Re: [RFC] aic94xx: attaching to the sas transport class

[Douglas Gilbert <dougg@torque.net>]

Stefan Richter wrote:
> James Bottomley wrote:
> 
>>Assuming the cages are all powered up, which is beyond
>>the driver control, then serialising the scan will allow one to
>>specify /dev/sda1 deterministically for root.
> 
> 
> Does SAS provide a persistent globally unique property of a device, or
> has SAS to rely on bus topology to uniqely identify devices?

That swings on what you mean by "device" but broadly:
Yes to the first part, but ... SAS is a transport and
SAS transport endpoints (i.e. "ports") have naa-5
world wide unique identifiers. That might sound great
but most SAS disks are dual ported, so that gives two
SAS (i.e. naa-5) addresses (typically consecutive).
A dual ported SAS disk could/should have up to
4 naa-5 addresses:
  - one for SAS port 1 (primary)
  - one for SAS port 2 (secondary)
  - one for the target device
  - and, one for the logical unit

A SAS (target) device might also be a bridge, for
example, as found in SAS expanders that support SATA
devices. SATA disks do not yet have mandated world
wide unique addresses (but provision has been made for
a naa-5 address). The SAS "port" address of a SATA
disk is actually the bridge it is plugged into (but
one could switch SATA disks).

I suspect what you would really like to identify uniquely
is the logical unit. Fuzziness remains due to the
possibility of bridges.

Doug Gilbert