The current test runner is only safe against parallel execution within
a single instance of the 'check' process, and only if -j is given a
value greater than 2. This prevents running multiple copies of the
'check' process for different test scenarios.
This change switches the output / socket directories to always include
the test name, image format and image protocol. This should allow full
parallelism of all distinct test scenarios. eg running both qcow2 and
raw tests at the same time, or both file and nbd tests at the same
time.
It would be possible to allow for parallelism of the same test scenario
by including the pid, but that would potentially let many directories
accumulate over time on failures, so is not done.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Acked-by: Hanna Czenczek <hreitz@redhat.com>
Tested-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20230303160727.3977246-7-berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20230315174331.2959-24-alex.bennee@linaro.org>
Currently the tests have their stdin inherited from the test harness,
meaning they are connected to a TTY. The QEMU processes spawned by
certain tests, however, modify TTY settings and if the test exits
abnormally the settings might not be restored.
The python test harness thus has some logic which will capture the
initial TTY settings and restore them once all tests are finished.
This does not, however, take into account the possibility of many
copies of the 'check' program running in parallel. With parallel
execution, a later invokation may save the TTY state that QEMU has
already modified, and thus restore bad state leaving the TTY
non-functional.
None of the I/O tests shnould actually be interactive requiring
user input and so they should not require a TTY at all. To avoid
this while TTY save/restore complexity we can connect the test
stdin to /dev/null instead.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Acked-by: Hanna Czenczek <hreitz@redhat.com>
Tested-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20230303160727.3977246-6-berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20230315174331.2959-23-alex.bennee@linaro.org>
Recently meson started complaining that TAP test reports don't include
the TAP protocol version. While this warning is bogus and has since been
removed from Meson, it looks like good practice to include this header
going forward. The GLib library test harness has started unconditionally
printing the version, so this brings the I/O tests into line.
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Acked-by: Hanna Czenczek <hreitz@redhat.com>
Tested-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20230303160727.3977246-5-berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20230315174331.2959-22-alex.bennee@linaro.org>
When running I/O tests using TAP output mode, we get a single TAP test
with a sub-test reported for each I/O test that is run. The output looks
something like this:
1..123
ok qcow2 011
ok qcow2 012
ok qcow2 013
ok qcow2 217
...
If everything runs or fails normally this is fine, but periodically we
have been seeing the test harness abort early before all 123 tests have
been run, just leaving a fairly useless message like
TAP parsing error: Too few tests run (expected 123, got 107)
we have no idea which tests were running at the time the test harness
abruptly exited. This change causes us to print a message about our
intent to run each test, so we have a record of what is active at the
time the harness exits abnormally.
1..123
# running qcow2 011
ok qcow2 011
# running qcow2 012
ok qcow2 012
# running qcow2 013
ok qcow2 013
# running qcow2 217
ok qcow2 217
...
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220509124134.867431-2-berrange@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
When stdout is not a terminal, the buffer may not be flushed at each end
of line, so we should flush after each test is done. This is especially
apparent when run by check-block, in two ways:
First, when running make check-block -jX with X > 1, progress indication
was missing, even though testrunner.py does theoretically print each
test's status once it has been run, even in multi-processing mode.
Flushing after each test restores this progress indication.
Second, sometimes make check-block failed altogether, with an error
message that "too few tests [were] run". I presume that's because one
worker process in the job pool did not get to flush its stdout before
the main process exited, and so meson did not get to see that worker's
test results. In any case, by flushing at the end of run_test(), the
problem has disappeared for me.
Signed-off-by: Hanna Reitz <hreitz@redhat.com>
Message-Id: <20220506134215.10086-1-hreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Quoting the TAP specification: "The plan tells how many tests will be
run [...]. It’s a check that the test file hasn’t stopped prematurely."
That's a good idea of course, so let's support that in the iotest
testrunner, too.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220223095816.2663005-1-thuth@redhat.com>
Signed-off-by: Hanna Reitz <hreitz@redhat.com>
When developing an I/O test it is typical to add some logic to the
test script, run it to view the output diff, and then apply the
output diff to the reference file. This can be drastically simplified
by letting the test runner update the reference file in place.
By setting 'QEMU_IOTESTS_REGEN=1', the test runner will report the
failure and show the diff, but at the same time update the reference
file. So next time the I/O test is run it will succeed.
Continuing to display the diff when updating the reference gives the
developer a chance to review what was changed.
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220304193610.3293146-7-berrange@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Drop the use of OUTPUT_DIR (test/qemu-iotests under the build
directory), and instead write test output files (.out.bad, .notrun, and
.casenotrun) to TEST_DIR.
With this, the same test can be run concurrently without the separate
instances interfering, because they will need separate TEST_DIRs anyway.
Running the same test separately is useful when running the iotests with
various format/protocol combinations in parallel, or when you just want
to aggressively exercise a single test (e.g. when it fails only
sporadically).
Putting this output into TEST_DIR means that it will stick around for
inspection after the test run is done (though running the same test in
the same TEST_DIR will overwrite it, just as it used to be); but given
that TEST_DIR is a scratch directory, it should be clear that users can
delete all of its content at any point. (And if TEST_DIR is on tmpfs,
it will just disappear on shutdown.) Contrarily, alternative approaches
that would put these output files into OUTPUT_DIR with some prefix to
differentiate between separate test runs might easily lead to cluttering
OUTPUT_DIR.
(This change means OUTPUT_DIR is no longer written to by the iotests, so
we can drop its usage altogether.)
Signed-off-by: Hanna Reitz <hreitz@redhat.com>
Message-Id: <20220221172909.762858-1-hreitz@redhat.com>
[hreitz: Simplified `Path(os.path.join(x, y))` to `Path(x, y)`, as
suggested by Vladimir; and rebased on 9086c76398
("tests/qemu-iotests: Rework the checks and spots using GNU
sed")]
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
In TAP mode, the stdout is reserved for the TAP protocol, so we
have to make sure to mark other lines with a comment '#' character
at the beginning to avoid that the TAP parser at the other end
gets confused.
To test this condition, run "configure" for example with:
--block-drv-rw-whitelist=copy-before-write,qcow2,raw,file,host_device,blkdebug,null-co,copy-on-read
so that iotest 041 will report that some tests are not run due to
the missing "quorum" driver. Without this change, "make check-block"
fails since the meson tap parser gets confused by these messages.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220223124353.3273898-1-thuth@redhat.com>
Signed-off-by: Hanna Reitz <hreitz@redhat.com>
When running in TAP mode, stdout is reserved for the TAP protocol.
To see the "diff" of the failed test, we have to print it to
stderr instead.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220209101530.3442837-8-thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Let "meson test" take care of showing the results of the individual tests,
consistently with other output from "make check V=1".
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
A lot of Optional[] types doesn't make code beautiful.
test_field_width defaults to 8, but that is never used in the code.
More over, if we want some default behavior for single call of
test_run(), it should just print the whole test name, not limiting or
expanding its width, so 8 is bad default.
So, just drop the default as unused for now.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20211210201450.101576-1-vsementsov@virtuozzo.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Add -j <JOBS> parameter, to run tests in several jobs simultaneously.
For realization - simply utilize multiprocessing.Pool class.
Notes:
1. Of course, tests can't run simultaneously in same TEST_DIR. So,
use subdirectories TEST_DIR/testname/ and SOCK_DIR/testname/
instead of simply TEST_DIR and SOCK_DIR
2. multiprocessing.Pool.starmap function doesn't support passing
context managers, so we can't simply pass "self". Happily, we need
self only for read-only access, and it just works if it is defined
in global space. So, add a temporary link TestRunner.shared_self
during run_tests().
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20211203122223.2780098-4-vsementsov@virtuozzo.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Tested-by: John Snow <jsnow@redhat.com>
Signed-off-by: Hanna Reitz <hreitz@redhat.com>
We are going to use do_run_test() in multiprocessing environment, where
we'll not be able to change original runner object.
Happily, the only thing we change is that last_elapsed and it's simple
to do it in run_tests() instead. All other accesses to self in
do_runt_test() and in run_test() are read-only.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20211203122223.2780098-3-vsementsov@virtuozzo.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Tested-by: John Snow <jsnow@redhat.com>
Signed-off-by: Hanna Reitz <hreitz@redhat.com>
We are going to modify these methods and will add more documentation in
further commit. As a preparation add basic documentation.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20211203122223.2780098-2-vsementsov@virtuozzo.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Tested-by: John Snow <jsnow@redhat.com>
Signed-off-by: Hanna Reitz <hreitz@redhat.com>
This makes it possible to see what is happening, even if the output of
"make check-block" is not sent to a tty (for example if it is sent to
grep or tee).
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
1. Ignore the new f-strings warning, we're not interested in doing a
full conversion at this time.
2. Just mute the unbalanced-tuple-unpacking warning, it's not a real
error in this case and muting the dozens of callsites is just not
worth it.
3. Add encodings to read_text().
Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20210923180715.4168522-7-jsnow@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
pylint 2.8 introduces consider-using-with error, suggesting
to use the 'with' block statement when possible.
Modify all subprocess.Popen call to use the 'with' statement,
except the one in __init__ of QemuIoInteractive class, since
it is assigned to a class field and used in other methods.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Message-Id: <20210510190449.65948-1-eesposit@redhat.com>
[mreitz: Disable bad-option-value warning in the iotests' pylintrc, so
that disabling consider-using-with in QemuIoInteractive will
not produce a warning in pre-2.8 pylint versions]
Signed-off-by: Max Reitz <mreitz@redhat.com>
Right now there is no easy way for "check" to print a reproducer command.
Because such a reproducer command line would be huge, we can instead teach
check to start a command of our choice. This can be for example a Python
unit test with arguments to only run a specific subtest.
Move the trailing empty line to print_env(), since it always looks better
and one caller was not adding it.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Tested-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Message-Id: <20210323181928.311862-5-pbonzini@redhat.com>
Message-Id: <20210503110110.476887-5-pbonzini@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
In the next patch, "check" will learn how to execute a test script without
going through TestRunner. To enable this, keep only the text output
and subprocess handling in the TestRunner; move into TestEnv the logic
to prepare for running a subprocess.
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Tested-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Message-Id: <20210323181928.311862-4-pbonzini@redhat.com>
Message-Id: <20210503110110.476887-4-pbonzini@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
For -makecheck, the old 'check' implementation skipped the output when
starting a test. It only had the condensed output at the end of a test.
testrunner.py prints the normal output when starting a test even for
-makecheck. This output contains '\r' at the end so that it can be
overwritten with the result at the end of the test. However, for
-makecheck this is shorter output in a different format, so effectively
we end up with garbled output that mixes both output forms.
Revert to the old behaviour of only printing a message after the test
had completed in -makecheck mode.
Fixes: d74c754c92
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20210201161024.127921-1-kwolf@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
We should indicate failure by exit code, not only output.
Reported-by: Peter Maydell
Fixes: f203080bbd
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210201085041.3079-1-vsementsov@virtuozzo.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
We should drop final '\n' before comparing with python3 shebang.
Fixes: d74c754c92
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210129090616.84145-1-vsementsov@virtuozzo.com>
Tested-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Add TestRunner class, which will run tests in a new python iotests
running framework.
There are some differences with current ./check behavior, most
significant are:
- Consider all tests self-executable, just run them, don't run python
by hand.
- Elapsed time is cached in json file
- Elapsed time precision increased a bit
- Instead of using "diff -w" which ignores all whitespace differences,
manually strip whitespace at line end then use python difflib, which
no longer ignores spacing mid-line
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210125185056.129513-5-vsementsov@virtuozzo.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
