Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 16 Mar 1995 17:03:29 +1000
From:      Bruce Evans <bde@zeta.org.au>
To:        phk@ref.tfs.com, terry@cs.weber.edu
Cc:        current@FreeBSD.org, rgrimes@gndrsh.aac.dev.com
Subject:   Re: newfs: sectors per cylinder (4096) disagrees with disk label (36)
Message-ID:  <199503160703.RAA19633@godzilla.zeta.org.au>

next in thread | raw e-mail | index | archive | help

>1)  The device-driver may use a geometry(A) to access the disk.

>2)  The UFS may be created with a geometry(B) != geometry(A)

>3)  The device driver may need to read the geometry(A) from the disklabel,
>    because the BIOS lies and the drive is too dumb to answer the question.
>    The disklabel is the only place this can come from.

Only on old devices that are currently unsupported because they also
require bad144 support.

Anyway, for sliced devices, geometry(A) is stored in the label for the
whole disk and is not necessarily related to the geometries(B) stored
in the BSD labeled slices on the disk.  You can put a 4096 * 1 geometry
in the slices is that is good for file systems.

>I don't see any point in having the device-drivers even know about geometry,
>considering that LFS for instance doesn't care a hoot, and I will strongly
>work to get the ties already in place broken.

The wd driver has to know for some old devices (those that don't allow
their translation to be programmed, a subset of the currently unsupported
devices).

Bruce



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199503160703.RAA19633>