Re: [PATCH] x86/pvh: fix identity mapping of low 1MB

["Roger Pau Monné" <roger.pau@citrix.com>]

On Fri, Oct 13, 2023 at 10:39:06AM +0200, Roger Pau Monné wrote:
> On Thu, Oct 12, 2023 at 12:48:20AM +0800, Andrew Cooper wrote:
> > On 11/10/2023 11:37 pm, Roger Pau Monne wrote:
> > But a question first.  Is this from debugging the XSA-442 fallout?  If
> > so, it's probably worth mentioning the hardware we saw this on (which
> > IIRC was fairly old AMD), and that XSA-442 unmasked a pre-existing bug. 
> > And we think it's USB/PS2 emulation?
> 
> Yeah, can add this all.  It is triggered by pinot{0,1} which is AMD
> Fam15h (AMD Opteron(tm) Processor 3350 HE).

Just re-checked this in order to be sure, and no, the IOMMU faults are
already there prior to XSA-442 for the dom0pvh tests, but they don't
cause issues to osstest, and hence we didn't realize.  I just came
across those as I was triaging a different issue and looking at the
serial log.

Hence I won't add any mentions to XSA-442 in the commit message.

Thanks, Roger.