Skip to content

runtime: fix idle bookkeeping when a task dump wakes parked workers - #8372

Merged
Darksonn merged 4 commits into
tokio-rs:masterfrom
yzfeng2020:runtime-taskdump-idle-accounting
Sep 8, 2026
Merged

Darksonn merged 4 commits into
tokio-rs:masterfrom
yzfeng2020:runtime-taskdump-idle-accounting

Conversation

@yzfeng2020

@yzfeng2020 yzfeng2020 commented Aug 19, 2026 •

Copy link
Copy Markdown
Contributor

Motivation

On the multi-threaded runtime, a task dump taken while a worker is idle corrupts the scheduler's idle bookkeeping, after which tasks spawned from outside the runtime never run.

Handle::dump calls notify_all, which unparks worker threads directly instead of going through Idle, so a parked worker is woken for tracing while the scheduler still counts it as a sleeper. transition_from_parked doesn't account
for that: with no tasks queued it returns without performing the logical unpark, so when the worker parks again after tracing it is pushed onto sleepers twice and num_unparked is decremented twice, underflowing. From then on
notify_should_wakeup is always false and no worker is ever notified of new work.

Solution

Treat a worker woken for tracing like a worker woken with tasks queued and perform the logical unpark. This mirrors transition_to_parked, which already refuses to park a worker that is about to be traced.

The regression tests in tokio/tests/dump.rs park every worker, take a dump, wait for the workers to be traced and re-park, then assert that a task spawned from outside the runtime still runs.

The current_thread scheduler is unaffected; it has no idle bookkeeping of this kind.

Requesting a task dump calls `notify_all`, which unparks every worker
thread directly rather than going through `Idle`. A worker that was
parked at that point is woken to be traced while the scheduler still
counts it as a sleeper.

`transition_from_parked` did not account for that. With no tasks
queued, the worker saw itself as still parked in `Idle` and returned
without performing the logical unpark. Once tracing finished and the
worker parked again, `transition_worker_to_parked` pushed it onto the
sleeper list a second time and decremented `num_unparked` a second
time, underflowing it. From then on `notify_should_wakeup` was always
false, so no worker was ever notified of newly spawned work and the
runtime silently stopped running tasks submitted from outside of it.

Treat a worker woken for tracing like a worker woken with tasks queued
and perform the logical unpark. This mirrors `transition_to_parked`,
which already refuses to park a worker that is about to be traced.

Co-authored-by: Isaac
@github-actions github-actions Bot added the R-loom-multi-thread Run loom multi-thread tests on this PR label Aug 19, 2026
`transition_to_parked` has refused to park a worker that is about to be
traced for some time, but its comment did not say so. Mention it, and
trim the comment added to `transition_from_parked` in the previous
commit to match.

Co-authored-by: Isaac
The bug fixed earlier in this PR was silent. `sleepers` is a `Vec`, so a
worker can be tracked as a sleeper twice, and `dec_num_unparked` is an
unchecked `fetch_sub`, so `num_unparked` wrapped rather than tripping on
the second decrement. The runtime then believed every worker was busy
and stopped notifying workers of new work, without any diagnostic.

Assert both halves in debug builds: a worker is not already tracked as a
sleeper when it parks, and `num_unparked` does not underflow. `Idle`
already asserts the same class of invariant when it notifies a worker.

Co-authored-by: Isaac

// A worker that is already tracked as a sleeper would be counted twice,
// which corrupts `num_unparked` and the sleeper list.
debug_assert!(!lock.idle.sleepers.contains(&worker));

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

hey @conradludgate, i added two debug_assert! in addition to the fix. lmk what you think, thanks!

@Darksonn Darksonn added A-tokio Area: The main tokio crate M-runtime Module: tokio/runtime M-taskdump --cfg tokio_taskdump C-enhancement Category: A PR with an enhancement or bugfix. labels Sep 8, 2026

@Darksonn Darksonn left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thank you.

@Darksonn
Darksonn enabled auto-merge (squash) September 8, 2026 09:04
@Darksonn
Darksonn merged commit 1309451 into tokio-rs:master Sep 8, 2026
93 of 95 checks passed
yzfeng2020 added a commit to yzfeng2020/tokio that referenced this pull request Sep 8, 2026
Resolve the conflict in tokio/tests/dump.rs: both sides appended a new
regression test at the end of the file. Keep both `wedged_worker_during_tracing`
from this branch and the `dump_of_parked_workers` module from tokio-rs#8372.
dadezzz pushed a commit to dadezzz/kube-dns-rs that referenced this pull request Oct 20, 2026
This PR contains the following updates:

| Package | Type | Update | Change |
|---|---|---|---|
| [tokio](https://tokio.rs) ([source](https://github.com/tokio-rs/tokio)) | dependencies | minor | `1.53.3` → `1.54.0` |

---

### Release Notes

<details>
<summary>tokio-rs/tokio (tokio)</summary>

### [`v1.54.0`](https://github.com/tokio-rs/tokio/releases/tag/tokio-1.54.0): Tokio v1.54.0

[Compare Source](tokio-rs/tokio@tokio-1.53.3...tokio-1.54.0)

### 1.54.0 (October 19th, 2026)

The 1.54.0 release includes all changes up to 1.53.3.

The MSRV is increased to 1.85. ([#&#8203;8514])

##### Added

- fs: add `File::set_times`, `File::set_modified`, and `fs::FileTimes` ([#&#8203;8504])
- net: implement `Default` for `named_pipe::{ClientOptions, ServerOptions}` ([#&#8203;8378])
- runtime: add `Builder::max_io_events_per_busy_tick` ([#&#8203;8410])
- runtime: stabilize `tokio::runtime::worker_index()` ([#&#8203;8431])
- sync: add `SemaphorePermit::semaphore()` ([#&#8203;8482])
- sync: add `blocking_acquire` methods to `Semaphore` ([#&#8203;8269])
- sync: add `watch::Receiver::is_closed` ([#&#8203;8481])
- wasm: support JSPI for Emscripten target ([#&#8203;8285], [#&#8203;8627])
- wasm: support the `wasm32-unknown-emscripten` target ([#&#8203;8281], [#&#8203;8422], [#&#8203;8528], [#&#8203;8628])

##### Changed

- fs: reject zero in `File::set_max_buf_size` ([#&#8203;8449])
- io: make `copy_buf` cooperative ([#&#8203;8459])
- io: panic when `BufReader` or `BufStream` is created with zero capacity ([#&#8203;8494])
- io: panic when in-memory pipe is created with zero capacity ([#&#8203;8397])
- io: reject zero buffer size in `copy_bidirectional_with_sizes` ([#&#8203;8454])
- io: support vectored writes in `Sink` ([#&#8203;8663])
- net: start `UnixStream::pair` and `UnixDatagram::pair` sockets as writable ([#&#8203;8444])
- net: start accepted sockets as readable and writable ([#&#8203;8435])
- net: support `UCred` on Fuchsia ([#&#8203;8257])
- process: use `mio::unix::pipe` for Unix child stdio ([#&#8203;8603])
- runtime: cache-pad the shared injection queue and idle state ([#&#8203;8622])
- runtime: decide auto-boxing of spawned futures at monomorphization time ([#&#8203;8414])
- runtime: make `enable_io` available whenever `rt` is enabled ([#&#8203;8387])
- runtime: move `block_in_place` setup into a non-generic function ([#&#8203;8315])
- runtime: move the multi-thread inject queue to its own mutex ([#&#8203;8382])
- runtime: panic when `max_io_events_per_tick` is zero ([#&#8203;8476])
- runtime: suggest `.await` in nested `block_on` panic message ([#&#8203;8501])
- sync: avoid atomic reader count in broadcast slots ([#&#8203;8357])
- sync: use acquire/release orderings in `Notify` ([#&#8203;8325])
- time: remove the runtime handle from timers ([#&#8203;8458])

##### Fixed

- fs: restore `File` internal state when op fails ([#&#8203;8291])
- io: complete zero-length memory stream operations immediately ([#&#8203;8323])
- io: flush the writer when the reader stalls in `copy_buf` ([#&#8203;8424])
- io: keep the shutdown bit when clearing readiness ([#&#8203;8435])
- io: preserve the full `ScheduledIo` tick in readiness events ([#&#8203;8331])
- io: read lines as bytes so a partial line survives an I/O error ([#&#8203;8400])
- io: retry on `ErrorKind::Interrupted` in more I/O operations ([#&#8203;8417], [#&#8203;8432], [#&#8203;8448], [#&#8203;8453], [#&#8203;8455], [#&#8203;8460])
- io: use `spawn_mandatory_blocking` for stdout and stderr ([#&#8203;8506])
- io: wait for pending operations before `BufReader` seeks ([#&#8203;8599])
- net: close `CreateFileW` handle on `SetNamedPipeHandleState` failure ([#&#8203;8418])
- net: fix peer credentials on NetBSD ([#&#8203;8589])
- net: saturate `TcpSocket` and `UnixSocket` listen backlog ([#&#8203;8511])
- net: treat `ENOTCONN` as success in `UnixStream::shutdown` ([#&#8203;8526])
- runtime: allow building with `--all-features` without `tokio_unstable` ([#&#8203;8401])
- runtime: avoid overflow in `shutdown_timeout()` ([#&#8203;8497])
- runtime: fix `spawn_blocking` hang when only scheduler workers exist ([#&#8203;8408])
- runtime: fix misleading signal driver panic message ([#&#8203;8419])
- sync: preserve FIFO lock order when coop budget is exhausted ([#&#8203;8371])
- sync: use `usize` for permit count in `SemaphorePermit` ([#&#8203;8405])
- time: avoid interval deadline overflow ([#&#8203;8385])
- time: handle wrapped top-level timer-wheel slots ([#&#8203;8334])

##### Unstable

- fs: fall back when `io_uring` has no IO driver ([#&#8203;8165])
- metrics: cap log histogram precision ([#&#8203;8496])
- runtime: add an opt-in sharded `spawn_blocking` queue ([#&#8203;8337])
- runtime: expose schedule latency in task hooks ([#&#8203;8282])
- taskdump: fix idle bookkeeping when a task dump wakes parked workers ([#&#8203;8372])
- taskdump: skip non-idle tasks during taskdumps ([#&#8203;8568])

##### Documented

- docs: fix stray braces and broken links in rustdoc ([#&#8203;8605])
- fs: document cancellation behavior of `tokio::fs` ([#&#8203;8265])
- net: add Safety docs for named pipe `from_raw_handle` ([#&#8203;8378])
- net: document that `TcpStream::connect` returns only the last error ([#&#8203;8421])
- runtime: add error documentation for `runtime::Builder` ([#&#8203;8297])
- runtime: link `available_parallelism` in worker count docs ([#&#8203;8625])
- sync: clarify panic behavior of `broadcast::channel()` ([#&#8203;8420])
- sync: fix typo in `Semaphore` documentation ([#&#8203;8521])

[#&#8203;8165]: tokio-rs/tokio#8165

[#&#8203;8257]: tokio-rs/tokio#8257

[#&#8203;8265]: tokio-rs/tokio#8265

[#&#8203;8269]: tokio-rs/tokio#8269

[#&#8203;8281]: tokio-rs/tokio#8281

[#&#8203;8282]: tokio-rs/tokio#8282

[#&#8203;8285]: tokio-rs/tokio#8285

[#&#8203;8291]: tokio-rs/tokio#8291

[#&#8203;8297]: tokio-rs/tokio#8297

[#&#8203;8315]: tokio-rs/tokio#8315

[#&#8203;8323]: tokio-rs/tokio#8323

[#&#8203;8325]: tokio-rs/tokio#8325

[#&#8203;8331]: tokio-rs/tokio#8331

[#&#8203;8334]: tokio-rs/tokio#8334

[#&#8203;8337]: tokio-rs/tokio#8337

[#&#8203;8357]: tokio-rs/tokio#8357

[#&#8203;8371]: tokio-rs/tokio#8371

[#&#8203;8372]: tokio-rs/tokio#8372

[#&#8203;8378]: tokio-rs/tokio#8378

[#&#8203;8382]: tokio-rs/tokio#8382

[#&#8203;8385]: tokio-rs/tokio#8385

[#&#8203;8387]: tokio-rs/tokio#8387

[#&#8203;8397]: tokio-rs/tokio#8397

[#&#8203;8400]: tokio-rs/tokio#8400

[#&#8203;8401]: tokio-rs/tokio#8401

[#&#8203;8405]: tokio-rs/tokio#8405

[#&#8203;8408]: tokio-rs/tokio#8408

[#&#8203;8410]: tokio-rs/tokio#8410

[#&#8203;8414]: tokio-rs/tokio#8414

[#&#8203;8417]: tokio-rs/tokio#8417

[#&#8203;8418]: tokio-rs/tokio#8418

[#&#8203;8419]: tokio-rs/tokio#8419

[#&#8203;8420]: tokio-rs/tokio#8420

[#&#8203;8421]: tokio-rs/tokio#8421

[#&#8203;8422]: tokio-rs/tokio#8422

[#&#8203;8424]: tokio-rs/tokio#8424

[#&#8203;8431]: tokio-rs/tokio#8431

[#&#8203;8432]: tokio-rs/tokio#8432

[#&#8203;8435]: tokio-rs/tokio#8435

[#&#8203;8444]: tokio-rs/tokio#8444

[#&#8203;8448]: tokio-rs/tokio#8448

[#&#8203;8449]: tokio-rs/tokio#8449

[#&#8203;8453]: tokio-rs/tokio#8453

[#&#8203;8454]: tokio-rs/tokio#8454

[#&#8203;8455]: tokio-rs/tokio#8455

[#&#8203;8458]: tokio-rs/tokio#8458

[#&#8203;8459]: tokio-rs/tokio#8459

[#&#8203;8460]: tokio-rs/tokio#8460

[#&#8203;8476]: tokio-rs/tokio#8476

[#&#8203;8481]: tokio-rs/tokio#8481

[#&#8203;8482]: tokio-rs/tokio#8482

[#&#8203;8494]: tokio-rs/tokio#8494

[#&#8203;8496]: tokio-rs/tokio#8496

[#&#8203;8497]: tokio-rs/tokio#8497

[#&#8203;8501]: tokio-rs/tokio#8501

[#&#8203;8504]: tokio-rs/tokio#8504

[#&#8203;8506]: tokio-rs/tokio#8506

[#&#8203;8511]: tokio-rs/tokio#8511

[#&#8203;8514]: tokio-rs/tokio#8514

[#&#8203;8521]: tokio-rs/tokio#8521

[#&#8203;8526]: tokio-rs/tokio#8526

[#&#8203;8528]: tokio-rs/tokio#8528

[#&#8203;8568]: tokio-rs/tokio#8568

[#&#8203;8589]: tokio-rs/tokio#8589

[#&#8203;8599]: tokio-rs/tokio#8599

[#&#8203;8603]: tokio-rs/tokio#8603

[#&#8203;8605]: tokio-rs/tokio#8605

[#&#8203;8622]: tokio-rs/tokio#8622

[#&#8203;8625]: tokio-rs/tokio#8625

[#&#8203;8627]: tokio-rs/tokio#8627

[#&#8203;8628]: tokio-rs/tokio#8628

[#&#8203;8663]: tokio-rs/tokio#8663

</details>

---

### Configuration

📅 **Schedule**: (UTC)

- Branch creation
  - At any time (no schedule defined)
- Automerge
  - At any time (no schedule defined)

🚦 **Automerge**: Disabled by config. Please merge this manually once you are satisfied.

♻ **Rebasing**: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

🔕 **Ignore**: Close this PR and you won't be reminded about this update again.

---

 - [ ] <!-- rebase-check -->If you want to rebase/retry this PR, check this box

---

This PR has been generated by [Mend Renovate CLI](https://github.com/renovatebot/renovate).
<!--renovate-debug:eyJjcmVhdGVkSW5WZXIiOiI0NC4xNjcuNSIsInVwZGF0ZWRJblZlciI6IjQ0LjE2Ny41IiwidGFyZ2V0QnJhbmNoIjoibWFpbiIsImxhYmVscyI6W119-->
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

A-tokio Area: The main tokio crate C-enhancement Category: A PR with an enhancement or bugfix. M-runtime Module: tokio/runtime M-taskdump --cfg tokio_taskdump R-loom-multi-thread Run loom multi-thread tests on this PR

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants