On 10/18/24 13:52, Lorenzo Stoakes wrote:
On Fri, Oct 18, 2024 at 12:32:37PM -0700, Jeff Xu wrote:
when they encouter a failure, the pattern I sketched in my earlier message, or switch to kselftest_harness.h (like I say I don't know if the fork()ing is an issue for these tests). If I had to have a macro it'd probably be something like mseal_assert().
I can go with mseal_assert, the original macro is used by mseal_test itself, and only intended as such.
If changing name to mseal_assert() is acceptable, this seems to be a minimum change and I'm happy with that.
No.
Jeff,
Please pay attention to the feedback you have been receiving so far from Mark and others about using the existing kselftest framework for reporting results and don't reinvent wheel.
We have two frameworks to choose from - they both have been in use for quiet sometime by tests. If there is a need to add new functions and fix the existing ones that should happen in kselftest_harness.h or kselftest.h. We keep fixing problem and enhancing them as needed.
With my kselfest and framework maintainer hat on, I don't want to see yet another framework emerging which is buried in tests.
thanks, -- Shuah