Re: [PATCH v8 4/4] scanf: break kunit into test cases

[Tamir Duberstein]

On Wed, Mar 5, 2025 at 10:01 AM Petr Mladek <pmladek@xxxxxxxx> wrote:
>
> On Fri 2025-02-14 11:20:01, Tamir Duberstein wrote:
> > Use `suite_init` and move some tests into `scanf_test_cases`. This
> > gives us nicer output in the event of a failure.
> >
> > Reviewed-by: David Gow <davidgow@xxxxxxxxxx>
> > Signed-off-by: Tamir Duberstein <tamird@xxxxxxxxx>
> > ---
> >  lib/tests/scanf_kunit.c | 95 ++++++++++++++++++++++++++-----------------------
> >  1 file changed, 51 insertions(+), 44 deletions(-)
> >
> > diff --git a/lib/tests/scanf_kunit.c b/lib/tests/scanf_kunit.c
> > index 3bbad9ebe437..fa215a7db366 100644
> > --- a/lib/tests/scanf_kunit.c
> > +++ b/lib/tests/scanf_kunit.c
> > @@ -4,14 +4,10 @@
> >   */
> >
> >  #include <kunit/test.h>
> > -#include <linux/bitops.h>
> > -#include <linux/kernel.h>
> >  #include <linux/module.h>
> > -#include <linux/overflow.h>
> > -#include <linux/printk.h>
> >  #include <linux/prandom.h>
> >  #include <linux/slab.h>
> > -#include <linux/string.h>
> > +#include <linux/sprintf.h>
> >
> >  #define BUF_SIZE 1024
>
> It would make more sense to do this clean up in the 3rd patch
> where some code was replaced by the kunit macros.
>
> Also I am not sure about the choice. It might make sense to remove
> <include/printk.h> because the pr_*() calls were removed.
> But what about the others? Did anyone request the clean up, please?
>
> I do not want to open a bike shadding because different people
> have different opinion.
>
> I would personally prefer to keep the explicit includes when the
> related API is still used. It helps to optimize nested includes
> in the header files which helps to speedup build. AFAIK, there
> are people working in this optimization and they might need
> to revert this change.

Yeah, I don't feel strongly. I'll just restore all the includes.

> > @@ -50,10 +46,9 @@ do {                                                                               \
> >       for (; n_args > 0; n_args--, expect++) {                                \
> >               typeof(*expect) got = *va_arg(ap, typeof(expect));              \
> >               if (got != *expect) {                                           \
> > -                     KUNIT_FAIL(test,                                        \
> > -                                "%s:%d: vsscanf(\"%s\", \"%s\", ...) expected " arg_fmt " got " arg_fmt, \
> > -                                file, line, str, fmt, *expect, got);         \
> > -                     return;                                                 \
> > +                     KUNIT_FAIL_AND_ABORT(test,                              \
> > +                                          "%s:%d: vsscanf(\"%s\", \"%s\", ...) expected " arg_fmt " got " arg_fmt, \
> > +                                          file, line, str, fmt, *expect, got); \
>
> I am just curious. Is there any particular reason why
> KUNIT_FAIL() is replaced with KUNIT_FAIL_AND_ABORT()?
>
> Did the move of some tests into KUNIT_CASE() increased the number of
> reported errors?
>
> Why is _ABORT() variant used in _check_numbers_template() and not in _test()?
>
> I do not have strong opinion. The change just looks a bit ad-hoc and
> inconsistent.
>
>
> >               }                                                               \
> >       }                                                                       \
> >  } while (0)

Honestly I don't remember. The effect is definitely to kill tests
earlier in the case of failure, but you're right to point out the
inconsistency with _test. The original code had the same behavior in
both cases, so I will restore that.

>
> Otherwise, the change looks good to me.
>
> Best Regards,
> Petr
>
> PS: I suggest to wait at least one or two days with the respin. Other
>     reviewers might want to add their own opinion.

Will do. Thanks for the reviews.
Tamir