Date: Fri, 14 Feb 1997 23:29:48 +0100 From: j@uriah.heep.sax.de (J Wunsch) To: dwhite@resnet.uoregon.edu (Doug White) Cc: hackers@freebsd.org Subject: Re: mfs & swap relationship Message-ID: <Mutt.19970214232948.j@uriah.heep.sax.de> In-Reply-To: <Pine.BSI.3.94.970213223902.612I-100000@localhost>; from Doug White on Feb 13, 1997 22:40:55 -0800 References: <Pine.BSI.3.94.970213223902.612I-100000@localhost>
next in thread | previous in thread | raw e-mail | index | archive | help
As Doug White wrote: > What's significant about mounting mfs's on the same partition as > a system swap partition? The mount_mfs(8) man page hints that this > partition is used for any spillover, but that doesn't make sense. Why > isolate it to a specific swap partition? It's not isolated, the entire swap is used to back the MFS. The only reason to prefer the swap partition is that this device node is used to `template' the UFS parameters from. Basically a moot point these days, now that we aren't using most of UFS's optimizations anymore. > And can you make an mfs that's attached to a different partition? Of course. It's only a little hard to mount a MFS on a machine that doesn't have any disk available at all. I've got patches in the queue that allow using /dev/null as a template (in which case mount_mfs invents some parameters, which is about the best it could do at all in case swapping goes over NFS). -- cheers, J"org joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Mutt.19970214232948.j>