kvarn 0.6.3

A forward-thinking fast web server designed to fit your needs, efficiently.
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

//! Limits traffic from a IP address to partially mitigate attacks.
//!
//! Kvarn's limiting is smart; when a client first makes to many requests,
//! a hard-coded `429 Too Many Requests` is sent back.
//! It the spam continues, the current connection and all future streams are blocked,
//! until the limit resets.
//!
//! The thresholds are configurable and have sensible defaults.
//!
//! After `reset_time` is elapsed, all stored request counts are cleared.
//! Longer `reset_time`s and higher `max_requests` can be less forgiving for clients,
//! but safer for the server, and vise versa.

use crate::prelude::*;
use threading::atomic;

/// Get a `429 Too Many Requests` response.
#[inline]
#[must_use]
pub fn get_too_many_requests() -> Response<Bytes> {
    let body = Bytes::from_static("<html>\
    <head>\
        <title>429 Too Many Requests</title>\
    </head>\
    <body>\
        <center>\
            <h1>429 Too Many Requests</h1>\
            <hr>\
            <p>You have requested resources from this server too many times. <i>Please Enhance Your Calm.</i></p>\
            <p>Try to access this page again in a minute. If this error persists, please contact the website administrator.</p>\
        </center>\
    </body>\
</html>".as_bytes());

    Response::builder()
        .status(StatusCode::TOO_MANY_REQUESTS)
        .header(
            "content-type",
            HeaderValue::from_static("text/html; charset=utf-8"),
        )
        .header("content-length", body.len().to_string())
        .header("content-encoding", "identity")
        .body(body)
        .unwrap()
}

/// The strength of limiting.
#[derive(Debug, Clone, Ord, PartialOrd, Eq, PartialEq)]
#[must_use]
pub enum Action {
    /// Request should continue as normal.
    Passed,
    /// Send a [`get_too_many_requests`] response.
    Send,
    /// Drop the connection immediately.
    Drop,
}

/// Data used to limit requests.
///
/// One instance of this is used per `Listener`.
/// It counts using a [`atomic::AtomicUsize`] and keeps the count of requests per
/// [`IpAddr`] in a [`Mutex`], for fast access times and cheap cloning.
#[derive(Debug, Clone)]
#[must_use]
pub struct Manager {
    // Arc so we can easily clone this
    connection_map: Arc<dashmap::DashMap<IpAddr, usize>>,
    max_requests: usize,
    check_every: usize,
    reset_seconds: f64,

    time: Arc<(atomic::AtomicU64, atomic::AtomicU32)>,
    iteration: Arc<atomic::AtomicUsize>,
}
impl LimitManager {
    /// Creates a new manager.
    ///
    /// Use [`LimitManager::default`] for sane defaults.
    ///
    /// The number of allowed requests per reset is `max_requests * check_every`.
    /// After `reset_seconds`, all data is cleared.
    ///
    /// As the math implies, increasing `max_requests` and lowering `reset_seconds`
    /// does nothing to the amount of accepted requests.
    /// Though, if you have large `reset_seconds`, it'll take longer
    /// for the limits to clear after the user has reached `max_requests`.
    pub fn new(max_requests: usize, check_every: usize, reset_seconds: f64) -> Self {
        let me = Self {
            connection_map: Arc::new(dashmap::DashMap::new()),
            max_requests,
            check_every,
            reset_seconds,

            time: Arc::new((atomic::AtomicU64::new(0), atomic::AtomicU32::new(0))),
            iteration: Arc::new(atomic::AtomicUsize::new(0)),
        };
        me.update_time();
        me
    }

    fn update_time(&self) {
        let dur = std::time::SystemTime::now()
            .duration_since(std::time::UNIX_EPOCH)
            .expect("we're before 1970!?");
        self.time.0.store(dur.as_secs(), atomic::Ordering::Relaxed);
        self.time
            .1
            .store(dur.subsec_nanos(), atomic::Ordering::Relaxed);
    }
    fn get_time(&self) -> std::time::SystemTime {
        let secs = self.time.0.load(atomic::Ordering::Relaxed);
        let nanos = self.time.1.load(atomic::Ordering::Relaxed);

        std::time::UNIX_EPOCH + Duration::new(secs, nanos)
    }

    /// Disables limiting of this manager.
    pub fn disable(&mut self) -> &mut Self {
        self.set_check_every(usize::MAX)
    }
    /// Sets the number of calls in between checking the request.
    ///
    /// This is here to not have to check every request.
    pub fn set_check_every(&mut self, check_every: usize) -> &mut Self {
        self.check_every = check_every;
        self
    }
    /// Sets the max requests in the current cycle
    /// (which resets after [`Self::set_reset_seconds`]).
    /// If the amount of requests from a IP address exceeds
    /// `max_requests` in the cycle, the request is denied.
    pub fn set_max_requests(&mut self, max_requests: usize) -> &mut Self {
        self.max_requests = max_requests;
        self
    }
    /// Sets the interval to clear all limits.
    ///
    /// See [`Self::new`] for considerations when making this value large.
    pub fn set_reset_seconds(&mut self, reset_seconds: f64) -> &mut Self {
        self.reset_seconds = reset_seconds;
        self
    }

    /// Registers a request from `addr`.
    ///
    /// This is called twice, once when a new connection is established, and once when a new request is made.
    ///
    /// Does not always lock the [`Mutex`], only once per `check_every`.
    /// It only [`atomic::AtomicUsize::fetch_add`] else, with [`atomic::Ordering::Relaxed`].
    /// This is less reliable, but faster. We do not require this to be to be exact.
    pub fn register(&self, addr: IpAddr) -> Action {
        // if self.check_every == usize::MAX, we are disabled, so just return Action::Passed.
        if self.check_every == usize::MAX
            || self.iteration.fetch_add(1, atomic::Ordering::Relaxed) + 1 < self.check_every
        {
            Action::Passed
        } else {
            self.iteration.store(0, atomic::Ordering::Release);
            if self
                .get_time()
                .elapsed()
                .unwrap_or(Duration::ZERO)
                .as_secs_f64()
                >= self.reset_seconds
            {
                self.update_time();
                self.connection_map.clear();
                Action::Passed
            } else {
                let requests = *self
                    .connection_map
                    .entry(addr)
                    .and_modify(|count| *count += 1)
                    .or_insert(1);
                if requests <= self.max_requests {
                    Action::Passed
                    // if the client goes past 3x usage, just drop connection
                } else if requests <= self.max_requests * 3 {
                    Action::Send
                } else {
                    Action::Drop
                }
            }
        }
    }
}
/// Default is `Self::new(10, 10, 10)`.
impl Default for LimitManager {
    #[inline]
    fn default() -> Self {
        Self::new(10, 10, 10.)
    }
}