From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([208.118.235.92]:57552) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1SWXoz-0004HG-PO for qemu-devel@nongnu.org; Mon, 21 May 2012 14:59:11 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1SWXoy-0006nS-27 for qemu-devel@nongnu.org; Mon, 21 May 2012 14:59:09 -0400 Received: from goliath.siemens.de ([192.35.17.28]:33376) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1SWXox-0006mk-OG for qemu-devel@nongnu.org; Mon, 21 May 2012 14:59:07 -0400 Message-ID: <4FBA9071.3010204@siemens.com> Date: Mon, 21 May 2012 15:58:57 -0300 From: Jan Kiszka MIME-Version: 1.0 References: <4FBA3F8B.9060103@siemens.com> <20120521173358.GA13690@redhat.com> In-Reply-To: <20120521173358.GA13690@redhat.com> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH 1/2] pci: Add pci_device_get_host_irq List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: "Michael S. Tsirkin" Cc: Alex Williamson , Marcelo Tosatti , qemu-devel , Avi Kivity On 2012-05-21 14:34, Michael S. Tsirkin wrote: > On Mon, May 21, 2012 at 10:13:47AM -0300, Jan Kiszka wrote: >> Add a PCI IRQ path discovery function that walks from a given device to >> the host bridge, returning the IRQ number that is reported to the >> attached interrupt controller. For this purpose, another PCI bridge >> callback function is introduced: map_host_irq. It is so far only >> implemented by the PIIX3, other host bridges can be added later on as >> required. >> >> Will be used for KVM PCI device assignment. >> >> Signed-off-by: Jan Kiszka > > interrupt injection is data path even for emulated devices. > So instead of special casing device assignment I would like to see all > devices converted to an API that caches irqs. > > This will likely mean that we can maintain the final > irq as part of the pci device structure, and > this api will simply return it. Yep, I definitely agree. It's just that such a design has to please even more users than PCI devices, thus will likely take longer to settle than the device assignment effort. Therefore I decided to rush forward with an intermediate approach first. Jan -- Siemens AG, Corporate Technology, CT T DE IT 1 Corporate Competence Center Embedded Linux