Re: possible SCSI device numbering solution

Gerd Knorr (kraxel@cs.tu-berlin.de)
Tue, 25 Jun 1996 23:30:54 +0200 (MET DST)


On Tue, 25 Jun 1996, Andrew E. Mileski wrote:

> > There is little point going through the disruption that changing it would
> > be, to change it again, so we may as well go to a 64 bit dev_t. One
> > thing tho, would we even need more than 65536 major devices? ie 16 bit
> > major, 48 bit minor (or 281474976710656 minor numbers (2.8*10^14) ;)
>
> I'm reminded of somebody saying, "they'll never need more than 640k". :-/
>
This really is'nt a problem. If you look back, the capacity of hard
disks increased, not the number of hard disks in a system. It's much more
likely that we run in trouble becauce 32 bits are to small to count the
all the sectors of a disk...

Gerd