Re: [PATCH] register_blkdev

[Oliver Xymoron]

On Sat, Mar 08, 2003 at 03:09:37PM +0000, Alan Cox wrote:
> On Sat, 2003-03-08 at 01:13, Linus Torvalds wrote:
> > On Fri, 7 Mar 2003, Andrew Morton wrote:
> > > 
> > > Some time back Linus expressed a preference for a 2^20 major / 2^12 minor split.
> > 
> > Other way around. 12 bits for major, 20 bits for minor.
> > 
> > Minor numbers tend to get used up more quickly, as shown by the current 
> > state of affairs, and also as trivially shown by things like pty-like 
> > virtual devices that pretty much scale arbitrarily with memory and users.
> 
> 20:12 is easier for the current behaviour. 12:20 with the ability to hand out
> sections of space has great potential for lumping things like "disks", 
> "serial ports" and so on together in more logical ways. 12:20 also makes the
> compatibility logic easier since all of the legacy space falls in "major 0"
> which becomes the remangler.
> 
> Is there any reason for not using CIDR like schemes as Al Viro proposed a long
> time back (I think it was Al anyway). That also sorts out the auditing problem

I think it was actually me, arguing with Viro. I was building a device
number registration layer on top of the new resource tree structure at
the time (now 3 years ago!). And in retrospect, I think 32:32
internally with an 8:8 legacy mangler is probably the sanest way to
go.

-- 
 "Love the dolphins," she advised him. "Write by W.A.S.T.E.."