Re: devfs persistence

Matthew Jacob (mjacob@feral.com)
Thu, 27 Apr 2000 21:41:53 -0700 (PDT)

> Right.
>
> > > We use labels to solve this.
> >
> > That's fine insofar as it goes, but if you have a naming schema that
> > has a bus address as part of it, if the address can change
> > dynamically, that can be an issue. I think that labels are a good
> > idea- you end up not caring what address something is at. And UUIDs
> > or other device unique information is good solid naming stuff.
>
> Why are you concerned about the naming scheme that depends on bus
> addresses? I was talking about specifying labels. How do bus addresses
> come into the picture?

The naming scheme on your website doc encodes bus addresses, as in:

------------

To uniquely identify any SCSI device requires the following information:

controller (host adapter)
bus (SCSI channel)
target (SCSI ID)
unit (Logical Unit Number)

All SCSI devices are placed under /dev/scsi (assuming devfs is mounted on
/dev). Hence, a SCSI device with the following parameters: c=1,b=2,t=3,u=4
would appear as:

/dev/scsi/host1/bus2/target3/lun4 device directory

-----------

This doesn't work so well if target can change, or the hierarchical lun space
is in fact the SSC class hierarchical lun space which is four 16 bit values
in a hierarchy.

I think what Rodger might be mentioning might be something more like this (as
I vastly distort Rodger's words!):

/dev/wwn/NNNNNNNNNNNNN

for the World Wide Name namespace, and

/dev/uuid/UUUUUUUUUUUUUU

for the UUID namespace, and

/dev/vpd/VVVVVVVVVVVVVVVV

for the VPD (drive serial number) namespace. While this isn't a device
physical address locator, it is persistent and easy to find and set up.

-matt

-
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/