Hello,
ncs 2.6.0
I'm encountering a weird bus data error when I'm calling psa_generate_random inside a work queue thread (it has name "api"):
00> [00:01:54.764,038] <err> os: ***** BUS FAULT ***** 00> [00:01:54.764,068] <err> os: Precise data bus error 00> [00:01:54.764,068] <err> os: BFAR Address: 0x40f04fc3 00> [00:01:54.764,099] <err> os: r0/a1: 0x40f04fb5 r1/a2: 0xffffffff r2/a3: 0x20003678 00> [00:01:54.764,099] <err> os: r3/a4: 0x00000000 r12/ip: 0xffffffff r14/lr: 0x00014281 00> [00:01:54.764,099] <err> os: xpsr: 0xa1000000 00> [00:01:54.764,129] <err> os: Faulting instruction address (r15/pc): 0x0002b4aa 00> [00:01:54.764,160] <err> os: >>> ZEPHYR FATAL ERROR 25: Unknown error on CPU 0 00> [00:01:54.764,190] <err> os: Current thread: 0x20003678 (api)
The code I'm running inside work queue is the following:
uint8_t test[16];
psa_status_t status;
status = psa_generate_random(test, ARRAY_SIZE(test));
if (status != PSA_SUCCESS) {
LOG_ERR("gen rand err: %d", status);
return;
}
LOG_HEXDUMP_INF(test, ARRAY_SIZE(test), "test data");
Faulting instruction address points to: workspace/zephyr/kernel/mutex.c:147 , which contains the following code (please see my comment in the code):
SYS_PORT_TRACING_OBJ_FUNC_BLOCKING(k_mutex, lock, mutex, timeout);
new_prio = new_prio_for_inheritance(_current->base.prio,
mutex->owner->base.prio); // THIS IS LINE #147
LOG_DBG("adjusting prio up on mutex %p", mutex);
r14/lr points to: workspace/nrfxlib/crypto/nrf_cc310_platform/src/nrf_cc3xx_platform_mutex_zephyr.c:306, which contains the following code (please see my comment again):
ret = k_mutex_lock(p_mutex, K_FOREVER);
if (ret == 0) { // !!!! This is line #306 !!!!
return NRF_CC3XX_PLATFORM_SUCCESS;
} else {
return NRF_CC3XX_PLATFORM_ERROR_MUTEX_FAILED;
}
If I comment psa_generate_random - everything works just fine.
The fun part of this bug - it appears only when the data published to the work queue is from USB CDC ACM. When the same data comes from BLE, the psa_generate_random (in the same place) works just fine. So I might assume it could be somehow related to USB.
Please help to resolve this issue.
Kind regards,
Oleh Hordiichuk
