Date: Sat, 13 Jan 1996 17:29:50 -0700 (MST) From: Terry Lambert <terry@lambert.org> To: msmith@atrad.adelaide.edu.au (Michael Smith) Cc: terry@lambert.org, msmith@atrad.adelaide.edu.au, gibbs@freefall.freebsd.org, phk@critter.tfs.com, jkh@time.cdrom.com, obrien@cs.ucdavis.edu, freebsd-hackers@FreeBSD.org Subject: Re: X for install Message-ID: <199601140029.RAA24502@phaeton.artisoft.com> In-Reply-To: <199601131402.AAA19555@genesis.atrad.adelaide.edu.au> from "Michael Smith" at Jan 14, 96 00:32:07 am
next in thread | previous in thread | raw e-mail | index | archive | help
> > The private address space has the advantage of isolating potentially
> > nasty ("scribble on the kernel") drivers, and the disadvantage of
> > additional protection domain crossing (ala NetWare 4.x "memory
> > protection").
>
> With all the hoohah on pipelining and speculative execution and all that
> crap, one never reads much about improvements in these issues (nor TLA and
> friends). Is this domain crossing still expensive on newer processors?
> (In context, I guess newer means P5+)
Doing protection domain crossing is more expensive than not doing it.
There *is* a method which uses the anonymity of pages in a "very large
address space" (say 64 bits), where it "just works" if you get the
right page and faults if you don't.
The magic is that with a very large space, you can "hide" pages with
a high statistical probability that it will be impossible to locate
a particular page without faulting instead, since there are many many
orders of magnitudes more unmapped pages than mapped ones.
I haven't seen this tried on anything less than a 64 bit processor.
Terry Lambert
terry@lambert.org
---
Any opinions in this posting are my own and not those of my present
or previous employers.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199601140029.RAA24502>