linux-ext4 - [PATCH] [RFC] Fix jbd2 to stop waking up sleeping disks on sync

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Date: Tue, 27 Feb 2024 16:25:46 -0500
From: Phillip Susi <phill@...susis.net>
To: tytso@....edu
Cc: linux-ext4@...r.kernel.org,
	Phillip Susi <phill@...susis.net>
Subject: [PATCH] [RFC] Fix jbd2 to stop waking up sleeping disks on sync

I noticed that every time I sync ( which happens automatically when
you suspend to ram ), ext4 issues a flush to the block device, even
though there have been no writes to flush.  This appears to be because
jbd2_trans_will_send_data_barrier() returns a 0 when no transaction
has been started.  The intent appears to be that a transaction that
has completed should return 0, and that when there is NO transaction,
it should return a 1, but the tests were in the wrong order, leading
to the 0 to be returned before checking for the absense of a
transaction at all.  Reversing the order allows my disk to remain in
runtime_pm when syncing.

I *think* this is correct, but I'm not very familliar with jbd2, so it
may have unintended consequences.  What do you think?
---
 fs/jbd2/journal.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index b6c114c11b97..be13dae767be 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -632,14 +632,16 @@ int jbd2_trans_will_send_data_barrier(journal_t *journal, tid_t tid)
 	if (!(journal->j_flags & JBD2_BARRIER))
 		return 0;
 	read_lock(&journal->j_state_lock);
-	/* Transaction already committed? */
-	if (tid_geq(journal->j_commit_sequence, tid))
-		goto out;
 	commit_trans = journal->j_committing_transaction;
 	if (!commit_trans || commit_trans->t_tid != tid) {
 		ret = 1;
 		goto out;
 	}
+	/* Transaction already committed? */
+	if (tid_geq(journal->j_commit_sequence, tid))
+	{
+		goto out;
+	}
 	/*
 	 * Transaction is being committed and we already proceeded to
 	 * submitting a flush to fs partition?
-- 
2.43.0