Re: [PATCH v1 2/2] KVM: arm64: allow the VM to select DEVICE_* and NORMAL_NC for IO memory

[Jason Gunthorpe]

On Fri, Oct 20, 2023 at 04:03:57PM +0200, Lorenzo Pieralisi wrote:
> > My point is if this becomes a real world concern we have a solid
> > answer on how to resolve it - fix the VFIO driver to have a stronger
> > barrier before reset.
> 
> Just to make sure I am parsing this correctly: this case above is
> related to a non-PCI VFIO device passthrough where a guest would want to
> map the device MMIO at stage-1 with normal-NC memory type (well, let's
> say with a memory attribute != device-nGnRE - that combined with the new
> stage-2 default might cause transactions ordering/grouping trouble with
> eg device resets), correct ? 

This is what I have understood was Will's concern, yes.

> IIRC, all requests related to honouring "write-combine" style
> stage-1 mappings were for PCI(e) devices but that's as far as what
> *I* was made aware of goes.

Yes, this is what I am aware of as well.

Though I do not object to the idea from the VFIO side that platform
devices would also have to support NormalNC too.

The theoretical missing peice is that someone would say they have a
SOC issue XYZ and thus their VFIO platform devices must fully block
NormalNC. I suggest if someone comes with this and really, really
wants VFIO, then we could use a VMA flag to indicate that KVM must not
upgrade it. Currently I have no knowledge of such a thing existing.

With PCI we've made the argument that if NormalNC is broken unsafe for
KVM in the SOC then probably so is Device-*. I think the same basic
argument holds for platform devices too. Thus I'm skeptical that
someone can come and say they have SOC issue XYZ and NormalNC is
broken but Device-* is perfectly safe.

> We will do that, I will start adding the recent discussions to the
> new documentation file. Side note: for those who attend LPC it would be
> useful to review the resulting documentation together there, it should
> happen around v6.7-rc1.

I will be there, let me know

Thanks,
Jason