Re: [PATCH 0/3] bitmap: Convert test_bitmap to kunit test

[David Gow]

On Tue, 30 Jul 2024 at 23:49, Shuah Khan <skhan@xxxxxxxxxxxxxxxxxxx> wrote:
>
> On 7/29/24 02:29, Muhammad Usama Anjum wrote:
> > On 7/27/24 12:26 AM, Shuah Khan wrote:
> >> On 7/26/24 05:06, Muhammad Usama Anjum wrote:
> >>> In this series, test_bitmap is being converted to kunit test. Multiple
> >>> patches will make the review process smooth.
> >>>
> >>> - Patch-1: Convert the tests in lib/test_bitmap.c to kunit
> >>> - Patch-2: Rename the lib/test_bitmap.c to lib/bitmap_kunit.c and other
> >>>              configuration options
> >>> - Patch-3: Remove the bitmap.sh selftest
> >>>
> >>> Muhammad Usama Anjum (3):
> >>>     bitmap: convert test_bitmap to KUnit test
> >>>     bitmap: Rename module
> >>>     selftests: lib: remove test_bitmap
> >>>
> >>>    MAINTAINERS                           |   2 +-
> >>>    lib/Kconfig.debug                     |  15 +-
> >>>    lib/Makefile                          |   2 +-
> >>>    lib/{test_bitmap.c => bitmap_kunit.c} | 624 ++++++++++++--------------
> >>>    tools/testing/selftests/lib/Makefile  |   2 +-
> >>>    tools/testing/selftests/lib/bitmap.sh |   3 -
> >>>    tools/testing/selftests/lib/config    |   1 -
> >>>    7 files changed, 295 insertions(+), 354 deletions(-)
> >>>    rename lib/{test_bitmap.c => bitmap_kunit.c} (70%)
> >>>    delete mode 100755 tools/testing/selftests/lib/bitmap.sh
> >>>
> >>
> >> Can you tell me how this conversion helps?
> >>
> >> It is removing the ability to run bitmap tests during boot.
> >> It doesn't make sense to blindly convert all test under lib
> >> to kunit - Nack on this change or any change that takes away
> >> the ability to run tests and makes them dependent on kunit.
> > Let's discuss this on discussion thread [1].
>
> So - it doesn't mean that it is a done deal. Each patch will be
> reviewed on individual basis. This test in particular clearly its
> use-case right in the config which was deleted without understanding
> it.
>
> -config TEST_BITMAP
> -       tristate "Test bitmap_*() family of functions at runtime"
> -       help
> -         Enable this option to test the bitmap functions at boot.
>
>
> This line above is the important piece of information which tells
> you how the test is intended to be used.
>
> 1. You can enable this option and boot the kernel to check for regressions.

Converting the test to KUnit _does not break this_. You can still
enable this option and boot the kernel to test for regressions if it's
a KUnit test.

> 2. You can load the module on a running kernel to check for health.

This is where the disagreement lies in my mind. While you can do this
with KUnit, the kernel does have to have been built with CONFIG_KUNIT
enabled (either built-in or as a module).

(Indeed, KUnit adds the ability to trigger a test at runtime even if
it's built-in, via debugfs, so it's possible to do so even if
CONFIG_MODULES is disabled. That's a pretty niche use-case, I admit,
but I'd guess that the bitmap functionailty breaking in a way which
disappears when CONFIG_KUNIT is enabled is as well.)

In general, though, my interpretation of "Test ... at runtime" in
config option descriptions has not precluded introducing the
CONFIG_KUNIT dependency. We already have several KUnit tests with that
in their description, several of which were ported to KUnit later.

>
> Converting it to kunit drops support for these two use-cases which
> are user-space regressions. You don't want to do that.
>
> -
> -         If unsure, say N.
> -
>   config TEST_UUID
>         tristate "Test functions located in the uuid module at runtime"
>
> @@ -2813,6 +2806,14 @@ config USERCOPY_KUNIT_TEST
>           on the copy_to/from_user infrastructure, making sure basic
>           user/kernel boundary testing is working.
>
> +config BITMAP_KUNIT_TEST
> +       tristate "KUnit Test for bitmap_*() family of functions"
> +       depends on KUNIT
> +       default KUNIT_ALL_TESTS
> +       help
> +         This builds the "bitmap_kunit" module that runs tests for
> +         bitmaps int the kernel making sure that there isn't any bug.
>
> And this isn't complete even. I don'ty see tristate in here.

(This looks like it's tristate to me?)

>
> I responded to the thread that started this flurry of conversion activity
> stating that converting tests without thinking through the use-cases
> isn't what we want to do.
>
> Reports aren't everything. The primary reason we have these tests is for
> developers to be able to test. Reports can be improved and shouldn't
> come at the expense of coverage and testing. Any patch that does that
> will be NACKed.
>
> thanks,
> -- Shuah

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature