Re: [PATCH v2] IOMMU: Prevent VT-d device IOTLB operations on wrong IOMMU

[Malcolm Crossley <malcolm.crossley@citrix.com>]

On 17/06/14 13:28, Malcolm Crossley wrote:
> PCIE ATS allows for device's to contain IOTLB's, the VT-d code was iterating
> around all ATS capable devices and issuing IOTLB operations for all IOMMU's,
> even though each ATS device's is only accessible via one particular IOMMU.

Sorry Paul, I just realised I forgot to fix the punctuation. Doing
things over lunch is sometimes a bad idea.
> 
> Issuing an IOMMU operation to a device not accessible via that IOMMU results
> in an IOMMU timeout because the device does not reply. VT-d IOMMU timeouts
> result in a Xen panic.
> 
> Therefore this bug prevents any Intel system with 2 or more ATS enabled IOMMU's,
> each with an ATS device connected to them, from booting Xen.
> 
> The patch add's a IOMMU pointer to the ATS device struct so the VT-d code can
> ensure it does not issue IOMMU ATS operations on the wrong IOMMU. A void
> pointer has to be used because AMD and Intel IOMMU implementations do not have
> a common IOMMU structure or indexing mechanism.
> 
> Optimised AMD's IOMMU code by leveraging the new IOMMU field in the ATS
> structure.
> 
> Signed-off-by: Malcolm Crossley <malcolm.crossley@citrix.com>
> Acked-by: Kevin Tian <kevin.tian@intel.com>
> 
> diff -r 4708591d8aa8 -r b27eda6dc2e2 xen/drivers/passthrough/amd/iommu_cmd.c
> --- a/xen/drivers/passthrough/amd/iommu_cmd.c
> +++ b/xen/drivers/passthrough/amd/iommu_cmd.c