs_add_dead_root(root);
+
+end_trans:
+	btrfs_end_transaction(trans);
+out:
+	btrfs_put_root(root);
+	return ret;
+}
+
 static noinline int btrfs_ioctl_snap_destroy(struct file *file,
 					     void __user *arg,
 					     bool destroy_v2)
@@ -2947,6 +3022,9 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file,
 					vol_args2->subvolid, 0, 0);
 			if (IS_ERR(dentry)) {
 				err = PTR_ERR(dentry);
+				if (err == -ENOENT)
+					err = remove_ghost_subvol(fs_info,
+							vol_args2->subvolid);
 				goto out_drop_write;
 			}
 
-- 
2.32.0


  parent reply	other threads:[~2021-06-28 10:16 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-06-28 10:16 [PATCH 0/3] btrfs: allow BTRFS_IOC_SNAP_DESTROY_V2 to remove ghost subvolume Qu Wenruo
2021-06-28 10:16 ` [PATCH 1/3] btrfs: return -EINVAL if some user wants to remove uuid/data_reloc tree Qu Wenruo
2021-06-28 10:59   ` Anand Jain
2021-06-28 10:16 ` [PATCH 2/3] btrfs: remove dead comment on btrfs_add_dead_root() Qu Wenruo
2021-06-28 11:01   ` Anand Jain
2021-06-28 10:16 ` Qu Wenruo [this message]
2021-06-28 16:22   ` [PATCH 3/3] btrfs: allow BTRFS_IOC_SNAP_DESTROY_V2 to remove ghost subvolume kernel test robot
2021-06-28 16:23   ` kernel test robot
2021-06-29  7:04   ` [kbuild] " Dan Carpenter
2021-06-30 13:16   ` David Sterba
2021-06-30 13:26     ` Qu Wenruo
2021-06-30 13:30       ` David Sterba
2021-06-30 13:35         ` Qu Wenruo
2021-07-20  4:05 ` [PATCH 0/3] " Zygo Blaxell
2021-07-20  4:33   ` Qu Wenruo
2021-07-20 15:41     ` Zygo Blaxell
2021-07-20 22:40       ` Qu Wenruo
2021-08-20  5:45 ` Qu Wenruo

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20210628101637.349718-4-wqu@suse.com \
    --to=wqu@suse.com \
    --cc=linux-btrfs@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).