linux-kernel - Re: [PATCH] fs: xfs: Fix possible null-pointer dereferences in xchk_da_btree_block_check

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20190729145838.GN1561054@magnolia>
Date:   Mon, 29 Jul 2019 07:58:38 -0700
From:   "Darrick J. Wong" <darrick.wong@...cle.com>
To:     Jia-Ju Bai <baijiaju1990@...il.com>
Cc:     bfoster@...hat.com, sandeen@...deen.net, linux-xfs@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] fs: xfs: Fix possible null-pointer dereferences in
 xchk_da_btree_block_check_sibling()

On Mon, Jul 29, 2019 at 03:25:04PM +0800, Jia-Ju Bai wrote:
> 
> 
> On 2019/7/29 12:20, Darrick J. Wong wrote:
> > On Mon, Jul 29, 2019 at 11:24:01AM +0800, Jia-Ju Bai wrote:
> > > In xchk_da_btree_block_check_sibling(), there is an if statement on
> > > line 274 to check whether ds->state->altpath.blk[level].bp is NULL:
> > >      if (ds->state->altpath.blk[level].bp)
> > > 
> > > When ds->state->altpath.blk[level].bp is NULL, it is used on line 281:
> > >      xfs_trans_brelse(..., ds->state->altpath.blk[level].bp);
> > >          struct xfs_buf_log_item	*bip = bp->b_log_item;
> > >          ASSERT(bp->b_transp == tp);
> > > 
> > > Thus, possible null-pointer dereferences may occur.
> > > 
> > > To fix these bugs, ds->state->altpath.blk[level].bp is checked before
> > > being used.
> > > 
> > > These bugs are found by a static analysis tool STCheck written by us.
> > > 
> > > Signed-off-by: Jia-Ju Bai <baijiaju1990@...il.com>
> > > ---
> > >   fs/xfs/scrub/dabtree.c | 4 +++-
> > >   1 file changed, 3 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/fs/xfs/scrub/dabtree.c b/fs/xfs/scrub/dabtree.c
> > > index 94c4f1de1922..33ff90c0dd70 100644
> > > --- a/fs/xfs/scrub/dabtree.c
> > > +++ b/fs/xfs/scrub/dabtree.c
> > > @@ -278,7 +278,9 @@ xchk_da_btree_block_check_sibling(
> > >   	/* Compare upper level pointer to sibling pointer. */
> > >   	if (ds->state->altpath.blk[level].blkno != sibling)
> > >   		xchk_da_set_corrupt(ds, level);
> > > -	xfs_trans_brelse(ds->dargs.trans, ds->state->altpath.blk[level].bp);
> > > +	if (ds->state->altpath.blk[level].bp)
> > > +		xfs_trans_brelse(ds->dargs.trans,
> > > +						ds->state->altpath.blk[level].bp);
> > Indentation here (in xfs we use two spaces)
> 
> Okay, I will fix this.
> 
> > 
> > Also, uh, shouldn't we set ds->state->altpath.blk[level].bp to NULL
> > since we've released the buffer?
> 
> So I should set ds->state->altpath.blk[level].bp to NULL at the end of the
> function xchk_da_btree_block_check_sibling()?
> Like:
>     if (ds->state->altpath.blk[level].bp)
>         xfs_trans_brelse(ds->dargs.trans,
>                 ds->state->altpath.blk[level].bp);
>     ds->state->altpath.blk[level].bp = NULL;

You could put the whole thing in a single if clause, e.g.

	if (ds->state->altpath.blk[level].bp) {
		xfs_trans_brelse(ds->dargs.trans,
				 ds->state->altpath.blk[level].bp);
		ds->state->altpath.blk[level].bp = NULL;
	}

--D

> 
> 
> Best wishes,
> Jia-Ju Bai