[1163] in linux-scsi channel archive
Re: adding scsi devices
daemon@ATHENA.MIT.EDU (Andy Poling)
Sun Dec 22 19:29:46 1996
Date: Sun, 22 Dec 1996 19:21:36 -0500 (EST)
From: Andy Poling <andy@realbig.com>
To: Keith Rohrer <kwrohrer@uiuc.edu>
cc: leisner@sdsp.mc.xerox.com, linux-kernel@vger.rutgers.edu,
linux-scsi@vger.rutgers.edu
In-Reply-To: <199612180708.BAA00598@jadrek.kwr.uiuc.edu>
On Wed, 18 Dec 1996, Keith Rohrer wrote:
> > The whole Linux scheme for naming devices on SCSI (and IDE) busses is
> > awkward - period. If I add a disk as SCSI ID 0 then, guess what? All of
> > the sudden none of my SCSI disks have the same device names as before
> > (despite the fact that none of their SCSI ID's have changed). Wonderful -
> > that's not confusing at all...
> Ah, so you want to be able to mount /dev/sd1050 or some such (we can get
> around separators if we use base 16 for the ID's)? Hmm. How many bits
> long are device (major, minor) numbers now, and how many major numbers
> would that take for scsi? Especially since we'd need sd*, st*, sc*,
> and sg*...
>
> Plus, if you plug in an additional controller and reboot, the new one
> might not be the last one... MCA and maybe PCI cards can be remembered
> by slot, but not ISA cards, plus you have to know how the BIOS will order
> the cards, or make lilo boot a device driver which in turn loads
> the kernel...
Yeah - other folks pointed out that Linux only uses 8 bits currently for the
major and minor numbers.
On the controllers issue though - that's really an issue of the drivers and
their probing methods. That is/can be controlled so that it is
predictable. E.g. my Adaptec is always probed for before my DPT EATA
controller. I can count on that remaining constant. :-)
-Andy
(this sig left intentionally blank - too many hats for one sig)
