These should eventually be per-smr I suspect but for now this is helpful.
Logic looks fine. I trust you've checked any performance effects.
In advance you count on entry, and then also on the slow path, whereas in poll you count either fast or slow. Did you mean to approach those differently?
|168 ↗||(On Diff #67646)|
Should they be initialized with EARLY_COUNTER? Since it seems that UMA SMR is available after SI_SUB_VM.