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 2/2] btrfs: do not commit transaction canceling a suspended replace
Date: Thu, 27 Jul 2023 14:07:54 +0800	[thread overview]
Message-ID: <18f1e6d4afa0db4aad56569bbab15b220f03236f.1690437675.git.wqu@suse.com> (raw)
In-Reply-To: <cover.1690437675.git.wqu@suse.com>

[BUG]
There is a very rare corner case that, if the filesystem falls into a
deadly ENOSPC trap (metadata is so full that committing a transaction
would trigger transaction abort and falls RO), and the user needs to
cancel a suspended dev-replace, it would fail.

This is because the dev-replace canceling itself would commit the
current transaction, and falls RO first.

[CAUSE]
There are two involved situations:

- Cancel a running dev-replace
  We just call btrfs_scrub_cancel(), it doesn't commit transaction
  anyway.

- Cancel a suspended dev-replace
  We only need to cleanup the various in-memory replace structure, which
  is no difference than the previous situation.

  But in this case we commit transaction, and may trigger the deadly
  ENOSPC trap.

[FIX]
Just follow the first case, do not commit transaction when not needed.

Link: https://lore.kernel.org/linux-btrfs/CA+W5K0pQyJH5zWxs4JxfHR06DSUWDOcDPNsKxbdKQ_CiUtpyUg@mail.gmail.com/
Signed-off-by: Qu Wenruo <wqu@suse.com>
---
 fs/btrfs/dev-replace.c | 10 ----------
 1 file changed, 10 deletions(-)

diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c
index fff22ed55c42..35590f17a5d7 100644
--- a/fs/btrfs/dev-replace.c
+++ b/fs/btrfs/dev-replace.c
@@ -1056,8 +1056,6 @@ int btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info)
 	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
 	struct btrfs_device *tgt_device = NULL;
 	struct btrfs_device *src_device = NULL;
-	struct btrfs_trans_handle *trans;
-	struct btrfs_root *root = fs_info->tree_root;
 	int result;
 	int ret;
 
@@ -1112,14 +1110,6 @@ int btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info)
 		/* Scrub for replace must not be running in suspended state */
 		btrfs_scrub_cancel(fs_info);
 
-		trans = btrfs_start_transaction(root, 0);
-		if (IS_ERR(trans)) {
-			mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
-			return PTR_ERR(trans);
-		}
-		ret = btrfs_commit_transaction(trans);
-		WARN_ON(ret);
-
 		btrfs_info_in_rcu(fs_info,
 		"suspended dev_replace from %s (devid %llu) to %s canceled",
 			btrfs_dev_name(src_device), src_device->devid,
-- 
2.41.0


  parent reply	other threads:[~2023-07-27  6:08 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-07-27  6:07 [PATCH 0/2] btrfs: do not commit transaction to avoid deadly ENOSPC trap Qu Wenruo
2023-07-27  6:07 ` [PATCH 1/2] btrfs: do not commit transaction after adding one device Qu Wenruo
2023-07-27  6:07 ` Qu Wenruo [this message]
2023-07-27 12:09   ` [PATCH 2/2] btrfs: do not commit transaction canceling a suspended replace David Sterba
2023-07-27 22:26     ` 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=18f1e6d4afa0db4aad56569bbab15b220f03236f.1690437675.git.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).