> >(One potential solution, of course, is to abandon device numbers entirely
> >and instead pass the entryname down into the open routine; a lot of work,
> >but it eliminates the table-version skew problem.  Of course, then you *can't*
> >call your root device /dev/bert (perhaps you can if it's a symlink, since the
> >extracted "real" name of the device would be available).  This also gets
> >rid of the problem of being unable to increase the number of partitions on
> >a disk drive owing to the minor device bit partitioning scheme.
> 
> What does this mean if you're booting NFS diskless clients?  How's NFS
> going to be taught about the new device node "number" format?

At least as far as I understood this proposal, the idea was to create a
separate filesystem for this that would be mounted on /dev.  The filesystem
would then get the device name to resolve to a driver and would pass the
name to the driver open routine, so the driver could in turn resolve things
like driver number or partition.

It's totally independent of whether the root filesystem is mounted via NFS,
FFS or anything else.
--
ws@TooLs.DE     (Wolfgang Solfrank, TooLs GmbH) 	+49-228-985800