Return to BSD News archive
Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!msunews!agate!howland.reston.ans.net!news.cac.psu.edu!news.pop.psu.edu!hudson.lm.com!netline-fddi.jpl.nasa.gov!nntp.et.byu.edu!news.byu.edu!hamblin.math.byu.edu!park.uvsc.edu!usenet
From: Terry Lambert <terry@cs.weber.edu>
Newsgroups: comp.os.386bsd.development
Subject: Re: DoubleSpace for UFS?
Date: 19 Feb 1995 22:52:35 GMT
Organization: Utah Valley State College, Orem, Utah
Lines: 32
Message-ID: <3i8i3j$pba@park.uvsc.edu>
References: <D47HrE.CK7@murdoch.acc.Virginia.EDU> <3i5rju$drn@park.uvsc.edu> <3i82r2$srh@olympus.bzn.vlt.eds.com>
NNTP-Posting-Host: hecate.artisoft.com
spannrin@olympus.bzn.vlt.eds.com (Craig Spannring) wrote:
] Terry Lambert <terry@cs.weber.edu> wrote:
] >act9m@mamba.cs.Virginia.EDU (Alan Chih-Chang Tai) wrote:
] >> As a class project, my team is considering implemeting an on-the-fly
] >> disk compression for the Unix File System.
]
] [snip]
]
] >> 3) would this involve mucking around with the kernel?
] >
] >Yes. Unless you went the user space route
]
] Wouldn't having it in user space break demand paged loading of binary
] executables? (Not really sure how BSD does virtual memory.)
No, you are confusing using the kernel space NFS client to access
a user space NFS server on the same machine with direct local access
to user space.
Even the second might work if you built a local kernel consumer
level interface. Actually, this was done in association with the
Ficus project (on which the BSD VFS is based) to allow for user
space file system developement, although I don't believe that a
debugged FS was expected to run anywhere by the kernel, since this
would add two sets of boundry crossing per layer access.
Terry Lambert
terry@cs.weber.edu
---
Any opinions in this posting are my own and not those of my present
or previous employers.