Re: [PATCH 3/3] ntfs: check overflow when iterates ATTR_RECORDs

[Hawkins Jiawei]

On Wed, 31 Aug 2022 at 18:13, Dan Carpenter <dan.carpenter@xxxxxxxxxx> wrote:
>
> On Wed, Aug 31, 2022 at 10:48:54AM +0800, Hawkins Jiawei wrote:
> > Kernel will iterates over ATTR_RECORDs in mft record in ntfs_attr_find().
> > Because the ATTR_RECORDs are next to each other, kernel can get the next
> > ATTR_RECORD from end address of current ATTR_RECORD, through current
> > ATTR_RECORD length field.
> >
> > The problem is that during iteration, when kernel calculates the end address
> > of current ATTR_RECORD, kernel may trigger an overflow bug in
> > executing `a = (ATTR_RECORD*)((u8*)a + le32_to_cpu(a->length))`. This
> > may wrap, leading to a forever iteration on 32bit systems.
> >
> > This patch solves it by adding an overflow checking on calculating end address
> > of current ATTR_RECORD during iteration.
> >
> > Suggested-by: Dan Carpenter <dan.carpenter@xxxxxxxxxx>
> > Link: https://lore.kernel.org/all/20220827105842.GM2030@kadam/
> > Signed-off-by: Hawkins Jiawei <yin31149@xxxxxxxxx>
> > ---
> >  fs/ntfs/attrib.c | 3 +++
> >  1 file changed, 3 insertions(+)
> >
> > diff --git a/fs/ntfs/attrib.c b/fs/ntfs/attrib.c
> > index 904734e34507..55e618c9e63e 100644
> > --- a/fs/ntfs/attrib.c
> > +++ b/fs/ntfs/attrib.c
> > @@ -617,6 +617,9 @@ static int ntfs_attr_find(const ATTR_TYPE type, const ntfschar *name,
> >                       return -ENOENT;
> >               if (unlikely(!a->length))
> >                       break;
> > +             /* check for wrap around */
> > +             if ((u8 *)a + le32_to_cpu(a->length) < (u8 *)a)
> > +                     break;
>
> Wouldn't it also be good to check that a + a->length <= mrec_end?
> It gets checked on the next iteration sure, but it just seems like a
> reasonable thing to check here.
Hi Dan,
Thanks for your suggestion, I will add this check in my v2 patch!