From mboxrd@z Thu Jan  1 00:00:00 1970
From: Pekka Enberg <penberg@kernel.org>
Subject: Re: [PATCH 1/2] tools/kvm/9p: Always include system header before
 our headers
Date: Sat, 18 Jun 2011 10:50:30 +0300
Message-ID: <1308383431.28516.6.camel@jaguar>
References: <1308334282-11865-1-git-send-email-aneesh.kumar@linux.vnet.ibm.com>
	 <1308339906.8711.13.camel@lappy> <1308340059.28516.3.camel@jaguar>
	 <87sjr71utd.fsf@linux.vnet.ibm.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 7bit
Cc: Sasha Levin <levinsasha928@gmail.com>, kvm@vger.kernel.org,
	mingo@elte.hu
To: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
Return-path: <kvm-owner@vger.kernel.org>
Received: from filtteri1.pp.htv.fi ([213.243.153.184]:36075 "EHLO
	filtteri1.pp.htv.fi" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1752299Ab1FRHud (ORCPT <rfc822;kvm@vger.kernel.org>);
	Sat, 18 Jun 2011 03:50:33 -0400
In-Reply-To: <87sjr71utd.fsf@linux.vnet.ibm.com>
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

On Sat, 2011-06-18 at 12:37 +0530, Aneesh Kumar K.V wrote:
> On Fri, 17 Jun 2011 22:47:39 +0300, Pekka Enberg <penberg@kernel.org> wrote:
> > On Fri, 2011-06-17 at 15:45 -0400, Sasha Levin wrote:
> > > On Fri, 2011-06-17 at 23:41 +0530, Aneesh Kumar K.V wrote:
> > > > Otherwise many things can break, such as mapping of stat to stat64
> > > 
> > > We always include our headers before system headers to make sure we
> > > don't have any hidden prerequisites or dependencies within the headers.
> > > 
> > > Why did you need to move system headers before our headers in this case?
> > 
> > The stat64() breakage seems rather odd since we do this in the Makefile:
> > 
> > DEFINES	+= -D_FILE_OFFSET_BITS=64
> > DEFINES	+= -D_GNU_SOURCE
> > 
> > Aneesh, what kind of build problems are you seeing?
> 
> Ok the below change also fix the issue for me. So it could be some
> mapping in the kernel headers
> 
> diff --git a/tools/kvm/virtio/9p.c b/tools/kvm/virtio/9p.c
> index 38a997d..730f4a1 100644
> --- a/tools/kvm/virtio/9p.c
> +++ b/tools/kvm/virtio/9p.c
> @@ -9,16 +9,16 @@
>  #include "kvm/threadpool.h"
>  #include "kvm/irq.h"
>  
> -#include <linux/virtio_ring.h>
> -#include <linux/virtio_9p.h>
> -#include <net/9p/9p.h>
> -
>  #include <fcntl.h>
>  #include <sys/types.h>
>  #include <sys/stat.h>
>  #include <pthread.h>
>  #include <dirent.h>
>  
> +#include <linux/virtio_ring.h>
> +#include <linux/virtio_9p.h>
> +#include <net/9p/9p.h>
> +
>  #define NUM_VIRT_QUEUES		1
>  #define VIRTIO_P9_QUEUE_SIZE	128
>  #define	VIRTIO_P9_TAG		"kvm_9p"

Yes, makes sense. Can you send that as a proper patch and I'll apply it?

The ordering of #includes is as follows:

  1. Our own headers. This helps in ensuring there's no missing
     #includes in the headers themselves.

  2. Library headers and glibc headers

  3. Linux kernel headers.

Code that doesn't follow that is a bug waiting to happen.

				Pekka