Hi Dan,
thanks for your comments and yes, this looks wrong. In Patch 5 of this series,
the whole crap is therefore being removed.
Marc
> On Wed, Aug 24, 2011 at 08:23:07PM +0200, Marc Dietrich wrote:
> > @@ -139,7 +159,7 @@ static void nvec_dispatch(struct work_struct *work)
> >
> > } else {
> >
> > parse_msg(nvec, msg);
>
> This isn't a comment on this patch, but just something I noticed in
> passing. parse_msg() dereferences msg so check on the next line is
> too late.
>
> > if((!msg) || (!msg->data))
>
> ^^^^^^
> Too late.
>
> > - dev_warn(nvec->dev, "attempt access zero
> > pointer");
> > + dev_warn(nvec->dev, "attempt access zero
> > pointer\n");
> >
> > else {
> >
> > kfree(msg->data);
>
> ^^^^^^^^^
> Another inconsistence dereference here (would oops if msg were NULL).
>
> > kfree(msg);
>
> regards,
> dan carpenter
_______________________________________________
devel mailing list
[email protected]
http://driverdev.linuxdriverproject.org/mailman/listinfo/devel
