Fang Deng | 5d518f4 | 2013-08-02 14:04:32 -0700 | [diff] [blame] | 1 | # Copyright (c) 2013 The Chromium OS Authors. All rights reserved. |
| 2 | # Use of this source code is governed by a BSD-style license that can be |
| 3 | # found in the LICENSE file. |
| 4 | # |
| 5 | # Expects to be run in an environment with sudo and no interactive password |
| 6 | # prompt, such as within the Chromium OS development chroot. |
| 7 | |
| 8 | |
| 9 | """This file provides core logic for servo verify/repair process.""" |
| 10 | |
| 11 | |
Fang Deng | 5d518f4 | 2013-08-02 14:04:32 -0700 | [diff] [blame] | 12 | import logging |
Raul E Rangel | 52ca2e8 | 2018-07-03 14:10:14 -0600 | [diff] [blame] | 13 | import os |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 14 | import re |
| 15 | import tarfile |
Garry Wang | c1288cf | 2019-12-17 14:58:00 -0800 | [diff] [blame] | 16 | import time |
Gregory Nisbet | 265a52c | 2019-12-10 20:38:42 -0800 | [diff] [blame] | 17 | import traceback |
Dana Goyette | 4dc0adc | 2019-05-06 14:51:53 -0700 | [diff] [blame] | 18 | import xmlrpclib |
Otabek Kasimov | 120b6fa | 2020-07-03 00:15:27 -0700 | [diff] [blame] | 19 | import calendar |
Fang Deng | 5d518f4 | 2013-08-02 14:04:32 -0700 | [diff] [blame] | 20 | |
| 21 | from autotest_lib.client.bin import utils |
Garry Wang | 79e9af6 | 2019-06-12 15:19:19 -0700 | [diff] [blame] | 22 | from autotest_lib.client.common_lib import error |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 23 | from autotest_lib.client.common_lib import hosts |
Garry Wang | 7b0e1b7 | 2020-03-25 19:08:59 -0700 | [diff] [blame] | 24 | from autotest_lib.client.common_lib import lsbrelease_utils |
Fang Deng | 5d518f4 | 2013-08-02 14:04:32 -0700 | [diff] [blame] | 25 | from autotest_lib.client.common_lib.cros import retry |
Christopher Wiley | cef1f90 | 2014-06-19 11:11:23 -0700 | [diff] [blame] | 26 | from autotest_lib.client.common_lib.cros.network import ping_runner |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 27 | from autotest_lib.server.cros.servo import servo |
Richard Barnette | d31580e | 2018-05-14 19:58:00 +0000 | [diff] [blame] | 28 | from autotest_lib.server.hosts import servo_repair |
Garry Wang | ebc015b | 2019-06-06 17:45:06 -0700 | [diff] [blame] | 29 | from autotest_lib.server.hosts import base_servohost |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 30 | from autotest_lib.server.hosts import servo_constants |
Otabek Kasimov | 4ea636e | 2020-04-14 23:35:06 -0700 | [diff] [blame] | 31 | from autotest_lib.server.cros.faft.utils import config |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 32 | from autotest_lib.client.common_lib import global_config |
Otabek Kasimov | 8475cce | 2020-07-14 12:11:31 -0700 | [diff] [blame] | 33 | from autotest_lib.site_utils.admin_audit import servo_updater |
Garry Wang | d736748 | 2020-02-27 13:52:40 -0800 | [diff] [blame] | 34 | |
Otabek Kasimov | 1596349 | 2020-06-23 21:10:51 -0700 | [diff] [blame] | 35 | try: |
| 36 | from chromite.lib import metrics |
| 37 | except ImportError: |
| 38 | metrics = utils.metrics_mock |
| 39 | |
Dan Shi | 3b2adf6 | 2015-09-02 17:46:54 -0700 | [diff] [blame] | 40 | _CONFIG = global_config.global_config |
Fang Deng | 5d518f4 | 2013-08-02 14:04:32 -0700 | [diff] [blame] | 41 | |
Otabek Kasimov | a7ba91a | 2020-03-09 08:31:01 -0700 | [diff] [blame] | 42 | |
Garry Wang | ebc015b | 2019-06-06 17:45:06 -0700 | [diff] [blame] | 43 | class ServoHost(base_servohost.BaseServoHost): |
| 44 | """Host class for a servo host(e.g. beaglebone, labstation) |
Dana Goyette | 0b6e640 | 2019-10-04 11:09:24 -0700 | [diff] [blame] | 45 | that with a servo instance for a specific port. |
| 46 | |
| 47 | @type _servo: servo.Servo | None |
| 48 | """ |
Fang Deng | 5d518f4 | 2013-08-02 14:04:32 -0700 | [diff] [blame] | 49 | |
Raul E Rangel | 52ca2e8 | 2018-07-03 14:10:14 -0600 | [diff] [blame] | 50 | DEFAULT_PORT = int(os.getenv('SERVOD_PORT', '9999')) |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 51 | |
Dan Shi | e5b3c51 | 2014-08-21 12:12:09 -0700 | [diff] [blame] | 52 | # Timeout for initializing servo signals. |
Wai-Hong Tam | 37b6ed3 | 2017-09-19 15:52:39 -0700 | [diff] [blame] | 53 | INITIALIZE_SERVO_TIMEOUT_SECS = 60 |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 54 | |
xixuan | 6cf6d2f | 2016-01-29 15:29:00 -0800 | [diff] [blame] | 55 | # Ready test function |
| 56 | SERVO_READY_METHOD = 'get_version' |
Fang Deng | 5d518f4 | 2013-08-02 14:04:32 -0700 | [diff] [blame] | 57 | |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 58 | # Directory prefix on the servo host where the servod logs are stored. |
| 59 | SERVOD_LOG_PREFIX = '/var/log/servod' |
| 60 | |
| 61 | # Exit code to use when symlinks for servod logs are not found. |
| 62 | NO_SYMLINKS_CODE = 9 |
| 63 | |
| 64 | # Directory in the job's results directory to dump the logs into. |
| 65 | LOG_DIR = 'servod' |
| 66 | |
| 67 | # Prefix for joint loglevel files in the logs. |
| 68 | JOINT_LOG_PREFIX = 'log' |
| 69 | |
| 70 | # Regex group to extract timestamp from logfile name. |
| 71 | TS_GROUP = 'ts' |
| 72 | |
| 73 | # This regex is used to extract the timestamp from servod logs. |
| 74 | # files always start with log. |
| 75 | TS_RE = (r'log.' |
| 76 | # The timestamp is of format %Y-%m-%d--%H-%M-%S.MS |
| 77 | r'(?P<%s>\d{4}(\-\d{2}){2}\-(-\d{2}){3}.\d{3})' |
| 78 | # The loglevel is optional depending on labstation version. |
| 79 | r'(.(INFO|DEBUG|WARNING))?' % TS_GROUP) |
| 80 | TS_EXTRACTOR = re.compile(TS_RE) |
| 81 | |
| 82 | # Regex group to extract MCU name from logline in servod logs. |
| 83 | MCU_GROUP = 'mcu' |
| 84 | |
| 85 | # Regex group to extract logline from MCU logline in servod logs. |
| 86 | LINE_GROUP = 'line' |
| 87 | |
| 88 | # This regex is used to extract the mcu and the line content from an |
| 89 | # MCU logline in servod logs. e.g. EC or servo_v4 console logs. |
| 90 | # Here is an example log-line: |
| 91 | # |
| 92 | # 2020-01-23 13:15:12,223 - servo_v4 - EC3PO.Console - DEBUG - |
| 93 | # console.py:219:LogConsoleOutput - /dev/pts/9 - cc polarity: cc1 |
| 94 | # |
| 95 | # Here is conceptually how they are formatted: |
| 96 | # |
| 97 | # <time> - <MCU> - EC3PO.Console - <LVL> - <file:line:func> - <pts> - |
| 98 | # <output> |
| 99 | # |
| 100 | # The log format starts with a timestamp |
| 101 | MCU_RE = (r'[\d\-]+ [\d:,]+ ' |
| 102 | # The mcu that is logging this is next. |
| 103 | r'- (?P<%s>\w+) - ' |
| 104 | # Next, we have more log outputs before the actual line. |
| 105 | # Information about the file line, logging function etc. |
| 106 | # Anchor on EC3PO Console, LogConsoleOutput and dev/pts. |
| 107 | # NOTE: if the log format changes, this regex needs to be |
| 108 | # adjusted. |
| 109 | r'EC3PO\.Console[\s\-\w\d:.]+LogConsoleOutput - /dev/pts/\d+ - ' |
| 110 | # Lastly, we get the MCU's console line. |
| 111 | r'(?P<%s>.+$)' % (MCU_GROUP, LINE_GROUP)) |
| 112 | MCU_EXTRACTOR = re.compile(MCU_RE) |
| 113 | |
| 114 | # Suffix to identify compressed logfiles. |
| 115 | COMPRESSION_SUFFIX = '.tbz2' |
| 116 | |
Ruben Rodriguez Buchillon | 5bac306 | 2020-03-25 21:32:58 -0700 | [diff] [blame] | 117 | # A suffix to mark servod log directories that came from instance that |
| 118 | # ran during this servo_host, but are not the last one running e.g. when |
| 119 | # an instance (on purpose, or due to a bug) restarted in the middle of the |
| 120 | # run. |
| 121 | OLD_LOG_SUFFIX = 'old' |
| 122 | |
Otabek Kasimov | c6f3041 | 2020-06-30 20:08:12 -0700 | [diff] [blame] | 123 | # Mapping servo board with their vid-pid |
| 124 | SERVO_VID_PID = { |
| 125 | 'servo_v4':'18d1:501b', |
| 126 | 'ccd_cr50':'18d1:5014', |
| 127 | 'servo_micro':'18d1:501a', |
| 128 | 'servo_v3':['18d1:5004', '0403:6014'], |
| 129 | } |
| 130 | |
Otabek Kasimov | 1596349 | 2020-06-23 21:10:51 -0700 | [diff] [blame] | 131 | # States of verifiers |
| 132 | # True - verifier run and passed |
| 133 | # False - verifier run and failed |
| 134 | # None - verifier did not run or dependency failed |
| 135 | VERIFY_SUCCESS = True |
| 136 | VERIFY_FAILED = False |
| 137 | VERIFY_NOT_RUN = None |
| 138 | |
Otabek Kasimov | cc9738e | 2020-02-14 16:17:15 -0800 | [diff] [blame] | 139 | def _init_attributes(self): |
| 140 | self._servo_state = None |
| 141 | self.servo_port = None |
| 142 | self.servo_board = None |
| 143 | self.servo_model = None |
| 144 | self.servo_serial = None |
Garry Wang | 000c6c0 | 2020-05-11 21:27:23 -0700 | [diff] [blame] | 145 | # The flag that indicate if a servo is connected to a smart usbhub. |
| 146 | # TODO(xianuowang@) remove this flag once all usbhubs in the lab |
| 147 | # get replaced. |
| 148 | self.smart_usbhub = None |
Otabek Kasimov | cc9738e | 2020-02-14 16:17:15 -0800 | [diff] [blame] | 149 | self._servo = None |
| 150 | self._servod_server_proxy = None |
Ruben Rodriguez Buchillon | 5bac306 | 2020-03-25 21:32:58 -0700 | [diff] [blame] | 151 | self._initial_instance_ts = None |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 152 | # Flag to make sure that multiple calls to close do not result in the |
| 153 | # logic executing multiple times. |
| 154 | self._closed = False |
Fang Deng | 5d518f4 | 2013-08-02 14:04:32 -0700 | [diff] [blame] | 155 | |
Richard Barnette | 17bfc6c | 2016-08-04 18:41:43 -0700 | [diff] [blame] | 156 | def _initialize(self, servo_host='localhost', |
Richard Barnette | e519dcd | 2016-08-15 17:37:17 -0700 | [diff] [blame] | 157 | servo_port=DEFAULT_PORT, servo_board=None, |
Nick Sanders | 2f3c985 | 2018-10-24 12:10:24 -0700 | [diff] [blame] | 158 | servo_model=None, servo_serial=None, is_in_lab=None, |
| 159 | *args, **dargs): |
Fang Deng | 5d518f4 | 2013-08-02 14:04:32 -0700 | [diff] [blame] | 160 | """Initialize a ServoHost instance. |
| 161 | |
| 162 | A ServoHost instance represents a host that controls a servo. |
| 163 | |
| 164 | @param servo_host: Name of the host where the servod process |
| 165 | is running. |
Raul E Rangel | 52ca2e8 | 2018-07-03 14:10:14 -0600 | [diff] [blame] | 166 | @param servo_port: Port the servod process is listening on. Defaults |
| 167 | to the SERVOD_PORT environment variable if set, |
| 168 | otherwise 9999. |
Kevin Cheng | 5f2ba6c | 2016-09-28 10:20:05 -0700 | [diff] [blame] | 169 | @param servo_board: Board that the servo is connected to. |
Nick Sanders | 2f3c985 | 2018-10-24 12:10:24 -0700 | [diff] [blame] | 170 | @param servo_model: Model that the servo is connected to. |
Dan Shi | 4d47852 | 2014-02-14 13:46:32 -0800 | [diff] [blame] | 171 | @param is_in_lab: True if the servo host is in Cros Lab. Default is set |
| 172 | to None, for which utils.host_is_in_lab_zone will be |
| 173 | called to check if the servo host is in Cros lab. |
Fang Deng | 5d518f4 | 2013-08-02 14:04:32 -0700 | [diff] [blame] | 174 | |
| 175 | """ |
| 176 | super(ServoHost, self)._initialize(hostname=servo_host, |
Garry Wang | ebc015b | 2019-06-06 17:45:06 -0700 | [diff] [blame] | 177 | is_in_lab=is_in_lab, *args, **dargs) |
Otabek Kasimov | cc9738e | 2020-02-14 16:17:15 -0800 | [diff] [blame] | 178 | self._init_attributes() |
Richard Barnette | 42f4db9 | 2018-08-23 15:05:15 -0700 | [diff] [blame] | 179 | self.servo_port = int(servo_port) |
Richard Barnette | e519dcd | 2016-08-15 17:37:17 -0700 | [diff] [blame] | 180 | self.servo_board = servo_board |
Nick Sanders | 2f3c985 | 2018-10-24 12:10:24 -0700 | [diff] [blame] | 181 | self.servo_model = servo_model |
Kevin Cheng | 643ce8a | 2016-09-15 15:42:12 -0700 | [diff] [blame] | 182 | self.servo_serial = servo_serial |
Wai-Hong Tam | 3a8a255 | 2019-11-19 14:28:04 +0800 | [diff] [blame] | 183 | |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 184 | # The location of the log files on the servo host for this instance. |
| 185 | self.remote_log_dir = '%s_%s' % (self.SERVOD_LOG_PREFIX, |
| 186 | self.servo_port) |
Garry Wang | 79e9af6 | 2019-06-12 15:19:19 -0700 | [diff] [blame] | 187 | # Path of the servo host lock file. |
| 188 | self._lock_file = (self.TEMP_FILE_DIR + str(self.servo_port) |
| 189 | + self.LOCK_FILE_POSTFIX) |
| 190 | # File path to declare a reboot request. |
| 191 | self._reboot_file = (self.TEMP_FILE_DIR + str(self.servo_port) |
| 192 | + self.REBOOT_FILE_POSTFIX) |
| 193 | |
| 194 | # Lock the servo host if it's an in-lab labstation to prevent other |
| 195 | # task to reboot it until current task completes. We also wait and |
| 196 | # make sure the labstation is up here, in the case of the labstation is |
| 197 | # in the middle of reboot. |
Garry Wang | 7c00b0f | 2019-06-25 17:28:17 -0700 | [diff] [blame] | 198 | self._is_locked = False |
Garry Wang | 42b4d86 | 2019-06-25 15:50:49 -0700 | [diff] [blame] | 199 | if (self.wait_up(self.REBOOT_TIMEOUT) and self.is_in_lab() |
| 200 | and self.is_labstation()): |
Garry Wang | 79e9af6 | 2019-06-12 15:19:19 -0700 | [diff] [blame] | 201 | self._lock() |
Garry Wang | ebc015b | 2019-06-06 17:45:06 -0700 | [diff] [blame] | 202 | |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 203 | self._repair_strategy = ( |
| 204 | servo_repair.create_servo_repair_strategy()) |
Richard Barnette | e519dcd | 2016-08-15 17:37:17 -0700 | [diff] [blame] | 205 | |
Dana Goyette | afa62fd | 2020-03-16 13:45:27 -0700 | [diff] [blame] | 206 | def __str__(self): |
| 207 | return "<%s '%s:%s'>" % ( |
| 208 | type(self).__name__, self.hostname, self.servo_port) |
| 209 | |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 210 | def connect_servo(self): |
Garry Wang | 8c8dc97 | 2020-06-09 13:41:51 -0700 | [diff] [blame] | 211 | """ Initialize and setup servo for later use. |
| 212 | """ |
| 213 | self.initilize_servo() |
| 214 | self.initialize_dut_for_servo() |
| 215 | |
| 216 | |
| 217 | def initilize_servo(self): |
Kevin Cheng | 5f2ba6c | 2016-09-28 10:20:05 -0700 | [diff] [blame] | 218 | """Establish a connection to the servod server on this host. |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 219 | |
| 220 | Initializes `self._servo` and then verifies that all network |
| 221 | connections are working. This will create an ssh tunnel if |
| 222 | it's required. |
Garry Wang | 8c8dc97 | 2020-06-09 13:41:51 -0700 | [diff] [blame] | 223 | """ |
| 224 | self._servo = servo.Servo(servo_host=self, |
| 225 | servo_serial=self.servo_serial) |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 226 | |
Garry Wang | 8c8dc97 | 2020-06-09 13:41:51 -0700 | [diff] [blame] | 227 | |
| 228 | def initialize_dut_for_servo(self): |
| 229 | """This method will do some setup for dut control, e.g. setup |
| 230 | main servo_v4 device, and also testing the connection between servo |
| 231 | and DUT. As a side effect of testing the connection, all signals on |
| 232 | the target servo are reset to default values, and the USB stick is |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 233 | set to the neutral (off) position. |
| 234 | """ |
Garry Wang | 8c8dc97 | 2020-06-09 13:41:51 -0700 | [diff] [blame] | 235 | if not self._servo: |
| 236 | raise hosts.AutoservVerifyError('Servo object needs to be' |
| 237 | ' initialized before initialize' |
| 238 | ' DUT.') |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 239 | timeout, _ = retry.timeout( |
Garry Wang | 8c8dc97 | 2020-06-09 13:41:51 -0700 | [diff] [blame] | 240 | self._servo.initialize_dut, |
| 241 | timeout_sec=self.INITIALIZE_SERVO_TIMEOUT_SECS) |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 242 | if timeout: |
Garry Wang | 8c8dc97 | 2020-06-09 13:41:51 -0700 | [diff] [blame] | 243 | raise hosts.AutoservVerifyError('Initialize dut for servo timed' |
| 244 | ' out.') |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 245 | |
| 246 | |
| 247 | def disconnect_servo(self): |
Kevin Cheng | 5f2ba6c | 2016-09-28 10:20:05 -0700 | [diff] [blame] | 248 | """Disconnect our servo if it exists. |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 249 | |
| 250 | If we've previously successfully connected to our servo, |
| 251 | disconnect any established ssh tunnel, and set `self._servo` |
| 252 | back to `None`. |
| 253 | """ |
| 254 | if self._servo: |
| 255 | # N.B. This call is safe even without a tunnel: |
| 256 | # rpc_server_tracker.disconnect() silently ignores |
| 257 | # unknown ports. |
| 258 | self.rpc_server_tracker.disconnect(self.servo_port) |
| 259 | self._servo = None |
Fang Deng | 5d518f4 | 2013-08-02 14:04:32 -0700 | [diff] [blame] | 260 | |
Garry Wang | c1288cf | 2019-12-17 14:58:00 -0800 | [diff] [blame] | 261 | |
Wai-Hong Tam | 3a8a255 | 2019-11-19 14:28:04 +0800 | [diff] [blame] | 262 | def _create_servod_server_proxy(self): |
| 263 | """Create a proxy that can be used to communicate with servod server. |
Fang Deng | 5d518f4 | 2013-08-02 14:04:32 -0700 | [diff] [blame] | 264 | |
| 265 | @returns: An xmlrpclib.ServerProxy that is connected to the servod |
| 266 | server on the host. |
Fang Deng | 5d518f4 | 2013-08-02 14:04:32 -0700 | [diff] [blame] | 267 | """ |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 268 | if (servo_constants.ENABLE_SSH_TUNNEL_FOR_SERVO |
| 269 | and not self.is_localhost()): |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 270 | return self.rpc_server_tracker.xmlrpc_connect( |
| 271 | None, self.servo_port, |
| 272 | ready_test_name=self.SERVO_READY_METHOD, |
Allen Li | 2b1a899 | 2018-11-27 14:17:18 -0800 | [diff] [blame] | 273 | timeout_seconds=60, |
Dana Goyette | afa62fd | 2020-03-16 13:45:27 -0700 | [diff] [blame] | 274 | request_timeout_seconds=3600, |
| 275 | server_desc=str(self)) |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 276 | else: |
| 277 | remote = 'http://%s:%s' % (self.hostname, self.servo_port) |
| 278 | return xmlrpclib.ServerProxy(remote) |
Fang Deng | 5d518f4 | 2013-08-02 14:04:32 -0700 | [diff] [blame] | 279 | |
| 280 | |
Wai-Hong Tam | 3a8a255 | 2019-11-19 14:28:04 +0800 | [diff] [blame] | 281 | def get_servod_server_proxy(self): |
| 282 | """Return a cached proxy if exists; otherwise, create a new one. |
| 283 | |
| 284 | @returns: An xmlrpclib.ServerProxy that is connected to the servod |
| 285 | server on the host. |
| 286 | """ |
| 287 | # Single-threaded execution, no race |
| 288 | if self._servod_server_proxy is None: |
| 289 | self._servod_server_proxy = self._create_servod_server_proxy() |
| 290 | return self._servod_server_proxy |
| 291 | |
| 292 | |
Richard Barnette | 1edbb16 | 2016-11-01 11:47:50 -0700 | [diff] [blame] | 293 | def verify(self, silent=False): |
| 294 | """Update the servo host and verify it's in a good state. |
| 295 | |
| 296 | @param silent If true, suppress logging in `status.log`. |
| 297 | """ |
Richard Barnette | abbdc25 | 2018-07-26 16:57:42 -0700 | [diff] [blame] | 298 | message = 'Beginning verify for servo host %s port %s serial %s' |
| 299 | message %= (self.hostname, self.servo_port, self.servo_serial) |
| 300 | self.record('INFO', None, None, message) |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 301 | try: |
Richard Barnette | 1edbb16 | 2016-11-01 11:47:50 -0700 | [diff] [blame] | 302 | self._repair_strategy.verify(self, silent) |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 303 | self._servo_state = servo_constants.SERVO_STATE_WORKING |
| 304 | self.record('INFO', None, None, |
| 305 | 'ServoHost verify set servo_state as WORKING') |
Garry Wang | 63b8c38 | 2020-03-11 22:28:40 -0700 | [diff] [blame] | 306 | except Exception as e: |
Otabek Kasimov | 120b6fa | 2020-07-03 00:15:27 -0700 | [diff] [blame] | 307 | if not self.is_localhost(): |
Otabek Kasimov | c6f3041 | 2020-06-30 20:08:12 -0700 | [diff] [blame] | 308 | self._servo_state = self.determine_servo_state() |
| 309 | self.record('INFO', None, None, |
| 310 | 'ServoHost verify set servo_state as %s' |
| 311 | % self._servo_state) |
Garry Wang | 63b8c38 | 2020-03-11 22:28:40 -0700 | [diff] [blame] | 312 | if self._is_critical_error(e): |
| 313 | raise |
Fang Deng | 5d518f4 | 2013-08-02 14:04:32 -0700 | [diff] [blame] | 314 | |
| 315 | |
Garry Wang | 7b0e1b7 | 2020-03-25 19:08:59 -0700 | [diff] [blame] | 316 | def get_image_name_from_usbkey(self, usbkey_dev): |
| 317 | """Mount usb drive and check ChromeOS image name on it if there is |
| 318 | one. This method assumes the image_usbkey_direction is already set |
| 319 | to servo side. |
| 320 | |
| 321 | @param: usbkey_dev usbkey dev path(e.g. /dev/sdb). |
| 322 | |
| 323 | @returns: image_name on the usbkey, e.g. nami-release/R82.10138.0.0, |
| 324 | or empty string if no test image detected, or unexpected |
| 325 | error occurred. |
Garry Wang | 7b0e1b7 | 2020-03-25 19:08:59 -0700 | [diff] [blame] | 326 | """ |
Garry Wang | 70e5d06 | 2020-04-03 18:01:05 -0700 | [diff] [blame] | 327 | logging.info('Checking ChromeOS image name on usbkey.') |
Garry Wang | 7b0e1b7 | 2020-03-25 19:08:59 -0700 | [diff] [blame] | 328 | usb_mount_path = '/media/servo_usb/%s' % self.servo_port |
| 329 | unmount_cmd = 'umount %s' % usb_mount_path |
| 330 | # ChromeOS root fs is in /dev/sdx3 |
| 331 | mount_cmd = 'mount -o ro %s3 %s' % (usbkey_dev, usb_mount_path) |
| 332 | # Unmount if there is an existing stale mount. |
| 333 | self.run(unmount_cmd, ignore_status=True) |
| 334 | # Create if the mount point is not existing. |
| 335 | self.run('mkdir -p %s' % usb_mount_path) |
| 336 | try: |
| 337 | # Attempt to mount the usb drive. |
| 338 | mount_result = self.run(mount_cmd, ignore_status=True) |
Garry Wang | 70e5d06 | 2020-04-03 18:01:05 -0700 | [diff] [blame] | 339 | if mount_result.exit_status != 0: |
| 340 | logging.error('Unexpected error occurred on mount usb drive.') |
Garry Wang | 7b0e1b7 | 2020-03-25 19:08:59 -0700 | [diff] [blame] | 341 | return '' |
Garry Wang | 70e5d06 | 2020-04-03 18:01:05 -0700 | [diff] [blame] | 342 | |
| 343 | release_content = self.run( |
| 344 | 'cat %s/etc/lsb-release' % usb_mount_path, |
| 345 | ignore_status=True).stdout.strip() |
| 346 | |
| 347 | if not re.search(r'RELEASE_TRACK=.*test', release_content): |
| 348 | logging.info('The image on usbkey is not a test image') |
| 349 | return '' |
| 350 | |
| 351 | return lsbrelease_utils.get_chromeos_release_builder_path( |
| 352 | lsb_release_content=release_content) |
Garry Wang | 7b0e1b7 | 2020-03-25 19:08:59 -0700 | [diff] [blame] | 353 | finally: |
Garry Wang | 70e5d06 | 2020-04-03 18:01:05 -0700 | [diff] [blame] | 354 | logging.debug('Image check compeleted, unmounting the usb drive.') |
Garry Wang | 7b0e1b7 | 2020-03-25 19:08:59 -0700 | [diff] [blame] | 355 | self.run(unmount_cmd, ignore_status=True) |
| 356 | |
| 357 | |
Garry Wang | 70e5d06 | 2020-04-03 18:01:05 -0700 | [diff] [blame] | 358 | def _probe_and_validate_usb_dev(self): |
| 359 | """This method probe the usb dev path by talking to servo, and then |
| 360 | validate the dev path is valid block device to servohost. |
| 361 | Possible output: |
| 362 | 1. Encounter error during probe usb dev, returns empty string. |
| 363 | 2. probe usb dev completed without error but cannot find usb dev, |
| 364 | raise AutoservRepairError. |
| 365 | 3. probe usb dev find a usb dev path, but failed validation in this |
| 366 | method, raise AutoservRepairError. |
Garry Wang | 7b0e1b7 | 2020-03-25 19:08:59 -0700 | [diff] [blame] | 367 | |
Garry Wang | 70e5d06 | 2020-04-03 18:01:05 -0700 | [diff] [blame] | 368 | @returns: A string of usb dev path(e.g. '/dev/sdb'), or empty string |
| 369 | if unexpected error occurred during probe. |
| 370 | @raises: AutoservRepairError if servo couldn't probe the usb dev path |
| 371 | (servo.probe_host_usb_dev() returns empty string), or the dev path is |
| 372 | not valid block device to servohost. |
Garry Wang | 7b0e1b7 | 2020-03-25 19:08:59 -0700 | [diff] [blame] | 373 | """ |
| 374 | logging.info('Validating image usbkey on servo.') |
Garry Wang | 7b0e1b7 | 2020-03-25 19:08:59 -0700 | [diff] [blame] | 375 | try: |
Garry Wang | 70e5d06 | 2020-04-03 18:01:05 -0700 | [diff] [blame] | 376 | usb_dev = self._servo.probe_host_usb_dev() |
Garry Wang | 7b0e1b7 | 2020-03-25 19:08:59 -0700 | [diff] [blame] | 377 | except Exception as e: |
| 378 | # We don't want any unexpected or transient servo communicating |
| 379 | # failure block usb repair, so capture all errors here. |
| 380 | logging.error(e, exc_info=True) |
| 381 | logging.error('Unexpected error occurred on get usbkey dev path,' |
| 382 | ' skipping usbkey validation.') |
| 383 | return '' |
| 384 | |
Garry Wang | 70e5d06 | 2020-04-03 18:01:05 -0700 | [diff] [blame] | 385 | if usb_dev: |
| 386 | # probe_host_usb_dev() sometimes return stale record, |
| 387 | # so we need to make sure the path exists in fdisk. |
| 388 | validate_cmd = 'fdisk -l | grep %s' % usb_dev |
Garry Wang | 1144118 | 2020-06-16 18:34:14 -0700 | [diff] [blame] | 389 | try: |
| 390 | resp = self.run(validate_cmd, ignore_status=True, timeout=60) |
| 391 | if resp.exit_status == 0: |
| 392 | return usb_dev |
Garry Wang | 70e5d06 | 2020-04-03 18:01:05 -0700 | [diff] [blame] | 393 | |
Garry Wang | 1144118 | 2020-06-16 18:34:14 -0700 | [diff] [blame] | 394 | logging.error('%s is reported from "image_usbkey_dev" control' |
| 395 | ' but not detected by fdisk!', usb_dev) |
| 396 | except error.AutoservRunError as e: |
| 397 | if 'Timeout encountered' in str(e): |
| 398 | logging.warning('Timeout encountered during fdisk run,' |
| 399 | ' skipping usbkey validation.') |
| 400 | return '' |
| 401 | raise |
Garry Wang | 70e5d06 | 2020-04-03 18:01:05 -0700 | [diff] [blame] | 402 | |
| 403 | raise hosts.AutoservRepairError( |
| 404 | 'No usbkey detected on servo, the usbkey may be either missing' |
| 405 | ' or broken. Please replace usbkey on the servo and retry.', |
| 406 | 'missing usbkey') |
| 407 | |
| 408 | |
Otabek Kasimov | 4ea636e | 2020-04-14 23:35:06 -0700 | [diff] [blame] | 409 | def is_ec_supported(self): |
Garry Wang | 9b8f234 | 2020-04-17 16:34:09 -0700 | [diff] [blame] | 410 | """Check if ec is supported on the servo_board""" |
Otabek Kasimov | 4ea636e | 2020-04-14 23:35:06 -0700 | [diff] [blame] | 411 | if self.servo_board: |
| 412 | try: |
| 413 | frm_config = config.Config(self.servo_board, self.servo_model) |
| 414 | return frm_config.chrome_ec |
| 415 | except Exception as e: |
| 416 | logging.error('Unexpected error when read from firmware' |
| 417 | ' configs; %s', str(e)) |
| 418 | return False |
| 419 | |
Garry Wang | 70e5d06 | 2020-04-03 18:01:05 -0700 | [diff] [blame] | 420 | def validate_image_usbkey(self): |
| 421 | """This method first validate if there is a recover usbkey on servo |
| 422 | that accessible to servohost, and second check if a ChromeOS image is |
| 423 | already on the usb drive and return the image_name so we can avoid |
| 424 | unnecessary download and flash to the recover usbkey on servo. |
| 425 | |
| 426 | Please note that, there is special error handling logic here: |
| 427 | 1. If unexpected error happens, we return empty string. So repair |
| 428 | actions will not get blocked. |
| 429 | 2. If no working usbkey present on servo, but no errors, we'll raise |
| 430 | AutoservRepairError here. |
| 431 | |
| 432 | @returns: image_name on the usbkey, e.g. nami-release/R82.10138.0.0, |
| 433 | or empty string if no test image detected, or unexpected |
| 434 | error occurred. |
| 435 | @raises: AutoservRepairError if the usbkey is not detected on servo. |
| 436 | """ |
| 437 | usb_dev = self._probe_and_validate_usb_dev() |
| 438 | if usb_dev: |
| 439 | return self.get_image_name_from_usbkey(usb_dev) |
| 440 | else: |
| 441 | return '' |
Garry Wang | 7b0e1b7 | 2020-03-25 19:08:59 -0700 | [diff] [blame] | 442 | |
| 443 | |
Richard Barnette | 1edbb16 | 2016-11-01 11:47:50 -0700 | [diff] [blame] | 444 | def repair(self, silent=False): |
| 445 | """Attempt to repair servo host. |
| 446 | |
| 447 | @param silent If true, suppress logging in `status.log`. |
| 448 | """ |
Richard Barnette | abbdc25 | 2018-07-26 16:57:42 -0700 | [diff] [blame] | 449 | message = 'Beginning repair for servo host %s port %s serial %s' |
| 450 | message %= (self.hostname, self.servo_port, self.servo_serial) |
| 451 | self.record('INFO', None, None, message) |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 452 | try: |
Richard Barnette | 1edbb16 | 2016-11-01 11:47:50 -0700 | [diff] [blame] | 453 | self._repair_strategy.repair(self, silent) |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 454 | self._servo_state = servo_constants.SERVO_STATE_WORKING |
| 455 | self.record('INFO', None, None, |
| 456 | 'ServoHost repair set servo_state as WORKING') |
Garry Wang | 464ff1e | 2019-07-18 17:20:34 -0700 | [diff] [blame] | 457 | # If target is a labstation then try to withdraw any existing |
| 458 | # reboot request created by this servo because it passed repair. |
| 459 | if self.is_labstation(): |
| 460 | self.withdraw_reboot_request() |
Garry Wang | 63b8c38 | 2020-03-11 22:28:40 -0700 | [diff] [blame] | 461 | except Exception as e: |
Otabek Kasimov | 120b6fa | 2020-07-03 00:15:27 -0700 | [diff] [blame] | 462 | if not self.is_localhost(): |
Otabek Kasimov | c6f3041 | 2020-06-30 20:08:12 -0700 | [diff] [blame] | 463 | self._servo_state = self.determine_servo_state() |
| 464 | self.record('INFO', None, None, |
| 465 | 'ServoHost repair set servo_state as %s' |
| 466 | % self._servo_state) |
Garry Wang | 63b8c38 | 2020-03-11 22:28:40 -0700 | [diff] [blame] | 467 | if self._is_critical_error(e): |
| 468 | self.disconnect_servo() |
| 469 | self.stop_servod() |
| 470 | raise |
| 471 | |
| 472 | |
| 473 | def _is_critical_error(self, error): |
| 474 | if (isinstance(error, hosts.AutoservVerifyDependencyError) |
| 475 | and not error.is_critical()): |
| 476 | logging.warning('Non-critical verify failure(s) detected during' |
| 477 | ' verify/repair servo, servo connection will' |
| 478 | ' still up but may not fully functional.' |
| 479 | ' Some repair actions and servo depended' |
| 480 | ' tests may not run.') |
| 481 | return False |
| 482 | logging.info('Critical verify failure(s) detected during repair/verify' |
| 483 | ' servo. Disconnecting servo and stop servod, all repair ' |
| 484 | 'action and tests that depends on servo will not run.') |
| 485 | return True |
Fang Deng | 5d518f4 | 2013-08-02 14:04:32 -0700 | [diff] [blame] | 486 | |
| 487 | |
Dan Shi | 4d47852 | 2014-02-14 13:46:32 -0800 | [diff] [blame] | 488 | def get_servo(self): |
| 489 | """Get the cached servo.Servo object. |
Fang Deng | 5d518f4 | 2013-08-02 14:04:32 -0700 | [diff] [blame] | 490 | |
Dan Shi | 4d47852 | 2014-02-14 13:46:32 -0800 | [diff] [blame] | 491 | @return: a servo.Servo object. |
Dana Goyette | 353d1d9 | 2019-06-27 10:43:59 -0700 | [diff] [blame] | 492 | @rtype: autotest_lib.server.cros.servo.servo.Servo |
Fang Deng | 5d518f4 | 2013-08-02 14:04:32 -0700 | [diff] [blame] | 493 | """ |
Dan Shi | 4d47852 | 2014-02-14 13:46:32 -0800 | [diff] [blame] | 494 | return self._servo |
| 495 | |
| 496 | |
Garry Wang | 79e9af6 | 2019-06-12 15:19:19 -0700 | [diff] [blame] | 497 | def request_reboot(self): |
| 498 | """Request servohost to be rebooted when it's safe to by touch a file. |
| 499 | """ |
| 500 | logging.debug('Request to reboot servohost %s has been created by ' |
Garry Wang | 464ff1e | 2019-07-18 17:20:34 -0700 | [diff] [blame] | 501 | 'servo with port # %s', self.hostname, self.servo_port) |
Garry Wang | 79e9af6 | 2019-06-12 15:19:19 -0700 | [diff] [blame] | 502 | self.run('touch %s' % self._reboot_file, ignore_status=True) |
| 503 | |
| 504 | |
Garry Wang | 464ff1e | 2019-07-18 17:20:34 -0700 | [diff] [blame] | 505 | def withdraw_reboot_request(self): |
| 506 | """Withdraw a servohost reboot request if exists by remove the flag |
| 507 | file. |
| 508 | """ |
| 509 | logging.debug('Withdrawing request to reboot servohost %s that created' |
| 510 | ' by servo with port # %s if exists.', |
| 511 | self.hostname, self.servo_port) |
| 512 | self.run('rm -f %s' % self._reboot_file, ignore_status=True) |
| 513 | |
| 514 | |
Garry Wang | c1288cf | 2019-12-17 14:58:00 -0800 | [diff] [blame] | 515 | def start_servod(self, quick_startup=False): |
| 516 | """Start the servod process on servohost. |
| 517 | """ |
Garry Wang | 2ac15ee | 2019-12-30 19:03:02 -0800 | [diff] [blame] | 518 | # Skip if running on the localhost.(crbug.com/1038168) |
| 519 | if self.is_localhost(): |
| 520 | logging.debug("Servohost is a localhost, skipping start servod.") |
| 521 | return |
| 522 | |
| 523 | cmd = 'start servod' |
Garry Wang | c1288cf | 2019-12-17 14:58:00 -0800 | [diff] [blame] | 524 | if self.servo_board: |
Garry Wang | 2ac15ee | 2019-12-30 19:03:02 -0800 | [diff] [blame] | 525 | cmd += ' BOARD=%s' % self.servo_board |
Garry Wang | c1288cf | 2019-12-17 14:58:00 -0800 | [diff] [blame] | 526 | if self.servo_model: |
| 527 | cmd += ' MODEL=%s' % self.servo_model |
Garry Wang | c1288cf | 2019-12-17 14:58:00 -0800 | [diff] [blame] | 528 | else: |
Garry Wang | 2ac15ee | 2019-12-30 19:03:02 -0800 | [diff] [blame] | 529 | logging.warning('Board for DUT is unknown; starting servod' |
| 530 | ' assuming a pre-configured board.') |
| 531 | |
| 532 | cmd += ' PORT=%d' % self.servo_port |
| 533 | if self.servo_serial: |
| 534 | cmd += ' SERIAL=%s' % self.servo_serial |
Garry Wang | d736748 | 2020-02-27 13:52:40 -0800 | [diff] [blame] | 535 | |
| 536 | # Start servod with dual_v4 if the DUT/servo from designated pools. |
| 537 | dut_host_info = self.get_dut_host_info() |
| 538 | if dut_host_info: |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 539 | if bool(dut_host_info.pools & |
| 540 | servo_constants.POOLS_SUPPORT_DUAL_V4): |
Garry Wang | d736748 | 2020-02-27 13:52:40 -0800 | [diff] [blame] | 541 | logging.debug('The DUT is detected in following designated' |
| 542 | ' pools %s,starting servod with DUAL_V4 option.', |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 543 | servo_constants.POOLS_SUPPORT_DUAL_V4) |
Garry Wang | d736748 | 2020-02-27 13:52:40 -0800 | [diff] [blame] | 544 | cmd += ' DUAL_V4=1' |
| 545 | |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 546 | # Remove the symbolic links from the logs. This helps ensure that |
| 547 | # a failed servod instantiation does not cause us to grab old logs |
| 548 | # by mistake. |
| 549 | self.remove_latest_log_symlinks() |
Garry Wang | cdd27b2 | 2020-01-13 14:59:11 -0800 | [diff] [blame] | 550 | self.run(cmd, timeout=60) |
Garry Wang | c1288cf | 2019-12-17 14:58:00 -0800 | [diff] [blame] | 551 | |
| 552 | # There's a lag between when `start servod` completes and when |
| 553 | # the _ServodConnectionVerifier trigger can actually succeed. |
| 554 | # The call to time.sleep() below gives time to make sure that |
| 555 | # the trigger won't fail after we return. |
| 556 | |
| 557 | # Normally servod on servo_v3 and labstation take ~10 seconds to ready, |
| 558 | # But in the rare case all servo on a labstation are in heavy use they |
| 559 | # may take ~30 seconds. So the timeout value will double these value, |
| 560 | # and we'll try quick start up when first time initialize servohost, |
| 561 | # and use standard start up timeout in repair. |
| 562 | if quick_startup: |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 563 | timeout = servo_constants.SERVOD_QUICK_STARTUP_TIMEOUT |
Garry Wang | c1288cf | 2019-12-17 14:58:00 -0800 | [diff] [blame] | 564 | else: |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 565 | timeout = servo_constants.SERVOD_STARTUP_TIMEOUT |
Garry Wang | c1288cf | 2019-12-17 14:58:00 -0800 | [diff] [blame] | 566 | logging.debug('Wait %s seconds for servod process fully up.', timeout) |
| 567 | time.sleep(timeout) |
Ruben Rodriguez Buchillon | 5bac306 | 2020-03-25 21:32:58 -0700 | [diff] [blame] | 568 | # Cache the initial instance timestamp to check against servod restarts |
| 569 | self._initial_instance_ts = self.get_instance_logs_ts() |
Garry Wang | c1288cf | 2019-12-17 14:58:00 -0800 | [diff] [blame] | 570 | |
| 571 | |
| 572 | def stop_servod(self): |
| 573 | """Stop the servod process on servohost. |
| 574 | """ |
Garry Wang | 2ac15ee | 2019-12-30 19:03:02 -0800 | [diff] [blame] | 575 | # Skip if running on the localhost.(crbug.com/1038168) |
| 576 | if self.is_localhost(): |
| 577 | logging.debug("Servohost is a localhost, skipping stop servod.") |
| 578 | return |
| 579 | |
Garry Wang | c1288cf | 2019-12-17 14:58:00 -0800 | [diff] [blame] | 580 | logging.debug('Stopping servod on port %s', self.servo_port) |
Garry Wang | cdd27b2 | 2020-01-13 14:59:11 -0800 | [diff] [blame] | 581 | self.run('stop servod PORT=%d' % self.servo_port, |
| 582 | timeout=60, ignore_status=True) |
Garry Wang | c1288cf | 2019-12-17 14:58:00 -0800 | [diff] [blame] | 583 | logging.debug('Wait %s seconds for servod process fully teardown.', |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 584 | servo_constants.SERVOD_TEARDOWN_TIMEOUT) |
| 585 | time.sleep(servo_constants.SERVOD_TEARDOWN_TIMEOUT) |
Garry Wang | c1288cf | 2019-12-17 14:58:00 -0800 | [diff] [blame] | 586 | |
| 587 | |
| 588 | def restart_servod(self, quick_startup=False): |
| 589 | """Restart the servod process on servohost. |
| 590 | """ |
| 591 | self.stop_servod() |
| 592 | self.start_servod(quick_startup) |
| 593 | |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 594 | |
Garry Wang | ffbd216 | 2020-04-17 16:13:48 -0700 | [diff] [blame] | 595 | def _process_servodtool_error(self, response): |
| 596 | """Helper function to handle non-zero servodtool response. |
| 597 | """ |
| 598 | if re.search(servo_constants.ERROR_MESSAGE_USB_HUB_NOT_COMPATIBLE, |
Garry Wang | ad24500 | 2020-05-15 15:20:23 -0700 | [diff] [blame] | 599 | response.stdout): |
Garry Wang | ffbd216 | 2020-04-17 16:13:48 -0700 | [diff] [blame] | 600 | logging.error('The servo is not plugged on a usb hub that supports' |
| 601 | ' power-cycle!') |
Garry Wang | 000c6c0 | 2020-05-11 21:27:23 -0700 | [diff] [blame] | 602 | # change the flag so we can update this label in later process. |
| 603 | self.smart_usbhub = False |
Garry Wang | ffbd216 | 2020-04-17 16:13:48 -0700 | [diff] [blame] | 604 | return |
| 605 | |
| 606 | if re.search(servo_constants.ERROR_MESSAGE_DEVICE_NOT_FOUND % |
| 607 | self.servo_serial, response.stdout): |
| 608 | logging.error('No servo with serial %s found!', self.servo_serial) |
| 609 | return |
| 610 | |
| 611 | logging.error('Unexpected error occurred from usbhub control, please' |
| 612 | ' file a bug and inform chrome-fleet-software@ team!') |
| 613 | |
| 614 | |
| 615 | def _get_servo_usb_devnum(self): |
| 616 | """Helper function to collect current usb devnum of servo. |
| 617 | """ |
Otabek Kasimov | 0919268 | 2020-06-01 18:17:44 -0700 | [diff] [blame] | 618 | # TODO remove try-except when fix crbug.com/1087964 |
| 619 | try: |
| 620 | cmd = 'servodtool device -s %s usb-path' % self.servo_serial |
| 621 | resp = self.run(cmd, ignore_status=True, timeout=30) |
| 622 | except Exception as e: |
| 623 | # Here we catch only timeout errors. |
| 624 | # Other errors is filtered by ignore_status=True |
| 625 | logging.debug('Attempt to get servo usb-path failed due to ' |
| 626 | 'timeout; %s', e) |
| 627 | return '' |
| 628 | |
Garry Wang | ffbd216 | 2020-04-17 16:13:48 -0700 | [diff] [blame] | 629 | if resp.exit_status != 0: |
| 630 | self._process_servodtool_error(resp) |
| 631 | return '' |
| 632 | usb_path = resp.stdout.strip() |
| 633 | logging.info('Usb path of servo %s is %s', self.servo_serial, usb_path) |
| 634 | |
| 635 | resp = self.run('cat %s/devnum' % usb_path, |
| 636 | ignore_status=True) |
| 637 | if resp.exit_status != 0: |
| 638 | self._process_servodtool_error(resp) |
| 639 | return '' |
| 640 | return resp.stdout.strip() |
| 641 | |
Garry Wang | 358aad4 | 2020-08-02 20:56:04 -0700 | [diff] [blame] | 642 | def reboot_servo_v3_on_need(self): |
| 643 | """Check and reboot servo_v3 based on below conditions. |
| 644 | 1. If there is an update pending on reboot. |
| 645 | 2. Servo_v3 has been up for more than 96 hours. |
| 646 | """ |
| 647 | if self.get_board() != 'beaglebone_servo': |
| 648 | logging.info('Servo reboot is only applicable for servo V3.') |
Otabek Kasimov | e6df810 | 2020-07-21 20:15:25 -0700 | [diff] [blame] | 649 | return |
| 650 | |
Garry Wang | 358aad4 | 2020-08-02 20:56:04 -0700 | [diff] [blame] | 651 | update_pending_reboot = (self._check_update_status() == |
| 652 | self.UPDATE_STATE.PENDING_REBOOT) |
| 653 | uptime_hours = float(self.check_uptime())/3600 |
| 654 | logging.info('Uptime of servo_v3: %s hour(s)', uptime_hours) |
| 655 | long_up_time = uptime_hours > 96 |
| 656 | |
| 657 | # Skip reboot if neither condition are met. |
| 658 | if not (update_pending_reboot or long_up_time): |
Otabek Kasimov | e6df810 | 2020-07-21 20:15:25 -0700 | [diff] [blame] | 659 | return |
| 660 | |
Garry Wang | 358aad4 | 2020-08-02 20:56:04 -0700 | [diff] [blame] | 661 | if update_pending_reboot: |
| 662 | message = 'Starting reboot servo_v3 because an update is pending.' |
| 663 | reboot_method = self._post_update_reboot |
| 664 | elif long_up_time: |
| 665 | message = 'Starting reboot servo_v3 because uptime > 96 hours.' |
| 666 | reboot_method = self._servo_host_reboot |
| 667 | self.record('INFO', None, None, message) |
| 668 | logging.info(message) |
Otabek Kasimov | e6df810 | 2020-07-21 20:15:25 -0700 | [diff] [blame] | 669 | try: |
Garry Wang | 358aad4 | 2020-08-02 20:56:04 -0700 | [diff] [blame] | 670 | reboot_method() |
Otabek Kasimov | e6df810 | 2020-07-21 20:15:25 -0700 | [diff] [blame] | 671 | message = 'Servo_v3 reboot completed successfully.' |
| 672 | except Exception as e: |
| 673 | logging.debug("Fail to reboot servo_v3; %s", e) |
| 674 | message = ('Servo_v3 reboot failed, please check debug log ' |
| 675 | 'for details.') |
| 676 | logging.info(message) |
| 677 | self.record('INFO', None, None, message) |
Garry Wang | ffbd216 | 2020-04-17 16:13:48 -0700 | [diff] [blame] | 678 | |
| 679 | def _reset_servo(self): |
| 680 | logging.info('Resetting servo through smart usbhub.') |
Otabek Kasimov | 0919268 | 2020-06-01 18:17:44 -0700 | [diff] [blame] | 681 | # TODO remove try-except when fix crbug.com/1087964 |
| 682 | try: |
| 683 | resp = self.run('servodtool device -s %s power-cycle' % |
| 684 | self.servo_serial, ignore_status=True, |
| 685 | timeout=30) |
| 686 | if resp.exit_status != 0: |
| 687 | self._process_servodtool_error(resp) |
| 688 | return False |
| 689 | except Exception as e: |
| 690 | # Here we catch only timeout errors. |
| 691 | # Other errors is filtered by ignore_status=True |
| 692 | logging.debug('Attempt to reset servo failed due to timeout;' |
| 693 | ' %s', e) |
Garry Wang | ffbd216 | 2020-04-17 16:13:48 -0700 | [diff] [blame] | 694 | return False |
| 695 | |
| 696 | logging.debug('Wait %s seconds for servo to come back from reset.', |
| 697 | servo_constants.SERVO_RESET_TIMEOUT_SECONDS) |
| 698 | time.sleep(servo_constants.SERVO_RESET_TIMEOUT_SECONDS) |
Garry Wang | 000c6c0 | 2020-05-11 21:27:23 -0700 | [diff] [blame] | 699 | # change the flag so we can update this label in later process. |
| 700 | self.smart_usbhub = True |
Garry Wang | ffbd216 | 2020-04-17 16:13:48 -0700 | [diff] [blame] | 701 | return True |
| 702 | |
| 703 | |
| 704 | def reset_servo(self): |
| 705 | """Reset(power-cycle) the servo via smart usbhub. |
| 706 | """ |
| 707 | if not self.is_labstation(): |
| 708 | logging.info('Servo reset is not applicable to servo_v3.') |
| 709 | return |
| 710 | |
| 711 | pre_reset_devnum = self._get_servo_usb_devnum() |
| 712 | logging.info('Servo usb devnum before reset: %s', pre_reset_devnum) |
| 713 | result = self._reset_servo() |
| 714 | if not result: |
Garry Wang | fd5c8b6 | 2020-06-08 15:36:54 -0700 | [diff] [blame] | 715 | message = ('Failed to reset servo with serial: %s. (Please ignore' |
| 716 | ' this error if the DUT is not connected to a smart' |
| 717 | ' usbhub).' % self.servo_serial) |
Garry Wang | ffbd216 | 2020-04-17 16:13:48 -0700 | [diff] [blame] | 718 | logging.warning(message) |
| 719 | self.record('INFO', None, None, message) |
| 720 | return |
| 721 | |
| 722 | post_reset_devnum = self._get_servo_usb_devnum() |
| 723 | logging.info('Servo usb devnum after reset: %s', post_reset_devnum) |
| 724 | if not (pre_reset_devnum and post_reset_devnum): |
| 725 | message = ('Servo reset completed but unable to verify' |
| 726 | ' devnum change!') |
| 727 | elif pre_reset_devnum != post_reset_devnum: |
| 728 | message = ('Reset servo with serial %s completed successfully!' |
| 729 | % self.servo_serial) |
| 730 | else: |
| 731 | message = 'Servo reset completed but devnum is still not changed!' |
| 732 | logging.info(message) |
| 733 | self.record('INFO', None, None, message) |
| 734 | |
| 735 | |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 736 | def _extract_compressed_logs(self, logdir, relevant_files): |
| 737 | """Decompress servod logs in |logdir|. |
| 738 | |
| 739 | @param logdir: directory containing compressed servod logs. |
| 740 | @param relevant_files: list of files in |logdir| to consider. |
| 741 | |
| 742 | @returns: tuple, (tarfiles, files) where |
| 743 | tarfiles: list of the compressed filenames that have been |
| 744 | extracted and deleted |
| 745 | files: list of the uncompressed files that were generated |
| 746 | """ |
| 747 | # For all tar-files, first extract them to the directory, and |
| 748 | # then let the common flow handle them. |
| 749 | tarfiles = [cf for cf in relevant_files if |
| 750 | cf.endswith(self.COMPRESSION_SUFFIX)] |
| 751 | files = [] |
| 752 | for f in tarfiles: |
| 753 | norm_name = os.path.basename(f)[:-len(self.COMPRESSION_SUFFIX)] |
| 754 | with tarfile.open(f) as tf: |
| 755 | # Each tarfile has only one member, as |
| 756 | # that's the compressed log. |
| 757 | member = tf.members[0] |
| 758 | # Manipulate so that it only extracts the basename, and not |
| 759 | # the directories etc. |
| 760 | member.name = norm_name |
| 761 | files.append(os.path.join(logdir, member.name)) |
| 762 | tf.extract(member, logdir) |
| 763 | # File has been extracted: remove the compressed file. |
| 764 | os.remove(f) |
| 765 | return tarfiles, files |
| 766 | |
| 767 | def _extract_mcu_logs(self, log_subdir): |
| 768 | """Extract MCU (EC, Cr50, etc) console output from servod debug logs. |
| 769 | |
| 770 | Using the MCU_EXTRACTOR regex (above) extract and split out MCU console |
| 771 | lines from the logs to generate invidiual console logs e.g. after |
| 772 | this method, you can find an ec.txt and servo_v4.txt in |log_dir| if |
| 773 | those MCUs had any console input/output. |
| 774 | |
| 775 | @param log_subdir: directory with log.DEBUG.txt main servod debug logs. |
| 776 | """ |
| 777 | # Extract the MCU for each one. The MCU logs are only in the .DEBUG |
| 778 | # files |
| 779 | mcu_lines_file = os.path.join(log_subdir, 'log.DEBUG.txt') |
| 780 | if not os.path.exists(mcu_lines_file): |
| 781 | logging.info('No DEBUG logs found to extract MCU logs from.') |
| 782 | return |
| 783 | mcu_files = {} |
| 784 | mcu_file_template = '%s.txt' |
| 785 | with open(mcu_lines_file, 'r') as f: |
| 786 | for line in f: |
| 787 | match = self.MCU_EXTRACTOR.match(line) |
| 788 | if match: |
| 789 | mcu = match.group(self.MCU_GROUP).lower() |
| 790 | line = match.group(self.LINE_GROUP) |
| 791 | if mcu not in mcu_files: |
| 792 | mcu_file = os.path.join(log_subdir, |
| 793 | mcu_file_template % mcu) |
| 794 | mcu_files[mcu] = open(mcu_file, 'a') |
| 795 | fd = mcu_files[mcu] |
| 796 | fd.write(line + '\n') |
| 797 | for f in mcu_files: |
| 798 | mcu_files[f].close() |
| 799 | |
| 800 | |
| 801 | def remove_latest_log_symlinks(self): |
| 802 | """Remove the conveninence symlinks 'latest' servod logs.""" |
| 803 | symlink_wildcard = '%s/latest*' % self.remote_log_dir |
| 804 | cmd = 'rm ' + symlink_wildcard |
| 805 | self.run(cmd, stderr_tee=None, ignore_status=True) |
| 806 | |
Ruben Rodriguez Buchillon | 5bac306 | 2020-03-25 21:32:58 -0700 | [diff] [blame] | 807 | def probe_servod_restart(self, instance_ts, outdir): |
| 808 | """Grab servod logs from previous instances if part of this session. |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 809 | |
Ruben Rodriguez Buchillon | 5bac306 | 2020-03-25 21:32:58 -0700 | [diff] [blame] | 810 | If since the last time this host called start_servod() servod crashed |
| 811 | and restarted, this helper finds those logs as well, and stores them |
| 812 | with the |OLD_LOG_SUFFIX| to investigate if necessary. |
Prasad Vuppalapu | 5bd9da1 | 2020-03-31 01:46:47 +0000 | [diff] [blame] | 813 | |
Ruben Rodriguez Buchillon | 5bac306 | 2020-03-25 21:32:58 -0700 | [diff] [blame] | 814 | It also issues a panicinfo command to servo devices after the restart |
| 815 | to try and collect reboot information for debugging. |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 816 | |
Ruben Rodriguez Buchillon | 5bac306 | 2020-03-25 21:32:58 -0700 | [diff] [blame] | 817 | @param instance_ts: the log timestamp that the current instance uses |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 818 | @param outdir: directory to create a subdirectory into to place the |
| 819 | servod logs into. |
| 820 | """ |
Ruben Rodriguez Buchillon | 5bac306 | 2020-03-25 21:32:58 -0700 | [diff] [blame] | 821 | if self._initial_instance_ts is None: |
| 822 | logging.info('No log timestamp grabbed successfully on servod ' |
| 823 | 'startup. Cannot check device restarts. Ignoring.') |
| 824 | return |
| 825 | if instance_ts == self._initial_instance_ts: |
| 826 | logging.debug('Servod appears to have run without restarting') |
| 827 | return |
| 828 | # Servod seems to have restarted (at least once). |_initial_instance_ts| |
| 829 | # is the first timestamp, and instance_ts is the current timestamp. Find |
| 830 | # all timestamps in between them, and grab the logs for each. |
| 831 | tss = self._find_instance_timestamps_between(self._initial_instance_ts, |
| 832 | instance_ts) |
| 833 | logging.info('Servod has restarted %d times between the start and the ' |
| 834 | 'end of this servo_host.', len(tss)) |
| 835 | logging.info('This might be an issue. Will extract all logs from each ' |
| 836 | 'instance.') |
| 837 | logging.info('Logs that are not the currently running (about to turn ' |
| 838 | 'down) instance are maked with a .%s in their folder.', |
| 839 | self.OLD_LOG_SUFFIX) |
| 840 | for ts in tss: |
| 841 | self.get_instance_logs(ts, outdir, old=True) |
| 842 | # Lastly, servod has restarted due to a potential issue. Try to get |
| 843 | # panic information from servo micro and servo v4 for the current logs. |
| 844 | # This can only happen if the |_servo| attribute is initialized. |
| 845 | if self._servo: |
| 846 | for mcu in ['servo_micro', 'servo_v4']: |
| 847 | ctrl = '%s_uart_cmd' % mcu |
| 848 | if self._servo.has_control(ctrl): |
| 849 | logging.info('Trying to retrieve %r panicinfo into logs', |
| 850 | mcu) |
| 851 | try: |
| 852 | self._servo.set_nocheck(ctrl, 'panicinfo') |
| 853 | except error.TestFail as e: |
| 854 | logging.error('Failed to generate panicinfo for %r ' |
| 855 | 'logs. %s', mcu, str(e)) |
| 856 | |
| 857 | def _find_instance_timestamps_between(self, start_ts, end_ts): |
| 858 | """Find all log timestamps between [start_ts, end_ts). |
| 859 | |
| 860 | @param start_ts: str, earliest log timestamp of interest |
| 861 | @param end_ts: str, latest log timestamp of interest |
| 862 | |
| 863 | @returns: list, all timestamps between start_ts and end_ts, end_ts |
| 864 | exclusive, on the servo_host. An empty list on errors |
| 865 | """ |
| 866 | # Simply get all timestamp, and then sort and remove |
| 867 | cmd = 'ls %s' % self.remote_log_dir |
| 868 | res = self.run(cmd, stderr_tee=None, ignore_status=True) |
| 869 | if res.exit_status != 0: |
| 870 | # Here we failed to find anything. |
| 871 | logging.info('Failed to find remote servod logs. Ignoring.') |
| 872 | return [] |
| 873 | logfiles = res.stdout.strip().split() |
| 874 | timestamps = set() |
| 875 | for logfile in logfiles: |
| 876 | ts_match = self.TS_EXTRACTOR.match(logfile) |
| 877 | if not ts_match: |
| 878 | # Simply ignore files that fail the check. It might be the |
| 879 | # 'latest' symlinks or random files. |
| 880 | continue |
| 881 | timestamps.add(ts_match.group(self.TS_GROUP)) |
| 882 | # At this point we have all unique timestamps. |
| 883 | timestamps = sorted(timestamps) |
| 884 | for ts in [start_ts, end_ts]: |
| 885 | if ts not in timestamps: |
| 886 | logging.error('Timestamp %r not in servod logs. Cannot query ' |
| 887 | 'for timestamps in between %r and %r', ts, |
| 888 | start_ts, end_ts) |
| 889 | return [] |
| 890 | return timestamps[timestamps.index(start_ts):timestamps.index(end_ts)] |
| 891 | |
| 892 | def get_instance_logs_ts(self): |
| 893 | """Retrieve the currently running servod instance's log timestamp |
| 894 | |
| 895 | @returns: str, timestamp for current instance, or None on failure |
| 896 | """ |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 897 | # First, extract the timestamp. This cmd gives the real filename of |
| 898 | # the latest aka current log file. |
| 899 | cmd = ('if [ -f %(dir)s/latest.DEBUG ];' |
| 900 | 'then realpath %(dir)s/latest.DEBUG;' |
| 901 | 'elif [ -f %(dir)s/latest ];' |
| 902 | 'then realpath %(dir)s/latest;' |
| 903 | 'else exit %(code)d;' |
| 904 | 'fi' % {'dir': self.remote_log_dir, |
| 905 | 'code': self.NO_SYMLINKS_CODE}) |
| 906 | res = self.run(cmd, stderr_tee=None, ignore_status=True) |
| 907 | if res.exit_status != 0: |
| 908 | if res.exit_status == self.NO_SYMLINKS_CODE: |
| 909 | logging.warning('servod log latest symlinks not found. ' |
| 910 | 'This is likely due to an error starting up ' |
| 911 | 'servod. Ignoring..') |
| 912 | else: |
| 913 | logging.warning('Failed to find servod logs on servo host.') |
| 914 | logging.warning(res.stderr.strip()) |
Ruben Rodriguez Buchillon | 5bac306 | 2020-03-25 21:32:58 -0700 | [diff] [blame] | 915 | return None |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 916 | fname = os.path.basename(res.stdout.strip()) |
| 917 | # From the fname, ought to extract the timestamp using the TS_EXTRACTOR |
Ruben Rodriguez Buchillon | e9aa2b0 | 2020-03-04 12:14:28 -0800 | [diff] [blame] | 918 | ts_match = self.TS_EXTRACTOR.match(fname) |
| 919 | if not ts_match: |
| 920 | logging.warning('Failed to extract timestamp from servod log file ' |
| 921 | '%r. Skipping. The servo host is using outdated ' |
| 922 | 'servod logging and needs to be updated.', fname) |
Ruben Rodriguez Buchillon | 5bac306 | 2020-03-25 21:32:58 -0700 | [diff] [blame] | 923 | return None |
| 924 | return ts_match.group(self.TS_GROUP) |
| 925 | |
| 926 | def get_instance_logs(self, instance_ts, outdir, old=False): |
| 927 | """Collect all logs with |instance_ts| and dump into a dir in |outdir| |
| 928 | |
| 929 | This method first collects all logs on the servo_host side pertaining |
| 930 | to this servod instance (port, instatiation). It glues them together |
| 931 | into combined log.[level].txt files and extracts all available MCU |
| 932 | console I/O from the logs into individual files e.g. servo_v4.txt |
| 933 | |
| 934 | All the output can be found in a directory inside |outdir| that |
| 935 | this generates based on |LOG_DIR|, the servod port, and the instance |
| 936 | timestamp on the servo_host side. |
| 937 | |
| 938 | @param instance_ts: log timestamp to grab logfiles for |
| 939 | @param outdir: directory to create a subdirectory into to place the |
| 940 | servod logs into. |
| 941 | @param old: bool, whether to append |OLD_LOG_SUFFIX| to output dir |
| 942 | """ |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 943 | # Create the local results log dir. |
| 944 | log_dir = os.path.join(outdir, '%s_%s.%s' % (self.LOG_DIR, |
| 945 | str(self.servo_port), |
| 946 | instance_ts)) |
Ruben Rodriguez Buchillon | 5bac306 | 2020-03-25 21:32:58 -0700 | [diff] [blame] | 947 | if old: |
| 948 | log_dir = '%s.%s' % (log_dir, self.OLD_LOG_SUFFIX) |
| 949 | logging.info('Saving servod logs to %r.', log_dir) |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 950 | os.mkdir(log_dir) |
| 951 | # Now, get all files with that timestamp. |
| 952 | cmd = 'find %s -maxdepth 1 -name "log.%s*"' % (self.remote_log_dir, |
| 953 | instance_ts) |
| 954 | res = self.run(cmd, stderr_tee=None, ignore_status=True) |
| 955 | files = res.stdout.strip().split() |
| 956 | try: |
| 957 | self.get_file(files, log_dir, try_rsync=False) |
Ruben Rodriguez Buchillon | 5bac306 | 2020-03-25 21:32:58 -0700 | [diff] [blame] | 958 | if not os.listdir(log_dir): |
| 959 | logging.info('No servod logs retrieved. Ignoring, and removing ' |
| 960 | '%r again.', log_dir) |
| 961 | os.rmdir(log_dir) |
| 962 | return |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 963 | except error.AutoservRunError as e: |
| 964 | result = e.result_obj |
| 965 | if result.exit_status != 0: |
| 966 | stderr = result.stderr.strip() |
| 967 | logging.warning("Couldn't retrieve servod logs. Ignoring: %s", |
| 968 | stderr or '\n%s' % result) |
Ruben Rodriguez Buchillon | 5bac306 | 2020-03-25 21:32:58 -0700 | [diff] [blame] | 969 | # Remove the log_dir as nothing was added to it. |
| 970 | os.rmdir(log_dir) |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 971 | return |
| 972 | local_files = [os.path.join(log_dir, f) for f in os.listdir(log_dir)] |
| 973 | # TODO(crrev.com/c/1793030): remove no-level case once CL is pushed |
| 974 | for level_name in ('DEBUG', 'INFO', 'WARNING', ''): |
| 975 | # Create the joint files for each loglevel. i.e log.DEBUG |
| 976 | joint_file = self.JOINT_LOG_PREFIX |
| 977 | if level_name: |
| 978 | joint_file = '%s.%s' % (self.JOINT_LOG_PREFIX, level_name) |
| 979 | # This helps with some online tools to avoid complaints about an |
| 980 | # unknown filetype. |
| 981 | joint_file = joint_file + '.txt' |
| 982 | joint_path = os.path.join(log_dir, joint_file) |
| 983 | files = [f for f in local_files if level_name in f] |
| 984 | if not files: |
| 985 | # TODO(crrev.com/c/1793030): remove no-level case once CL |
| 986 | # is pushed |
| 987 | continue |
| 988 | # Extract compressed logs if any. |
| 989 | compressed, extracted = self._extract_compressed_logs(log_dir, |
| 990 | files) |
| 991 | files = list(set(files) - set(compressed)) |
| 992 | files.extend(extracted) |
| 993 | # Need to sort. As they all share the same timestamp, and |
| 994 | # loglevel, the index itself is sufficient. The highest index |
| 995 | # is the oldest file, therefore we need a descending sort. |
| 996 | def sortkey(f, level=level_name): |
| 997 | """Custom sortkey to sort based on rotation number int.""" |
| 998 | if f.endswith(level_name): return 0 |
| 999 | return int(f.split('.')[-1]) |
| 1000 | |
| 1001 | files.sort(reverse=True, key=sortkey) |
| 1002 | # Just rename the first file rather than building from scratch. |
| 1003 | os.rename(files[0], joint_path) |
| 1004 | with open(joint_path, 'a') as joint_f: |
| 1005 | for logfile in files[1:]: |
| 1006 | # Transfer the file to the joint file line by line. |
| 1007 | with open(logfile, 'r') as log_f: |
| 1008 | for line in log_f: |
| 1009 | joint_f.write(line) |
| 1010 | # File has been written over. Delete safely. |
| 1011 | os.remove(logfile) |
| 1012 | # Need to remove all files form |local_files| so we don't |
| 1013 | # analyze them again. |
| 1014 | local_files = list(set(local_files) - set(files) - set(compressed)) |
| 1015 | # Lastly, extract MCU logs from the joint logs. |
| 1016 | self._extract_mcu_logs(log_dir) |
| 1017 | |
Garry Wang | 79e9af6 | 2019-06-12 15:19:19 -0700 | [diff] [blame] | 1018 | def _lock(self): |
| 1019 | """lock servohost by touching a file. |
| 1020 | """ |
| 1021 | logging.debug('Locking servohost %s by touching %s file', |
| 1022 | self.hostname, self._lock_file) |
| 1023 | self.run('touch %s' % self._lock_file, ignore_status=True) |
Garry Wang | 7c00b0f | 2019-06-25 17:28:17 -0700 | [diff] [blame] | 1024 | self._is_locked = True |
Garry Wang | 79e9af6 | 2019-06-12 15:19:19 -0700 | [diff] [blame] | 1025 | |
| 1026 | |
| 1027 | def _unlock(self): |
| 1028 | """Unlock servohost by removing the lock file. |
| 1029 | """ |
| 1030 | logging.debug('Unlocking servohost by removing %s file', |
| 1031 | self._lock_file) |
| 1032 | self.run('rm %s' % self._lock_file, ignore_status=True) |
Garry Wang | 7c00b0f | 2019-06-25 17:28:17 -0700 | [diff] [blame] | 1033 | self._is_locked = False |
Garry Wang | 79e9af6 | 2019-06-12 15:19:19 -0700 | [diff] [blame] | 1034 | |
| 1035 | |
Congbin Guo | a1f9cba | 2018-07-03 11:36:59 -0700 | [diff] [blame] | 1036 | def close(self): |
Congbin Guo | fc3b896 | 2019-03-22 17:38:46 -0700 | [diff] [blame] | 1037 | """Close the associated servo and the host object.""" |
Ruben Rodriguez Buchillon | 5bac306 | 2020-03-25 21:32:58 -0700 | [diff] [blame] | 1038 | # NOTE: throughout this method there are multiple attempts to catch |
| 1039 | # all errors. This is WAI as log grabbing should not fail tests. |
| 1040 | # However, the goal is to catch and handle/process all errors, thus |
| 1041 | # we print the traceback and ask for a bug. |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 1042 | if self._closed: |
| 1043 | logging.debug('ServoHost is already closed.') |
| 1044 | return |
Ruben Rodriguez Buchillon | 5bac306 | 2020-03-25 21:32:58 -0700 | [diff] [blame] | 1045 | instance_ts = self.get_instance_logs_ts() |
| 1046 | # TODO(crbug.com/1011516): once enabled, remove the check against |
| 1047 | # localhost and instead check against log-rotiation enablement. |
| 1048 | logs_available = (instance_ts is not None and |
| 1049 | self.job and |
| 1050 | not self.is_localhost()) |
| 1051 | if logs_available: |
| 1052 | # Probe whether there was a servod restart, and grab those old |
| 1053 | # logs as well. |
| 1054 | try: |
| 1055 | self.probe_servod_restart(instance_ts, self.job.resultdir) |
| 1056 | except (error.AutoservRunError, error.TestFail) as e: |
| 1057 | logging.info('Failed to grab servo logs due to: %s. ' |
| 1058 | 'This error is forgiven.', str(e)) |
| 1059 | except Exception as e: |
| 1060 | logging.error('Unexpected error probing for old logs. %s. ' |
| 1061 | 'Forgiven. Please file a bug and fix or catch ' |
| 1062 | 'in log probing function', str(e), |
| 1063 | exc_info=True) |
Congbin Guo | a1f9cba | 2018-07-03 11:36:59 -0700 | [diff] [blame] | 1064 | if self._servo: |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 1065 | outdir = None if not self.job else self.job.resultdir |
Congbin Guo | 2e5e2a2 | 2018-07-27 10:32:48 -0700 | [diff] [blame] | 1066 | # In some cases when we run as lab-tools, the job object is None. |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 1067 | self._servo.close(outdir) |
| 1068 | |
Ruben Rodriguez Buchillon | 5bac306 | 2020-03-25 21:32:58 -0700 | [diff] [blame] | 1069 | if logs_available: |
| 1070 | # Grab current (not old like above) logs after the servo instance |
| 1071 | # was closed out. |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 1072 | try: |
Ruben Rodriguez Buchillon | 5bac306 | 2020-03-25 21:32:58 -0700 | [diff] [blame] | 1073 | self.get_instance_logs(instance_ts, self.job.resultdir) |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 1074 | except error.AutoservRunError as e: |
| 1075 | logging.info('Failed to grab servo logs due to: %s. ' |
| 1076 | 'This error is forgiven.', str(e)) |
Ruben Rodriguez Buchillon | 5bac306 | 2020-03-25 21:32:58 -0700 | [diff] [blame] | 1077 | except Exception as e: |
| 1078 | logging.error('Unexpected error grabbing servod logs. %s. ' |
| 1079 | 'Forgiven. Please file a bug and fix or catch ' |
| 1080 | 'in log grabbing function', str(e), exc_info=True) |
Congbin Guo | a1f9cba | 2018-07-03 11:36:59 -0700 | [diff] [blame] | 1081 | |
Garry Wang | 7c00b0f | 2019-06-25 17:28:17 -0700 | [diff] [blame] | 1082 | if self._is_locked: |
| 1083 | # Remove the lock if the servohost has been locked. |
Garry Wang | 79e9af6 | 2019-06-12 15:19:19 -0700 | [diff] [blame] | 1084 | try: |
| 1085 | self._unlock() |
| 1086 | except error.AutoservSSHTimeout: |
| 1087 | logging.error('Unlock servohost failed due to ssh timeout.' |
| 1088 | ' It may caused by servohost went down during' |
| 1089 | ' the task.') |
Garry Wang | c1288cf | 2019-12-17 14:58:00 -0800 | [diff] [blame] | 1090 | # We want always stop servod after task to minimum the impact of bad |
| 1091 | # servod process interfere other servods.(see crbug.com/1028665) |
Garry Wang | 4c624bc | 2020-01-27 16:34:43 -0800 | [diff] [blame] | 1092 | try: |
| 1093 | self.stop_servod() |
| 1094 | except error.AutoservRunError as e: |
| 1095 | logging.info("Failed to stop servod due to:\n%s\n" |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 1096 | "This error is forgiven.", str(e)) |
Garry Wang | c1288cf | 2019-12-17 14:58:00 -0800 | [diff] [blame] | 1097 | |
Congbin Guo | a1f9cba | 2018-07-03 11:36:59 -0700 | [diff] [blame] | 1098 | super(ServoHost, self).close() |
Ruben Rodriguez Buchillon | 93084d0 | 2020-01-21 15:17:36 -0800 | [diff] [blame] | 1099 | # Mark closed. |
| 1100 | self._closed = True |
Congbin Guo | a1f9cba | 2018-07-03 11:36:59 -0700 | [diff] [blame] | 1101 | |
| 1102 | |
Otabek Kasimov | cc9738e | 2020-02-14 16:17:15 -0800 | [diff] [blame] | 1103 | def get_servo_state(self): |
Otabek Kasimov | a7ba91a | 2020-03-09 08:31:01 -0700 | [diff] [blame] | 1104 | return self._servo_state |
Otabek Kasimov | cc9738e | 2020-02-14 16:17:15 -0800 | [diff] [blame] | 1105 | |
Otabek Kasimov | c6f3041 | 2020-06-30 20:08:12 -0700 | [diff] [blame] | 1106 | def _get_host_metrics_data(self): |
| 1107 | return {'port': self.servo_port, |
Otabek Kasimov | 0ea4736 | 2020-07-11 20:55:09 -0700 | [diff] [blame] | 1108 | 'host': self.get_dut_hostname() or self.hostname, |
Otabek Kasimov | c6f3041 | 2020-06-30 20:08:12 -0700 | [diff] [blame] | 1109 | 'board': self.servo_board or ''} |
| 1110 | |
| 1111 | def _is_servo_device_connected(self, servo_type, serial): |
| 1112 | """Check if device is connected to the labstation. |
| 1113 | |
| 1114 | Works for all servo devices connected to the labstation. |
| 1115 | For servo_v3 please use 'self._is_servo_board_present_on_servo_v3' |
| 1116 | |
| 1117 | @param servo_type: The type of servo device. Expecting value can be |
| 1118 | servo_v4 or servo_micro. |
| 1119 | @param serial: The serial number of the device to detect it. |
| 1120 | """ |
| 1121 | vid_pid = self.SERVO_VID_PID.get(servo_type) |
| 1122 | if not vid_pid or not serial: |
| 1123 | # device cannot detected without VID/PID or serial number |
| 1124 | return False |
| 1125 | logging.debug('Started to detect %s', servo_type) |
| 1126 | try: |
| 1127 | cmd = 'lsusb -v -d %s |grep iSerial |grep %s' % (vid_pid, serial) |
| 1128 | result = self.run(cmd, ignore_status=True, timeout=30) |
| 1129 | if result.exit_status == 0 and result.stdout.strip(): |
| 1130 | logging.debug('The %s is plugged in to the host.', servo_type) |
| 1131 | return True |
| 1132 | logging.debug('%s device is not detected; %s', servo_type, result) |
| 1133 | return False |
| 1134 | except Exception as e: |
| 1135 | # can be triggered by timeout issue due running the script |
| 1136 | metrics.Counter( |
| 1137 | 'chromeos/autotest/repair/servo_detection/timeout' |
| 1138 | ).increment(fields=self._get_host_metrics_data()) |
| 1139 | logging.error('%s device is not detected; %s', servo_type, str(e)) |
| 1140 | return None |
| 1141 | |
| 1142 | def _is_servo_board_present_on_servo_v3(self): |
| 1143 | """Check if servo board is detected on servo_v3""" |
| 1144 | vid_pids = self.SERVO_VID_PID['servo_v3'] |
| 1145 | if not vid_pids or len(vid_pids) == 0: |
| 1146 | # device cannot detected without VID/PID |
| 1147 | return False |
| 1148 | logging.debug('Started to detect servo board on servo_v3') |
| 1149 | not_detected = 'The servo board is not detected on servo_v3' |
| 1150 | try: |
| 1151 | cmd = 'lsusb | grep "%s"' % "\|".join(vid_pids) |
| 1152 | result = self.run(cmd, ignore_status=True, timeout=30) |
| 1153 | if result.exit_status == 0 and result.stdout.strip(): |
| 1154 | logging.debug('The servo board is detected on servo_v3') |
| 1155 | return True |
| 1156 | logging.debug('%s; %s', not_detected, result) |
| 1157 | return False |
| 1158 | except Exception as e: |
| 1159 | # can be triggered by timeout issue due running the script |
| 1160 | metrics.Counter( |
| 1161 | 'chromeos/autotest/repair/servo_detection/timeout' |
| 1162 | ).increment(fields=self._get_host_metrics_data()) |
| 1163 | logging.error('%s; %s', not_detected, str(e)) |
| 1164 | return None |
| 1165 | |
Otabek Kasimov | 120b6fa | 2020-07-03 00:15:27 -0700 | [diff] [blame] | 1166 | def _is_main_device_not_detected_on_servo_v4(self): |
| 1167 | """Check if servod cannot find main device on servo. |
| 1168 | |
| 1169 | The check based on reading servod logs for servo_v4. |
| 1170 | """ |
| 1171 | if not self._initial_instance_ts: |
| 1172 | # latest log not found |
| 1173 | return False |
| 1174 | logging.debug('latest log for servod created at %s', |
| 1175 | self._initial_instance_ts) |
| 1176 | try: |
| 1177 | log_created = calendar.timegm(time.strptime( |
| 1178 | self._initial_instance_ts, |
| 1179 | "%Y-%m-%d--%H-%M-%S.%f")) |
| 1180 | except ValueError as e: |
| 1181 | logging.debug('Cannot read time from log file name: %s', |
| 1182 | self._initial_instance_ts) |
| 1183 | return False |
| 1184 | min_time_created = calendar.timegm(time.gmtime()) |
| 1185 | if min_time_created > log_created + 3600: |
| 1186 | # the log file is old we cannot use it |
| 1187 | logging.debug('log file was created more than hour ago, too old') |
| 1188 | return False |
| 1189 | logging.debug('latest log was created not longer then 1 hour ago') |
| 1190 | |
| 1191 | # check if servod can detect main device by servo_v4 |
| 1192 | message = 'ERROR - No servo micro or CCD detected for board' |
| 1193 | cmd = ('cat /var/log/servod_%s/log.%s.INFO |grep "%s"' |
| 1194 | % (self.servo_port, self._initial_instance_ts, message)) |
| 1195 | result = self.run(cmd, ignore_status=True) |
| 1196 | if result.stdout.strip(): |
| 1197 | logging.info('Servod cannot detect main device on the servo; ' |
| 1198 | 'Can be caused by bad hardware of servo or ' |
| 1199 | 'issue on the DUT side.') |
| 1200 | return True |
| 1201 | logging.debug('The main device is detected') |
| 1202 | return False |
| 1203 | |
Otabek Kasimov | 1596349 | 2020-06-23 21:10:51 -0700 | [diff] [blame] | 1204 | def get_verify_state(self, tag): |
| 1205 | """Return the state of servo verifier. |
| 1206 | |
| 1207 | @returns: bool or None |
| 1208 | """ |
| 1209 | return self._repair_strategy.verifier_is_good(tag) |
| 1210 | |
| 1211 | def determine_servo_state(self): |
| 1212 | """Determine servo state based on the failed verifier. |
| 1213 | |
| 1214 | @returns: servo state value |
| 1215 | The state detecting based on first fail verifier or collecting of |
| 1216 | them. |
| 1217 | """ |
| 1218 | ssh = self.get_verify_state('servo_ssh') |
| 1219 | disk_space = self.get_verify_state('disk_space') |
| 1220 | start_servod = self.get_verify_state('servod_job') |
| 1221 | create_servo = self.get_verify_state('servod_connection') |
| 1222 | init_servo = self.get_verify_state('servod_control') |
| 1223 | pwr_button = self.get_verify_state('pwr_button') |
| 1224 | lid_open = self.get_verify_state('lid_open') |
| 1225 | ec_board = self.get_verify_state('ec_board') |
| 1226 | ccd_testlab = self.get_verify_state('ccd_testlab') |
| 1227 | |
| 1228 | if not ssh: |
| 1229 | return servo_constants.SERVO_STATE_NO_SSH |
| 1230 | |
Otabek Kasimov | 066bdb8 | 2020-08-12 15:57:44 -0700 | [diff] [blame^] | 1231 | if (start_servod == self.VERIFY_FAILED |
| 1232 | or create_servo == self.VERIFY_FAILED): |
| 1233 | # sometimes servo can start with out present servo |
Otabek Kasimov | c6f3041 | 2020-06-30 20:08:12 -0700 | [diff] [blame] | 1234 | if self.is_labstation(): |
| 1235 | if not self.servo_serial: |
| 1236 | return servo_constants.SERVO_STATE_WRONG_CONFIG |
| 1237 | if self._is_servo_device_connected( |
| 1238 | 'servo_v4', |
| 1239 | self.servo_serial) == False: |
| 1240 | return servo_constants.SERVO_STATE_NOT_CONNECTED |
| 1241 | elif self._is_servo_board_present_on_servo_v3() == False: |
| 1242 | return servo_constants.SERVO_STATE_NOT_CONNECTED |
Otabek Kasimov | 066bdb8 | 2020-08-12 15:57:44 -0700 | [diff] [blame^] | 1243 | |
| 1244 | if start_servod == self.VERIFY_FAILED: |
Otabek Kasimov | c6f3041 | 2020-06-30 20:08:12 -0700 | [diff] [blame] | 1245 | return servo_constants.SERVO_STATE_SERVOD_ISSUE |
| 1246 | |
Otabek Kasimov | 120b6fa | 2020-07-03 00:15:27 -0700 | [diff] [blame] | 1247 | if create_servo == self.VERIFY_FAILED: |
| 1248 | if (self.is_labstation() |
| 1249 | and self._is_main_device_not_detected_on_servo_v4()): |
| 1250 | servo_type = None |
| 1251 | if self.get_dut_host_info(): |
| 1252 | servo_type = self.get_dut_host_info().get_label_value( |
| 1253 | servo_constants.SERVO_TYPE_LABEL_PREFIX) |
| 1254 | if servo_type and 'servo_micro' in servo_type: |
| 1255 | serial = self.get_servo_micro_serial_number() |
| 1256 | logging.debug('servo_micro serial: %s', serial) |
| 1257 | if self._is_servo_device_detected('servo_micro', |
| 1258 | serial): |
| 1259 | return servo_constants.SERVO_STATE_BAD_RIBBON_CABLE |
| 1260 | # Device can be not detected because of DUT |
| 1261 | # TODO (otabek) update after b/159755652 and b/159754985 |
| 1262 | metrics.Counter( |
| 1263 | 'chromeos/autotest/repair/servo_state/needs_replacement' |
| 1264 | ).increment(fields=self._get_host_metrics_data()) |
| 1265 | elif not self.is_labstation(): |
| 1266 | # Here need logic to check if flex cable is connected |
| 1267 | pass |
| 1268 | |
Otabek Kasimov | 1596349 | 2020-06-23 21:10:51 -0700 | [diff] [blame] | 1269 | # one of the reason why servo can not initialized |
| 1270 | if ccd_testlab == self.VERIFY_FAILED: |
| 1271 | return servo_constants.SERVO_STATE_CCD_TESTLAB_ISSUE |
| 1272 | |
Otabek Kasimov | 120b6fa | 2020-07-03 00:15:27 -0700 | [diff] [blame] | 1273 | if (create_servo == self.VERIFY_FAILED |
| 1274 | or init_servo == self.VERIFY_FAILED): |
Otabek Kasimov | 1596349 | 2020-06-23 21:10:51 -0700 | [diff] [blame] | 1275 | return servo_constants.SERVO_STATE_SERVOD_ISSUE |
| 1276 | |
| 1277 | if pwr_button == self.VERIFY_FAILED: |
| 1278 | return servo_constants.SERVO_STATE_BAD_RIBBON_CABLE |
| 1279 | if lid_open == self.VERIFY_FAILED: |
| 1280 | return servo_constants.SERVO_STATE_LID_OPEN_FAILED |
| 1281 | if ec_board == self.VERIFY_FAILED: |
| 1282 | return servo_constants.SERVO_STATE_EC_BROKEN |
| 1283 | |
Otabek Kasimov | 1596349 | 2020-06-23 21:10:51 -0700 | [diff] [blame] | 1284 | metrics.Counter( |
| 1285 | 'chromeos/autotest/repair/unknown_servo_state' |
Otabek Kasimov | c6f3041 | 2020-06-30 20:08:12 -0700 | [diff] [blame] | 1286 | ).increment(fields=self._get_host_metrics_data()) |
Otabek Kasimov | 1596349 | 2020-06-23 21:10:51 -0700 | [diff] [blame] | 1287 | logging.info('We do not have special state for this failure yet :)') |
| 1288 | return servo_constants.SERVO_STATE_BROKEN |
| 1289 | |
Otabek Kasimov | cc9738e | 2020-02-14 16:17:15 -0800 | [diff] [blame] | 1290 | |
Richard Barnette | ea3e460 | 2016-06-10 12:36:41 -0700 | [diff] [blame] | 1291 | def make_servo_hostname(dut_hostname): |
| 1292 | """Given a DUT's hostname, return the hostname of its servo. |
| 1293 | |
| 1294 | @param dut_hostname: hostname of a DUT. |
| 1295 | |
| 1296 | @return hostname of the DUT's servo. |
| 1297 | |
| 1298 | """ |
| 1299 | host_parts = dut_hostname.split('.') |
| 1300 | host_parts[0] = host_parts[0] + '-servo' |
| 1301 | return '.'.join(host_parts) |
| 1302 | |
| 1303 | |
| 1304 | def servo_host_is_up(servo_hostname): |
Kevin Cheng | 5f2ba6c | 2016-09-28 10:20:05 -0700 | [diff] [blame] | 1305 | """Given a servo host name, return if it's up or not. |
Richard Barnette | ea3e460 | 2016-06-10 12:36:41 -0700 | [diff] [blame] | 1306 | |
| 1307 | @param servo_hostname: hostname of the servo host. |
| 1308 | |
| 1309 | @return True if it's up, False otherwise |
| 1310 | """ |
| 1311 | # Technically, this duplicates the SSH ping done early in the servo |
| 1312 | # proxy initialization code. However, this ping ends in a couple |
| 1313 | # seconds when if fails, rather than the 60 seconds it takes to decide |
| 1314 | # that an SSH ping has timed out. Specifically, that timeout happens |
| 1315 | # when our servo DNS name resolves, but there is no host at that IP. |
| 1316 | logging.info('Pinging servo host at %s', servo_hostname) |
| 1317 | ping_config = ping_runner.PingConfig( |
| 1318 | servo_hostname, count=3, |
| 1319 | ignore_result=True, ignore_status=True) |
| 1320 | return ping_runner.PingRunner().ping(ping_config).received > 0 |
| 1321 | |
| 1322 | |
Richard Barnette | e519dcd | 2016-08-15 17:37:17 -0700 | [diff] [blame] | 1323 | def _map_afe_board_to_servo_board(afe_board): |
| 1324 | """Map a board we get from the AFE to a servo appropriate value. |
| 1325 | |
| 1326 | Many boards are identical to other boards for servo's purposes. |
| 1327 | This function makes that mapping. |
| 1328 | |
| 1329 | @param afe_board string board name received from AFE. |
| 1330 | @return board we expect servo to have. |
| 1331 | |
| 1332 | """ |
| 1333 | KNOWN_SUFFIXES = ['-freon', '_freon', '_moblab', '-cheets'] |
| 1334 | BOARD_MAP = {'gizmo': 'panther'} |
| 1335 | mapped_board = afe_board |
| 1336 | if afe_board in BOARD_MAP: |
| 1337 | mapped_board = BOARD_MAP[afe_board] |
| 1338 | else: |
| 1339 | for suffix in KNOWN_SUFFIXES: |
| 1340 | if afe_board.endswith(suffix): |
| 1341 | mapped_board = afe_board[0:-len(suffix)] |
| 1342 | break |
| 1343 | if mapped_board != afe_board: |
| 1344 | logging.info('Mapping AFE board=%s to %s', afe_board, mapped_board) |
| 1345 | return mapped_board |
| 1346 | |
| 1347 | |
Prathmesh Prabhu | b481023 | 2018-09-07 13:24:08 -0700 | [diff] [blame] | 1348 | def get_servo_args_for_host(dut_host): |
Kevin Cheng | 5f2ba6c | 2016-09-28 10:20:05 -0700 | [diff] [blame] | 1349 | """Return servo data associated with a given DUT. |
Richard Barnette | ea3e460 | 2016-06-10 12:36:41 -0700 | [diff] [blame] | 1350 | |
Richard Barnette | ea3e460 | 2016-06-10 12:36:41 -0700 | [diff] [blame] | 1351 | @param dut_host Instance of `Host` on which to find the servo |
| 1352 | attributes. |
Prathmesh Prabhu | f605dd3 | 2018-08-28 17:09:04 -0700 | [diff] [blame] | 1353 | @return `servo_args` dict with host and an optional port. |
Richard Barnette | ea3e460 | 2016-06-10 12:36:41 -0700 | [diff] [blame] | 1354 | """ |
Prathmesh Prabhu | cba4429 | 2018-08-28 17:44:45 -0700 | [diff] [blame] | 1355 | info = dut_host.host_info_store.get() |
| 1356 | servo_args = {k: v for k, v in info.attributes.iteritems() |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 1357 | if k in servo_constants.SERVO_ATTR_KEYS} |
Richard Barnette | ea3e460 | 2016-06-10 12:36:41 -0700 | [diff] [blame] | 1358 | |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 1359 | if servo_constants.SERVO_PORT_ATTR in servo_args: |
Prathmesh Prabhu | cba4429 | 2018-08-28 17:44:45 -0700 | [diff] [blame] | 1360 | try: |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 1361 | servo_args[servo_constants.SERVO_PORT_ATTR] = int( |
| 1362 | servo_args[servo_constants.SERVO_PORT_ATTR]) |
Prathmesh Prabhu | cba4429 | 2018-08-28 17:44:45 -0700 | [diff] [blame] | 1363 | except ValueError: |
| 1364 | logging.error('servo port is not an int: %s', |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 1365 | servo_args[servo_constants.SERVO_PORT_ATTR]) |
Prathmesh Prabhu | cba4429 | 2018-08-28 17:44:45 -0700 | [diff] [blame] | 1366 | # Reset servo_args because we don't want to use an invalid port. |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 1367 | servo_args.pop(servo_constants.SERVO_HOST_ATTR, None) |
Prathmesh Prabhu | cba4429 | 2018-08-28 17:44:45 -0700 | [diff] [blame] | 1368 | |
| 1369 | if info.board: |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 1370 | servo_board = _map_afe_board_to_servo_board(info.board) |
| 1371 | servo_args[servo_constants.SERVO_BOARD_ATTR] = servo_board |
Nick Sanders | 2f3c985 | 2018-10-24 12:10:24 -0700 | [diff] [blame] | 1372 | if info.model: |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 1373 | servo_args[servo_constants.SERVO_MODEL_ATTR] = info.model |
| 1374 | return servo_args if servo_constants.SERVO_HOST_ATTR in servo_args else None |
Richard Barnette | ea3e460 | 2016-06-10 12:36:41 -0700 | [diff] [blame] | 1375 | |
| 1376 | |
Prathmesh Prabhu | efb1b48 | 2018-08-28 17:15:05 -0700 | [diff] [blame] | 1377 | def _tweak_args_for_ssp_moblab(servo_args): |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 1378 | if (servo_args[servo_constants.SERVO_HOST_ATTR] |
| 1379 | in ['localhost', '127.0.0.1']): |
| 1380 | servo_args[servo_constants.SERVO_HOST_ATTR] = _CONFIG.get_config_value( |
Prathmesh Prabhu | efb1b48 | 2018-08-28 17:15:05 -0700 | [diff] [blame] | 1381 | 'SSP', 'host_container_ip', type=str, default=None) |
| 1382 | |
| 1383 | |
Dan Shi | 023aae3 | 2016-05-25 11:13:01 -0700 | [diff] [blame] | 1384 | def create_servo_host(dut, servo_args, try_lab_servo=False, |
Gregory Nisbet | de13e2a | 2019-12-09 22:44:00 -0800 | [diff] [blame] | 1385 | try_servo_repair=False, dut_host_info=None): |
Kevin Cheng | 5f2ba6c | 2016-09-28 10:20:05 -0700 | [diff] [blame] | 1386 | """Create a ServoHost object for a given DUT, if appropriate. |
Dan Shi | 4d47852 | 2014-02-14 13:46:32 -0800 | [diff] [blame] | 1387 | |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 1388 | This function attempts to create and verify or repair a `ServoHost` |
| 1389 | object for a servo connected to the given `dut`, subject to various |
| 1390 | constraints imposed by the parameters: |
| 1391 | * When the `servo_args` parameter is not `None`, a servo |
| 1392 | host must be created, and must be checked with `repair()`. |
| 1393 | * Otherwise, if a servo exists in the lab and `try_lab_servo` is |
| 1394 | true: |
| 1395 | * If `try_servo_repair` is true, then create a servo host and |
| 1396 | check it with `repair()`. |
| 1397 | * Otherwise, if the servo responds to `ping` then create a |
| 1398 | servo host and check it with `verify()`. |
Fang Deng | e545abb | 2014-12-30 18:43:47 -0800 | [diff] [blame] | 1399 | |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 1400 | In cases where `servo_args` was not `None`, repair failure |
| 1401 | exceptions are passed back to the caller; otherwise, exceptions |
Richard Barnette | 07c2e1d | 2016-10-26 14:24:28 -0700 | [diff] [blame] | 1402 | are logged and then discarded. Note that this only happens in cases |
| 1403 | where we're called from a test (not special task) control file that |
| 1404 | has an explicit dependency on servo. In that case, we require that |
| 1405 | repair not write to `status.log`, so as to avoid polluting test |
| 1406 | results. |
| 1407 | |
| 1408 | TODO(jrbarnette): The special handling for servo in test control |
| 1409 | files is a thorn in my flesh; I dearly hope to see it cut out before |
| 1410 | my retirement. |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 1411 | |
| 1412 | Parameters for a servo host consist of a host name, port number, and |
| 1413 | DUT board, and are determined from one of these sources, in order of |
| 1414 | priority: |
Richard Barnette | ea3e460 | 2016-06-10 12:36:41 -0700 | [diff] [blame] | 1415 | * Servo attributes from the `dut` parameter take precedence over |
| 1416 | all other sources of information. |
| 1417 | * If a DNS entry for the servo based on the DUT hostname exists in |
| 1418 | the CrOS lab network, that hostname is used with the default |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 1419 | port and the DUT's board. |
Richard Barnette | ea3e460 | 2016-06-10 12:36:41 -0700 | [diff] [blame] | 1420 | * If no other options are found, the parameters will be taken |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 1421 | from the `servo_args` dict passed in from the caller. |
Richard Barnette | ea3e460 | 2016-06-10 12:36:41 -0700 | [diff] [blame] | 1422 | |
| 1423 | @param dut An instance of `Host` from which to take |
| 1424 | servo parameters (if available). |
| 1425 | @param servo_args A dictionary with servo parameters to use if |
| 1426 | they can't be found from `dut`. If this |
| 1427 | argument is supplied, unrepaired exceptions |
| 1428 | from `verify()` will be passed back to the |
| 1429 | caller. |
| 1430 | @param try_lab_servo If not true, servo host creation will be |
| 1431 | skipped unless otherwise required by the |
| 1432 | caller. |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 1433 | @param try_servo_repair If true, check a servo host with |
| 1434 | `repair()` instead of `verify()`. |
Otabek Kasimov | 8475cce | 2020-07-14 12:11:31 -0700 | [diff] [blame] | 1435 | @param dut_host_info: A HostInfo object of the DUT that connected |
| 1436 | to this servo. |
Dan Shi | 4d47852 | 2014-02-14 13:46:32 -0800 | [diff] [blame] | 1437 | |
| 1438 | @returns: A ServoHost object or None. See comments above. |
| 1439 | |
| 1440 | """ |
Richard Barnette | 07c2e1d | 2016-10-26 14:24:28 -0700 | [diff] [blame] | 1441 | servo_dependency = servo_args is not None |
Richard Barnette | 07c2e1d | 2016-10-26 14:24:28 -0700 | [diff] [blame] | 1442 | if dut is not None and (try_lab_servo or servo_dependency): |
Prathmesh Prabhu | b481023 | 2018-09-07 13:24:08 -0700 | [diff] [blame] | 1443 | servo_args_override = get_servo_args_for_host(dut) |
Richard Barnette | ea3e460 | 2016-06-10 12:36:41 -0700 | [diff] [blame] | 1444 | if servo_args_override is not None: |
Prathmesh Prabhu | efb1b48 | 2018-08-28 17:15:05 -0700 | [diff] [blame] | 1445 | if utils.in_moblab_ssp(): |
| 1446 | _tweak_args_for_ssp_moblab(servo_args_override) |
Prathmesh Prabhu | 88bf605 | 2018-08-28 16:21:26 -0700 | [diff] [blame] | 1447 | logging.debug( |
| 1448 | 'Overriding provided servo_args (%s) with arguments' |
| 1449 | ' determined from the host (%s)', |
| 1450 | servo_args, |
| 1451 | servo_args_override, |
| 1452 | ) |
Richard Barnette | ea3e460 | 2016-06-10 12:36:41 -0700 | [diff] [blame] | 1453 | servo_args = servo_args_override |
Prathmesh Prabhu | cba4429 | 2018-08-28 17:44:45 -0700 | [diff] [blame] | 1454 | |
Richard Barnette | ea3e460 | 2016-06-10 12:36:41 -0700 | [diff] [blame] | 1455 | if servo_args is None: |
Prathmesh Prabhu | 88bf605 | 2018-08-28 16:21:26 -0700 | [diff] [blame] | 1456 | logging.debug('No servo_args provided, and failed to find overrides.') |
Otabek Kasimov | a7ba91a | 2020-03-09 08:31:01 -0700 | [diff] [blame] | 1457 | if try_lab_servo or servo_dependency: |
Otabek Kasimov | 646812c | 2020-06-23 20:01:36 -0700 | [diff] [blame] | 1458 | return None, servo_constants.SERVO_STATE_MISSING_CONFIG |
Otabek Kasimov | a7ba91a | 2020-03-09 08:31:01 -0700 | [diff] [blame] | 1459 | else: |
| 1460 | # For regular test case which not required the servo |
| 1461 | return None, None |
| 1462 | |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 1463 | servo_hostname = servo_args.get(servo_constants.SERVO_HOST_ATTR) |
| 1464 | servo_port = servo_args.get(servo_constants.SERVO_PORT_ATTR) |
Otabek Kasimov | a7ba91a | 2020-03-09 08:31:01 -0700 | [diff] [blame] | 1465 | if not _is_servo_host_information_exist(servo_hostname, servo_port): |
| 1466 | logging.debug( |
| 1467 | 'Servo connection info missed hostname: %s , port: %s', |
| 1468 | servo_hostname, servo_port) |
Otabek Kasimov | 646812c | 2020-06-23 20:01:36 -0700 | [diff] [blame] | 1469 | return None, servo_constants.SERVO_STATE_MISSING_CONFIG |
Otabek Kasimov | a7ba91a | 2020-03-09 08:31:01 -0700 | [diff] [blame] | 1470 | if not is_servo_host_information_valid(servo_hostname, servo_port): |
| 1471 | logging.debug( |
| 1472 | 'Servo connection info is incorrect hostname: %s , port: %s', |
| 1473 | servo_hostname, servo_port) |
Garry Wang | 11b5e87 | 2020-03-11 15:14:08 -0700 | [diff] [blame] | 1474 | return None, servo_constants.SERVO_STATE_WRONG_CONFIG |
Richard Barnette | 07c2e1d | 2016-10-26 14:24:28 -0700 | [diff] [blame] | 1475 | if (not servo_dependency and not try_servo_repair and |
Otabek Kasimov | a7ba91a | 2020-03-09 08:31:01 -0700 | [diff] [blame] | 1476 | not servo_host_is_up(servo_hostname)): |
Prathmesh Prabhu | 88bf605 | 2018-08-28 16:21:26 -0700 | [diff] [blame] | 1477 | logging.debug('ServoHost is not up.') |
Otabek Kasimov | 646812c | 2020-06-23 20:01:36 -0700 | [diff] [blame] | 1478 | return None, servo_constants.SERVO_STATE_NO_SSH |
Prathmesh Prabhu | 88bf605 | 2018-08-28 16:21:26 -0700 | [diff] [blame] | 1479 | |
Garry Wang | ebc015b | 2019-06-06 17:45:06 -0700 | [diff] [blame] | 1480 | newhost = ServoHost(**servo_args) |
Garry Wang | ffbd216 | 2020-04-17 16:13:48 -0700 | [diff] [blame] | 1481 | |
Otabek Kasimov | e6df810 | 2020-07-21 20:15:25 -0700 | [diff] [blame] | 1482 | # Reset or reboot servo device only during AdminRepair tasks. |
| 1483 | if try_servo_repair: |
| 1484 | if newhost._is_locked: |
| 1485 | # Reset servo if the servo is locked, as we check if the servohost |
| 1486 | # is up, if the servohost is labstation and if the servohost is in |
| 1487 | # lab inside the locking logic. |
| 1488 | newhost.reset_servo() |
| 1489 | else: |
Garry Wang | 358aad4 | 2020-08-02 20:56:04 -0700 | [diff] [blame] | 1490 | try: |
| 1491 | newhost.reboot_servo_v3_on_need() |
Garry Wang | 1f0d533 | 2020-08-10 19:32:32 -0700 | [diff] [blame] | 1492 | except Exception as e: |
| 1493 | logging.info('[Non-critical] Unexpected error while trying to' |
| 1494 | ' reboot servo_v3, skipping the reboot; %s', e) |
Otabek Kasimov | e6df810 | 2020-07-21 20:15:25 -0700 | [diff] [blame] | 1495 | |
Otabek Kasimov | 2b50cdb | 2020-07-06 19:16:06 -0700 | [diff] [blame] | 1496 | if dut: |
| 1497 | newhost.set_dut_hostname(dut.hostname) |
Otabek Kasimov | 9e90ae1 | 2020-08-14 03:01:19 -0700 | [diff] [blame] | 1498 | if dut_host_info: |
| 1499 | newhost.set_dut_host_info(dut_host_info) |
Garry Wang | ffbd216 | 2020-04-17 16:13:48 -0700 | [diff] [blame] | 1500 | |
Otabek Kasimov | 8475cce | 2020-07-14 12:11:31 -0700 | [diff] [blame] | 1501 | if try_lab_servo or try_servo_repair: |
| 1502 | try: |
| 1503 | logging.info("Check and update servo firmware.") |
| 1504 | servo_updater.update_servo_firmware( |
| 1505 | newhost, |
| 1506 | force_update=False) |
| 1507 | except Exception as e: |
| 1508 | logging.error("Servo device update error: %s", e) |
| 1509 | |
Garry Wang | cdd27b2 | 2020-01-13 14:59:11 -0800 | [diff] [blame] | 1510 | try: |
| 1511 | newhost.restart_servod(quick_startup=True) |
| 1512 | except error.AutoservSSHTimeout: |
| 1513 | logging.warning("Restart servod failed due ssh connection " |
| 1514 | "to servohost timed out. This error is forgiven" |
| 1515 | " here, we will retry in servo repair process.") |
| 1516 | except error.AutoservRunError as e: |
| 1517 | logging.warning("Restart servod failed due to:\n%s\n" |
| 1518 | "This error is forgiven here, we will retry" |
| 1519 | " in servo repair process.", str(e)) |
Garry Wang | ebc015b | 2019-06-06 17:45:06 -0700 | [diff] [blame] | 1520 | |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 1521 | # Note that the logic of repair() includes everything done |
| 1522 | # by verify(). It's sufficient to call one or the other; |
| 1523 | # we don't need both. |
Richard Barnette | 07c2e1d | 2016-10-26 14:24:28 -0700 | [diff] [blame] | 1524 | if servo_dependency: |
| 1525 | newhost.repair(silent=True) |
Otabek Kasimov | a7ba91a | 2020-03-09 08:31:01 -0700 | [diff] [blame] | 1526 | return newhost, newhost.get_servo_state() |
Prathmesh Prabhu | 88bf605 | 2018-08-28 16:21:26 -0700 | [diff] [blame] | 1527 | |
| 1528 | if try_servo_repair: |
| 1529 | try: |
| 1530 | newhost.repair() |
| 1531 | except Exception: |
| 1532 | logging.exception('servo repair failed for %s', newhost.hostname) |
Richard Barnette | 9a26ad6 | 2016-06-10 12:03:08 -0700 | [diff] [blame] | 1533 | else: |
| 1534 | try: |
Prathmesh Prabhu | 88bf605 | 2018-08-28 16:21:26 -0700 | [diff] [blame] | 1535 | newhost.verify() |
Kevin Cheng | 5f2ba6c | 2016-09-28 10:20:05 -0700 | [diff] [blame] | 1536 | except Exception: |
Prathmesh Prabhu | 88bf605 | 2018-08-28 16:21:26 -0700 | [diff] [blame] | 1537 | logging.exception('servo verify failed for %s', newhost.hostname) |
Otabek Kasimov | a7ba91a | 2020-03-09 08:31:01 -0700 | [diff] [blame] | 1538 | return newhost, newhost.get_servo_state() |
Otabek Kasimov | 7267a7a | 2020-03-04 11:18:45 -0800 | [diff] [blame] | 1539 | |
| 1540 | |
Otabek Kasimov | a7ba91a | 2020-03-09 08:31:01 -0700 | [diff] [blame] | 1541 | def _is_servo_host_information_exist(hostname, port): |
Otabek Kasimov | 7267a7a | 2020-03-04 11:18:45 -0800 | [diff] [blame] | 1542 | if hostname is None or len(hostname.strip()) == 0: |
| 1543 | return False |
Otabek Kasimov | a7ba91a | 2020-03-09 08:31:01 -0700 | [diff] [blame] | 1544 | if port is None: |
Otabek Kasimov | 7267a7a | 2020-03-04 11:18:45 -0800 | [diff] [blame] | 1545 | return False |
Otabek Kasimov | a7ba91a | 2020-03-09 08:31:01 -0700 | [diff] [blame] | 1546 | if not type(port) is int: |
| 1547 | try: |
| 1548 | int(port) |
| 1549 | except ValueError: |
| 1550 | return False |
| 1551 | |
Otabek Kasimov | 7267a7a | 2020-03-04 11:18:45 -0800 | [diff] [blame] | 1552 | return True |
| 1553 | |
| 1554 | |
Otabek Kasimov | a7ba91a | 2020-03-09 08:31:01 -0700 | [diff] [blame] | 1555 | def is_servo_host_information_valid(hostname, port): |
Garry Wang | 9b8f234 | 2020-04-17 16:34:09 -0700 | [diff] [blame] | 1556 | """Check if provided servo attributes are valid. |
| 1557 | |
| 1558 | @param hostname Hostname of the servohost. |
| 1559 | @param port servo port number. |
| 1560 | |
| 1561 | @returns: A bool value to indicate if provided servo attribute valid. |
| 1562 | """ |
Otabek Kasimov | a7ba91a | 2020-03-09 08:31:01 -0700 | [diff] [blame] | 1563 | if not _is_servo_host_information_exist(hostname, port): |
Otabek Kasimov | 7267a7a | 2020-03-04 11:18:45 -0800 | [diff] [blame] | 1564 | return False |
| 1565 | # checking range and correct of the port |
Otabek Kasimov | a7ba91a | 2020-03-09 08:31:01 -0700 | [diff] [blame] | 1566 | port_int = int(port) |
Otabek Kasimov | 7267a7a | 2020-03-04 11:18:45 -0800 | [diff] [blame] | 1567 | if port_int < 1 or port_int > 65000: |
| 1568 | return False |
| 1569 | # we expecting host contain only latters, digits and '-' or '_' |
Otabek Kasimov | a7ba91a | 2020-03-09 08:31:01 -0700 | [diff] [blame] | 1570 | if not re.match('[a-zA-Z0-9-_\.]*$', hostname) or len(hostname) < 5: |
Otabek Kasimov | 7267a7a | 2020-03-04 11:18:45 -0800 | [diff] [blame] | 1571 | return False |
| 1572 | return True |