From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from [140.186.70.92] (port=40402 helo=eggs.gnu.org)
	by lists.gnu.org with esmtp (Exim 4.43) id 1P5fr4-00074i-Ts
	for qemu-devel@nongnu.org; Tue, 12 Oct 2010 10:29:27 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <anthony@codemonkey.ws>) id 1P5fr3-0003Na-SI
	for qemu-devel@nongnu.org; Tue, 12 Oct 2010 10:29:26 -0400
Received: from mail-qy0-f173.google.com ([209.85.216.173]:42795)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <anthony@codemonkey.ws>) id 1P5fr3-0003NV-PP
	for qemu-devel@nongnu.org; Tue, 12 Oct 2010 10:29:25 -0400
Received: by qyk36 with SMTP id 36so715697qyk.4
	for <qemu-devel@nongnu.org>; Tue, 12 Oct 2010 07:29:25 -0700 (PDT)
Message-ID: <4CB46F85.8060007@codemonkey.ws>
Date: Tue, 12 Oct 2010 09:24:05 -0500
From: Anthony Liguori <anthony@codemonkey.ws>
MIME-Version: 1.0
Subject: Re: [Qemu-devel] qdev: Some ISA devices don't handle second
	instantiation gracefully
References: <m3ocazlglv.fsf@blackfin.pond.sub.org>	<m3aamjk1lu.fsf@blackfin.pond.sub.org>
	<4CB46239.60500@codemonkey.ws>
	<m3fwwbfrd8.fsf@blackfin.pond.sub.org>
In-Reply-To: <m3fwwbfrd8.fsf@blackfin.pond.sub.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
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.nongnu.org/archive/html/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: Markus Armbruster <armbru@redhat.com>
Cc: qemu-devel@nongnu.org, "H. Peter Anvin" <hpa@linux.intel.com>, Gerd Hoffmann <kraxel@redhat.com>, Alexander Graf <agraf@suse.de>, "Richard W. M. Jones" <rjones@redhat.com>

On 10/12/2010 08:54 AM, Markus Armbruster wrote:
>
>> I think the proper thing to do is remove all exit(1)s and propagate
>> errors instead.
>>      
> exit() is good enough during startup, i.e. -device.  It's wrong for hot
> plug; anything to be used in a hot plug path must propagate errors.  We
> could keep exiting in code that's only used by non-hotpluggable devices.
>    

I'm not really suggesting that we move exit()s out of all devices right 
now.  Am just suggesting that we bump them up one level.

>> A simple approach would be to make register_ioport_{read,write}()
>> return an int, then do a query-replace on the source tree to make all
>> invocations of it simply check the return value and exit if it's
>> non-zero.
>>      
> In similar cases, we've used a simple FOO_nofail() wrapper in places
> that want to exit.
>
> I'll see what I can do.
>    

I prefer propagating the exit but am not deeply opposed to nofail().

Regards,

Anthony Liguori