pub fn populate_run_ext_all(stats: &mut ScenarioStats, samples: &SampleSeries)Expand description
Run the FULL run-level ext_metrics population sequence into stats — the
single source of truth shared by the eval layer (evaluate_vm_result) and
crate::vmm::VmResult::run_metric, so the two produce byte-identical
run-level ext maps for the same run. Reads samples + stats.phases +
stats.cgroups, writes stats.ext_metrics, in the canonical order:
derive_phase_metrics— derive per-cgroup and pooled phase scalars, including the CPU-denominatediteration_ratecomponents.populate_run_ext_metrics— theread_sample-wired registry family (over every freeze insamples), then the whole-run wall + IRQ rates.populate_run_ext_metrics_from_phases— the phase-only ext metrics whoseread_sampleisNone(avg_imbalance_ratio,iteration_rate,system_time_ns,user_time_ns, the per-CPU IRQ spatial maxes, and the per-cgroup PSI-irq spatial maxes).populate_run_pooled_iteration_rate— add any whole-run cgroup carriers to the canonicaliteration_ratepool.populate_run_pooled_taobench— the whole-run taobench throughput + hit Rates pooled cross-cgroup (fromstats.cgroups[].taobench_whole).populate_run_pooled_taobench_distribution— the taobench whole-run open-loop serve-latency*_wholepercentiles (union of the per-phase per-cgroup servePlatStatshistograms, percentile re-derived over the union).populate_run_pooled_schbench— the schbench whole-run loop Counter + role-separate run-delay gate-Rates (fromstats.phases[].per_cgroup[].schbenchraw pairs, summed over phases+cgroups).populate_run_pooled_schbench_distribution— the schbench whole-run latency/rps*_wholepercentiles (union of the per-phase per-cgroupPlatStatshistograms, percentile re-derived over the union).populate_run_distribution_metrics— theDistribution/WorstLowest/WakeLatencyTailRatio/WorstCrossNodeRatiore-pools (fromstats.phases[].per_cgroupraw samples +stats.cgroups).
ORDER IS LOAD-BEARING: step 1 must precede step 3 so the canonical
CPU-denominated phase rate exists before the run fold. Step 2 must also
precede step 3 so whole-run IRQ-counter deltas land before step 3’s
contains_key skip; steps 4-9 fold the per-cgroup roll-up after the
per-phase families.
stats.phases may be pre- or post-derived. Step 1 is idempotent and
deliberately owns this derivation so callers cannot accidentally omit the
guest CPU-time components required by iteration_rate.