guest_epa_quota.get("cpu-pinning-policy") == "DEDICATED"
and not epa_vcpu_set
):
+ # Pinning policy "REQUIRE" uses threads as host should support SMT architecture
+ # Pinning policy "ISOLATE" uses cores as host should not support SMT architecture
+ # Pinning policy "PREFER" uses threads in case host supports SMT architecture
numa[
"cores"
- if guest_epa_quota.get("cpu-thread-pinning-policy") != "PREFER"
+ if guest_epa_quota.get("cpu-thread-pinning-policy") == "ISOLATE"
else "threads"
] = max(vcpu_count, 1)
local_epa_vcpu_set = True