FreeRTOS 7.4.1 sometimes halts in printf after vTaskStartScheduler starts (beagleboard)

liujinsuiyue wrote on Friday, October 11, 2013:


I ported FreeRTOS 7.4.1 to beagleboard with simple uart support.
My problem is that as long as the scheduler not starts the printf function works fine, but the system hangs sometimes after i implement a printf after the scheduler start.
For example if printf(“the test begins\n”); then only “the tes” will show on the console. But sometimes, let’s say 1 out of 10 times restart the system works fine.

any ideas about this problem? Any suggestion would be appreciated.

rtel wrote on Friday, October 11, 2013:

On small systems printf() nearly always overflows the stack if you are using Newlib or equivalent. For that reason a really tiny third party version is provided with FreeRTOS. If you search FreeRTOS/Demo for a file called printf-stdarg.c you will find lots of copies.

Do you have stack overflow detection turned on?

However, on your hardware I’m guessing you have plenty of stack - so it is more likely to be a reentracy issue. Do you have any kind of mutual exclusion around the printf() function or the UART?

If you are using Newlib then there is some basic support for its impure_ptr implementation to try and make it reentrant, but you would need to update your FreeRTOS version to get access to it. It is a FreeRTOSConfig.h option.


liujinsuiyue wrote on Friday, October 11, 2013:

Hi Richard,

Thank you for your reply.
Forgot to mention: the system also tend to halt when i call vTaskDelay(). I think that means the problem is not related to printf function.
FYI, i use a similar third party printf as printf-stdarg.c in this project.
I will turn on the stack overflow detection to see if that is the problem.
Any other opinions?

Best regards.

liujinsuiyue wrote on Friday, October 11, 2013:


When i put the code like the following:
then the system will not hang because of printf but still stop when calling vTaskDelay().