Re: [Qemu-devel] [PULL 08/41] intel_iommu: support device iotlb descriptor

[Jason Wang]

On 2017年02月17日 14:18, Liu, Yi L wrote:
> > -----Original Message-----
> > From: Jason Wang [mailto:address@hidden
> > Sent: Thursday, February 16, 2017 1:44 PM
> > To: Liu, Yi L <address@hidden>; Michael S. Tsirkin <address@hidden>; qemu-
> > address@hidden
> > Cc: Peter Maydell <address@hidden>; Eduardo Habkost
> > <address@hidden>; Peter Xu <address@hidden>; Paolo Bonzini
> > <address@hidden>; Richard Henderson <address@hidden>; Tian, Kevin
> > <address@hidden>; Lan, Tianyu <address@hidden>; Alex Williamson
> > <address@hidden>
> > Subject: Re: [Qemu-devel] [PULL 08/41] intel_iommu: support device iotlb
> > descriptor
> >
> >
> >
> > On 2017年02月16日 13:36, Liu, Yi L wrote:
> > > > -----Original Message-----
> > > > From: Qemu-devel
> > > > [mailto:address@hidden
> > > > On Behalf Of Michael S. Tsirkin
> > > > Sent: Tuesday, January 10, 2017 1:40 PM
> > > > To: address@hidden
> > > > Cc: Peter Maydell <address@hidden>; Eduardo Habkost
> > > > <address@hidden>; Jason Wang <address@hidden>; Peter Xu
> > > > <address@hidden>; Paolo Bonzini <address@hidden>; Richard
> > > > Henderson <address@hidden>
> > > > Subject: [Qemu-devel] [PULL 08/41] intel_iommu: support device iotlb
> > > > descriptor
> > > >
> > > > From: Jason Wang <address@hidden>
> > > >
> > > > This patch enables device IOTLB support for intel iommu. The major
> > > > work is to implement QI device IOTLB descriptor processing and notify
> > > > the device through iommu notifier.
> > > Hi Jason/Michael,
> > >
> > > Recently Peter Xu's patch also touched intel-iommu emulation. His
> > > patch shadows second-level page table by capturing iotlb flush from
> > > guest. It would result in page table updating in host. Does this patch
> > > also use the same map/umap API provided by VFIO?
> > Yes, it depends on the iommu notifier too.
> >
> > > If it is, then I think it would also update page table in host. It
> > > looks to be a duplicate update. Pls refer to the following snapshot
> > > captured from section 6.5.2.5 of vtd spec.
> > >
> > > "Since translation requests from a device may be serviced by hardware
> > > from the IOTLB, software must always request IOTLB invalidation
> > > (iotlb_inv_dsc) before requesting corresponding Device-TLB
> > > (dev_tlb_inv_dsc) invalidation."
> > >
> > > Maybe for device-iotlb, we need a separate API which just pass down
> > > the invalidate info without updating page table. Any thoughts?
> > cc Alex.
> >
> > If we want ATS to be visible for guest (but I'm not sure if VFIO support this), 
> > we
> > probably need another notifier or a new flag.
> Jason, for assigned device, I think guest could see ATS if the assigned device
> supports ATS. I can see it when passthru iGPU.
>
> Regards,
> Yi L

Good to know this.

If I understand your suggestion correctly, you want a dedicated API to 
flush a hardware device IOTLB. I'm not sure this is really needed.

There's some discussion of similar issue in the past (when ATS is used 
for virtio-net/vhost), looks like we could solve this by not trigger the 
UNMAP notifier unless it was device IOTLB inv desc if ATS is enabled for 
the device? With this remote IOMMU/IOTLB can only get invalidation 
request once. For VFIO, the under layer IOMMU can deal with hardware 
device IOTLB without any extra efforts.

Does this make sense?

Thanks