Re: [PATCH v2 kunit-next 2/3] kunit: add "run" debugfs file to run suites, display results

From: Brendan Higgins
Date: Thu Jan 30 2020 - 21:43:25 EST


On Thu, Jan 23, 2020 at 10:47 AM Alan Maguire <alan.maguire@xxxxxxxxxx> wrote:
>
> Add /sys/kernel/debug/kunit/<suite>/run file which will run the
> specified suite and show results.
>
> Signed-off-by: Alan Maguire <alan.maguire@xxxxxxxxxx>

If you don't mind, I would like to see the device tree unit test from
Frank before we accept this patch. I definitely like your approach
here, but this would break with KUnit test cases which depend on
__init code and data. I just figure that it would be easier for us to
solve the __init problem now if we have a working example that uses it
rather than having someone who wants to write a test which depends on
__init having to fix this after the fact. Let me know if this is a
problem for you.

> ---
> lib/kunit/debugfs.c | 33 +++++++++++++++++++++++++++++++++
> 1 file changed, 33 insertions(+)
>
> diff --git a/lib/kunit/debugfs.c b/lib/kunit/debugfs.c
> index 578843c..1ea3fbc 100644
> --- a/lib/kunit/debugfs.c
> +++ b/lib/kunit/debugfs.c
> @@ -13,6 +13,7 @@
>
> #define KUNIT_DEBUGFS_ROOT "kunit"
> #define KUNIT_DEBUGFS_RESULTS "results"
> +#define KUNIT_DEBUGFS_RUN "run"
>
> /*
> * Create a debugfs representation of test suites:
> @@ -20,6 +21,7 @@
> * Path Semantics
> * /sys/kernel/debug/kunit/<testsuite>/results Show results of last run for
> * testsuite
> + * /sys/kernel/debug/kunit/<testsuite>/run Run testsuite and show results
> *
> */
>
> @@ -67,6 +69,18 @@ static int debugfs_print_results(struct seq_file *seq, void *v)
> return 0;
> }
>
> +/*
> + * /sys/kernel/debug/kunit/<testsuite>/run (re)runs suite and shows all results.
> + */
> +static int debugfs_run_print_results(struct seq_file *seq, void *v)
> +{
> + struct kunit_suite *suite = (struct kunit_suite *)seq->private;
> +
> + kunit_run_tests(suite);
> +
> + return debugfs_print_results(seq, v);
> +}
> +
> static int debugfs_release(struct inode *inode, struct file *file)
> {
> return single_release(inode, file);
> @@ -88,6 +102,22 @@ static int debugfs_results_open(struct inode *inode, struct file *file)
> .release = debugfs_release,
> };
>
> +static int debugfs_run_open(struct inode *inode, struct file *file)
> +{
> + struct kunit_suite *suite;
> +
> + suite = (struct kunit_suite *)inode->i_private;
> +
> + return single_open(file, debugfs_run_print_results, suite);
> +}
> +
> +static const struct file_operations debugfs_run_fops = {
> + .open = debugfs_run_open,
> + .read = seq_read,
> + .llseek = seq_lseek,
> + .release = debugfs_release,
> +};
> +
> void kunit_debugfs_create_suite(struct kunit_suite *suite)
> {
> /* First add /sys/kernel/debug/kunit/<testsuite> */
> @@ -96,6 +126,9 @@ void kunit_debugfs_create_suite(struct kunit_suite *suite)
> debugfs_create_file(KUNIT_DEBUGFS_RESULTS, S_IFREG | 0444,
> suite->debugfs,
> suite, &debugfs_results_fops);
> + debugfs_create_file(KUNIT_DEBUGFS_RUN, S_IFREG | 0444,
> + suite->debugfs,
> + suite, &debugfs_run_fops);

Should anyone be able to read this? I think I agree since I am of the
opinion that people shouldn't build or load tests into a production
environment, but still I think it should be brought up.

I was actually talking to David the other day and we had the idea that
maybe KUnit should taint the kernel after tests run or after a
failure. Maybe that might communicate to a user that after running
tests the kernel shouldn't be used for production purposes.
(Obviously, I don't expect you to make that change here, the point of
anyone being able to cause tests to run just made me think of it.)
What do you think?

> }
>
> void kunit_debugfs_destroy_suite(struct kunit_suite *suite)
> --
> 1.8.3.1
>