Re: [Xen-devel] VT-d faults with Integrated Intel graphics on 4.6

[Malcolm Crossley <malcolm.crossley@xxxxxxxxxx>]

On 25/08/15 15:43, Konrad Rzeszutek Wilk wrote:
> On Tue, Aug 25, 2015 at 02:55:31PM +0800, Chen, Tiejun wrote:
>> On 8/25/2015 8:19 AM, Tamas K Lengyel wrote:
>>> Hi everyone,
>>> I saw some people passingly mention this on the list before but just in
>>> case it has been missed, my serial is also being spammed with the following
>>> printouts with both Xen 4.6 RC1 and the latest staging build:
>>>
>>> ...
>>> (XEN) [VT-D]DMAR:[DMA Read] Request device [0000:00:02.0] fault addr
>>> 33487d7000, iommu reg = ffff82c000201000
>>> (XEN) [VT-D]DMAR: reason 06 - PTE Read access is not set
>>> (XEN) [VT-D]DMAR:[DMA Read] Request device [0000:00:02.0] fault addr
>>> 33487d7000, iommu reg = ffff82c000201000
>>> (XEN) [VT-D]DMAR: reason 06 - PTE Read access is not set
>>> (XEN) [VT-D]DMAR:[DMA Read] Request device [0000:00:02.0] fault addr
>>> 33487d7000, iommu reg = ffff82c000201000
>>> (XEN) [VT-D]DMAR: reason 06 - PTE Read access is not set
>>> (XEN) [VT-D]DMAR:[DMA Read] Request device [0000:00:02.0] fault addr
>>> 33487d7000, iommu reg = ffff82c000201000
>>> (XEN) [VT-D]DMAR: reason 06 - PTE Read access is not set
>>> (XEN) [VT-D]DMAR:[DMA Read] Request device [0000:00:02.0] fault addr
>>> 2610742000, iommu reg = ffff82c000201000
>>> (XEN) [VT-D]DMAR: reason 07 - Next page table ptr is invalid
>>> ...
>>>
>>

I think this problem is caused by missing IOMMU mappings for the RMRR regions
if the domain does not have shared EPT enabled. This includes PV Dom 0.

I have posted a patch to fix the issue:

http://lists.xen.org/archives/html/xen-devel/2015-08/msg02090.html



>> What's your platform? BDW? And how much memory is set to your guest OS?
> 
> Is see this as well. But oddly enough - only when I use the AMT feature
> (normally I just use serial console on the machine).
> 
> The platform is             /DQ67SW, BIOS
> SWQ6710H.86A.0066.2012.1105.1504 11/05/2012
> 
> There is no guest OS - this is initial domain. And I boot with 2GB:
>  Released 0 page(s)
> 
> Xen: [mem 0x0000000000000000-0x0000000000099fff] usable
> Xen: [mem 0x000000000009a800-0x00000000000fffff] reserved
> Xen: [mem 0x0000000000100000-0x000000001fffffff] usable
> Xen: [mem 0x0000000020000000-0x00000000201fffff] reserved
> Xen: [mem 0x0000000020200000-0x000000003fffffff] usable
> Xen: [mem 0x0000000040000000-0x00000000401fffff] reserved
> Xen: [mem 0x0000000040200000-0x0000000080465fff] usable
> Xen: [mem 0x0000000080466000-0x000000009e855fff] unusable
> Xen: [mem 0x000000009e856000-0x000000009e85efff] ACPI data
> Xen: [mem 0x000000009e85f000-0x000000009e8a9fff] ACPI NVS
> Xen: [mem 0x000000009e8aa000-0x000000009e8b1fff] unusable
> Xen: [mem 0x000000009e8b2000-0x000000009e9a4fff] reserved
> Xen: [mem 0x000000009e9a5000-0x000000009e9a6fff] unusable
> Xen: [mem 0x000000009e9a7000-0x000000009ebc5fff] reserved
> Xen: [mem 0x000000009ebc6000-0x000000009ebc6fff] unusable
> Xen: [mem 0x000000009ebc7000-0x000000009ebd6fff] reserved
> Xen: [mem 0x000000009ebd7000-0x000000009ebf4fff] ACPI NVS
> Xen: [mem 0x000000009ebf5000-0x000000009ec18fff] reserved
> Xen: [mem 0x000000009ec19000-0x000000009ec5bfff] ACPI NVS
> Xen: [mem 0x000000009ec5c000-0x000000009ee7bfff] reserved
> Xen: [mem 0x000000009ee7c000-0x000000009effffff] unusable
> Xen: [mem 0x000000009f800000-0x00000000bf9fffff] reserved
> Xen: [mem 0x00000000fec00000-0x00000000fec00fff] reserved
> Xen: [mem 0x00000000fed1c000-0x00000000fed3ffff] reserved
> Xen: [mem 0x00000000fed90000-0x00000000fed91fff] reserved
> Xen: [mem 0x00000000fee00000-0x00000000feefffff] reserved
> Xen: [mem 0x00000000ff000000-0x00000000ffffffff] reserved
> Xen: [mem 0x0000000100000000-0x000000043e5fffff] unusable
> 
>>
>> Just at first glance to fault address, this seems be issued from some 
>> known erratas on BDS and SKL.
> 
> I am runnig v4.2-rc8.
>>
>> Thanks
>> Tiejun
>>
>>> The device in question is an integrated Intel graphics card:
>>>
>>> 00:02.0 VGA compatible controller: Intel Corporation 2nd Generation Core
>>> Processor Family Integrated Graphics Controller (rev 09)
> 
> Same device..

This is a sandy bridge device which doesn't support shared EPT for GPU device.


>>>
>>> The only way I found to stop the messages from making my serial connection
>>> useless was by assigning the device to xen-pciback.

This will cause the GPU to be reset and the reset stopped the GPU accessing the
RMRR region.

>>>
>>> Cheers,
>>> Tamas
>>>
>>>
>>>
>>> _______________________________________________
>>> Xen-devel mailing list
>>> Xen-devel@xxxxxxxxxxxxx
>>> http://lists.xen.org/xen-devel
>>>
>>
>> _______________________________________________
>> Xen-devel mailing list
>> Xen-devel@xxxxxxxxxxxxx
>> http://lists.xen.org/xen-devel
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@xxxxxxxxxxxxx
> http://lists.xen.org/xen-devel
> 


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel