Date: Mon, 15 Oct 2001 16:44:31 -0500 From: Alfred Perlstein <bright@mu.org> To: Garrett Wollman <wollman@khavrinen.lcs.mit.edu> Cc: cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/sys/vm vnode_pager.c Message-ID: <20011015164431.G59854@elvis.mu.org> In-Reply-To: <200110152140.f9FLeOl32078@khavrinen.lcs.mit.edu>; from wollman@khavrinen.lcs.mit.edu on Mon, Oct 15, 2001 at 05:40:24PM -0400 References: <xzpelo4r90f.fsf@flood.ping.uio.no> <200110152105.aa38680@salmon.maths.tcd.ie> <20011015153606.F59854@elvis.mu.org> <200110152140.f9FLeOl32078@khavrinen.lcs.mit.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
* Garrett Wollman <wollman@khavrinen.lcs.mit.edu> [011015 16:40] wrote: > <<On Mon, 15 Oct 2001 15:36:06 -0500, Alfred Perlstein <bright@mu.org> said: > > > No, the zone allocator sucks, one memory is pinned for a zone it > > can not be reused. > > Actually, that is a good and intentional feature. (See all the papers > on TSM.) That doesn't necessarily mean that it's useful or > appropriate for *nodes, though. It's barely acceptable when it's needed, it's really bad because any pressure on the memory system even if we were to implement hooks to ask subssystems to free back memory would not be applicable to anything using zalloc because of this. I currently don't have a better idea, but I really don't like the current TSM stuff for this reason, BSD/os has managed to go for a long time without needing to implement such things (at least for sockets) and I hope that we can get away from it or at least contain its spread to other subsystems. -- -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 cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20011015164431.G59854>