From owner-freebsd-questions@FreeBSD.ORG Thu Apr 9 16:58:15 2015 Return-Path: Delivered-To: freebsd-questions@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTPS id 1CADD7A5; Thu, 9 Apr 2015 16:58:15 +0000 (UTC) Received: from cosmo.uchicago.edu (cosmo.uchicago.edu [128.135.52.97]) by mx1.freebsd.org (Postfix) with ESMTP id E60F4768; Thu, 9 Apr 2015 16:58:14 +0000 (UTC) Received: by cosmo.uchicago.edu (Postfix, from userid 48) id 4C476CB8C8E; Thu, 9 Apr 2015 11:58:08 -0500 (CDT) Received: from 128.135.70.2 (SquirrelMail authenticated user valeri) by cosmo.uchicago.edu with HTTP; Thu, 9 Apr 2015 11:58:08 -0500 (CDT) Message-ID: <55353.128.135.70.2.1428598688.squirrel@cosmo.uchicago.edu> In-Reply-To: References: <5526408F.1090005@asim.lip6.fr> <20150409114941.692910293a8120de4dafd400@yahoo.es> <20150409114906.6dd4379c@archlinux> <55264DC5.9070606@asim.lip6.fr> <55267A63.4040503@qeng-ho.org> <55269509.1080103@asim.lip6.fr> Date: Thu, 9 Apr 2015 11:58:08 -0500 (CDT) Subject: Re: How FreeBSD manage more than 4GB on 32 bits architecture From: "Valeri Galtsev" To: "Adrian Chadd" Reply-To: galtsev@kicp.uchicago.edu User-Agent: SquirrelMail/1.4.8-5.el5.centos.7 MIME-Version: 1.0 Content-Type: text/plain;charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Priority: 3 (Normal) Importance: Normal Cc: Pierre-Yves =?iso-8859-1?Q?P=C3=A9neau?= , FreeBSD Questions X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.18-1 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 09 Apr 2015 16:58:15 -0000 On Thu, April 9, 2015 11:19 am, Adrian Chadd wrote: > Hi! > > This sounds... very exciting. I'd start by taking a look at how PAE > works - you may be able to reuse all those pieces and just implement > more than the 4 bits of physical address space. The interesting bits > of PAE aren't just the extension of the physical addresses, but all > the management of mapping things into virtual address space as > required. The only problem with all these exercises is: on 32 bit CPU box a single process still can only have 4GB of address space (according to my definitely not deep enough knowledge of kernel)... Just my $0.02. Valeri > > I haven't looked at the MIPS side of things though. IIRC, on MIPS32 we > only implement up to what the direct map KSEG supports, so a little > PAE ish work would be required there to make things work right. > > So, I'd start by looking at how PAE on i386 works. I think there's a > little bit of KVA exhaustion going on in -HEAD with i386 PAE that can > be fixed with a little bit of attention. The rest of the basic PAE > stuff should still work fine. Then if you need it for MIPS, you'll > need to sketch out how you expect PAE to work on MIPS, and then make > up a project for that. :) > > > > -adrian > _______________________________________________ > freebsd-questions@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-questions > To unsubscribe, send any mail to > "freebsd-questions-unsubscribe@freebsd.org" > ++++++++++++++++++++++++++++++++++++++++ Valeri Galtsev Sr System Administrator Department of Astronomy and Astrophysics Kavli Institute for Cosmological Physics University of Chicago Phone: 773-702-4247 ++++++++++++++++++++++++++++++++++++++++