Posix semaphore between different processes

[Ulrich Drepper]

Kaz Kylheku <kaz@ashi.footprints.net> writes:

> Why? Function pointers be used to bind a different set of operations for a
> process-shared semaphore at run time.  Going to the kernel for inter-thread
> semaphore operations would be a drag.

We always go though the kernel, we use signals and this is much more
costly.  Of course there are optimizations possible and this is using
an m:n model which is honestly considered.

-- 
---------------.                          ,-.   1325 Chesapeake Terrace
Ulrich Drepper  \    ,-------------------'   \  Sunnyvale, CA 94089 USA
Red Hat          `--' drepper at redhat.com   `------------------------