[libvirt] [PATCH v3 25/26] qemu: Isolate hostdevs on pSeries guests

[David Gibson]

On Thu, 29 Jun 2017 17:37:30 +0200
Andrea Bolognani <abologna at redhat.com> wrote:

> On Wed, 2017-06-28 at 18:22 -0400, Laine Stump wrote:
>  [...]  
>  [...]  
> 
> Right, I hadn't considered that case. I'll make sure it is
> handled correctly.
> 
>  [...]  
> 
> Hm, that's quite a pickle.
> 
> Choosing a synthetic isolation group that is guaranteed not
> to clash with any host device or other VFs sounds like it
> might require some annoying bookkeeping.
> 
> I'm not sure isolating SR-IOV VFs is really needed, though,
> because by virtue of being subject to their own layer of
> virtualization / abstraction they might being unable to take
> advantage of EEH and similar isolation-dependent features.
> 
> David, what's your take on this?

So, I still need to talk to the IBMers some more about this, but I'm
pretty sure we want VF isolation.  I believe EEH should work on VFs,
but it will break if there is >1 VF on the vPHB (since every VF should
be in its own IOMMU group).

And, yes, things would be far nicer if EEH still worked with multiple
(host) IOMMU groups on the vPHB, but doing so is a real PITA, and
no-one is working on it.

> 
>  [...]  
>  [...]  
> 
> I don't think PCIe Root Ports provide any isolation; not of
> the kind we're talking about here, at least. So the concepts
> implemented here don't apply AFAICT.
> 
> -- 
> Andrea Bolognani / Red Hat / Virtualization


-- 
David Gibson <dgibson at redhat.com>
Principal Software Engineer, Virtualization, Red Hat
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 819 bytes
Desc: OpenPGP digital signature
URL: <http://listman.redhat.com/archives/libvir-list/attachments/20170704/57897929/attachment-0001.sig>