feat(causal): offcpu audit bucket — the @50 deficit located (RFC 007)

GPU sweep decomposed the ~23ms/700ms @50 injection deficit: every
ledger bucket is ~zero (drop park 0, discards 0, drop yield ~0.3ms),
books balance at absorbed+forgiven = 4x injected in all 48 cells, and
the 0%-cell contamination signature is absent. The probe pins the
residual: eff 0.933-0.943, a constant 22-27µs missing per site entry
= ~4.9 slice-expiry yields/entry x ~5.6µs runqueue wait. The
"deficit" is runnable off-CPU time inside the site — wall time the
probe's ground truth counts but on-CPU attribution correctly skips
(Coz model: speeding the site's code does not shrink queue-wait).

Measure-only reclassification, no behaviour change: a yield in the
target site stashes (tsc, experiment epoch) on the slot; the next
on_resume counts the gap into OFFCPU_IN_SITE_{CYCLES,N} (would-be
delta terms, MAX_SAMPLE_CYCLES-capped) iff the epoch still matches
and the word is live — a gap straddling end()/a same-word begin()
(live in the probe's 50,50 schedule) is dropped, never a leaked
cooldown. Parks excluded: blocked time is forgiveness territory.
New offcpu column in render_ledger_audit; LedgerCounters and
ExperimentResult grow the two fields. Fidelity footer now states
the on-CPU basis (deliberate wording change to the pinned summary;
the substring pin test still holds). +2 tests (counted gap; epoch
straddle) + render assert.

src/causal.rs

@@ -48,6 +48,12 @@ mod inner {
     /// across scheduler threads (jar Q7, v1: plain Relaxed atomics).
     static EXPERIMENT: AtomicU64 = AtomicU64::new(0);
 
+    /// Monotone experiment-window counter, bumped by every `begin()`. Lets
+    /// the offcpu-gap stash (RFC 007) tell apart two windows with an
+    /// identical site+pct word — live in the attrib probe's 50,50 schedule
+    /// — so a gap straddling `end()`/`begin()` never counts a cooldown.
+    static EXPERIMENT_EPOCH: AtomicU64 = AtomicU64::new(0);
+
     /// Global virtual-delay ledger, in TSC cycles: the total delay every
     /// actor *should* have experienced since startup. Grows while a sample
     /// lands in the experiment's target site; each actor's `Slot` ledger
@@ -103,6 +109,13 @@ mod inner {
     static DISCARD_OVERMAX_N: AtomicU64 = AtomicU64::new(0);
     /// In-site samples dropped because the thread's clock was unarmed.
     static DISCARD_UNARMED_N: AtomicU64 = AtomicU64::new(0);
+    /// Would-be attribution over runnable off-CPU gaps inside the target
+    /// site (yield-descheduled -> resumed within the same window). Not a
+    /// loss: on-CPU-only attribution is the Coz model — queue-wait is not
+    /// shrunk by speeding the site's code — but counted so the audit books
+    /// close against wall in-site time (the located @50 "deficit").
+    static OFFCPU_IN_SITE_CYCLES: AtomicU64 = AtomicU64::new(0);
+    static OFFCPU_IN_SITE_N: AtomicU64 = AtomicU64::new(0);
 
     fn sites() -> &'static Mutex<Vec<&'static str>> {
         SITES.get_or_init(|| Mutex::new(Vec::new()))
@@ -387,10 +400,30 @@ mod inner {
     ///   its debt: it must pay by spinning at its next check. Forgiving on
     ///   every resume would make any yield-cadence actor delay-immune and
     ///   experiments inert (found live on a 24-core run: nothing slowed).
+    /// - If the deschedule was a *yield* in the live experiment's target
+    ///   site, count the off-CPU gap it opened into the offcpu audit bucket
+    ///   (RFC 007: the located @50 deficit — runnable queue-wait is wall
+    ///   time in-site that on-CPU attribution correctly skips). Same-window
+    ///   only, enforced by the experiment epoch; measure-only.
     /// - Arm this thread's sample clock so the first interval of the resume
     ///   excludes scheduler time.
     #[inline]
     pub(crate) fn on_resume(slot: &crate::runtime::Slot) {
+        let (desched_tsc, desched_epoch) = slot.take_causal_desched();
+        if desched_tsc != 0 && desched_epoch == EXPERIMENT_EPOCH.load(Ordering::Relaxed) {
+            // Same epoch ⇒ no `begin()` since the stash; a nonzero word ⇒
+            // no `end()` either — the gap closed inside its own window.
+            let exp = EXPERIMENT.load(Ordering::Relaxed);
+            if exp != 0 {
+                let pct = exp & 0xffff_ffff;
+                let gap = preempt::rdtsc()
+                    .saturating_sub(desched_tsc)
+                    .min(MAX_SAMPLE_CYCLES);
+                OFFCPU_IN_SITE_CYCLES
+                    .fetch_add(gap.saturating_mul(pct) / 100, Ordering::Relaxed);
+                OFFCPU_IN_SITE_N.fetch_add(1, Ordering::Relaxed);
+            }
+        }
         if slot.take_causal_parked() {
             let global = GLOBAL_DELAY.load(Ordering::Relaxed);
             let mine = slot.causal_delay();
@@ -417,6 +450,9 @@ mod inner {
     /// Slice-expiry yields sample at the same checkpoint that deschedules
     /// them, so their tails are ~zero by construction; a fat yield bucket
     /// therefore points at explicit `yield_now` calls or requeued parks.
+    ///
+    /// Yields additionally stash the deschedule instant on the slot so
+    /// `on_resume` can count the runnable off-CPU gap (offcpu bucket).
     pub(crate) fn on_deschedule(slot: &crate::runtime::Slot, real_park: bool) {
         let exp = EXPERIMENT.load(Ordering::Relaxed);
         if exp == 0 {
@@ -427,11 +463,19 @@ mod inner {
         if pct == 0 || slot.causal_site() != target {
             return;
         }
+        let now = preempt::rdtsc();
+        if !real_park {
+            // Runnable gap opens here; `on_resume` closes and counts it
+            // (offcpu bucket). Parks are excluded: blocked time is already
+            // represented by forgiveness, and blocked wall time is not
+            // queue-wait.
+            slot.set_causal_desched(now, EXPERIMENT_EPOCH.load(Ordering::Relaxed));
+        }
         let last = LAST_SAMPLE_TSC.with(|c| c.get());
         if last == 0 {
             return;
         }
-        let interval = preempt::rdtsc().saturating_sub(last).min(MAX_SAMPLE_CYCLES);
+        let interval = now.saturating_sub(last).min(MAX_SAMPLE_CYCLES);
         let would_be = interval.saturating_mul(pct) / 100;
         if real_park {
             DROP_PARK_N.fetch_add(1, Ordering::Relaxed);
@@ -447,6 +491,7 @@ mod inner {
     // -----------------------------------------------------------------------
 
     fn begin(site: u32, pct: u32) {
+        EXPERIMENT_EPOCH.fetch_add(1, Ordering::Relaxed);
         EXPERIMENT.store(((site as u64) << 32) | pct as u64, Ordering::Relaxed);
     }
 
@@ -474,6 +519,8 @@ mod inner {
         pub discard_overmax_cycles: u64,
         pub discard_overmax_n: u64,
         pub discard_unarmed_n: u64,
+        pub offcpu_in_site_cycles: u64,
+        pub offcpu_in_site_n: u64,
     }
 
     impl LedgerCounters {
@@ -497,6 +544,10 @@ mod inner {
                     .saturating_sub(before.discard_overmax_cycles),
                 discard_overmax_n: self.discard_overmax_n.saturating_sub(before.discard_overmax_n),
                 discard_unarmed_n: self.discard_unarmed_n.saturating_sub(before.discard_unarmed_n),
+                offcpu_in_site_cycles: self
+                    .offcpu_in_site_cycles
+                    .saturating_sub(before.offcpu_in_site_cycles),
+                offcpu_in_site_n: self.offcpu_in_site_n.saturating_sub(before.offcpu_in_site_n),
             }
         }
     }
@@ -513,6 +564,8 @@ mod inner {
             discard_overmax_cycles: DISCARD_OVERMAX_CYCLES.load(Ordering::Relaxed),
             discard_overmax_n: DISCARD_OVERMAX_N.load(Ordering::Relaxed),
             discard_unarmed_n: DISCARD_UNARMED_N.load(Ordering::Relaxed),
+            offcpu_in_site_cycles: OFFCPU_IN_SITE_CYCLES.load(Ordering::Relaxed),
+            offcpu_in_site_n: OFFCPU_IN_SITE_N.load(Ordering::Relaxed),
         }
     }
 
@@ -592,6 +645,8 @@ mod inner {
         pub discard_overmax_cycles: u64,
         pub discard_overmax_n: u64,
         pub discard_unarmed_n: u64,
+        pub offcpu_in_site_cycles: u64,
+        pub offcpu_in_site_n: u64,
     }
 
     /// Run the plan synchronously on the calling (OS) thread: for every
@@ -671,6 +726,8 @@ mod inner {
                     discard_overmax_cycles: audit.discard_overmax_cycles,
                     discard_overmax_n: audit.discard_overmax_n,
                     discard_unarmed_n: audit.discard_unarmed_n,
+                    offcpu_in_site_cycles: audit.offcpu_in_site_cycles,
+                    offcpu_in_site_n: audit.offcpu_in_site_n,
                 });
                 controller_sleep(plan.cooldown);
             }
@@ -753,9 +810,10 @@ mod inner {
     /// site buys you nothing — the RFC's headline answer.
     ///
     /// Ends with a one-line fidelity note (RFC 007 Validation): reported
-    /// impacts are conservative — the controller undershoots ideal injection
-    /// at high speedup pcts, so gains are lower bounds; site *rankings* are
-    /// unaffected.
+    /// impacts are conservative — attribution counts on-CPU site time only,
+    /// so runnable queue-wait inside the site (the located @50 "deficit",
+    /// eff ≈ 0.93 live) is never injected and gains are lower bounds; site
+    /// *rankings* are unaffected.
     pub fn render_summary(results: &[ExperimentResult]) -> String {
         use std::fmt::Write;
         let mut s = String::new();
@@ -789,7 +847,7 @@ mod inner {
         if !results.is_empty() {
             let _ = writeln!(
                 s,
-                "note: impacts are lower bounds — undershoot grows with speedup pct; rankings unaffected"
+                "note: impacts are lower bounds — site time counts on-CPU only (runnable queue-wait is not attributed); rankings unaffected"
             );
         }
         s
@@ -812,7 +870,8 @@ mod inner {
             let _ = writeln!(
                 s,
                 "site {:<22} @{:>2}%  injected {:>7.1}ms  absorbed {:>7.1}ms  forgiven {:>7.1}ms  \
-                 drop park {:>6.2}ms/{:<4} yield {:>6.2}ms/{:<4} discard >max {:>6.2}ms/{:<3} unarmed {}",
+                 drop park {:>6.2}ms/{:<4} yield {:>6.2}ms/{:<4} offcpu {:>6.2}ms/{:<5} \
+                 discard >max {:>6.2}ms/{:<3} unarmed {}",
                 r.site,
                 r.speedup_pct,
                 ms(r.injected_cycles),
@@ -822,6 +881,8 @@ mod inner {
                 r.drop_park_n,
                 ms(r.drop_yield_cycles),
                 r.drop_yield_n,
+                ms(r.offcpu_in_site_cycles),
+                r.offcpu_in_site_n,
                 ms(r.discard_overmax_cycles),
                 r.discard_overmax_n,
                 r.discard_unarmed_n

src/runtime.rs

@@ -467,6 +467,18 @@ pub(crate) struct Slot {
     /// history. Written by the owning scheduler thread at the deschedule /
     /// resume boundary only — same single-writer discipline as `causal_delay`.
     causal_parked: AtomicBool,
+    /// RFC 007 — TSC at this actor's last *runnable* (yield) deschedule
+    /// while it sat in the live experiment's target site; 0 = no gap
+    /// pending. Paired with `causal_desched_epoch`; written on the
+    /// deschedule path, consumed at the next resume — same single-writer
+    /// discipline as `causal_delay`.
+    causal_desched_tsc: AtomicU64,
+    /// RFC 007 — experiment epoch live at that deschedule (see
+    /// `causal::EXPERIMENT_EPOCH`): the resume counts the gap only into
+    /// the same window, so one straddling `end()` — or a later `begin()`
+    /// with an identical site+pct word — is dropped instead of leaking a
+    /// cooldown across windows.
+    causal_desched_epoch: AtomicU64,
     /// Cold lifecycle data. See [`SlotCold`].
     pub(crate) cold: RawMutex<SlotCold>,
 }
@@ -484,6 +496,8 @@ impl Slot {
             causal_site: AtomicU32::new(0),
             causal_delay: AtomicU64::new(0),
             causal_parked: AtomicBool::new(true),
+            causal_desched_tsc: AtomicU64::new(0),
+            causal_desched_epoch: AtomicU64::new(0),
             cold: RawMutex::new(SlotCold {
                 actor: None,
                 waiters: Vec::new(),
@@ -577,6 +591,8 @@ impl Slot {
         // True, not false: the first resume of a fresh actor is a "wake" —
         // it must not owe the process's accumulated delay history.
         self.causal_parked.store(true, Ordering::Relaxed);
+        self.causal_desched_tsc.store(0, Ordering::Relaxed);
+        self.causal_desched_epoch.store(0, Ordering::Relaxed);
     }
 
     /// RFC 007 — mark that this actor's deschedule was a genuine park.
@@ -597,6 +613,24 @@ impl Slot {
         self.causal_parked.swap(false, Ordering::Relaxed)
     }
 
+    /// RFC 007 — stash the runnable-deschedule instant and the experiment
+    /// epoch it happened under (offcpu-gap audit; see `causal::on_resume`).
+    #[cfg_attr(not(feature = "smarm-causal"), allow(dead_code))]
+    #[inline]
+    pub(crate) fn set_causal_desched(&self, tsc: u64, epoch: u64) {
+        self.causal_desched_tsc.store(tsc, Ordering::Relaxed);
+        self.causal_desched_epoch.store(epoch, Ordering::Relaxed);
+    }
+
+    /// RFC 007 — consume the stash: `(tsc, epoch)`; `(0, _)` = none pending.
+    #[cfg_attr(not(feature = "smarm-causal"), allow(dead_code))]
+    #[inline]
+    pub(crate) fn take_causal_desched(&self) -> (u64, u64) {
+        let tsc = self.causal_desched_tsc.swap(0, Ordering::Relaxed);
+        let epoch = self.causal_desched_epoch.load(Ordering::Relaxed);
+        (tsc, epoch)
+    }
+
     /// RFC 007 — current causal site id (0 = none). Single-writer: only the
     /// on-CPU actor's thread writes, via the site-guard enter/exit.
     #[cfg_attr(not(feature = "smarm-causal"), allow(dead_code))]

tests/causal.rs

@@ -902,10 +902,78 @@ fn ledger_audit_renders_per_cell() {
         discard_overmax_cycles: 5_000,
         discard_overmax_n: 1,
         discard_unarmed_n: 0,
+        offcpu_in_site_cycles: 20_000,
+        offcpu_in_site_n: 3,
     }];
     let s = render_ledger_audit(&results);
     assert!(s.contains("ledger audit"), "missing header: {s}");
     assert!(s.contains("audit-render-site"), "missing site line: {s}");
     assert!(s.contains("absorbed"), "missing absorbed column: {s}");
     assert!(s.contains("forgiven"), "missing forgiven column: {s}");
+    assert!(s.contains("offcpu"), "missing offcpu column: {s}");
+}
+
+/// A runnable (yield) deschedule inside the target site opens an off-CPU
+/// gap; the resume closing it lands in the same experiment window, so the
+/// audit counts it. RFC 007: this is the located @50 deficit — runnable
+/// queue-wait is real wall time in-site that on-CPU attribution correctly
+/// skips; the bucket exists so the books close instead of leaving a
+/// silent ~7% residual.
+#[test]
+fn audit_counts_offcpu_runnable_gap_in_target_site() {
+    let _s = serial();
+    smarm::init(smarm::Config::exact(1)).run(|| {
+        smarm::causal::begin_experiment_for_test("audit-offcpu-site", 50);
+        let c0 = smarm::causal::ledger_counters();
+        let h = smarm::spawn(|| {
+            let _g = smarm::causal_site!("audit-offcpu-site");
+            smarm::check!(); // arm
+            smarm::yield_now(); // runnable gap: deschedule -> resume
+            smarm::check!();
+        });
+        h.join().unwrap();
+        let c1 = smarm::causal::ledger_counters();
+        smarm::causal::end_experiment_for_test();
+
+        assert!(
+            c1.offcpu_in_site_n > c0.offcpu_in_site_n,
+            "in-site runnable gap recorded no offcpu event"
+        );
+        assert!(
+            c1.offcpu_in_site_cycles > c0.offcpu_in_site_cycles,
+            "in-site runnable gap carried no cycles"
+        );
+    });
+}
+
+/// An off-CPU gap that straddles `end()` — even into a later window with
+/// an identical site+pct word (live in the attrib probe's 50,50 schedule)
+/// — is dropped, not counted: the stash carries the experiment *epoch*,
+/// so a straddling resume can never leak a cooldown across windows.
+#[test]
+fn audit_offcpu_gap_across_windows_not_counted() {
+    let _s = serial();
+    smarm::init(smarm::Config::exact(1)).run(|| {
+        smarm::causal::begin_experiment_for_test("audit-offcpu-straddle", 50);
+        let c0 = smarm::causal::ledger_counters();
+        let h = smarm::spawn(|| {
+            let _g = smarm::causal_site!("audit-offcpu-straddle");
+            smarm::check!();
+            smarm::yield_now(); // stash carries window A's epoch
+            smarm::check!();
+        });
+        // FIFO on exact(1): let h run to its yield, then close window A and
+        // open a same-word window B before h resumes.
+        smarm::yield_now();
+        smarm::causal::end_experiment_for_test();
+        smarm::causal::begin_experiment_for_test("audit-offcpu-straddle", 50);
+        h.join().unwrap();
+        let c1 = smarm::causal::ledger_counters();
+        smarm::causal::end_experiment_for_test();
+
+        assert_eq!(
+            c1.offcpu_in_site_n, c0.offcpu_in_site_n,
+            "gap straddling windows must not be counted"
+        );
+    });
 }