| Message ID | 20220525154442.1438081-1-dlatypov@google.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | kunit: more assertion reworking | expand |
Hi Daniel, On Wed, May 25, 2022 at 5:44 PM Daniel Latypov <dlatypov@google.com> wrote: > > Note: this does change the function signature of > kunit_do_failed_assertion, so we'd need to update the rust wrapper in https://github.com/Rust-for-Linux/linux/blob/rust/rust/kernel/kunit.rs > But otherwise, I don't think this series changes anything on the > rust-side. Thanks for the heads up! I can take care of that. Cheers, Miguel
On Wed, May 25, 2022 at 8:56 AM Miguel Ojeda <miguel.ojeda.sandonis@gmail.com> wrote: > > Hi Daniel, > > On Wed, May 25, 2022 at 5:44 PM Daniel Latypov <dlatypov@google.com> wrote: > > > > Note: this does change the function signature of > > kunit_do_failed_assertion, so we'd need to update the rust wrapper in https://github.com/Rust-for-Linux/linux/blob/rust/rust/kernel/kunit.rs > > But otherwise, I don't think this series changes anything on the > > rust-side. > > Thanks for the heads up! I can take care of that. FYI This probably won't need to happen soon. I'd mentioned the idea of patch #1, which changes kunit_do_failed_assertion(), and Brendan didn't seem keen on it before. But the diff should be simple, e.g. just something like this: @@ -38,9 +38,7 @@ }); static CONDITION: &'static $crate::str::CStr = $crate::c_str!(stringify!($cond)); static ASSERTION: UnaryAssert = UnaryAssert($crate::bindings::kunit_unary_assert { - assert: $crate::bindings::kunit_assert { - format: Some($crate::bindings::kunit_unary_assert_format), - }, + assert: $crate::bindings::kunit_assert {}, condition: CONDITION.as_char_ptr(), expected_true: true, }); @@ -67,6 +65,7 @@ core::ptr::addr_of!(LOCATION.0), $crate::bindings::kunit_assert_type_KUNIT_ASSERTION, core::ptr::addr_of!(ASSERTION.0.assert), + Some($crate::bindings::kunit_unary_assert_format), core::ptr::null(), ); } The only tricky bit will be coordinating the changes :) Daniel
This is a follow up to these three series: https://lore.kernel.org/all/20220113165931.451305-1-dlatypov@google.com/ https://lore.kernel.org/all/20220118223506.1701553-1-dlatypov@google.com/ https://lore.kernel.org/all/20220125210011.3817742-1-dlatypov@google.com/ The two goals of those series were a) reduce the size of struct kunit_assert and friends. (struct kunit_assert went from 48 => 8 bytes on UML.) b) simplify the internal code, mostly by deleting macros This series goes further a) sizeof(struct kunit_assert) = 0 now b) e.g. we delete another class of macros (KUNIT_INIT_*_ASSERT_STRUCT) Note: this does change the function signature of kunit_do_failed_assertion, so we'd need to update the rust wrapper in https://github.com/Rust-for-Linux/linux/blob/rust/rust/kernel/kunit.rs But otherwise, I don't think this series changes anything on the rust-side. Daniel Latypov (4): rfc: kunit: remove format func from struct kunit_assert, get it to 0 bytes rfc: kunit: rename base KUNIT_ASSERTION macro to _KUNIT_FAILED rfc: kunit: eliminate KUNIT_INIT_*_ASSERT_STRUCT macros kunit: delcare kunit_assert structs as const include/kunit/assert.h | 74 ++----------------------- include/kunit/test.h | 123 ++++++++++++++++++++++------------------- lib/kunit/test.c | 7 ++- 3 files changed, 76 insertions(+), 128 deletions(-) base-commit: e7eaffce47b7db72b077630dbe836f0c4132496d