[Xen-devel] PVH PCI passthrough for DomUs

[Xen-devel] PVH PCI passthrough for DomUs

[Roman Shaposhnik]

Hi!

I've just tried this with Xen 4.13.0 and it seems like that is still
not supported.

This makes me curious if anybody is working on this and whether
there's anything we can do to help accelerate the effort.

Thanks,
Roman.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] PVH PCI passthrough for DomUs

[Wei Liu]

Cc Roger

On Sun, Jan 19, 2020 at 11:30:42PM -0800, Roman Shaposhnik wrote:
> Hi!
> 
> I've just tried this with Xen 4.13.0 and it seems like that is still
> not supported.
> 
> This makes me curious if anybody is working on this and whether
> there's anything we can do to help accelerate the effort.
> 
> Thanks,
> Roman.
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xenproject.org
> https://lists.xenproject.org/mailman/listinfo/xen-devel

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] PVH PCI passthrough for DomUs

[Roger Pau Monné]

On Mon, Jan 27, 2020 at 12:27:18PM +0000, Wei Liu wrote:
> Cc Roger

Thanks :).

> On Sun, Jan 19, 2020 at 11:30:42PM -0800, Roman Shaposhnik wrote:
> > Hi!
> > 
> > I've just tried this with Xen 4.13.0 and it seems like that is still
> > not supported.

No, there hasn't been much progress on this sadly.

> > This makes me curious if anybody is working on this and whether
> > there's anything we can do to help accelerate the effort.

The first step would be to get vPCI hooked into the ioreq machinery,
so that a domain can have devices on the emulated PCI bus handled by
vPCI while others are handled by external ioreq emulators. I've posted
a v3 of this work on September:

https://lists.xenproject.org/archives/html/xen-devel/2019-09/msg03278.html

But I haven't got time to go over the comments and post a new version.

Once that's done the remaining step would be to make vPCI safe for
unprivileged guests. We need to assure that guests can only write to
specific bits of the config space, and need to limit the capabilities
that are exposed to the ones Xen knows to be safe to handle. This can
be worked by multiple people concurrently IMO, but requires step 1
(integration with ioreq) to be finished first.

I'm more than happy for someone to pick any of those tasks, including
the integration of vPCI with the ioreq machinery. If not, I expect I
will be able to do some work on this in a couple of weeks, but that
depends on nothing else getting on fire, and me being able to flush my
queue of pending patches.

Would you be up to pick some of these tasks?

I can try to speedup the vPCI ioreq integration if there's people
willing to work on the remaining steps, I haven't done so because I
didn't see much interest in general, and I was expecting to be the
only one working on the remaining steps anyway.

Regards, Roger.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] PVH PCI passthrough for DomUs

[Roger Pau Monné]

Forgot to set 'To:' correctly.

On Mon, Jan 27, 2020 at 03:28:36PM +0100, Roger Pau Monné wrote:
> On Mon, Jan 27, 2020 at 12:27:18PM +0000, Wei Liu wrote:
> > Cc Roger
> 
> Thanks :).
> 
> > On Sun, Jan 19, 2020 at 11:30:42PM -0800, Roman Shaposhnik wrote:
> > > Hi!
> > > 
> > > I've just tried this with Xen 4.13.0 and it seems like that is still
> > > not supported.
> 
> No, there hasn't been much progress on this sadly.
> 
> > > This makes me curious if anybody is working on this and whether
> > > there's anything we can do to help accelerate the effort.
> 
> The first step would be to get vPCI hooked into the ioreq machinery,
> so that a domain can have devices on the emulated PCI bus handled by
> vPCI while others are handled by external ioreq emulators. I've posted
> a v3 of this work on September:
> 
> https://lists.xenproject.org/archives/html/xen-devel/2019-09/msg03278.html
> 
> But I haven't got time to go over the comments and post a new version.
> 
> Once that's done the remaining step would be to make vPCI safe for
> unprivileged guests. We need to assure that guests can only write to
> specific bits of the config space, and need to limit the capabilities
> that are exposed to the ones Xen knows to be safe to handle. This can
> be worked by multiple people concurrently IMO, but requires step 1
> (integration with ioreq) to be finished first.
> 
> I'm more than happy for someone to pick any of those tasks, including
> the integration of vPCI with the ioreq machinery. If not, I expect I
> will be able to do some work on this in a couple of weeks, but that
> depends on nothing else getting on fire, and me being able to flush my
> queue of pending patches.
> 
> Would you be up to pick some of these tasks?
> 
> I can try to speedup the vPCI ioreq integration if there's people
> willing to work on the remaining steps, I haven't done so because I
> didn't see much interest in general, and I was expecting to be the
> only one working on the remaining steps anyway.
> 
> Regards, Roger.
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xenproject.org
> https://lists.xenproject.org/mailman/listinfo/xen-devel

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] PVH PCI passthrough for DomUs

[Roman Shaposhnik]

Thanks for all the background information -- this is very much appreciated!

Looking at the level of effort on this one, we ultimately decided to
stick with HVM for our usecase in Project EVE for now.

If there's a customer pressure -- we'll definitely look into picking it back up.

Thanks,
Roman.

On Mon, Jan 27, 2020 at 6:52 AM Roger Pau Monné <roger.pau@citrix.com> wrote:
>
> Forgot to set 'To:' correctly.
>
> On Mon, Jan 27, 2020 at 03:28:36PM +0100, Roger Pau Monné wrote:
> > On Mon, Jan 27, 2020 at 12:27:18PM +0000, Wei Liu wrote:
> > > Cc Roger
> >
> > Thanks :).
> >
> > > On Sun, Jan 19, 2020 at 11:30:42PM -0800, Roman Shaposhnik wrote:
> > > > Hi!
> > > >
> > > > I've just tried this with Xen 4.13.0 and it seems like that is still
> > > > not supported.
> >
> > No, there hasn't been much progress on this sadly.
> >
> > > > This makes me curious if anybody is working on this and whether
> > > > there's anything we can do to help accelerate the effort.
> >
> > The first step would be to get vPCI hooked into the ioreq machinery,
> > so that a domain can have devices on the emulated PCI bus handled by
> > vPCI while others are handled by external ioreq emulators. I've posted
> > a v3 of this work on September:
> >
> > https://lists.xenproject.org/archives/html/xen-devel/2019-09/msg03278.html
> >
> > But I haven't got time to go over the comments and post a new version.
> >
> > Once that's done the remaining step would be to make vPCI safe for
> > unprivileged guests. We need to assure that guests can only write to
> > specific bits of the config space, and need to limit the capabilities
> > that are exposed to the ones Xen knows to be safe to handle. This can
> > be worked by multiple people concurrently IMO, but requires step 1
> > (integration with ioreq) to be finished first.
> >
> > I'm more than happy for someone to pick any of those tasks, including
> > the integration of vPCI with the ioreq machinery. If not, I expect I
> > will be able to do some work on this in a couple of weeks, but that
> > depends on nothing else getting on fire, and me being able to flush my
> > queue of pending patches.
> >
> > Would you be up to pick some of these tasks?
> >
> > I can try to speedup the vPCI ioreq integration if there's people
> > willing to work on the remaining steps, I haven't done so because I
> > didn't see much interest in general, and I was expecting to be the
> > only one working on the remaining steps anyway.
> >
> > Regards, Roger.
> >
> > _______________________________________________
> > Xen-devel mailing list
> > Xen-devel@lists.xenproject.org
> > https://lists.xenproject.org/mailman/listinfo/xen-devel

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel