From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43)
	id 1M1el2-0006qC-KR
	for qemu-devel@nongnu.org; Wed, 06 May 2009 06:53:48 -0400
Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43)
	id 1M1ekx-0006ns-1m
	for qemu-devel@nongnu.org; Wed, 06 May 2009 06:53:47 -0400
Received: from [199.232.76.173] (port=50433 helo=monty-python.gnu.org)
	by lists.gnu.org with esmtp (Exim 4.43) id 1M1ekw-0006nZ-Bs
	for qemu-devel@nongnu.org; Wed, 06 May 2009 06:53:42 -0400
Received: from mail2.shareable.org ([80.68.89.115]:43556)
	by monty-python.gnu.org with esmtps (TLS-1.0:RSA_AES_256_CBC_SHA1:32)
	(Exim 4.60) (envelope-from <jamie@shareable.org>) id 1M1ekv-0000cd-Th
	for qemu-devel@nongnu.org; Wed, 06 May 2009 06:53:42 -0400
Date: Wed, 6 May 2009 11:53:37 +0100
From: Jamie Lokier <jamie@shareable.org>
Subject: Re: [Qemu-devel] [PATCH] linux-user: implement pipe2 syscall
Message-ID: <20090506105337.GB23364@shareable.org>
References: <20090505133048.GA29646@kos.to>
	<20090505225809.GJ7574@shareable.org>
	<20090506080023.GA7230@kos.to> <4A0155F6.9000801@opensuse.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <4A0155F6.9000801@opensuse.org>
List-Id: qemu-devel.nongnu.org
List-Unsubscribe: <http://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.gnu.org/pipermail/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <http://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Martin Mohring <martin.mohring@opensuse.org>
Cc: Riku Voipio <riku.voipio@iki.fi>, qemu-devel@nongnu.org

Martin Mohring wrote:
> >> The point of pipe2() with FD_CLOEXEC is to be atomic: make sure
> >> another thread can never see the file descriptor with FD_CLOEXEC not set.
> >   
> How do we want to handle if "O_CLOEXEC" is not defined on the host OS,
> because kernel too old (Debian Etch, CentOS 5)? Should it then return
> ENOSYS?

You can emulate it on any host OS using a mutex or read-write lock:

    - Acquire the lock for reading before calling the hosts's
      open/pipe/socket/etc.
    - Release the lock after calling the host's fcntl to set FD_CLOEXEC.

    - Acquire the lock for writing around emulation of fork, exec,
      clone and unshare.

There are ways to do it more efficiently without a lock, but I don't
think they are worth bothering with here.

-- Jamie