Use devil's adb with each bisect iteration, also start/stop daemons.

scripts/slave/recipes/bisection/android_bisect.py

 # Copyright (c) 2015 The Chromium Authors. All rights reserved.
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 
 from recipe_engine.types import freeze
 
 DEPS = [
     'build/adb',
     'auto_bisect',
     'bisect_tester',
@@ skipping 104 matching lines @@
   update_step = api.auto_bisect.ensure_checkout()
   api.path.c.dynamic_paths['catapult'] = api.m.auto_bisect.working_dir.join(
       'catapult')
   api.chromium_android.clean_local_files()
 
   bot_db = api.chromium_tests.create_bot_db_from_master_dict(mastername,
                                                              master_dict)
 
   api.chromium_android.use_devil_adb()
 
+  # Make sure the bisect build and individual test runs are wrapped.
+  api.auto_bisect.build_context_mgr = (
+      api.chromium_android.android_build_wrapper)
+  api.auto_bisect.test_context_mgr = api.chromium_android.android_test_wrapper
+
   api.auto_bisect.start_try_job(api, update_step=update_step, bot_db=bot_db,
                                 do_not_nest_wait_for_revision=True)
 
 
 def GenTests(api):
   config_json_main = {
       'command': ('./tools/perf/run_benchmark -v --browser=android-chrome '
                 'sunspider'),
       'max_time_minutes': '25',
       'client_operation_id': '12345726327327',
@@ skipping 568 matching lines @@
     # multiple_device_status
     api.step_data('Debug Info', retcode=1) +
     api.override_step_data('device_status (3)',
         api.json.output(working_device)) +
     api.step_data('Expanding revision range.for revisions %s:%s (2)' % (
             good_revision_hash, bad_revision_hash),
         stdout=api.json.output([[bad_revision_hash, 'ignored'], [
             good_revision_hash, 'ignored']])) +
     api.step_data('Post bisect results',
         stdout=api.json.output({'status_code': 200})))