esp-idf
esp-idf copied to clipboard
espressif
uart deadlock while using automatic light sleep (IDFGH-11870)
Answers checklist.
- [X] I have read the documentation ESP-IDF Programming Guide and the issue is not addressed there.
- [X] I have updated my IDF branch (master or release) to the latest version and checked that the issue is present there.
- [X] I have searched the issue tracker for a similar issue and not found a similar issue.
IDF version.
v5.3-dev-1353-gb3f7e2c8a4 v5.2-beta1-263-ge49823f10c v4.4.6-374-gbb8dd9d35b
Espressif SoC revision.
ESP32-S3 (QFN56) (rev v0.1) ESP32-D0WD-V3 (rev v3.1) ESP32-C3 (QFN32) (rev v0.3)
Operating System used.
Windows
How did you build your project?
VS Code IDE
If you are using Windows, please specify command line type.
PowerShell
Development Kit.
ESP32-S3/C3 DevKits Wemos D1-Mini ESP32 ESP32/S3 custom boards
Power Supply used.
External 3.3V, USB (irrelevant)
What is the expected behavior?
Uart is working with automatic light sleep enabled
What is the actual behavior?
Uart locks indefinitely with automatic light sleep enabled
Steps to reproduce.
app main:
#include "driver/uart.h"
#include "esp_pm.h"
#include "esp_timer.h"
#include "freertos/FreeRTOS.h"
#include "freertos/task.h"
#include "hal/uart_types.h"
const uart_port_t uart_nr = (uart_port_t)0;
const int rx_buffer_size = 256;
const int tx_buffer_size = 0;
const int queue_size = 20;
const int baud_rate = 115200;
const int rx_flow_ctrl_thresh = 112; // value taken from arduino core
QueueHandle_t uart_event_queue;
extern "C" void app_main()
{
uart_config_t uart_config;
uart_config.data_bits = uart_word_length_t::UART_DATA_8_BITS;
uart_config.parity = uart_parity_t::UART_PARITY_DISABLE;
uart_config.stop_bits = uart_stop_bits_t::UART_STOP_BITS_1;
uart_config.flow_ctrl = uart_hw_flowcontrol_t::UART_HW_FLOWCTRL_DISABLE;
uart_config.rx_flow_ctrl_thresh = rx_flow_ctrl_thresh;
uart_config.baud_rate = baud_rate;
uart_config.source_clk = UART_SCLK_DEFAULT;
ESP_ERROR_CHECK(uart_driver_install(uart_nr, rx_buffer_size, tx_buffer_size, queue_size, &uart_event_queue, 0));
ESP_ERROR_CHECK(uart_param_config(uart_nr, &uart_config));
esp_pm_config_t pm_config = {
.max_freq_mhz = 160,
.min_freq_mhz = 160,
.light_sleep_enable = true,
};
ESP_ERROR_CHECK(esp_pm_configure(&pm_config));
uart_wait_tx_idle_polling(uart_nr);
while (true)
{
char buf[1024];
int size = sprintf(buf, "Time from boot: %lld\n", (esp_timer_get_time() / 1000ULL));
uart_write_bytes(uart_nr, buf, size);
// uart_wait_tx_idle_polling(uart_nr);
vTaskDelay(1000 / portTICK_PERIOD_MS);
}
}
sdkconfig
CONFIG_PM_ENABLE=y
CONFIG_FREERTOS_HZ=1000
CONFIG_FREERTOS_USE_TICKLESS_IDLE=y
Debug Logs.
ESP-ROM:esp32c3-api1-20210207
Build:Feb 7 2021
rst:0x1 (POWERON),boot:0xc (SPI_FAST_FLASH_BOOT)
SPIWP:0xee
mode:DIO, clock div:1
load:0x3fcd5820,len:0x1714
load:0x403cc710,len:0x968
load:0x403ce710,len:0x2f9c
entry 0x403cc710
I (30) boot: ESP-IDF v5.1.2-602-gdb1e54a0c5 2nd stage bootloader
I (30) boot: compile time Jan 10 2024 19:56:27
I (31) boot: chip revision: v0.3
I (35) boot.esp32c3: SPI Speed : 80MHz
I (39) boot.esp32c3: SPI Mode : DIO
I (44) boot.esp32c3: SPI Flash Size : 2MB
I (49) boot: Enabling RNG early entropy source...
I (54) boot: Partition Table:
I (58) boot: ## Label Usage Type ST Offset Length
I (65) boot: 0 nvs WiFi data 01 02 00009000 00006000
I (73) boot: 1 phy_init RF data 01 01 0000f000 00001000
I (80) boot: 2 factory factory app 00 00 00010000 00100000
I (87) boot: End of partition table
I (92) esp_image: segment 0: paddr=00010020 vaddr=3c020020 size=09d40h ( 40256) map
I (107) esp_image: segment 1: paddr=00019d68 vaddr=3fc8d600 size=011d8h ( 4568) load
I (109) esp_image: segment 2: paddr=0001af48 vaddr=40380000 size=050d0h ( 20688) load
I (121) esp_image: segment 3: paddr=00020020 vaddr=42000020 size=1b1b0h (111024) map
I (143) esp_image: segment 4: paddr=0003b1d8 vaddr=403850d0 size=084c8h ( 33992) load
I (150) esp_image: segment 5: paddr=000436a8 vaddr=50000010 size=00004h ( 4) load
I (154) boot: Loaded app from partition at offset 0x10000
I (155) boot: Disabling RNG early entropy source...
I (172) cpu_start: Unicore app
I (172) cpu_start: Pro cpu up.
I (181) cpu_start: Pro cpu start user code
I (181) cpu_start: cpu freq: 160000000 Hz
I (181) cpu_start: Application information:
I (184) cpu_start: Project name: uart-lightsleep-deadlock
I (190) cpu_start: App version: 5a04b5d-dirty
I (196) cpu_start: Compile time: Jan 10 2024 19:56:11
I (202) cpu_start: ELF file SHA256: 788c280a5ec37e35...
I (208) cpu_start: ESP-IDF: v5.1.2-602-gdb1e54a0c5
I (214) cpu_start: Min chip rev: v0.3
I (219) cpu_start: Max chip rev: v1.99
I (224) cpu_start: Chip rev: v0.3
I (228) heap_init: Initializing. RAM available for dynamic allocation:
I (236) heap_init: At 3FC8F6E0 len 00030920 (194 KiB): DRAM
I (242) heap_init: At 3FCC0000 len 0001C710 (113 KiB): DRAM/RETENTION
I (249) heap_init: At 3FCDC710 len 00002950 (10 KiB): DRAM/RETENTION/STACK
I (256) heap_init: At 50000014 len 00001FD4 (7 KiB): RTCRAM
I (264) spi_flash: detected chip: generic
I (267) spi_flash: flash io: dio
W (271) spi_flash: Detected size(4096k) larger than the size in the binary image header(2048k). Using the size in the binary image header.
I (285) sleep: Configure to isolate all GPIO pins in sleep state
I (291) sleep: Enable automatic switching of GPIO sleep configuration
I (310) app_start: Starting scheduler on CPU0
I (310) main_]�I (316) pm: Frequency switching config: CPU_MAX: 160, APB_MAX: 160, APB_MIN: 160, Light sleep: ENABLED
Time from boot: 327
Time from boot: 132
More Information.
A problem seems to be in the uart.c line 1202:
while (size) {
//semaphore for tx_fifo available
1202 >> if (pdTRUE == xSemaphoreTake(p_uart_obj[uart_num]->tx_fifo_sem, (TickType_t)portMAX_DELAY)) { <<<<
uint32_t sent = uart_enable_tx_write_fifo(uart_num, (const uint8_t *) src, size);
if (sent < size) {
p_uart_obj[uart_num]->tx_waiting_fifo = true;
uart_enable_tx_intr(uart_num, 1, UART_EMPTY_THRESH_DEFAULT);
}
size -= sent;
src += sent;
}
}
cyclic xSemaphoreTake(p_uart_obj[uart_num]->tx_fifo_sem, with indefinite timeout. It obtains the lock, and on the next iteration tries to obtain it again and eventually hangs forever.
I've moved that xSemaphoreTake before while - and it started working as expected (on all three CPUs). However im not sure if this would be correct, because if we should wait for some ISR to happen - maybe lightsleep prevents this interrupt from trigger?
Taking into account Sleep Modes: UART Output Handling and available UART wakeup sources (TX detection), probably better solution might be wrapping that while loop with a esp_pm_lock_acquire and esp_pm_lock_release with a ESP_PM_NO_LIGHT_SLEEP lock type. Such kind of a fix also works as expected.
@tshcherban You are right! We are aware of this issue, the MR to add the PM lock to UART driver is under internal review.
@tshcherban You are right! We are aware of this issue, the MR to add the PM lock to UART driver is under internal review.
@songruo It has been 3 Months since your last reply.How is the status now?
@tshcherban You are right! We are aware of this issue, the MR to add the PM lock to UART driver is under internal review.
@songruo It has been 3 Months since your last reply.How is the status now?
@songruo Are you still working on this issue? Or someone else will help to fix it? @igrr
@tshcherban You are right! We are aware of this issue, the MR to add the PM lock to UART driver is under internal review.
@songruo When will this get fixed?
@tshcherban You are right! We are aware of this issue, the MR to add the PM lock to UART driver is under internal review.
@igrr So this never get fix and @songruo no longer response... Can someone else from espressif help to update the status?
It has been fixed in 235bb6f on master. Backports to earlier versions is undergoing.
Does APB_MAX prevent ESP from automatic lightsleep? If not - this issue is not done/closed
Does APB_MAX prevent ESP from automatic lightsleep?
Yes, APB_MAX lock should prevent entry into light sleep as well. In light sleep, peripherals and the APB are clock-gated (or even powered down, on chips which support peripheral state retention), so requesting the APB to run at the maximum available frequency also prevents entry into light sleep.
It has been fixed in 235bb6f on master. Backports to earlier versions is undergoing.
@songruo I still cannot find this fix in release branches. I'm wondering if v5.2.3 will include this fix?
@AxelLin Here you go https://github.com/espressif/esp-idf/commit/9b0d8032379d20fd28419382704881bdc09bcd6e