On 200205 1318, Darren Kenny wrote: > On Wed, Jan 29, 2020 at 05:34:24AM +0000, Bulekov, Alexander wrote: > > Signed-off-by: Alexander Bulekov <alx...@bu.edu> > > Reviewed-by: Stefan Hajnoczi <stefa...@redhat.com> > > --- > > <snip> > > + return allocate_objects(qts, current_path + 1, p_alloc); > > +} > > + > > +static const char *qos_build_main_args(void) > > +{ > > + char **path = fuzz_path_vec; > > Is it possible that fuzz_path_vec is not valid here? Specifically, > how likely is it that walk_path() won't ever set it, or that it > results in a possible previous value being used since we don't reset > it before calling qos_graph_foreach_test_path() in > qos_get_cmdline().
Hi Darren, Maybe this can happen if the target doesn't support a particular device. I added a check -Alex > Thanks, > > Darren. >