Re: [RFC bpf-next v1 5/8] selftests/bpf: no need to track next_match_pos in struct test_loader

[Eduard Zingerman <eddyz87@gmail.com>]

On Mon, 2024-07-01 at 17:41 -0700, Andrii Nakryiko wrote:

[...]

> >  static void emit_verifier_log(const char *log_buf, bool force)
> > @@ -450,23 +449,23 @@ static void validate_case(struct test_loader *tester,
> >                           struct bpf_program *prog,
> >                           int load_err)
> >  {
> > -       int i, j, err;
> > -       char *match;
> >         regmatch_t reg_match[1];
> > +       const char *match;
> > +       const char *log = tester->log_buf;
> > +       int i, j, err;
> > 
> >         for (i = 0; i < subspec->expect_msg_cnt; i++) {
> >                 struct expect_msg *msg = &subspec->expect_msgs[i];
> > 
> >                 if (msg->substr) {
> > -                       match = strstr(tester->log_buf + tester->next_match_pos, msg->substr);
> > +                       match = strstr(log, msg->substr);
> >                         if (match)
> > -                               tester->next_match_pos = match - tester->log_buf + strlen(msg->substr);
> > +                               log += strlen(msg->substr);
> >                 } else {
> > -                       err = regexec(&msg->regex,
> > -                                     tester->log_buf + tester->next_match_pos, 1, reg_match, 0);
> > +                       err = regexec(&msg->regex, log, 1, reg_match, 0);
> >                         if (err == 0) {
> > -                               match = tester->log_buf + tester->next_match_pos + reg_match[0].rm_so;
> > -                               tester->next_match_pos += reg_match[0].rm_eo;
> > +                               match = log + reg_match[0].rm_so;
> > +                               log += reg_match[0].rm_eo;
> 
> invert and simplify:
> 
> log += reg_match[0].rm_eo;
> match = log;
> 
> ?

The 'match' is at 'log + rm_so' (start offset).
The 'log'   is at 'log + rm_eo' (end offset).

The brilliance of standard library naming.

> 
> >                         } else {
> >                                 match = NULL;
> >                         }
> 
> how about we move this to the beginning of iteration (before `if
> (msg->substr)`) and so we'll assume the match is NULL on regexec
> failing?

Ok, but this would require explicit match re-initialization to NULL at
each iteration.

[...]