Re: [PATCH bpf-next v2] selftests/bpf: Skip callback tests if jit is disabled in test_verifier

[Tiezhu Yang]

On 01/12/2024 12:21 PM, Hou Tao wrote:
> Hi,
>
> On 1/12/2024 9:57 AM, Tiezhu Yang wrote:
> > If CONFIG_BPF_JIT_ALWAYS_ON is not set and bpf_jit_enable is 0, there
> > exist 6 failed tests.

..

> > +static bool is_jit_enabled(void)
> > +{
> > +	const char *jit_sysctl = "/proc/sys/net/core/bpf_jit_enable";
> > +	bool enabled = false;
> > +	int sysctl_fd;
> > +
> > +	sysctl_fd = open(jit_sysctl, 0, O_RDONLY);
>
> It should be open(jit_sysctl, O_RDONLY).

Yes, this function comes from test_progs.c, I think
it is better to move it to testing_helpers.c with
this change.

> > +	if (sysctl_fd != -1) {
> > +		char tmpc;
> > +
> > +		if (read(sysctl_fd, &tmpc, sizeof(tmpc)) == 1)
> > +			enabled = (tmpc != '0');
> > +		close(sysctl_fd);
> > +	}
> > +
> > +	return enabled;
> > +}
> > +
> >  static int null_terminated_insn_len(struct bpf_insn *seq, int max_len)
> >  {
> >  	int i;
> > @@ -1662,6 +1691,16 @@ static void do_test_single(struct bpf_test *test, bool unpriv,
> >  		goto close_fds;
> >  	}
> >
> > +	if (!is_jit_enabled()) {
>
> Is it necessary to check whether jit is enabled or not each time ? Could
> we just check it only once just like unpriv_disabled does ?

Yes, it looks better, will modify the related code.

> > +		for (i = 0; i < prog_len; i++, prog++) {
>
> Is it better to only check pseudo_func only when both fd_prog < 0 and
> saved_errno == EINVAL are true, so unnecessary check can be skipped ?

Yes, will do it like this:

  if (fd_prog < 0 && saved_errno == EINVAL && jit_disabled)

Thanks,
Tiezhu