Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 8 Oct 1997 20:57:56 +0000 (GMT)
From:      Terry Lambert <tlambert@primenet.com>
To:        dg@root.com
Cc:        dag-erli@ifi.uio.no, gordon@drogon.net, mike@smith.net.au, hackers@FreeBSD.ORG
Subject:   Re: Wheres all my memory going?
Message-ID:  <199710082057.NAA11442@usr05.primenet.com>
In-Reply-To: <199710081859.LAA22364@implode.root.com> from "David Greenman" at Oct 8, 97 11:59:22 am

next in thread | previous in thread | raw e-mail | index | archive | help
>    FreeBSD and other 4.4BSD derived OSes don't include the space consumed by
> mmap'ed files (except the one being execed) in the 'SIZE'. The virtual size
> as reported is the sum of text+data+bss+heap+stack. This is a "bug" of sorts,
> but is difficult to fix.

You mean "clean pages which have not triggered copy-on-write are not
counted", right?

I don't think this is a bug.  I think clean pages should not be counted,
if the file is being used as swap store.  That said, I think there are
several cases where it would be *very* desirable to be able to flag
clean pages as "must be resident in swap".  Mostly, these are dataless
configurations, such that the client machine does not hang on a page-in
of a page for a running application should the NFS server go down.  I
admit that this is not very typical usage, but I think that has more
to do with support technology (like read-only mounting of / and /usr)
than relative desirability of the configuration.


					Terry Lambert
					terry@lambert.org
---
Any opinions in this posting are my own and not those of my present
or previous employers.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199710082057.NAA11442>