stable.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: <gregkh@linuxfoundation.org>
To: jack@suse.cz, alexander.levin@microsoft.com, gregkh@linuxfoundation.org
Cc: <stable@vger.kernel.org>, <stable-commits@vger.kernel.org>
Subject: Patch "reiserfs: Make cancel_old_flush() reliable" has been added to the 3.18-stable tree
Date: Mon, 19 Mar 2018 10:13:19 +0100	[thread overview]
Message-ID: <1521450799201161@kroah.com> (raw)


This is a note to let you know that I've just added the patch titled

    reiserfs: Make cancel_old_flush() reliable

to the 3.18-stable tree which can be found at:
    http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary

The filename of the patch is:
     reiserfs-make-cancel_old_flush-reliable.patch
and it can be found in the queue-3.18 subdirectory.

If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@vger.kernel.org> know about it.


>From foo@baz Mon Mar 19 10:11:52 CET 2018
From: Jan Kara <jack@suse.cz>
Date: Wed, 5 Apr 2017 14:09:48 +0200
Subject: reiserfs: Make cancel_old_flush() reliable

From: Jan Kara <jack@suse.cz>


[ Upstream commit 71b0576bdb862e964a82c73327cdd1a249c53e67 ]

Currently canceling of delayed work that flushes old data using
cancel_old_flush() does not prevent work from being requeued. Thus
in theory new work can be queued after cancel_old_flush() from
reiserfs_freeze() has run. This will become larger problem once
flush_old_commits() can requeue the work itself.

Fix the problem by recording in sbi->work_queue that flushing work is
canceled and should not be requeued.

Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Sasha Levin <alexander.levin@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 fs/reiserfs/journal.c  |    2 +-
 fs/reiserfs/reiserfs.h |    1 +
 fs/reiserfs/super.c    |   21 +++++++++++++++------
 3 files changed, 17 insertions(+), 7 deletions(-)

--- a/fs/reiserfs/journal.c
+++ b/fs/reiserfs/journal.c
@@ -1961,7 +1961,7 @@ static int do_journal_release(struct rei
 	 * will be requeued because superblock is being shutdown and doesn't
 	 * have MS_ACTIVE set.
 	 */
-	cancel_delayed_work_sync(&REISERFS_SB(sb)->old_work);
+	reiserfs_cancel_old_flush(sb);
 	/* wait for all commits to finish */
 	cancel_delayed_work_sync(&SB_JOURNAL(sb)->j_work);
 
--- a/fs/reiserfs/reiserfs.h
+++ b/fs/reiserfs/reiserfs.h
@@ -2946,6 +2946,7 @@ int reiserfs_allocate_list_bitmaps(struc
 				   struct reiserfs_list_bitmap *, unsigned int);
 
 void reiserfs_schedule_old_flush(struct super_block *s);
+void reiserfs_cancel_old_flush(struct super_block *s);
 void add_save_link(struct reiserfs_transaction_handle *th,
 		   struct inode *inode, int truncate);
 int remove_save_link(struct inode *inode, int truncate);
--- a/fs/reiserfs/super.c
+++ b/fs/reiserfs/super.c
@@ -89,7 +89,9 @@ static void flush_old_commits(struct wor
 	s = sbi->s_journal->j_work_sb;
 
 	spin_lock(&sbi->old_work_lock);
-	sbi->work_queued = 0;
+	/* Avoid clobbering the cancel state... */
+	if (sbi->work_queued == 1)
+		sbi->work_queued = 0;
 	spin_unlock(&sbi->old_work_lock);
 
 	reiserfs_sync_fs(s, 1);
@@ -116,21 +118,22 @@ void reiserfs_schedule_old_flush(struct
 	spin_unlock(&sbi->old_work_lock);
 }
 
-static void cancel_old_flush(struct super_block *s)
+void reiserfs_cancel_old_flush(struct super_block *s)
 {
 	struct reiserfs_sb_info *sbi = REISERFS_SB(s);
 
-	cancel_delayed_work_sync(&REISERFS_SB(s)->old_work);
 	spin_lock(&sbi->old_work_lock);
-	sbi->work_queued = 0;
+	/* Make sure no new flushes will be queued */
+	sbi->work_queued = 2;
 	spin_unlock(&sbi->old_work_lock);
+	cancel_delayed_work_sync(&REISERFS_SB(s)->old_work);
 }
 
 static int reiserfs_freeze(struct super_block *s)
 {
 	struct reiserfs_transaction_handle th;
 
-	cancel_old_flush(s);
+	reiserfs_cancel_old_flush(s);
 
 	reiserfs_write_lock(s);
 	if (!(s->s_flags & MS_RDONLY)) {
@@ -151,7 +154,13 @@ static int reiserfs_freeze(struct super_
 
 static int reiserfs_unfreeze(struct super_block *s)
 {
+	struct reiserfs_sb_info *sbi = REISERFS_SB(s);
+
 	reiserfs_allow_writes(s);
+	spin_lock(&sbi->old_work_lock);
+	/* Allow old_work to run again */
+	sbi->work_queued = 0;
+	spin_unlock(&sbi->old_work_lock);
 	return 0;
 }
 
@@ -2164,7 +2173,7 @@ error_unlocked:
 	if (sbi->commit_wq)
 		destroy_workqueue(sbi->commit_wq);
 
-	cancel_delayed_work_sync(&REISERFS_SB(s)->old_work);
+	reiserfs_cancel_old_flush(s);
 
 	reiserfs_free_bitmap_cache(s);
 	if (SB_BUFFER_WITH_SB(s))


Patches currently in stable-queue which might be from jack@suse.cz are

queue-3.18/reiserfs-make-cancel_old_flush-reliable.patch

                 reply	other threads:[~2018-03-19  9:15 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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=1521450799201161@kroah.com \
    --to=gregkh@linuxfoundation.org \
    --cc=alexander.levin@microsoft.com \
    --cc=jack@suse.cz \
    --cc=stable-commits@vger.kernel.org \
    --cc=stable@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).