Subject: Re: kernel ip_randomid() and libc randomid(3) still "broken"
To: Simon Burge <>
From: Thor Lancelot Simon <>
List: tech-net
Date: 11/14/2003 18:14:44
On Sat, Nov 15, 2003 at 10:11:55AM +1100, Simon Burge wrote:
> A couple of months ago there was a discussion about ip_randomid() and
> the related randomid(3) libc function being able to return consecutive
> IDs every now and then.  Should we disable these functions until they
> are fixed to not show this behaviour?  I've added a regression test
> for this too.

Yes.  Didn't we also pretty much reach consensus that most of the uses
of these functions that Itojun added to our tree should be disabled by
default?  I must admit that I am mystified as to why they are still
there.  The code is broken; its security benefit is questionable; in
at least one case, a far better -- because it doesn't repeat -- and
cheaper -- because it avoids the math -- approach is well-known and in
use in Solaris.  Why are we stuck with the status quo as of months ago?

 Thor Lancelot Simon	                            
   But as he knew no bad language, he had called him all the names of common
 objects that he could think of, and had screamed: "You lamp!  You towel!  You
 plate!" and so on.              --Sigmund Freud