Re: Style question: Should one check for NULL pointers?

From: Alan Stern <stern@rowland.harvard.edu>
To: Horst von Brand <vonbrand@inf.utfsm.cl>
Cc: Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Subject: Re: Style question: Should one check for NULL pointers?
Date: Sun, 13 Jul 2003 17:42:03 -0400 (EDT)	[thread overview]
Message-ID: <Pine.LNX.4.44L0.0307131738230.8445-100000@netrider.rowland.org> (raw)
In-Reply-To: <200307121840.h6CIeKIj004212@eeyore.valparaiso.cl>

On Sat, 12 Jul 2003, Horst von Brand wrote:

> Alan Stern <stern@rowland.harvard.edu> said:
> 
> [...]
> 
> > But if you look very far through the kernel sources you will see many 
> > occurrences of code similar to this:
> > 
> > 	static void release(struct xxx *ptr)
> > 	{
> > 		if (!ptr)
> > 			return;
> > 	...
> > 
> > I can't see any reason for keeping something like that.
> 
> Just like free(3)

NO!  Not just like free().  The documentation for free() explicitly states 
that NULL pointers are valid input and result in no action.  A 
release()-type function, by contrast, is called back from core system code 
that guarantees it will always pass a pointer to the currently-registered 
owner of the corresponding resource.  If the owner were NULL, the 
release() function wouldn't have been called in the first place.

Alan Stern