[patch 2/2] PPC64: Janitorial: Remove unused, deprecated code

Linas Vepstas linas at austin.ibm.com
Fri Aug 19 05:54:49 EST 2005


On Thu, Aug 18, 2005 at 11:19:44AM -0500, Nathan Lynch was heard to remark:
> linas at austin.ibm.com wrote:
> > -/**
> > - * Returns all nodes linked together
> > - */
> > -struct device_node *
> > -find_all_nodes(void)
> > -{
> > -	struct device_node *head, **prevp, *np;
> > -
> > -	prevp = &head;
> > -	for (np = allnodes; np != 0; np = np->allnext) {
> > -		*prevp = np;
> > -		prevp = &np->next;
> > -	}
> > -	*prevp = NULL;
> > -	return head;
> > -}
> > -EXPORT_SYMBOL(find_all_nodes);
> 
> I don't think we're supposed to just rip out exported symbols without
> notice anymore.

Oh, why not?  I grepped, I can't find any users of this code anywhere.
Surely, there aren't any proprietary drivers calling this.

> This is only one of a family of unofficially deprecated functions for
> traversing the OF device tree.  

I was thinking of asking if I should get rid of the other uses of the 
other deprecated functions as well.  So.. Should I?

> Perhaps we should add this group of
> functions to Documentation/feature-removal-schedule.txt to make it
> official, and then we can get rid of them in a few months.

Hmm, just looked at that file.  Surely, the intent of that file 
is to document big things that affect many other people.  By contrast,
I would have thought that this patch affected just about no one.

I found it somewhat bizarre to see dates in that file, and not version
numbers.  Wouldn't one want to have the removal of big chunks coincide
with releases of new major versions? e.g. 2.6.14 or 2.6.15?


--linas



More information about the Linuxppc64-dev mailing list