cpython/Lib/test/libregrtest/runtest_mp.py

558 lines
18 KiB
Python
Raw Normal View History

import faulthandler
import json
import os.path
import queue
import shlex
import signal
import subprocess
import sys
import tempfile
import threading
import time
import traceback
from typing import NamedTuple, NoReturn, Literal, Any
from test import support
from test.support import os_helper
from test.libregrtest.cmdline import Namespace
from test.libregrtest.main import Regrtest
from test.libregrtest.runtest import (
runtest, is_failed, TestResult, Interrupted, Timeout, ChildError,
PROGRESS_MIN_TIME, Passed, EnvChanged)
from test.libregrtest.setup import setup_tests
from test.libregrtest.utils import format_duration, print_warning
# Display the running tests if nothing happened last N seconds
PROGRESS_UPDATE = 30.0 # seconds
assert PROGRESS_UPDATE >= PROGRESS_MIN_TIME
# Kill the main process after 5 minutes. It is supposed to write an update
# every PROGRESS_UPDATE seconds. Tolerate 5 minutes for Python slowest
# buildbot workers.
MAIN_PROCESS_TIMEOUT = 5 * 60.0
assert MAIN_PROCESS_TIMEOUT >= PROGRESS_UPDATE
# Time to wait until a worker completes: should be immediate
JOIN_TIMEOUT = 30.0 # seconds
USE_PROCESS_GROUP = (hasattr(os, "setsid") and hasattr(os, "killpg"))
def must_stop(result: TestResult, ns: Namespace) -> bool:
if isinstance(result, Interrupted):
return True
if ns.failfast and is_failed(result, ns):
return True
return False
def parse_worker_args(worker_args) -> tuple[Namespace, str]:
ns_dict, test_name = json.loads(worker_args)
ns = Namespace(**ns_dict)
return (ns, test_name)
def run_test_in_subprocess(testname: str, ns: Namespace, tmp_dir: str) -> subprocess.Popen:
ns_dict = vars(ns)
worker_args = (ns_dict, testname)
worker_args = json.dumps(worker_args)
if ns.python is not None:
# The "executable" may be two or more parts, e.g. "node python.js"
executable = shlex.split(ns.python)
else:
executable = [sys.executable]
cmd = [*executable, *support.args_from_interpreter_flags(),
'-u', # Unbuffered stdout and stderr
'-m', 'test.regrtest',
'--worker-args', worker_args]
env = dict(os.environ)
if tmp_dir is not None:
env['TMPDIR'] = tmp_dir
env['TEMP'] = tmp_dir
env['TMP'] = tmp_dir
# Running the child from the same working directory as regrtest's original
# invocation ensures that TEMPDIR for the child is the same when
# sysconfig.is_python_build() is true. See issue 15300.
kw = {'env': env}
if USE_PROCESS_GROUP:
kw['start_new_session'] = True
return subprocess.Popen(cmd,
stdout=subprocess.PIPE,
# bpo-45410: Write stderr into stdout to keep
# messages order
stderr=subprocess.STDOUT,
universal_newlines=True,
close_fds=(os.name != 'nt'),
cwd=os_helper.SAVEDCWD,
**kw)
def run_tests_worker(ns: Namespace, test_name: str) -> NoReturn:
setup_tests(ns)
result = runtest(ns, test_name)
print() # Force a newline (just in case)
# Serialize TestResult as dict in JSON
print(json.dumps(result, cls=EncodeTestResult), flush=True)
sys.exit(0)
# We do not use a generator so multiple threads can call next().
class MultiprocessIterator:
"""A thread-safe iterator over tests for multiprocess mode."""
def __init__(self, tests_iter):
self.lock = threading.Lock()
self.tests_iter = tests_iter
def __iter__(self):
return self
def __next__(self):
with self.lock:
if self.tests_iter is None:
raise StopIteration
return next(self.tests_iter)
def stop(self):
with self.lock:
self.tests_iter = None
class MultiprocessResult(NamedTuple):
result: TestResult
# bpo-45410: stderr is written into stdout to keep messages order
stdout: str
error_msg: str
ExcStr = str
QueueOutput = tuple[Literal[False], MultiprocessResult] | tuple[Literal[True], ExcStr]
class ExitThread(Exception):
pass
class TestWorkerProcess(threading.Thread):
def __init__(self, worker_id: int, runner: "MultiprocessTestRunner") -> None:
super().__init__()
self.worker_id = worker_id
self.pending = runner.pending
self.output = runner.output
self.ns = runner.ns
self.timeout = runner.worker_timeout
self.regrtest = runner.regrtest
self.current_test_name = None
self.start_time = None
self._popen = None
self._killed = False
self._stopped = False
def __repr__(self) -> str:
info = [f'TestWorkerProcess #{self.worker_id}']
if self.is_alive():
info.append("running")
else:
info.append('stopped')
test = self.current_test_name
if test:
info.append(f'test={test}')
popen = self._popen
if popen is not None:
dt = time.monotonic() - self.start_time
info.extend((f'pid={self._popen.pid}',
f'time={format_duration(dt)}'))
return '<%s>' % ' '.join(info)
def _kill(self) -> None:
popen = self._popen
if popen is None:
return
if self._killed:
return
self._killed = True
if USE_PROCESS_GROUP:
what = f"{self} process group"
else:
what = f"{self}"
print(f"Kill {what}", file=sys.stderr, flush=True)
try:
if USE_PROCESS_GROUP:
os.killpg(popen.pid, signal.SIGKILL)
else:
popen.kill()
except ProcessLookupError:
# popen.kill(): the process completed, the TestWorkerProcess thread
# read its exit status, but Popen.send_signal() read the returncode
# just before Popen.wait() set returncode.
pass
except OSError as exc:
print_warning(f"Failed to kill {what}: {exc!r}")
def stop(self) -> None:
# Method called from a different thread to stop this thread
self._stopped = True
self._kill()
def mp_result_error(
self,
test_result: TestResult,
stdout: str = '',
err_msg=None
) -> MultiprocessResult:
test_result.duration_sec = time.monotonic() - self.start_time
return MultiprocessResult(test_result, stdout, err_msg)
def _run_process(self, test_name: str, tmp_dir: str) -> tuple[int, str, str]:
self.start_time = time.monotonic()
self.current_test_name = test_name
try:
popen = run_test_in_subprocess(test_name, self.ns, tmp_dir)
self._killed = False
self._popen = popen
except:
self.current_test_name = None
raise
try:
if self._stopped:
# If kill() has been called before self._popen is set,
# self._popen is still running. Call again kill()
# to ensure that the process is killed.
self._kill()
raise ExitThread
try:
# bpo-45410: stderr is written into stdout
stdout, _ = popen.communicate(timeout=self.timeout)
retcode = popen.returncode
assert retcode is not None
except subprocess.TimeoutExpired:
if self._stopped:
# kill() has been called: communicate() fails on reading
# closed stdout
raise ExitThread
# On timeout, kill the process
self._kill()
# None means TIMEOUT for the caller
retcode = None
# bpo-38207: Don't attempt to call communicate() again: on it
# can hang until all child processes using stdout
# pipes completes.
stdout = ''
except OSError:
if self._stopped:
# kill() has been called: communicate() fails
# on reading closed stdout
raise ExitThread
raise
else:
stdout = stdout.strip()
return (retcode, stdout)
except:
self._kill()
raise
finally:
self._wait_completed()
self._popen = None
self.current_test_name = None
def _runtest(self, test_name: str) -> MultiprocessResult:
# Don't check for leaked temporary files and directories if Python is
# run on WASI. WASI don't pass environment variables like TMPDIR to
# worker processes.
if not support.is_wasi:
# gh-93353: Check for leaked temporary files in the parent process,
# since the deletion of temporary files can happen late during
# Python finalization: too late for libregrtest.
tmp_dir = tempfile.mkdtemp(prefix="test_python_")
tmp_dir = os.path.abspath(tmp_dir)
try:
retcode, stdout = self._run_process(test_name, tmp_dir)
finally:
tmp_files = os.listdir(tmp_dir)
os_helper.rmtree(tmp_dir)
else:
retcode, stdout = self._run_process(test_name, None)
tmp_files = ()
if retcode is None:
return self.mp_result_error(Timeout(test_name), stdout)
err_msg = None
if retcode != 0:
err_msg = "Exit code %s" % retcode
else:
stdout, _, result = stdout.rpartition("\n")
stdout = stdout.rstrip()
if not result:
err_msg = "Failed to parse worker stdout"
else:
try:
# deserialize run_tests_worker() output
result = json.loads(result, object_hook=decode_test_result)
except Exception as exc:
err_msg = "Failed to parse worker JSON: %s" % exc
if err_msg is not None:
return self.mp_result_error(ChildError(test_name), stdout, err_msg)
if tmp_files:
msg = (f'\n\n'
f'Warning -- {test_name} leaked temporary files '
f'({len(tmp_files)}): {", ".join(sorted(tmp_files))}')
stdout += msg
if isinstance(result, Passed):
result = EnvChanged.from_passed(result)
return MultiprocessResult(result, stdout, err_msg)
def run(self) -> None:
while not self._stopped:
try:
try:
test_name = next(self.pending)
except StopIteration:
break
mp_result = self._runtest(test_name)
self.output.put((False, mp_result))
if must_stop(mp_result.result, self.ns):
break
except ExitThread:
break
except BaseException:
self.output.put((True, traceback.format_exc()))
break
def _wait_completed(self) -> None:
popen = self._popen
# stdout must be closed to ensure that communicate() does not hang
popen.stdout.close()
try:
popen.wait(JOIN_TIMEOUT)
except (subprocess.TimeoutExpired, OSError) as exc:
print_warning(f"Failed to wait for {self} completion "
f"(timeout={format_duration(JOIN_TIMEOUT)}): "
f"{exc!r}")
def wait_stopped(self, start_time: float) -> None:
# bpo-38207: MultiprocessTestRunner.stop_workers() called self.stop()
# which killed the process. Sometimes, killing the process from the
# main thread does not interrupt popen.communicate() in
# TestWorkerProcess thread. This loop with a timeout is a workaround
# for that.
#
# Moreover, if this method fails to join the thread, it is likely
# that Python will hang at exit while calling threading._shutdown()
# which tries again to join the blocked thread. Regrtest.main()
# uses EXIT_TIMEOUT to workaround this second bug.
while True:
# Write a message every second
self.join(1.0)
if not self.is_alive():
break
dt = time.monotonic() - start_time
self.regrtest.log(f"Waiting for {self} thread "
f"for {format_duration(dt)}")
if dt > JOIN_TIMEOUT:
print_warning(f"Failed to join {self} in {format_duration(dt)}")
break
def get_running(workers: list[TestWorkerProcess]) -> list[TestWorkerProcess]:
running = []
for worker in workers:
current_test_name = worker.current_test_name
if not current_test_name:
continue
dt = time.monotonic() - worker.start_time
if dt >= PROGRESS_MIN_TIME:
text = '%s (%s)' % (current_test_name, format_duration(dt))
running.append(text)
return running
class MultiprocessTestRunner:
def __init__(self, regrtest: Regrtest) -> None:
self.regrtest = regrtest
self.log = self.regrtest.log
self.ns = regrtest.ns
self.output: queue.Queue[QueueOutput] = queue.Queue()
self.pending = MultiprocessIterator(self.regrtest.tests)
if self.ns.timeout is not None:
# Rely on faulthandler to kill a worker process. This timouet is
# when faulthandler fails to kill a worker process. Give a maximum
# of 5 minutes to faulthandler to kill the worker.
self.worker_timeout = min(self.ns.timeout * 1.5,
self.ns.timeout + 5 * 60)
else:
self.worker_timeout = None
self.workers = None
def start_workers(self) -> None:
self.workers = [TestWorkerProcess(index, self)
for index in range(1, self.ns.use_mp + 1)]
msg = f"Run tests in parallel using {len(self.workers)} child processes"
if self.ns.timeout:
msg += (" (timeout: %s, worker timeout: %s)"
% (format_duration(self.ns.timeout),
format_duration(self.worker_timeout)))
self.log(msg)
for worker in self.workers:
worker.start()
def stop_workers(self) -> None:
start_time = time.monotonic()
for worker in self.workers:
worker.stop()
for worker in self.workers:
worker.wait_stopped(start_time)
def _get_result(self) -> QueueOutput | None:
use_faulthandler = (self.ns.timeout is not None)
timeout = PROGRESS_UPDATE
# bpo-46205: check the status of workers every iteration to avoid
# waiting forever on an empty queue.
while any(worker.is_alive() for worker in self.workers):
if use_faulthandler:
faulthandler.dump_traceback_later(MAIN_PROCESS_TIMEOUT,
exit=True)
# wait for a thread
try:
return self.output.get(timeout=timeout)
except queue.Empty:
pass
# display progress
running = get_running(self.workers)
if running and not self.ns.pgo:
self.log('running: %s' % ', '.join(running))
# all worker threads are done: consume pending results
try:
return self.output.get(timeout=0)
except queue.Empty:
return None
def display_result(self, mp_result: MultiprocessResult) -> None:
result = mp_result.result
text = str(result)
if mp_result.error_msg is not None:
# CHILD_ERROR
text += ' (%s)' % mp_result.error_msg
elif (result.duration_sec >= PROGRESS_MIN_TIME and not self.ns.pgo):
text += ' (%s)' % format_duration(result.duration_sec)
running = get_running(self.workers)
if running and not self.ns.pgo:
text += ' -- running: %s' % ', '.join(running)
self.regrtest.display_progress(self.test_index, text)
def _process_result(self, item: QueueOutput) -> bool:
"""Returns True if test runner must stop."""
if item[0]:
# Thread got an exception
format_exc = item[1]
print_warning(f"regrtest worker thread failed: {format_exc}")
return True
self.test_index += 1
mp_result = item[1]
self.regrtest.accumulate_result(mp_result.result)
self.display_result(mp_result)
if mp_result.stdout:
print(mp_result.stdout, flush=True)
if must_stop(mp_result.result, self.ns):
return True
return False
def run_tests(self) -> None:
self.start_workers()
self.test_index = 0
try:
while True:
item = self._get_result()
if item is None:
break
stop = self._process_result(item)
if stop:
break
except KeyboardInterrupt:
print()
self.regrtest.interrupted = True
finally:
if self.ns.timeout is not None:
faulthandler.cancel_dump_traceback_later()
# Always ensure that all worker processes are no longer
# worker when we exit this function
self.pending.stop()
self.stop_workers()
def run_tests_multiprocess(regrtest: Regrtest) -> None:
MultiprocessTestRunner(regrtest).run_tests()
class EncodeTestResult(json.JSONEncoder):
"""Encode a TestResult (sub)class object into a JSON dict."""
def default(self, o: Any) -> dict[str, Any]:
if isinstance(o, TestResult):
result = vars(o)
result["__test_result__"] = o.__class__.__name__
return result
return super().default(o)
def decode_test_result(d: dict[str, Any]) -> TestResult | dict[str, Any]:
"""Decode a TestResult (sub)class object from a JSON dict."""
if "__test_result__" not in d:
return d
cls_name = d.pop("__test_result__")
for cls in get_all_test_result_classes():
if cls.__name__ == cls_name:
return cls(**d)
def get_all_test_result_classes() -> set[type[TestResult]]:
prev_count = 0
classes = {TestResult}
while len(classes) > prev_count:
prev_count = len(classes)
to_add = []
for cls in classes:
to_add.extend(cls.__subclasses__())
classes.update(to_add)
return classes