Re: [PATCH] ww_mutex: convert self-test to KUnit

[Tamir Duberstein <tamird@gmail.com>]

Hi Dan,

On Wed, Feb 12, 2025 at 6:53 AM Dan Carpenter <dan.carpenter@linaro.org> wrote:
>
> Hi Tamir,
>
> kernel test robot noticed the following build warnings:
>
> url:    https://github.com/intel-lab-lkp/linux/commits/Tamir-Duberstein/ww_mutex-convert-self-test-to-KUnit/20250211-000245
> base:   a64dcfb451e254085a7daee5fe51bf22959d52d3
> patch link:    https://lore.kernel.org/r/20250210-ww_mutex-kunit-convert-v1-1-972f0201f71e%40gmail.com
> patch subject: [PATCH] ww_mutex: convert self-test to KUnit
> config: i386-randconfig-141-20250212 (https://download.01.org/0day-ci/archive/20250212/202502121806.CS6r741y-lkp@intel.com/config)
> compiler: clang version 19.1.3 (https://github.com/llvm/llvm-project ab51eccf88f5321e7c60591c5546b254b6afab99)
>
> If you fix the issue in a separate patch/commit (i.e. not just a new version of
> the same patch/commit), kindly add following tags
> | Reported-by: kernel test robot <lkp@intel.com>
> | Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
> | Closes: https://lore.kernel.org/r/202502121806.CS6r741y-lkp@intel.com/
>
> smatch warnings:
> kernel/locking/ww_mutex_kunit.c:238 test_abba_gen_params() warn: shift has higher precedence than mask
> kernel/locking/ww_mutex_kunit.c:249 test_abba() warn: shift has higher precedence than mask
>
> vim +238 kernel/locking/ww_mutex_kunit.c
>
> 70207686e492fb kernel/locking/test-ww_mutex.c  Chris Wilson      2016-12-01  231
> daf92a37bd1117 kernel/locking/ww_mutex_kunit.c Tamir Duberstein  2025-02-10  232  static const void *test_abba_gen_params(const void *prev, char *desc)
> daf92a37bd1117 kernel/locking/ww_mutex_kunit.c Tamir Duberstein  2025-02-10  233  {
> daf92a37bd1117 kernel/locking/ww_mutex_kunit.c Tamir Duberstein  2025-02-10  234        static unsigned int storage;
> daf92a37bd1117 kernel/locking/ww_mutex_kunit.c Tamir Duberstein  2025-02-10  235        const unsigned int *next = gen_range(&storage, 0b00, 0b11, prev);
> daf92a37bd1117 kernel/locking/ww_mutex_kunit.c Tamir Duberstein  2025-02-10  236
> daf92a37bd1117 kernel/locking/ww_mutex_kunit.c Tamir Duberstein  2025-02-10  237        if (next != NULL) {
> daf92a37bd1117 kernel/locking/ww_mutex_kunit.c Tamir Duberstein  2025-02-10 @238                const bool trylock = *next & 0b01 >> 0;
> daf92a37bd1117 kernel/locking/ww_mutex_kunit.c Tamir Duberstein  2025-02-10  239                const bool resolve = *next & 0b10 >> 1;
>
> The shifts here are weird...  A zero shift is strange but even the 1 shift
> is odd.  The current code is equivalent to:
>
>         const bool resolve = *next & (0b10 >> 1);
>
> But changing it to:
>
>         const bool resolve = (*next & 0b10) >> 1;
>
> Doesn't make sense either...  Probably that makes less sense actually.
> What are you trying to communicate with this code?

Yeah, the bit shifting here is not necessary. I'll replace this with a
proper bitfield.

> daf92a37bd1117 kernel/locking/ww_mutex_kunit.c Tamir Duberstein  2025-02-10  240
> daf92a37bd1117 kernel/locking/ww_mutex_kunit.c Tamir Duberstein  2025-02-10  241                snprintf(desc, KUNIT_PARAM_DESC_SIZE, "trylock=%d,resolve=%d", trylock, resolve);
> daf92a37bd1117 kernel/locking/ww_mutex_kunit.c Tamir Duberstein  2025-02-10  242        }
> daf92a37bd1117 kernel/locking/ww_mutex_kunit.c Tamir Duberstein  2025-02-10  243        return next;
> daf92a37bd1117 kernel/locking/ww_mutex_kunit.c Tamir Duberstein  2025-02-10  244  }
> daf92a37bd1117 kernel/locking/ww_mutex_kunit.c Tamir Duberstein  2025-02-10  245
> daf92a37bd1117 kernel/locking/ww_mutex_kunit.c Tamir Duberstein  2025-02-10  246  static void test_abba(struct kunit *test)
> 70207686e492fb kernel/locking/test-ww_mutex.c  Chris Wilson      2016-12-01  247  {
> daf92a37bd1117 kernel/locking/ww_mutex_kunit.c Tamir Duberstein  2025-02-10  248        const unsigned int *param = test->param_value;
> daf92a37bd1117 kernel/locking/ww_mutex_kunit.c Tamir Duberstein  2025-02-10 @249        const bool trylock = *param & 0b01 >> 0;
> daf92a37bd1117 kernel/locking/ww_mutex_kunit.c Tamir Duberstein  2025-02-10  250        const bool resolve = *param & 0b10 >> 1;
>
> Same.
>
> 70207686e492fb kernel/locking/test-ww_mutex.c  Chris Wilson      2016-12-01  251        struct test_abba abba;
> 70207686e492fb kernel/locking/test-ww_mutex.c  Chris Wilson      2016-12-01  252        struct ww_acquire_ctx ctx;
> daf92a37bd1117 kernel/locking/ww_mutex_kunit.c Tamir Duberstein  2025-02-10  253        int err;
> 70207686e492fb kernel/locking/test-ww_mutex.c  Chris Wilson      2016-12-01  254
> 70207686e492fb kernel/locking/test-ww_mutex.c  Chris Wilson      2016-12-01  255        ww_mutex_init(&abba.a_mutex, &ww_class);
> 70207686e492fb kernel/locking/test-ww_mutex.c  Chris Wilson      2016-12-01  256        ww_mutex_init(&abba.b_mutex, &ww_class);
> 70207686e492fb kernel/locking/test-ww_mutex.c  Chris Wilson      2016-12-01  257        INIT_WORK_ONSTACK(&abba.work, test_abba_work);
>
> --
> 0-DAY CI Kernel Test Service
> https://github.com/intel/lkp-tests/wiki
>

As an aside, how can I compile with the warning settings used by
kernel test robot?

Thanks.
Tamir