From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:38568) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1aj7Ha-00035c-2Y for qemu-devel@nongnu.org; Thu, 24 Mar 2016 11:34:47 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1aj7HW-0007lE-1R for qemu-devel@nongnu.org; Thu, 24 Mar 2016 11:34:46 -0400 Received: from mail-wm0-x244.google.com ([2a00:1450:400c:c09::244]:36476) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1aj7HV-0007l8-PA for qemu-devel@nongnu.org; Thu, 24 Mar 2016 11:34:41 -0400 Received: by mail-wm0-x244.google.com with SMTP id l68so13604104wml.3 for ; Thu, 24 Mar 2016 08:34:41 -0700 (PDT) References: <1457014732-19428-1-git-send-email-marcel@redhat.com> <56E7C13B.6010601@cn.fujitsu.com> <56E7F597.9060506@redhat.com> <56EA80A5.50101@cn.fujitsu.com> From: Marcel Apfelbaum Message-ID: <56F40904.7080208@gmail.com> Date: Thu, 24 Mar 2016 17:34:28 +0200 MIME-Version: 1.0 In-Reply-To: <56EA80A5.50101@cn.fujitsu.com> Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH V4] hw/pxb: add chassis_nr property Reply-To: marcel@redhat.com List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Cao jin , Marcel Apfelbaum , qemu-devel@nongnu.org Cc: ehabkost@redhat.com, laine@redhat.com, mst@redhat.com On 03/17/2016 12:02 PM, Cao jin wrote: > hi > > On 03/15/2016 07:44 PM, Marcel Apfelbaum wrote: >> On 03/15/2016 10:00 AM, Cao jin wrote: > >>> >>> And I have another personal question: In qemu design, it seems every >>> pci bridge reside in a separate chassis, what`s benefit? why don`t >>> put them all in the main chassis? >> >> Please have a look on pci-to-pci bridge specification, chapter 13, slot >> numbering. >> > > thanks for the hint. > I still have a question: in docs/pci_expander_bridge.txt, it says: create a TYPE_PCI_BRIDGE_DEV to enable hotplug support. > But I didn`t see it can hotplug as following step: > > 1: ./qemu-system-x86_64 -device pxb,id=br,bus=pci.0,bus_nr=2 -hda linux.img -smp 2 --enable-kvm -m 1024 -monitor stdio > > 2: in monitor, type: device_add e1000,bus=br > result: no message output to monitor, and don`t see e1000 nic in guest > > Is is a bug or I test it in a wrong way? Hi, You used hot-plug correctly, it doesn't work yet since I am still working on this feature. The bridge gives us the basics environment for hot-plug, what is missing is ACPI "bsel" mechanism. Thanks, Marcel >