linux-ext4.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Mauricio Faria de Oliveira <mfo@canonical.com>
To: Jan Kara <jack@suse.cz>
Cc: linux-ext4@vger.kernel.org,
	dann frazier <dann.frazier@canonical.com>,
	Mauricio Faria de Oliveira <mauricio.foliveira@gmail.com>,
	Jan Kara <jack@suse.com>
Subject: [RFC PATCH v2 2/5] jbd2: introduce journal callbacks j_submit|finish_inode_data_buffers
Date: Sun,  9 Aug 2020 22:02:05 -0300	[thread overview]
Message-ID: <20200810010210.3305322-3-mfo@canonical.com> (raw)
In-Reply-To: <20200810010210.3305322-1-mfo@canonical.com>

Add the callbacks as opt-in to override the default behavior for
the transaction's inode list, instead of moving that code around.

This is important as not only ext4 uses the inode list: ocfs2 too,
via jbd2_journal_inode_ranged_write(), and maybe out-of-tree code.

To opt-out of the default behavior (i.e., to do nothing), one has
to opt-in with a no-op function.
---
 fs/jbd2/commit.c     | 21 ++++++++++++++++-----
 include/linux/jbd2.h | 21 ++++++++++++++++++++-
 2 files changed, 36 insertions(+), 6 deletions(-)

diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c
index 51f713089e35..b98d227b50d8 100644
--- a/fs/jbd2/commit.c
+++ b/fs/jbd2/commit.c
@@ -237,10 +237,14 @@ static int journal_submit_data_buffers(journal_t *journal,
 		 * instead of writepages. Because writepages can do
 		 * block allocation  with delalloc. We need to write
 		 * only allocated blocks here.
+		 * This can be overriden with a custom callback.
 		 */
 		trace_jbd2_submit_inode_data(jinode->i_vfs_inode);
-		err = journal_submit_inode_data_buffers(mapping, dirty_start,
-				dirty_end);
+		if (journal->j_submit_inode_data_buffers)
+			err = journal->j_submit_inode_data_buffers(jinode);
+		else
+			err = journal_submit_inode_data_buffers(mapping,
+					dirty_start, dirty_end);
 		if (!ret)
 			ret = err;
 		spin_lock(&journal->j_list_lock);
@@ -274,9 +278,16 @@ static int journal_finish_inode_data_buffers(journal_t *journal,
 			continue;
 		jinode->i_flags |= JI_COMMIT_RUNNING;
 		spin_unlock(&journal->j_list_lock);
-		err = filemap_fdatawait_range_keep_errors(
-				jinode->i_vfs_inode->i_mapping, dirty_start,
-				dirty_end);
+		/*
+		 * Wait for the inode data buffers writeout.
+		 * This can be overriden with a custom callback.
+		 */
+		if (journal->j_finish_inode_data_buffers)
+			err = journal->j_finish_inode_data_buffers(jinode);
+		else
+			err = filemap_fdatawait_range_keep_errors(
+					jinode->i_vfs_inode->i_mapping,
+					dirty_start, dirty_end);
 		if (!ret)
 			ret = err;
 		spin_lock(&journal->j_list_lock);
diff --git a/include/linux/jbd2.h b/include/linux/jbd2.h
index d56128df2aff..24efe88eda1b 100644
--- a/include/linux/jbd2.h
+++ b/include/linux/jbd2.h
@@ -628,7 +628,8 @@ struct transaction_s
 	struct journal_head	*t_shadow_list;
 
 	/*
-	 * List of inodes whose data we've modified in data=ordered mode.
+	 * List of inodes whose data we've modified in data=ordered mode
+	 * or whose pages we should write-protect in data=journaled mode.
 	 * [j_list_lock]
 	 */
 	struct list_head	t_inode_list;
@@ -1110,6 +1111,24 @@ struct journal_s
 	void			(*j_commit_callback)(journal_t *,
 						     transaction_t *);
 
+	/**
+	 * @j_submit_inode_data_buffers:
+	 *
+	 * This function is called before flushing metadata buffers.
+	 * This overrides the default behavior (writeout data buffers.)
+	 */
+	int			(*j_submit_inode_data_buffers)
+					(struct jbd2_inode *);
+
+	/**
+	 * @j_finish_inode_data_buffers:
+	 *
+	 * This function is called after flushing metadata buffers.
+	 * This overrides the default behavior (wait writeout.)
+	 */
+	int			(*j_finish_inode_data_buffers)
+					(struct jbd2_inode *);
+
 	/*
 	 * Journal statistics
 	 */
-- 
2.17.1


  parent reply	other threads:[~2020-08-10  1:02 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-08-10  1:02 [RFC PATCH v2 0/5] ext4/jbd2: data=journal: write-protect pages on transaction commit Mauricio Faria de Oliveira
2020-08-10  1:02 ` [RFC PATCH v2 1/5] jbd2: test case for ext4 data=journal/mmap() journal corruption Mauricio Faria de Oliveira
2020-08-18 14:38   ` Jan Kara
2020-08-19  1:15     ` Mauricio Faria de Oliveira
2020-08-10  1:02 ` Mauricio Faria de Oliveira [this message]
2020-08-18 14:52   ` [RFC PATCH v2 2/5] jbd2: introduce journal callbacks j_submit|finish_inode_data_buffers Jan Kara
2020-08-19  1:20     ` Mauricio Faria de Oliveira
2020-08-19  9:16       ` Jan Kara
2020-08-10  1:02 ` [RFC PATCH v2 3/5] ext4: data=journal: write-protect pages on submit inode data buffers callback Mauricio Faria de Oliveira
2020-08-19  8:44   ` Jan Kara
2020-08-19 10:41     ` Jan Kara
2020-08-20 22:55       ` Mauricio Faria de Oliveira
2020-08-21 10:26         ` Jan Kara
2020-08-10  1:02 ` [RFC PATCH v2 4/5] ext4: data=journal: add inode to transaction inode list in ext4_page_mkwrite() Mauricio Faria de Oliveira
2020-08-10  1:02 ` [RFC PATCH v2 5/5] ext4/jbd2: debugging messages Mauricio Faria de Oliveira
2020-08-19  8:46   ` Jan Kara
2020-08-10  1:02 ` [RFC PATCH v2/SETUP SCRIPT] Mauricio Faria de Oliveira
2020-08-10  1:02 ` [RFC PATCH v2/TEST CASE] Mauricio Faria de Oliveira
2020-08-19  9:27 ` [RFC PATCH v2 0/5] ext4/jbd2: data=journal: write-protect pages on transaction commit Jan Kara

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=20200810010210.3305322-3-mfo@canonical.com \
    --to=mfo@canonical.com \
    --cc=dann.frazier@canonical.com \
    --cc=jack@suse.com \
    --cc=jack@suse.cz \
    --cc=linux-ext4@vger.kernel.org \
    --cc=mauricio.foliveira@gmail.com \
    /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).