Re: [PATCH 1/2] x86/resctrl, Documentation: Keep mbm_assign_mode "default" on boot

[Reinette Chatre]

Hi Boris,

On 7/30/26 5:34 PM, Borislav Petkov wrote:
> On Thu, Jul 30, 2026 at 08:40:29PM +0000, Luck, Tony wrote:
>>> In either case, creating more monitoring groups than available counters can
>>> lead to innacuracies during measurement.
>>
>> Is that magic "64" enumerated anywhere? Or are users expected to just know it somehow (or find out
>> the hard way when they get odd results when using more than 64 RMIDs).
> 
> How is this done on Intel?
> 
> How does the resctrl code handle counter exhaustion?

The underlying implementations appear to be different between AMD and Intel.
On Intel side only the RMIDs (not separate hardware counters) need to be considered.
resctrl supports this RMID limit by placing a limit on how many monitoring groups
the user can create.

> 
> Do tools get some -E<error> thing in that last command status?
> 

The Intel/RDT spec does leave room for this to occur but a return of "Unavailable"
when reading a monitoring event is not encountered on current Intel hardware. The user
can always expect a value when reading a monitoring event (on current hardware).

Reinette