Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 26 Jan 1999 09:39:36 -0500 (EST)
From:      Erez Zadok <ezk@cs.columbia.edu>
To:        Eivind Eklund <eivind@FreeBSD.ORG>
Cc:        Jan Pechanec <pechy@hp735.cvut.cz>, Wolfram Schneider <wosch@panke.de.freebsd.org>, lists@gal.netlab.sk, questions@FreeBSD.ORG, freebsd-isp@FreeBSD.ORG, freebsd-fs@FreeBSD.ORG
Subject:   Re: Need non-case sensitive fs 
Message-ID:  <199901261439.JAA04364@shekel.mcl.cs.columbia.edu>
In-Reply-To: Your message of "Tue, 26 Jan 1999 10:55:31 %2B0100." <19990126105531.J382@bitbox.follo.net> 

next in thread | previous in thread | raw e-mail | index | archive | help
In message <19990126105531.J382@bitbox.follo.net>, Eivind Eklund writes:
> On Mon, Jan 25, 1999 at 05:01:25PM +0100, Jan Pechanec wrote:
> > 	To solve the problem, I think the best solution would be to
> > use stackable layering. This is a problem that somebody could already
> > solved (search for heidemann and UCLA), if not and you have somebody
> > that can program in C, let him add some code to the null layer
> > template (see mount_null). One hundred lines of code, I estimate.
> 
> nullfs does not work as of present.   You can possibly get it to work by
> using tegge's patches as a brute-force interrim solution; I'm working on
> getting stacking layers to work properly again.

Re: brute force, I've got a working nullfs and several f/s built on top of
it.  It circumvents the whole VM buffer-cache thing by using synchronous
writes, and using read/write for putpage/getpage.  It may be slower than
what it should be, but in practice it was only a few % slower than an async
implementation on linux (identical h/w of course).

Until these big problems are formally resolved (yes I'm working on them
too), users may wish to start implementing their favorite f/s based on my
*working* implementation.  Later on they can do whatever changes might be
necessary to comply with a fixed stackable interface.  I believe that doing
a case-insensitive f/s using my template nullfs (which I call wrapfs) would
be easy.

You can get freebsd wrapfs software from:

	http://www.cs.columbia.edu/~ezk/research/software/

and a few papers that mention the freebsd port in

	http://www.cs.columbia.edu/~ezk/research/wip.html

> Eivind.

Eivind et al.  I think there are already several people who are trying to do
the same thing --- fix stacking support in freebsd.  I would suggest that we
all get together and do a mini-design that would be approved by the rest of
the freebsd-fs community, and then embark on an implementation.  I'm willing
to help of course.

Erez.

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199901261439.JAA04364>