Subject: Re: Another changer, another changer problem
To: Christoph Badura <>
From: Gandhi woulda smacked you <>
List: current-users
Date: 10/02/1998 01:07:18
On Fri, 2 Oct 1998, Christoph Badura wrote:

 * (Greg A. Woods) writes:
 * >The former two
 * >are probably not worthy of use, but the latter still is though I'll
 * >probably keep it in my AT&T 3B2 where it has full OS support, including
 * >explict device naming such as "c0t0d0s0" which means "controller 0,
 * >target 0, disk(LUN) 0, slice 0" (something I wish NetBSD had).
 * Excuse me?
 * 	sd0 at scsibus0 target 0 lun 0
 * in the kernel config file works just fine.  Noone is preventing you from
 * changing to /dev and doing "ln sd0a c0t0d0s0" either.

Talk about masochistic.  Hate those long names living in /dev/dsk and
/dev/rdsk :-P.

Although, it would be possible to write a parser to look at dmesg output
at boot time and rebuild /dev after the root filesystem check finished,
(although the nodes would be needed beforehand, ick), but this is a
whole 'nother pointer to functions returning pointers to cans of worms
which just got close()d about a year ago; I guess one of the worms
is still slithering about and procreating (get the Triox and Benzene,

Anyway, if we want Slowlaris, we know where we can find it and I'd
really hate to have to type the mouthfuls required by /dev/dsk/c0t0d0s0,
however obvious the references.  I know where my disks live (and have
undone all sd0 <--> sd3 mappings including in the EEPROM).

Friends don't let friends use Microsoft.