Re: [PATCH] target/arm: build smbios 19 table

[Peter Maydell]

On Mon, 21 Nov 2022 at 11:43, Peter Maydell <peter.maydell@linaro.org> wrote:
>
> On Mon, 21 Nov 2022 at 11:24, Mihai Carabas <mihai.carabas@oracle.com> wrote:
> >
> > La 21.11.2022 13:02, Peter Maydell a scris:
> > > On Sun, 20 Nov 2022 at 17:53, Mihai Carabas <mihai.carabas@oracle.com> 
> > > wrote:
> > >> La 18.11.2022 21:11, Peter Maydell a scris:
> > >>> On Fri, 18 Nov 2022 at 17:37, Mihai Carabas <mihai.carabas@oracle.com> 
> > >>> wrote:
> > >>>> Use the base_memmap to build the SMBIOS 19 table which provides the 
> > >>>> address
> > >>>> mapping for a Physical Memory Array (from spec [1] chapter 7.20).
> > >>>>
> > >>>> This was present on i386 from commit 
> > >>>> c97294ec1b9e36887e119589d456557d72ab37b5
> > >>>> ("SMBIOS: Build aggregate smbios tables and entry point").
> > >>>>
> > >>>> [1] 
> > >>>> https://urldefense.com/v3/__https://www.dmtf.org/sites/default/files/standards/documents/DSP0134_3.5.0.pdf__;!!ACWV5N9M2RV99hQ!KF2xmQw9nxPvqvNCgDleyVHv4MoZseoZFHmR1veww7O2BmRxSH1spOCNWX-c-FvzcaR_o8PunXSWWH2ECvFqlR4E7vw$
> > >>>>
> > >>>> Signed-off-by: Mihai Carabas <mihai.carabas@oracle.com>
> > >>> Is this a bug fix, or a new feature? What are the consequences
> > >>> of it being missing? Is this important enough to go into the 7.2
> > >>> release? (My default position would be "no", given this has been
> > >>> like this on the virt board for a very long time.)
> > >>
> > >> This is required by ARM SystemReady Virtual Environment [1]. As
> > >> described in the Arm SystemReady Requirements Specification v2.0
> > >>    [2] page 9, 2.5.1 SystemReady Virtual Environment (VE) v1.0
> > >> requirements,: "FirmwareTestSuite (FWTS) must still be used" -> fwts
> > >> checks for the presence of SMBIOS type 19 table and fails the test in
> > >> this case.
> > > OK, so it's a spec requirement. Are there any actual realworld
> > > guests that don't work because we get this wrong ?
> >
> > We do not have a clear example. The thing we hit was the ARM SystemReady
> > certification based on fwts.
>
> Thanks for clarifying; in that case given we're quite far along
> in the 7.2 release cycle I think we shouldn't try to get this
> patch in to that release but instead put it in for 8.0.

I checked how x86 sets up this smbios table, and it doesn't
put any pluggable DIMMs in it, only the initial fixed RAM,
so I'm happy that virt doesn't need to do anything special
there. I've applied this patch to a target-arm-for-8.0 branch
that will eventually turn into target-arm when 7.2 is released
in a few weeks time.

I've tweaked the commit message to make it clear that this
is fixing a spec issue which doesn't to our knowledge
cause any guest OS problems.

thanks
-- PMM