u-boot/test/py/conftest.py

521 lines
16 KiB
Python
Raw Normal View History

# Copyright (c) 2015 Stephen Warren
# Copyright (c) 2015-2016, NVIDIA CORPORATION. All rights reserved.
#
# SPDX-License-Identifier: GPL-2.0
# Implementation of pytest run-time hook functions. These are invoked by
# pytest at certain points during operation, e.g. startup, for each executed
# test, at shutdown etc. These hooks perform functions such as:
# - Parsing custom command-line options.
# - Pullilng in user-specified board configuration.
# - Creating the U-Boot console test fixture.
# - Creating the HTML log file.
# - Monitoring each test's results.
# - Implementing custom pytest markers.
import atexit
import errno
import os
import os.path
import pytest
from _pytest.runner import runtestprotocol
import ConfigParser
import re
import StringIO
import sys
# Globals: The HTML log file, and the connection to the U-Boot console.
log = None
console = None
def mkdir_p(path):
"""Create a directory path.
This includes creating any intermediate/parent directories. Any errors
caused due to already extant directories are ignored.
Args:
path: The directory path to create.
Returns:
Nothing.
"""
try:
os.makedirs(path)
except OSError as exc:
if exc.errno == errno.EEXIST and os.path.isdir(path):
pass
else:
raise
def pytest_addoption(parser):
"""pytest hook: Add custom command-line options to the cmdline parser.
Args:
parser: The pytest command-line parser.
Returns:
Nothing.
"""
parser.addoption('--build-dir', default=None,
help='U-Boot build directory (O=)')
parser.addoption('--result-dir', default=None,
help='U-Boot test result/tmp directory')
parser.addoption('--persistent-data-dir', default=None,
help='U-Boot test persistent generated data directory')
parser.addoption('--board-type', '--bd', '-B', default='sandbox',
help='U-Boot board type')
parser.addoption('--board-identity', '--id', default='na',
help='U-Boot board identity/instance')
parser.addoption('--build', default=False, action='store_true',
help='Compile U-Boot before running tests')
test/py: support running sandbox under gdbserver Implement command--line option --gdbserver COMM, which does two things: a) Run the sandbox process under gdbserver, using COMM as gdbserver's communication channel. b) Disables all timeouts, so that if U-Boot is halted under the debugger, tests don't fail. If the user gives up in the middle of a debugging session, they can simply CTRL-C the test script to abort it. This allows easy debugging of test failures without having to manually re-create the failure conditions. Usage is: Window 1: ./test/py/test.py --bd sandbox --gdbserver localhost:1234 Window 2: gdb ./build-sandbox/u-boot -ex 'target remote localhost:1234' When using this option, it likely makes sense to use pytest's -k option to limit the set of tests that are executed. Simply running U-Boot directly under gdb (rather than gdbserver) was also considered. However, this was rejected because: a) gdb's output would then be processed by the test script, and likely confuse it causing false failures. b) pytest by default hides stdout from tests, which would prevent the user from interacting with gdb. While gdb can be told to redirect the debugee's stdio to a separate PTY, this would appear to leave gdb's stdio directed at the test scripts and the debugee's stdio directed elsewhere, which is the opposite of the desired effect. Perhaps some complicated PTY muxing and process hierarchy could invert this. However, the current scheme is simple to implement and use, so it doesn't seem worth complicating matters. c) Using gdbserver allows arbitrary debuggers to be used, even those with a GUI. If the test scripts invoked the debugger themselves, they'd have to know how to execute arbitary applications. While the user could hide this all in a wrapper script, this feels like extra complication. An interesting future idea might be a --gdb-screen option, which could spawn both U-Boot and gdb separately, and spawn the screen into a newly created window under screen. Similar options could be envisaged for creating a new xterm/... too. --gdbserver currently only supports sandbox, and not real hardware. That's primarily because the test hooks are responsible for all aspects of hardware control, so there's nothing for the test scripts themselves can do to enable gdbserver on real hardware. We might consider introducing a separate --disable-timeouts option to support use of debuggers on real hardware, and having --gdbserver imply that option. Signed-off-by: Stephen Warren <swarren@nvidia.com>
2016-02-04 23:11:50 +00:00
parser.addoption('--gdbserver', default=None,
help='Run sandbox under gdbserver. The argument is the channel '+
'over which gdbserver should communicate, e.g. localhost:1234')
def pytest_configure(config):
"""pytest hook: Perform custom initialization at startup time.
Args:
config: The pytest configuration.
Returns:
Nothing.
"""
global log
global console
global ubconfig
test_py_dir = os.path.dirname(os.path.abspath(__file__))
source_dir = os.path.dirname(os.path.dirname(test_py_dir))
board_type = config.getoption('board_type')
board_type_filename = board_type.replace('-', '_')
board_identity = config.getoption('board_identity')
board_identity_filename = board_identity.replace('-', '_')
build_dir = config.getoption('build_dir')
if not build_dir:
build_dir = source_dir + '/build-' + board_type
mkdir_p(build_dir)
result_dir = config.getoption('result_dir')
if not result_dir:
result_dir = build_dir
mkdir_p(result_dir)
persistent_data_dir = config.getoption('persistent_data_dir')
if not persistent_data_dir:
persistent_data_dir = build_dir + '/persistent-data'
mkdir_p(persistent_data_dir)
test/py: support running sandbox under gdbserver Implement command--line option --gdbserver COMM, which does two things: a) Run the sandbox process under gdbserver, using COMM as gdbserver's communication channel. b) Disables all timeouts, so that if U-Boot is halted under the debugger, tests don't fail. If the user gives up in the middle of a debugging session, they can simply CTRL-C the test script to abort it. This allows easy debugging of test failures without having to manually re-create the failure conditions. Usage is: Window 1: ./test/py/test.py --bd sandbox --gdbserver localhost:1234 Window 2: gdb ./build-sandbox/u-boot -ex 'target remote localhost:1234' When using this option, it likely makes sense to use pytest's -k option to limit the set of tests that are executed. Simply running U-Boot directly under gdb (rather than gdbserver) was also considered. However, this was rejected because: a) gdb's output would then be processed by the test script, and likely confuse it causing false failures. b) pytest by default hides stdout from tests, which would prevent the user from interacting with gdb. While gdb can be told to redirect the debugee's stdio to a separate PTY, this would appear to leave gdb's stdio directed at the test scripts and the debugee's stdio directed elsewhere, which is the opposite of the desired effect. Perhaps some complicated PTY muxing and process hierarchy could invert this. However, the current scheme is simple to implement and use, so it doesn't seem worth complicating matters. c) Using gdbserver allows arbitrary debuggers to be used, even those with a GUI. If the test scripts invoked the debugger themselves, they'd have to know how to execute arbitary applications. While the user could hide this all in a wrapper script, this feels like extra complication. An interesting future idea might be a --gdb-screen option, which could spawn both U-Boot and gdb separately, and spawn the screen into a newly created window under screen. Similar options could be envisaged for creating a new xterm/... too. --gdbserver currently only supports sandbox, and not real hardware. That's primarily because the test hooks are responsible for all aspects of hardware control, so there's nothing for the test scripts themselves can do to enable gdbserver on real hardware. We might consider introducing a separate --disable-timeouts option to support use of debuggers on real hardware, and having --gdbserver imply that option. Signed-off-by: Stephen Warren <swarren@nvidia.com>
2016-02-04 23:11:50 +00:00
gdbserver = config.getoption('gdbserver')
if gdbserver and board_type != 'sandbox':
raise Exception('--gdbserver only supported with sandbox')
import multiplexed_log
log = multiplexed_log.Logfile(result_dir + '/test-log.html')
if config.getoption('build'):
if build_dir != source_dir:
o_opt = 'O=%s' % build_dir
else:
o_opt = ''
cmds = (
['make', o_opt, '-s', board_type + '_defconfig'],
['make', o_opt, '-s', '-j8'],
)
with log.section('make'):
runner = log.get_runner('make', sys.stdout)
for cmd in cmds:
runner.run(cmd, cwd=source_dir)
runner.close()
log.status_pass('OK')
class ArbitraryAttributeContainer(object):
pass
ubconfig = ArbitraryAttributeContainer()
ubconfig.brd = dict()
ubconfig.env = dict()
modules = [
(ubconfig.brd, 'u_boot_board_' + board_type_filename),
(ubconfig.env, 'u_boot_boardenv_' + board_type_filename),
(ubconfig.env, 'u_boot_boardenv_' + board_type_filename + '_' +
board_identity_filename),
]
for (dict_to_fill, module_name) in modules:
try:
module = __import__(module_name)
except ImportError:
continue
dict_to_fill.update(module.__dict__)
ubconfig.buildconfig = dict()
for conf_file in ('.config', 'include/autoconf.mk'):
dot_config = build_dir + '/' + conf_file
if not os.path.exists(dot_config):
raise Exception(conf_file + ' does not exist; ' +
'try passing --build option?')
with open(dot_config, 'rt') as f:
ini_str = '[root]\n' + f.read()
ini_sio = StringIO.StringIO(ini_str)
parser = ConfigParser.RawConfigParser()
parser.readfp(ini_sio)
ubconfig.buildconfig.update(parser.items('root'))
ubconfig.test_py_dir = test_py_dir
ubconfig.source_dir = source_dir
ubconfig.build_dir = build_dir
ubconfig.result_dir = result_dir
ubconfig.persistent_data_dir = persistent_data_dir
ubconfig.board_type = board_type
ubconfig.board_identity = board_identity
test/py: support running sandbox under gdbserver Implement command--line option --gdbserver COMM, which does two things: a) Run the sandbox process under gdbserver, using COMM as gdbserver's communication channel. b) Disables all timeouts, so that if U-Boot is halted under the debugger, tests don't fail. If the user gives up in the middle of a debugging session, they can simply CTRL-C the test script to abort it. This allows easy debugging of test failures without having to manually re-create the failure conditions. Usage is: Window 1: ./test/py/test.py --bd sandbox --gdbserver localhost:1234 Window 2: gdb ./build-sandbox/u-boot -ex 'target remote localhost:1234' When using this option, it likely makes sense to use pytest's -k option to limit the set of tests that are executed. Simply running U-Boot directly under gdb (rather than gdbserver) was also considered. However, this was rejected because: a) gdb's output would then be processed by the test script, and likely confuse it causing false failures. b) pytest by default hides stdout from tests, which would prevent the user from interacting with gdb. While gdb can be told to redirect the debugee's stdio to a separate PTY, this would appear to leave gdb's stdio directed at the test scripts and the debugee's stdio directed elsewhere, which is the opposite of the desired effect. Perhaps some complicated PTY muxing and process hierarchy could invert this. However, the current scheme is simple to implement and use, so it doesn't seem worth complicating matters. c) Using gdbserver allows arbitrary debuggers to be used, even those with a GUI. If the test scripts invoked the debugger themselves, they'd have to know how to execute arbitary applications. While the user could hide this all in a wrapper script, this feels like extra complication. An interesting future idea might be a --gdb-screen option, which could spawn both U-Boot and gdb separately, and spawn the screen into a newly created window under screen. Similar options could be envisaged for creating a new xterm/... too. --gdbserver currently only supports sandbox, and not real hardware. That's primarily because the test hooks are responsible for all aspects of hardware control, so there's nothing for the test scripts themselves can do to enable gdbserver on real hardware. We might consider introducing a separate --disable-timeouts option to support use of debuggers on real hardware, and having --gdbserver imply that option. Signed-off-by: Stephen Warren <swarren@nvidia.com>
2016-02-04 23:11:50 +00:00
ubconfig.gdbserver = gdbserver
ubconfig.dtb = build_dir + '/arch/sandbox/dts/test.dtb'
env_vars = (
'board_type',
'board_identity',
'source_dir',
'test_py_dir',
'build_dir',
'result_dir',
'persistent_data_dir',
)
for v in env_vars:
os.environ['U_BOOT_' + v.upper()] = getattr(ubconfig, v)
if board_type.startswith('sandbox'):
import u_boot_console_sandbox
console = u_boot_console_sandbox.ConsoleSandbox(log, ubconfig)
else:
import u_boot_console_exec_attach
console = u_boot_console_exec_attach.ConsoleExecAttach(log, ubconfig)
re_ut_test_list = re.compile(r'_u_boot_list_2_(dm|env)_test_2_\1_test_(.*)\s*$')
def generate_ut_subtest(metafunc, fixture_name):
"""Provide parametrization for a ut_subtest fixture.
Determines the set of unit tests built into a U-Boot binary by parsing the
list of symbols generated by the build process. Provides this information
to test functions by parameterizing their ut_subtest fixture parameter.
Args:
metafunc: The pytest test function.
fixture_name: The fixture name to test.
Returns:
Nothing.
"""
fn = console.config.build_dir + '/u-boot.sym'
try:
with open(fn, 'rt') as f:
lines = f.readlines()
except:
lines = []
lines.sort()
vals = []
for l in lines:
m = re_ut_test_list.search(l)
if not m:
continue
vals.append(m.group(1) + ' ' + m.group(2))
ids = ['ut_' + s.replace(' ', '_') for s in vals]
metafunc.parametrize(fixture_name, vals, ids=ids)
def generate_config(metafunc, fixture_name):
"""Provide parametrization for {env,brd}__ fixtures.
If a test function takes parameter(s) (fixture names) of the form brd__xxx
or env__xxx, the brd and env configuration dictionaries are consulted to
find the list of values to use for those parameters, and the test is
parametrized so that it runs once for each combination of values.
Args:
metafunc: The pytest test function.
fixture_name: The fixture name to test.
Returns:
Nothing.
"""
subconfigs = {
'brd': console.config.brd,
'env': console.config.env,
}
parts = fixture_name.split('__')
if len(parts) < 2:
return
if parts[0] not in subconfigs:
return
subconfig = subconfigs[parts[0]]
vals = []
val = subconfig.get(fixture_name, [])
# If that exact name is a key in the data source:
if val:
# ... use the dict value as a single parameter value.
vals = (val, )
else:
# ... otherwise, see if there's a key that contains a list of
# values to use instead.
vals = subconfig.get(fixture_name+ 's', [])
def fixture_id(index, val):
try:
return val['fixture_id']
except:
return fixture_name + str(index)
ids = [fixture_id(index, val) for (index, val) in enumerate(vals)]
metafunc.parametrize(fixture_name, vals, ids=ids)
def pytest_generate_tests(metafunc):
"""pytest hook: parameterize test functions based on custom rules.
Check each test function parameter (fixture name) to see if it is one of
our custom names, and if so, provide the correct parametrization for that
parameter.
Args:
metafunc: The pytest test function.
Returns:
Nothing.
"""
for fn in metafunc.fixturenames:
if fn == 'ut_subtest':
generate_ut_subtest(metafunc, fn)
continue
generate_config(metafunc, fn)
test/py: move U-Boot respawn trigger to the test core Prior to this change, U-Boot was lazilly (re-)spawned if/when a test attempted to interact with it, and no active connection existed. This approach was simple, yet had the disadvantage that U-Boot might be spawned in the middle of a test function, e.g. after the test had already performed actions such as creating data files, etc. In that case, this could cause the log to contain the sequence (1) some test logs, (2) U-Boot's boot process, (3) the rest of that test's logs. This isn't optimally readable. This issue will affect the upcoming DFU and enhanced UMS tests. This change converts u_boot_console to be a function-scoped fixture, so that pytest attempts to re-create the object for each test invocation. This allows the fixture factory function to ensure that U-Boot is spawned prior to every test. In practice, the same object is returned each time so there is essentially no additional overhead due to this change. This allows us to remove: - The explicit ensure_spawned() call from test_sleep, since the core now ensures that the spawn happens before the test code is executed. - The laxy calls to ensure_spawned() in the u_boot_console_* implementations. The one downside is that test_env's "state_ttest_env" fixture must be converted to a function-scoped fixture too, since a module-scoped fixture cannot use a function-scoped fixture. To avoid overhead, we use the same trick of returning the same object each time. Signed-off-by: Stephen Warren <swarren@nvidia.com> Acked-by: Simon Glass <sjg@chromium.org>
2016-01-22 19:30:08 +00:00
@pytest.fixture(scope='function')
def u_boot_console(request):
"""Generate the value of a test's u_boot_console fixture.
Args:
request: The pytest request.
Returns:
The fixture value.
"""
test/py: move U-Boot respawn trigger to the test core Prior to this change, U-Boot was lazilly (re-)spawned if/when a test attempted to interact with it, and no active connection existed. This approach was simple, yet had the disadvantage that U-Boot might be spawned in the middle of a test function, e.g. after the test had already performed actions such as creating data files, etc. In that case, this could cause the log to contain the sequence (1) some test logs, (2) U-Boot's boot process, (3) the rest of that test's logs. This isn't optimally readable. This issue will affect the upcoming DFU and enhanced UMS tests. This change converts u_boot_console to be a function-scoped fixture, so that pytest attempts to re-create the object for each test invocation. This allows the fixture factory function to ensure that U-Boot is spawned prior to every test. In practice, the same object is returned each time so there is essentially no additional overhead due to this change. This allows us to remove: - The explicit ensure_spawned() call from test_sleep, since the core now ensures that the spawn happens before the test code is executed. - The laxy calls to ensure_spawned() in the u_boot_console_* implementations. The one downside is that test_env's "state_ttest_env" fixture must be converted to a function-scoped fixture too, since a module-scoped fixture cannot use a function-scoped fixture. To avoid overhead, we use the same trick of returning the same object each time. Signed-off-by: Stephen Warren <swarren@nvidia.com> Acked-by: Simon Glass <sjg@chromium.org>
2016-01-22 19:30:08 +00:00
console.ensure_spawned()
return console
anchors = {}
tests_not_run = []
tests_failed = []
tests_xpassed = []
tests_xfailed = []
tests_skipped = []
tests_passed = []
def pytest_itemcollected(item):
"""pytest hook: Called once for each test found during collection.
This enables our custom result analysis code to see the list of all tests
that should eventually be run.
Args:
item: The item that was collected.
Returns:
Nothing.
"""
tests_not_run.append(item.name)
def cleanup():
"""Clean up all global state.
Executed (via atexit) once the entire test process is complete. This
includes logging the status of all tests, and the identity of any failed
or skipped tests.
Args:
None.
Returns:
Nothing.
"""
if console:
console.close()
if log:
with log.section('Status Report', 'status_report'):
log.status_pass('%d passed' % len(tests_passed))
if tests_skipped:
log.status_skipped('%d skipped' % len(tests_skipped))
for test in tests_skipped:
anchor = anchors.get(test, None)
log.status_skipped('... ' + test, anchor)
if tests_xpassed:
log.status_xpass('%d xpass' % len(tests_xpassed))
for test in tests_xpassed:
anchor = anchors.get(test, None)
log.status_xpass('... ' + test, anchor)
if tests_xfailed:
log.status_xfail('%d xfail' % len(tests_xfailed))
for test in tests_xfailed:
anchor = anchors.get(test, None)
log.status_xfail('... ' + test, anchor)
if tests_failed:
log.status_fail('%d failed' % len(tests_failed))
for test in tests_failed:
anchor = anchors.get(test, None)
log.status_fail('... ' + test, anchor)
if tests_not_run:
log.status_fail('%d not run' % len(tests_not_run))
for test in tests_not_run:
anchor = anchors.get(test, None)
log.status_fail('... ' + test, anchor)
log.close()
atexit.register(cleanup)
def setup_boardspec(item):
"""Process any 'boardspec' marker for a test.
Such a marker lists the set of board types that a test does/doesn't
support. If tests are being executed on an unsupported board, the test is
marked to be skipped.
Args:
item: The pytest test item.
Returns:
Nothing.
"""
mark = item.get_marker('boardspec')
if not mark:
return
required_boards = []
for board in mark.args:
if board.startswith('!'):
if ubconfig.board_type == board[1:]:
pytest.skip('board not supported')
return
else:
required_boards.append(board)
if required_boards and ubconfig.board_type not in required_boards:
pytest.skip('board not supported')
def setup_buildconfigspec(item):
"""Process any 'buildconfigspec' marker for a test.
Such a marker lists some U-Boot configuration feature that the test
requires. If tests are being executed on an U-Boot build that doesn't
have the required feature, the test is marked to be skipped.
Args:
item: The pytest test item.
Returns:
Nothing.
"""
mark = item.get_marker('buildconfigspec')
if not mark:
return
for option in mark.args:
if not ubconfig.buildconfig.get('config_' + option.lower(), None):
pytest.skip('.config feature not enabled')
def pytest_runtest_setup(item):
"""pytest hook: Configure (set up) a test item.
Called once for each test to perform any custom configuration. This hook
is used to skip the test if certain conditions apply.
Args:
item: The pytest test item.
Returns:
Nothing.
"""
anchors[item.name] = log.start_section(item.name)
setup_boardspec(item)
setup_buildconfigspec(item)
def pytest_runtest_protocol(item, nextitem):
"""pytest hook: Called to execute a test.
This hook wraps the standard pytest runtestprotocol() function in order
to acquire visibility into, and record, each test function's result.
Args:
item: The pytest test item to execute.
nextitem: The pytest test item that will be executed after this one.
Returns:
A list of pytest reports (test result data).
"""
reports = runtestprotocol(item, nextitem=nextitem)
failure_cleanup = False
test_list = tests_passed
msg = 'OK'
msg_log = log.status_pass
for report in reports:
if report.outcome == 'failed':
if hasattr(report, 'wasxfail'):
test_list = tests_xpassed
msg = 'XPASSED'
msg_log = log.status_xpass
else:
failure_cleanup = True
test_list = tests_failed
msg = 'FAILED:\n' + str(report.longrepr)
msg_log = log.status_fail
break
if report.outcome == 'skipped':
if hasattr(report, 'wasxfail'):
failure_cleanup = True
test_list = tests_xfailed
msg = 'XFAILED:\n' + str(report.longrepr)
msg_log = log.status_xfail
break
test_list = tests_skipped
msg = 'SKIPPED:\n' + str(report.longrepr)
msg_log = log.status_skipped
if failure_cleanup:
console.drain_console()
test_list.append(item.name)
tests_not_run.remove(item.name)
try:
msg_log(msg)
except:
# If something went wrong with logging, it's better to let the test
# process continue, which may report other exceptions that triggered
# the logging issue (e.g. console.log wasn't created). Hence, just
# squash the exception. If the test setup failed due to e.g. syntax
# error somewhere else, this won't be seen. However, once that issue
# is fixed, if this exception still exists, it will then be logged as
# part of the test's stdout.
import traceback
print 'Exception occurred while logging runtest status:'
traceback.print_exc()
# FIXME: Can we force a test failure here?
log.end_section(item.name)
if failure_cleanup:
console.cleanup_spawn()
return reports