On Sat, 13 Sep 2003, Matthias Buelow wrote:

: >vfork() was put back into NetBSD for a reason: it wins.
:
: By enough a margin to keep that inelegant crude hack in the API?

Yes, until a flexible enough replacement such as the (relatively recently
standardized!) posix_spawn() is commonplace.

: I mean, it was a workaround for a CPU bug -- the disability of
: COW on a certain VAX CPU due to microcode errors,

Even COW handling has a huge context dup overhead with fork(2) that is not
there with vfork(2).  Yes, even on i386.

-- 
-- Todd Vierling <tv@duh.org> <tv@pobox.com>