Re: [RFC: 2.6 patch] kernel/sys.c: remove unused exports

[Andrew Morton <akpm@osdl.org>]

On Thu, 29 Jun 2006 21:58:28 +0200
Adrian Bunk <bunk@stusta.de> wrote:

> On Thu, Jun 29, 2006 at 12:44:00PM -0700, Andrew Morton wrote:
> > On Thu, 29 Jun 2006 12:36:08 -0700
> > Andrew Morton <akpm@osdl.org> wrote:
> > 
> > > On Thu, 29 Jun 2006 21:19:40 +0200
> > > Adrian Bunk <bunk@stusta.de> wrote:
> > > 
> > > > This patch removes the following unused exports:
> > > > - EXPORT_SYMBOL:
> > > >   - in_egroup_p
> > > > - EXPORT_SYMBOL_GPL's:
> > > >   - kernel_restart
> > > >   - kernel_halt
> > > 
> > > Switch 'em to EXPORT_UNUSED_SYMBOL and I'll stop dropping your patches ;)
> > > 
> > 
> > If doing this, I'd suggest it be done thusly:
> > 
> > EXPORT_UNUSED_SYMBOL(in_egroup_p);	/* June 2006 */
> > 
> > to aid later decision-making.
> 
> I had some bad experiences with following processes you suggest the 
> doesn't remove the symbol immediately:
> 
> As you wanted me to do, I scheduled the EXPORT_SYMBOL(insert_resource) 
> for removal on 2 May 2005 with both __deprecated_for_modules and an 
> entry in feature-removal-schedule.txt with the target date April 2006.
> 
> On 11 Apr 2006, I sent the patch to implement this scheduled removal.
> 
> As of today, the latter patch is still stuck in -mm (which isn't better 
> than having it dropped) although it's long overdue.

Blame Greg ;)

> Do you understand why I distrust your "to aid later decision-making"?

You'll cope.

> Can you state publically "If there's still no in-kernel user after six 
> months, the removal is automatically ACK'ed."?

6 or 12.  We haven't decided.  6 sounds OK.  If nobody complains.  If
they do, we rethink a particular export.