On Wed 10-08-22 15:20:48, Baokun Li wrote: > 在 2022/8/5 21:07, Lukas Czerner 写道: > > On Fri, Aug 05, 2022 at 08:39:47PM +0800, Baokun Li wrote: > > > I caught a null-ptr-deref bug as follows: > > > ================================================================== > > > KASAN: null-ptr-deref in range [0x0000000000000068-0x000000000000006f] > > > CPU: 1 PID: 1589 Comm: umount Not tainted 5.10.0-02219-dirty #339 > > > RIP: 0010:ext4_write_info+0x53/0x1b0 > > > [...] > > > Call Trace: > > > dquot_writeback_dquots+0x341/0x9a0 > > > ext4_sync_fs+0x19e/0x800 > > > __sync_filesystem+0x83/0x100 > > > sync_filesystem+0x89/0xf0 > > > generic_shutdown_super+0x79/0x3e0 > > > kill_block_super+0xa1/0x110 > > > deactivate_locked_super+0xac/0x130 > > > deactivate_super+0xb6/0xd0 > > > cleanup_mnt+0x289/0x400 > > > __cleanup_mnt+0x16/0x20 > > > task_work_run+0x11c/0x1c0 > > > exit_to_user_mode_prepare+0x203/0x210 > > > syscall_exit_to_user_mode+0x5b/0x3a0 > > > do_syscall_64+0x59/0x70 > > > entry_SYSCALL_64_after_hwframe+0x44/0xa9 > > > ================================================================== > > > > > > Above issue may happen as follows: > > > ------------------------------------- > > > exit_to_user_mode_prepare > > > task_work_run > > > __cleanup_mnt > > > cleanup_mnt > > > deactivate_super > > > deactivate_locked_super > > > kill_block_super > > > generic_shutdown_super > > > shrink_dcache_for_umount > > > dentry = sb->s_root > > > sb->s_root = NULL <--- Here set NULL > > > sync_filesystem > > > __sync_filesystem > > > sb->s_op->sync_fs > ext4_sync_fs > > > dquot_writeback_dquots > > > sb->dq_op->write_info > ext4_write_info > > > ext4_journal_start(d_inode(sb->s_root), EXT4_HT_QUOTA, 2) > > > d_inode(sb->s_root) > > > s_root->d_inode <--- Null pointer dereference > > > > > > To solve this problem, we use ext4_journal_start_sb directly > > > to avoid s_root being used. > > Are we syncing the file system after the superblock shutdown and getting > > away with it? This does not look good. Do you have a reproducer? > > > > Thanks! > > -Lukas > > Hi, Lukas! > > This problem is triggered by a pressure test when I reproduce another > problem. > > So I didn't have the reproducer. > > I looked at the error stack code and found that it seems there is something > wrong. > > Moreover, it's really weird to write code like > "ext4_journal_start(d_inode(sb->s_root), ...)". I agree. That looks like some leftover from the past. Feel free to add: Reviewed-by: Jan Kara <jack@xxxxxxx> Honza > > > > Signed-off-by: Baokun Li <libaokun1@xxxxxxxxxx> > > > --- > > > fs/ext4/super.c | 2 +- > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > diff --git a/fs/ext4/super.c b/fs/ext4/super.c > > > index 9a66abcca1a8..0ce4565422f6 100644 > > > --- a/fs/ext4/super.c > > > +++ b/fs/ext4/super.c > > > @@ -6653,7 +6653,7 @@ static int ext4_write_info(struct super_block *sb, int type) > > > handle_t *handle; > > > /* Data block + inode block */ > > > - handle = ext4_journal_start(d_inode(sb->s_root), EXT4_HT_QUOTA, 2); > > > + handle = ext4_journal_start_sb(sb, EXT4_HT_QUOTA, 2); > > > if (IS_ERR(handle)) > > > return PTR_ERR(handle); > > > ret = dquot_commit_info(sb, type); > > > -- > > > 2.31.1 > > > > > . > > > Thanks! > > -- > With Best Regards, > Baokun Li > -- Jan Kara <jack@xxxxxxxx> SUSE Labs, CR
