fix(newlib): usleep returning early

This commit updates usleep() to always sleep for the required sleep
period or more. This fixes a bug where the usleep() could sleep for less
than the request sleep period.

Closes https://github.com/espressif/esp-idf/pull/15132
This commit is contained in:
Stephen Noonan
2025-01-03 16:47:19 -05:00
committed by Sudeep Mohanty
parent 21a3ace68d
commit b47f25ac23

View File

@ -1,5 +1,5 @@
/* /*
* SPDX-FileCopyrightText: 2015-2022 Espressif Systems (Shanghai) CO LTD * SPDX-FileCopyrightText: 2015-2025 Espressif Systems (Shanghai) CO LTD
* *
* SPDX-License-Identifier: Apache-2.0 * SPDX-License-Identifier: Apache-2.0
*/ */
@ -201,10 +201,26 @@ int usleep(useconds_t us)
if (us < us_per_tick) { if (us < us_per_tick) {
esp_rom_delay_us((uint32_t) us); esp_rom_delay_us((uint32_t) us);
} else { } else {
/* since vTaskDelay(1) blocks for anywhere between 0 and portTICK_PERIOD_MS, /* vTaskDelay may return up to (n-1) tick periods due to the tick ISR
* round up to compensate. being asynchronous to the call. We must sleep at least the specified
*/ time, or longer. Checking the monotonic clock allows making an
vTaskDelay((us + us_per_tick - 1) / us_per_tick); additional call to vTaskDelay when needed to ensure minimal time is
actually slept. Adding `us_per_tick - 1` prevents ever passing 0 to
vTaskDelay().
*/
uint64_t now_us = esp_time_impl_get_time();
uint64_t target_us = now_us + us;
do {
vTaskDelay((((target_us - now_us) + us_per_tick - 1) / us_per_tick));
now_us = esp_time_impl_get_time();
/* It is possible that the time left until the target time is less
* than a tick period. However, we let usleep() to sleep for an
* entire tick period. This, could result in usleep() sleeping for
* a longer time than the requested time but that does not violate
* the spec of usleep(). Additionally, it allows FreeRTOS to schedule
* other tasks while the current task is sleeping.
*/
} while (now_us < target_us);
} }
return 0; return 0;
} }