linux-kernel - [PATCH] bcachefs: WARNING in bch2_fs_journal

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]

Message-Id: <20240717220237.1246673-1-cam.alvarez.i@gmail.com>
Date: Wed, 17 Jul 2024 18:02:39 -0400
From: Camila Alvarez <cam.alvarez.i@...il.com>
To: Kent Overstreet <kent.overstreet@...ux.dev>,
	syzbot+4093905737cf289b6b38@...kaller.appspotmail.com
Cc: Brian Foster <bfoster@...hat.com>,
	linux-bcachefs@...r.kernel.org,
	linux-kernel@...r.kernel.org,
	Camila Alvarez <cam.alvarez.i@...il.com>,
	syzkaller-bugs@...glegroups.com
Subject: [PATCH] bcachefs: WARNING in bch2_fs_journal_stop

last_seq and last_empty_seq suffered from an off by one error when the
journal has no entries.

The indexes were fixed and an assertion is added to check that the
last_empty_seq is always kept under the next valid seq number.

Reported-by: syzbot+4093905737cf289b6b38@...kaller.appspotmail.com
Signed-off-by: Camila Alvarez <cam.alvarez.i@...il.com>
---
 fs/bcachefs/journal.c | 23 +++++++++++++----------
 1 file changed, 13 insertions(+), 10 deletions(-)

diff --git a/fs/bcachefs/journal.c b/fs/bcachefs/journal.c
index 10b19791ec98..7bbbf4b149e9 100644
--- a/fs/bcachefs/journal.c
+++ b/fs/bcachefs/journal.c
@@ -1201,7 +1201,7 @@ int bch2_fs_journal_start(struct journal *j, u64 cur_seq)
 	struct journal_replay *i, **_i;
 	struct genradix_iter iter;
 	bool had_entries = false;
-	u64 last_seq = cur_seq, nr, seq;
+        u64 last_written_seq = cur_seq - 1, last_seq = cur_seq - 1, nr, seq;
 
 	genradix_for_each_reverse(&c->journal_entries, iter, _i) {
 		i = *_i;
@@ -1209,11 +1209,11 @@ int bch2_fs_journal_start(struct journal *j, u64 cur_seq)
 		if (journal_replay_ignore(i))
 			continue;
 
-		last_seq = le64_to_cpu(i->j.last_seq);
+		last_written_seq = le64_to_cpu(i->j.last_seq);
 		break;
 	}
 
-	nr = cur_seq - last_seq;
+	nr = cur_seq - last_written_seq;
 
 	if (nr + 1 > j->pin.size) {
 		free_fifo(&j->pin);
@@ -1224,14 +1224,14 @@ int bch2_fs_journal_start(struct journal *j, u64 cur_seq)
 		}
 	}
 
-	j->replay_journal_seq	= last_seq;
+	j->replay_journal_seq	= last_written_seq;
 	j->replay_journal_seq_end = cur_seq;
-	j->last_seq_ondisk	= last_seq;
-	j->flushed_seq_ondisk	= cur_seq - 1;
-	j->seq_ondisk		= cur_seq - 1;
-	j->pin.front		= last_seq;
+	j->last_seq_ondisk	= last_written_seq;
+	j->flushed_seq_ondisk	= last_seq;
+	j->seq_ondisk		= last_seq;
+	j->pin.front		= last_written_seq;
 	j->pin.back		= cur_seq;
-	atomic64_set(&j->seq, cur_seq - 1);
+	atomic64_set(&j->seq, last_seq);
 
 	fifo_for_each_entry_ptr(p, &j->pin, seq)
 		journal_pin_list_init(p, 1);
@@ -1261,7 +1261,10 @@ int bch2_fs_journal_start(struct journal *j, u64 cur_seq)
 	}
 
 	if (!had_entries)
-		j->last_empty_seq = cur_seq;
+		j->last_empty_seq = last_seq;
+
+	WARN(j->last_empty_seq >= cur_seq, "journal startup error: last empty seq %llu is higher or equal than the next seq number to be used (%llu)",
+	      j->last_empty_seq, cur_seq);
 
 	spin_lock(&j->lock);
 
-- 
2.34.1