Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 22 Apr 95 22:09:13 MDT
From:      terry@cs.weber.edu (Terry Lambert)
To:        phk@ref.tfs.com (Poul-Henning Kamp)
Cc:        rashid@haven.ios.com, smace@metal-mail.neosoft.com, hackers@FreeBSD.org
Subject:   Re: 2.x and large memory configs
Message-ID:  <9504230409.AA10351@cs.weber.edu>
In-Reply-To: <199504230342.UAA25498@ref.tfs.com> from "Poul-Henning Kamp" at Apr 22, 95 08:42:08 pm

next in thread | previous in thread | raw e-mail | index | archive | help
> > This issue is likely to surface again if FreeBSD decides to support
> > PCMCIA memory cards being ejected.  8-).
> 
> I doubt we would even think about that.  Considering that you get no warning
> before it's ejected, and that they oftern suffer a big accesstime penalty,
> it is not a good idea to use PCMCIA for main-memory.
> 
> We could add it as a "vm-cache", for inactive pages or something, that way
> nothing would turn purple because those pages vanished...

I was thinking of RAM disk myself, and sysctl'ing it back down before
you eject it manually.

Kinda brings up the problem of page classification for inactive vs. active
and where RAM disks fit.

This is on the order of "do I need bounce buffers" and "Is this a loadable
driver that's going to stay around or not do I can avoid fragging memory".

An intersting problem in what might be termed "page migration".  Of course
with migration, you'd migrate the user's pages in their buffers down
instead of bouncing them.

An interesting idea to consider, anyway.


					Terry Lambert
					terry@cs.weber.edu
---
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?9504230409.AA10351>