| Message ID | 20210611035725.1248874-1-davidgow@google.com |
|---|---|
| State | Accepted |
| Commit | 384426bd101cb3cd580b18de19d4891ec5ca5bf9 |
| Headers | show |
| Series | kunit: Fix result propagation for parameterised tests | expand |
On 6/11/21 2:29 AM, Marco Elver wrote: > On Fri, 11 Jun 2021 at 05:57, David Gow <davidgow@google.com> wrote: >> >> When one parameter of a parameterised test failed, its failure would be >> propagated to the overall test, but not to the suite result (unless it >> was the last parameter). >> >> This is because test_case->success was being reset to the test->success >> result after each parameter was used, so a failing test's result would >> be overwritten by a non-failing result. The overall test result was >> handled in a third variable, test_result, but this was disacarded after >> the status line was printed. >> >> Instead, just propagate the result after each parameter run. >> >> Signed-off-by: David Gow <davidgow@google.com> >> Fixes: fadb08e7c750 ("kunit: Support for Parameterized Testing") > > Reviewed-by: Marco Elver <elver@google.com> > > Would Cc: stable be appropriate? > > Thanks, > -- Marco > >> --- >> >> This is fixing quite a serious bug where some test suites would appear >> to succeed even if some of their component tests failed. It'd be nice to >> get this into kunit-fixes ASAP. >> Will apply this with cc stable. >> (This will require a rework of some of the skip tests work, for which >> I'll send out a new version soon.) >> Thanks for the heads up. I will wait for new version. thanks, -- Shuah
On Sat, Jun 12, 2021 at 1:44 AM Shuah Khan <skhan@linuxfoundation.org> wrote: > > On 6/11/21 2:29 AM, Marco Elver wrote: > > On Fri, 11 Jun 2021 at 05:57, David Gow <davidgow@google.com> wrote: > >> > >> When one parameter of a parameterised test failed, its failure would be > >> propagated to the overall test, but not to the suite result (unless it > >> was the last parameter). > >> > >> This is because test_case->success was being reset to the test->success > >> result after each parameter was used, so a failing test's result would > >> be overwritten by a non-failing result. The overall test result was > >> handled in a third variable, test_result, but this was disacarded after > >> the status line was printed. > >> > >> Instead, just propagate the result after each parameter run. > >> > >> Signed-off-by: David Gow <davidgow@google.com> > >> Fixes: fadb08e7c750 ("kunit: Support for Parameterized Testing") > > > > Reviewed-by: Marco Elver <elver@google.com> > > > > Would Cc: stable be appropriate? > > > > Thanks, > > -- Marco > > > >> --- > >> > >> This is fixing quite a serious bug where some test suites would appear > >> to succeed even if some of their component tests failed. It'd be nice to > >> get this into kunit-fixes ASAP. > >> > > Will apply this with cc stable. > Thanks! > >> (This will require a rework of some of the skip tests work, for which > >> I'll send out a new version soon.) > >> > > Thanks for the heads up. I will wait for new version. > Thanks: I've sent out v4 which fixes this: https://lore.kernel.org/linux-kselftest/20210611070802.1318911-1-davidgow@google.com/ It's rebased on top of this patch, so depends on it, and also depends on the first two patches in the "Do not typecheck binary assertions" series: https://lore.kernel.org/linux-kselftest/20210513193204.816681-1-davidgow@google.com/ > thanks, > -- Shuah
diff --git a/lib/kunit/test.c b/lib/kunit/test.c index 2f6cc0123232..17973a4a44c2 100644 --- a/lib/kunit/test.c +++ b/lib/kunit/test.c @@ -376,7 +376,7 @@ static void kunit_run_case_catch_errors(struct kunit_suite *suite, context.test_case = test_case; kunit_try_catch_run(try_catch, &context); - test_case->success = test->success; + test_case->success &= test->success; } int kunit_run_tests(struct kunit_suite *suite) @@ -388,7 +388,7 @@ int kunit_run_tests(struct kunit_suite *suite) kunit_suite_for_each_test_case(suite, test_case) { struct kunit test = { .param_value = NULL, .param_index = 0 }; - bool test_success = true; + test_case->success = true; if (test_case->generate_params) { /* Get initial param. */ @@ -398,7 +398,6 @@ int kunit_run_tests(struct kunit_suite *suite) do { kunit_run_case_catch_errors(suite, test_case, &test); - test_success &= test_case->success; if (test_case->generate_params) { if (param_desc[0] == '\0') { @@ -420,7 +419,7 @@ int kunit_run_tests(struct kunit_suite *suite) } } while (test.param_value); - kunit_print_ok_not_ok(&test, true, test_success, + kunit_print_ok_not_ok(&test, true, test_case->success, kunit_test_case_num(suite, test_case), test_case->name); }
When one parameter of a parameterised test failed, its failure would be propagated to the overall test, but not to the suite result (unless it was the last parameter). This is because test_case->success was being reset to the test->success result after each parameter was used, so a failing test's result would be overwritten by a non-failing result. The overall test result was handled in a third variable, test_result, but this was disacarded after the status line was printed. Instead, just propagate the result after each parameter run. Signed-off-by: David Gow <davidgow@google.com> Fixes: fadb08e7c750 ("kunit: Support for Parameterized Testing") --- This is fixing quite a serious bug where some test suites would appear to succeed even if some of their component tests failed. It'd be nice to get this into kunit-fixes ASAP. (This will require a rework of some of the skip tests work, for which I'll send out a new version soon.) Cheers, -- David lib/kunit/test.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-)