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

[Luck, Tony]

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

For the L3 event counters all the RMIDs are backed by counters, so no issues.

For AET the IP is not on the core, so available RMIDs may not match what is
enumerated by CPUID leaf 0xF. Enumeration for AET is done with DVSEC on
the PCIe device. It includes how many RMIDs are supported.  If AET has more,
then the extra can't be used. If AET has less, then AET code forces resctrl to
limit how many MON groups can be created so that events always have a counter.

-Tony