kernel/linux.git/tools/testing/selftests/cgroup, branch v6.18.21

selftests: cgroup: Use values_close_report in test_cpu

2025-10-15T15:00:59+00:00

Convert test_cpu to use the newly added values_close_report() helper to print detailed diagnostics when a tolerance check fails. This provides clearer insight into deviations while run in the CI. Signed-off-by: Sebastian Chlad Signed-off-by: Tejun Heo

selftests: cgroup: add values_close_report helper

2025-10-15T15:00:49+00:00

Some cgroup selftests, such as test_cpu, occasionally fail by a very small margin and if run in the CI context, it is useful to have detailed diagnostic output to understand the deviation. Introduce a values_close_report() helper which performs the same comparison as values_close(), but prints detailed information when the values differ beyond the allowed tolerance. Signed-off-by: Sebastian Chlad Signed-off-by: Tejun Heo

selftests: cgroup: Make test_pids backwards compatible

2025-08-27T15:59:52+00:00

The predicates in test expect event counting from 73e75e6fc352b ("cgroup/pids: Separate semantics of pids.events related to pids.max") and the test would fail on older kernels. We want to have one version of tests for all, so detect the feature and skip the test on old kernels. (The test could even switch to check v1 semantics based on the flag but keep it simple for now.) Fixes: 9f34c566027b6 ("selftests: cgroup: Add basic tests for pids controller") Signed-off-by: Michal Koutný Tested-by: Sebastian Chlad Signed-off-by: Tejun Heo

cgroup: selftests: Add tests for freezer time

2025-08-22T17:50:52+00:00

Test cgroup v2 freezer time stat. Freezer time accounting should be independent of other cgroups in the hierarchy and should increase iff a cgroup is CGRP_FREEZE (regardless of whether it reaches CGRP_FROZEN). Skip these tests on systems without freeze time accounting. Signed-off-by: Tiffany Yang Cc: Michal Koutný Signed-off-by: Tejun Heo

selftests/cgroup: fix cpu.max tests

2025-07-17T18:12:19+00:00

Current cpu.max tests (both the normal one and the nested one) are broken. They setup cpu.max with 1000 us quota and the default period (100,000 us). A cpu hog is run for a duration of 1s as per wall clock time. This corresponds to 10 periods, hence an expected usage of 10,000 us. We want the measured usage (as per cpu.stat) to be close to 10,000 us. Previously, this approximate equality test was done by `!values_close(usage_usec, expected_usage_usec, 95)`: if the absolute difference between usage_usec and expected_usage_usec is greater than 95% of their sum, then we pass. And expected_usage_usec was set to 1,000,000 us. Mathematically, this translates to the following being true for pass: |usage - expected_usage| > (usage + expected_usage)*0.95 If usage > expected_usage: usage - expected_usage > (usage + expected_usage)*0.95 0.05*usage > 1.95*expected_usage usage > 39*expected_usage = 39s If usage < expected_usage: expected_usage - usage > (usage + expected_usage)*0.95 0.05*expected_usage > 1.95*usage usage < 0.0256*expected_usage = 25,600 us Combined, Pass if usage < 25,600 us or > 39 s, which makes no sense given that all we need is for usage_usec to be close to 10,000 us. Fix this by explicitly calcuating the expected usage duration based on the configured quota, default period, and the duration, and compare usage_usec and expected_usage_usec using values_close() with a 10% error margin. Also, use snprintf to get the quota string to write to cpu.max instead of hardcoding the quota, ensuring a single source of truth. Remove the check comparing user_usec and expected_usage_usec, since on running this test modified with printfs, it's seen that user_usec and usage_usec can regularly exceed the theoretical expected_usage_usec: $ sudo ./test_cpu user: 10485, usage: 10485, expected: 10000 ok 1 test_cpucg_max user: 11127, usage: 11127, expected: 10000 ok 2 test_cpucg_max_nested $ sudo ./test_cpu user: 10286, usage: 10286, expected: 10000 ok 1 test_cpucg_max user: 10404, usage: 11271, expected: 10000 ok 2 test_cpucg_max_nested Hence, a values_close() check of usage_usec and expected_usage_usec is sufficient. Fixes: a79906570f9646ae17 ("cgroup: Add test_cpucg_max_nested() testcase") Fixes: 889ab8113ef1386c57 ("cgroup: Add test_cpucg_max() testcase") Acked-by: Michal Koutný Signed-off-by: Shashank Balaji Signed-off-by: Tejun Heo

selftests: cgroup: Fix missing newline in test_zswap_writeback_one

2025-07-12T17:35:30+00:00

Fixes malformed test output due to missing newline Signed-off-by: Sebastian Chlad Signed-off-by: Tejun Heo

selftests: cgroup: Allow longer timeout for kmem_dead_cgroups cleanup

2025-07-12T17:34:49+00:00

The test_kmem_dead_cgroups test currently assumes that RCU and memory reclaim will complete within 5 seconds. In some environments this timeout may be insufficient, leading to spurious test failures. This patch introduces max_time set to 20 which is then used in the test. After 5th sec the debug message is printed to indicate the cleanup is still ongoing. In the system under test with 16 CPUs the original test was failing most of the time and the cleanup time took usually approx. 6sec. Further tests were conducted with and without do_rcu_barrier and the results (respectively) are as follow: quantiles 0 0.25 0.5 0.75 1 1 2 3 8 20 (mean = 4.7667) 3 5 8 8 20 (mean = 7.6667) Acked-by: Michal Koutny Signed-off-by: Sebastian Chlad Signed-off-by: Tejun Heo

selftests: cgroup: Fix compilation on pre-cgroupns kernels

2025-06-17T18:13:25+00:00

The test would be skipped because of nsdelegate, so the defined value is not used (0 is always acceptable). Signed-off-by: Michal Koutný Signed-off-by: Tejun Heo

selftests: cgroup: Optionally set up v1 environment

2025-06-17T18:13:16+00:00

Use the missing mount of the unifier hierarchy as a hint of legacy system and prepare our own named v1 hierarchy for tests. The code is only in test_core.c and not cgroup_util.c because other selftests are related to controllers which will be exposed on v2 hierarchy but named hierarchies are only v1 thing. Signed-off-by: Michal Koutný Signed-off-by: Tejun Heo

selftests: cgroup: Add support for named v1 hierarchies in test_core

2025-06-17T18:13:09+00:00

This comes useful when using selftests from mainline on older kernels/setups that still rely on cgroup v1. The core tests that rely on v2 specific features are skipped, the remaining ones are adjusted to work with a v1 hierarchy. The expected output on v1 system: ok 1 # SKIP test_cgcore_internal_process_constraint ok 2 # SKIP test_cgcore_top_down_constraint_enable ok 3 # SKIP test_cgcore_top_down_constraint_disable ok 4 # SKIP test_cgcore_no_internal_process_constraint_on_threads ok 5 # SKIP test_cgcore_parent_becomes_threaded ok 6 # SKIP test_cgcore_invalid_domain ok 7 # SKIP test_cgcore_populated ok 8 test_cgcore_proc_migration ok 9 test_cgcore_thread_migration ok 10 test_cgcore_destroy ok 11 test_cgcore_lesser_euid_open ok 12 # SKIP test_cgcore_lesser_ns_open Signed-off-by: Michal Koutný Signed-off-by: Tejun Heo