1 # Autotest Best Practices 2 When the Chrome OS team started using autotest, we tried our best to figure out 3 how to fit our code and our tests into the upstream style with little guidance 4 and poor documentation. This went poorly. With the benefit of hindsight, 5 were going to lay out some best-practices that wed like to enforce going 6 forward. In many cases, there is legacy code that contradicts this style; we 7 should go through and refactor that code to fit these guidelines as time 8 allows. 9 10 ## Upstream Documentation 11 12 There is a sizeable volume of general Autotest documentation available on 13 github: 14 https://github.com/autotest/autotest/wiki 15 16 ## Coding style 17 18 Basically PEP-8. See [docs/coding-style.md](coding-style.md) 19 20 ## Where should my code live? 21 22 | Type of Code | Relative Path | 23 |---------------------------|-------------------------| 24 | client-side tests | client/site_tests/ | 25 | server-side tests | server/site_tests | 26 | common library code | client/common_lib/cros/ | 27 | server-only library code | server/cros | 28 29 30 ## Writing tests 31 32 An autotest is really defined by its control file. A control file contains 33 important metadata about the test (name, author, description, duration, what 34 suite its in, etc) and then pulls in and executes the actual test code. This 35 test code can be shared among multiple distinct test cases by parameterizing it 36 and passing those parameters in from separate control files. 37 38 Autotests *must*: 39 40 * Be self-contained: assume nothing about the condition of the device 41 * Be hermetic: requiring the Internet to be reachable in order for your test 42 to succeed is unacceptable. 43 * Be automatic: avoid user interaction and run-time specification of input 44 values. 45 * Be integration tests: if you can test the feature in a unit test (or a 46 chrome browser test), do so. 47 * Prefer object composition to inheritance: avoid subclassing test.test to 48 implement common functionality for multiple tests. Instead, create a class 49 that your tests can instantiate to perform common operations. This enables 50 us to write tests that use both PyAuto and Servo without dealing with 51 multiple inheritance, for example. 52 * Be deterministic: a test should not validate the timing of some operation. 53 Instead, write a test that records the timing in performance keyvals so that 54 we can track the numbers over time. 55 56 Autotests *must not*: 57 58 * Put significant logic in the control file: control files are really just 59 python, so one can put arbitrary logic in there. Dont. Run your test 60 code, perhaps with some parameters. 61 62 Autotests *may*: 63 64 * Share parameterized fixtures: a test is defined by a control file. Control 65 files import and run test code, and can pass simple parameters to the code 66 they run through a well-specified interface. 67 68 Autotest has a notion of both client-side tests and server-side tests. Code in 69 a client-side test runs only on the device under test (DUT), and as such isnt 70 capable of maintaining state across reboots or handling a failed suspend/resume 71 and the like. If possible, an autotest should be written as a client-side 72 test. A server test runs on the autotest server, but gets assigned a DUT 73 just like a client-side test. It can use various autotest primitives (and 74 library code written by the CrOS team) to manipulate that device. Most, if not 75 all, tests that use Servo or remote power management should be server-side 76 tests, as an example. 77 78 Adding a test involves putting a control file and a properly-written test 79 wrapper in the right place in the source tree. There are conventions that must 80 be followed, and a variety of primitives available for use. When writing any 81 code, whether client-side test, server-side test, or library, have a strong 82 bias towards using autotest utility code. This keeps the codebase consistent. 83 84 85 ## Writing a test 86 87 This section explains considerations and requirements for any autotest, whether 88 client or server. 89 90 ### Control files 91 92 Upstream documentation 93 Our local conventions for autotest control files deviate from the above a bit, 94 but the indication about which fields are mandatory still holds. 95 96 | Variable | Required | Value | 97 |--------------|----------|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| 98 | AUTHOR | Yes | A comma-delimited string of at least one responsible engineer and a backup engineer -- or at worst a backup mailing list. i.e. AUTHOR = msb, snanda | 99 | DEPENDENCIES | No | list of tags known to the HW test lab. | 100 | DOC | Yes | Long description of the test, pass/fail criteria | 101 | NAME | Yes | Display name of the test. Generally this is the directory where your test lives e.g. hardware_TPMCheck. If you are using multiple run_test calls in the same control file or multiple control files with one test wrapper in the same suite, problems arise with the displaying of your test name. crosbug.com/35795. When in doubt ask. | 102 | SYNC\_COUNT | No | Integer >= 1. Number of simultaneous devices needed for a test run. | 103 | TIME | Yes | Test duration: 'FAST' (<1m), 'MEDIUM' (<10m), 'LONG' (<20m), 'LENGTHY' (>30m) | 104 | TEST\_TYPE | Yes | Client or Server | 105 | ATTRIBUTES | No | Comma separated list of attribute tags to apply to this control file, used in composing suites. For instance, 'suite:foo, suite:bar, subsystem:baz'. | 106 107 ### Running tests in suites 108 109 Make sure that the suite name is listed in `site_utils/attribute_whitelist.txt`, 110 then add the appropriate attribute to the ATTRIBUTES field in tests that make 111 up the test suite. For instance: 112 113 ``` 114 ... 115 ATTRIBUTES = 'suite:suite-a, suite:suite-b' 116 ... 117 ``` 118 119 would indicate that the control file above should be run as part of both 120 `suite-a` and `suite-b`. 121 122 ### Pure python 123 124 Lie, cheat and steal to keep your tests in pure python. It will be easier to 125 debug failures, it will be easier to generate meaningful error output, it will 126 be simpler to get your tests installed and run, and it will be simpler for the 127 lab team to build tools that allow you to quickly iterate. 128 129 Shelling out to existing command-line tools is done fairly often, and isnt a 130 terrible thing. The test author can wind up having to do a lot of output 131 parsing, which is often brittle, but this can be a decent tradeoff in lieu of 132 having to reimplement large pieces of functionality in python. 133 134 Note that you will need to be sure that any commands you use are installed on 135 the host. For a client-side test, the host means the DUT. For a 136 server-side test, the host typically means the system running autoserv; 137 however, if you use SiteHost.run(), the command will run on the DUT. On the 138 server, your tests will have access to all tools common to both a typical CrOS 139 chroot environment and standard Goobuntu. 140 141 If you want to use a tool on the DUT, it may be appropriate to include it as a 142 dependency of the chromeos-base/chromeos-test package. This ensures that the 143 tool is pre-installed on every test image for every device, and will always be 144 available for use. Otherwise, the tool must be installed as an autotest dep. 145 146 _Never install your own shell scripts and call them._ Anything you can do in 147 shell, you can do in python. 148 149 ### Reporting failures 150 151 Autotest supports several kinds of failure statuses: 152 153 | Status | Exception | Reason | 154 |----------|-------------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| 155 | WARN | error.TestWarn | error.TestWarn should be used when side effects to the test running are encountered but are not directly related to the test running. For example, if you are testing Wifi and powerd crashes. *Currently* there are not any clear usecases for this and error.TestWarn should be generally avoided until further notice. | 156 | TEST\_NA | error.TestNAError | This test does not apply in the current environment. | 157 | ERROR | error.TestError | The test was unable to validate the desired behavior. | 158 | FAIL | error.TestFail | The test determined the desired behavior failed to occur. | 159 160 161 ### Considerations when writing client-side tests 162 163 All client-side tests authored at Google must live in the client/site\_tests sub-directory of the autotest source tree. 164 165 ###Compiling and executing binaries 166 167 It is possible to compile source thats included with your test and use the 168 products at test runtime. The build infrastructure will compile this code for 169 the appropriate target architecture and package it up along with the rest of 170 your tests resources, but this increases your development iteration time as 171 you need to actually re-build and re-package your test to deploy it to the 172 device. While we hope to improve tooling support for this use case in the 173 future, avoiding this issue is the ideal. 174 175 If you cant avoid this, heres how to get your code compiled and installed as 176 a part of your test: 177 1. Create a src/ directory next to your control file. 178 2. Put your source, including its Makefile, in src/ 179 3. define a method in your test class called setup(self) that takes no arguments. 180 4. setup(self) should perform all tasks necessary to build your tool. There are some helpful utility functions in client/common_lib/utils.py. Trivial example: 181 182 ``` 183 def setup(self): 184 os.chdir(self.srcdir) 185 utils.make('OUT_DIR=.') 186 ``` 187 188 ### Reusing code (fixtures) 189 190 Any autotest is, essentially, a single usage of a re-usable test fixture. This 191 is because run\_once() in your test wrapper can take any arguments you want. As 192 such, multiple control files can re-use the same wrapper -- and should, where 193 it makes sense. 194 195 ### Considerations when writing server-side tests 196 197 All server-side tests authored at Google must live in the server/site\_tests 198 sub-directory of the autotest source tree. 199 200 It should be even easier to keep the server-side of a test in pure python, as 201 you should simply be driving the DUT and verifying state. 202 203 ### When/why to write a server-side test 204 205 Server-side tests are appropriate when some operation in the test can't be 206 executed on the DUT. The prototypical example is rebooting the DUT. Other 207 examples include tests that manipulate the network around the DUT (e.g. WiFi 208 tests), tests that power off the DUT, and tests that rely on a Servo attached 209 to the DUT. 210 211 One simple criterion for whether to write a server-side test is this: Is the 212 DUT an object that the test must manipulate? If the answer is yes, then a 213 server-side test makes sense. 214 215 ### Control files for server-side tests 216 217 Server-side tests commonly operate on the DUT as an object. Autotest 218 represents the DUT with an instance of class Host; the instance is constructed 219 and passed to the test from the control file. Creating the host object in the 220 control file can be done using certain definitions present in the global 221 environment of every control file: 222 223 * Function hosts.create\_host() will create a host object from a string with 224 the name of the host (an IP address as a string is also acceptable). 225 * Variable machines is a list of the host names available to the test. 226 227 Below is a sample fragment for a control file that runs a simple server side test in parallel on all the hosts specified for the test. The fragment is a complete control file, except for the missing boilerplate comments and documentation definitions required in all control files. 228 229 ``` 230 def run(machine): 231 host = hosts.create_host(machine) 232 job.run_test("platform_ServerTest", host=host) 233 234 parallel_simple(run, machines) 235 ``` 236 237 Note: The sample above relies on a common convention that the run\_once() 238 method of a server-side test defines an argument named host with a default 239 value, e.g. 240 241 ``` 242 def run_once(self, host=None): 243 # test code goes here. 244 ``` 245 246 ### Operations on Host objects 247 248 A Host object supports various methods to operate on a DUT. Below is a short list of important methods supported by instances of Host: 249 250 * run(command) - run a shell command on the host 251 * reboot() - reboot the host, and wait for it to be back on the network 252 * wait_up() - wait for the host to be active on the network 253 * wait_down() - wait until the host is no longer on the network, or until it is known to have rebooted. 254 255 More details, including a longer list of available methods, and more about how 256 they work can be found in the Autotest documentation for autoserv and Autotest 257 documentation for Host. 258 259 ### Servo-based tests 260 261 For server-side tests that use a servo-attached DUT, the host object has a 262 servo attribute. If Autotest determines that the DUT has a Servo attached, the 263 servo attribute will be a valid instance of a Servo client object; otherwise 264 the attribute will be None. 265 266 For a DUT in the lab, Autotest will automatically determine whether there is a 267 servo available; however, if a test requires Servo, its control file must have 268 additional code to guarantee a properly initialized servo object on the host. 269 270 Below is a code snippet outlining the requirements; portions of the control file have been omitted for brevity: 271 272 ``` 273 # ... Standard boilerplate variable assignments... 274 DEPENDENCIES = "servo" 275 # ... more standard boilerplate... 276 277 args_dict = utils.args_to_dict(args) 278 servo_args = hosts.SiteHost.get_servo_arguments(args_dict) 279 280 def run(machine): 281 host = hosts.create_host(machine, servo_args=servo_args) 282 job.run_test("platform_SampleServoTest", host=host) 283 284 parallel_simple(run, machines) 285 ``` 286 287 The `DEPENDENCIES` setting guarantees that if the test is scheduled in the lab, 288 it will be assigned to a DUT that has a servo. 289 290 The setting of `servo_args` guarantees two distinct things: First, it forces 291 checks that will make sure that the Servo is functioning properly; this 292 guarantees that the host's `servo` attribute will not be None. Second, the code 293 allows you to pass necessary servo specific command-line arguments to 294 `test_that`. 295 296 If the test control file follows the formula above, the test can be reliably called in a variety of ways: 297 * When used for hosts in the lab, the hosts servo object will use the servo attached to the host, and the test can assume that the servo object is not None. 298 * If you start servod manually on your desktop using the default port, you can use test_that without any special options. 299 * If you need to specify a non-default host or port number (e.g. because servod is remote, or because you have more than one servo board), you can specify them with commands like these: 300 301 ``` 302 test_that --args=servo_host=... 303 test_that --args=servo_port=... 304 test_that --args=servo_host=... servo_port=... ... 305 ``` 306 307 ### Calling client-side tests from a server-side test 308 309 Commonly, server-side tests need to do more on the DUT than simply run short 310 shell commands. In those cases, a client-side test should be written and 311 invoked from the server-side test. In particular, a client side test allows 312 the client side code to be written in Python that uses standard Autotest 313 infrastructure, such as various utility modules or the logging infrastructure. 314 315 Below is a short snippet showing the standard form for calling a client-side 316 test from server-side code: 317 318 ``` 319 from autotest_lib.server import autotest 320 321 # ... inside some function, e.g. in run_once() 322 client_at = autotest.Autotest(host) 323 client_at.run_test("platform_ClientTest") 324 ``` 325 326 ### Writing library code 327 328 There is a large quantity of Chromium OS specific code in the autotest 329 codebase. Much of this exists to provide re-usable modules that enable tests 330 to talk to system services. The guidelines from above apply here as well. 331 This code should be as pure python as possible, though it is reasonable to 332 shell out to command line tools from time to time. In some cases weve done 333 this where we could (now) use the services DBus APIs directly. If youre 334 adding code to allow tests to communicate with your service, it is strongly 335 recommended that you use DBus where possible, instead of munging config files 336 directly or using command-line tools. 337 338 Currently, our library code lives in a concerning variety of places in the 339 autotest tree. This is due to a poor initial understanding of how to do 340 things, and new code should follow the following conventions instead: 341 342 * Used only in server-side tests: server/cros 343 * Used in both server- and client-side tests, or only client: 344 client/common\_lib/cros 345 346 ### Adding test deps 347 348 This does not refer to the optional `DEPENDENCIES` field in test control files. 349 Rather, this section discusses how and when to use code/data/tools that are not 350 pre-installed on test images, and should (or can) not be included right in with 351 the test source. 352 353 Unfortunately, there is no hard-and-fast rule here. Generally, if this is some 354 small tool or blob of data you need for a single test, you should include it as 355 discussed above in Writing client-side tests. If youre writing the tool, and 356 it has use for developers as well as in one or more tests that youre writing, 357 then make it a first-class CrOS project. Write an ebuild, write unit tests, 358 and then add it to the test image by default. This can be done by RDEPENDing 359 on your new test package from the chromeos-test ebuild. 360 361 If your code/data falls in the middle (useful to several tests, not to devs), 362 and/or is large (hundreds of megabytes as opposed to tens) then using an 363 autotest dep may be the right choice. Conceptually, an autotest test dep is 364 simply another kind of archive that the autotest infrastructure knows how to 365 fetch and unpack. There are two components to including a dependency from an 366 autotest test -- setup during build time and installing it on your DUT when 367 running a test. The setup phase must be run from your tests setup() method 368 like so: 369 370 ``` 371 def setup(self): 372 self.job.setup_dep([mydep]) 373 logging.debug(mydep is at %s % (os.path.join(self.autodir, 374 deps/mydep)) 375 ``` 376 377 The above gets run when you build the test. 378 379 The other half of this equation is actually installing the dependency so you 380 can use it while running a test. To do this, add the following to either your 381 run\_once or initialize methods: 382 383 ``` 384 dep = dep_name 385 dep_dir = os.path.join(self.autodir, 'deps', dep=dep) 386 self.job.install_pkg(dep, 'dep', dep_dir) 387 ``` 388 389 390 You can now reference the content of your dep using dep_dir. 391 392 Now that you know how to include a dep, the next question is how to write one. 393 Before you read further, you should check out client/deps/\* for many examples 394 of deps in our autotest tree. 395 396 ### Create a dep from a third-party package 397 398 There are many examples of how to do this in the client/deps directory already. 399 The key component is to check in a tarball of the version of the dependency 400 youd like to include under client/deps/your\_dep. 401 402 All deps require a control file and an actual python module by the same name. 403 They will also need a copy of common.py to import utils.update\_version. Both 404 the control and common are straightforward, the python module does all the 405 magic. 406 407 The deps python module follows a standard convention: a setup function and a 408 call to utils.update\_version. update\_version is used instead of directly 409 calling setup as it maintains additional versioning logic ensuring setup is 410 only done 1x per dep. The following is its method signature: 411 412 ``` 413 def update_version(srcdir, preserve_srcdir, new_version, install, 414 *args, **dargs) 415 ``` 416 417 418 Notably, install should be a pointer to your setup function and `*args` should 419 be filled in with params to said setup function. 420 421 If you are using a tarball, your setup function should look something like: 422 423 ``` 424 def setup(tarball, my_dir) 425 utils.extract_tarball_to_dir(tarball, my_dir) 426 os.chdir(my_dir) 427 utils.make() # this assumes your tarball has a Makefile. 428 ``` 429 430 And you would invoke this with: 431 432 ``` 433 utils.update_version(os.getcwd(), True, version, setup, tarball_path, 434 os.getcwd()) 435 ``` 436 437 438 Note: The developer needs to call this because def setup is a function they are 439 defining that can take any number of arguments or install the dep in any way 440 they see fit. The above example uses tarballs but some are distributed as 441 straight source under the src dir so their setup function only takes a top 442 level path. We could avoid this by forcing a convention but that would be 443 artificially constraining the deps mechanism. 444 445 Once youve created the dep, you will also have to add the dep to the 446 autotest-deps package in chromiumos-overlay/chromeos-base/autotest-deps, 447 cros\_workon start it, and re-emerge it. 448 449 ### Create a dep from other chrome-os packages 450 451 One can also create autotest deps from code that lives in other CrOS packages, 452 or from build products generated by other packages. This is similar as above 453 but you can reference code using the `CHROMEOS_ROOT` env var that points to the 454 root of the CrOS source checkout, or the SYSROOT env var (which points to 455 /build/<board>) to refer to build products. Again, read the above. Heres an 456 example of the former with the files I want in 457 chromeos\_tree/chromite/my\_dep/\* where this will be the python code in 458 autotest/files/client/deps/my\_dep/my\_dep.py module. 459 460 ``` 461 import common, os, shutil 462 from autotest_lib.client.bin import utils 463 464 version = 1 465 466 def setup(setup_dir): 467 my_dep_dir = os.path.join(os.environ['CHROMEOS_ROOT'], 'chromite', 468 'buildbot') 469 shutil.copytree(my_dep_dir, setup_dir) 470 471 472 work_dir = os.path.join(os.getcwd(), 'src') 473 utils.update_version(os.getcwd(), True, version, setup, work_dir) 474 ``` 475
