Re: [RFC PATCH v2 00/24] KVM: combined patchset for MBEC/GMET support

[Jon Kohler]

> On Apr 7, 2026, at 7:18 AM, David Riley <d.riley@xxxxxxxxxxx> wrote:
> 
> On 4/7/26 1:53 PM, David Riley wrote:
>> [...]
>> Results
>> 
>> I compared the VM-Exit volume by running the workload with the feature
>> flags (+vmx-mbec / +gmet) enabled vs. disabled. The table below shows
>> the average number of samples (Total count of VM-Exit Events) over
>> the 3 runs for each mode (idle/load).
>> 
>> Platform  Mode  #Avg Exits Disabled  #Avg Exits Enabled  Reduction
>> Intel          Idle          631722                   189460             70.01%
>> Intel          Load   10020824                    775194     92.26%
>> AMD         Idle          249200                    118042         52.63%
>> AMD         Load        425217                    306945     27.81%
>> 
> sorry for the broken formatting. The following should be better.
> 
> Platform  Mode  #Avg Exits Disabled  #Avg Exits Enabled  Reduction
> 
> Intel     Idle    631722             189460               70.01%
> Intel     Load  10020824             775194               92.26%
> AMD       Idle    249200             118042               52.63%
> AMD       Load    425217             306945               27.81%
> 
>> Observations
>> 
>> Intel MBEC:
>> The reduction in VM-exits is massive (over 92% under load).
>> 
>> AMD GMET:
>> There is a measurable benefit (~28% reduction), though it appears less
>> significant than the Intel results. However, it is worth noting that the
>> AMD platform maintains a much lower baseline of VM-exits even with
>> the feature disabled, suggesting more efficient native handling of
>> these transitions compared to the Intel baseline in this specific test
>> environment.

Indeed, I’ve seen similar things in our testing, wherein HVCI without GMET
isn't a hot mess like it is without MBEC, but it definitely is still
noticeable. That finding is in-line with our understanding too. Good to
see that you found similar results.

>> 
>> On both architectures, I did not notice any stability issues with the
>> patches applied.
>> 
>> Let me know if there are more specific tests you're interested in.

Only other thing that would be worth doing whilst stress testing is
live migration back n forth to make sure things don’t blow up. I haven’t
seen anything blow up though, but worth a double check if you’ve got the
extra hardware to do so.

Also it would be worth while to make sure KVM unit tests pass with 
our patch series. They did for me, but its always good to get more eyes.

>> Tested-by: David Riley <d.riley@xxxxxxxxxxx>

Thanks for giving this a run on both sets of hardware, this is really
valuable.