[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: devfs support in ofpath: criteria for patches



On Thu, May 24, 2001 at 05:19:34PM -0700, Andrew Sharp wrote:
> That's correct.
> 
> Scsibus0, id0, lun0 (if disk) == /dev/sda
> Scsibus0, id0, lun1 (if disk) == /dev/sd[a+1]
> .
> .
> Scsibus0, id1, lun0 (if disk) == /dev/sdN
> .
> .
> scsibus1, id0, lun0 (if disk) == /dev/sdN+1

thought so.  so all that tom needs to do is assign the nodename
following this logic, however he will also need to test whether that
device is a Direct-Acces or not.  this can be done using the same
wizardry i already use in scsiinfo(). (parsing /proc/scsi/scsi) 

it would make ofpath slower on devfs systems, but thats the price you
pay for using devfs ;-)

> and so on.  If your external drive is on the same bus as your
> internal drive, and you set the id on the external to 0 and the id
> on the internal to 3, then guess what?  Your external will be
> /dev/sda and your internal will be /dev/sdb.  That's the way it
> works.  You are in complete control of how the assignments are
> made.  If you're using a scsi setup where the id's are assigned
> automatically by position on the bus (so called SCA connectors on
> the drive(s)), then you can just move the drives around 'till
> they're in the order you want.
> 
> a

-- 
Ethan Benson
http://www.alaska.net/~erbenson/

Attachment: pgpaQdRxeGm6u.pgp
Description: PGP signature


Reply to: