This reverts commit fd8ca83e9a.
https://github.com/zephyrproject-rtos/zephyr/pull/69705
Introduced a regression in main in which
tests/subsys/logging/log_timestamp
started failing. (See
https://github.com/zephyrproject-rtos/zephyr/issues/72344
for more info).
Let's revert the PR. It can be submitted after with the issue
fixed.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
48 lines
1.1 KiB
C
48 lines
1.1 KiB
C
/*
|
|
* Copyright (c) 2018 Intel Corporation
|
|
*
|
|
* SPDX-License-Identifier: Apache-2.0
|
|
*/
|
|
|
|
#include <zephyr/ztest.h>
|
|
|
|
/**
|
|
* @brief Test delay during boot
|
|
* @defgroup kernel_init_tests Init
|
|
* @ingroup all_tests
|
|
* @{
|
|
*/
|
|
|
|
/**
|
|
* @brief This module verifies the delay specified during boot.
|
|
*/
|
|
ZTEST(boot_delay, test_bootdelay)
|
|
{
|
|
if (CONFIG_SYS_CLOCK_HW_CYCLES_PER_SEC > 1000000000) {
|
|
/* Systems with very fast counters (like the x86 TSC)
|
|
* and long firmware startup (often 10+ seconds on a
|
|
* EFI PC!) can easily roll this over during startup,
|
|
* and there's no way to detect that case with a 32
|
|
* bit OS API. Just skip it if we have a GHz-scale
|
|
* counter.
|
|
*/
|
|
ztest_test_skip();
|
|
}
|
|
|
|
uint32_t current_cycles = k_cycle_get_32();
|
|
|
|
/* compare this with the boot delay specified */
|
|
zassert_true(k_cyc_to_ns_floor64(current_cycles) >=
|
|
(NSEC_PER_MSEC * CONFIG_BOOT_DELAY),
|
|
"boot delay not executed: %d < %d",
|
|
(uint32_t)k_cyc_to_ns_floor64(current_cycles),
|
|
(NSEC_PER_MSEC * CONFIG_BOOT_DELAY));
|
|
}
|
|
|
|
extern void *common_setup(void);
|
|
ZTEST_SUITE(boot_delay, NULL, common_setup, NULL, NULL, NULL);
|
|
|
|
/**
|
|
* @}
|
|
*/
|