hostapd/tests/hwsim/vm/parallel-vm.py

533 lines
19 KiB
Python
Raw Normal View History

#!/usr/bin/env python2
#
# Parallel VM test case executor
# Copyright (c) 2014-2015, Jouni Malinen <j@w1.fi>
#
# This software may be distributed under the terms of the BSD license.
# See README for more details.
import curses
import fcntl
import logging
import os
import subprocess
import sys
import time
logger = logging.getLogger()
# Test cases that take significantly longer time to execute than average.
long_tests = [ "ap_roam_open",
"wpas_mesh_password_mismatch_retry",
"wpas_mesh_password_mismatch",
"hostapd_oom_wpa2_psk_connect",
"ap_hs20_fetch_osu_stop",
"ap_roam_wpa2_psk",
"ibss_wpa_none_ccmp",
"nfc_wps_er_handover_pk_hash_mismatch_sta",
"go_neg_peers_force_diff_freq",
"p2p_cli_invite",
"sta_ap_scan_2b",
"ap_pmf_sta_unprot_deauth_burst",
"ap_bss_add_remove_during_ht_scan",
"wext_scan_hidden",
"autoscan_exponential",
"nfc_p2p_client",
"wnm_bss_keep_alive",
"ap_inactivity_disconnect",
"scan_bss_expiration_age",
"autoscan_periodic",
"discovery_group_client",
"concurrent_p2pcli",
"ap_bss_add_remove",
"wpas_ap_wps",
"wext_pmksa_cache",
"ibss_wpa_none",
"ap_ht_40mhz_intolerant_ap",
"ibss_rsn",
"discovery_pd_retries",
"ap_wps_setup_locked_timeout",
"ap_vht160",
"dfs_radar",
"dfs",
"grpform_cred_ready_timeout",
"hostapd_oom_wpa2_eap_connect",
"wpas_ap_dfs",
"autogo_many",
"hostapd_oom_wpa2_eap",
"ibss_open",
"proxyarp_open_ebtables",
"radius_failover",
"obss_scan_40_intolerant",
"dbus_connect_oom",
"proxyarp_open",
"ap_wps_iteration",
"ap_wps_iteration_error",
"ap_wps_pbc_timeout",
"ap_wps_http_timeout",
"p2p_go_move_reg_change",
"p2p_go_move_active",
"p2p_go_move_scm",
"p2p_go_move_scm_peer_supports",
"p2p_go_move_scm_peer_does_not_support",
"p2p_go_move_scm_multi" ]
def get_failed(vm):
failed = []
for i in range(num_servers):
failed += vm[i]['failed']
return failed
def vm_read_stdout(vm, i):
global total_started, total_passed, total_failed, total_skipped
global rerun_failures
ready = False
try:
out = vm['proc'].stdout.read()
except:
return False
logger.debug("VM[%d] stdout.read[%s]" % (i, out))
pending = vm['pending'] + out
lines = []
while True:
pos = pending.find('\n')
if pos < 0:
break
line = pending[0:pos].rstrip()
pending = pending[(pos + 1):]
logger.debug("VM[%d] stdout full line[%s]" % (i, line))
if line.startswith("READY"):
ready = True
elif line.startswith("PASS"):
ready = True
total_passed += 1
elif line.startswith("FAIL"):
ready = True
total_failed += 1
vals = line.split(' ')
if len(vals) < 2:
logger.info("VM[%d] incomplete FAIL line: %s" % (i, line))
name = line
else:
name = vals[1]
logger.debug("VM[%d] test case failed: %s" % (i, name))
vm['failed'].append(name)
elif line.startswith("NOT-FOUND"):
ready = True
total_failed += 1
logger.info("VM[%d] test case not found" % i)
elif line.startswith("SKIP"):
ready = True
total_skipped += 1
elif line.startswith("START"):
total_started += 1
if len(vm['failed']) == 0:
vals = line.split(' ')
if len(vals) >= 2:
vm['fail_seq'].append(vals[1])
vm['out'] += line + '\n'
lines.append(line)
vm['pending'] = pending
return ready
def show_progress(scr):
global num_servers
global vm
global dir
global timestamp
global tests
global first_run_failures
global total_started, total_passed, total_failed, total_skipped
total_tests = len(tests)
logger.info("Total tests: %d" % total_tests)
scr.leaveok(1)
scr.addstr(0, 0, "Parallel test execution status", curses.A_BOLD)
for i in range(0, num_servers):
scr.addstr(i + 1, 0, "VM %d:" % (i + 1), curses.A_BOLD)
scr.addstr(i + 1, 10, "starting VM")
scr.addstr(num_servers + 1, 0, "Total:", curses.A_BOLD)
scr.addstr(num_servers + 1, 20, "TOTAL={} STARTED=0 PASS=0 FAIL=0 SKIP=0".format(total_tests))
scr.refresh()
completed_first_pass = False
rerun_tests = []
while True:
running = False
first_running = False
updated = False
for i in range(0, num_servers):
if completed_first_pass:
continue
if vm[i]['first_run_done']:
continue
if not vm[i]['proc']:
continue
if vm[i]['proc'].poll() is not None:
vm[i]['proc'] = None
scr.move(i + 1, 10)
scr.clrtoeol()
log = '{}/{}.srv.{}/console'.format(dir, timestamp, i + 1)
with open(log, 'r') as f:
if "Kernel panic" in f.read():
scr.addstr("kernel panic")
logger.info("VM[%d] kernel panic" % i)
else:
scr.addstr("unexpected exit")
logger.info("VM[%d] unexpected exit" % i)
updated = True
continue
running = True
first_running = True
try:
err = vm[i]['proc'].stderr.read()
vm[i]['err'] += err
logger.debug("VM[%d] stderr.read[%s]" % (i, err))
except:
pass
if vm_read_stdout(vm[i], i):
scr.move(i + 1, 10)
scr.clrtoeol()
updated = True
if not tests:
vm[i]['first_run_done'] = True
scr.addstr("completed first round")
logger.info("VM[%d] completed first round" % i)
continue
else:
name = tests.pop(0)
vm[i]['proc'].stdin.write(name + '\n')
scr.addstr(name)
logger.debug("VM[%d] start test %s" % (i, name))
if not first_running and not completed_first_pass:
logger.info("First round of testing completed")
if tests:
logger.info("Unexpected test cases remaining from first round: " + str(tests))
raise Exception("Unexpected test cases remaining from first round")
completed_first_pass = True
for name in get_failed(vm):
if rerun_failures:
rerun_tests.append(name)
first_run_failures.append(name)
for i in range(num_servers):
if not completed_first_pass:
continue
if not vm[i]['proc']:
continue
if vm[i]['proc'].poll() is not None:
vm[i]['proc'] = None
scr.move(i + 1, 10)
scr.clrtoeol()
log = '{}/{}.srv.{}/console'.format(dir, timestamp, i + 1)
with open(log, 'r') as f:
if "Kernel panic" in f.read():
scr.addstr("kernel panic")
logger.info("VM[%d] kernel panic" % i)
else:
scr.addstr("completed run")
logger.info("VM[%d] completed run" % i)
updated = True
continue
running = True
try:
err = vm[i]['proc'].stderr.read()
vm[i]['err'] += err
logger.debug("VM[%d] stderr.read[%s]" % (i, err))
except:
pass
ready = False
if vm[i]['first_run_done']:
vm[i]['first_run_done'] = False
ready = True
else:
ready = vm_read_stdout(vm[i], i)
if ready:
scr.move(i + 1, 10)
scr.clrtoeol()
updated = True
if not rerun_tests:
vm[i]['proc'].stdin.write('\n')
scr.addstr("shutting down")
logger.info("VM[%d] shutting down" % i)
else:
name = rerun_tests.pop(0)
vm[i]['proc'].stdin.write(name + '\n')
scr.addstr(name + "(*)")
logger.debug("VM[%d] start test %s (*)" % (i, name))
if not running:
break
if updated:
scr.move(num_servers + 1, 10)
scr.clrtoeol()
scr.addstr("{} %".format(int(100.0 * (total_passed + total_failed + total_skipped) / total_tests)))
scr.addstr(num_servers + 1, 20, "TOTAL={} STARTED={} PASS={} FAIL={} SKIP={}".format(total_tests, total_started, total_passed, total_failed, total_skipped))
failed = get_failed(vm)
if len(failed) > 0:
scr.move(num_servers + 2, 0)
scr.clrtoeol()
scr.addstr("Failed test cases: ")
count = 0
for f in failed:
count += 1
if count > 30:
scr.addstr('...')
scr.clrtoeol()
break
scr.addstr(f)
scr.addstr(' ')
scr.move(0, 35)
scr.clrtoeol()
if rerun_tests:
scr.addstr("(RETRY FAILED %d)" % len(rerun_tests))
elif rerun_failures:
pass
elif first_run_failures:
scr.addstr("(RETRY FAILED)")
scr.refresh()
time.sleep(0.25)
scr.refresh()
time.sleep(0.3)
def main():
import argparse
import os
global num_servers
global vm
global dir
global timestamp
global tests
global first_run_failures
global total_started, total_passed, total_failed, total_skipped
global rerun_failures
total_started = 0
total_passed = 0
total_failed = 0
total_skipped = 0
debug_level = logging.INFO
rerun_failures = True
timestamp = int(time.time())
scriptsdir = os.path.dirname(os.path.realpath(sys.argv[0]))
p = argparse.ArgumentParser(description='run multiple testing VMs in parallel')
p.add_argument('num_servers', metavar='number of VMs', type=int, choices=range(1, 100),
help="number of VMs to start")
p.add_argument('-f', dest='testmodules', metavar='<test module>',
help='execute only tests from these test modules',
type=str, nargs='+')
p.add_argument('-1', dest='no_retry', action='store_const', const=True, default=False,
help="don't retry failed tests automatically")
p.add_argument('--debug', dest='debug', action='store_const', const=True, default=False,
help="enable debug logging")
p.add_argument('--codecov', dest='codecov', action='store_const', const=True, default=False,
help="enable code coverage collection")
p.add_argument('--shuffle-tests', dest='shuffle', action='store_const', const=True, default=False,
help="shuffle test cases to randomize order")
p.add_argument('--short', dest='short', action='store_const', const=True,
default=False,
help="only run short-duration test cases")
p.add_argument('--long', dest='long', action='store_const', const=True,
default=False,
help="include long-duration test cases")
p.add_argument('--valgrind', dest='valgrind', action='store_const',
const=True, default=False,
help="run tests under valgrind")
p.add_argument('params', nargs='*')
args = p.parse_args()
dir = os.environ.get('HWSIM_TEST_LOG_DIR', '/tmp/hwsim-test-logs')
try:
os.makedirs(dir)
except:
pass
num_servers = args.num_servers
rerun_failures = not args.no_retry
if args.debug:
debug_level = logging.DEBUG
extra_args = []
if args.valgrind:
extra_args += [ '--valgrind' ]
if args.long:
extra_args += [ '--long' ]
if args.codecov:
print "Code coverage - build separate binaries"
logdir = os.path.join(dir, str(timestamp))
os.makedirs(logdir)
subprocess.check_call([os.path.join(scriptsdir, 'build-codecov.sh'),
logdir])
codecov_args = ['--codecov_dir', logdir]
codecov = True
else:
codecov_args = []
codecov = False
first_run_failures = []
if args.params:
tests = args.params
else:
tests = []
cmd = [ os.path.join(os.path.dirname(scriptsdir), 'run-tests.py'),
'-L' ]
if args.testmodules:
cmd += [ "-f" ]
cmd += args.testmodules
lst = subprocess.Popen(cmd, stdout=subprocess.PIPE)
for l in lst.stdout.readlines():
name = l.split(' ')[0]
tests.append(name)
if len(tests) == 0:
sys.exit("No test cases selected")
if args.shuffle:
from random import shuffle
shuffle(tests)
elif num_servers > 2 and len(tests) > 100:
# Move test cases with long duration to the beginning as an
# optimization to avoid last part of the test execution running a long
# duration test case on a single VM while all other VMs have already
# completed their work.
for l in long_tests:
if l in tests:
tests.remove(l)
tests.insert(0, l)
if args.short:
tests = [t for t in tests if t not in long_tests]
logger.setLevel(debug_level)
log_handler = logging.FileHandler('parallel-vm.log')
log_handler.setLevel(debug_level)
fmt = "%(asctime)s %(levelname)s %(message)s"
log_formatter = logging.Formatter(fmt)
log_handler.setFormatter(log_formatter)
logger.addHandler(log_handler)
vm = {}
for i in range(0, num_servers):
print("\rStarting virtual machine {}/{}".format(i + 1, num_servers)),
logger.info("Starting virtual machine {}/{}".format(i + 1, num_servers))
cmd = [os.path.join(scriptsdir, 'vm-run.sh'), '--delay', str(i),
'--timestamp', str(timestamp),
'--ext', 'srv.%d' % (i + 1),
'-i'] + codecov_args + extra_args
vm[i] = {}
vm[i]['first_run_done'] = False
vm[i]['proc'] = subprocess.Popen(cmd,
stdin=subprocess.PIPE,
stdout=subprocess.PIPE,
stderr=subprocess.PIPE)
vm[i]['out'] = ""
vm[i]['pending'] = ""
vm[i]['err'] = ""
vm[i]['failed'] = []
vm[i]['fail_seq'] = []
for stream in [ vm[i]['proc'].stdout, vm[i]['proc'].stderr ]:
fd = stream.fileno()
fl = fcntl.fcntl(fd, fcntl.F_GETFL)
fcntl.fcntl(fd, fcntl.F_SETFL, fl | os.O_NONBLOCK)
print
curses.wrapper(show_progress)
with open('{}/{}-parallel.log'.format(dir, timestamp), 'w') as f:
for i in range(0, num_servers):
f.write('VM {}\n{}\n{}\n'.format(i, vm[i]['out'], vm[i]['err']))
failed = get_failed(vm)
if first_run_failures:
print "To re-run same failure sequence(s):"
for i in range(0, num_servers):
if len(vm[i]['failed']) == 0:
continue
print "./parallel-vm.py -1 1",
skip = len(vm[i]['fail_seq'])
skip -= min(skip, 30)
for t in vm[i]['fail_seq']:
if skip > 0:
skip -= 1
continue
print t,
print
print "Failed test cases:"
for f in first_run_failures:
print f,
logger.info("Failed: " + f)
print
double_failed = []
for name in failed:
double_failed.append(name)
for test in first_run_failures:
double_failed.remove(test)
if not rerun_failures:
pass
elif failed and not double_failed:
print "All failed cases passed on retry"
logger.info("All failed cases passed on retry")
elif double_failed:
print "Failed even on retry:"
for f in double_failed:
print f,
logger.info("Failed on retry: " + f)
print
res = "TOTAL={} PASS={} FAIL={} SKIP={}".format(total_started,
total_passed,
total_failed,
total_skipped)
print(res)
logger.info(res)
print "Logs: " + dir + '/' + str(timestamp)
logger.info("Logs: " + dir + '/' + str(timestamp))
for i in range(0, num_servers):
if len(vm[i]['pending']) > 0:
logger.info("Unprocessed stdout from VM[%d]: '%s'" %
(i, vm[i]['pending']))
log = '{}/{}.srv.{}/console'.format(dir, timestamp, i + 1)
with open(log, 'r') as f:
if "Kernel panic" in f.read():
print "Kernel panic in " + log
logger.info("Kernel panic in " + log)
if codecov:
print "Code coverage - preparing report"
for i in range(num_servers):
subprocess.check_call([os.path.join(scriptsdir,
'process-codecov.sh'),
logdir + ".srv.%d" % (i + 1),
str(i)])
subprocess.check_call([os.path.join(scriptsdir, 'combine-codecov.sh'),
logdir])
print "file://%s/index.html" % logdir
logger.info("Code coverage report: file://%s/index.html" % logdir)
if double_failed or (failed and not rerun_failures):
logger.info("Test run complete - failures found")
sys.exit(2)
if failed:
logger.info("Test run complete - failures found on first run; passed on retry")
sys.exit(1)
logger.info("Test run complete - no failures")
sys.exit(0)
if __name__ == "__main__":
main()