tests/posix/timer.test_timer fails on nrf51_pca10028

[inakypg]

commit 4e8438b

***** delaying boot 1000ms (per build configuration) *****
***** Booting Zephyr OS 1.12.0-rc1 (delayed boot 1000ms) *****
Running test suite test_posix_timer
===================================================================
starting test - test_timer
POSIX timer test
Timer fires every 0 secs and  100000000 nsecs
Time remaining to fire 0 secs and  890000000 nsecs
Handler Signal value :20 for 1 times
Handler Signal value :20 for 2 times
Handler Signal value :20 for 3 times
Handler Signal value :20 for 4 times
Handler Signal value :20 for 5 times
Handler Signal value :20 for 6 times
Handler Signal value :20 for 7 times
Handler Signal value :20 for 8 times
Handler Signal value :20 for 9 times
Handler Signal value :20 for 10 times
Handler Signal value :20 for 11 times
Handler Signal value :20 for 12 times

    Assertion failed at tests/posix/timer/src/posix_timer.c:87: test_timer: (total_secs_timer not equal to secs_elapsed)
POSIX timer test has failed
FAIL - test_timer
===================================================================
===================================================================
RunID: ci-180524-0741-1662:knj5
PROJECT EXECUTION FAILED

[pizi-nordic]

Hi.

It looks that the problem is here:

int clock_gettime(clockid_t clock_id, struct timespec *ts)
    s64_t elapsed_msecs, elapsed_secs;
    s64_t elapsed_nsec, elapsed_cycles;

    if (clock_id != CLOCK_MONOTONIC) {
        errno = EINVAL;
        return -1;
    }

    elapsed_msecs = k_uptime_get();
    elapsed_secs = elapsed_msecs / MSEC_PER_SEC;

    elapsed_cycles = (s64_t) (k_cycle_get_32() %
                  sys_clock_hw_cycles_per_sec);
    elapsed_nsec = (s64_t) ((elapsed_cycles * NSEC_PER_SEC) /
                sys_clock_hw_cycles_per_sec);

    ts->tv_sec = (s32_t) elapsed_secs;
    ts->tv_nsec = (s32_t) elapsed_nsec;
    return 0;
}

As you see, the output of the function is calculated basing on 2 separate clock sources - the k_uptime_get() and k_cycle_get_32(). As result, the timestamp is not obtained in atomic way. There could be a timer interrupt between k_uptime_get() and k_cycle_get_32() which might result in dropping one second of time (elapsed_cycles will drop to zero but elapsed_secs will still hold old value). Disabling interrupts does not help much, as it only freezes time returned by k_uptime_get().

[nashif]

@ramakrishnapallala fyi

[nashif]

fixed by #8124