dekopon-http-host 0.46.2

Statically linked bounded HTTP primitive for Dekopon broker hosts
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
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
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
use std::{
    fs::File,
    io::{self, Write as _},
    path::PathBuf,
    sync::{
        Arc,
        atomic::{AtomicU64, AtomicUsize, Ordering},
    },
};
use tempfile::TempPath;
use thiserror::Error;

/// These variants are deliberately sanitized so they can cross the provider boundary without
/// leaking paths or internal state.
#[derive(Debug, Error)]
pub enum AssetIoError {
    #[error("broker asset byte budget is exhausted")]
    OverBudget,
    #[error("broker asset I/O failed ({kind:?})")]
    Io { kind: io::ErrorKind },
    #[error("broker asset worker failed")]
    Worker,
}

impl From<io::Error> for AssetIoError {
    fn from(error: io::Error) -> Self {
        if error.kind() == io::ErrorKind::StorageFull {
            Self::OverBudget
        } else {
            Self::Io { kind: error.kind() }
        }
    }
}

#[derive(Debug)]
struct Budget {
    maximum: u64,
    used: AtomicU64,
}

#[derive(Debug, Default)]
struct Jobs {
    active: AtomicUsize,
    idle: tokio::sync::Notify,
}

struct Pending(Arc<Jobs>);
impl Drop for Pending {
    fn drop(&mut self) {
        if self.0.active.fetch_sub(1, Ordering::AcqRel) == 1 {
            self.0.idle.notify_waiters();
        }
    }
}

// A cancelled waiter's result is dropped before its completion notification, so files and budget
// are reclaimed before drain can return.
pub(crate) struct Completed<T> {
    pub(crate) value: T,
    _pending: Pending,
}

#[derive(Clone, Debug, Default)]
pub struct AssetJobs(Arc<Jobs>);

impl AssetJobs {
    pub(crate) fn spawn<T: Send + 'static>(
        &self,
        work: impl FnOnce() -> T + Send + 'static,
    ) -> tokio::task::JoinHandle<Completed<T>> {
        self.0.active.fetch_add(1, Ordering::AcqRel);
        let pending = Pending(Arc::clone(&self.0));
        let span = tracing::Span::current();
        tokio::task::spawn_blocking(move || Completed {
            value: span.in_scope(work),
            _pending: pending,
        })
    }
    pub async fn run<T: Send + 'static>(
        &self,
        work: impl FnOnce() -> Result<T, AssetIoError> + Send + 'static,
    ) -> Result<T, AssetIoError> {
        self.spawn(work)
            .await
            .map_err(|failure| {
                tracing::error!(error = %failure, "asset worker failed");
                AssetIoError::Worker
            })?
            .value
    }
    pub async fn drain(&self) {
        loop {
            let idle = self.0.idle.notified();
            tokio::pin!(idle);
            idle.as_mut().enable();
            if self.0.active.load(Ordering::Acquire) == 0 {
                return;
            }
            idle.await;
        }
    }
}

#[derive(Clone, Debug)]
pub struct AssetDirectory {
    root: PathBuf,
    budget: Arc<Budget>,
    jobs: AssetJobs,
}

impl AssetDirectory {
    pub fn new(root: PathBuf, maximum: u64) -> Self {
        Self {
            root,
            jobs: AssetJobs::default(),
            budget: Arc::new(Budget {
                maximum,
                used: AtomicU64::new(0),
            }),
        }
    }

    pub fn invocation(&self, jobs: AssetJobs) -> Self {
        Self {
            root: self.root.clone(),
            budget: Arc::clone(&self.budget),
            jobs,
        }
    }

    pub async fn drain(&self) {
        self.jobs.drain().await;
    }

    pub async fn allocate(&self) -> Result<Spool, AssetIoError> {
        let root = self.root.clone();
        let budget = Arc::clone(&self.budget);
        let jobs = self.jobs.clone();
        self.jobs
            .run(move || {
                let (file, path) = tempfile::NamedTempFile::new_in(root)?.into_parts();
                Ok(Spool {
                    sink: Sink::File { fd: file, path },
                    reservation: Reservation {
                        budget,
                        bytes: 0,
                        #[cfg(test)]
                        release_probe: None,
                    },
                    jobs,
                })
            })
            .await
    }
}

#[derive(Debug)]
struct Reservation {
    budget: Arc<Budget>,
    bytes: u64,
    #[cfg(test)]
    release_probe: Option<tests::ReleaseProbe>,
}

impl Reservation {
    fn grow(&mut self, bytes: u64) -> Result<(), AssetIoError> {
        self.budget
            .used
            .fetch_update(Ordering::Relaxed, Ordering::Relaxed, |current| {
                current
                    .checked_add(bytes)
                    .filter(|next| *next <= self.budget.maximum)
            })
            .map_err(|_current| AssetIoError::OverBudget)?;
        self.bytes += bytes;
        Ok(())
    }
}

impl Drop for Reservation {
    fn drop(&mut self) {
        #[cfg(test)]
        if let Some(probe) = &self.release_probe {
            probe.verify();
        }
        self.budget.used.fetch_sub(self.bytes, Ordering::Relaxed);
    }
}

/// Field order matters here: the file closes and unlinks before its accounting is released, since
/// fields drop in declaration order.
#[derive(Debug)]
pub struct Spool {
    sink: Sink,
    reservation: Reservation,
    jobs: AssetJobs,
}

#[derive(Debug)]
enum Sink {
    File { fd: File, path: TempPath },
}

impl Spool {
    pub fn len(&self) -> u64 {
        self.reservation.bytes
    }

    pub fn is_empty(&self) -> bool {
        self.len() == 0
    }

    pub async fn write(mut self, bytes: Vec<u8>) -> Result<Self, AssetIoError> {
        self.reservation.grow(bytes.len() as u64)?;
        let jobs = self.jobs.clone();
        jobs.run(move || {
            match &mut self.sink {
                Sink::File { fd, .. } => fd.write_all(&bytes)?,
            }
            Ok(self)
        })
        .await
    }

    pub async fn finish(self) -> Result<AssetFile, AssetIoError> {
        let jobs = self.jobs.clone();
        jobs.run(move || {
            // A failed reopen must drop the intact sink before releasing its reservation.
            let file = match &self.sink {
                Sink::File { path, .. } => File::open(path)?,
            };
            let Self {
                sink: Sink::File { fd: writer, path },
                reservation,
                jobs,
            } = self;
            drop(writer);
            // This preserves the same close-before-release ordering on the fallible unlink path as
            // well.
            let output = AssetFile {
                file,
                reservation,
                jobs,
            };
            path.close()?;
            Ok(output)
        })
        .await
    }
}

/// The file is already unlinked from disk, but its byte budget stays charged until broker ownership
/// of it ends.
#[derive(Debug)]
pub struct AssetFile {
    file: File,
    reservation: Reservation,
    jobs: AssetJobs,
}

impl AssetFile {
    /// Callers must only use positional reads on this descriptor, since it may be shared
    /// concurrently with other readers.
    pub fn file(&self) -> &File {
        &self.file
    }
    pub fn len(&self) -> u64 {
        self.reservation.bytes
    }
    pub fn is_empty(&self) -> bool {
        self.len() == 0
    }
}

#[derive(Clone, Debug)]
pub struct AssetReader {
    owner: Arc<ReaderFile>,
    jobs: AssetJobs,
}
#[derive(Debug)]
enum ReaderFile {
    Input(File),
    Output(Arc<AssetFile>),
}
impl AssetReader {
    pub fn input(file: File, jobs: AssetJobs) -> Self {
        Self {
            owner: Arc::new(ReaderFile::Input(file)),
            jobs,
        }
    }
    pub fn output(file: Arc<AssetFile>) -> Self {
        Self {
            jobs: file.jobs.clone(),
            owner: Arc::new(ReaderFile::Output(file)),
        }
    }
    pub fn output_file(&self) -> Option<Arc<AssetFile>> {
        match self.owner.as_ref() {
            ReaderFile::Input(_) => None,
            ReaderFile::Output(file) => Some(Arc::clone(file)),
        }
    }
    pub(crate) fn file(&self) -> &File {
        match self.owner.as_ref() {
            ReaderFile::Input(file) => file,
            ReaderFile::Output(file) => file.file(),
        }
    }
    pub(crate) fn jobs(&self) -> &AssetJobs {
        &self.jobs
    }
    pub async fn read<T: Send + 'static>(
        &self,
        work: impl FnOnce(&File) -> Result<T, AssetIoError> + Send + 'static,
    ) -> Result<T, AssetIoError> {
        let reader = self.clone();
        self.jobs.run(move || work(reader.file())).await
    }
}

#[cfg(test)]
mod tests {
    use super::*;
    use std::os::unix::fs::FileExt as _;

    #[derive(Debug)]
    pub(super) struct ReleaseProbe {
        descriptor: PathBuf,
        identity: (u64, u64),
        path: PathBuf,
        observed: Arc<std::sync::atomic::AtomicBool>,
    }

    impl ReleaseProbe {
        pub(super) fn verify(&self) {
            use std::os::unix::fs::MetadataExt as _;
            assert_ne!(
                std::fs::metadata(&self.descriptor)
                    .ok()
                    .map(|metadata| (metadata.dev(), metadata.ino())),
                Some(self.identity),
                "writer must close before reservation release"
            );
            assert_eq!(
                std::fs::symlink_metadata(&self.path).unwrap_err().kind(),
                io::ErrorKind::NotFound,
                "temporary path must be removed before reservation release"
            );
            self.observed.store(true, Ordering::Release);
        }
    }

    #[tokio::test]
    async fn failed_readonly_reopen_closes_and_unlinks_before_releasing_capacity() {
        use std::os::{
            fd::AsRawFd as _,
            unix::fs::{MetadataExt as _, symlink},
        };
        let root = tempfile::tempdir().unwrap();
        let directory = AssetDirectory::new(root.path().to_owned(), 4);
        let mut writer = directory
            .allocate()
            .await
            .unwrap()
            .write(b"four".to_vec())
            .await
            .unwrap();
        let Sink::File { fd, path } = &writer.sink;
        let metadata = fd.metadata().unwrap();
        let descriptor = PathBuf::from(format!("/dev/fd/{}", fd.as_raw_fd()));
        let visible = std::fs::metadata(&descriptor).unwrap();
        assert_eq!(visible.ino(), metadata.ino());
        let identity = (visible.dev(), visible.ino());
        let _inode_pin = fd.try_clone().unwrap();
        let path = path.to_path_buf();
        std::fs::remove_file(&path).unwrap();
        symlink(root.path().join("missing-reopen-target"), &path).unwrap();
        let mut bytes = [0; 4];
        fd.read_exact_at(&mut bytes, 0).unwrap();
        assert_eq!(&bytes, b"four");
        assert_eq!(directory.budget.used.load(Ordering::Relaxed), 4);
        let observed = Arc::new(std::sync::atomic::AtomicBool::new(false));
        writer.reservation.release_probe = Some(ReleaseProbe {
            descriptor,
            identity,
            path,
            observed: Arc::clone(&observed),
        });
        assert!(matches!(
            writer.finish().await,
            Err(AssetIoError::Io {
                kind: io::ErrorKind::NotFound
            })
        ));
        directory.drain().await;
        assert!(observed.load(Ordering::Acquire));
        assert_eq!(std::fs::read_dir(root.path()).unwrap().count(), 0);
        assert_eq!(directory.budget.used.load(Ordering::Relaxed), 0);
        directory
            .allocate()
            .await
            .unwrap()
            .write(vec![0; 4])
            .await
            .unwrap();
    }

    #[tokio::test]
    async fn the_exact_budget_is_accepted_and_one_byte_more_is_refused() {
        let root = tempfile::tempdir().unwrap();
        let directory = AssetDirectory::new(root.path().to_owned(), 4);
        let writer = directory.allocate().await.unwrap();
        let writer = writer.write(b"four".to_vec()).await.unwrap();
        assert!(matches!(
            writer.write(b"!".to_vec()).await,
            Err(AssetIoError::OverBudget)
        ));
        assert_eq!(directory.budget.used.load(Ordering::Relaxed), 0);
        assert_eq!(std::fs::read_dir(root.path()).unwrap().count(), 0);
    }

    #[tokio::test]
    async fn finished_outputs_are_read_only_unlinked_and_charged_until_closed() {
        let root = tempfile::tempdir().unwrap();
        let directory = AssetDirectory::new(root.path().to_owned(), 4);
        let writer = directory.allocate().await.unwrap();
        let writer = writer.write(b"four".to_vec()).await.unwrap();
        let output = writer.finish().await.unwrap();
        assert_eq!(std::fs::read_dir(root.path()).unwrap().count(), 0);
        assert_eq!(directory.budget.used.load(Ordering::Relaxed), 4);
        assert!(output.file().write_at(b"!", 0).is_err());
        let mut bytes = [0; 4];
        output.file().read_exact_at(&mut bytes, 0).unwrap();
        assert_eq!(&bytes, b"four");
        drop(output);
        assert_eq!(directory.budget.used.load(Ordering::Relaxed), 0);
    }
    #[tokio::test(flavor = "multi_thread")]
    async fn cancelled_writers_stay_charged_until_the_worker_closes_and_unlinks_them() {
        let root = tempfile::tempdir().unwrap();
        let directory = AssetDirectory::new(root.path().to_owned(), 4);
        let writer = directory
            .allocate()
            .await
            .unwrap()
            .write(b"four".to_vec())
            .await
            .unwrap();
        let jobs = directory.jobs.clone();
        let (started, running) = tokio::sync::oneshot::channel();
        let (release, wait) = std::sync::mpsc::channel();
        let task = tokio::spawn(async move {
            jobs.run(move || {
                started.send(()).unwrap();
                wait.recv().unwrap();
                Ok(writer)
            })
            .await
        });
        running.await.unwrap();
        task.abort();
        assert!(task.await.unwrap_err().is_cancelled());
        assert_eq!(directory.budget.used.load(Ordering::Relaxed), 4);
        assert_eq!(std::fs::read_dir(root.path()).unwrap().count(), 1);
        release.send(()).unwrap();
        tokio::time::timeout(std::time::Duration::from_secs(1), directory.drain())
            .await
            .unwrap();
        assert_eq!(directory.budget.used.load(Ordering::Relaxed), 0);
        assert_eq!(std::fs::read_dir(root.path()).unwrap().count(), 0);
    }

    #[test]
    fn disk_full_is_the_same_typed_fail_fast_budget_error() {
        assert!(matches!(
            AssetIoError::from(io::Error::from(io::ErrorKind::StorageFull)),
            AssetIoError::OverBudget
        ));
    }
    #[tokio::test(flavor = "multi_thread")]
    async fn cancelled_readers_keep_capacity_and_terminal_drain_waits_for_actual_close() {
        let root = tempfile::tempdir().unwrap();
        let directory = AssetDirectory::new(root.path().to_owned(), 4);
        let output = directory
            .allocate()
            .await
            .unwrap()
            .write(b"four".to_vec())
            .await
            .unwrap()
            .finish()
            .await
            .unwrap();
        let reader = AssetReader::output(Arc::new(output));
        let (started, running) = tokio::sync::oneshot::channel();
        let (release, wait) = std::sync::mpsc::channel();
        let task = tokio::spawn(async move {
            reader
                .read(move |file| {
                    started.send(()).unwrap();
                    wait.recv().unwrap();
                    let mut bytes = [0; 4];
                    file.read_exact_at(&mut bytes, 0)?;
                    assert_eq!(&bytes, b"four");
                    Ok(())
                })
                .await
        });
        running.await.unwrap();
        task.abort();
        assert!(task.await.unwrap_err().is_cancelled());
        assert_eq!(directory.budget.used.load(Ordering::Acquire), 4);
        assert_eq!(std::fs::read_dir(root.path()).unwrap().count(), 0);
        assert!(matches!(
            directory.allocate().await.unwrap().write(vec![0]).await,
            Err(AssetIoError::OverBudget)
        ));
        assert!(
            tokio::time::timeout(std::time::Duration::from_millis(25), directory.drain())
                .await
                .is_err()
        );
        release.send(()).unwrap();
        tokio::time::timeout(std::time::Duration::from_secs(1), directory.drain())
            .await
            .unwrap();
        assert_eq!(directory.budget.used.load(Ordering::Acquire), 0);
        directory
            .allocate()
            .await
            .unwrap()
            .write(vec![0; 4])
            .await
            .unwrap();
    }

    #[cfg(target_os = "linux")]
    #[tokio::test]
    async fn kernel_enospc_fails_as_over_budget_and_releases_the_writer() {
        let root = tempfile::tempdir().unwrap();
        let directory = AssetDirectory::new(root.path().to_owned(), 4);
        let mut writer = directory.allocate().await.unwrap();
        let Sink::File { fd, .. } = &mut writer.sink;
        *fd = std::fs::OpenOptions::new()
            .write(true)
            .open("/dev/full")
            .unwrap();
        assert!(matches!(
            writer.write(b"four".to_vec()).await,
            Err(AssetIoError::OverBudget)
        ));
        assert_eq!(directory.budget.used.load(Ordering::Relaxed), 0);
        assert_eq!(std::fs::read_dir(root.path()).unwrap().count(), 0);
    }
}