On some platforms, the console is not the first serial port. To make
this work, the first serial port in QEMU must be set to "null".
Add support for this by adding an optional "serial" parameter, which
defaults to "stdio", and can be overridden by platform-specific
configuration.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
We don't run a linter regularly over kunit.py code (the default settings
on most don't like kernel style, e.g. tabs) so some of these imports
didn't get removed when they stopped being used.
Signed-off-by: Daniel Latypov <dlatypov@google.com>
Reviewed-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
The main function contains a wide if-elif block that handles different
subcommands. It's possible to make code refactoring to extract
subcommands handlers.
Fixed commit summary line.
Shuah Khan <skhan@linuxfoundation.org>
Signed-off-by: Alexander Pantyukhin <apantykhin@gmail.com>
Reviewed-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
The build_tests function contained double checking for not success
result. It is fixed in the current patch. Additional small
simplifications of code like using ternary if were applied (avoid using
the same operation by calculation times differ in two places).
Signed-off-by: Alexander Pantyukhin <apantykhin@gmail.com>
Reviewed-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Currently, kunit_parser.py is stripping all leading whitespace to make
parsing easier. But this means we can't accurately show kernel output
for failing tests or when the kernel crashes.
Embarassingly, this affects even KUnit's own output, e.g.
[13:40:46] Expected 2 + 1 == 2, but
[13:40:46] 2 + 1 == 3 (0x3)
[13:40:46] not ok 1 example_simple_test
[13:40:46] [FAILED] example_simple_test
After this change, here's what the output in context would look like
[13:40:46] =================== example (4 subtests) ===================
[13:40:46] # example_simple_test: initializing
[13:40:46] # example_simple_test: EXPECTATION FAILED at lib/kunit/kunit-example-test.c:29
[13:40:46] Expected 2 + 1 == 2, but
[13:40:46] 2 + 1 == 3 (0x3)
[13:40:46] [FAILED] example_simple_test
[13:40:46] [SKIPPED] example_skip_test
[13:40:46] [SKIPPED] example_mark_skipped_test
[13:40:46] [PASSED] example_all_expect_macros_test
[13:40:46] # example: initializing suite
[13:40:46] # example: pass:1 fail:1 skip:2 total:4
[13:40:46] # Totals: pass:1 fail:1 skip:2 total:4
[13:40:46] ===================== [FAILED] example =====================
This example shows one minor cosmetic defect this approach has.
The test counts lines prevent us from dedenting the suite-level output.
But at the same time, any form of non-KUnit output would do the same
unless it happened to be indented as well.
Signed-off-by: Daniel Latypov <dlatypov@google.com>
Reviewed-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
We print the "test log" on failure.
This is meant to be all the kernel output that happened during the test.
But we also include the special KTAP lines in it, which are often
redundant.
E.g. we include the "not ok" line in the log, right before we print
that the test case failed...
[13:51:48] Expected 2 + 1 == 2, but
[13:51:48] 2 + 1 == 3 (0x3)
[13:51:48] not ok 1 example_simple_test
[13:51:48] [FAILED] example_simple_test
More full example after this patch:
[13:51:48] =================== example (4 subtests) ===================
[13:51:48] # example_simple_test: initializing
[13:51:48] # example_simple_test: EXPECTATION FAILED at lib/kunit/kunit-example-test.c:29
[13:51:48] Expected 2 + 1 == 2, but
[13:51:48] 2 + 1 == 3 (0x3)
[13:51:48] [FAILED] example_simple_test
Signed-off-by: Daniel Latypov <dlatypov@google.com>
Reviewed-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Change the KUnit parser to be able to parse test output that complies with
the KTAP version 1 specification format found here:
https://kernel.org/doc/html/latest/dev-tools/ktap.html. Ensure the parser
is able to parse tests with the original KUnit test output format as
well.
KUnit parser now accepts any of the following test output formats:
Original KUnit test output format:
TAP version 14
1..1
# Subtest: kunit-test-suite
1..3
ok 1 - kunit_test_1
ok 2 - kunit_test_2
ok 3 - kunit_test_3
# kunit-test-suite: pass:3 fail:0 skip:0 total:3
# Totals: pass:3 fail:0 skip:0 total:3
ok 1 - kunit-test-suite
KTAP version 1 test output format:
KTAP version 1
1..1
KTAP version 1
1..3
ok 1 kunit_test_1
ok 2 kunit_test_2
ok 3 kunit_test_3
ok 1 kunit-test-suite
New KUnit test output format (changes made in the next patch of
this series):
KTAP version 1
1..1
KTAP version 1
# Subtest: kunit-test-suite
1..3
ok 1 kunit_test_1
ok 2 kunit_test_2
ok 3 kunit_test_3
# kunit-test-suite: pass:3 fail:0 skip:0 total:3
# Totals: pass:3 fail:0 skip:0 total:3
ok 1 kunit-test-suite
Signed-off-by: Rae Moar <rmoar@google.com>
Reviewed-by: Daniel Latypov <dlatypov@google.com>
Reviewed-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
When --raw_output is set (to any value), we don't actually parse the
test results. So asking to print the test results as json doesn't make
sense.
We internally create a fake test with one passing subtest, so --json
would actually print out something misleading.
This patch:
* Rewords the flag descriptions so hopefully this is more obvious.
* Also updates --raw_output's description to note the default behavior
is to print out only "KUnit" results (actually any KTAP results)
* also renames and refactors some related logic for clarity (e.g.
test_result => test, it's a kunit_parser.Test object).
Notably, this patch does not make it an error to specify --json and
--raw_output together. This is an edge case, but I know of at least one
wrapper around kunit.py that always sets --json. You'd never be able to
use --raw_output with that wrapper.
Signed-off-by: Daniel Latypov <dlatypov@google.com>
Reviewed-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
We currently tell people we "couldn't find any KTAP output" with no
indication as to what this might mean.
After this patch, we get:
$ ./tools/testing/kunit/kunit.py parse /dev/null
============================================================
[ERROR] Test: <missing>: Could not find any KTAP output. Did any KUnit tests run?
============================================================
Testing complete. Ran 0 tests: errors: 1
Note: we could try and generate a more verbose message like
> Please check .kunit/test.log to see the raw kernel output.
or the like, but we'd need to know what the build dir was to know where
test.log actually lives.
This patch tries to make a more minimal improvement.
Signed-off-by: Daniel Latypov <dlatypov@google.com>
Reviewed-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Let's verify that the parser isn't reporting any errors for valid
inputs.
This change also
* does result.status checking on one line
* makes sure we consistently do it outside of the `with` block
Signed-off-by: Daniel Latypov <dlatypov@google.com>
Reviewed-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Since we're using Python 3.7+, we can use dataclasses to tersen the
code.
It also lets us create pre-populated TestCounts() objects and compare
them in our unit test. (Before, you could only create empty ones).
Signed-off-by: Daniel Latypov <dlatypov@google.com>
Reviewed-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
E.g. all the hw_breakpoint tests are failing right now.
So if I run `kunit.py run --altests --arch=x86_64`, then I see
> Testing complete. Ran 408 tests: passed: 392, failed: 9, skipped: 7
Seeing which 9 tests failed out of the hundreds is annoying.
If my terminal doesn't have scrollback support, I have to resort to
looking at `.kunit/test.log` for the `not ok` lines.
Teach kunit.py to print a summarized list of failures if the # of tests
reachs an arbitrary threshold (>=100 tests).
To try and keep the output from being too long/noisy, this new logic
a) just reports "parent_test failed" if every child test failed
b) won't print anything if there are >10 failures (also arbitrary).
With this patch, we get an extra line of output showing:
> Testing complete. Ran 408 tests: passed: 392, failed: 9, skipped: 7
> Failures: hw_breakpoint
This also works with parameterized tests, e.g. if I add a fake failure
> Failures: kcsan.test_atomic_builtins_missing_barrier.threads=6
Note: we didn't have enough tests for this to be a problem before.
But with commit 980ac3ad05 ("kunit: tool: rename all_test_uml.config,
use it for --alltests"), --alltests works and thus running >100 tests
will probably become more common.
Signed-off-by: Daniel Latypov <dlatypov@google.com>
Reviewed-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Currently, if you run
$ ./tools/testing/kunit/kunit_tool_test.py
you'll see a lot of output from the parser as we feed it testdata.
This makes the output hard to read and fairly confusing, esp. since our
testdata includes example failures, which get printed out in red.
Silence that output so real failures are easier to see.
Signed-off-by: Daniel Latypov <dlatypov@google.com>
Reviewed-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
When running a RISC-V test kernel under QEMU, we need an OpenSBI BIOS
file. In the original QEMU support patchset, kunit_tool would optionally
download this file from GitHub if it didn't exist, using wget.
These days, it can usually be found in the distro's qemu-system-riscv
package, and is located in /usr/share/qemu on all the distros I tried
(Debian, Arch, OpenSUSE). Use this file, and thereby don't do any
downloading in kunit_tool.
In addition, we used to shell out to whatever 'wget' was in the path,
which could have potentially been used to trick the developer into
running another binary. By not using wget at all, we nicely sidestep
this issue.
Cc: Xu Panda <xu.panda@zte.com.cn>
Fixes: 87c9c16317 ("kunit: tool: add support for QEMU")
Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: David Gow <davidgow@google.com>
Tested-by: Daniel Latypov <dlatypov@google.com>
Reviewed-by: Brendan Higgins <brendanhiggins@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Context:
1. all_tests_uml.config used to be UML specific back when users to
manually specify CONFIG_VIRTIO_UML=y to enable CONFIG_PCI=y.
2. --alltests used allyesconfig along with a curated list of options to
disable. It's only ever worked for brief periods of time and has
perennially been broken due to compile issues.
Now all_tests_uml.config should work across ~all architectures.
Let's instead use this to implement --alltests.
Note: if anyone was using all_tests_uml.config, this change breaks them.
I think that's unlikely since it was added in 5.19 and was a lot to
type: --kunitconfig=tools/testing/kunit/configs/all_tests_uml.config.
We could make it a symlink to the new name, but I don't think the
caution is warranted here.
Signed-off-by: Daniel Latypov <dlatypov@google.com>
Reviewed-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Commit 6fc3a8636a ("kunit: tool: Enable virtio/PCI by default on UML")
made it so we enable these options by default for UML.
Specifying them here is now redundant.
Signed-off-by: Daniel Latypov <dlatypov@google.com>
Reviewed-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
This patch adds the kunit.enable module parameter that will need to be
set to true in addition to KUNIT being enabled for KUnit tests to run.
The default value is true giving backwards compatibility. However, for
the production+testing use case the new config option
KUNIT_DEFAULT_ENABLED can be set to N requiring the tester to opt-in
by passing kunit.enable=1 to the kernel.
Signed-off-by: Joe Fradley <joefradley@google.com>
Reviewed-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
With
$ kunit.py run --raw_output=all ...
you get the raw output from the kernel, e.g. something like
> TAP version 14
> 1..26
> # Subtest: time_test_cases
> 1..1
> ok 1 - time64_to_tm_test_date_range
> ok 1 - time_test_cases
But --raw_output=kunit or equivalently --raw_output, you get
> TAP version 14
> 1..26
> # Subtest: time_test_cases
> 1..1
> ok 1 - time64_to_tm_test_date_range
> ok 1 - time_test_cases
It looks less readable in my opinion, and it also isn't "raw output."
This is due to sharing code with kunit_parser.py, which wants to strip
leading whitespace since it uses anchored regexes.
We could update the kunit_parser.py code to tolerate leaading spaces,
but this patch takes the easier way out and adds a bool flag.
Signed-off-by: Daniel Latypov <dlatypov@google.com>
Reviewed-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
There are several tests which depend on PCI, and hence need a bunch of
extra options to run under UML. This makes it awkward to give
configuration instructions (whether in documentation, or as part of a
.kunitconfig file), as two separate, incompatible sets of config options
are required for UML and "most other architectures".
For non-UML architectures, it's possible to add default kconfig options
via the qemu_config python files, but there's no equivalent for UML. Add
a new tools/testing/kunit/configs/arch_uml.config file containing extra
kconfig options to use on UML.
Tested-by: José Expósito <jose.exposito89@gmail.com>
Reviewed-by: Daniel Latypov <dlatypov@google.com>
Signed-off-by: David Gow <davidgow@google.com>
Reviewed-by: Brendan Higgins <brendanhiggins@google.com>
Signed-off-by: Daniel Latypov <dlatypov@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
It's come up a few times that it would be useful to have --kunitconfig
be repeatable [1][2].
This could be done before with a bit of shell-fu, e.g.
$ find fs/ -name '.kunitconfig' -exec cat {} + | \
./tools/testing/kunit/kunit.py run --kunitconfig=/dev/stdin
or equivalently:
$ cat fs/ext4/.kunitconfig fs/fat/.kunitconfig | \
./tools/testing/kunit/kunit.py run --kunitconfig=/dev/stdin
But this can be fairly clunky to use in practice.
And having explicit support in kunit.py opens the door to having more
config fragments of interest, e.g. options for PCI on UML [1], UML
coverage [2], variants of tests [3].
There's another argument to be made that users can just use multiple
--kconfig_add's, but this gets very clunky very fast (e.g. [2]).
Note: there's a big caveat here that some kconfig options might be
incompatible. We try to give a clearish error message in the simple case
where the same option appears multiple times with conflicting values,
but more subtle ones (e.g. mutually exclusive options) will be
potentially very confusing for the user. I don't know we can do better.
Note 2: if you want to combine a --kunitconfig with the default, you
either have to do to specify the current build_dir
> --kunitconfig=.kunit --kunitconfig=additional.config
or
> --kunitconfig=tools/testing/kunit/configs/default.config --kunitconifg=additional.config
each of which have their downsides (former depends on --build_dir,
doesn't work if you don't have a .kunitconfig yet), etc.
Example with conflicting values:
> $ ./tools/testing/kunit/kunit.py config --kunitconfig=lib/kunit --kunitconfig=/dev/stdin <<EOF
> CONFIG_KUNIT_TEST=n
> CONFIG_KUNIT=m
> EOF
> ...
> kunit_kernel.ConfigError: Multiple values specified for 2 options in kunitconfig:
> CONFIG_KUNIT=y
> vs from /dev/stdin
> CONFIG_KUNIT=m
>
> CONFIG_KUNIT_TEST=y
> vs from /dev/stdin
> # CONFIG_KUNIT_TEST is not set
[1] https://lists.freedesktop.org/archives/dri-devel/2022-June/357616.html
[2] https://lore.kernel.org/linux-kselftest/CAFd5g45f3X3xF2vz2BkTHRqOC4uW6GZxtUUMaP5mwwbK8uNVtA@mail.gmail.com/
[3] https://lore.kernel.org/linux-kselftest/CANpmjNOdSy6DuO6CYZ4UxhGxqhjzx4tn0sJMbRqo2xRFv9kX6Q@mail.gmail.com/
Signed-off-by: Daniel Latypov <dlatypov@google.com>
Reviewed-by: Brendan Higgins <brendanhiggins@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
