Context: It's difficult to map a given .kunitconfig => set of enabled tests. Letting kunit.py figure that out would be useful.

This patch: * is intended to be an implementation detail used only by kunit.py * adds a kunit.action module param with one valid non-null value, "list" * for the "list" action, it simply prints out "<suite>.<test>" * leaves the kunit.py changes to make use of this for another patch.

Note: kunit.filter_glob is respected for this and all future actions.

Hack: we print a TAP header (but no test plan) to allow kunit.py to use the same code to pick up KUnit output that it does for normal tests. Since this is intended to be an implementation detail, it seems fine for now. Maybe in the future we output each test as SKIPPED or the like.

Go with a more generic "action" param, since it seems like we might eventually have more modes besides just running or listing tests, e.g. * perhaps a benchmark mode that reruns test cases and reports timing * perhaps a deflake mode that reruns test cases that failed * perhaps a mode where we randomize test order to try and catch hermeticity bugs like "test a only passes if run after test b"

Tested: $ ./tools/testing/kunit/kunit.py run --kernel_arg=kunit.action=list --raw_output=kunit ... TAP version 14 1..1 example.example_simple_test example.example_skip_test example.example_mark_skipped_test reboot: System halted

Signed-off-by: Daniel Latypov dlatypov@google.com Reviewed-by: David Gow davidgow@google.com --- lib/kunit/executor.c | 45 +++++++++++++++++++++++++++++++++++++++----- 1 file changed, 40 insertions(+), 5 deletions(-)

diff --git a/lib/kunit/executor.c b/lib/kunit/executor.c index bab3ab940acc..ce1f57a44ab9 100644 --- a/lib/kunit/executor.c +++ b/lib/kunit/executor.c @@ -15,9 +15,16 @@ extern struct kunit_suite * const * const __kunit_suites_end[]; #if IS_BUILTIN(CONFIG_KUNIT)

static char *filter_glob_param; +static char *action_param; + module_param_named(filter_glob, filter_glob_param, charp, 0); MODULE_PARM_DESC(filter_glob, "Filter which KUnit test suites/tests run at boot-time, e.g. list* or list*.*del_test"); +module_param_named(action, action_param, charp, 0); +MODULE_PARM_DESC(action, + "Changes KUnit executor behavior, valid values are:\n" + "<none>: run the tests like normal\n" + "'list' to list test names instead of running them.\n");

/* glob_match() needs NULL terminated strings, so we need a copy of filter_glob_param. */ struct kunit_test_filter { @@ -196,9 +203,35 @@ static void kunit_print_tap_header(struct suite_set *suite_set) pr_info("1..%d\n", num_of_suites); }

-int kunit_run_all_tests(void) +static void kunit_exec_run_tests(struct suite_set *suite_set) { struct kunit_suite * const * const *suites; + + kunit_print_tap_header(suite_set); + + for (suites = suite_set->start; suites < suite_set->end; suites++) + __kunit_test_suites_init(*suites); +} + +static void kunit_exec_list_tests(struct suite_set *suite_set) +{ + unsigned int i; + struct kunit_suite * const * const *suites; + struct kunit_case *test_case; + + /* Hack: print a tap header so kunit.py can find the start of KUnit output. */ + pr_info("TAP version 14\n"); + + for (suites = suite_set->start; suites < suite_set->end; suites++) + for (i = 0; (*suites)[i] != NULL; i++) { + kunit_suite_for_each_test_case((*suites)[i], test_case) { + pr_info("%s.%s\n", (*suites)[i]->name, test_case->name); + } + } +} + +int kunit_run_all_tests(void) +{ struct suite_set suite_set = { .start = __kunit_suites_start, .end = __kunit_suites_end, @@ -207,10 +240,12 @@ int kunit_run_all_tests(void) if (filter_glob_param) suite_set = kunit_filter_suites(&suite_set, filter_glob_param);

- kunit_print_tap_header(&suite_set); - - for (suites = suite_set.start; suites < suite_set.end; suites++) - __kunit_test_suites_init(*suites); + if (!action_param) + kunit_exec_run_tests(&suite_set); + else if (strcmp(action_param, "list") == 0) + kunit_exec_list_tests(&suite_set); + else + pr_err("kunit executor: unknown action '%s'\n", action_param);

if (filter_glob_param) { /* a copy was made of each array */ kunit_free_suite_set(suite_set);

This is a long standing bug in kunit tool. Since these files were added, run_kernel() has always yielded lines.

That means, the call to run_kernel() returns before the kernel finishes executing tests, potentially before a single line of output is even produced.

So code like this time_start = time.time() result = linux.run_kernel(...) time_end = time.time()

would only measure the time taken for python to give back the generator object.

From a caller's perspective, the only way to know the kernel has exited

is for us to consume all the output from the `result` generator object. Alternatively, we could change run_kernel() to try and do its own book keeping and return the total time, but that doesn't seem worth it.

This change makes us record `time_end` after we're doing parsing all the output (which should mean we've consumed all of it, or errored out). That means we're including in the parsing time as well, but that should be quite small, and it's better than claiming it took 0s to run tests.

Let's use this as an example: $ ./tools/testing/kunit/kunit.py run --kunitconfig=lib/kunit example

Before: Elapsed time: 7.684s total, 0.001s configuring, 4.692s building, 0.000s running

After: Elapsed time: 6.283s total, 0.001s configuring, 3.202s building, 3.079s running

Signed-off-by: Daniel Latypov dlatypov@google.com --- tools/testing/kunit/kunit.py | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/tools/testing/kunit/kunit.py b/tools/testing/kunit/kunit.py index 31eec9f6ecc3..5e717594df5b 100755 --- a/tools/testing/kunit/kunit.py +++ b/tools/testing/kunit/kunit.py @@ -101,12 +101,14 @@ def exec_tests(linux: kunit_kernel.LinuxSourceTree, request: KunitExecRequest, filter_glob=request.filter_glob, build_dir=request.build_dir)

+ result = parse_tests(parse_request, run_result) + + # run_kernel() doesn't block on the kernel exiting. + # That only happens after we get the last line of output from `run_result`. + # So exec_time here actually contains parsing + execution time, which is fine. test_end = time.time() exec_time = test_end - test_start

- # Named tuples are immutable, so we rebuild them here manually - result = parse_tests(parse_request, run_result) - return KunitResult(status=result.status, result=result.result, elapsed_time=exec_time)

def parse_tests(request: KunitParseRequest, input_data: Iterable[str]) -> KunitResult:

The new --run_isolated flag makes the tool boot the kernel once per suite or test, preventing leftover state from one suite to impact the other. This can be useful as a starting point to debugging test hermeticity issues.

Note: it takes a lot longer, so people should not use it normally.

Consider the following very simplified example:

bool disable_something_for_test = false; void function_being_tested() { ... if (disable_something_for_test) return; ... }

static void test_before(struct kunit *test) { disable_something_for_test = true; function_being_tested(); /* oops, we forgot to reset it back to false */ }

static void test_after(struct kunit *test) { /* oops, now "fixing" test_before can cause test_after to fail! */ function_being_tested(); }

Presented like this, the issues are obvious, but it gets a lot more complicated to track down as the amount of test setup and helper functions increases.

Another use case is memory corruption. It might not be surfaced as a failure/crash in the test case or suite that caused it. I've noticed in kunit's own unit tests, the 3rd suite after might be the one to finally crash after an out-of-bounds write, for example.

Example usage:

Per suite: $ ./tools/testing/kunit/kunit.py run --kunitconfig=lib/kunit --run_isolated=suite ... Starting KUnit Kernel (1/7)... ============================================================ ======== [PASSED] kunit_executor_test ======== .... Testing complete. 5 tests run. 0 failed. 0 crashed. 0 skipped. Starting KUnit Kernel (2/7)... ============================================================ ======== [PASSED] kunit-try-catch-test ======== ...

Per test: $ ./tools/testing/kunit/kunit.py run --kunitconfig=lib/kunit --run_isolated=test Starting KUnit Kernel (1/23)... ============================================================ ======== [PASSED] kunit_executor_test ======== [PASSED] parse_filter_test ============================================================ Testing complete. 1 tests run. 0 failed. 0 crashed. 0 skipped. Starting KUnit Kernel (2/23)... ============================================================ ======== [PASSED] kunit_executor_test ======== [PASSED] filter_subsuite_test ...

It works with filters as well: $ ./tools/testing/kunit/kunit.py run --kunitconfig=lib/kunit --run_isolated=suite example ... Starting KUnit Kernel (1/1)... ============================================================ ======== [PASSED] example ======== ...

It also handles test filters, '*.*skip*' runs these 3 tests: kunit_status.kunit_status_mark_skipped_test example.example_skip_test example.example_mark_skipped_test

Signed-off-by: Daniel Latypov dlatypov@google.com Reviewed-by: David Gow davidgow@google.com --- tools/testing/kunit/kunit.py | 95 ++++++++++++++++++++------ tools/testing/kunit/kunit_tool_test.py | 40 +++++++++++ 2 files changed, 114 insertions(+), 21 deletions(-)

diff --git a/tools/testing/kunit/kunit.py b/tools/testing/kunit/kunit.py index 5e717594df5b..b9d63f558765 100755 --- a/tools/testing/kunit/kunit.py +++ b/tools/testing/kunit/kunit.py @@ -16,7 +16,7 @@ assert sys.version_info >= (3, 7), "Python version is too old"

from collections import namedtuple from enum import Enum, auto -from typing import Iterable +from typing import Iterable, List

import kunit_config import kunit_json @@ -31,13 +31,13 @@ KunitBuildRequest = namedtuple('KunitBuildRequest', ['jobs', 'build_dir', 'alltests', 'make_options']) KunitExecRequest = namedtuple('KunitExecRequest', - ['timeout', 'build_dir', 'alltests', - 'filter_glob', 'kernel_args']) + ['timeout', 'build_dir', 'alltests', + 'filter_glob', 'kernel_args', 'run_isolated']) KunitParseRequest = namedtuple('KunitParseRequest', ['raw_output', 'build_dir', 'json']) KunitRequest = namedtuple('KunitRequest', ['raw_output','timeout', 'jobs', 'build_dir', 'alltests', 'filter_glob', - 'kernel_args', 'json', 'make_options']) + 'kernel_args', 'run_isolated', 'json', 'make_options'])

KernelDirectoryPath = sys.argv[0].split('tools/testing/kunit/')[0]

@@ -91,23 +91,68 @@ def build_tests(linux: kunit_kernel.LinuxSourceTree, 'built kernel successfully', build_end - build_start)

+def _list_tests(linux: kunit_kernel.LinuxSourceTree, request: KunitExecRequest) -> List[str]: + args = ['kunit.action=list'] + if request.kernel_args: + args.extend(request.kernel_args) + + output = linux.run_kernel(args=args, + timeout=None if request.alltests else request.timeout, + filter_glob=request.filter_glob, + build_dir=request.build_dir) + lines = kunit_parser.extract_tap_lines(output) + # Hack! Drop the dummy TAP version header that the executor prints out. + lines.pop() + return list(lines) + +def _suites_from_test_list(tests: List[str]) -> List[str]: + """Extracts all the suites from an ordered list of tests.""" + suites = [] # type: List[str] + for t in tests: + parts = t.split('.', maxsplit=2) + if len(parts) != 2: + raise ValueError(f'internal KUnit error, test name should be of the form "<suite>.<test>", got "{t}"') + suite, case = parts + if not suites or suites[-1] != suite: + suites.append(suite) + return suites + + + def exec_tests(linux: kunit_kernel.LinuxSourceTree, request: KunitExecRequest, parse_request: KunitParseRequest) -> KunitResult: - kunit_parser.print_with_timestamp('Starting KUnit Kernel ...') - test_start = time.time() - run_result = linux.run_kernel( - args=request.kernel_args, - timeout=None if request.alltests else request.timeout, - filter_glob=request.filter_glob, - build_dir=request.build_dir) - - result = parse_tests(parse_request, run_result) - - # run_kernel() doesn't block on the kernel exiting. - # That only happens after we get the last line of output from `run_result`. - # So exec_time here actually contains parsing + execution time, which is fine. - test_end = time.time() - exec_time = test_end - test_start + filter_globs = [request.filter_glob] + if request.run_isolated: + tests = _list_tests(linux, request) + if request.run_isolated == 'test': + filter_globs = tests + if request.run_isolated == 'suite': + filter_globs = _suites_from_test_list(tests) + # Apply the test-part of the user's glob, if present. + if '.' in request.filter_glob: + test_glob = request.filter_glob.split('.', maxsplit=2)[1] + filter_globs = [g + '.'+ test_glob for g in filter_globs] + + overall_status = kunit_parser.TestStatus.SUCCESS + exec_time = 0.0 + for i, filter_glob in enumerate(filter_globs): + kunit_parser.print_with_timestamp('Starting KUnit Kernel ({}/{})...'.format(i+1, len(filter_globs))) + + test_start = time.time() + run_result = linux.run_kernel( + args=request.kernel_args, + timeout=None if request.alltests else request.timeout, + filter_glob=filter_glob, + build_dir=request.build_dir) + + result = parse_tests(parse_request, run_result) + # run_kernel() doesn't block on the kernel exiting. + # That only happens after we get the last line of output from `run_result`. + # So exec_time here actually contains parsing + execution time, which is fine. + test_end = time.time() + exec_time += test_end - test_start + + overall_status = kunit_parser.max_status(overall_status, result.status)

return KunitResult(status=result.status, result=result.result, elapsed_time=exec_time)

@@ -168,7 +213,7 @@ def run_tests(linux: kunit_kernel.LinuxSourceTree,

exec_request = KunitExecRequest(request.timeout, request.build_dir, request.alltests, request.filter_glob, - request.kernel_args) + request.kernel_args, request.run_isolated) parse_request = KunitParseRequest(request.raw_output, request.build_dir, request.json) @@ -252,6 +297,12 @@ def add_exec_opts(parser) -> None: parser.add_argument('--kernel_args', help='Kernel command-line parameters. Maybe be repeated', action='append') + parser.add_argument('--run_isolated', help='If set, boot the kernel for each ' + 'individual suite/test. This is can be useful for debugging ' + 'a non-hermetic test, one that might pass/fail based on ' + 'what ran before it.', + type=str, + choices=['suite', 'test']),

def add_parse_opts(parser) -> None: parser.add_argument('--raw_output', help='If set don't format output from kernel. ' @@ -325,6 +376,7 @@ def main(argv, linux=None): cli_args.alltests, cli_args.filter_glob, cli_args.kernel_args, + cli_args.run_isolated, cli_args.json, cli_args.make_options) result = run_tests(linux, request) @@ -380,7 +432,8 @@ def main(argv, linux=None): cli_args.build_dir, cli_args.alltests, cli_args.filter_glob, - cli_args.kernel_args) + cli_args.kernel_args, + cli_args.run_isolated) parse_request = KunitParseRequest(cli_args.raw_output, cli_args.build_dir, cli_args.json) diff --git a/tools/testing/kunit/kunit_tool_test.py b/tools/testing/kunit/kunit_tool_test.py index 619c4554cbff..1ff35c08d2c8 100755 --- a/tools/testing/kunit/kunit_tool_test.py +++ b/tools/testing/kunit/kunit_tool_test.py @@ -477,6 +477,46 @@ class KUnitMainTest(unittest.TestCase): args=['a=1','b=2'], build_dir='.kunit', filter_glob='', timeout=300) self.print_mock.assert_any_call(StrContains('Testing complete.'))

+ def test_list_tests(self): + want = ['suite.test1', 'suite.test2', 'suite2.test1'] + self.linux_source_mock.run_kernel.return_value = ['TAP version 14'] + want + + got = kunit._list_tests(self.linux_source_mock, + kunit.KunitExecRequest(300, '.kunit', False, 'suite*', None, 'suite')) + + self.assertEqual(got, want) + # Should respect the user's filter glob when listing tests. + self.linux_source_mock.run_kernel.assert_called_once_with( + args=['kunit.action=list'], build_dir='.kunit', filter_glob='suite*', timeout=300) + + + @mock.patch.object(kunit, '_list_tests') + def test_run_isolated_by_suite(self, mock_tests): + mock_tests.return_value = ['suite.test1', 'suite.test2', 'suite2.test1'] + kunit.main(['exec', '--run_isolated=suite', 'suite*.test*'], self.linux_source_mock) + + # Should respect the user's filter glob when listing tests. + mock_tests.assert_called_once_with(mock.ANY, + kunit.KunitExecRequest(300, '.kunit', False, 'suite*.test*', None, 'suite')) + self.linux_source_mock.run_kernel.assert_has_calls([ + mock.call(args=None, build_dir='.kunit', filter_glob='suite.test*', timeout=300), + mock.call(args=None, build_dir='.kunit', filter_glob='suite2.test*', timeout=300), + ]) + + @mock.patch.object(kunit, '_list_tests') + def test_run_isolated_by_test(self, mock_tests): + mock_tests.return_value = ['suite.test1', 'suite.test2', 'suite2.test1'] + kunit.main(['exec', '--run_isolated=test', 'suite*'], self.linux_source_mock) + + # Should respect the user's filter glob when listing tests. + mock_tests.assert_called_once_with(mock.ANY, + kunit.KunitExecRequest(300, '.kunit', False, 'suite*', None, 'test')) + self.linux_source_mock.run_kernel.assert_has_calls([ + mock.call(args=None, build_dir='.kunit', filter_glob='suite.test1', timeout=300), + mock.call(args=None, build_dir='.kunit', filter_glob='suite.test2', timeout=300), + mock.call(args=None, build_dir='.kunit', filter_glob='suite2.test1', timeout=300), + ]) +

if __name__ == '__main__': unittest.main()