Well, as soon as doors make it into the linux kernel, I propose that
we eliminate mmap(). ;-)
(Humor aside:) I certainly must agree that dealing directly with
shared memory is the way to go for speed, however I've been mulling
over various test applications for doors where, for correctness and
security, you really want to pass a copy-on-write shared memory
buffer. Something like a video-capture server for image processing,
where the server continually sends frames to connected clients: A
slow client would be assured of a consistent image (which is hard to
implement with mmap() or shmat()) and fast clients would tend to share
pages.
- J
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu