RE: [PATCH v2 00/22] intel_iommu: expose Shared Virtual Addressing to VMs

[Liu, Yi L]

> From: Peter Xu <address@hidden>
> Sent: Friday, April 3, 2020 2:13 AM
> To: Liu, Yi L <address@hidden>
> Subject: Re: [PATCH v2 00/22] intel_iommu: expose Shared Virtual Addressing to
> VMs
> 
> On Sun, Mar 29, 2020 at 09:24:39PM -0700, Liu Yi L wrote:
> > Tests: basci vSVA functionality test,
> 
> Could you elaborate what's the functionality test?  Does that contains
> at least some IOs go through the SVA-capable device so the nested page
> table is used?  I thought it was a yes, but after I notice that the
> BIND message flags seems to be wrong, I really think I should ask this
> loud..

as just replied, in the verification, only the SRE bit is used. So it's not
spotted. In my functionality test, I've passthru a SVA-capable device
and issue SVA transactions.

> > VM reboot/shutdown/crash,
> 
> What's the VM crash test?

it's ctrl+c to kill the VM.

> > kernel build in
> > guest, boot VM with vSVA disabled, full comapilation with all archs.
> 
> I believe I've said similar things, but...  I'd appreciate if you can
> also smoke on 2nd-level only with the series applied.

yeah, you mean the legacy case, I booted with such config.

Regards,
Yi Liu