On Apr 24, 2007, at 7:20 PM, Thor Lancelot Simon wrote:

> Just out of curiosity, if the implementation uses an additional  
> thread in
> the user process, why is any kernel support required at all?

Maybe it uses an additional kernel thread in that user process,  
without requiring a pthread?  Anyway, if that's the case, it's good,  
because then we can change the implementation without exposing the  
details to user space.

I, for one, would like to see AIO that did not require a thread that  
could block, but rather use of completely asynchronous APIs that  
already exist in our kernel.

-- thorpej