Date: Fri, 25 Oct 2002 20:39:40 +0100 (BST) From: Mark Valentine <mark@thuvia.demon.co.uk> To: John Baldwin <jhb@freebsd.org> Cc: Poul-Henning Kamp <phk@critter.freebsd.dk>, Poul-Henning Kamp <phk@critter.freebsd.dk>, "M. Warner Losh" <imp@bsdimp.com>, cvs-committers@freebsd.org, cvs-all@freebsd.org, Mikhail Teterin <mi+celsius@aldan.algebra.com> Subject: Re: cvs commit: src/lib/libdisk Makefile chunk.c write_alpha_dis Message-ID: <200210251939.g9PJde9E079205@dotar.thuvia.org> In-Reply-To: <XFMail.20021025145539.jhb@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
> From: John Baldwin <jhb@freebsd.org>
> Date: Fri 25 Oct, 2002
> Subject: Re: cvs commit: src/lib/libdisk Makefile chunk.c write_alpha_dis
> So according to you (using this argument of mount or the kernel or somebody
> looking at foo0a and going "oh, you really mean foo0s3a so I'll fix it up
> for you") I should be able to do "mount /dev/ad0 /windows" and it should
> just do the right thing, right? That's exactly the same thing that you are
> asking for. Or better yet, just 'mount -t msdos /windows' and someone
> (mount? kernel?) is supposed to magically walk all the disk devices and
> find some msdos filesystem and mount it.
/dev/ad0 is an insufficient specification for "the Windows partition on
the first IDE disk", so no, that's not what I expect.
The correct specification would be closer to what QNX uses, which is
something like /dev/hd0t6 ("the primary DOS partition on the first disk");
a better scheme would be more intuitive.
The reason for this is that DOS and friends use the partition type as the
primary key into the partition table, not a stored index. These tools
don't seem to think anything of re-arranging the partition table, which
is what makes the "slice" naming scheme fall down.
Currently, /dev/ad0a specifies "the partition labelled 'a' in the BSD
disklabel on the first IDE disk; if the disk has a DOS partition table,
looks for the BSD disklabel in a partition of type 0xa5".
GEOM is currently lacking a way of saying this; Julian pointed out a
GEOM naming scheme to specify this, but it's incompatible with the current
scheme.
> Seriously. What you want is logical volume names. GEOM does not provide
> that. The old kernel code didn't either. What you want is to add an
> extra layer to do the physical <-> logical mapping.
I don't want to add unnecessary layers; simply a way to specify what I
currently can so that my systems boot more reliably.
Cheers,
Mark.
--
Mark Valentine, Thuvia Labs <mark@thuvia.co.uk> <http://www.thuvia.co.uk>
"Tigers will do ANYTHING for a tuna fish sandwich." Mark Valentine uses
"We're kind of stupid that way." *munch* *munch* and endorses FreeBSD
-- <http://www.calvinandhobbes.com> <http://www.freebsd.org>
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200210251939.g9PJde9E079205>
