Message ID | e75010281350ff3a4380006218f81e1233fa4e6b.1610554432.git.andreyknvl@google.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | kasan: HW_TAGS tests support and fixes | expand |
On Wed, 13 Jan 2021 at 17:22, Andrey Konovalov <andreyknvl@google.com> wrote: > > It might not be obvious to the compiler that the expression must be > executed between writing and reading to fail_data. In this case, the > compiler might reorder or optimize away some of the accesses, and > the tests will fail. > > Add compiler barriers around the expression in KUNIT_EXPECT_KASAN_FAIL > and use READ/WRITE_ONCE() for accessing fail_data fields. > > Link: https://linux-review.googlesource.com/id/I046079f48641a1d36fe627fc8827a9249102fd50 > Signed-off-by: Andrey Konovalov <andreyknvl@google.com> Reviewed-by: Marco Elver <elver@google.com> > --- > lib/test_kasan.c | 17 ++++++++++++----- > mm/kasan/report.c | 2 +- > 2 files changed, 13 insertions(+), 6 deletions(-) > > diff --git a/lib/test_kasan.c b/lib/test_kasan.c > index 5c8aa3a5ce93..283feda9882a 100644 > --- a/lib/test_kasan.c > +++ b/lib/test_kasan.c > @@ -68,23 +68,30 @@ static void kasan_test_exit(struct kunit *test) > * normally auto-disabled. When this happens, this test handler reenables > * tag checking. As tag checking can be only disabled or enabled per CPU, this > * handler disables migration (preemption). > + * > + * Since the compiler doesn't see that the expression can change the fail_data > + * fields, it can reorder or optimize away the accesses to those fields. > + * Use READ/WRITE_ONCE() for the accesses and compiler barriers around the > + * expression to prevent that. > */ > #define KUNIT_EXPECT_KASAN_FAIL(test, expression) do { \ > if (IS_ENABLED(CONFIG_KASAN_HW_TAGS)) \ > migrate_disable(); \ > - fail_data.report_expected = true; \ > - fail_data.report_found = false; \ > + WRITE_ONCE(fail_data.report_expected, true); \ > + WRITE_ONCE(fail_data.report_found, false); \ > kunit_add_named_resource(test, \ > NULL, \ > NULL, \ > &resource, \ > "kasan_data", &fail_data); \ > + barrier(); \ > expression; \ > + barrier(); \ > KUNIT_EXPECT_EQ(test, \ > - fail_data.report_expected, \ > - fail_data.report_found); \ > + READ_ONCE(fail_data.report_expected), \ > + READ_ONCE(fail_data.report_found)); \ > if (IS_ENABLED(CONFIG_KASAN_HW_TAGS)) { \ > - if (fail_data.report_found) \ > + if (READ_ONCE(fail_data.report_found)) \ > hw_enable_tagging(); \ > migrate_enable(); \ > } \ > diff --git a/mm/kasan/report.c b/mm/kasan/report.c > index e93d7973792e..234f35a84f19 100644 > --- a/mm/kasan/report.c > +++ b/mm/kasan/report.c > @@ -331,7 +331,7 @@ static void kasan_update_kunit_status(struct kunit *cur_test) > } > > kasan_data = (struct kunit_kasan_expectation *)resource->data; > - kasan_data->report_found = true; > + WRITE_ONCE(kasan_data->report_found, true); > kunit_put_resource(resource); > } > #endif /* IS_ENABLED(CONFIG_KUNIT) */ > -- > 2.30.0.284.gd98b1dd5eaa7-goog >
diff --git a/lib/test_kasan.c b/lib/test_kasan.c index 5c8aa3a5ce93..283feda9882a 100644 --- a/lib/test_kasan.c +++ b/lib/test_kasan.c @@ -68,23 +68,30 @@ static void kasan_test_exit(struct kunit *test) * normally auto-disabled. When this happens, this test handler reenables * tag checking. As tag checking can be only disabled or enabled per CPU, this * handler disables migration (preemption). + * + * Since the compiler doesn't see that the expression can change the fail_data + * fields, it can reorder or optimize away the accesses to those fields. + * Use READ/WRITE_ONCE() for the accesses and compiler barriers around the + * expression to prevent that. */ #define KUNIT_EXPECT_KASAN_FAIL(test, expression) do { \ if (IS_ENABLED(CONFIG_KASAN_HW_TAGS)) \ migrate_disable(); \ - fail_data.report_expected = true; \ - fail_data.report_found = false; \ + WRITE_ONCE(fail_data.report_expected, true); \ + WRITE_ONCE(fail_data.report_found, false); \ kunit_add_named_resource(test, \ NULL, \ NULL, \ &resource, \ "kasan_data", &fail_data); \ + barrier(); \ expression; \ + barrier(); \ KUNIT_EXPECT_EQ(test, \ - fail_data.report_expected, \ - fail_data.report_found); \ + READ_ONCE(fail_data.report_expected), \ + READ_ONCE(fail_data.report_found)); \ if (IS_ENABLED(CONFIG_KASAN_HW_TAGS)) { \ - if (fail_data.report_found) \ + if (READ_ONCE(fail_data.report_found)) \ hw_enable_tagging(); \ migrate_enable(); \ } \ diff --git a/mm/kasan/report.c b/mm/kasan/report.c index e93d7973792e..234f35a84f19 100644 --- a/mm/kasan/report.c +++ b/mm/kasan/report.c @@ -331,7 +331,7 @@ static void kasan_update_kunit_status(struct kunit *cur_test) } kasan_data = (struct kunit_kasan_expectation *)resource->data; - kasan_data->report_found = true; + WRITE_ONCE(kasan_data->report_found, true); kunit_put_resource(resource); } #endif /* IS_ENABLED(CONFIG_KUNIT) */
It might not be obvious to the compiler that the expression must be executed between writing and reading to fail_data. In this case, the compiler might reorder or optimize away some of the accesses, and the tests will fail. Add compiler barriers around the expression in KUNIT_EXPECT_KASAN_FAIL and use READ/WRITE_ONCE() for accessing fail_data fields. Link: https://linux-review.googlesource.com/id/I046079f48641a1d36fe627fc8827a9249102fd50 Signed-off-by: Andrey Konovalov <andreyknvl@google.com> --- lib/test_kasan.c | 17 ++++++++++++----- mm/kasan/report.c | 2 +- 2 files changed, 13 insertions(+), 6 deletions(-)