[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CABVgOSmbbAyckSvKREmUDBrZJtErQpxaNjXH0vaH1oZjkVt3JA@mail.gmail.com>
Date: Sat, 9 Dec 2023 15:57:43 +0800
From: David Gow <davidgow@...gle.com>
To: Rae Moar <rmoar@...gle.com>
Cc: shuah@...nel.org, dlatypov@...gle.com, brendan.higgins@...ux.dev,
sadiyakazi@...gle.com, keescook@...omium.org, arnd@...db.de,
linux-kselftest@...r.kernel.org, linux-arch@...r.kernel.org,
kunit-dev@...glegroups.com, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 4/6] kunit: add is_init test attribute
On Tue, 5 Dec 2023 at 06:19, Rae Moar <rmoar@...gle.com> wrote:
>
> Add is_init test attribute of type bool. Add to_string, get, and filter
> methods to lib/kunit/attributes.c.
>
> Mark each of the tests in the init section with the is_init=true attribute.
>
> Add is_init to the attributes documentation.
>
> Signed-off-by: Rae Moar <rmoar@...gle.com>
> ---
Would it be possible to not have this in kunit_attributes? I know it's
required for the run-after-boot stuff later, but I'd love this to be
(a) just generated at runtime, or (b) stored only at a suite or
suite-set level. It seems like a bit of a waste to store this
per-test-case, and to have it potentially accessible or overwritable
by users.
Otherwise, this looks good (and I appreciate the automatic setting of
this when merging the suite sets.
Maybe if we always kept the init suites in a separate set, we could
just use pointer comparisons to generate this; otherwise let's make
this a suite-level-only attribute (inherited by tests).
-- David
Download attachment "smime.p7s" of type "application/pkcs7-signature" (4003 bytes)
Powered by blists - more mailing lists