tests/kernel/tickless/tickless_concept fails on nRF5x with 1000msec boot delay

[pswarnak]

The test tests/kernel/tickless/tickless_concept results in Assertion failure in test_tickless_slice subtestcase, when boot delay of 1000msec is added.

Arch: arm
Board: nrf51/52 - nrf51_pca10028:arm, nrf52_pca10040:arm, nrf52840_pca10056:arm
Zephyr commit: 6a01c69
Toolchain: ZEPHYR_TOOLCHAIN_VARIANT=zephyr
Additional Info: Add CONFIG_BOOT_DELAY=1000 in tests/kernel/tickless/tickless_concept/prj.conf

Error Console Log:

***** delaying boot 1000ms (per build configuration) *****
***** Booting Zephyr OS v1.13.0-rc1-69-g6a01c69 (delayed boot 1000ms) *****
Running test suite tickless_concept
===================================================================
starting test - test_tickless_sysclock
time 1027, 1237
time 1247, 1446
PASS - test_tickless_sysclock
===================================================================
starting test - test_tickless_slice
elapsed slice 110, expected: <99, 109>

    Assertion failed at /home/pswarnak/workspace/1.13_execution/zephyr/tests/kernel/tickless/tickless_concept/src/main.c:59: thread_tslice: t <= SLICE_SIZE_LIMIT is false

elapsed slice 10, expected: <99, 109>

    Assertion failed at /home/pswarnak/workspace/1.13_execution/zephyr/tests/kernel/tickless/tickless_concept/src/main.c:57: thread_tslice: t >= SLICE_SIZE is false

elapsed slice 20, expected: <99, 109>

    Assertion failed at /home/pswarnak/workspace/1.13_execution/zephyr/tests/kernel/tickless/tickless_concept/src/main.c:57: thread_tslice: t >= SLICE_SIZE is false

elapsed slice 20, expected: <99, 109>

    Assertion failed at /home/pswarnak/workspace/1.13_execution/zephyr/tests/kernel/tickless/tickless_concept/src/main.c:57: thread_tslice: t >= SLICE_SIZE is false

FAIL - test_tickless_slice
===================================================================
===================================================================
RunID: :r7eb
PROJECT EXECUTION FAILED

Steps to reproduce:

cd zephyr/tests/kernel/tickless/tickless_concept
rm -rf build && mkdir build && cd build
cmake -D BOARD=nrf51_pca10028 ../
make BOARD=nrf51_pca10028 flash

Check the console log.

[pswarnak]

Hi @pizi-nordic,
Can you please check the failure, #9137 still does not resolve the failure, when boot delay of 1000msec is added.

[pswarnak]

Just figured out that tests/kernel/sched/schedule_api/testcase.yaml#kernel.sched.slice_reset also fails by adding 1000msec boot delay on nRF5x boards. Please find console output below:

***** delaying boot 1000ms (per build configuration) *****
***** Booting Zephyr OS 1.13.0-rc1 (delayed boot 1000ms) *****
Running test suite threads_scheduling
===================================================================
starting test - test_priority_cooperative
PASS - test_priority_cooperative
===================================================================
starting test - test_priority_preemptible
PASS - test_priority_preemptible
===================================================================
starting test - test_yield_cooperative
PASS - test_yield_cooperative
===================================================================
starting test - test_sleep_cooperative
PASS - test_sleep_cooperative
===================================================================
starting test - test_sleep_wakeup_preemptible
PASS - test_sleep_wakeup_preemptible
===================================================================
starting test - test_pending_thread_wakeup
PASS - test_pending_thread_wakeup
===================================================================
starting test - test_time_slicing_preemptible
PASS - test_time_slicing_preemptible
===================================================================
starting test - test_time_slicing_disable_preemptible
PASS - test_time_slicing_disable_preemptible
===================================================================
starting test - test_lock_preemptible
PASS - test_lock_preemptible
===================================================================
starting test - test_unlock_preemptible
PASS - test_unlock_preemptible
===================================================================
starting test - test_sched_is_preempt_thread
PASS - test_sched_is_preempt_thread
===================================================================
starting test - test_slice_reset

    Assertion failed at /home/jenkins/workspace/zephyr-master-tcf-v0.11-branch/LABEL/verify/SHARD/2-3/ZEPHYR_GCC_VARIANT/zephyr/zephyr.git/tests/kernel/sched/schedule_api/src/test_sched_timeslice_reset.c:46: thread_tslice: t <= expected_slice_max is false


    Assertion failed at /home/jenkins/workspace/zephyr-master-tcf-v0.11-branch/LABEL/verify/SHARD/2-3/ZEPHYR_GCC_VARIANT/zephyr/zephyr.git/tests/kernel/sched/schedule_api/src/test_sched_timeslice_reset.c:45: thread_tslice: t >= expected_slice_min is false


    Assertion failed at /home/jenkins/workspace/zephyr-master-tcf-v0.11-branch/LABEL/verify/SHARD/2-3/ZEPHYR_GCC_VARIANT/zephyr/zephyr.git/tests/kernel/sched/schedule_api/src/test_sched_timeslice_reset.c:45: thread_tslice: t >= expected_slice_min is false

FAIL - test_slice_reset
===================================================================
starting test - test_slice_scheduling
PASS - test_slice_scheduling
===================================================================
starting test - test_priority_scheduling
ABCDEFGHI
ABCDEFGHI
ABCDEFGHI
ABCDEFGHI
ABCDEFGHI
PASS - test_priority_scheduling
===================================================================
starting test - test_wakeup_expired_timer_thread
PASS - test_wakeup_expired_timer_thread
===================================================================
===================================================================
RunID: ci-180828-1458-2110:mp9z
PROJECT EXECUTION FAILED

[andyross]

Should be fixed as of the timer rework. I just retried this on a nrf52840_pca10056, with and without BOOT_DELAY=1000 and it was reliable across 20 tries each.

[carlescufi]

@pizi-nordic can you please re-run and close if fixed?

[pizi-nordic]

I have ran 100 iterations without delay and 100 iterations with delay. There was no even single failure. Closing.