server/hosts/site_host.py

# Copyright (c) 2012 The Chromium OS Authors. All rights reserved.
# Use of this source code is governed by a BSD-style license that can be
# found in the LICENSE file.

import functools
import logging
import re
import socket
import subprocess
import time
import xmlrpclib

from autotest_lib.client.bin import utils
from autotest_lib.client.common_lib import error
from autotest_lib.client.common_lib import global_config
from autotest_lib.client.common_lib.cros import autoupdater
from autotest_lib.client.common_lib.cros import dev_server
from autotest_lib.client.common_lib.cros import retry
from autotest_lib.client.cros import constants
from autotest_lib.server import autoserv_parser
from autotest_lib.server import autotest
from autotest_lib.server import site_host_attributes
from autotest_lib.server.cros import servo
from autotest_lib.server.cros.dynamic_suite import constants as ds_constants
from autotest_lib.server.cros.dynamic_suite import tools
from autotest_lib.server.hosts import remote
from autotest_lib.site_utils.rpm_control_system import rpm_client

# Importing frontend.afe.models requires a full Autotest
# installation (with the Django modules), not just the source
# repository.  Most developers won't have the full installation, so
# the imports below will fail for them.
#
# The fix is to catch import exceptions, and set `models` to `None`
# on failure.  This has the side effect that
# SiteHost._get_board_from_afe() will fail:  That will manifest as
# failures during Repair jobs leaving the DUT as "Repair Failed".
# In practice, you can't test Repair jobs without a full
# installation, so that kind of failure isn't expected.
try:
    # pylint: disable=W0611
    from autotest_lib.frontend import setup_django_environment
    from autotest_lib.frontend.afe import models
except:
    models = None


def _make_servo_hostname(hostname):
    host_parts = hostname.split('.')
    host_parts[0] = host_parts[0] + '-servo'
    return '.'.join(host_parts)


def _get_lab_servo(target_hostname):
    """Instantiate a Servo for |target_hostname| in the lab.

    Assuming that |target_hostname| is a device in the CrOS test
    lab, create and return a Servo object pointed at the servo
    attached to that DUT.  The servo in the test lab is assumed
    to already have servod up and running on it.

    @param target_hostname: device whose servo we want to target.
    @return an appropriately configured Servo instance.
    """
    servo_host = _make_servo_hostname(target_hostname)
    if utils.host_is_in_lab_zone(servo_host):
        try:
            return servo.Servo(servo_host=servo_host)
        except: # pylint: disable=W0702
            # TODO(jrbarnette):  Long-term, if we can't get to
            # a servo in the lab, we want to fail, so we should
            # pass any exceptions along.  Short-term, we're not
            # ready to rely on servo, so we ignore failures.
            pass
    return None


def make_ssh_command(user='root', port=22, opts='', hosts_file=None,
                     connect_timeout=None, alive_interval=None):
    """Override default make_ssh_command to use options tuned for Chrome OS.

    Tuning changes:
      - ConnectTimeout=30; maximum of 30 seconds allowed for an SSH connection
      failure.  Consistency with remote_access.sh.

      - ServerAliveInterval=180; which causes SSH to ping connection every
      180 seconds. In conjunction with ServerAliveCountMax ensures that if the
      connection dies, Autotest will bail out quickly. Originally tried 60 secs,
      but saw frequent job ABORTS where the test completed successfully.

      - ServerAliveCountMax=3; consistency with remote_access.sh.

      - ConnectAttempts=4; reduce flakiness in connection errors; consistency
      with remote_access.sh.

      - UserKnownHostsFile=/dev/null; we don't care about the keys. Host keys
      change with every new installation, don't waste memory/space saving them.

      - SSH protocol forced to 2; needed for ServerAliveInterval.

    @param user User name to use for the ssh connection.
    @param port Port on the target host to use for ssh connection.
    @param opts Additional options to the ssh command.
    @param hosts_file Ignored.
    @param connect_timeout Ignored.
    @param alive_interval Ignored.
    """
    base_command = ('/usr/bin/ssh -a -x %s -o StrictHostKeyChecking=no'
                    ' -o UserKnownHostsFile=/dev/null -o BatchMode=yes'
                    ' -o ConnectTimeout=30 -o ServerAliveInterval=180'
                    ' -o ServerAliveCountMax=3 -o ConnectionAttempts=4'
                    ' -o Protocol=2 -l %s -p %d')
    return base_command % (opts, user, port)



def add_label_detector(label_function_list, label_list=None, label=None):
    """Decorator used to group functions together into the provided list.
    @param label_function_list: List of label detecting functions to add
                                decorated function to.
    @param label_list: List of detectable labels to add detectable labels to.
                       (Default: None)
    @param label: Label string that is detectable by this detection function
                  (Default: None)
    """
    def add_func(func):
        """
        @param func: The function to be added as a detector.
        """
        label_function_list.append(func)
        if label and label_list is not None:
            label_list.append(label)
        return func
    return add_func


class SiteHost(remote.RemoteHost):
    """Chromium OS specific subclass of Host."""

    _parser = autoserv_parser.autoserv_parser

    # Time to wait for new kernel to be marked successful after
    # auto update.
    _KERNEL_UPDATE_TIMEOUT = 120

    # Timeout values (in seconds) associated with various Chrome OS
    # state changes.
    #
    # In general, a good rule of thumb is that the timeout can be up
    # to twice the typical measured value on the slowest platform.
    # The times here have not necessarily been empirically tested to
    # meet this criterion.
    #
    # SLEEP_TIMEOUT:  Time to allow for suspend to memory.
    # RESUME_TIMEOUT: Time to allow for resume after suspend, plus
    #   time to restart the netwowrk.
    # BOOT_TIMEOUT: Time to allow for boot from power off.  Among
    #   other things, this must account for the 30 second dev-mode
    #   screen delay and time to start the network,
    # USB_BOOT_TIMEOUT: Time to allow for boot from a USB device,
    #   including the 30 second dev-mode delay and time to start the
    #   network,
    # SHUTDOWN_TIMEOUT: Time to allow for shut down.
    # REBOOT_TIMEOUT: Combination of shutdown and reboot times.
    # _INSTALL_TIMEOUT: Time to allow for chromeos-install.

    SLEEP_TIMEOUT = 2
    RESUME_TIMEOUT = 5
    BOOT_TIMEOUT = 45
    USB_BOOT_TIMEOUT = 150
    SHUTDOWN_TIMEOUT = 5
    REBOOT_TIMEOUT = SHUTDOWN_TIMEOUT + BOOT_TIMEOUT
    _INSTALL_TIMEOUT = 240

    # _USB_POWER_TIMEOUT: Time to allow for USB to power toggle ON and OFF.
    # _POWER_CYCLE_TIMEOUT: Time to allow for manual power cycle.
    _USB_POWER_TIMEOUT = 5
    _POWER_CYCLE_TIMEOUT = 10

    _DEFAULT_SERVO_URL_FORMAT = ('/static/servo-images/'
                                 '%(board)s_test_image.bin')

    # TODO(jrbarnette):  Servo repair is restricted to x86-alex,
    # because the existing servo client code won't work on other
    # boards.  http://crosbug.com/36973
    _SERVO_REPAIR_WHITELIST = [ 'x86-alex' ]


    _RPM_RECOVERY_BOARDS = global_config.global_config.get_config_value('CROS',
            'rpm_recovery_boards', type=str).split(',')

    _MAX_POWER_CYCLE_ATTEMPTS = 6
    _LAB_MACHINE_FILE = '/mnt/stateful_partition/.labmachine'
    _RPM_HOSTNAME_REGEX = ('chromeos[0-9]+(-row[0-9]+)?-rack[0-9]+[a-z]*-'
                           'host[0-9]+')
    _LIGHTSENSOR_FILES = ['in_illuminance0_input',
                          'in_illuminance0_raw',
                          'illuminance0_input']
    _LIGHTSENSOR_SEARCH_DIR = '/sys/bus/iio/devices'
    _LABEL_FUNCTIONS = []
    _DETECTABLE_LABELS = []
    label_decorator = functools.partial(add_label_detector, _LABEL_FUNCTIONS,
                                        _DETECTABLE_LABELS)

    # Constants used in ping_wait_up() and ping_wait_down().
    #
    # _PING_WAIT_COUNT is the approximate number of polling
    # cycles to use when waiting for a host state change.
    #
    # _PING_STATUS_DOWN and _PING_STATUS_UP are names used
    # for arguments to the internal _ping_wait_for_status()
    # method.
    _PING_WAIT_COUNT = 40
    _PING_STATUS_DOWN = False
    _PING_STATUS_UP = True

    # Allowed values for the power_method argument.

    # POWER_CONTROL_RPM: Passed as default arg for power_off/on/cycle() methods.
    # POWER_CONTROL_SERVO: Used in set_power() and power_cycle() methods.
    # POWER_CONTROL_MANUAL: Used in set_power() and power_cycle() methods.
    POWER_CONTROL_RPM = 'RPM'
    POWER_CONTROL_SERVO = 'servoj10'
    POWER_CONTROL_MANUAL = 'manual'

    POWER_CONTROL_VALID_ARGS = (POWER_CONTROL_RPM,
                                POWER_CONTROL_SERVO,
                                POWER_CONTROL_MANUAL)

    @staticmethod
    def get_servo_arguments(args_dict):
        """Extract servo options from `args_dict` and return the result.

        Take the provided dictionary of argument options and return
        a subset that represent standard arguments needed to
        construct a servo object for a host.  The intent is to
        provide standard argument processing from run_remote_tests
        for tests that require a servo to operate.

        Recommended usage:
        ~~~~~~~~
            args_dict = utils.args_to_dict(args)
            servo_args = hosts.SiteHost.get_servo_arguments(args_dict)
            host = hosts.create_host(machine, servo_args=servo_args)
        ~~~~~~~~

        @param args_dict Dictionary from which to extract the servo
          arguments.
        """
        servo_args = {}
        for arg in ('servo_host', 'servo_port'):
            if arg in args_dict:
                servo_args[arg] = args_dict[arg]
        return servo_args


    def _initialize(self, hostname, servo_args=None, *args, **dargs):
        """Initialize superclasses, and |self.servo|.

        For creating the host servo object, there are three
        possibilities:  First, if the host is a lab system known to
        have a servo board, we connect to that servo unconditionally.
        Second, if we're called from a control file that requires
        servo features for testing, it will pass settings for
        `servo_host`, `servo_port`, or both.  If neither of these
        cases apply, `self.servo` will be `None`.

        """
        super(SiteHost, self)._initialize(hostname=hostname,
                                          *args, **dargs)
        # self.env is a dictionary of environment variable settings
        # to be exported for commands run on the host.
        # LIBC_FATAL_STDERR_ can be useful for diagnosing certain
        # errors that might happen.
        self.env['LIBC_FATAL_STDERR_'] = '1'
        self._xmlrpc_proxy_map = {}
        self.servo = _get_lab_servo(hostname)
        if not self.servo and servo_args is not None:
            self.servo = servo.Servo(**servo_args)


    def get_repair_image_name(self):
        """Generate a image_name from variables in the global config.

        @returns a str of $board-version/$BUILD.

        """
        stable_version = global_config.global_config.get_config_value(
                'CROS', 'stable_cros_version')
        build_pattern = global_config.global_config.get_config_value(
                'CROS', 'stable_build_pattern')
        board = self._get_board_from_afe()
        if board is None:
            raise error.AutoservError('DUT has no board attribute, '
                                      'cannot be repaired.')
        return build_pattern % (board, stable_version)


    def clear_cros_version_labels_and_job_repo_url(self):
        """Clear cros_version labels and host attribute job_repo_url."""
        host_model = models.Host.objects.get(hostname=self.hostname)
        for label in host_model.labels.iterator():
            if not label.name.startswith(ds_constants.VERSION_PREFIX):
                continue
            label = models.Label.smart_get(label.name)
            label.host_set.remove(host_model)

        host_model.set_or_delete_attribute('job_repo_url', None)


    def machine_install(self, update_url=None, force_update=False,
                        local_devserver=False, repair=False):
        """Install the DUT.

        @param update_url: The url to use for the update
                pattern: http://$devserver:###/update/$build
                If update_url is None and repair is True we will install the
                stable image listed in global_config under
                CROS.stable_cros_version.
        @param force_update: Force an update even if the version installed
                is the same. Default:False
        @param local_devserver: Used by run_remote_test to allow people to
                use their local devserver. Default: False
        @param repair: Whether or not we are in repair mode. This adds special
                cases for repairing a machine like starting update_engine.
                Setting repair to True sets force_update to True as well.
                default: False
        @raises autoupdater.ChromiumOSError

        """
        if not update_url and self._parser.options.image:
            update_url = self._parser.options.image
        elif not update_url and not repair:
            raise autoupdater.ChromiumOSError(
                'Update failed. No update URL provided.')
        elif not update_url and repair:
            image_name = self.get_repair_image_name()
            devserver = dev_server.ImageServer.resolve(image_name)
            logging.info('Staging repair build: %s', image_name)
            devserver.trigger_download(image_name, synchronous=False)
            self.clear_cros_version_labels_and_job_repo_url()
            update_url = tools.image_url_pattern() % (devserver.url(),
                                                      image_name)


        # In case the system is in a bad state, we always reboot the machine
        # before machine_install.
        self.reboot(timeout=60, wait=True)

        if repair:
            self.run('stop update-engine; start update-engine')
            force_update = True
        # Attempt to update the system.
        updater = autoupdater.ChromiumOSUpdater(update_url, host=self,
                                                local_devserver=local_devserver)
        if updater.run_update(force_update):
            # Figure out active and inactive kernel.
            active_kernel, inactive_kernel = updater.get_kernel_state()

            # Ensure inactive kernel has higher priority than active.
            if (updater.get_kernel_priority(inactive_kernel)
                    < updater.get_kernel_priority(active_kernel)):
                raise autoupdater.ChromiumOSError(
                    'Update failed. The priority of the inactive kernel'
                    ' partition is less than that of the active kernel'
                    ' partition.')

            update_engine_log = '/var/log/update_engine.log'
            logging.info('Dumping %s', update_engine_log)
            self.run('cat %s' % update_engine_log)
            # Updater has returned successfully; reboot the host.
            self.reboot(timeout=60, wait=True)
            # Touch the lab machine file to leave a marker that distinguishes
            # this image from other test images.
            self.run('touch %s' % self._LAB_MACHINE_FILE)

            # Following the reboot, verify the correct version.
            updater.check_version()

            # Figure out newly active kernel.
            new_active_kernel, _ = updater.get_kernel_state()

            # Ensure that previously inactive kernel is now the active kernel.
            if new_active_kernel != inactive_kernel:
                raise autoupdater.ChromiumOSError(
                    'Update failed. New kernel partition is not active after'
                    ' boot.')

            host_attributes = site_host_attributes.HostAttributes(self.hostname)
            if host_attributes.has_chromeos_firmware:
                # Wait until tries == 0 and success, or until timeout.
                utils.poll_for_condition(
                    lambda: (updater.get_kernel_tries(new_active_kernel) == 0
                             and updater.get_kernel_success(new_active_kernel)),
                    exception=autoupdater.ChromiumOSError(
                        'Update failed. Timed out waiting for system to mark'
                        ' new kernel as successful.'),
                    timeout=self._KERNEL_UPDATE_TIMEOUT, sleep_interval=5)

            # Kick off the autoreboot script as the _LAB_MACHINE_FILE was
            # missing on the first boot.
            self.run('start autoreboot')

        # Clean up any old autotest directories which may be lying around.
        for path in global_config.global_config.get_config_value(
                'AUTOSERV', 'client_autodir_paths', type=list):
            self.run('rm -rf ' + path)


    def _get_board_from_afe(self):
        """Retrieve this host's board from its labels in the AFE.

        Looks for a host label of the form "board:<board>", and
        returns the "<board>" part of the label.  `None` is returned
        if there is not a single, unique label matching the pattern.

        @returns board from label, or `None`.
        """
        host_model = models.Host.objects.get(hostname=self.hostname)
        board_labels = filter(lambda l: l.name.startswith('board:'),
                              host_model.labels.all())
        board_name = None
        if len(board_labels) == 1:
            board_name = board_labels[0].name.split(':', 1)[1]
        elif len(board_labels) == 0:
            logging.error('Host %s does not have a board label.',
                          self.hostname)
        else:
            logging.error('Host %s has multiple board labels.',
                          self.hostname)
        return board_name


    def _install_repair(self):
        """Attempt to repair this host using upate-engine.

        If the host is up, try installing the DUT with a stable
        "repair" version of Chrome OS as defined in the global_config
        under CROS.stable_cros_version.

        @returns True if successful, False if update_engine failed.

        """
        if not self.is_up():
            return False

        logging.info('Attempting to reimage machine to repair image.')
        try:
            self.machine_install(repair=True)
        except autoupdater.ChromiumOSError:
            logging.info('Repair via install failed.')
            return False

        return True


    def _servo_repair(self, board):
        """Attempt to repair this host using an attached Servo.

        Re-install the OS on the DUT by 1) installing a test image
        on a USB storage device attached to the Servo board,
        2) booting that image in recovery mode, and then
        3) installing the image.

        """
        server = dev_server.ImageServer.devserver_url_for_servo(board)
        image = server + (self._DEFAULT_SERVO_URL_FORMAT %
                          { 'board': board })
        self.servo.install_recovery_image(image)
        if not self.wait_up(timeout=self.USB_BOOT_TIMEOUT):
            raise error.AutoservError('DUT failed to boot from USB'
                                      ' after %d seconds' %
                                        self.USB_BOOT_TIMEOUT)
        self.run('chromeos-install --yes',
                 timeout=self._INSTALL_TIMEOUT)
        self.servo.power_long_press()
        self.servo.set('usb_mux_sel1', 'servo_sees_usbkey')
        self.servo.power_short_press()
        if not self.wait_up(timeout=self.BOOT_TIMEOUT):
            raise error.AutoservError('DUT failed to reboot installed '
                                      'test image after %d seconds' %
                                        self.BOOT_TIMEOUT)


    def _powercycle_to_repair(self):
        """Utilize the RPM Infrastructure to bring the host back up.

        If the host is not up/repaired after the first powercycle we utilize
        auto fallback to the last good install by powercycling and rebooting the
        host 6 times.
        """
        logging.info('Attempting repair via RPM powercycle.')
        failed_cycles = 0
        self.power_cycle()
        while not self.wait_up(timeout=self.BOOT_TIMEOUT):
            failed_cycles += 1
            if failed_cycles >= self._MAX_POWER_CYCLE_ATTEMPTS:
                raise error.AutoservError('Powercycled host %s %d times; '
                                          'device did not come back online.' %
                                            (self.hostname, failed_cycles))
            self.power_cycle()
        if failed_cycles == 0:
            logging.info('Powercycling was successful first time.')
        else:
            logging.info('Powercycling was successful after %d failures.',
                         failed_cycles)


    def repair_full(self):
        """Repair a host for repair level NO_PROTECTION.

        This overrides the base class function for repair; it does
        not call back to the parent class, but instead offers a
        simplified implementation based on the capabilities in the
        Chrome OS test lab.

        Repair follows this sequence:
          1. If the DUT passes `self.verify()`, do nothing.
          2. If the DUT can be power-cycled via RPM, try to repair
             by power-cycling.

        As with the parent method, the last operation performed on
        the DUT must be to call `self.verify()`; if that call fails,
        the exception it raises is passed back to the caller.
        """
        try:
            self.verify()
        except:
            host_board = self._get_board_from_afe()
            if host_board is None:
                logging.error('host %s has no board; failing repair',
                              self.hostname)
                raise

            reimage_success = self._install_repair()
            # TODO(scottz): All repair pathways should be executed until we've
            # exhausted all options. Below we favor servo over powercycle when
            # we really should be falling back to power if servo fails.
            if (not reimage_success and self.servo and
                    host_board in self._SERVO_REPAIR_WHITELIST):
                self._servo_repair(host_board)
            elif (self.has_power() and
                  host_board in self._RPM_RECOVERY_BOARDS):
                self._powercycle_to_repair()
            else:
                logging.error('host %s has no servo and no RPM control; '
                              'failing repair', self.hostname)
                raise
            self.verify()


    def close(self):
        super(SiteHost, self).close()
        self.xmlrpc_disconnect_all()


    def cleanup(self):
        client_at = autotest.Autotest(self)
        self.run('rm -f %s' % constants.CLEANUP_LOGS_PAUSED_FILE)
        try:
            client_at.run_static_method('autotest_lib.client.cros.cros_ui',
                                        '_clear_login_prompt_state')
            self.run('restart ui')
            client_at.run_static_method('autotest_lib.client.cros.cros_ui',
                                        '_wait_for_login_prompt')
        except (error.AutotestRunError, error.AutoservRunError):
            logging.warn('Unable to restart ui, rebooting device.')
            # Since restarting the UI fails fall back to normal Autotest
            # cleanup routines, i.e. reboot the machine.
            super(SiteHost, self).cleanup()


    # TODO (sbasi) crosbug.com/35656
    # Renamed the sitehost cleanup method so we don't go down this pathway.
    # def cleanup(self):
    def cleanup_poweron(self):
        """Special cleanup method to make sure hosts always get power back."""
        super(SiteHost, self).cleanup()
        if self.has_power():
            try:
                self.power_on()
            except rpm_client.RemotePowerException:
                # If cleanup has completed but there was an issue with the RPM
                # Infrastructure, log an error message rather than fail cleanup
                logging.error('Failed to turn Power On for this host after '
                              'cleanup through the RPM Infrastructure.')


    def reboot(self, **dargs):
        """
        This function reboots the site host. The more generic
        RemoteHost.reboot() performs sync and sleeps for 5
        seconds. This is not necessary for Chrome OS devices as the
        sync should be finished in a short time during the reboot
        command.
        """
        if 'reboot_cmd' not in dargs:
            dargs['reboot_cmd'] = ('((reboot & sleep 10; reboot -f &)'
                                   ' </dev/null >/dev/null 2>&1 &)')
        # Enable fastsync to avoid running extra sync commands before reboot.
        if 'fastsync' not in dargs:
            dargs['fastsync'] = True
        super(SiteHost, self).reboot(**dargs)


    def verify_software(self):
        """Verify working software on a Chrome OS system.

        Tests for the following conditions:
         1. All conditions tested by the parent version of this
            function.
         2. Sufficient space in /mnt/stateful_partition.
         3. update_engine answers a simple status request over DBus.

        """
        super(SiteHost, self).verify_software()
        self.check_diskspace(
            '/mnt/stateful_partition',
            global_config.global_config.get_config_value(
                'SERVER', 'gb_diskspace_required', type=int,
                default=20))
        self.run('update_engine_client --status')
        # Makes sure python is present, loads and can use built in functions.
        # We have seen cases where importing cPickle fails with undefined
        # symbols in cPickle.so.
        self.run('python -c "import cPickle"')


    def xmlrpc_connect(self, command, port, command_name=None,
                       ready_test_name=None, timeout_seconds=10):
        """Connect to an XMLRPC server on the host.

        The `command` argument should be a simple shell command that
        starts an XMLRPC server on the given `port`.  The command
        must not daemonize, and must terminate cleanly on SIGTERM.
        The command is started in the background on the host, and a
        local XMLRPC client for the server is created and returned
        to the caller.

        Note that the process of creating an XMLRPC client makes no
        attempt to connect to the remote server; the caller is
        responsible for determining whether the server is running
        correctly, and is ready to serve requests.

        Optionally, the caller can pass ready_test_name, a string
        containing the name of a method to call on the proxy.  This
        method should take no parameters and return successfully only
        when the server is ready to process client requests.  When
        ready_test_name is set, xmlrpc_connect will block until the
        proxy is ready, and throw a TestError if the server isn't
        ready by timeout_seconds.

        @param command Shell command to start the server.
        @param port Port number on which the server is expected to
                    be serving.
        @param command_name String to use as input to `pkill` to
            terminate the XMLRPC server on the host.
        @param ready_test_name String containing the name of a
            method defined on the XMLRPC server.
        @param timeout_seconds Number of seconds to wait
            for the server to become 'ready.'  Will throw a
            TestFail error if server is not ready in time.

        """
        self.xmlrpc_disconnect(port)

        # Chrome OS on the target closes down most external ports
        # for security.  We could open the port, but doing that
        # would conflict with security tests that check that only
        # expected ports are open.  So, to get to the port on the
        # target we use an ssh tunnel.
        local_port = utils.get_unused_port()
        tunnel_options = '-n -N -q -L %d:localhost:%d' % (local_port, port)
        ssh_cmd = make_ssh_command(opts=tunnel_options)
        tunnel_cmd = '%s %s' % (ssh_cmd, self.hostname)
        logging.debug('Full tunnel command: %s', tunnel_cmd)
        tunnel_proc = subprocess.Popen(tunnel_cmd, shell=True, close_fds=True)
        logging.debug('Started XMLRPC tunnel, local = %d'
                      ' remote = %d, pid = %d',
                      local_port, port, tunnel_proc.pid)

        # Start the server on the host.  Redirection in the command
        # below is necessary, because 'ssh' won't terminate until
        # background child processes close stdin, stdout, and
        # stderr.
        remote_cmd = '( %s ) </dev/null >/dev/null 2>&1 & echo $!' % command
        remote_pid = self.run(remote_cmd).stdout.rstrip('\n')
        logging.debug('Started XMLRPC server on host %s, pid = %s',
                      self.hostname, remote_pid)

        self._xmlrpc_proxy_map[port] = (command_name, tunnel_proc)
        rpc_url = 'http://localhost:%d' % local_port
        proxy = xmlrpclib.ServerProxy(rpc_url, allow_none=True)
        if ready_test_name is not None:
            @retry.retry((socket.error, xmlrpclib.ProtocolError),
                         timeout_min=timeout_seconds/60.0,
                         delay_sec=0.1)
            def ready_test():
                """ Call proxy.ready_test_name(). """
                getattr(proxy, ready_test_name)()
            successful = False
            try:
                logging.info('Waiting %d seconds for XMLRPC server '
                             'to start.', timeout_seconds)
                ready_test()
                successful = True
            except retry.TimeoutException:
                raise error.TestError('Unable to start XMLRPC server after '
                                      '%d seconds.' % timeout_seconds)
            finally:
                if not successful:
                    logging.error('Failed to start XMLRPC server.')
                    self.xmlrpc_disconnect(port)
        logging.info('XMLRPC server started successfully.')
        return proxy

    def xmlrpc_disconnect(self, port):
        """Disconnect from an XMLRPC server on the host.

        Terminates the remote XMLRPC server previously started for
        the given `port`.  Also closes the local ssh tunnel created
        for the connection to the host.  This function does not
        directly alter the state of a previously returned XMLRPC
        client object; however disconnection will cause all
        subsequent calls to methods on the object to fail.

        This function does nothing if requested to disconnect a port
        that was not previously connected via `self.xmlrpc_connect()`

        @param port Port number passed to a previous call to
                    `xmlrpc_connect()`
        """
        if port not in self._xmlrpc_proxy_map:
            return
        entry = self._xmlrpc_proxy_map[port]
        remote_name = entry[0]
        tunnel_proc = entry[1]
        if remote_name:
            # We use 'pkill' to find our target process rather than
            # a PID, because the host may have rebooted since
            # connecting, and we don't want to kill an innocent
            # process with the same PID.
            #
            # 'pkill' helpfully exits with status 1 if no target
            # process  is found, for which run() will throw an
            # exception.  We don't want that, so we the ignore
            # status.
            self.run("pkill -f '%s'" % remote_name, ignore_status=True)

        if tunnel_proc.poll() is None:
            tunnel_proc.terminate()
            logging.debug('Terminated tunnel, pid %d', tunnel_proc.pid)
        else:
            logging.debug('Tunnel pid %d terminated early, status %d',
                          tunnel_proc.pid, tunnel_proc.returncode)
        del self._xmlrpc_proxy_map[port]


    def xmlrpc_disconnect_all(self):
        """Disconnect all known XMLRPC proxy ports."""
        for port in self._xmlrpc_proxy_map.keys():
            self.xmlrpc_disconnect(port)


    def _ping_check_status(self, status):
        """Ping the host once, and return whether it has a given status.

        @param status Check the ping status against this value.
        @return True iff `status` and the result of ping are the same
                (i.e. both True or both False).

        """
        ping_val = utils.ping(self.hostname, tries=1, deadline=1)
        return not (status ^ (ping_val == 0))

    def _ping_wait_for_status(self, status, timeout):
        """Wait for the host to have a given status (UP or DOWN).

        Status is checked by polling.  Polling will not last longer
        than the number of seconds in `timeout`.  The polling
        interval will be long enough that only approximately
        _PING_WAIT_COUNT polling cycles will be executed, subject
        to a maximum interval of about one minute.

        @param status Waiting will stop immediately if `ping` of the
                      host returns this status.
        @param timeout Poll for at most this many seconds.
        @return True iff the host status from `ping` matched the
                requested status at the time of return.

        """
        # _ping_check_status() takes about 1 second, hence the
        # "- 1" in the formula below.
        poll_interval = min(int(timeout / self._PING_WAIT_COUNT), 60) - 1
        end_time = time.time() + timeout
        while time.time() <= end_time:
            if self._ping_check_status(status):
                return True
            if poll_interval > 0:
                time.sleep(poll_interval)

        # The last thing we did was sleep(poll_interval), so it may
        # have been too long since the last `ping`.  Check one more
        # time, just to be sure.
        return self._ping_check_status(status)

    def ping_wait_up(self, timeout):
        """Wait for the host to respond to `ping`.

        N.B.  This method is not a reliable substitute for
        `wait_up()`, because a host that responds to ping will not
        necessarily respond to ssh.  This method should only be used
        if the target DUT can be considered functional even if it
        can't be reached via ssh.

        @param timeout Minimum time to allow before declaring the
                       host to be non-responsive.
        @return True iff the host answered to ping before the timeout.

        """
        return self._ping_wait_for_status(self._PING_STATUS_UP, timeout)

    def ping_wait_down(self, timeout):
        """Wait until the host no longer responds to `ping`.

        This function can be used as a slightly faster version of
        `wait_down()`, by avoiding potentially long ssh timeouts.

        @param timeout Minimum time to allow for the host to become
                       non-responsive.
        @return True iff the host quit answering ping before the
                timeout.

        """
        return self._ping_wait_for_status(self._PING_STATUS_DOWN, timeout)

    def test_wait_for_sleep(self):
        """Wait for the client to enter low-power sleep mode.

        The test for "is asleep" can't distinguish a system that is
        powered off; to confirm that the unit was asleep, it is
        necessary to force resume, and then call
        `test_wait_for_resume()`.

        This function is expected to be called from a test as part
        of a sequence like the following:

        ~~~~~~~~
            boot_id = host.get_boot_id()
            # trigger sleep on the host
            host.test_wait_for_sleep()
            # trigger resume on the host
            host.test_wait_for_resume(boot_id)
        ~~~~~~~~

        @exception TestFail The host did not go to sleep within
                            the allowed time.
        """
        if not self.ping_wait_down(timeout=self.SLEEP_TIMEOUT):
            raise error.TestFail(
                'client failed to sleep after %d seconds' %
                    self.SLEEP_TIMEOUT)


    def test_wait_for_resume(self, old_boot_id):
        """Wait for the client to resume from low-power sleep mode.

        The `old_boot_id` parameter should be the value from
        `get_boot_id()` obtained prior to entering sleep mode.  A
        `TestFail` exception is raised if the boot id changes.

        See @ref test_wait_for_sleep for more on this function's
        usage.

        @param old_boot_id A boot id value obtained before the
                               target host went to sleep.

        @exception TestFail The host did not respond within the
                            allowed time.
        @exception TestFail The host responded, but the boot id test
                            indicated a reboot rather than a sleep
                            cycle.
        """
        if not self.wait_up(timeout=self.RESUME_TIMEOUT):
            raise error.TestFail(
                'client failed to resume from sleep after %d seconds' %
                    self.RESUME_TIMEOUT)
        else:
            new_boot_id = self.get_boot_id()
            if new_boot_id != old_boot_id:
                raise error.TestFail(
                    'client rebooted, but sleep was expected'
                    ' (old boot %s, new boot %s)'
                        % (old_boot_id, new_boot_id))


    def test_wait_for_shutdown(self):
        """Wait for the client to shut down.

        The test for "has shut down" can't distinguish a system that
        is merely asleep; to confirm that the unit was down, it is
        necessary to force boot, and then call test_wait_for_boot().

        This function is expected to be called from a test as part
        of a sequence like the following:

        ~~~~~~~~
            boot_id = host.get_boot_id()
            # trigger shutdown on the host
            host.test_wait_for_shutdown()
            # trigger boot on the host
            host.test_wait_for_boot(boot_id)
        ~~~~~~~~

        @exception TestFail The host did not shut down within the
                            allowed time.
        """
        if not self.ping_wait_down(timeout=self.SHUTDOWN_TIMEOUT):
            raise error.TestFail(
                'client failed to shut down after %d seconds' %
                    self.SHUTDOWN_TIMEOUT)


    def test_wait_for_boot(self, old_boot_id=None):
        """Wait for the client to boot from cold power.

        The `old_boot_id` parameter should be the value from
        `get_boot_id()` obtained prior to shutting down.  A
        `TestFail` exception is raised if the boot id does not
        change.  The boot id test is omitted if `old_boot_id` is not
        specified.

        See @ref test_wait_for_shutdown for more on this function's
        usage.

        @param old_boot_id A boot id value obtained before the
                               shut down.

        @exception TestFail The host did not respond within the
                            allowed time.
        @exception TestFail The host responded, but the boot id test
                            indicated that there was no reboot.
        """
        if not self.wait_up(timeout=self.REBOOT_TIMEOUT):
            raise error.TestFail(
                'client failed to reboot after %d seconds' %
                    self.REBOOT_TIMEOUT)
        elif old_boot_id:
            if self.get_boot_id() == old_boot_id:
                raise error.TestFail(
                    'client is back up, but did not reboot'
                    ' (boot %s)' % old_boot_id)


    @staticmethod
    def check_for_rpm_support(hostname):
        """For a given hostname, return whether or not it is powered by an RPM.

        @return None if this host does not follows the defined naming format
                for RPM powered DUT's in the lab. If it does follow the format,
                it returns a regular expression MatchObject instead.
        """
        return re.match(SiteHost._RPM_HOSTNAME_REGEX, hostname)


    def has_power(self):
        """For this host, return whether or not it is powered by an RPM.

        @return True if this host is in the CROS lab and follows the defined
                naming format.
        """
        return SiteHost.check_for_rpm_support(self.hostname)


    def _set_power(self, state, power_method):
        """Sets the power to the host via RPM, Servo or manual.

        @param state Specifies which power state to set to DUT
        @param power_method Specifies which method of power control to
                            use. By default "RPM" will be used. Valid values
                            are the strings "RPM", "manual", "servoj10".

        """
        ACCEPTABLE_STATES = ['ON', 'OFF']

        if state.upper() not in ACCEPTABLE_STATES:
            raise error.TestError('State must be one of: %s.'
                                   % (ACCEPTABLE_STATES,))

        if power_method == self.POWER_CONTROL_SERVO:
            logging.info('Setting servo port J10 to %s', state)
            self.servo.set('prtctl3_pwren', state.lower())
            time.sleep(self._USB_POWER_TIMEOUT)
        elif power_method == self.POWER_CONTROL_MANUAL:
            logging.info('You have %d seconds to set the AC power to %s.',
                         self._POWER_CYCLE_TIMEOUT, state)
            time.sleep(self._POWER_CYCLE_TIMEOUT)
        else:
            if not self.has_power():
                raise error.TestFail('DUT does not have RPM connected.')
            rpm_client.set_power(self.hostname, state.upper())


    def power_off(self, power_method=POWER_CONTROL_RPM):
        """Turn off power to this host via RPM, Servo or manual.

        @param power_method Specifies which method of power control to
                            use. By default "RPM" will be used. Valid values
                            are the strings "RPM", "manual", "servoj10".

        """
        self._set_power('OFF', power_method)


    def power_on(self, power_method=POWER_CONTROL_RPM):
        """Turn on power to this host via RPM, Servo or manual.

        @param power_method Specifies which method of power control to
                            use. By default "RPM" will be used. Valid values
                            are the strings "RPM", "manual", "servoj10".

        """
        self._set_power('ON', power_method)


    def power_cycle(self, power_method=POWER_CONTROL_RPM):
        """Cycle power to this host by turning it OFF, then ON.

        @param power_method Specifies which method of power control to
                            use. By default "RPM" will be used. Valid values
                            are the strings "RPM", "manual", "servoj10".

        """
        if power_method in (self.POWER_CONTROL_SERVO,
                            self.POWER_CONTROL_MANUAL):
            self.power_off(power_method=power_method)
            time.sleep(self._POWER_CYCLE_TIMEOUT)
            self.power_on(power_method=power_method)
        else:
            rpm_client.set_power(self.hostname, 'CYCLE')


    def get_platform(self):
        """Determine the correct platform label for this host.

        @returns a string representing this host's platform.
        """
        crossystem = utils.Crossystem(self)
        crossystem.init()
        # Extract fwid value and use the leading part as the platform id.
        # fwid generally follow the format of {platform}.{firmware version}
        # Example: Alex.X.YYY.Z or Google_Alex.X.YYY.Z
        platform = crossystem.fwid().split('.')[0].lower()
        # Newer platforms start with 'Google_' while the older ones do not.
        return platform.replace('google_', '')


    @label_decorator()
    def get_board(self):
        """Determine the correct board label for this host.

        @returns a string representing this host's board.
        """
        release_info = utils.parse_cmd_output('cat /etc/lsb-release',
                                              run_method=self.run)
        board = release_info['CHROMEOS_RELEASE_BOARD']
        # Devices in the lab generally have the correct board name but our own
        # development devices have {board_name}-signed-{key_type}. The board
        # name may also begin with 'x86-' which we need to keep.
        if 'x86' not in board:
            return 'board:%s' % board.split('-')[0]
        return 'board:%s' % '-'.join(board.split('-')[0:2])


    @label_decorator('lightsensor')
    def has_lightsensor(self):
        """Determine the correct board label for this host.

        @returns the string 'lightsensor' if this host has a lightsensor or
                 None if it does not.
        """
        search_cmd = "find -L %s -maxdepth 4 | egrep '%s'" % (
            self._LIGHTSENSOR_SEARCH_DIR, '|'.join(self._LIGHTSENSOR_FILES))
        try:
            # Run the search cmd following the symlinks. Stderr_tee is set to
            # None as there can be a symlink loop, but the command will still
            # execute correctly with a few messages printed to stderr.
            self.run(search_cmd, stdout_tee=None, stderr_tee=None)
            return 'lightsensor'
        except error.AutoservRunError:
            # egrep exited with a return code of 1 meaning none of the possible
            # lightsensor files existed.
            return None


    @label_decorator('bluetooth')
    def has_bluetooth(self):
        """Determine the correct board label for this host.

        @returns the string 'bluetooth' if this host has bluetooth or
                 None if it does not.
        """
        try:
            self.run('test -d /sys/class/bluetooth/hci0')
            # test exited with a return code of 0.
            return 'bluetooth'
        except error.AutoservRunError:
            # test exited with a return code 1 meaning the directory did not
            # exist.
            return None


    def get_labels(self):
        """Return a list of labels for this given host.

        This is the main way to retrieve all the automatic labels for a host
        as it will run through all the currently implemented label functions.
        """
        labels = []
        for label_function in self._LABEL_FUNCTIONS:
            label = label_function(self)
            if label:
                labels.append(label)
        return labels