> Jon M. Taylor wrote:
> > On Fri, 7 Aug 1998, Alan Cox wrote:
> > > > P.S. I'm not arguing for devfs per se; simply arguing for a "canonical,
> > > > never-out-of-date /dev" and "devices named by SCSI id".
> > > Devices named by SCSI ID is IMHO a partial solution. I should be able to
> > > take put a jaz disk of /home on any jaz drive on the machine and boot the
> > > system happily.
> >
> > You'll have to be the only person that can pull this trick though.
> > If anyone else wants to do this, how is the kernel supposed to know whose
> > /home to mount? If the kernel doesn't know in advance where the /home is
> > located, it will have to search for it and it might find someone elses'
> > on another Jaz drive somewhere.
> Yeah, the only real problem inherent in UUIDs is that they inherently
> required a discovery phase, which probably involves searching (a) ALL
> your drives and partitions, (b) looking at a device list ("only search
> *these* devices"), or (c) adding another level of indirection by
> requiring a device->UUID map.
> Going from the abstract (UUID) to the concrete (/dev/...) will be the
> bear.
> Jeff
It is not so bad. There is this old patch that gives you /proc/partitions.
Given that and a typical system with a few dozen partitions, a few dozen
disk accesses will discover the right partition.
(Putting things in a file defeats the idea that this should work also
after rebooting with different disks, so (b) with an actual file is bad,
but with /proc/partitions it works fine.)
-
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.altern.org/andrebalsa/doc/lkml-faq.html