On Wed, Feb 28, 2024 at 3:44 PM Eduard Zingerman <eddyz87@xxxxxxxxx> wrote:
>
> On Wed, 2024-02-28 at 15:40 -0800, Andrii Nakryiko wrote:
> [...]
>
> > > +static libbpf_print_fn_t old_print_cb;
> > > +static bool msg_found;
> > > +
> > > +static int print_cb(enum libbpf_print_level level, const char *fmt, va_list args)
> > > +{
> > > + old_print_cb(level, fmt, args);
> > > + if (level == LIBBPF_WARN && strncmp(fmt, EXPECTED_MSG, strlen(EXPECTED_MSG)) == 0)
> > > + msg_found = true;
> > > +
> > > + return 0;
> > > +}
> > > +
> > > +static void test_bad_struct_ops(void)
> > > +{
> > > + struct bad_struct_ops *skel;
> > > + int err;
> > > +
> > > + old_print_cb = libbpf_set_print(print_cb);
> > > + skel = bad_struct_ops__open_and_load();
> >
> > we want to check that the load step failed specifically, right? So
> > please split open from load, make sure that open succeeds, but load
> > fails
>
> Ok
>
> >
> > > + err = errno;
> > > + libbpf_set_print(old_print_cb);
> > > + if (!ASSERT_NULL(skel, "bad_struct_ops__open_and_load"))
> > > + return;
> > > +
> > > + ASSERT_EQ(err, EINVAL, "errno should be EINVAL");
> > > + ASSERT_TRUE(msg_found, "expected message");
> > > +
> > > + bad_struct_ops__destroy(skel);
> > > +}
> > > +
> > > +void serial_test_bad_struct_ops(void)
> >
> > why does it have to be a serial test?
>
> Because it hijacks libbpf print callback.
each non-serial test runs in its own *process*, there is no
multi-threading here, so it's fine if non-serial test temporarily
hijacks print callback, as long as it restores it properly before
finishing
>
> [...]
