linux-kernel-mentees.lists.linuxfoundation.org archive mirror
 help / color / mirror / Atom feed
* [PATCH] fs: fix WARNING in mark_buffer_dirty (4)
       [not found] <CACT4Y+bUtBuhD7_BAN+NavEfhBNOavqF0CJkrZ+Gc4pYeLiy+g@mail.gmail.com>
@ 2022-08-21 12:10 ` Hawkins Jiawei
  0 siblings, 0 replies; only message in thread
From: Hawkins Jiawei @ 2022-08-21 12:10 UTC (permalink / raw)
  To: dvyukov
  Cc: paskripkin, syzkaller-bugs, linux-kernel, willy, syzkaller, viro,
	linux-fsdevel, Hawkins Jiawei, linux-kernel-mentees,
	syzbot+2af3bc9585be7f23f290

Syzkaller reports bug as follows:
------------[ cut here ]------------
WARNING: CPU: 0 PID: 3684 at fs/buffer.c:1081 mark_buffer_dirty+0x59d/0xa20 fs/buffer.c:1081
[...]
Call Trace:
 <TASK>
 minix_put_super+0x199/0x500 fs/minix/inode.c:49
 generic_shutdown_super+0x14c/0x400 fs/super.c:462
 kill_block_super+0x97/0xf0 fs/super.c:1394
 deactivate_locked_super+0x94/0x160 fs/super.c:332
 deactivate_super+0xad/0xd0 fs/super.c:363
 cleanup_mnt+0x3a2/0x540 fs/namespace.c:1186
 task_work_run+0xdd/0x1a0 kernel/task_work.c:177
 ptrace_notify+0x114/0x140 kernel/signal.c:2353
 ptrace_report_syscall include/linux/ptrace.h:420 [inline]
 ptrace_report_syscall_exit include/linux/ptrace.h:482 [inline]
 syscall_exit_work kernel/entry/common.c:249 [inline]
 syscall_exit_to_user_mode_prepare+0x129/0x280 kernel/entry/common.c:276
 __syscall_exit_to_user_mode_work kernel/entry/common.c:281 [inline]
 syscall_exit_to_user_mode+0x9/0x50 kernel/entry/common.c:294
 do_syscall_64+0x42/0xb0 arch/x86/entry/common.c:86
 entry_SYSCALL_64_after_hwframe+0x63/0xcd
 [...]
 </TASK>
------------------------------------

During VFS releasing the minix's superblock, kernel will calls
sync_filesystem() to write out and wait upon all dirty data
associated with this superblock.

Yet the problem is that this write may fail, then kernel will
clear BH_Uptodate flag in superblock's struct buffer_head
in end_buffer_async_write(). When kernel returns from
sync_filesystem() and calls sop->put_super()
(which is minix_put_super()), it will triggers the warning
for struct buffer_head is not uptodate in mark_buffer_dirty().

This patch solves it by handling sync_filesystem() write error
in minix_put_super(), before calling mark_buffer_dirty()

Reported-and-tested-by: syzbot+2af3bc9585be7f23f290@syzkaller.appspotmail.com
Signed-off-by: Hawkins Jiawei <yin31149@gmail.com>
---
 fs/minix/inode.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

diff --git a/fs/minix/inode.c b/fs/minix/inode.c
index da8bdd1712a7..8e9a8057dcfe 100644
--- a/fs/minix/inode.c
+++ b/fs/minix/inode.c
@@ -42,17 +42,27 @@ static void minix_put_super(struct super_block *sb)
 {
 	int i;
 	struct minix_sb_info *sbi = minix_sb(sb);
+	struct buffer_head *sbh = sbi->s_sbh;
 
 	if (!sb_rdonly(sb)) {
 		if (sbi->s_version != MINIX_V3)	 /* s_state is now out from V3 sb */
 			sbi->s_ms->s_state = sbi->s_mount_state;
-		mark_buffer_dirty(sbi->s_sbh);
+
+		lock_buffer(sbh);
+		if (buffer_write_io_error(sbh)) {
+			clear_buffer_write_io_error(sbh);
+			set_buffer_uptodate(sbh);
+			printk("MINIX-fs warning: superblock detected "
+			       "previous I/O error\n");
+		}
+		mark_buffer_dirty(sbh);
+		unlock_buffer(sbh);
 	}
 	for (i = 0; i < sbi->s_imap_blocks; i++)
 		brelse(sbi->s_imap[i]);
 	for (i = 0; i < sbi->s_zmap_blocks; i++)
 		brelse(sbi->s_zmap[i]);
-	brelse (sbi->s_sbh);
+	brelse (sbh);
 	kfree(sbi->s_imap);
 	sb->s_fs_info = NULL;
 	kfree(sbi);
-- 
2.25.1

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] only message in thread

only message in thread, other threads:[~2022-08-21 12:11 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
     [not found] <CACT4Y+bUtBuhD7_BAN+NavEfhBNOavqF0CJkrZ+Gc4pYeLiy+g@mail.gmail.com>
2022-08-21 12:10 ` [PATCH] fs: fix WARNING in mark_buffer_dirty (4) Hawkins Jiawei

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).