Avalanche of interruptions and stack sizes

[Rafael Morales]

Dear Sebastian,

Thank you for your response.

Finally, the problem was not in the interruption and the stack size, it 
was in the data generated by the interruption that generates an EDAC 
problem.
So, I was confused about the source of the problem.

In any case, thank you for your support.

Greetings from Spain.
> On 28/08/14 08:28, Rafael Morales wrote:
>> Dear all,
>>
>> I'm working with Leon3 at 25Mhz, RTEMS 4.10 and a list of devices that
>> generates interruptions.
>>
>> Nominally, there are 18 tasks running and the program is stable, but 
>> at certain
>> time, one of the devices generates an avalanche of interruptions and the
>> program stops suddenly.
>>
>> The only way to stabilize  program in is to increase the BSP stack 
>> (If I'm not
>> wrong it also increases the ISR and task stack).
>>
>> But, why?,  Which is the effect of a avalanche of interruption in the 
>> stack size?
>>
>> Any idea?
>
> This can have several reasons and it is hard to determine the error 
> source with the information provided by you (e.g. what means "stops 
> suddenly"?).  Each thread needs enough stack space to provide one 
> interrupt frame.  In case you use nested interrupts, then the 
> interrupt stack must be large enough for the deepest nesting.  In case 
> spurious interrupts lead to interrupt nesting, then this is a severe 
> problem.
>