linux-ext4 - [PATCH -next 6/6] ext4: fix possible store wrong check interval value in disk when umount

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

Message-ID: <20210906144754.2601607-7-yebin10@huawei.com>
Date:   Mon, 6 Sep 2021 22:47:54 +0800
From:   Ye Bin <yebin10@...wei.com>
To:     <tytso@....edu>, <adilger.kernel@...ger.ca>,
        <linux-ext4@...r.kernel.org>
CC:     <linux-kernel@...r.kernel.org>, <jack@...e.cz>,
        Ye Bin <yebin10@...wei.com>
Subject: [PATCH -next 6/6] ext4: fix possible store wrong check interval value in disk when umount

Test follow steps:
1. mkfs.ext4 /dev/sda -O mmp
2. mount /dev/sda  /mnt
3. wait for about 1 minute
4. umount mnt
5. debugfs /dev/sda
6. dump_mmp
7. fsck.ext4 /dev/sda

I found 'check_interval' is range in [5, 10]. And sometime run fsck
print "MMP interval is 10 seconds and total wait time is 42 seconds.
Please wait...".
kmmpd:
...
	if (diff < mmp_update_interval * HZ)
		schedule_timeout_interruptible(mmp_update_interval * HZ - diff);
	 diff = jiffies - last_update_time;
...
	mmp_check_interval = max(min(EXT4_MMP_CHECK_MULT * diff / HZ,
				EXT4_MMP_MAX_CHECK_INTERVAL),
			        EXT4_MMP_MIN_CHECK_INTERVAL);
	mmp->mmp_check_interval = cpu_to_le16(mmp_check_interval);
...
We will call ext4_stop_mmpd to stop kmmpd kthread when umount, and
schedule_timeout_interruptible will be interrupted, so 'diff' maybe
little than mmp_update_interval. Then mmp_check_interval will range
in [EXT4_MMP_MAX_CHECK_INTERVAL, EXT4_MMP_CHECK_MULT * diff / HZ].
To solve this issue, if 'diff' little then mmp_update_interval * HZ
just break loop, don't update check interval.

Signed-off-by: Ye Bin <yebin10@...wei.com>
---
 fs/ext4/mmp.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/fs/ext4/mmp.c b/fs/ext4/mmp.c
index eed854bb6194..41c0a03019db 100644
--- a/fs/ext4/mmp.c
+++ b/fs/ext4/mmp.c
@@ -205,6 +205,14 @@ static int kmmpd(void *data)
 			schedule_timeout_interruptible(mmp_update_interval *
 						       HZ - diff);
 			diff = jiffies - last_update_time;
+			/* If 'diff' little 'than mmp_update_interval * HZ', it
+			 * means someone call ext4_stop_mmpd to stop kmmpd
+			 * kthread. We don't need to update mmp_check_interval
+			 * any more, as 'diff' is not exact value.
+			 */
+			if (unlikely(diff < mmp_update_interval * HZ &&
+			    kthread_should_stop()))
+				break;
 		}
 
 		/*
-- 
2.31.1