Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 24 Apr 2001 09:14:31 -0700
From:      Alfred Perlstein <bright@wintelcom.net>
To:        Poul-Henning Kamp <phk@critter.freebsd.dk>
Cc:        Garrett Wollman <wollman@khavrinen.lcs.mit.edu>, current@FreeBSD.ORG
Subject:   Re: vm_mtx
Message-ID:  <20010424091431.J1790@fw.wintelcom.net>
In-Reply-To: <59442.988126516@critter>; from phk@critter.freebsd.dk on Tue, Apr 24, 2001 at 05:35:16PM %2B0200
References:  <200104241527.LAA95978@khavrinen.lcs.mit.edu> <59442.988126516@critter>

next in thread | previous in thread | raw e-mail | index | archive | help
* Poul-Henning Kamp <phk@critter.freebsd.dk> [010424 08:36] wrote:
> In message <200104241527.LAA95978@khavrinen.lcs.mit.edu>, Garrett Wollman write
> s:
> ><<On Mon, 23 Apr 2001 21:51:31 -0700, Alfred Perlstein <bright@wintelcom.net> said:
> >
> >> You can find the work I've done so far to make a giant vm mutex
> >> here:
> >
> >The Mach code we originally inherited was supposed to already by
> >multiprocessor safe.  Did we manage to eliminate that capability?
> 
> I'm sure you are fully aware of the implications of the strategically
> placed "supposed" in your own sentence.  I have never heard anybody
> get Mach code multithreaded yet.

It's not just that, looking at the old code it doesn't seem to deal
very well actually performing the IO, there's also other issues
that Alan Cox and Matt Dillon explained to me where in certain
locations vm_object lists are traversed in forward and reverse
order, this causes a lock order problem, it could be fixed by
possibly sharing a lock across object chains, but for now moving
to a giant lock is still a step in the right direction because
as I've mentioned, with everything still under Giant we've still
got decent performance.

-- 
-Alfred Perlstein - [alfred@freebsd.org]
Instead of asking why a piece of software is using "1970s technology,"
start asking why software is ignoring 30 years of accumulated wisdom.

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20010424091431.J1790>