[lkml]   [1998]   [Jan]   [12]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: devfs
On Sun, Jan 11, 1998 at 05:43:54PM -0600, Mitch Adair wrote:
> To my thinking, any combination of letter and numbers is equally
> valid... once you've figured out what they all stand for ;)
> I was just voting for the one that is closest to what I am
> used to working with - the Solaris naming scheme. Since the
> positions are the same its not that bad - just doesn't roll of the
> tongue quite like c0b0t0d0s0 :)

Here's an idea. If CONFIG_SCSI_MULTI_LUN is not defined then leave out the
'dX' pair. And it would even be nice if the 'bX' pair could be left out
where it is not supported. In most systems, c0t0s0 is /dev/sda and is

Another thing is it would be nice if we could have a nicer way of
identifying the controller. It's a toss up between not caring about
controllers (most people only have 1 anyway) or situations where people
have two . I used to have a T130B as well as an NCR53c710 'cos I was
supporting the T130B code. Then I removed the T130B. Chances are my device
names would change.

With PCI we could use PCI based info to identify them. Or maybe by name. If
you have two of the same controller then the c0 c1 syntax makes sense. So
ncr53cxx_0t0s0 is my /dev/sda and ncr53c400_0t4s0 could be /dev/scd0.
Whereas the proposed scheme would allocate the controller number based
probably on the order they are listed in the kernel. Or, in a modularised
kernel, in the order the modules are loaded. And that depends on the order
things get mounted in a kerneld situation. That's all quite ugly (although
it doesn't affect me in the slightest now since I only have one scsi

[ Kevin Lentin Email: ]
[ finger for PGP public key block. ]
[ KeyId: 06808EED FingerPrint: 6024308DE1F84314 811B511DBA6FD596 ]

 \ /
  Last update: 2005-03-22 13:41    [W:0.061 / U:5.996 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site