Re: [PATCH] ocfs2: dereference before NULL check in ocfs2_direct_IO_get_blocks()

[Arjan van de Ven <arjan@infradead.org>]

On Sat, 2006-06-03 at 14:16 -0700, Joel Becker wrote:
> On Sat, Jun 03, 2006 at 11:15:58PM +0400, Alexey Dobriyan wrote:
> > On Sat, Jun 03, 2006 at 11:34:39AM -0400, Florin Malita wrote:
> > > 'bh_result' & 'inode' are explicitly checked against NULL so they
> > > shouldn't be dereferenced prior to that.
> > >
> > > Coverity ID: 1273, 1274.
> > 
> > AFAICS, the patch is BS, as usual with this type of patches.
> > 
> > Can "inode" and "bh_result" be NULL here? I bet they can't.
> 
> 	This is a common result of this sort of scan.  The scan merely
> provides good information, not a perfect patch.  There are two
> possibilities:
> 
> 	1) The scan is right, and the dereference is dangerous.  The
> 	   patch is correct.
> 	2) The dereference is not dangerous ("can't happen"), and the
> 	   later check for NULL is spurious.  A correct patch would
> 	   merely remove the check.
> 
> 	This is clearly a case of (2), but I bet that (1) is seen just
> as often.

and in case of (2); newer gcc is often smart enough to do that
automatically :)

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/