On Thu 22-09-22 16:13:59, Zhihao Cheng wrote:
> 在 2022/9/21 21:37, Jan Kara 写道:
> > > --- a/fs/quota/quota_tree.c
> > > +++ b/fs/quota/quota_tree.c
> > > @@ -71,6 +71,35 @@ static ssize_t write_blk(struct qtree_mem_dqinfo *info, uint blk, char *buf)
> > > return ret;
> > > }
> > > +static inline int do_check_range(struct super_block *sb, uint val, uint max_val)
> > > +{
> > > + if (val >= max_val) {
> > > + quota_error(sb, "Getting block too big (%u >= %u)",
> > > + val, max_val);
> > > + return -EUCLEAN;
> > > + }
> > > +
> > > + return 0;
> > > +}
> >
> > I'd already provide min_val and the string for the message here as well (as
> > you do in patch 2). It is less churn in the next patch and free blocks
> > checking actually needs that as well. See below.
> >
> > > +
> > > +static int check_free_block(struct qtree_mem_dqinfo *info,
> > > + struct qt_disk_dqdbheader *dh)
> > > +{
> > > + int err = 0;
> > > + uint nextblk, prevblk;
> > > +
> > > + nextblk = le32_to_cpu(dh->dqdh_next_free);
> > > + err = do_check_range(info->dqi_sb, nextblk, info->dqi_blocks);
> > > + if (err)
> > > + return err;
> > > + prevblk = le32_to_cpu(dh->dqdh_prev_free);
> > > + err = do_check_range(info->dqi_sb, prevblk, info->dqi_blocks);
> > > + if (err)
> > > + return err;
> >
> > The free block should actually be > QT_TREEOFF so I'd add the check to
> > do_check_range().
>
> 'dh->dqdh_next_free' may be updated when quota entry removed,
> 'dh->dqdh_next_free' can be used for next new quota entris.
> Before sending v2, I found 'dh->dqdh_next_free' and 'dh->dqdh_prev_free' can
> easily be zero in newly allocated blocks when continually creating files
> onwed by different users:
> find_free_dqentry
> get_free_dqblk
> write_blk(info, info->dqi_blocks, buf) // zero'd qt_disk_dqdbheader
> blk = info->dqi_blocks++ // allocate new one block
> info->dqi_free_entry = blk // will be used for new quota entries
>
> find_free_dqentry
> if (info->dqi_free_entry)
> blk = info->dqi_free_entry
> read_blk(info, blk, buf) // dh->dqdh_next_free = dh->dqdh_prev_free =
> 0
>
> I think it's normal when 'dh->dqdh_next_free' or 'dh->dqdh_prev_free' equals
> to 0.
Good point! 0 means "not present". So any block number (either in free list
or pointed from the quota tree) should be either 0 or > QT_TREEOFF.
Honza
--
Jan Kara <jack@xxxxxxxx>
SUSE Labs, CR
