Date: Tue, 6 Jan 1998 01:44:50 -0800 (PST) From: Julian Elischer <julian@whistle.com> To: David Greenman <dg@root.com> Cc: "John S. Dyson" <toor@dyson.iquest.net>, cvs-committers@FreeBSD.ORG, cvs-all@FreeBSD.ORG, cvs-sys@FreeBSD.ORG Subject: Re: John's latest VM commit. Message-ID: <Pine.BSF.3.95.980106013941.10245A-100000@current1.whistle.com> In-Reply-To: <199801060905.BAA27275@implode.root.com>
next in thread | previous in thread | raw e-mail | index | archive | help
> > After thinking about what I said for a bit, I think I might be exagerating > a little - on some systems where the vnode/object cache is sufficiently > large, you could easily have more of those than default/swap objects. It > depends on the processes - how many, memory mappings, etc., and how many > cached files are in the system. It would be more accurate to say simply that > there are a sizeable number of non-vnode objects... > your point is well taken. However as you said before. It could be argued that even anon objects should be represented by an appropriate vnode (maybe lazily evaluated) construction. surely an 'fs type' that handled swap would produce a more uniform interface, as that would mean that all paging operations always went through the same interface... Certainly though this is not something that will give us a great win for little work. It's more of an accademic exercise. julian
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.95.980106013941.10245A-100000>