kunit: tool: actually track how long it took to run tests
authorDaniel Latypov <dlatypov@google.com>
Thu, 30 Sep 2021 22:20:47 +0000 (15:20 -0700)
committerShuah Khan <skhan@linuxfoundation.org>
Tue, 19 Oct 2021 20:18:50 +0000 (14:18 -0600)
commit5f6aa6d82e45d27f00f39844749ae812c7fd6ab3
tree9036c9cb370707947a0d11b9af0f0bc4429d88ae
parent7ef925ea819469b1341a9b704c3e72bacf62000c
kunit: tool: actually track how long it took to run tests

This is a long standing bug in kunit tool.
Since these files were added, run_kernel() has always yielded lines.

That means, the call to run_kernel() returns before the kernel finishes
executing tests, potentially before a single line of output is even
produced.

So code like this
  time_start = time.time()
  result = linux.run_kernel(...)
  time_end = time.time()

would only measure the time taken for python to give back the generator
object.

From a caller's perspective, the only way to know the kernel has exited
is for us to consume all the output from the `result` generator object.
Alternatively, we could change run_kernel() to try and do its own book
keeping and return the total time, but that doesn't seem worth it.

This change makes us record `time_end` after we're done parsing all the
output (which should mean we've consumed all of it, or errored out).
That means we're including in the parsing time as well, but that should
be quite small, and it's better than claiming it took 0s to run tests.

Let's use this as an example:
$ ./tools/testing/kunit/kunit.py run --kunitconfig=lib/kunit example

Before:
Elapsed time: 7.684s total, 0.001s configuring, 4.692s building, 0.000s running

After:
Elapsed time: 6.283s total, 0.001s configuring, 3.202s building, 3.079s running

Signed-off-by: Daniel Latypov <dlatypov@google.com>
Reviewed-by: David Gow <davidgow@google.com>
Reviewed-by: Brendan Higgins <brendanhiggins@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
tools/testing/kunit/kunit.py