Re: [PATCH v3 05/18] dmaengine: idxd: add IMS support in base driver

[Thomas Gleixner]

On Wed, Sep 30 2020 at 18:07, Ashok Raj wrote:
> On Wed, Sep 30, 2020 at 11:57:22PM +0200, Thomas Gleixner wrote:
>
> Devices exposed to guest need host OS support for programming interrupt
> entries in the IOMMU interrupt remapping table. VFIO provides those 
> services for standard interrupt schemes like MSI/MSIx for instance. 
> Since IMS is device specific VFIO can't provide an intercept when 
> IMS entries are programmed by the guest OS.

Why is IMS exposed to the guest in the first place? You expose a
subdevice to a guest, right? And that subdevice should not even know
that IMS exists simply because IMS is strictly host specific.

The obvious emulation here is to make the subdevice look like a PCI
device and expose emulated MSIX (not MSI) which is intercepted when
accessing the MSIX table and then redirected to the proper place along
with IRTE and PASID and whatever.

>> Also this stuff is host side and not guest side. I seriously doubt that
>> you want to hand in the whole PCI device which contains the IMS
>
> You are right, but nothing prevents a user from simply taking a full PCI
> device and assign to guest. 

You surely can and should prevent that because it makes no sense and
cannot work.

That's why you want a generic check for 'this device is magic SIOV or
whatever' and not something burried deep into a driver..

Thanks,

        tglx