From mboxrd@z Thu Jan  1 00:00:00 1970
From: "J. Bruce Fields" <bfields@fieldses.org>
Date: Sun, 17 Feb 2008 01:27:06 +0000
Subject: Re: git-nfsd on ia64
Message-Id: <20080217012706.GB22252@fieldses.org>
List-Id: <linux-ia64.vger.kernel.org>
References: <20080215223943.d12b8eb5.akpm@linux-foundation.org>
In-Reply-To: <20080215223943.d12b8eb5.akpm@linux-foundation.org>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: linux-ia64@vger.kernel.org

On Fri, Feb 15, 2008 at 10:43:19PM -0800, Andrew Morton wrote:
> On Fri, 15 Feb 2008 22:39:43 -0800 Andrew Morton <akpm@linux-foundation.org> wrote:
> 
> > Cunningly, current git-nfsd causes this, with ia64 allmodconfig:
> > 
> > .config:3397:warning: trying to assign nonexistent symbol NFSD_TCP
> > In file included from arch/ia64/ia32/sys_ia32.c:59:
> > arch/ia64/ia32/ia32priv.h:297:1: warning: "elf_check_arch" redefined
> > In file included from include/linux/elf.h:7,
> >                  from include/linux/module.h:14,
> >                  from include/linux/textsearch.h:9,
> >                  from include/linux/skbuff.h:26,
> >                  from include/linux/icmpv6.h:79,
> >                  from include/linux/ipv6.h:197,
> >                  from include/net/ipv6.h:18,
> >                  from include/linux/sunrpc/svcauth.h:18,
> >                  from include/linux/sunrpc/svc.h:18,
> >                  from arch/ia64/ia32/sys_ia32.c:37:
> > include/asm/elf.h:19:1: warning: this is the location of the previous definition
> 
> Easy fix:
> 
> --- a/include/linux/sunrpc/svcauth.h~git-nfsd-fix
> +++ a/include/linux/sunrpc/svcauth.h
> @@ -15,7 +15,6 @@
>  #include <linux/sunrpc/msg_prot.h>
>  #include <linux/sunrpc/cache.h>
>  #include <linux/hash.h>
> -#include <net/ipv6.h>
>  
>  #define SVC_CRED_NGROUPS	32
>  struct svc_cred {
> @@ -25,6 +24,7 @@ struct svc_cred {
>  };
>  
>  struct svc_rqst;		/* forward decl */
> +struct in6_addr;
>  
>  /* Authentication is done in the context of a domain.
>   *
> _
> 
> 
> This should have been done in the first place, IMO.  Avoid header file
> dependencies, faster compile, avoid crap like the above.

OK.  So as a general rule, we should skip the include any time when a
few structure declarations would suffice?

Hm.  It also seems wrong that sys_ia32.c would need to include so many
nfsd/sunrpc headers.  It looks like that should be easy to fix with a
better organization of the nfsd headers--I'll do that, assuming it's
considered worthwhile....

--b.