Received-SPF: pass (google.com: best guess record for domain of linux-ext4-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Tue, 31 Mar 2020 16:30:35 +0200
From:   Jan Kara <jack@suse.cz>
To:     Lukas Czerner <lczerner@redhat.com>
Cc:     Jan Kara <jack@suse.cz>, Ted Tso <tytso@mit.edu>,
        linux-ext4@vger.kernel.org
Subject: Re: [PATCH 2/2] ext2fs: Fix off-by-one in dx_grow_tree()
Message-ID: <20200331143035.GB13528@quack2.suse.cz>
References: <20200330090932.29445-1-jack@suse.cz>
 <20200330090932.29445-3-jack@suse.cz>
 <20200330132712.ckevhpof4vrsx5rw@work>
 <20200330145531.GF26544@quack2.suse.cz>
 <20200331113303.huhzo3jxdnhoupwv@work>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20200331113303.huhzo3jxdnhoupwv@work>
User-Agent: Mutt/1.10.1 (2018-07-13)
Sender: linux-ext4-owner@vger.kernel.org
Precedence: bulk

On Tue 31-03-20 13:33:03, Lukas Czerner wrote:
> On Mon, Mar 30, 2020 at 04:55:31PM +0200, Jan Kara wrote:
> > On Mon 30-03-20 15:27:12, Lukas Czerner wrote:
> > > On Mon, Mar 30, 2020 at 11:09:32AM +0200, Jan Kara wrote:
> > > > There is an off-by-one error in dx_grow_tree() when checking whether we
> > > > can add another level to the tree. Thus we can grow tree too much
> > > > leading to possible crashes in the library or corrupted filesystem. Fix
> > > > the bug.
> > > 
> > > Looks good, thanks
> > > 
> > > Reviewed-by: Lukas Czerner <lczerner@redhat.com>
> > > 
> > > Don't we have basically the same off-by-one in
> > > e2fsck/pass1.c handle_htree() ?
> > > 
> > >        if ((root->indirect_levels > ext2_dir_htree_level(fs)) &&
> > >            fix_problem(ctx, PR_1_HTREE_DEPTH, pctx))
> > 
> > I don't think so. It is indeed correct for root->indirect_levels to be
> > equal to ext2_dir_htree_level(). However dx_grow_tree() is going to
> > increase root->indirect_levels by 1 which is where tree would become
> > invalid...
> > 
> > 								Honza
> 
> Hmm, are you sure ? I think the names are very confusing
> 
> root->indirect_levels is zero based, while ext2_dir_htree_level()
> returns the maximum number of levels (that is 3 by default). If I am
> right then indirect_levels must always be smaller then
> ext2_dir_htree_level() and that is how we use it everywhere else - the
> palce I am pointing out is an exception and I think it's a bug.
> 
> Indeed it looks like the bug got introduced in
> 3f0cf647539970474be8f607017ca7eccfc2fbbe
> 
> -       if ((root->indirect_levels > 1) &&
> +       if ((root->indirect_levels > ext2_dir_htree_level(fs)) &&
> 
> Or am I missing something ?

Ah, you're indeed right! e2fsck/pass2.c even has a correct version of the
condition. Just the condition in pass1.c is wrong.

								Honza

> 
> -Lukas
> 
> > 
> > > > 
> > > > Signed-off-by: Jan Kara <jack@suse.cz>
> > > > ---
> > > >  lib/ext2fs/link.c | 2 +-
> > > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > > 
> > > > diff --git a/lib/ext2fs/link.c b/lib/ext2fs/link.c
> > > > index 7b5bb022117c..469eea8cd06d 100644
> > > > --- a/lib/ext2fs/link.c
> > > > +++ b/lib/ext2fs/link.c
> > > > @@ -473,7 +473,7 @@ static errcode_t dx_grow_tree(ext2_filsys fs, ext2_ino_t dir,
> > > >  		    ext2fs_le16_to_cpu(info->frames[i].head->limit))
> > > >  			break;
> > > >  	/* Need to grow tree depth? */
> > > > -	if (i < 0 && info->levels > ext2_dir_htree_level(fs))
> > > > +	if (i < 0 && info->levels >= ext2_dir_htree_level(fs))
> > > >  		return EXT2_ET_DIR_NO_SPACE;
> > > >  	lblk = size / fs->blocksize;
> > > >  	size += fs->blocksize;
> > > > -- 
> > > > 2.16.4
> > > > 
> > > 
> > -- 
> > Jan Kara <jack@suse.com>
> > SUSE Labs, CR
> > 
> 
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR