Re: [Gm2] GM2 coroutines

[Gaius Mulley]

> Benjamin Kowarsch <address@hidden> writes:
>
>> Hi
>>
>> Recently I was looking at the status of GM2 again to see if we might be able 
>> to
>> use it for embedded work under FreeBSD.
>>
>> From a paper by Gaius about GM2, it would seem to me that coroutines had been
>> implemented using the pthread library. I presume this was an interim shortcut
>> since of course coroutines are not meant to be preempted.
>>
>> For what we do, it is of paramount importance to have true coroutines instead
>> of pthreads. Therefore, I am wondering if true coroutines have since been
>> implemented, or if not, whether there are any plans to do so.

Hi Benjamin,

just to clarify and answer the questions directly - the implementation
provides the PIM coroutine procedures TRANSFER, IOTRANSFER, NEWPROCESS
and LISTEN, they are all non-preemptive.  In fact I thought this was a
slight weakness of the implementation in that IOTRANSFER will only
return to the original callee when another process calls LISTEN or
calls a procedure which is in a module with a lower interrupt priority
or explicitly calls a runtime support procedure function
TurnInterrupts (used to build LISTEN).  All other code is treated as
being atomic.  Still I believe this satisfies the semantics of
IOTRANSFER, TRANSFER et al.

regards,
Gaius