From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:45050)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1YDyKX-0000LX-Ha
	for qemu-devel@nongnu.org; Wed, 21 Jan 2015 11:40:34 -0500
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1YDyKS-0003Y4-Ko
	for qemu-devel@nongnu.org; Wed, 21 Jan 2015 11:40:33 -0500
Received: from mx1.redhat.com ([209.132.183.28]:47484)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1YDyKS-0003Xy-9Z
	for qemu-devel@nongnu.org; Wed, 21 Jan 2015 11:40:28 -0500
Message-ID: <54BFD677.2070905@redhat.com>
Date: Wed, 21 Jan 2015 17:40:23 +0100
From: Paolo Bonzini <pbonzini@redhat.com>
MIME-Version: 1.0
References: <1421857693-12014-1-git-send-email-jslaby@suse.cz>
In-Reply-To: <1421857693-12014-1-git-send-email-jslaby@suse.cz>
Content-Type: text/plain; charset=windows-1252
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] [PATCH v4 1/1] pci-host: add educational driver
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: Jiri Slaby <jslaby@suse.cz>, qemu-devel@nongnu.org



On 21/01/2015 17:28, Jiri Slaby wrote:
> +        if (atomic_fetch_or(&edu->status, EDU_STATUS_COMPUTING) & EDU_STATUS_COMPUTING) {

Theoretically the other thread could see EDU_STATUS_COMPUTING here and
not enter the condvar wait.  So...

> +            break;
> +        }
> +        qemu_mutex_lock(&edu->thr_mutex);
> +        edu->fact = val;
> +        qemu_cond_signal(&edu->thr_cond);
> +        qemu_mutex_unlock(&edu->thr_mutex);

... just one change:

   if (atomic_read(&edu->status) & EDU_STATUS_COMPUTING) {
       break;
   }
   /* EDU_STATUS_COMPUTING cannot go 0->1 concurrently, because
    * it is only set in this function and it is under the iothread
    * mutex.
    */
   qemu_mutex_lock(&edu->thr_mutex);
   edu->fact = val;
   atomic_or(&edu->status), EDU_STATUS_COMPUTING);
   qemu_cond_signal(&edu->thr_cond);
   qemu_mutex_unlock(&edu->thr_mutex);

If you are okay with this change, I'll apply the patch to my tree.

Paolo