From mboxrd@z Thu Jan  1 00:00:00 1970
From: Anthony Liguori <anthony@codemonkey.ws>
Subject: Re: [Qemu-devel] [RFC] Replace posix-aio with custom thread pool
Date: Fri, 12 Dec 2008 08:35:57 -0600
Message-ID: <494276CD.6060904@codemonkey.ws>
References: <1228512061-25398-1-git-send-email-aliguori@us.ibm.com> <493E941D.4000608@redhat.com> <493E965E.5050701@us.ibm.com> <20081210164401.GF18814@random.random> <493FFAB6.2000106@codemonkey.ws> <493FFC8E.9080802@redhat.com> <49400F69.8080707@codemonkey.ws> <20081210190810.GG18814@random.random> <20081212142435.GL6809@random.random>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: Gerd Hoffmann <kraxel@redhat.com>, qemu-devel@nongnu.org,
	kvm-devel <kvm@vger.kernel.org>
To: Andrea Arcangeli <aarcange@redhat.com>
Return-path: <kvm-owner@vger.kernel.org>
Received: from mail-gx0-f12.google.com ([209.85.217.12]:45931 "EHLO
	mail-gx0-f12.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1756850AbYLLOgD (ORCPT <rfc822;kvm@vger.kernel.org>);
	Fri, 12 Dec 2008 09:36:03 -0500
Received: by gxk5 with SMTP id 5so919695gxk.13
        for <kvm@vger.kernel.org>; Fri, 12 Dec 2008 06:36:01 -0800 (PST)
In-Reply-To: <20081212142435.GL6809@random.random>
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

Andrea Arcangeli wrote:
> On Wed, Dec 10, 2008 at 08:08:10PM +0100, Andrea Arcangeli wrote:
>   
>> On Wed, Dec 10, 2008 at 12:50:17PM -0600, Anthony Liguori wrote:
>>     
>>> But opening twice means that you lose coherency with NFS.
>>>       
>> Not sure why. They're not running from different nfs clients. If this
>>     
>
> I just got confirmation from Trond that from nfs point of view,
> opening the file multiple times or duping it, is the same as I
> expected (all it matters is that the inode is the same so the
> pagecache radix tree is the same etc..). So opening the file each time
> a thread starts,

I've been thinking about this, the problems I see are:

1) It's impossible to accept a file descriptor for a block device 
(possibly not a problem)

2) You'd have to open all the file descriptors at once.  Otherwise, you 
get really strange behavior if the file gets deleted while the guest is 
running (for instance, with -snapshot).

Regards,

Anthony Liguori

>  would provide a local f_pos and it would work fine on
> older/current kernels too. In theory other nfs clients should also
> behave. Not sure what's best, if to hack around the bdrv api and open
> the file multiple times or wait for preadv/pwritev.
>