Re: [PATCH v2] hw/smbios: fix memory corruption for large guests due to handle overlap

[Ani Sinha]

On Fri, 4 Feb 2022, Ani Sinha wrote:

> On Fri, Feb 4, 2022 at 17:48 Igor Mammedov <imammedo@redhat.com> wrote:
>
> > On Fri, 4 Feb 2022 12:05:58 +0100
> > Gerd Hoffmann <kraxel@redhat.com> wrote:
> >
> > >   Hi,
> > >
> > > > Another question is why we split memory on 16Gb chunks, to begin with.
> > > > Maybe instead of doing so, we should just add 1 type17 entry describing
> > > > whole system RAM size. In which case we don't need this dance around
> > > > handle offsets anymore.
> > >
> > > Maybe to make the entries look like they do on physical hardware?
> > > i.e. DIMM size is a power of two?  Also physical 1TB DIMMs just
> > > don't exist?
> >
> > Does it have to be a DIMM, we can make it Other/Unknown/Row of chips/Chip
> > to more close to builtin memory that's is our main ram is.
>
>
> My concern here is even though the spec has provisions for those form
> factors I wonder if the guests only expect dimm ? Will it break some guest
> operating system?

I am not sure if we have come to any conclusion here. However, I have sent
a v2 patch series with Igor's suggestion addressed.