freenet 0.2.117

Freenet core software
Documentation 
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

//! Fixed-cardinality diagnostics for the production broadcast scheduler.
//!
//! These are lifetime counters sampled by the existing five-minute
//! `router_snapshot` task and exported in the wide diagnostic every 30 minutes.
//! They deliberately contain no peer or contract identifiers, and recording
//! them never changes queue behavior.

use std::sync::atomic::{AtomicU64, Ordering};

use serde::{Deserialize, Serialize};

/// Cumulative production-queue measurements used to decide the scheduler fix.
#[derive(Debug, Clone, Copy, Default, PartialEq, Eq, Serialize, Deserialize)]
#[cfg_attr(test, derive(arbitrary::Arbitrary))]
pub(crate) struct BroadcastQueueEfficiencySnapshot {
    pub capacity_evictions: u64,
    pub dedup_replacements: u64,
    pub enqueues_while_pair_active: u64,
    pub large_head_blocking_incidents: u64,
    pub large_head_blocked_millis: u64,
    pub small_entry_millis_blocked: u64,
    pub queued_large_actual_small: u64,
    pub queued_small_actual_large: u64,
    pub queued_large_actual_small_bytes: u64,
    pub queued_small_actual_large_bytes: u64,
    pub scheduled_small: u64,
    pub scheduled_large: u64,
    pub scheduled_small_state_bytes: u64,
    pub scheduled_large_state_bytes: u64,
    pub active_tracking_overflow: u64,
}

/// Lock-free sink shared by the queue's enqueue, scheduling, and delivery paths.
#[derive(Debug, Default)]
pub(crate) struct BroadcastQueueEfficiencyMetrics {
    capacity_evictions: AtomicU64,
    dedup_replacements: AtomicU64,
    enqueues_while_pair_active: AtomicU64,
    large_head_blocking_incidents: AtomicU64,
    large_head_blocked_millis: AtomicU64,
    small_entry_millis_blocked: AtomicU64,
    queued_large_actual_small: AtomicU64,
    queued_small_actual_large: AtomicU64,
    queued_large_actual_small_bytes: AtomicU64,
    queued_small_actual_large_bytes: AtomicU64,
    scheduled_small: AtomicU64,
    scheduled_large: AtomicU64,
    scheduled_small_state_bytes: AtomicU64,
    scheduled_large_state_bytes: AtomicU64,
    active_tracking_overflow: AtomicU64,
}

// The simulation feature replaces the production scheduler, so its mutation
// sites are intentionally absent while snapshots remain part of the node API.
#[cfg_attr(feature = "simulation_tests", allow(dead_code))]
impl BroadcastQueueEfficiencyMetrics {
    pub(crate) fn record_capacity_eviction(&self) {
        self.capacity_evictions.fetch_add(1, Ordering::Relaxed);
    }

    pub(crate) fn record_dedup_replacement(&self) {
        self.dedup_replacements.fetch_add(1, Ordering::Relaxed);
    }

    pub(crate) fn record_enqueue_while_active(&self) {
        self.enqueues_while_pair_active
            .fetch_add(1, Ordering::Relaxed);
    }

    pub(crate) fn record_large_head_block(
        &self,
        blocked_millis: u64,
        small_entry_millis_blocked: u64,
    ) {
        self.large_head_blocking_incidents
            .fetch_add(1, Ordering::Relaxed);
        self.large_head_blocked_millis
            .fetch_add(blocked_millis, Ordering::Relaxed);
        self.small_entry_millis_blocked
            .fetch_add(small_entry_millis_blocked, Ordering::Relaxed);
    }

    pub(crate) fn record_scheduled(&self, large: bool, state_bytes: usize) {
        let state_bytes = u64::try_from(state_bytes).unwrap_or(u64::MAX);
        let (count, bytes) = if large {
            (&self.scheduled_large, &self.scheduled_large_state_bytes)
        } else {
            (&self.scheduled_small, &self.scheduled_small_state_bytes)
        };
        count.fetch_add(1, Ordering::Relaxed);
        bytes.fetch_add(state_bytes, Ordering::Relaxed);
    }

    pub(crate) fn record_queued_large_actual_small(&self, payload_bytes: usize) {
        self.queued_large_actual_small
            .fetch_add(1, Ordering::Relaxed);
        self.queued_large_actual_small_bytes.fetch_add(
            u64::try_from(payload_bytes).unwrap_or(u64::MAX),
            Ordering::Relaxed,
        );
    }

    pub(crate) fn record_queued_small_actual_large(&self, payload_bytes: usize) {
        self.queued_small_actual_large
            .fetch_add(1, Ordering::Relaxed);
        self.queued_small_actual_large_bytes.fetch_add(
            u64::try_from(payload_bytes).unwrap_or(u64::MAX),
            Ordering::Relaxed,
        );
    }

    pub(crate) fn record_active_tracking_overflow(&self) {
        self.active_tracking_overflow
            .fetch_add(1, Ordering::Relaxed);
    }

    pub(crate) fn snapshot(&self) -> BroadcastQueueEfficiencySnapshot {
        BroadcastQueueEfficiencySnapshot {
            capacity_evictions: self.capacity_evictions.load(Ordering::Relaxed),
            dedup_replacements: self.dedup_replacements.load(Ordering::Relaxed),
            enqueues_while_pair_active: self.enqueues_while_pair_active.load(Ordering::Relaxed),
            large_head_blocking_incidents: self
                .large_head_blocking_incidents
                .load(Ordering::Relaxed),
            large_head_blocked_millis: self.large_head_blocked_millis.load(Ordering::Relaxed),
            small_entry_millis_blocked: self.small_entry_millis_blocked.load(Ordering::Relaxed),
            queued_large_actual_small: self.queued_large_actual_small.load(Ordering::Relaxed),
            queued_small_actual_large: self.queued_small_actual_large.load(Ordering::Relaxed),
            queued_large_actual_small_bytes: self
                .queued_large_actual_small_bytes
                .load(Ordering::Relaxed),
            queued_small_actual_large_bytes: self
                .queued_small_actual_large_bytes
                .load(Ordering::Relaxed),
            scheduled_small: self.scheduled_small.load(Ordering::Relaxed),
            scheduled_large: self.scheduled_large.load(Ordering::Relaxed),
            scheduled_small_state_bytes: self.scheduled_small_state_bytes.load(Ordering::Relaxed),
            scheduled_large_state_bytes: self.scheduled_large_state_bytes.load(Ordering::Relaxed),
            active_tracking_overflow: self.active_tracking_overflow.load(Ordering::Relaxed),
        }
    }
}

/// Production has one queue per process. Simulation uses a separate fan-out
/// path, matching the existing `BROADCAST_STREAM_METRICS` scope.
pub(crate) static BROADCAST_QUEUE_EFFICIENCY_METRICS: BroadcastQueueEfficiencyMetrics =
    BroadcastQueueEfficiencyMetrics {
        capacity_evictions: AtomicU64::new(0),
        dedup_replacements: AtomicU64::new(0),
        enqueues_while_pair_active: AtomicU64::new(0),
        large_head_blocking_incidents: AtomicU64::new(0),
        large_head_blocked_millis: AtomicU64::new(0),
        small_entry_millis_blocked: AtomicU64::new(0),
        queued_large_actual_small: AtomicU64::new(0),
        queued_small_actual_large: AtomicU64::new(0),
        queued_large_actual_small_bytes: AtomicU64::new(0),
        queued_small_actual_large_bytes: AtomicU64::new(0),
        scheduled_small: AtomicU64::new(0),
        scheduled_large: AtomicU64::new(0),
        scheduled_small_state_bytes: AtomicU64::new(0),
        scheduled_large_state_bytes: AtomicU64::new(0),
        active_tracking_overflow: AtomicU64::new(0),
    };

#[cfg(test)]
mod tests {
    use super::BroadcastQueueEfficiencyMetrics;

    #[test]
    fn snapshot_reconciles_all_queue_harm_counters() {
        let metrics = BroadcastQueueEfficiencyMetrics::default();
        metrics.record_capacity_eviction();
        metrics.record_dedup_replacement();
        metrics.record_enqueue_while_active();
        metrics.record_large_head_block(250, 750);
        metrics.record_scheduled(false, 10);
        metrics.record_scheduled(true, 20);
        metrics.record_queued_large_actual_small(30);
        metrics.record_queued_small_actual_large(40);
        metrics.record_active_tracking_overflow();

        let snapshot = metrics.snapshot();
        assert_eq!(snapshot.capacity_evictions, 1);
        assert_eq!(snapshot.dedup_replacements, 1);
        assert_eq!(snapshot.enqueues_while_pair_active, 1);
        assert_eq!(snapshot.large_head_blocking_incidents, 1);
        assert_eq!(snapshot.large_head_blocked_millis, 250);
        assert_eq!(snapshot.small_entry_millis_blocked, 750);
        assert_eq!(snapshot.queued_large_actual_small, 1);
        assert_eq!(snapshot.queued_small_actual_large, 1);
        assert_eq!(snapshot.queued_large_actual_small_bytes, 30);
        assert_eq!(snapshot.queued_small_actual_large_bytes, 40);
        assert_eq!(snapshot.scheduled_small, 1);
        assert_eq!(snapshot.scheduled_large, 1);
        assert_eq!(snapshot.scheduled_small_state_bytes, 10);
        assert_eq!(snapshot.scheduled_large_state_bytes, 20);
        assert_eq!(snapshot.active_tracking_overflow, 1);
    }
}