From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1LjD7p-0006lm-G2 for qemu-devel@nongnu.org; Mon, 16 Mar 2009 09:45:05 -0400 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1LjD7l-0006lD-2Z for qemu-devel@nongnu.org; Mon, 16 Mar 2009 09:45:05 -0400 Received: from [199.232.76.173] (port=58171 helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1LjD7k-0006lA-Sy for qemu-devel@nongnu.org; Mon, 16 Mar 2009 09:45:00 -0400 Received: from verein.lst.de ([213.95.11.210]:51686) by monty-python.gnu.org with esmtps (TLS-1.0:DHE_RSA_3DES_EDE_CBC_SHA1:24) (Exim 4.60) (envelope-from ) id 1LjD7k-0006dt-Ax for qemu-devel@nongnu.org; Mon, 16 Mar 2009 09:45:00 -0400 Received: from verein.lst.de (localhost [127.0.0.1]) by verein.lst.de (8.12.3/8.12.3/Debian-7.1) with ESMTP id n2GDiwIF004649 (version=TLSv1/SSLv3 cipher=EDH-RSA-DES-CBC3-SHA bits=168 verify=NO) for ; Mon, 16 Mar 2009 14:44:58 +0100 Received: (from hch@localhost) by verein.lst.de (8.12.3/8.12.3/Debian-6.6) id n2GDiwjA004647 for qemu-devel@nongnu.org; Mon, 16 Mar 2009 14:44:58 +0100 Date: Mon, 16 Mar 2009 14:44:58 +0100 From: Christoph Hellwig Subject: Re: [Qemu-devel] [PATCH 2/5] Convert vectored aio emulation to use a dedicated pool Message-ID: <20090316134458.GA4612@lst.de> References: <1237209071-26942-1-git-send-email-avi@redhat.com> <1237209071-26942-3-git-send-email-avi@redhat.com> <20090316133202.GB4200@lst.de> <49BE5692.1030509@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <49BE5692.1030509@redhat.com> Reply-To: qemu-devel@nongnu.org List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org On Mon, Mar 16, 2009 at 03:39:30PM +0200, Avi Kivity wrote: > Right, when you don't use a synthetic aiocb, the need goes away. > > We'll still have one for the block-dma-helpers, due to the need to > fragment a request when bounce resources are exhausted. The pools might also help with the independent scsi-generic implementation, by not requiring it to be tied to the block layer while still beeing able to use posix-aio-compat.c. Still need to sort out the setup issue, but we might just keep a dummy BlockDriverState for it for now.