This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
Home Products Support Community News


[Xen-devel] Re: AMD iommu , pci passthrough problems assigning devices t

To: Wei Wang2 <wei.wang2@xxxxxxx>
Subject: [Xen-devel] Re: AMD iommu , pci passthrough problems assigning devices to HVM
From: Sander Eikelenboom <linux@xxxxxxxxxxxxxx>
Date: Wed, 11 Aug 2010 13:26:09 +0200
Cc: "Przywara, Andre" <Andre.Przywara@xxxxxxx>, Joerg Roedel <joro@xxxxxxxxxx>, "Huang2, Wei" <Wei.Huang2@xxxxxxx>, "xen-devel@xxxxxxxxxxxxxxxxxxx" <xen-devel@xxxxxxxxxxxxxxxxxxx>
Delivery-date: Wed, 11 Aug 2010 04:27:37 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
In-reply-to: <201008111212.01455.wei.wang2@xxxxxxx>
List-help: <mailto:xen-devel-request@lists.xensource.com?subject=help>
List-id: Xen developer discussion <xen-devel.lists.xensource.com>
List-post: <mailto:xen-devel@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=unsubscribe>
Organization: Eikelenboom IT services
References: <11610190890.20100810142614@xxxxxxxxxxxxxx> <201008111212.01455.wei.wang2@xxxxxxx>
Sender: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
Hello Wei,

It's an pci express usb3/xhci add-on card, there also seemed to be an alias 
problem with another device, see this thread on LKML, it also contains the 
result of amd_iommu_dump on baremetal 2.6.35 kernel.
At first the kernel didn't boot at all, though xen-unstable did load, so i 
guess there are a few differences in the amd iommu implementation and parsing 
between the linux kernel and xen.


Would be nice to have some good explanation of the problem, so i could contact 
MSI to get the bios fixed.


Wednesday, August 11, 2010, 12:11:58 PM, you wrote:

> Hi,
> In your debug output I saw:
> (XEN) AMD-Vi: Fail to find iommu for device04:00.0
> This means no iommu attached to 04:00.0, so it cannot be assigned to any hvm 
> guest. What is this device in your system? An add-on device or on board 
> device? It looks like BIOS failed to create an IVRS entry for this device.

> Wei

> On Tuesday 10 August 2010 14:26:14 Sander Eikelenboom wrote:
>> Hi,
>> I'm running a machine with:
>> - Hypervisor: xen-unstable
>> - Dom0: xen/next-  commit c3f5aba942c7516a5c76c47345ff0a473daf2542
>> Hardware:
>> -MSI 890FXA-GD70 with AMD IOMMU
>> -Phenom X6
>> I'm trying to do PCI passthrough to a HVM, this fails with:
>> Error: Failed to assign device to IOMMU
>> (0000:04:00.0@100,msitranslate=1,power_mgmt=0) Error: pci: to avoid
>> potential security issue, 0000:08:00.0 is not allowed to be assigned to
>> guest since it is behind PCIe switch that does not support or enable ACS.
>> I have found this thread https://bugzilla.redhat.com/show_bug.cgi?id=526713
>> regarding the ACS problem, which says the pci bridges need these
>> capabilities: ScrValid+, ReqRedir+, CmpltRedir+, and UpstreamFwd+
>> When i do a lspci -vvv on my system i see these capabilities, but i do see
>> some disturbing debug message on boot when amd_iommu_debug is on.
>> Attached:
>> - xm info
>> - xm dmesg (booted with amd_iommu_debug on)
>> - dmesg
>> - lspci -vvv
>> --
>> Sander

Best regards,
 Sander                            mailto:linux@xxxxxxxxxxxxxx

Xen-devel mailing list

<Prev in Thread] Current Thread [Next in Thread>