From: Keir Fraser Subject: Re: [rfc 00/18] ioemu: use devfn instead of slots as the unit for passthrough Date: Mon, 23 Feb 2009 03:31:30 -0800 Message-ID: References: <20090223065530.GA22192@verge.net.au> Mime-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20090223065530.GA22192@verge.net.au> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Simon Horman , Yuji Shimada Cc: "xen-devel@lists.xensource.com" , Ian Jackson List-Id: xen-devel@lists.xenproject.org On 22/02/2009 22:55, "Simon Horman" wrote: >> The another solution is expanding GSI to 127. I don't sure it is >> possible, but sharing virtual GSI will not occur. > > That thought crossed my mind too, I will investigate further. > But I think that ideally it would need to be expanded to 143 > as the first 16 GSI are currently reserved for ISA. Can't xend or qemu do a search of PCI slot space to pick a virtual devfn:intx that doesn't conflict (or that purposely does conflict, if there are any cases where you want that)? 32 non-legacy GSIs should be enough to avoid any aliasing if a bit of effort is put in to avoid it. -- Keir