From ffjeld at common-lisp.net  Wed Jan 16 16:41:59 2008
From: ffjeld at common-lisp.net (Frode Vatvedt Fjeld)
Date: 16 Jan 2008 17:41:59 +0100
Subject: [movitz-devel] Re: A question about the tmp directory,
	and ext2 support
References: <Pine.LNX.4.64.0801161734100.5917@Psilocybe.Update.UU.SE>
Message-ID: <2hy7apiuy0.fsf@vserver.cs.uit.no>

Andreas Davour <ante at Update.UU.SE> writes:

> I found the tmp directory and the files there looks interesting, since
> if I haven't missunderstood the image I've managed to dump, they lack
> support for harddrives and file systems. How do you build muerte with
> ext2 support, really?
> 
> Any hints?

There is in principle no ext2 support. The tmp directory holds various
bits and pieces of (contributed) code that's not really integrated
with the rest of the system (yet). See the ext2 file as a starting
point for you to write proper ext2 support for movitz :-)

-- 
Frode Vatvedt Fjeld



From ffjeld at common-lisp.net  Wed Jan 16 16:49:43 2008
From: ffjeld at common-lisp.net (Frode Vatvedt Fjeld)
Date: 16 Jan 2008 17:49:43 +0100
Subject: [movitz-devel] Re: A question about the tmp directory,
	and ext2	support
References: <Pine.LNX.4.64.0801161734100.5917@Psilocybe.Update.UU.SE>
	<2hy7apiuy0.fsf@vserver.cs.uit.no>
	<Pine.LNX.4.64.0801161745310.5917@Psilocybe.Update.UU.SE>
Message-ID: <2htzldiul4.fsf@vserver.cs.uit.no>

Andreas Davour <ante at Update.UU.SE> writes:

> Then I still have the problem I don't know how to integrate anything
> with the rest of the system...

Currently there's only a primitive provide/require mechanism. You
provide an initial lisp-file to create-image (los0.lisp is the
default), which pulls in other stuff by require statements,
recursively. This should probably be changed/improved to use something
in the defsystem style.

-- 
Frode Vatvedt Fjeld



From ffjeld at common-lisp.net  Wed Jan 16 16:58:59 2008
From: ffjeld at common-lisp.net (Frode Vatvedt Fjeld)
Date: 16 Jan 2008 17:58:59 +0100
Subject: [movitz-devel] Re: A question about the tmp directory,
	and ext2	support
References: <Pine.LNX.4.64.0801161734100.5917@Psilocybe.Update.UU.SE>
	<2hy7apiuy0.fsf@vserver.cs.uit.no>
	<Pine.LNX.4.64.0801161745310.5917@Psilocybe.Update.UU.SE>
	<2htzldiul4.fsf@vserver.cs.uit.no>
Message-ID: <2hprw1iu5o.fsf@vserver.cs.uit.no>

Frode Vatvedt Fjeld <ffjeld at common-lisp.net> writes:

> Currently there's only a primitive provide/require mechanism. [..]

And I should probably mention, there's not much in the way of an
"architecture" beyond the basic run-time system. So basically, you'd
write plain functions that can e.g. be called from the REPL.

It's a design goal to keep a layer of "architecture-less" functions in
movitz, allowing architectures/applications/systems/whatever maximum
flexibility atop this layer. The los0.lisp file (which really just
does boot-up and sets up the rather simple interactive REPL) is the
one example of such an architectures/applications/systems/whatever
(I'd call this an "application" I guess, since it's not really
intended to have more stuff build over it).

-- 
Frode Vatvedt Fjeld



From mbealby at gmail.com  Wed Jan 16 18:26:59 2008
From: mbealby at gmail.com (Martin Bealby)
Date: Wed, 16 Jan 2008 18:26:59 +0000
Subject: [movitz-devel] Re: A question about the tmp directory,
	and ext2 support
In-Reply-To: <Pine.LNX.4.64.0801161745310.5917@Psilocybe.Update.UU.SE>
	(Andreas Davour's message of "Wed\,
	16 Jan 2008 17\:47\:11 +0100 \(CET\)")
References: <Pine.LNX.4.64.0801161734100.5917@Psilocybe.Update.UU.SE>
	<2hy7apiuy0.fsf@vserver.cs.uit.no>
	<Pine.LNX.4.64.0801161745310.5917@Psilocybe.Update.UU.SE>
Message-ID: <8763xtzkwc.fsf@muscle.bealbywm.plus.com>

Andreas Davour <ante at Update.UU.SE> writes:

>> There is in principle no ext2 support. The tmp directory holds various
>> bits and pieces of (contributed) code that's not really integrated
>> with the rest of the system (yet). See the ext2 file as a starting
>> point for you to write proper ext2 support for movitz :-)
>
> Great, it's an opportunity! :)
>
> Then I still have the problem I don't know how to integrate anything with the
> rest of the system...
>
> I'll take a closer look at the source and see if I can get a handle of things
> to post more specific questions.
>
> /Andreas
>

Hi Andreas, I'm glad you are willing to contribute to Movitz.

Both Shawn and myself track the unofficial git repository[1] of movitz
as everyone can then have commit access for exchanging ideas.

In the git repository you will find some hard drive sector reading /
writing code[2] in there that was created by myself back in October.  I
believe there was an issue with the code in tmp (unknown author?) so I
decided to write a (simpler) replacement.  I performed some preliminary
testing (in qemu) and it appeared to work fine.

This could be a useful platform for building up a filesystem layer upon.

I hope this helps,
Martin

[1] http://repo.or.cz/w/movitz-core.git
[2] http://repo.or.cz/w/movitz-core.git?a=blob;f=losp/x86-pc/ata.lisp


From i_am_weasel at kittymail.com  Fri Jan 18 08:16:44 2008
From: i_am_weasel at kittymail.com (eric weasil)
Date: Fri, 18 Jan 2008 16:16:44 +0800
Subject: [movitz-devel] Re: A question about the tmp directory,
	and ext2 support
Message-ID: <20080118081645.0DBAD7B94D@ws5-10.us4.outblaze.com>



Martin Bealby
> 
> Andreas Davour <ante at Update.UU.SE> writes:
> 
> >> There is in principle no ext2 support. The tmp directory holds various
> >> bits and pieces of (contributed) code that's not really integrated
> >> with the rest of the system (yet). See the ext2 file as a starting
> >> point for you to write proper ext2 support for movitz :-)
> >
> > Great, it's an opportunity! :)
> >
> > Then I still have the problem I don't know how to integrate anything with the
> > rest of the system...
> >
> > I'll take a closer look at the source and see if I can get a handle of things
> > to post more specific questions.
> >
> > /Andreas
> >
> 
> Hi Andreas, I'm glad you are willing to contribute to Movitz.
> 
> Both Shawn and myself track the unofficial git repository[1] of movitz
> as everyone can then have commit access for exchanging ideas.
> 
> In the git repository you will find some hard drive sector reading /
> writing code[2] in there that was created by myself back in October.  I
> believe there was an issue with the code in tmp (unknown author?) so I
> decided to write a (simpler) replacement.  I performed some preliminary
> testing (in qemu) and it appeared to work fine.
> 
> This could be a useful platform for building up a filesystem layer upon.
> 
> I hope this helps,
> Martin

Hi everybody,

I did also write a couple of lines IDE-harddisk interaction code (which is not the hard part in harddisk interfacing). The hard part (IMHO) is to figure out which disks are present and which parameters apply to them. I ended up whith a simple read-sectors and write-sectors commmand (tested) and a init functions wich detect the presence of your disks. I wrote also a partition-table-reader and began to read some ext2fs specs.
The disk interface seems to be currently the most wanted improvement to movitz. We have now 3 attempts. I guess its time for ext2 :).

Erik Soehnel


-- 
_______________________________________________
Get your free email from http://mail.kittymail.com