FreeRTOS mutex example crashes

[Mat Py]

Hi there,
I'm using the ESP-IDF with PlatformIO for a small project on ESP32. I should use mutexes but I keep getting a runtime error. To understand FreeRTOS mutexes from scratch, I created a test project with code from the freeRTOS book:

Code: [Select all] [Expand/Collapse]

1. #include "freertos/FreeRTOS.h"
2. #include "freertos/task.h"
3. #include "freertos/semphr.h"
4. #include "stdio.h"
5.  
6. SemaphoreHandle_t xMutex = NULL;
7. static void prvNewPrintString( const char *pcString );
8. static void prvPrintTask( void *pvParameters );
9.  
10. static void prvPrintTask( void *pvParameters )
11. {
12.     char *pcStringToPrint;
13.     const TickType_t xMaxBlockTimeTicks = 0x20;
14.     pcStringToPrint = ( char * ) pvParameters;
15.     for( ;; )
16.     {
17.         prvNewPrintString( pcStringToPrint );
18.         vTaskDelay( ( rand() % xMaxBlockTimeTicks ) );
19.     }
20. }
21.  
22. static void prvNewPrintString( const char *pcString )
23. {
24.     xSemaphoreTake( xMutex, portMAX_DELAY );
25.     {
26.         printf( "%s", pcString );
27.         fflush( stdout );
28.     }
29.     xSemaphoreGive( xMutex );
30. }
31.  
32. void app_main() {
33.     xMutex = xSemaphoreCreateMutex();
34.     if( xMutex != NULL )
35.     {
36.         xTaskCreate( prvPrintTask, "Print1", 1000,
37.         "Task 1 ***************************************\r\n", 1, NULL );
38.         xTaskCreate( prvPrintTask, "Print2", 1000,
39.         "Task 2 ---------------------------------------\r\n", 2, NULL );
40.     }
41.     for( ;; );
42. }

GeSHi © Codebox Plus Extension

The only modification I made is removing

Code: Select all

vTaskStartScheduler();

since app_main is already running with the scheduler.
Unfortunately, this code crashes. Here is the panic message:

Code: [Select all] [Expand/Collapse]

1. I (0) cpu_start: Starting scheduler on APP CPU.
2. Guru Meditation Error: Core  1 panic'ed (IllegalInstruction). Exception was unhandled.
3. Memory dump at 0x400824e8: e5f875a1 f01d0227 52006136
4. Core  1 register dump:
5. PC      : 0x400824ee  PS      : 0x00060230  A0      : 0x3ffb6ed4  A1      : 0x3ffb6650
6. A2      : 0x3ffae978  A3      : 0x00000004  A4      : 0x00060e20  A5      : 0x00060e23  
7. A6      : 0x00000001  A7      : 0x00000000  A8      : 0x800824ee  A9      : 0x3ffb6630  
8. A10     : 0x3ffb21e4  A11     : 0x00060023  A12     : 0x00060020  A13     : 0x00000000
9. A14     : 0x00000000  A15     : 0x00000000  SAR     : 0x00000000  EXCCAUSE: 0x00000000
10. EXCVADDR: 0x00000000  LBEG    : 0x00000000  LEND    : 0x00000000  LCOUNT  : 0x00000000  
11.  
12. Backtrace:0x400824eb:0x3ffb6650 0x3ffb6ed1:0x3ffb6670 |<-CORRUPTED

GeSHi © Codebox Plus Extension

Even if the ESP-IDF FreeRTOS implementation is slightly different than the original to handle both cores, this example shouldn't run properly?
I have also no idea how to decode the register dump and the backtrace is not very helpful.

Can you tell me what's wrong here?
Thanks,
MatPy

[burtrum]

Hi,
When using printf, a task stack of 1000 is considered small. Try a bigger stack, 2048.

Burtrum

[Mat Py]

Thank you, this is the solution. I added some additional "debug" printf then the stack overflow showed clearly in the backtrace, unlike what I got before with no information. Adding more stack space to the tasks solved the problem.
And the error in my project was not related to this one...