This is a note to let you know that I've just added the patch titled
f2fs: fix to drop meta_inode's page cache in f2fs_put_super()
to the 6.6-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:
f2fs-fix-to-drop-meta_inode-s-page-cache-in-f2fs_put.patch
and it can be found in the queue-6.6 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@xxxxxxxxxxxxxxx> know about it.
commit 3880a6f9f34f01e52e73411d6c748768f5d9ca7a
Author: Chao Yu <chao@xxxxxxxxxx>
Date: Tue Sep 5 12:57:53 2023 +0800
f2fs: fix to drop meta_inode's page cache in f2fs_put_super()
[ Upstream commit a4639380bbe66172df329f8b54aa7d2e943f0f64 ]
syzbot reports a kernel bug as below:
F2FS-fs (loop1): detect filesystem reference count leak during umount, type: 10, count: 1
kernel BUG at fs/f2fs/super.c:1639!
CPU: 0 PID: 15451 Comm: syz-executor.1 Not tainted 6.5.0-syzkaller-09338-ge0152e7481c6 #0
RIP: 0010:f2fs_put_super+0xce1/0xed0 fs/f2fs/super.c:1639
Call Trace:
generic_shutdown_super+0x161/0x3c0 fs/super.c:693
kill_block_super+0x3b/0x70 fs/super.c:1646
kill_f2fs_super+0x2b7/0x3d0 fs/f2fs/super.c:4879
deactivate_locked_super+0x9a/0x170 fs/super.c:481
deactivate_super+0xde/0x100 fs/super.c:514
cleanup_mnt+0x222/0x3d0 fs/namespace.c:1254
task_work_run+0x14d/0x240 kernel/task_work.c:179
resume_user_mode_work include/linux/resume_user_mode.h:49 [inline]
exit_to_user_mode_loop kernel/entry/common.c:171 [inline]
exit_to_user_mode_prepare+0x210/0x240 kernel/entry/common.c:204
__syscall_exit_to_user_mode_work kernel/entry/common.c:285 [inline]
syscall_exit_to_user_mode+0x1d/0x60 kernel/entry/common.c:296
do_syscall_64+0x44/0xb0 arch/x86/entry/common.c:86
entry_SYSCALL_64_after_hwframe+0x63/0xcd
In f2fs_put_super(), it tries to do sanity check on dirty and IO
reference count of f2fs, once there is any reference count leak,
it will trigger panic.
The root case is, during f2fs_put_super(), if there is any IO error
in f2fs_wait_on_all_pages(), we missed to truncate meta_inode's page
cache later, result in panic, fix this case.
Fixes: 20872584b8c0 ("f2fs: fix to drop all dirty meta/node pages during umount()")
Reported-by: syzbot+ebd7072191e2eddd7d6e@xxxxxxxxxxxxxxxxxxxxxxxxx
Closes: https://lore.kernel.org/linux-f2fs-devel/000000000000a14f020604a62a98@xxxxxxxxxx
Signed-off-by: Chao Yu <chao@xxxxxxxxxx>
Signed-off-by: Jaegeuk Kim <jaegeuk@xxxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index 12790bc6e0739..bc303a0522155 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@ -1662,7 +1662,7 @@ static void f2fs_put_super(struct super_block *sb)
f2fs_wait_on_all_pages(sbi, F2FS_WB_CP_DATA);
- if (err) {
+ if (err || f2fs_cp_error(sbi)) {
truncate_inode_pages_final(NODE_MAPPING(sbi));
truncate_inode_pages_final(META_MAPPING(sbi));
}
