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

   1 #!/usr/bin/env python2
   2 #
   3 # Parallel VM test case executor
   4 # Copyright (c) 2014-2015, Jouni Malinen <j@w1.fi>
   5 #
   6 # This software may be distributed under the terms of the BSD license.
   7 # See README for more details.
   8
   9 import curses
  10 import fcntl
  11 import logging
  12 import os
  13 import subprocess
  14 import sys
  15 import time
  16
  17 logger = logging.getLogger()
  18
  19 # Test cases that take significantly longer time to execute than average.
  20 long_tests = [ "ap_roam_open",
  21                "wpas_mesh_password_mismatch_retry",
  22                "wpas_mesh_password_mismatch",
  23                "hostapd_oom_wpa2_psk_connect",
  24                "ap_hs20_fetch_osu_stop",
  25                "ap_roam_wpa2_psk",
  26                "ibss_wpa_none_ccmp",
  27                "nfc_wps_er_handover_pk_hash_mismatch_sta",
  28                "go_neg_peers_force_diff_freq",
  29                "p2p_cli_invite",
  30                "sta_ap_scan_2b",
  31                "ap_pmf_sta_unprot_deauth_burst",
  32                "ap_bss_add_remove_during_ht_scan",
  33                "wext_scan_hidden",
  34                "autoscan_exponential",
  35                "nfc_p2p_client",
  36                "wnm_bss_keep_alive",
  37                "ap_inactivity_disconnect",
  38                "scan_bss_expiration_age",
  39                "autoscan_periodic",
  40                "discovery_group_client",
  41                "concurrent_p2pcli",
  42                "ap_bss_add_remove",
  43                "wpas_ap_wps",
  44                "wext_pmksa_cache",
  45                "ibss_wpa_none",
  46                "ap_ht_40mhz_intolerant_ap",
  47                "ibss_rsn",
  48                "discovery_pd_retries",
  49                "ap_wps_setup_locked_timeout",
  50                "ap_vht160",
  51                "dfs_radar",
  52                "dfs",
  53                "grpform_cred_ready_timeout",
  54                "hostapd_oom_wpa2_eap_connect",
  55                "wpas_ap_dfs",
  56                "autogo_many",
  57                "hostapd_oom_wpa2_eap",
  58                "ibss_open",
  59                "proxyarp_open_ebtables",
  60                "radius_failover",
  61                "obss_scan_40_intolerant",
  62                "dbus_connect_oom",
  63                "proxyarp_open",
  64                "ap_wps_iteration",
  65                "ap_wps_iteration_error",
  66                "ap_wps_pbc_timeout",
  67                "ap_wps_http_timeout",
  68                "p2p_go_move_reg_change",
  69                "p2p_go_move_active",
  70                "p2p_go_move_scm",
  71                "p2p_go_move_scm_peer_supports",
  72                "p2p_go_move_scm_peer_does_not_support",
  73                "p2p_go_move_scm_multi" ]
  74
  75 def get_failed(vm):
  76     failed = []
  77     for i in range(num_servers):
  78         failed += vm[i]['failed']
  79     return failed
  80
  81 def vm_read_stdout(vm, i):
  82     global total_started, total_passed, total_failed, total_skipped
  83     global rerun_failures
  84
  85     ready = False
  86     try:
  87         out = vm['proc'].stdout.read()
  88     except:
  89         return False
  90     logger.debug("VM[%d] stdout.read[%s]" % (i, out))
  91     pending = vm['pending'] + out
  92     lines = []
  93     while True:
  94         pos = pending.find('\n')
  95         if pos < 0:
  96             break
  97         line = pending[0:pos].rstrip()
  98         pending = pending[(pos + 1):]
  99         logger.debug("VM[%d] stdout full line[%s]" % (i, line))
 100         if line.startswith("READY"):
 101             ready = True
 102         elif line.startswith("PASS"):
 103             ready = True
 104             total_passed += 1
 105         elif line.startswith("FAIL"):
 106             ready = True
 107             total_failed += 1
 108             vals = line.split(' ')
 109             if len(vals) < 2:
 110                 logger.info("VM[%d] incomplete FAIL line: %s" % (i, line))
 111                 name = line
 112             else:
 113                 name = vals[1]
 114             logger.debug("VM[%d] test case failed: %s" % (i, name))
 115             vm['failed'].append(name)
 116         elif line.startswith("NOT-FOUND"):
 117             ready = True
 118             total_failed += 1
 119             logger.info("VM[%d] test case not found" % i)
 120         elif line.startswith("SKIP"):
 121             ready = True
 122             total_skipped += 1
 123         elif line.startswith("START"):
 124             total_started += 1
 125             if len(vm['failed']) == 0:
 126                 vals = line.split(' ')
 127                 if len(vals) >= 2:
 128                     vm['fail_seq'].append(vals[1])
 129         vm['out'] += line + '\n'
 130         lines.append(line)
 131     vm['pending'] = pending
 132     return ready
 133
 134 def show_progress(scr):
 135     global num_servers
 136     global vm
 137     global dir
 138     global timestamp
 139     global tests
 140     global first_run_failures
 141     global total_started, total_passed, total_failed, total_skipped
 142
 143     total_tests = len(tests)
 144     logger.info("Total tests: %d" % total_tests)
 145
 146     scr.leaveok(1)
 147     scr.addstr(0, 0, "Parallel test execution status", curses.A_BOLD)
 148     for i in range(0, num_servers):
 149         scr.addstr(i + 1, 0, "VM %d:" % (i + 1), curses.A_BOLD)
 150         scr.addstr(i + 1, 10, "starting VM")
 151     scr.addstr(num_servers + 1, 0, "Total:", curses.A_BOLD)
 152     scr.addstr(num_servers + 1, 20, "TOTAL={} STARTED=0 PASS=0 FAIL=0 SKIP=0".format(total_tests))
 153     scr.refresh()
 154
 155     completed_first_pass = False
 156     rerun_tests = []
 157
 158     while True:
 159         running = False
 160         first_running = False
 161         updated = False
 162
 163         for i in range(0, num_servers):
 164             if completed_first_pass:
 165                 continue
 166             if vm[i]['first_run_done']:
 167                 continue
 168             if not vm[i]['proc']:
 169                 continue
 170             if vm[i]['proc'].poll() is not None:
 171                 vm[i]['proc'] = None
 172                 scr.move(i + 1, 10)
 173                 scr.clrtoeol()
 174                 log = '{}/{}.srv.{}/console'.format(dir, timestamp, i + 1)
 175                 with open(log, 'r') as f:
 176                     if "Kernel panic" in f.read():
 177                         scr.addstr("kernel panic")
 178                         logger.info("VM[%d] kernel panic" % i)
 179                     else:
 180                         scr.addstr("unexpected exit")
 181                         logger.info("VM[%d] unexpected exit" % i)
 182                 updated = True
 183                 continue
 184
 185             running = True
 186             first_running = True
 187             try:
 188                 err = vm[i]['proc'].stderr.read()
 189                 vm[i]['err'] += err
 190                 logger.debug("VM[%d] stderr.read[%s]" % (i, err))
 191             except:
 192                 pass
 193
 194             if vm_read_stdout(vm[i], i):
 195                 scr.move(i + 1, 10)
 196                 scr.clrtoeol()
 197                 updated = True
 198                 if not tests:
 199                     vm[i]['first_run_done'] = True
 200                     scr.addstr("completed first round")
 201                     logger.info("VM[%d] completed first round" % i)
 202                     continue
 203                 else:
 204                     name = tests.pop(0)
 205                     vm[i]['proc'].stdin.write(name + '\n')
 206                     scr.addstr(name)
 207                     logger.debug("VM[%d] start test %s" % (i, name))
 208
 209         if not first_running and not completed_first_pass:
 210             logger.info("First round of testing completed")
 211             if tests:
 212                 logger.info("Unexpected test cases remaining from first round: " + str(tests))
 213                 raise Exception("Unexpected test cases remaining from first round")
 214             completed_first_pass = True
 215             for name in get_failed(vm):
 216                 if rerun_failures:
 217                     rerun_tests.append(name)
 218                 first_run_failures.append(name)
 219
 220         for i in range(num_servers):
 221             if not completed_first_pass:
 222                 continue
 223             if not vm[i]['proc']:
 224                 continue
 225             if vm[i]['proc'].poll() is not None:
 226                 vm[i]['proc'] = None
 227                 scr.move(i + 1, 10)
 228                 scr.clrtoeol()
 229                 log = '{}/{}.srv.{}/console'.format(dir, timestamp, i + 1)
 230                 with open(log, 'r') as f:
 231                     if "Kernel panic" in f.read():
 232                         scr.addstr("kernel panic")
 233                         logger.info("VM[%d] kernel panic" % i)
 234                     else:
 235                         scr.addstr("completed run")
 236                         logger.info("VM[%d] completed run" % i)
 237                 updated = True
 238                 continue
 239
 240             running = True
 241             try:
 242                 err = vm[i]['proc'].stderr.read()
 243                 vm[i]['err'] += err
 244                 logger.debug("VM[%d] stderr.read[%s]" % (i, err))
 245             except:
 246                 pass
 247
 248             ready = False
 249             if vm[i]['first_run_done']:
 250                 vm[i]['first_run_done'] = False
 251                 ready = True
 252             else:
 253                 ready = vm_read_stdout(vm[i], i)
 254             if ready:
 255                 scr.move(i + 1, 10)
 256                 scr.clrtoeol()
 257                 updated = True
 258                 if not rerun_tests:
 259                     vm[i]['proc'].stdin.write('\n')
 260                     scr.addstr("shutting down")
 261                     logger.info("VM[%d] shutting down" % i)
 262                 else:
 263                     name = rerun_tests.pop(0)
 264                     vm[i]['proc'].stdin.write(name + '\n')
 265                     scr.addstr(name + "(*)")
 266                     logger.debug("VM[%d] start test %s (*)" % (i, name))
 267
 268         if not running:
 269             break
 270
 271         if updated:
 272             scr.move(num_servers + 1, 10)
 273             scr.clrtoeol()
 274             scr.addstr("{} %".format(int(100.0 * (total_passed + total_failed + total_skipped) / total_tests)))
 275             scr.addstr(num_servers + 1, 20, "TOTAL={} STARTED={} PASS={} FAIL={} SKIP={}".format(total_tests, total_started, total_passed, total_failed, total_skipped))
 276             failed = get_failed(vm)
 277             if len(failed) > 0:
 278                 scr.move(num_servers + 2, 0)
 279                 scr.clrtoeol()
 280                 scr.addstr("Failed test cases: ")
 281                 count = 0
 282                 for f in failed:
 283                     count += 1
 284                     if count > 30:
 285                         scr.addstr('...')
 286                         scr.clrtoeol()
 287                         break
 288                     scr.addstr(f)
 289                     scr.addstr(' ')
 290
 291             scr.move(0, 35)
 292             scr.clrtoeol()
 293             if rerun_tests:
 294                 scr.addstr("(RETRY FAILED %d)" % len(rerun_tests))
 295             elif rerun_failures:
 296                 pass
 297             elif first_run_failures:
 298                 scr.addstr("(RETRY FAILED)")
 299
 300             scr.refresh()
 301
 302         time.sleep(0.25)
 303
 304     scr.refresh()
 305     time.sleep(0.3)
 306
 307 def main():
 308     import argparse
 309     import os
 310     global num_servers
 311     global vm
 312     global dir
 313     global timestamp
 314     global tests
 315     global first_run_failures
 316     global total_started, total_passed, total_failed, total_skipped
 317     global rerun_failures
 318
 319     total_started = 0
 320     total_passed = 0
 321     total_failed = 0
 322     total_skipped = 0
 323
 324     debug_level = logging.INFO
 325     rerun_failures = True
 326     timestamp = int(time.time())
 327
 328     scriptsdir = os.path.dirname(os.path.realpath(sys.argv[0]))
 329
 330     p = argparse.ArgumentParser(description='run multiple testing VMs in parallel')
 331     p.add_argument('num_servers', metavar='number of VMs', type=int, choices=range(1, 100),
 332                    help="number of VMs to start")
 333     p.add_argument('-f', dest='testmodules', metavar='<test module>',
 334                    help='execute only tests from these test modules',
 335                    type=str, nargs='+')
 336     p.add_argument('-1', dest='no_retry', action='store_const', const=True, default=False,
 337                    help="don't retry failed tests automatically")
 338     p.add_argument('--debug', dest='debug', action='store_const', const=True, default=False,
 339                    help="enable debug logging")
 340     p.add_argument('--codecov', dest='codecov', action='store_const', const=True, default=False,
 341                    help="enable code coverage collection")
 342     p.add_argument('--shuffle-tests', dest='shuffle', action='store_const', const=True, default=False,
 343                    help="shuffle test cases to randomize order")
 344     p.add_argument('--short', dest='short', action='store_const', const=True,
 345                    default=False,
 346                    help="only run short-duration test cases")
 347     p.add_argument('--long', dest='long', action='store_const', const=True,
 348                    default=False,
 349                    help="include long-duration test cases")
 350     p.add_argument('--valgrind', dest='valgrind', action='store_const',
 351                    const=True, default=False,
 352                    help="run tests under valgrind")
 353     p.add_argument('params', nargs='*')
 354     args = p.parse_args()
 355     num_servers = args.num_servers
 356     rerun_failures = not args.no_retry
 357     if args.debug:
 358         debug_level = logging.DEBUG
 359     extra_args = []
 360     if args.valgrind:
 361         extra_args += [ '--valgrind' ]
 362     if args.long:
 363         extra_args += [ '--long' ]
 364     if args.codecov:
 365         print "Code coverage - build separate binaries"
 366         logdir = "/tmp/hwsim-test-logs/" + str(timestamp)
 367         os.makedirs(logdir)
 368         subprocess.check_call([os.path.join(scriptsdir, 'build-codecov.sh'),
 369                                logdir])
 370         codecov_args = ['--codecov_dir', logdir]
 371         codecov = True
 372     else:
 373         codecov_args = []
 374         codecov = False
 375
 376     first_run_failures = []
 377     if args.params:
 378         tests = args.params
 379     else:
 380         tests = []
 381         cmd = [ os.path.join(os.path.dirname(scriptsdir), 'run-tests.py'),
 382                 '-L' ]
 383         if args.testmodules:
 384             cmd += [ "-f" ]
 385             cmd += args.testmodules
 386         lst = subprocess.Popen(cmd, stdout=subprocess.PIPE)
 387         for l in lst.stdout.readlines():
 388             name = l.split(' ')[0]
 389             tests.append(name)
 390     if len(tests) == 0:
 391         sys.exit("No test cases selected")
 392
 393     dir = '/tmp/hwsim-test-logs'
 394     try:
 395         os.mkdir(dir)
 396     except:
 397         pass
 398
 399     if args.shuffle:
 400         from random import shuffle
 401         shuffle(tests)
 402     elif num_servers > 2 and len(tests) > 100:
 403         # Move test cases with long duration to the beginning as an
 404         # optimization to avoid last part of the test execution running a long
 405         # duration test case on a single VM while all other VMs have already
 406         # completed their work.
 407         for l in long_tests:
 408             if l in tests:
 409                 tests.remove(l)
 410                 tests.insert(0, l)
 411     if args.short:
 412         tests = [t for t in tests if t not in long_tests]
 413
 414     logger.setLevel(debug_level)
 415     log_handler = logging.FileHandler('parallel-vm.log')
 416     log_handler.setLevel(debug_level)
 417     fmt = "%(asctime)s %(levelname)s %(message)s"
 418     log_formatter = logging.Formatter(fmt)
 419     log_handler.setFormatter(log_formatter)
 420     logger.addHandler(log_handler)
 421
 422     vm = {}
 423     for i in range(0, num_servers):
 424         print("\rStarting virtual machine {}/{}".format(i + 1, num_servers)),
 425         logger.info("Starting virtual machine {}/{}".format(i + 1, num_servers))
 426         cmd = [os.path.join(scriptsdir, 'vm-run.sh'), '--delay', str(i),
 427                '--timestamp', str(timestamp),
 428                '--ext', 'srv.%d' % (i + 1),
 429                '-i'] + codecov_args + extra_args
 430         vm[i] = {}
 431         vm[i]['first_run_done'] = False
 432         vm[i]['proc'] = subprocess.Popen(cmd,
 433                                          stdin=subprocess.PIPE,
 434                                          stdout=subprocess.PIPE,
 435                                          stderr=subprocess.PIPE)
 436         vm[i]['out'] = ""
 437         vm[i]['pending'] = ""
 438         vm[i]['err'] = ""
 439         vm[i]['failed'] = []
 440         vm[i]['fail_seq'] = []
 441         for stream in [ vm[i]['proc'].stdout, vm[i]['proc'].stderr ]:
 442             fd = stream.fileno()
 443             fl = fcntl.fcntl(fd, fcntl.F_GETFL)
 444             fcntl.fcntl(fd, fcntl.F_SETFL, fl | os.O_NONBLOCK)
 445     print
 446
 447     curses.wrapper(show_progress)
 448
 449     with open('{}/{}-parallel.log'.format(dir, timestamp), 'w') as f:
 450         for i in range(0, num_servers):
 451             f.write('VM {}\n{}\n{}\n'.format(i, vm[i]['out'], vm[i]['err']))
 452
 453     failed = get_failed(vm)
 454
 455     if first_run_failures:
 456         print "To re-run same failure sequence(s):"
 457         for i in range(0, num_servers):
 458             if len(vm[i]['failed']) == 0:
 459                 continue
 460             print "./parallel-vm.py -1 1",
 461             skip = len(vm[i]['fail_seq'])
 462             skip -= min(skip, 30)
 463             for t in vm[i]['fail_seq']:
 464                 if skip > 0:
 465                     skip -= 1
 466                     continue
 467                 print t,
 468             print
 469         print "Failed test cases:"
 470         for f in first_run_failures:
 471             print f,
 472             logger.info("Failed: " + f)
 473         print
 474     double_failed = []
 475     for name in failed:
 476         double_failed.append(name)
 477     for test in first_run_failures:
 478         double_failed.remove(test)
 479     if not rerun_failures:
 480         pass
 481     elif failed and not double_failed:
 482         print "All failed cases passed on retry"
 483         logger.info("All failed cases passed on retry")
 484     elif double_failed:
 485         print "Failed even on retry:"
 486         for f in double_failed:
 487             print f,
 488             logger.info("Failed on retry: " + f)
 489         print
 490     res = "TOTAL={} PASS={} FAIL={} SKIP={}".format(total_started,
 491                                                     total_passed,
 492                                                     total_failed,
 493                                                     total_skipped)
 494     print(res)
 495     logger.info(res)
 496     print "Logs: " + dir + '/' + str(timestamp)
 497     logger.info("Logs: " + dir + '/' + str(timestamp))
 498
 499     for i in range(0, num_servers):
 500         if len(vm[i]['pending']) > 0:
 501             logger.info("Unprocessed stdout from VM[%d]: '%s'" %
 502                         (i, vm[i]['pending']))
 503         log = '{}/{}.srv.{}/console'.format(dir, timestamp, i + 1)
 504         with open(log, 'r') as f:
 505             if "Kernel panic" in f.read():
 506                 print "Kernel panic in " + log
 507                 logger.info("Kernel panic in " + log)
 508
 509     if codecov:
 510         print "Code coverage - preparing report"
 511         for i in range(num_servers):
 512             subprocess.check_call([os.path.join(scriptsdir,
 513                                                 'process-codecov.sh'),
 514                                    logdir + ".srv.%d" % (i + 1),
 515                                    str(i)])
 516         subprocess.check_call([os.path.join(scriptsdir, 'combine-codecov.sh'),
 517                                logdir])
 518         print "file://%s/index.html" % logdir
 519         logger.info("Code coverage report: file://%s/index.html" % logdir)
 520
 521     if double_failed or (failed and not rerun_failures):
 522         logger.info("Test run complete - failures found")
 523         sys.exit(2)
 524     if failed:
 525         logger.info("Test run complete - failures found on first run; passed on retry")
 526         sys.exit(1)
 527     logger.info("Test run complete - no failures")
 528     sys.exit(0)
 529
 530 if __name__ == "__main__":
 531     main()