quota 0.3.3

Fastest Lane-parallel Rate-limiter for Rust
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
200
201
202
203
204

use crate::{CurrentQuota, QuotaPolicy, QuotaResult, QuotaTooLow};
use std::sync::atomic::{AtomicU64, Ordering};

#[inline]
pub(super) fn consume_tat(
    tat_ns: &AtomicU64,
    effective_now_ns: u64,
    cost: u64,
    ns_per_token: f64,
    token_ns: u64,
    burst_tokens: f64,
    burst_ns: u64,
    burst_after_one_ns: u64,
) -> QuotaResult {
    let mut observed_tat_ns = tat_ns.load(Ordering::Relaxed);

    loop {
        let current_tat_ns = observed_tat_ns;

        if cost == 0 {
            return Ok(CurrentQuota::gcra(
                current_tat_ns,
                effective_now_ns,
                burst_ns,
                ns_per_token,
                burst_tokens,
            ));
        }

        let cost_ns = if cost == 1 {
            token_ns
        } else {
            duration_ns(cost as f64, ns_per_token)
        };
        let allowed_tat_ns = if cost == 1 {
            effective_now_ns.saturating_add(burst_after_one_ns)
        } else if cost_ns > burst_ns {
            return Err(QuotaTooLow::new(CurrentQuota::gcra(
                current_tat_ns,
                effective_now_ns,
                burst_ns,
                ns_per_token,
                burst_tokens,
            )));
        } else {
            effective_now_ns.saturating_add(burst_ns - cost_ns)
        };
        if current_tat_ns > allowed_tat_ns {
            return Err(QuotaTooLow::new(CurrentQuota::gcra(
                current_tat_ns,
                effective_now_ns,
                burst_ns,
                ns_per_token,
                burst_tokens,
            )));
        }

        let new_tat_ns = current_tat_ns.max(effective_now_ns).saturating_add(cost_ns);

        match tat_ns.compare_exchange_weak(
            current_tat_ns,
            new_tat_ns,
            Ordering::Relaxed,
            Ordering::Relaxed,
        ) {
            Ok(_) => {
                return Ok(CurrentQuota::gcra(
                    new_tat_ns,
                    effective_now_ns,
                    burst_ns,
                    ns_per_token,
                    burst_tokens,
                ));
            }
            Err(actual) => observed_tat_ns = actual,
        }
    }
}

#[inline]
pub(super) fn check_tat(
    tat_ns: &AtomicU64,
    effective_now_ns: u64,
    cost: u64,
    ns_per_token: f64,
    token_ns: u64,
    burst_ns: u64,
    burst_after_one_ns: u64,
) -> bool {
    if cost == 0 {
        return true;
    }

    let cost_ns = if cost == 1 {
        token_ns
    } else {
        duration_ns(cost as f64, ns_per_token)
    };
    let allowed_tat_ns = if cost == 1 {
        effective_now_ns.saturating_add(burst_after_one_ns)
    } else if cost_ns > burst_ns {
        return false;
    } else {
        effective_now_ns.saturating_add(burst_ns - cost_ns)
    };
    let mut observed_tat_ns = tat_ns.load(Ordering::Relaxed);

    loop {
        let current_tat_ns = observed_tat_ns;

        if current_tat_ns > allowed_tat_ns {
            return false;
        }

        let new_tat_ns = current_tat_ns.max(effective_now_ns).saturating_add(cost_ns);

        match tat_ns.compare_exchange_weak(
            current_tat_ns,
            new_tat_ns,
            Ordering::Relaxed,
            Ordering::Relaxed,
        ) {
            Ok(_) => return true,
            Err(actual) => observed_tat_ns = actual,
        }
    }
}

#[inline]
pub(super) fn check_one_tat(
    tat_ns: &AtomicU64,
    effective_now_ns: u64,
    token_ns: u64,
    burst_after_one_ns: u64,
) -> bool {
    let allowed_tat_ns = effective_now_ns.saturating_add(burst_after_one_ns);
    let mut observed_tat_ns = tat_ns.load(Ordering::Relaxed);

    loop {
        let current_tat_ns = observed_tat_ns;

        if current_tat_ns > allowed_tat_ns {
            return false;
        }

        let new_tat_ns = current_tat_ns
            .max(effective_now_ns)
            .saturating_add(token_ns);

        match tat_ns.compare_exchange_weak(
            current_tat_ns,
            new_tat_ns,
            Ordering::Relaxed,
            Ordering::Relaxed,
        ) {
            Ok(_) => return true,
            Err(actual) => observed_tat_ns = actual,
        }
    }
}

#[inline]
pub(super) fn burst_tokens(
    policy: &QuotaPolicy,
    initial_available_tokens: u64,
    rate_ns: f64,
) -> f64 {
    if rate_ns <= 0.0 {
        return initial_available_tokens as f64;
    }

    policy
        .capacity()
        .max(initial_available_tokens as f64)
        .max(0.0)
}

#[inline]
pub(super) fn ns_per_token(rate_ns: f64) -> f64 {
    if rate_ns.is_finite() && rate_ns > 0.0 {
        (1.0 / rate_ns).max(1.0)
    } else {
        1.0
    }
}

#[inline]
pub(super) fn duration_ns(tokens: f64, ns_per_token: f64) -> u64 {
    if tokens <= 0.0 {
        return 0;
    }

    (tokens * ns_per_token).ceil().min((u64::MAX - 1) as f64) as u64
}

#[inline]
pub(super) fn initial_tat_ns(
    burst_tokens: f64,
    initial_available_tokens: u64,
    ns_per_token: f64,
) -> u64 {
    let initial_tokens = (initial_available_tokens as f64).clamp(0.0, burst_tokens);
    duration_ns(burst_tokens - initial_tokens, ns_per_token)
}