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 14/16] btrfs: extract kthread code into its own init/exit helpers
Date: Thu, 22 Sep 2022 08:06:31 +0800	[thread overview]
Message-ID: <90da76842c5fe2562299af8493aa429777709dbb.1663804335.git.wqu@suse.com> (raw)
In-Reply-To: <cover.1663804335.git.wqu@suse.com>

There are several changes involved:

- Change the timing of btrfs_cleanup_transaction()
  That call is to address any unfinished transaction mostly caused by
  the cleaner/commit kthread.

  Thus at exit function and error handling path, we should stop all
  kthread, then cleanup the unfinished transaction.

  Not calling it before stopping cleaner thread.

- Remove the filemap_write_and_wait() call
  Now we have open_ctree_btree_inode_exit() call, which will invalidate
  all dirty pages of btree inode.
  Thus there is no need to writeback those dirtied tree blocks anymore.

Signed-off-by: Qu Wenruo <wqu@suse.com>
---
 fs/btrfs/disk-io.c | 79 +++++++++++++++++++++++++++++-----------------
 1 file changed, 50 insertions(+), 29 deletions(-)

diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index a152899fa21a..7d46c9442b0f 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -3804,6 +3804,52 @@ static int open_ctree_fs_root_init(struct btrfs_fs_info *fs_info)
 	return 0;
 }
 
+static int open_ctree_kthread_init(struct btrfs_fs_info *fs_info)
+{
+	int ret;
+
+	fs_info->cleaner_kthread = kthread_run(cleaner_kthread, fs_info,
+					       "btrfs-cleaner");
+	if (IS_ERR(fs_info->cleaner_kthread)) {
+		ret = PTR_ERR(fs_info->cleaner_kthread);
+		return ret;
+	}
+
+	fs_info->transaction_kthread = kthread_run(transaction_kthread,
+						   fs_info->tree_root,
+						   "btrfs-transaction");
+	if (IS_ERR(fs_info->transaction_kthread)) {
+		kthread_stop(fs_info->cleaner_kthread);
+
+		/*
+		 * Cleanup thread may have already started a trans.
+		 * The dirtied tree blocks will be invalidated at
+		 * open_ctree_btree_inode_exit() thus we don't need to bother.
+		 */
+		btrfs_cleanup_transaction(fs_info);
+
+		ret = PTR_ERR(fs_info->cleaner_kthread);
+		return ret;
+	}
+	/*
+	 * Mount does not set all options immediately, we can do it now and do
+	 * not have to wait for transaction commit
+	 */
+	btrfs_apply_pending_changes(fs_info);
+	return 0;
+}
+
+static void open_ctree_kthread_exit(struct btrfs_fs_info *fs_info)
+{
+	kthread_stop(fs_info->transaction_kthread);
+	kthread_stop(fs_info->cleaner_kthread);
+	/*
+	 * Cleanup any unfinished transaction started by transaction/cleaner
+	 * kthread.
+	 */
+	btrfs_cleanup_transaction(fs_info);
+}
+
 struct init_sequence {
 	int (*init_func)(struct btrfs_fs_info *fs_info);
 	void (*exit_func)(struct btrfs_fs_info *fs_info);
@@ -3848,6 +3894,9 @@ static const struct init_sequence open_ctree_seq[] = {
 		 */
 		.init_func = open_ctree_fs_root_init,
 		.exit_func = btrfs_free_fs_roots,
+	}, {
+		.init_func = open_ctree_kthread_init,
+		.exit_func = open_ctree_kthread_exit,
 	}
 };
 
@@ -3873,26 +3922,9 @@ int __cold open_ctree(struct super_block *sb, char *options)
 		open_ctree_res[i] = true;
 	}
 
-	fs_info->cleaner_kthread = kthread_run(cleaner_kthread, fs_info,
-					       "btrfs-cleaner");
-	if (IS_ERR(fs_info->cleaner_kthread))
-		goto fail;
-
-	fs_info->transaction_kthread = kthread_run(transaction_kthread,
-						   fs_info->tree_root,
-						   "btrfs-transaction");
-	if (IS_ERR(fs_info->transaction_kthread))
-		goto fail_cleaner;
-
-	/*
-	 * Mount does not set all options immediately, we can do it now and do
-	 * not have to wait for transaction commit
-	 */
-	btrfs_apply_pending_changes(fs_info);
-
 	ret = btrfs_read_qgroup_config(fs_info);
 	if (ret)
-		goto fail_trans_kthread;
+		goto fail;
 
 	if (btrfs_build_ref_tree(fs_info))
 		btrfs_err(fs_info, "couldn't build ref tree");
@@ -3944,17 +3976,6 @@ int __cold open_ctree(struct super_block *sb, char *options)
 
 fail_qgroup:
 	btrfs_free_qgroup_config(fs_info);
-fail_trans_kthread:
-	kthread_stop(fs_info->transaction_kthread);
-	btrfs_cleanup_transaction(fs_info);
-fail_cleaner:
-	kthread_stop(fs_info->cleaner_kthread);
-
-	/*
-	 * make sure we're done with the btree inode before we stop our
-	 * kthreads
-	 */
-	filemap_write_and_wait(fs_info->btree_inode->i_mapping);
 
 fail:
 	for (i = ARRAY_SIZE(open_ctree_seq) - 1; i >= 0; i--) {
-- 
2.37.3


  parent reply	other threads:[~2022-09-22  0:07 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-09-22  0:06 [PATCH 00/16] btrfs: make open_ctree() init/exit sequence strictly matched Qu Wenruo
2022-09-22  0:06 ` [PATCH 01/16] btrfs: make btrfs module init/exit match their sequence Qu Wenruo
2022-09-22  0:06 ` [PATCH 02/16] btrfs: initialize fs_info->sb at the very beginning of open_ctree() Qu Wenruo
2022-09-22  0:06 ` [PATCH 03/16] btrfs: remove @fs_devices argument from open_ctree() Qu Wenruo
2022-09-22  0:06 ` [PATCH 04/16] btrfs: extract btree inode init code into its own init/exit helpers Qu Wenruo
2022-09-22  0:06 ` [PATCH 05/16] btrfs: extract super block read code into its own init helper Qu Wenruo
2022-09-22  0:06 ` [PATCH 06/16] btrfs: extract mount options and features init " Qu Wenruo
2022-09-22  0:28   ` Qu Wenruo
2022-09-22  0:06 ` [PATCH 07/16] btrfs: move btrfs_init_workqueus() and btrfs_stop_all_workers() into open_ctree_seq[] Qu Wenruo
2022-09-22  0:06 ` [PATCH 08/16] btrfs: extract chunk tree read code into its own init/exit helpers Qu Wenruo
2022-09-22  0:06 ` [PATCH 09/16] btrfs: extract tree roots and zone info initialization into " Qu Wenruo
2022-09-22  0:06 ` [PATCH 10/16] btrfs: extract mount time checks and items load code into its init helper Qu Wenruo
2022-09-22  0:06 ` [PATCH 11/16] btrfs: extract sysfs init into its own helper Qu Wenruo
2022-09-22  0:06 ` [PATCH 12/16] btrfs: extra block groups read code into its own init/exit helpers Qu Wenruo
2022-09-22  0:06 ` [PATCH 13/16] btrfs: move the fs root related " Qu Wenruo
2022-09-22  0:06 ` Qu Wenruo [this message]
2022-09-22  0:06 ` [PATCH 15/16] btrfs: move qgroup init/exit code into open_ctree_seq[] array Qu Wenruo
2022-09-22  0:06 ` [PATCH 16/16] btrfs: introduce a debug mount option to do error injection for each stage of open_ctree() 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=90da76842c5fe2562299af8493aa429777709dbb.1663804335.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).