[ On Tue, April 14, 1998 at 19:52:24 (+0200), Stefan Grefen wrote: ]
> Subject: Re: Real vfork() (was: third results) 
>
> My $.02:
> Keep vfork as it is now, until we have a cheap enough COW version of it.
> Document that the existing behaviour will go away and give applications a
> way of figuring out which one is active. (A well known define for source
> code and a sysctl for binaries ...).
> We have quick vfork now, and implementing a way to detect it shouldn't
> take too much time. Than start working on the vm system ...

I don't think I'd go quite that far.  I think we should at minimum back
down on the shared address space (eg. what would it cost to
write-protect most of the parent address space and then unprotect it
before unblocking the parent, perhaps through a single per-process flag?).

Even more interesting would be to make the vfork() semantics switchable
at run-time through sysctl.  Then the user-land would really get a
regular workout and hopefully we'd be protected from evil programmers
abusing the nasty side-effects.

-- 
							Greg A. Woods

+1 416 443-1734      VE3TCP      <gwoods@acm.org>      <robohack!woods>
Planix, Inc. <woods@planix.com>; Secrets of the Weird <woods@weird.com>