1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
// Core scheduler round execution and coroutine stepping.
impl ProtocolMachine {
/// Execute one scheduler round: advance at most one ready coroutine.
///
/// # Errors
///
/// Returns a `ProtocolMachineError` if a coroutine faults.
#[allow(clippy::too_many_lines)]
pub(crate) fn kernel_step_round(
&mut self,
handler: &dyn EffectHandler,
n: usize,
) -> Result<StepResult, ProtocolMachineError> {
#[cfg(debug_assertions)]
debug_assert!(self.wf_vm_state().is_ok());
if n == 0 {
return Err(ProtocolMachineError::InvalidConcurrency { n });
}
self.last_sched_step = None;
self.last_pre_dispatch_state = None;
self.clock.advance();
if self.all_done() {
self.last_pre_dispatch_state = self.refinement_slice().ok();
return Ok(StepResult::AllDone);
}
// Event ordering contract: topology effects ingress first each round,
// before unblocking and scheduler selection.
self.ingest_topology_events(handler)?;
self.prune_expired_timeouts();
self.try_unblock_senders();
self.try_unblock_receivers();
self.evaluate_progress_contracts()?;
self.ensure_ready_eligibility();
self.last_pre_dispatch_state = self.refinement_slice().ok();
#[cfg(debug_assertions)]
self.debug_assert_ready_eligibility_consistent();
if !self.sched.has_eligible_ready() {
return Ok(StepResult::Stuck);
}
let coroutines = &self.coroutines;
let coro_slots = &self.coro_slots;
let Some(coro_id) = self.sched.pick_eligible_runnable(|id| {
coro_slots
.get(&id)
.and_then(|idx| coroutines.get(*idx))
.or_else(|| coroutines.get(id).filter(|coro| coro.id == id))
.or_else(|| coroutines.iter().find(|coro| coro.id == id))
.is_some_and(|coro| !coro.progress_tokens.is_empty())
}) else {
return Ok(StepResult::Stuck);
};
#[cfg(debug_assertions)]
self.eligible_ready.remove(&coro_id);
let result = self.exec_instr(coro_id, handler);
match result {
Ok(ExecOutcome::Continue) => {
self.last_sched_step = Some(SchedStepDebug {
selected_coro: coro_id,
exec_status: SchedExecStatus::Continue,
});
self.sched.reschedule(coro_id);
self.sync_ready_eligibility_for(coro_id);
}
Ok(ExecOutcome::Blocked(reason)) => {
let yielded = matches!(reason, BlockReason::Spawn);
self.last_sched_step = Some(SchedStepDebug {
selected_coro: coro_id,
exec_status: if yielded {
SchedExecStatus::Yielded
} else {
SchedExecStatus::Blocked
},
});
if yielded {
self.sched.reschedule(coro_id);
self.sync_ready_eligibility_for(coro_id);
} else {
self.sched.mark_blocked(coro_id, reason);
#[cfg(debug_assertions)]
self.eligible_ready.remove(&coro_id);
}
}
Ok(ExecOutcome::Halted) => {
self.last_sched_step = Some(SchedStepDebug {
selected_coro: coro_id,
exec_status: SchedExecStatus::Halted,
});
self.sched.mark_done(coro_id);
#[cfg(debug_assertions)]
self.eligible_ready.remove(&coro_id);
self.obs_trace.push(
ObsEvent::Halted {
tick: self.clock.tick,
coro_id,
},
&self.config.observability_retention,
);
}
Err(fault) => {
self.last_sched_step = Some(SchedStepDebug {
selected_coro: coro_id,
exec_status: SchedExecStatus::Faulted,
});
let Some(idx) = self.coro_index(coro_id) else {
return Err(ProtocolMachineError::Fault { coro_id, fault });
};
let session = self.coroutines[idx].session_id;
self.obs_trace.push(
ObsEvent::FailureBranchEntered {
tick: self.clock.tick,
session,
coro_id,
fault: fault.clone(),
},
&self.config.observability_retention,
);
self.obs_trace.push(
ObsEvent::Faulted {
tick: self.clock.tick,
coro_id,
fault: fault.clone(),
},
&self.config.observability_retention,
);
self.coroutines[idx].status = CoroStatus::Faulted(fault.clone());
self.sched.mark_done(coro_id);
#[cfg(debug_assertions)]
self.eligible_ready.remove(&coro_id);
return Err(ProtocolMachineError::Fault { coro_id, fault });
}
}
self.evaluate_progress_contracts()?;
if self.all_done() {
#[cfg(debug_assertions)]
self.debug_assert_ready_eligibility_consistent();
#[cfg(debug_assertions)]
debug_assert!(self.wf_vm_state().is_ok());
Ok(StepResult::AllDone)
} else {
#[cfg(debug_assertions)]
self.debug_assert_ready_eligibility_consistent();
#[cfg(debug_assertions)]
debug_assert!(self.wf_vm_state().is_ok());
Ok(StepResult::Continue)
}
}
/// Execute one scheduler step: pick a coroutine, run one instruction.
///
/// # Errors
///
/// Returns a `ProtocolMachineError` if a coroutine faults.
pub fn step(&mut self, handler: &dyn EffectHandler) -> Result<StepResult, ProtocolMachineError> {
self.step_round(handler, 1)
}
/// Execute one scheduler round through the canonical kernel API.
///
/// # Errors
///
/// Returns a `ProtocolMachineError` if a coroutine faults.
pub fn step_round(
&mut self,
handler: &dyn EffectHandler,
n: usize,
) -> Result<StepResult, ProtocolMachineError> {
ProtocolMachineKernel::step_round(self, handler, n)
}
/// Run the ProtocolMachine until all coroutines complete or an error occurs, with concurrency N.
///
/// `max_rounds` prevents infinite loops.
///
/// # Errors
///
/// Returns a `ProtocolMachineError` if any coroutine faults.
pub fn run_concurrent(
&mut self,
handler: &dyn EffectHandler,
max_rounds: usize,
concurrency: usize,
) -> Result<RunStatus, ProtocolMachineError> {
ProtocolMachineKernel::run_concurrent(self, handler, max_rounds, concurrency)
}
/// Run the ProtocolMachine until all coroutines complete or an error occurs.
///
/// `max_steps` prevents infinite loops.
///
/// # Errors
///
/// Returns a `ProtocolMachineError` if any coroutine faults.
pub fn run(
&mut self,
handler: &dyn EffectHandler,
max_steps: usize,
) -> Result<RunStatus, ProtocolMachineError> {
ProtocolMachineKernel::run(self, handler, max_steps)
}
/// Run with replayed effect outcomes captured from a prior execution.
///
/// The `fallback` handler is only used for optional hooks not encoded in
/// replay entries.
///
/// # Errors
///
/// Returns a `ProtocolMachineError` if replay data is exhausted/mismatched or a coroutine faults.
pub fn run_replay(
&mut self,
fallback: &dyn EffectHandler,
replay_trace: &[EffectTraceEntry],
max_steps: usize,
) -> Result<RunStatus, ProtocolMachineError> {
self.run_replay_shared(
fallback,
Arc::<[EffectTraceEntry]>::from(replay_trace),
max_steps,
)
}
/// Run with replayed effect outcomes using shared trace storage.
///
/// Accepts an `Arc`-backed trace to avoid cloning when callers already hold
/// shared replay buffers.
///
/// # Errors
///
/// Returns a `ProtocolMachineError` if replay data is exhausted/mismatched or a coroutine faults.
pub fn run_replay_shared(
&mut self,
fallback: &dyn EffectHandler,
replay_trace: Arc<[EffectTraceEntry]>,
max_steps: usize,
) -> Result<RunStatus, ProtocolMachineError> {
let replay = ReplayEffectHandler::with_fallback(replay_trace, fallback);
self.run(&replay, max_steps)
}
/// Run concurrently with replayed effect outcomes.
///
/// # Errors
///
/// Returns a `ProtocolMachineError` if replay data is exhausted/mismatched or a coroutine faults.
pub fn run_concurrent_replay(
&mut self,
fallback: &dyn EffectHandler,
replay_trace: &[EffectTraceEntry],
max_rounds: usize,
concurrency: usize,
) -> Result<RunStatus, ProtocolMachineError> {
self.run_concurrent_replay_shared(
fallback,
Arc::<[EffectTraceEntry]>::from(replay_trace),
max_rounds,
concurrency,
)
}
/// Run concurrently with replayed outcomes using shared trace storage.
///
/// # Errors
///
/// Returns a `ProtocolMachineError` if replay data is exhausted/mismatched or a coroutine faults.
pub fn run_concurrent_replay_shared(
&mut self,
fallback: &dyn EffectHandler,
replay_trace: Arc<[EffectTraceEntry]>,
max_rounds: usize,
concurrency: usize,
) -> Result<RunStatus, ProtocolMachineError> {
let replay = ReplayEffectHandler::with_fallback(replay_trace, fallback);
self.run_concurrent(&replay, max_rounds, concurrency)
}
/// Get the observable trace.
#[must_use]
pub fn trace(&self) -> &[ObsEvent] {
self.obs_trace.as_slice()
}
/// Reap closed sessions once all associated coroutines are terminal.
pub fn reap_closed_sessions(&mut self) -> Vec<ClosedSessionSummary> {
let eligible: Vec<SessionId> = self
.sessions
.closed_session_ids()
.into_iter()
.filter(|sid| {
self.coroutines
.iter()
.filter(|coro| coro.session_id == *sid)
.all(Coroutine::is_terminal)
})
.collect();
if eligible.is_empty() {
return Vec::new();
}
for sid in &eligible {
self.monitor.remove_kind(*sid);
self.resource_states.remove(sid);
self.communication_consumption.prune_session(*sid);
}
self.coroutines.retain(|coro| {
!(eligible.contains(&coro.session_id) && coro.is_terminal())
});
self.rebuild_coroutine_indexes();
self.sessions.reap_sessions(&eligible)
}
/// Lean-aligned observable trace accessor.
#[must_use]
pub fn obs_trace(&self) -> &[ObsEvent] {
self.obs_trace.as_slice()
}
/// Number of interned role symbols.
#[must_use]
pub fn role_symbol_count(&self) -> usize {
self.role_symbols.len()
}
/// Number of interned label symbols.
#[must_use]
pub fn label_symbol_count(&self) -> usize {
self.label_symbols.len()
}
/// Number of interned handler symbols.
#[must_use]
pub fn handler_symbol_count(&self) -> usize {
self.handler_symbols.len()
}
/// Number of interned edge symbols.
#[must_use]
pub fn edge_symbol_count(&self) -> usize {
self.edge_symbols.len()
}
/// Access ProtocolMachine configuration.
#[must_use]
pub fn config(&self) -> &ProtocolMachineConfig {
&self.config
}
/// Last scheduler-dispatched step metadata, if any coroutine ran.
#[must_use]
pub fn last_sched_step(&self) -> Option<&SchedStepDebug> {
self.last_sched_step.as_ref()
}
/// Scheduler-dispatched step counter.
#[must_use]
pub fn scheduler_step_count(&self) -> usize {
self.sched.step_count()
}
/// Number of coroutine records in the ProtocolMachine.
#[must_use]
pub fn coroutine_count(&self) -> usize {
self.coroutines.len()
}
/// Next session identifier reserved for allocation.
#[must_use]
pub fn next_session_id(&self) -> SessionId {
self.sessions.next_session_id()
}
/// Number of active sessions in the ProtocolMachine.
#[must_use]
pub fn session_count(&self) -> usize {
self.sessions.active_count()
}
/// Number of sessions still resident in the ProtocolMachine, including closed ones.
#[must_use]
pub fn live_session_count(&self) -> usize {
self.sessions.live_count()
}
}