Re: [RFC PATCH v1.1 03/11] mm/damon/tests/core-kunit: test damon_rand()

[SeongJae Park <sj@kernel.org>]

On Thu, 25 Jun 2026 05:13:51 +0000 sashiko-bot@kernel.org wrote:

> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
> - [Medium] The KUnit test uses a non-deterministic seed with tight probability bounds, which will cause sporadic test failures.
> --
> 
> commit bc9f1073820b74c3fbb6ac52d5ca363629f99f3d
> Author: SeongJae Park <sj@kernel.org>
> 
> mm/damon/tests/core-kunit: test damon_rand()
> 
> Commit 9012c4e647df ("mm/damon: replace damon_rand() with a per-ctx
> lockless PRNG") optimized DAMON for better performance.  Add a kunit
> test for ensuring the pseudo randomness quality.
> 
> > diff --git a/mm/damon/tests/core-kunit.h b/mm/damon/tests/core-kunit.h
> > index 1cfb8c176b873..756f3b9e2ed3b 100644
> > --- a/mm/damon/tests/core-kunit.h
> > +++ b/mm/damon/tests/core-kunit.h
> 
> [ ... ]
> 
> > @@ -1460,6 +1460,26 @@ static void damon_test_is_last_region(struct kunit *test)
> >  	damon_free_target(t);
> >  }
> >  
> > +static void damon_test_rand(struct kunit *test)
> > +{
> > +	struct damon_ctx ctx;
> > +	int counts[10] = {};
> > +	int i;
> > +
> > +	prandom_seed_state(&ctx.rnd_state, get_random_u64());
> > +	for (i = 0; i < 10000; i++) {
> > +		unsigned long rnd = damon_rand(&ctx, 0, 10);
> > +
> > +		KUNIT_EXPECT_GE(test, rnd, 0);
> > +		KUNIT_EXPECT_LE(test, rnd, 9);
> > +		counts[rnd]++;
> > +	}
> > +	for (i = 0; i < 10; i++) {
> > +		KUNIT_EXPECT_GE(test, counts[i], 900);
> > +		KUNIT_EXPECT_LE(test, counts[i], 1100);
> 
> [Severity: Medium]
> The test uses a true random seed to initialize the PRNG state, making the
> test output non-deterministic. Is it possible that these tight expectation
> bounds will cause sporadic test failures in automated CI environments?
> 
> Statistically, a binomial distribution with n=10,000 and p=0.1 has a
> standard deviation of 30. The allowed variation of +/- 100 is roughly
> +/- 3.33 standard deviations. The probability of any single bin falling
> outside this range is around 0.086%.
> 
> Over 10 bins, does this mean the test could spuriously fail about 0.86%
> of the time (roughly 1 in 117 runs)?

Good catch.  I will modify the test to widen the bound to [800, 1200].


Thanks,
SJ

[...]