Re: [Qemu-devel] [PATCH v4 07/11] hmat acpi: Build Memory Side Cache Information Structure(s) in ACPI HMAT

[Igor Mammedov <imammedo@redhat.com>]

On Thu, 6 Jun 2019 11:00:33 +0800
Tao Xu <tao3.xu@intel.com> wrote:

> On 6/5/2019 8:12 PM, Igor Mammedov wrote:
> > On Wed, 5 Jun 2019 14:04:10 +0800
> > Tao Xu <tao3.xu@intel.com> wrote:
> >   
> >> On 6/4/2019 11:04 PM, Igor Mammedov wrote:  
> >>> On Wed,  8 May 2019 14:17:22 +0800
> >>> Tao Xu <tao3.xu@intel.com> wrote:
> >>>      
> ...
> >>>> +
> >>>> +                /* SMBIOS Handles */
> >>>> +                /* TBD: set smbios handles */
> >>>> +                build_append_int_noprefix(table_data, 0, 2 * n);  
> >>> Is memory side cache structure useful at all without pointing to SMBIOS entries?
> >>>      
> >> They are not useful yet, and the kernel 5.1 HMAT sysfs doesn't show
> >> SMBIOS entries. We can update it if it useful in the future.  
> > 
> > In that case I'd suggest to drop it for now until this table is properly
> > populated and ready for consumption. (i.e. drop this patch and corresponding
> > CLI 9/11 patch).
> >   
> 
> But the kernel HMAT can read othe Memory Side Cache Information except 
> SMBIOS entries and the host HMAT tables also haven’t SMBIOS Handles it 
> also shows Number of SMBIOS handles (n) as 0. So I am wondering if it is 
> better to setting "SMBIOS handles (n)" as 0, remove TODO and comment the 
> reason why set it 0?

My understanding is that SMBIOS handles are used to associate side cache
descriptions with RAM pointed by SMBIOS handles, so that OS would be
able to figure out what RAM modules are cached by what cache.
Hence I suspect that side cache table is useless in the best case without
valid references to SMBIOS handles.
(I might be totally mistaken but the matter requires clarification before
we commit to it)