Re: [RFC] 2.5.8 sort kernel tables

[David Weinehall]

On Thu, Apr 18, 2002 at 08:16:00PM +0200, Kai Henningsen wrote:
> [e-mail] (William Lee Irwin III)  wrote on 18.04.02 in <[e-mail]>:
> 
> > On Thu, Apr 18, 2002 at 07:46:26PM +1000, Keith Owens wrote:
> > > The use of __init and __exit sections breaks the assumption that tables
> > > such as __ex_table are sorted, it has already broken the dbe table in
> > > mips on 2.5.  This patch against 2.5.8 adds a generic sort routine and
> > > sorts the i386 exception table.
> > > This sorting needs to be extended to several other tables, to all
> > > architectures, to modutils (insmod loads some of these tables for
> > > modules) and back ported to 2.4.  Before I spend the rest of the time,
> > > any objections?
> >
> > It doesn't have to be an O(n lg(n)) method but could you use something
> > besides bubblesort? Insertion sort, selection sort, etc. are just as
> > easy and they don't have the horrific stigma of being "the worst sorting
> > algorithm ever" etc.
> 
> Surely the worst (working) sort is randomsort? (Check if sorted. If not,  
> pick two entries at random, exchange, retry.)

I've always been a fan of bit-decay sort.

while (!sorted(data))
	;

/David
  _                                                                 _
 // David Weinehall <[e-mail]> /> Northern lights wander      \\
//  Maintainer of the v2.0 kernel   //  Dance across the winter sky //
\>  http://www.acc.umu.se/~tao/    </   Full colour fire           </