Re: [PATCH 1/4] befs: check allocation_group number before use

[Salah Triki]

On Thu, Aug 11, 2016 at 11:56:16AM +0100, Luis de Bethencourt wrote:
> On 10/08/16 23:12, Salah Triki wrote:
> > Check that the allocation group number is not greater or equal to the
> > number of allocations group in the file system and return BEF_ERR in the
> > case of error.
> > 
> > Signed-off-by: Salah Triki <salah.triki@xxxxxxxxx>
> > ---
> >  fs/befs/befs.h | 5 +++++
> >  1 file changed, 5 insertions(+)
> > 
> > diff --git a/fs/befs/befs.h b/fs/befs/befs.h
> > index 55f3ea2..6daf4c4 100644
> > --- a/fs/befs/befs.h
> > +++ b/fs/befs/befs.h
> > @@ -121,6 +121,11 @@ BEFS_I(const struct inode *inode)
> >  static inline befs_blocknr_t
> >  iaddr2blockno(struct super_block *sb, const befs_inode_addr *iaddr)
> >  {
> > +	if (iaddr->allocation_group >= BEFS_SB(sb)->num_ags) {
> > +		befs_error(sb, "BEFS: Invalid allocation group %u, max is %u",
> > +			   iaddr->allocation_group, BEFS_SB(sb)->num_ags);
> > +		return BEFS_ERR;
> > +	}
> >  	return ((iaddr->allocation_group << BEFS_SB(sb)->ag_shift) +
> >  		iaddr->start);
> >  }
> > 
> 
> Hi Salah,
> 
> To understand why would we check for this. When can this error happen? I mean,
> when can iaddr2blockno() be called with an out of bounds allocation group?
> 
> Thanks,
> Luis

Hi Luis,

allocation group number is set to blockno >> BEFS_SB(sb)->ag_shift by 
blockno2iaddr(), so if ag_shift is not valid, an out of bound may occur. 
By now, thanx to your question ;), I think it's better to check the validity 
of ag_shift when the superblock is loaded.

Nack.

Thanx for the question :)
Salah