Re: [RFC][PATCH][2.5] Possible race in wait_task_zombie and finish_task_switch

[Mika Penttilä]

Zwane Mwaikambo wrote:

>On Sun, 25 May 2003, Ingo Molnar wrote:
>
>  
>
>>On Sun, 25 May 2003, Zwane Mwaikambo wrote:
>>
>>    
>>
>>>	if (prev->state & (TASK_DEAD | TASK_ZOMBIE))
>>>		put_task_struct(prev);
>>>      
>>>
>>we initialize tsk->usage with 2 in fork() - are you sure the removal of
>>the above code will not result in a memory leak?
>>    
>>
>
>Isn't current the forked task? Also we initialise the forked task's state 
>to TASK_UNINTERRUPTIBLE.
>
>	Zwane
>  
>

The put_task_struct() above is for dropping our own reference after we 
exit. The other reference gets dropped by who is waiting for us.

--Mika