[dm-devel] Re: [PATCH] 1/1: Device-Mapper: Remove 1024 devices limitation

[Alasdair G Kergon]

On Tue, Jul 06, 2004 at 10:35:52PM +0100, Alasdair G Kergon wrote:
> Yes, but the comments imply that the thing it found might in some
> circumstances not be the thing you asked it to look for 

To clarify, we make two sorts of requests for ids [minor numbers]:
  (a) What is the lowest unused id number?
  (b) Is id number N used or not?

If I interpret things correctly, in case (b) if N is larger than the
the highest id number currently stored it might still return non-NULL 
(if lower order bits match lower order bits of a different stored id).

[Would it be sufficient for the caller to keep track of the highest 
allocated ID and check before calling, or is the masking sometimes
stronger than that?]

Alasdair
-- 
agk at redhat.com