Re: Blocking on multiple threads with timeout

[Andrew Hacking <ahacking@xxxxxxxxxxxxxxx>]

On Mon, 2005-02-28 at 10:58, Jonathon McKitrick wrote:
> : Instead, you may want to re-consider the software that *allows* your
> : worker threads to block indefinitely, because it probably won't be
> 
> Well, on the first round, we check for the device we are talking to (serial,
> GPIB, IP, USB, etc.).  That is where a timeout is most likely to occur.  But
> this operation is performed serially, so no threads are spawned.  But after
> that, we spawn threads to initialize each device (which is full of motors
> and such) and this can take a minute or more.  Basically, I should be able
> to write the init thread to timeout if the 'ok' reply has not been received
> after the 'init' command has been sent to the device.  Is that what you are
> saying?
> 

Yes, pass a timeout to your init function (or via some other means if
required), and ensure that the code does not block indefinitely.  Your
main thread can just perform a pthread_join() on all the worker threads,
and *know* that it wont block much longer than the timeout.

> : operations that can block indefinitely.  This is much easier and more
> : reliable than the above high wire act, and often avoids the need for a
> : "manager" thread altogether.
> 
> I'd like to start the init threads, put the id's in an array, block until
> they are all done, then continue.  Each thread should have its own timeout
> handling internally.
> 
Sounds fine.


-Andrew.