Re: System unique identifier.....

Matthew Jacob (mjacob@feral.com)
Tue, 22 Jun 1999 11:52:18 -0700 (PWT)

My original posting said why I was doing this- I've a need to generate at
least a 64 bit unique WWN for a NODE id for fibre channel. This is for
fibre channel ports that don't have assigned (in NVRAM) WWNs. Like I said,
Solaris has a hack that takes the first ethernet mac addr and or's in 1 <<
60 (and controller instance number if you're making a port vs. a node WWN)
for this purpose. I could do this also for Linux, but it raised the
metaquestion of "I need a system unique identifier". Yes, it's a
heavyweight object, but that's no reason to not make it available to
kernel (at boot time too) subsystems that need it.

On Tue, 22 Jun 1999 tytso@mit.edu wrote:

> Date: Mon, 21 Jun 1999 11:19:27 -0700 (PDT)
> From: Matthew Jacob <mjacob@feral.com>
>
> This is all well and good, and yes I'll definitely use this. The question
> though is "Is this available to a driver as a linux driver ABI type
> call?". The answer has to be "not yet" I'd guess.
>
> Are you looking for the ability to generate a 128 UUID in kernel space?
> If so, my first reaction would be why.... a UUID is needed when you need
> an ID which is universally unique across all time and all space and all
> machines in the world. What are the properties of an ID which you are
> looking for, and why does the driver need to assign such a thing?
> 128-bit UUID are rather heavyweight objects, and if you don't need all
> of the properties which they guarantee, something ligherweight might be
> more appropriate.
>
>
> - Ted
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.rutgers.edu
> Please read the FAQ at http://www.tux.org/lkml/
>

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/