The KUNIT_EXPECT_KASAN_FAIL() macro currently uses KUNIT_EXPECT_EQ() to
compare fail_data.report_expected and fail_data.report_found. This
always gave a somewhat useless error message on failure, but the
addition of extra compile-time checking with READ_ONCE() has caused it
to get much longer, and be truncated before anything useful is displayed.
Instead, just check fail_data.report_found by hand (we've just test
report_expected to 'true'), and print a better failure message with
KUNIT_FAIL()
Beforehand, a failure in:
KUNIT_EXPECT_KASAN_FAIL(test, ((volatile char *)area)[3100]);
would looked like:
[22:00:34] [FAILED] vmalloc_oob
[22:00:34] # vmalloc_oob: EXPECTATION FAILED at lib/test_kasan.c:991
[22:00:34] Expected ({ do { extern void __compiletime_assert_705(void) __attribute__((__error__("Unsupported access size for {READ,WRITE}_ONCE()."))); if (!((sizeof(fail_data.report_expected) == sizeof(char) || sizeof(fail_data.repp
[22:00:34] not ok 45 - vmalloc_oob
With this change, it instead looks like:
[22:04:04] [FAILED] vmalloc_oob
[22:04:04] # vmalloc_oob: EXPECTATION FAILED at lib/test_kasan.c:993
[22:04:04] KASAN failure expected in "((volatile char *)area)[3100]", but none occurred
[22:04:04] not ok 45 - vmalloc_oob
Signed-off-by: David Gow <[email protected]>
---
Stumbled across this because the vmalloc_oob test is failing (i.e.,
KASAN isn't picking up an error) under qemu on my system, and the
message above was horrifying. (I'll file a Bugzilla bug for the test
failure today.)
Cheers,
-- David
lib/test_kasan.c | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/lib/test_kasan.c b/lib/test_kasan.c
index cacbbbdef768..deda13c9d9ff 100644
--- a/lib/test_kasan.c
+++ b/lib/test_kasan.c
@@ -98,9 +98,11 @@ static void kasan_test_exit(struct kunit *test)
barrier(); \
expression; \
barrier(); \
- KUNIT_EXPECT_EQ(test, \
- READ_ONCE(fail_data.report_expected), \
- READ_ONCE(fail_data.report_found)); \
+ if (READ_ONCE(fail_data.report_found) == false) { \
+ KUNIT_FAIL(test, KUNIT_SUBTEST_INDENT "KASAN failure " \
+ "expected in \"" #expression \
+ "\", but none occurred"); \
+ } \
if (IS_ENABLED(CONFIG_KASAN_HW_TAGS)) { \
if (READ_ONCE(fail_data.report_found)) \
kasan_enable_tagging_sync(); \
--
2.32.0.rc1.229.g3e70b5a671-goog
On Fri, 4 Jun 2021 at 07:26, 'David Gow' via kasan-dev
<[email protected]> wrote:
> The KUNIT_EXPECT_KASAN_FAIL() macro currently uses KUNIT_EXPECT_EQ() to
> compare fail_data.report_expected and fail_data.report_found. This
> always gave a somewhat useless error message on failure, but the
> addition of extra compile-time checking with READ_ONCE() has caused it
> to get much longer, and be truncated before anything useful is displayed.
>
> Instead, just check fail_data.report_found by hand (we've just test
> report_expected to 'true'), and print a better failure message with
> KUNIT_FAIL()
>
> Beforehand, a failure in:
> KUNIT_EXPECT_KASAN_FAIL(test, ((volatile char *)area)[3100]);
> would looked like:
> [22:00:34] [FAILED] vmalloc_oob
> [22:00:34] # vmalloc_oob: EXPECTATION FAILED at lib/test_kasan.c:991
> [22:00:34] Expected ({ do { extern void __compiletime_assert_705(void) __attribute__((__error__("Unsupported access size for {READ,WRITE}_ONCE()."))); if (!((sizeof(fail_data.report_expected) == sizeof(char) || sizeof(fail_data.repp
> [22:00:34] not ok 45 - vmalloc_oob
>
> With this change, it instead looks like:
> [22:04:04] [FAILED] vmalloc_oob
> [22:04:04] # vmalloc_oob: EXPECTATION FAILED at lib/test_kasan.c:993
> [22:04:04] KASAN failure expected in "((volatile char *)area)[3100]", but none occurred
> [22:04:04] not ok 45 - vmalloc_oob
>
> Signed-off-by: David Gow <[email protected]>
> ---
>
> Stumbled across this because the vmalloc_oob test is failing (i.e.,
> KASAN isn't picking up an error) under qemu on my system, and the
> message above was horrifying. (I'll file a Bugzilla bug for the test
> failure today.)
>
> Cheers,
> -- David
>
> lib/test_kasan.c | 8 +++++---
> 1 file changed, 5 insertions(+), 3 deletions(-)
>
> diff --git a/lib/test_kasan.c b/lib/test_kasan.c
> index cacbbbdef768..deda13c9d9ff 100644
> --- a/lib/test_kasan.c
> +++ b/lib/test_kasan.c
> @@ -98,9 +98,11 @@ static void kasan_test_exit(struct kunit *test)
> barrier(); \
> expression; \
> barrier(); \
> - KUNIT_EXPECT_EQ(test, \
> - READ_ONCE(fail_data.report_expected), \
What do we have fail_data.report_expected for? Could we remove it now?
I think it's unused now.
> - READ_ONCE(fail_data.report_found)); \
> + if (READ_ONCE(fail_data.report_found) == false) { \
if (!READ_ONCE(fail_data.report_found)) {
?
> + KUNIT_FAIL(test, KUNIT_SUBTEST_INDENT "KASAN failure " \
> + "expected in \"" #expression \
> + "\", but none occurred"); \
> + } \
Thanks,
-- Marco
On Fri, Jun 4, 2021 at 3:55 PM Marco Elver <[email protected]> wrote:
>
> On Fri, 4 Jun 2021 at 07:26, 'David Gow' via kasan-dev
> <[email protected]> wrote:
> > The KUNIT_EXPECT_KASAN_FAIL() macro currently uses KUNIT_EXPECT_EQ() to
> > compare fail_data.report_expected and fail_data.report_found. This
> > always gave a somewhat useless error message on failure, but the
> > addition of extra compile-time checking with READ_ONCE() has caused it
> > to get much longer, and be truncated before anything useful is displayed.
> >
> > Instead, just check fail_data.report_found by hand (we've just test
> > report_expected to 'true'), and print a better failure message with
> > KUNIT_FAIL()
> >
> > Beforehand, a failure in:
> > KUNIT_EXPECT_KASAN_FAIL(test, ((volatile char *)area)[3100]);
> > would looked like:
> > [22:00:34] [FAILED] vmalloc_oob
> > [22:00:34] # vmalloc_oob: EXPECTATION FAILED at lib/test_kasan.c:991
> > [22:00:34] Expected ({ do { extern void __compiletime_assert_705(void) __attribute__((__error__("Unsupported access size for {READ,WRITE}_ONCE()."))); if (!((sizeof(fail_data.report_expected) == sizeof(char) || sizeof(fail_data.repp
> > [22:00:34] not ok 45 - vmalloc_oob
> >
> > With this change, it instead looks like:
> > [22:04:04] [FAILED] vmalloc_oob
> > [22:04:04] # vmalloc_oob: EXPECTATION FAILED at lib/test_kasan.c:993
> > [22:04:04] KASAN failure expected in "((volatile char *)area)[3100]", but none occurred
> > [22:04:04] not ok 45 - vmalloc_oob
> >
> > Signed-off-by: David Gow <[email protected]>
> > ---
> >
> > Stumbled across this because the vmalloc_oob test is failing (i.e.,
> > KASAN isn't picking up an error) under qemu on my system, and the
> > message above was horrifying. (I'll file a Bugzilla bug for the test
> > failure today.)
> >
> > Cheers,
> > -- David
> >
> > lib/test_kasan.c | 8 +++++---
> > 1 file changed, 5 insertions(+), 3 deletions(-)
> >
> > diff --git a/lib/test_kasan.c b/lib/test_kasan.c
> > index cacbbbdef768..deda13c9d9ff 100644
> > --- a/lib/test_kasan.c
> > +++ b/lib/test_kasan.c
> > @@ -98,9 +98,11 @@ static void kasan_test_exit(struct kunit *test)
> > barrier(); \
> > expression; \
> > barrier(); \
> > - KUNIT_EXPECT_EQ(test, \
> > - READ_ONCE(fail_data.report_expected), \
>
> What do we have fail_data.report_expected for? Could we remove it now?
> I think it's unused now.
>
I thought this was being used in kasan_update_kunit_status() (in
mm/kasan/report.c), but it looks like I was mistaken. We should be
able to get rid of it, then/
> > - READ_ONCE(fail_data.report_found)); \
> > + if (READ_ONCE(fail_data.report_found) == false) { \
>
> if (!READ_ONCE(fail_data.report_found)) {
> ?
>
I'll change this for v2.
> > + KUNIT_FAIL(test, KUNIT_SUBTEST_INDENT "KASAN failure " \
> > + "expected in \"" #expression \
> > + "\", but none occurred"); \
> > + } \
>
> Thanks,
> -- Marco