[H-GEN] pthreads + communication in concurrent environments

[Martin Pool]

[ Humbug *General* list - semi-serious discussions about Humbug and
Unix-related topics. ]

At 03:33 PM 26/08/99 +1000, you wrote:
>so my quesions follow:  Has anyone on the list had good/bad experiences
>with various Unix/posixish message passing systems, particularly in
>contrast to systems based on shared memory and synchronisation
>primatives?  

There was a discussion a while ago of implementing a QNX-like single-copy
message-passing system on top of Linux.  It's quite a cool design: I'd be
interested to hear from people who've used it.  Since this requires a copy
directly from the address-space of one process into that of another it
can't be done directly by Linux because it only maps a single process at a
time when in kernel mode.  One workaround is to copy into a temporary
buffer somewhere and then out again, so it can be emulated perfectly well
on top of existing IPC, but will be slower.  Another is to make the entire
address space of all the communicating processes shared, but that's pretty
ugly.

The trend in large distributed Java systems seems to be to concentrate
concurrency control in a few places and then make statements about whole
swathes of the system like "called only from the Swing thread" or "called
by several CORBA threads concurrently."


--
Martin Pool

--
This is list (humbug) general handled by majordomo at lists.humbug.org.au .
Postings only from subscribed addresses of lists general or general-post.