I only know how to do it from IDE. You have to read user manual it is 
better written there than I can do here.

Rodrigo Campos wrote:
> I will try to have access and try kernel tracing. By any chance do you have some "hello world" example I can try ? :)
>
> Sorry, I'm not sure if I  understand what you mean with pthread_exit. I was not clear, but I'm not using pthread_exit 
in any place of the code, so it shouldn't be pthread_exit either.
>
>
>
>
> Thanks a lot, really
>
> Rodrigo
>
>   
>> You can run kernel tracing. But you need "access" for this. And you 
>> ignored my comment about pthread_exit
>>
>> Rodrigo Campos wrote:
>>     
>>> I don't have access to the machine right now.
>>> But I'm quite sure pthread_cancel is not being called, since I've added 
>>>       
>> prints (write + fsync) and when the problem happens the prints are not shown.
>>     
>>> Any other ideas ?
>>>
>>>   
>>>       
>>>> Can you debug it? Put breakpoints on pthread_cancel, pthread_exit.
>>>>
>>>>
>>>> Rodrigo Campos wrote:
>>>>     
>>>>         
>>>>> Hi,
>>>>>
>>>>> I have a process that runs many threads. One of the threads is created by 
>>>>>           
>> a 
>>     
>>>>>       
>>>>>           
>>>> library it (dynamically) links to.
>>>>     
>>>>         
>>>>> The process start running OK (pidin shows all threads) and it seems to run
>>>>>           
>>  
>>     
>>>>>       
>>>>>           
>>>> OK until some point of time where suddenly some threads are not there. 
>>>>         
>> There 
>>     
>>>> are many threads that disappear (pidin does not show it) and, if it gives 
>>>>         
>> more
>>     
>>>>  information to anyone, the thread created by the library it links to is 
>>>>         
>> gone 
>>     
>>>> too.
>>>>     
>>>>         
>>>>> There are parts of the code that pthread_cancel() that threads, but I've 
>>>>>       
>>>>>           
>>>> added prints (write + fsync) and is not being executed. So I *really* don't
>>>>         
>>  
>>     
>>>> understand what could be happening to make the threads disappear.
>>>>     
>>>>         
>>>>> Also, always the remaining threads are in "REPLY" state
>>>>>
>>>>> Does this rings any bell to any of you guys ? What can make a thread *
>>>>>       
>>>>>           
>>>> disappear* ?
>>>>     
>>>>         
>>>>> Is there something (I don't really know) to configure that makes a null de
>>>>>           
>> -
>>     
>>>>>       
>>>>>           
>>>> reference instead of SEGFAULT all the process only a thread or something 
>>>>         
>> like 
>>     
>>>> that ?
>>>>     
>>>>         
>>>>> Or some suggestion how to debug further this problem ?
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> Thanks a lot,
>>>>> Rodrigo
>>>>>
>>>>>
>>>>>
>>>>>...