Re: [PATCH v3 0/3] Introduce KUNIT_EXPECT_MEMEQ and KUNIT_EXPECT_MEMNEQ macros

[David Gow <davidgow@google.com>]

[-- Attachment #1: Type: text/plain, Size: 6753 bytes --]

On Fri, Aug 5, 2022 at 8:18 PM Maíra Canal <mairacanal@riseup.net> wrote:
>
> On 8/5/22 01:44, David Gow wrote:
> > On Thu, Aug 4, 2022 at 5:59 AM Maíra Canal <mairacanal@riseup.net> wrote:
> >>
> >> Currently, in order to compare memory blocks in KUnit, the KUNIT_EXPECT_EQ or
> >> KUNIT_EXPECT_FALSE macros are used in conjunction with the memcmp function,
> >> such as:
> >>   KUNIT_EXPECT_EQ(test, memcmp(foo, bar, size), 0);
> >>
> >> Although this usage produces correct results for the test cases, if the
> >> expectation fails the error message is not very helpful, indicating only the
> >> return of the memcmp function.
> >>
> >> Therefore, create a new set of macros KUNIT_EXPECT_MEMEQ and
> >> KUNIT_EXPECT_MEMNEQ that compare memory blocks until a determined size. In
> >> case of expectation failure, those macros print the hex dump of the memory
> >> blocks, making it easier to debug test failures for memory blocks.
> >>
> >> Other than the style changes, this v3 brings alignment to the bytes, making
> >> it easier to identify the faulty bytes. So, on the previous version, the
> >> output from a failure would be:
> >> [14:27:42] # xrgb8888_to_rgb565_test: EXPECTATION FAILED at drivers/gpu/drm/tests/drm_format_helper_test.c:248
> >> [14:27:42] Expected dst == result->expected, but
> >> [14:27:42] dst ==
> >> [14:27:42] 33 0a <60> 12 00 a8 00 00 <00> 00 8e 6b <33> 0a 60 12
> >> [14:27:42] 00 00 <00> 00 00 a8 <8e> 6b 33 0a 00 00 <00> 00
> >> [14:27:42] result->expected ==
> >> [14:27:42] 33 0a <61> 12 00 a8 00 00 <01> 00 8e 6b <31> 0a 60 12
> >> [14:27:42] 00 00 <01> 00 00 a8 <81> 6b 33 0a 00 00 <01> 00
> >>
> >> Now, with the alignment, the output is:
> >> [14:27:42] # xrgb8888_to_rgb565_test: EXPECTATION FAILED at drivers/gpu/drm/tests/drm_format_helper_test.c:248
> >> [14:27:42] Expected dst == result->expected, but
> >> [14:27:42] dst ==
> >> [14:27:42] 33  0a <60> 12  00  a8  00  00 <00> 00  8e  6b <33> 0a  60  12
> >> [14:27:42] 00  00 <00> 00  00  a8 <8e> 6b  33  0a  00  00 <00> 00
> >> [14:27:42] result->expected ==
> >> [14:27:42] 33  0a <61> 12  00  a8  00  00 <01> 00  8e  6b <31> 0a  60  12
> >> [14:27:42] 00  00 <01> 00  00  a8 <81> 6b  33  0a  00  00 <01> 00
> >>
> >> Moreover, on the raw output, there were some indentation problems. Those
> >> problems were solved with the use of KUNIT_SUBSUBTEST_INDENT.
> >>
> >> The first patch of the series introduces the KUNIT_EXPECT_MEMEQ and
> >> KUNIT_EXPECT_MEMNEQ. The second patch adds an example of memory block
> >> expectations on the kunit-example-test.c. And the last patch replaces the
> >> KUNIT_EXPECT_EQ for KUNIT_EXPECT_MEMEQ on the existing occurrences.
> >>
> >> Best Regards,
> >> - Maíra Canal
> >>
> >> v1 -> v2: https://lore.kernel.org/linux-kselftest/2a0dcd75-5461-5266-2749-808f638f4c50@riseup.net/T/#m402cc72eb01fb3b88d6706cf7d1705fdd51e5da2
> >>
> >> - Change "determinated" to "specified" (Daniel Latypov).
> >> - Change the macro KUNIT_EXPECT_ARREQ to KUNIT_EXPECT_MEMEQ, in order to make
> >> it easier for users to infer the right size unit (Daniel Latypov).
> >> - Mark the different bytes on the failure message with a <> (Daniel Latypov).
> >> - Replace a constant number of array elements for ARRAY_SIZE() (André Almeida).
> >> - Rename "array" and "expected" variables to "array1" and "array2" (Daniel Latypov).
> >>
> >> v2 -> v3: https://lore.kernel.org/linux-kselftest/20220802212621.420840-1-mairacanal@riseup.net/T/#t
> >>
> >> - Make the bytes aligned at output.
> >> - Add KUNIT_SUBSUBTEST_INDENT to the output for the indentation (Daniel Latypov).
> >> - Line up the trailing \ at macros using tabs (Daniel Latypov).
> >> - Line up the params to the functions (Daniel Latypov).
> >> - Change "Increament" to "Augment" (Daniel Latypov).
> >> - Use sizeof() for array sizes (Daniel Latypov).
> >>
> >> Maíra Canal (3):
> >>   kunit: Introduce KUNIT_EXPECT_MEMEQ and KUNIT_EXPECT_MEMNEQ macros
> >>   kunit: Add KUnit memory block assertions to the example_all_expect_macros_test
> >>   kunit: Use KUNIT_EXPECT_MEMEQ macro
> >>
> >>  .../gpu/drm/tests/drm_format_helper_test.c    |  6 +-
> >>  include/kunit/assert.h                        | 34 +++++++++
> >>  include/kunit/test.h                          | 76 +++++++++++++++++++
> >>  lib/kunit/assert.c                            | 56 ++++++++++++++
> >>  lib/kunit/kunit-example-test.c                |  7 ++
> >>  net/core/dev_addr_lists_test.c                |  4 +-
> >>  6 files changed, 178 insertions(+), 5 deletions(-)
> >>
> >> --
> >> 2.37.1
> >>
> >> --
> >> You received this message because you are subscribed to the Google Groups "KUnit Development" group.
> >> To unsubscribe from this group and stop receiving emails from it, send an email to kunit-dev+unsubscribe@googlegroups.com.
> >> To view this discussion on the web visit https://groups.google.com/d/msgid/kunit-dev/20220803215855.258704-1-mairacanal%40riseup.net.
> >
> > These patches look pretty good to me overall, but I was unable to
> > apply v3 to test -- it looks like the mail client has wrapped some
> > lines or something...
> >
> > davidgow@slicestar:~/linux-kselftest$ git am
> > ./v3_20220803_mairacanal_introduce_kunit_expect_memeq_and_kunit_expect_memneq_macros.mbx
> > Applying: kunit: Introduce KUNIT_EXPECT_MEMEQ and KUNIT_EXPECT_MEMNEQ macros
> > error: corrupt patch at line 24
> > Patch failed at 0001 kunit: Introduce KUNIT_EXPECT_MEMEQ and
> > KUNIT_EXPECT_MEMNEQ macros
> >
> > Checkpatch also picks up an issue:
> > ERROR: patch seems to be corrupt (line wrapped?)
> > #62: FILE: include/kunit/assert.h:255:
> >                                    const struct va_format *message,
> >
> > v2 applied clearnly, so it seems to be specific to v3.
>
> I'll check this issue and submit a v4. Thank you!
>

Thanks!

> >
> > In general, I like the patches, though. While I think there are a few
> > places it'd be slightly suboptimale if it's being used to compare more
> > structured data, such as the prospect of comparing padding between
> > elements, as well as the output formatting not being ideal. It's
> > perfect for the cases where memcmp() otherwise would be used, though.
>
> Do you any take on how to make the output formatting more ideal?
>

I don't actually think we need to change any of the formatting in this
patch, I'm just noting that usinng MEMEQ()/MEMNEQ() might not be the
best choice for comparing, e.g., structs (and that comparing their
members individually might be better there).
_Maybe_ that's something that could be mentioned in the documentation,
but I wouldn't change the code at all.

Cheers,
-- David

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4003 bytes --]