From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:42579)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <den@openvz.org>) id 1abUXq-0000zB-BI
	for qemu-devel@nongnu.org; Thu, 03 Mar 2016 09:48:03 -0500
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <den@openvz.org>) id 1abUXk-0004zN-O1
	for qemu-devel@nongnu.org; Thu, 03 Mar 2016 09:48:02 -0500
Received: from mx2.parallels.com ([199.115.105.18]:42710)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <den@openvz.org>) id 1abUXk-0004zB-I0
	for qemu-devel@nongnu.org; Thu, 03 Mar 2016 09:47:56 -0500
References: <1457012886-7626-1-git-send-email-den@openvz.org>
	<56D840D4.4010109@redhat.com> <56D841C6.1060303@openvz.org>
	<56D84487.5070009@redhat.com> <56D84555.7090003@openvz.org>
	<56D8470C.5000201@redhat.com> <56D84955.6070508@openvz.org>
	<56D84B85.70408@redhat.com>
From: "Denis V. Lunev" <den@openvz.org>
Message-ID: <56D84E93.90904@openvz.org>
Date: Thu, 3 Mar 2016 17:47:47 +0300
MIME-Version: 1.0
In-Reply-To: <56D84B85.70408@redhat.com>
Content-Type: text/plain; charset="windows-1252"; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] [PATCH 1/1] log: fix hanged connect from
	virt-manager to libvirt
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/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: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Paolo Bonzini <pbonzini@redhat.com>
Cc: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>, qemu-devel@nongnu.org, Dimitris Aragiorgis <dimara@arrikto.com>

On 03/03/2016 05:34 PM, Paolo Bonzini wrote:
>
> On 03/03/2016 15:25, Denis V. Lunev wrote:
>>> Actually, the patch in v1 is fine.  My worry after looking at your patch
>>> was that I didn't have the dup2(stdout, stderr) case.  However, with my
>>> change you can never call qemu_log_close if is_daemonized(), because
>>> even the monitor command "logfile" cannot set logfilename to NULL.
>> IMHO you are wrong.
> Possible. :)
>
>> void qemu_set_log_filename(const char *filename)
>> {
>>      g_free(logfilename);
>>      logfilename = g_strdup(filename);
>>      qemu_log_close();
>>      qemu_set_log(qemu_loglevel);
>> }
>>
>> static void hmp_logfile(Monitor *mon, const QDict *qdict)
>> {
>>      qemu_set_log_filename(qdict_get_str(qdict, "filename"));
>> }
>>
>> This means that we will have qemu_log_close()
>> called in ANY case, even daemonized.
>>  From my point of view stderr will continue to be mapped
>> to the old file if we request to stop logging either by
>> zero mask or by setting empty filename.
> Yes, but filename will never be NULL in hmp_logfile, will it?  It's
> declared as 'F' in hmp-commands.hx, not as 'F?'.  If this is changed, I
> agree that the dup2 needs to be added.
>
> A different issue is that do_qemu_set_log *exits* instead of printing an
> Error if fopen fails. In my opinion, in this case logging should not
> even be disabled in this case.  But it can and should be fixed separately.
>
> The code is really ugly, it is old and used to be just a debugging aid.
>
> Paolo
ok, you are finally right. Your original code is correct and will
work even in these corner case.

Will it make sense to send a patch which will replace that code
to my one as followup + fix if 'fopen' will fail?

Den