From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:49032) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1aiIu2-0002Lg-DT for qemu-devel@nongnu.org; Tue, 22 Mar 2016 05:47:07 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1aiItz-0003vZ-7G for qemu-devel@nongnu.org; Tue, 22 Mar 2016 05:47:06 -0400 Sender: Paolo Bonzini References: <56E93A22.1080102@de.ibm.com> <56E93ECE.10103@redhat.com> <56E9425C.8030201@de.ibm.com> <56E957AD.2050005@redhat.com> <56E961EA.4090908@de.ibm.com> <56E9638B.5090204@redhat.com> <20160317003906.GA23821@ad.usersys.redhat.com> <56EA8EEE.2020801@linux.vnet.ibm.com> <20160321105718.GA7710@ad.usersys.redhat.com> <56F0EFCA.4080003@linux.vnet.ibm.com> <20160322071804.GC24999@ad.usersys.redhat.com> <20160322100706.597dab1b.cornelia.huck@de.ibm.com> From: Paolo Bonzini Message-ID: <56F11492.5040103@redhat.com> Date: Tue, 22 Mar 2016 10:46:58 +0100 MIME-Version: 1.0 In-Reply-To: <20160322100706.597dab1b.cornelia.huck@de.ibm.com> Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH 0/4] Tweaks around virtio-blk start/stop List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Cornelia Huck , Fam Zheng Cc: Kevin Wolf , qemu-block@nongnu.org, tu bo , "Michael S. Tsirkin" , qemu-devel@nongnu.org, Christian Borntraeger , Stefan Hajnoczi On 22/03/2016 10:07, Cornelia Huck wrote: > So far, we had the best results with my refactoring + the mutex/bh > change. Two problems: > > - We don't really understand yet why my refactoring helps, but passing > the assign value is a good canditate (and it's agreed that this fixes a > bug, I think.) > - There's some problem with the bh, if I understood Stefan correctly. They can be fixed with just an extra object_ref/object_unref. I didn't understand that Tu Bo also needed the BH fix, and with that information it makes sense. Passing the assign value ensures that ioeventfd remains always assigned. With the CPU threads out of the picture, the BH becomes enough to make everything thread-safe. Paolo