Subject: Re: rpc xid randomness
To: David Laight <david@l8s.co.uk>
From: Steven M. Bellovin <smb@research.att.com>
List: tech-userlevel
Date: 09/07/2003 16:53:30
In message <20030907204608.D1404@snowdrop.l8s.co.uk>, David Laight writes:
>> why are we using (poorly-designed) pseudorandom number instead of
>> sequential number right now?
>
>I suspect because seqential xids get confused when a system reboots
>and immediately starts again the the same point.
>
>date ones are a problem if more that one has to be allocated in the
>same tick.
The obvious solution there is to use a counter that's initialized from
the clock at reboot time. We can't reboot in less than a tick...
--Steve Bellovin, http://www.research.att.com/~smb