Re: [PATCH 3/4] tools: bpftool: fix potential NULL pointer dereference in do_load

[Y Song]

On Wed, Nov 21, 2018 at 9:30 AM Jakub Kicinski
<jakub.kicinski@xxxxxxxxxxxxx> wrote:
>
> On Wed, 21 Nov 2018 09:18:06 -0800, Y Song wrote:
> > On Tue, Nov 20, 2018 at 11:42 PM Wen Yang <wen.yang99@xxxxxxxxxx> wrote:
> > >
> > > This patch fixes a possible null pointer dereference in
> > > do_load, detected by the semantic patch
> > > deref_null.cocci, with the following warning:
> > >
> > > ./tools/bpf/bpftool/prog.c:1021:23-25: ERROR: map_replace is NULL but dereferenced.
> > >
> > > The following code has potential null pointer references:
> > >  881             map_replace = reallocarray(map_replace, old_map_fds + 1,
> > >  882                                        sizeof(*map_replace));
> > >  883             if (!map_replace) {
> > >  884                     p_err("mem alloc failed");
> > >  885                     goto err_free_reuse_maps;
> > >  886             }
> > >
> > > ...
> > > 1019 err_free_reuse_maps:
> > > 1020         for (i = 0; i < old_map_fds; i++)
> > > 1021                 close(map_replace[i].fd);
> > > 1022         free(map_replace);
> >
> > I did not see any issues here.
> > We have code looks like:
> >  867         struct map_replace *map_replace = NULL;
> >  868         struct bpf_program *prog = NULL, *pos;
> >  869         unsigned int old_map_fds = 0;
> > ...
> >  948                         map_replace = reallocarray(map_replace,
> > old_map_fds + 1,
> >  949                                                    sizeof(*map_replace));
> >  950                         if (!map_replace) {
> >  951                                 p_err("mem alloc failed");
> >  952                                 goto err_free_reuse_maps;
> >  953                         }
> >  954                         map_replace[old_map_fds].idx = idx;
> >  955                         map_replace[old_map_fds].name = name;
> >  956                         map_replace[old_map_fds].fd = fd;
> >  957                         old_map_fds++;
> > ...
> >
> > old_map_fds becomes non zero if and only if map_replace is not NULL.
>
> Note that this is a realloc in a loop, and map_replace will become NULL
> again if realloc fails.  We should check the return value of realloc()
> without loosing the pointer to the old value.  No?

Or right. Agreed. The right fix seems to restore the old map_replace
in the error path
and jump to err_free_reuse_maps if reallocarray fails.

>
> > > Signed-off-by: Wen Yang <wen.yang99@xxxxxxxxxx>
> > > Reviewed-by: Tan Hu <tan.hu@xxxxxxxxxx>
> > > CC: Julia Lawall <julia.lawall@xxxxxxx>
> > > ---
> > >  tools/bpf/bpftool/prog.c | 5 +++--
> > >  1 file changed, 3 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/tools/bpf/bpftool/prog.c b/tools/bpf/bpftool/prog.c
> > > index 5302ee2..de42187 100644
> > > --- a/tools/bpf/bpftool/prog.c
> > > +++ b/tools/bpf/bpftool/prog.c
> > > @@ -1017,8 +1017,9 @@ static int do_load(int argc, char **argv)
> > >  err_close_obj:
> > >         bpf_object__close(obj);
> > >  err_free_reuse_maps:
> > > -       for (i = 0; i < old_map_fds; i++)
> > > -               close(map_replace[i].fd);
> > > +       if (map_replace)
> > > +               for (i = 0; i < old_map_fds; i++)
> > > +                       close(map_replace[i].fd);
> > >         free(map_replace);
> > >         return -1;
> > >  }
> > > --
> > > 2.9.5
> > >
>