Re: [PATCH v2 1/3] rust: kunit: add #[should_panic] support

[Nicolás Antinori]

On Mon Nov 2, 2026 at 12:51 PM -03, Miguel Ojeda wrote:
> On Mon, Nov 2, 2026 at 3:58 PM Nicolás Antinori
> <nico.antinori.7@xxxxxxxxx> wrote:
>>
>> To solve this I created a new config "RUST_KUNIT_SHOULD_PANIC_ATTRIBUTE"
>> that depends on !UML && !PANIC_ON_OOPS && BUG. If the test is annotated
>> with `#[should_panic]` and CONFIG_RUST_KUNIT_SHOULD_PANIC_ATTRIBUTE=n,
>> the test won't run and won't show up in the summary.
>
> I was wondering if we needed the flexibility, i.e. if we could just
> simply disable them when the conditions are not satisfied, but I guess
> given `CONFIG_KUNIT_FAULT_TEST`'s existing description, it may make
> sense to have a new Kconfig symbol 

I first thought of disabling `#[should_panic]` tests by directly
checking the conditions in the macro code. I ended up creating a new
config because I thought it would be cleaner codewise, if a new
constraint arises in the future, we just have to add it to the 
`depends on` clause instead of modifying the macro code.

> -- do the ones added here look
> similar in the kernel log to those fault ones?

Yes! they are pretty similar, for `KUNIT_FAULT_TEST`:

KTAP version 1
1..1
    KTAP version 1
    # Subtest: kunit_fault
    # module: kunit_test
    1..1
BUG: kernel NULL pointer dereference, address: 0000000000000000
#PF: supervisor write access in kernel mode
#PF: error_code(0x0002) - not-present page
PGD 0 P4D 0
Oops: Oops: 0002 [#1] SMP NOPTI
CPU: 0 UID: 0 PID: 30 Comm: kunit_try_catch Tainted: G                 N   7.3.0-18137-ga2943ba718b4-dirty #357 PREEMPT(lazy)
Tainted: [N]=TEST
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 0.1 11/11/2019

... <Registers and stack trace> ...

note: kunit_try_catch[30] exited with irqs disabled
    # kunit_test_fault_null_dereference: try faulted: last line seen lib/kunit/kunit-test.c:100
    ok 1 kunit_test_fault_null_dereference
ok 1 kunit_fault

For `RUST_KUNIT_TESTS` (left only one that panics to shrink the log, but
they are all pretty much the same):

KTAP version 1
1..1
    KTAP version 1
    # Subtest: rust_kernel_kunit
    # speed: normal
    1..1
rust_kernel: panicked at ../rust/kernel/kunit.rs:399:9:
This test should panic and pass
------------[ cut here ]------------
kernel BUG at rust/helpers/bug.c:7!
Oops: invalid opcode: 0000 [#1] SMP NOPTI
CPU: 0 UID: 0 PID: 29 Comm: kunit_try_catch Tainted: G                 N   7.3.0-18137-ga2943ba718b4-dirty #357 PREEMPT(lazy)
Tainted: [N]=TEST
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 0.1 11/11/2019

... <Registers and stack trace> ...

    # rust_test_kunit_panic_in_kunit_test.speed: normal
    ok 1 rust_test_kunit_panic_in_kunit_test
ok 1 rust_kernel_kunit

The only one that is different in `RUST_KUNIT_TESTS` is:

#[test]
#[should_panic]
fn rust_test_assert_false_success_with_should_panic() {
    assert!(false)
}

That logs:

KTAP version 1
1..1
    KTAP version 1
    # Subtest: rust_kernel_kunit
    # speed: normal
    1..1
    # rust_test_assert_false_success_with_should_panic.speed: normal
    ok 1 rust_test_assert_false_success_with_should_panic
ok 1 rust_kernel_kunit

Because a BUG() is never called, the failed assertion is catched in the
`kunit_assert!` macro and, if the test should panic, the KUnit thread is
aborted without throwing a failed asssertion.

>
> (And I was wondering because, in general, adding Kconfig symbols has a
> price: users may get confused etc.).

On top of the reasons I gave above, I tried to follow the same as
`CONFIG_RUST_BUG_KUNIT_TEST` or `CONFIG_KUNIT_FAULT_TEST` are doing, but
in this case in particular, the config is not for enabling a test suite
but for enabling a feature that can affect multiple suites (in this
patch series the ones behind `CONFIG_RUST_KUNIT_TESTS` and
`CONFIG_RUST_BITMAP_KUNIT_TESTS`).

If this is prone to causing confusion, I think I am going to check the
contraints direclty in the code. 

Thank you for the feedback!

>
> Cheers,
> Miguel