linux-security-module.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: David Gow <davidgow@google.com>
To: "Mickaël Salaün" <mic@digikod.net>
Cc: Brendan Higgins <brendanhiggins@google.com>,
	Kees Cook <keescook@chromium.org>,  Rae Moar <rmoar@google.com>,
	Shuah Khan <skhan@linuxfoundation.org>,
	 Alan Maguire <alan.maguire@oracle.com>,
	Borislav Petkov <bp@alien8.de>,
	 Dave Hansen <dave.hansen@linux.intel.com>,
	"H . Peter Anvin" <hpa@zytor.com>,
	 Ingo Molnar <mingo@redhat.com>,
	James Morris <jamorris@linux.microsoft.com>,
	 Luis Chamberlain <mcgrof@kernel.org>,
	 "Madhavan T . Venkataraman" <madvenka@linux.microsoft.com>,
	Marco Pagani <marpagan@redhat.com>,
	 Paolo Bonzini <pbonzini@redhat.com>,
	Sean Christopherson <seanjc@google.com>,
	Stephen Boyd <sboyd@kernel.org>,
	 Thara Gopinath <tgopinath@microsoft.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	 Vitaly Kuznetsov <vkuznets@redhat.com>,
	Wanpeng Li <wanpengli@tencent.com>,
	 Zahra Tarkhani <ztarkhani@microsoft.com>,
	kvm@vger.kernel.org,  linux-hardening@vger.kernel.org,
	linux-hyperv@vger.kernel.org,  linux-kernel@vger.kernel.org,
	linux-security-module@vger.kernel.org,
	 linux-um@lists.infradead.org, x86@kernel.org
Subject: Re: [PATCH v2 6/7] kunit: Print last test location on fault
Date: Tue, 12 Mar 2024 12:54:48 +0800	[thread overview]
Message-ID: <CABVgOSk_vea-LrPwJet6hQ4D3PBQOLVg32nZ_gE4c9kgGDEEnQ@mail.gmail.com> (raw)
In-Reply-To: <20240301194037.532117-7-mic@digikod.net>

[-- Attachment #1: Type: text/plain, Size: 9294 bytes --]

On Sat, 2 Mar 2024 at 03:40, Mickaël Salaün <mic@digikod.net> wrote:
>
> This helps identify the location of test faults.
>
> Cc: Brendan Higgins <brendanhiggins@google.com>
> Cc: David Gow <davidgow@google.com>
> Cc: Rae Moar <rmoar@google.com>
> Cc: Shuah Khan <skhan@linuxfoundation.org>
> Reviewed-by: Kees Cook <keescook@chromium.org>
> Signed-off-by: Mickaël Salaün <mic@digikod.net>
> Link: https://lore.kernel.org/r/20240301194037.532117-7-mic@digikod.net
> ---

I like the idea of this, but am a little bit worried about how
confusing it might be, given that the location only updates on those
particular macros.

Maybe the answer is to make the __KUNIT_SAVE_LOC() macro, or something
equivalent, a supported API.

One possibility would be to have a KUNIT_MARKER() macro. If we really
wanted to, we could expand it to take a string so we can have a more
user-friendly KUNIT_MARKER(test, "parsing packet") description of
where things went wrong. Another could be to extend this to use the
code tagging framework[1], if that lands.

That being said, I think this is still an improvement without any of
those features. I've left a few comments below. Let me know what you
think.

Cheers,
-- David

[1]: https://lwn.net/Articles/906660/
>
> Changes since v1:
> * Added Kees's Reviewed-by.
> ---
>  include/kunit/test.h  | 24 +++++++++++++++++++++---
>  lib/kunit/try-catch.c | 10 +++++++---
>  2 files changed, 28 insertions(+), 6 deletions(-)
>
> diff --git a/include/kunit/test.h b/include/kunit/test.h
> index fcb4a4940ace..f3aa66eb0087 100644
> --- a/include/kunit/test.h
> +++ b/include/kunit/test.h
> @@ -301,6 +301,8 @@ struct kunit {
>         struct list_head resources; /* Protected by lock. */
>
>         char status_comment[KUNIT_STATUS_COMMENT_SIZE];
> +       /* Saves the last seen test. Useful to help with faults. */
> +       struct kunit_loc last_seen;
>  };
>
>  static inline void kunit_set_failure(struct kunit *test)
> @@ -567,6 +569,15 @@ void __printf(2, 3) kunit_log_append(struct string_stream *log, const char *fmt,
>  #define kunit_err(test, fmt, ...) \
>         kunit_printk(KERN_ERR, test, fmt, ##__VA_ARGS__)
>
> +/*
> + * Must be called at the beginning of each KUNIT_*_ASSERTION().
> + * Cf. KUNIT_CURRENT_LOC.
> + */
> +#define _KUNIT_SAVE_LOC(test) do {                                            \
> +       WRITE_ONCE(test->last_seen.file, __FILE__);                            \
> +       WRITE_ONCE(test->last_seen.line, __LINE__);                            \
> +} while (0)

Can we get rid of the leading '_', make this public, and document it?
If we want to rename it to KUNIT_MARKER() or similar, that might work
better, too.

> +
>  /**
>   * KUNIT_SUCCEED() - A no-op expectation. Only exists for code clarity.
>   * @test: The test context object.
> @@ -575,7 +586,7 @@ void __printf(2, 3) kunit_log_append(struct string_stream *log, const char *fmt,
>   * words, it does nothing and only exists for code clarity. See
>   * KUNIT_EXPECT_TRUE() for more information.
>   */
> -#define KUNIT_SUCCEED(test) do {} while (0)
> +#define KUNIT_SUCCEED(test) _KUNIT_SAVE_LOC(test)
>
>  void __noreturn __kunit_abort(struct kunit *test);
>
> @@ -601,14 +612,16 @@ void __kunit_do_failed_assertion(struct kunit *test,
>  } while (0)
>
>
> -#define KUNIT_FAIL_ASSERTION(test, assert_type, fmt, ...)                     \
> +#define KUNIT_FAIL_ASSERTION(test, assert_type, fmt, ...) do {                \
> +       _KUNIT_SAVE_LOC(test);                                                 \
>         _KUNIT_FAILED(test,                                                    \
>                       assert_type,                                             \
>                       kunit_fail_assert,                                       \
>                       kunit_fail_assert_format,                                \
>                       {},                                                      \
>                       fmt,                                                     \
> -                     ##__VA_ARGS__)
> +                     ##__VA_ARGS__);                                          \
> +} while (0)
>
>  /**
>   * KUNIT_FAIL() - Always causes a test to fail when evaluated.
> @@ -637,6 +650,7 @@ void __kunit_do_failed_assertion(struct kunit *test,
>                               fmt,                                             \
>                               ...)                                             \
>  do {                                                                          \
> +       _KUNIT_SAVE_LOC(test);                                                 \
>         if (likely(!!(condition_) == !!expected_true_))                        \
>                 break;                                                         \
>                                                                                \
> @@ -698,6 +712,7 @@ do {                                                                               \
>                 .right_text = #right,                                          \
>         };                                                                     \
>                                                                                \
> +       _KUNIT_SAVE_LOC(test);                                                 \
>         if (likely(__left op __right))                                         \
>                 break;                                                         \
>                                                                                \
> @@ -758,6 +773,7 @@ do {                                                                               \
>                 .right_text = #right,                                          \
>         };                                                                     \
>                                                                                \
> +       _KUNIT_SAVE_LOC(test);                                                 \
>         if (likely((__left) && (__right) && (strcmp(__left, __right) op 0)))   \
>                 break;                                                         \
>                                                                                \
> @@ -791,6 +807,7 @@ do {                                                                               \
>                 .right_text = #right,                                          \
>         };                                                                     \
>                                                                                \
> +       _KUNIT_SAVE_LOC(test);                                                 \
>         if (likely(__left && __right))                                         \
>                 if (likely(memcmp(__left, __right, __size) op 0))              \
>                         break;                                                 \
> @@ -815,6 +832,7 @@ do {                                                                               \
>  do {                                                                          \
>         const typeof(ptr) __ptr = (ptr);                                       \
>                                                                                \
> +       _KUNIT_SAVE_LOC(test);                                                 \
>         if (!IS_ERR_OR_NULL(__ptr))                                            \
>                 break;                                                         \
>                                                                                \
> diff --git a/lib/kunit/try-catch.c b/lib/kunit/try-catch.c
> index c6ee4db0b3bd..2ec21c6918f3 100644
> --- a/lib/kunit/try-catch.c
> +++ b/lib/kunit/try-catch.c
> @@ -91,9 +91,13 @@ void kunit_try_catch_run(struct kunit_try_catch *try_catch, void *context)
>
>         if (exit_code == -EFAULT)
>                 try_catch->try_result = 0;
> -       else if (exit_code == -EINTR)
> -               kunit_err(test, "try faulted\n");
> -       else if (exit_code == -ETIMEDOUT)
> +       else if (exit_code == -EINTR) {
> +               if (test->last_seen.file)
> +                       kunit_err(test, "try faulted after %s:%d\n",
> +                                 test->last_seen.file, test->last_seen.line);

It's possibly a bit confusing to just say "after file:line",
particularly if we then loop or call a function "higher up" in the
file. Maybe something like "try faulted: last line seen %s:%d" is
clearer.

> +               else
> +                       kunit_err(test, "try faulted before the first test\n");

I don't like using "test" here, as it introduces ambiguity between
"kunit tests" and "assertions/expectations" if we call them both
tests. Maybe just "try faulted" here, or "try faulted (no markers
seen)" or similar?


> +       } else if (exit_code == -ETIMEDOUT)
>                 kunit_err(test, "try timed out\n");
>         else if (exit_code)
>                 kunit_err(test, "Unknown error: %d\n", exit_code);
> --
> 2.44.0
>

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4014 bytes --]

  reply	other threads:[~2024-03-12  4:55 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-03-01 19:40 [PATCH v2 0/7] Handle faults in KUnit tests Mickaël Salaün
2024-03-01 19:40 ` [PATCH v2 1/7] kunit: Handle thread creation error Mickaël Salaün
2024-03-05 20:57   ` Rae Moar
2024-03-12  4:53   ` David Gow
2024-03-01 19:40 ` [PATCH v2 2/7] kunit: Fix kthread reference Mickaël Salaün
2024-03-05 20:57   ` Rae Moar
2024-03-12  4:53   ` David Gow
2024-03-01 19:40 ` [PATCH v2 3/7] kunit: Fix timeout message Mickaël Salaün
2024-03-05 20:58   ` Rae Moar
2024-03-12  4:53   ` David Gow
2024-03-01 19:40 ` [PATCH v2 4/7] kunit: Handle test faults Mickaël Salaün
2024-03-11 21:21   ` Rae Moar
2024-03-12 12:15     ` Mickaël Salaün
2024-03-12  5:05   ` David Gow
2024-03-12 12:15     ` Mickaël Salaün
2024-03-01 19:40 ` [PATCH v2 5/7] kunit: Fix KUNIT_SUCCESS() calls in iov_iter tests Mickaël Salaün
2024-03-12  4:54   ` David Gow
2024-03-01 19:40 ` [PATCH v2 6/7] kunit: Print last test location on fault Mickaël Salaün
2024-03-12  4:54   ` David Gow [this message]
2024-03-12 12:15     ` Mickaël Salaün
2024-03-01 19:40 ` [PATCH v2 7/7] kunit: Add tests for fault Mickaël Salaün
2024-03-12  4:54   ` David Gow

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=CABVgOSk_vea-LrPwJet6hQ4D3PBQOLVg32nZ_gE4c9kgGDEEnQ@mail.gmail.com \
    --to=davidgow@google.com \
    --cc=alan.maguire@oracle.com \
    --cc=bp@alien8.de \
    --cc=brendanhiggins@google.com \
    --cc=dave.hansen@linux.intel.com \
    --cc=hpa@zytor.com \
    --cc=jamorris@linux.microsoft.com \
    --cc=keescook@chromium.org \
    --cc=kvm@vger.kernel.org \
    --cc=linux-hardening@vger.kernel.org \
    --cc=linux-hyperv@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-security-module@vger.kernel.org \
    --cc=linux-um@lists.infradead.org \
    --cc=madvenka@linux.microsoft.com \
    --cc=marpagan@redhat.com \
    --cc=mcgrof@kernel.org \
    --cc=mic@digikod.net \
    --cc=mingo@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=rmoar@google.com \
    --cc=sboyd@kernel.org \
    --cc=seanjc@google.com \
    --cc=skhan@linuxfoundation.org \
    --cc=tglx@linutronix.de \
    --cc=tgopinath@microsoft.com \
    --cc=vkuznets@redhat.com \
    --cc=wanpengli@tencent.com \
    --cc=x86@kernel.org \
    --cc=ztarkhani@microsoft.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).