The signal driver uses a `UnixStream` to receive signal events.
Previously, the signal driver used `PollEvented` internally to receive
events on the `UnixStream`. However, using `PollEvented` from
within a runtime driver created a circular link between the runtime and
the `PollEvented` instance.
This patch replaces `PollEvented` usage in favor of accessing the I/O
driver directly. The I/O driver now reserves a token for signal-related
events and tracks signal readiness internally. The signal driver queries
the I/O driver to check for signal-related readiness.
The signal feature only requires a driver with unix platforms. The
unix signal driver uses the I/O driver. A future refactor would like to
make the signal driver use internal APIs of the I/O driver. To do this,
the signal driver must be moved to the runtime module.
This patch removes a handle to the internal runtime driver handle held
by the runtime. This is another step towards reducing the number of Arc
refs across the runtime internals. Specifically, this change is part of
an effort to remove an Arc in the time driver itself.
This excludes the initial let statement of the proc-macro expansion
from receiving the last statement spans to aid completions in rust-analyzer.
The current span confuses rust-analyzer as it will map the tail expression
tokens to the let keyword (as this is the first token it finds with the
same span) which currently breaks completions. This commit should not
degrade the initial intent of the span reusages, as the type mismatch
parts are still spanned appropriately.
The `LocalSet` implementation includes a conditional compilation clause
that removes the `const` statement from the `thread_local` definition.
However, there already is an internal macro that does this:
`tokio_thread_local`.
This patch removes the conditional compilation in favor of using the
`tokio_thread_local` macro. This also fixes a conditional compilation
issue with an internal utility (`RcCell`).
Tokio maintains an internal thread_local macro that abstracts some
conditional build logic. Before this patch, the macro was named the same
as the `std` macro (`thread_local`). This resulted in confusion as to
whether or not the internal macro or the std macro was being called.
This patch renames the internal macro to `tokio_thread_local` making it
more obvious.
Motivation
Currently, when a task spawned on a `LocalSet` is woken by an I/O driver
or time driver running on the same thread as the `LocalSet`, the task is
pushed to the `LocalSet`'s locked remote run queue rather than to its
unsynchronized local run queue. This is unfortunate, as it
negates some of the performance benefits of having an unsynchronized
local run queue. Instead, tasks are only woken to the local queue when
they are woken by other tasks also running on the local set.
This occurs because the local queue is only used when the `CONTEXT`
thread-local contains a Context that's the same as the task's
`Schedule` instance (an `Arc<Shared>`)'s Context. When the `LocalSet`
is not being polled, the thread-local is unset, and the local run queue
cannot be accessed by the `Schedule` implementation for `Arc<Shared>`.
Solution
This branch fixes this by moving the local run queue into Shared along
with the remote run queue. When an `Arc<Shared>`'s Schedule impl wakes
a task and the `CONTEXT` thread-local is None (indicating we are not
currently polling the LocalSet on this thread), we now check if the
current thread's `ThreadId` matches that of the thread the `LocalSet`
was created on, and push the woken task to the local queue if it was.
Moving the local run queue into `Shared` is somewhat unfortunate, as it
means we now have a single field on the `Shared` type, which must not be
accessed from other threads and must add an unsafe impl `Sync` for `Shared`.
However, it's the only viable way to wake to the local queue
from the Schedule impl for `Arc<Shared>`, so I figured it was worth
the additional unsafe code. I added a debug assertion to check that the
local queue is only accessed from the thread that owns the `LocalSet`.
## Motivation
It turns out that `LocalSet` has...never actually passed Miri. This is
because all of the tests for `LocalSet` are defined as integration tests
in `tokio/tests/task_local_set.rs`, rather than lib tests in
`tokio/src`, and we never run Miri for integration tests.
PR #5095 added a new test reproducing an unrelated bug in `LocalSet`,
which had to be implemented as a lib test, as it must make internal
assertions about the state of the `LocalSet` that cannot be tested with
just the public API. This test failed under Miri, and it turns out that
the reason for this is unrelated to the change in #5095 --- `LocalSet`
uses the `LocalOwnedTasks` type, which turns out to contain a stacked
borrows violation due to converting an `&Header` into an
`NonNull<Header>` when removing a task from the `LocalOwnedTasks` list.
## Solution
Fortunately, this was actually quite easy to fix. The non-`Local`
version of `OwnedTasks` already uses `Task::header_ptr()` to get a
`NonNull<Header>` in its version of `remove`, which avoids this issue.
Therefore, the fix was as simple as updating `LocalOwnedTasks` to do the
same.
I've also added a very simple lib test in `src/task/local_set.rs` that
just creates a `LocalSet` and spawns a single task on it under a
current-thread runtime. This test fails under Miri without the fix in
this PR: https://github.com/tokio-rs/tokio/actions/runs/3237072694/jobs/5303682530
We should probably keep the test so that we're always testing at least
the most trivial `LocalSet` usage under Miri...
The original tests for the `Builder::rng_seed` added in #4910 were a bit
fragile. There have already been a couple of instances where internal
refactoring caused the tests to fail and need to be modified. While it
is expected that internal refactoring may cause the random values to
change, this shouldn't cause the tests to break.
The tests should be more robust and not be affected by internal
refactoring or changes in the Rust compiler version. The tests are
changed to perform the same operation in 2 runtimes created with the
same seed, the expectation is that the values that result from each
runtime are the same.
When 64-bit atomics are supported, use 32-bit queue indices. This
greatly improves resilience to ABA and has no impact on performance on
64-bit platforms.
Fixes: #5041
Functions that may panic can be annotated with `#[track_caller]` so that
in the event of a panic, the function where the user called the
panicking function is shown instead of the file and line within Tokio
source.
This change adds `#[track_caller]` to two public APIs in tokio task
module which weren't added in #4848.
* `tokio::task::block_in_place`
* `tokio::task::spawn_local`
These APIs had call stacks that went through closures, which is a use
case not supported by `#[track_caller]`. These two cases have been
refactored so that the `panic!` call is no longer inside a closure.
Tests have been added for these two new cases.
Refs: #4413
This patch removes all internal `Unpark` runtime types. The runtime now
uses the `Handle` types (`runtime::io::Handle` and
`runtime::time::Handle`) to signal threads to unpark.
Without separate `Unpark` types, future patches will be able to remove
more `Arc`s used inside various drivers.
This patch removes the driver Unpark handle out of the multi-thread
parker and passes a reference in when it is needed. This is a first step
towards getting rid of the separate driver unpark handle in favor of
just using the regular driver handle. Because the regular driver handle
is owned at a higher level (at the top of the worker struct).
This is the equivalent as tokio-rs/tokio#5023, but for the MultiThread
scheduler. This patch updates `MultiThread` to use `Arc<Handle>` for all
of its cross-thread needs instead of `Arc<Shared>`.
The effect of this change is that the multi-thread scheduler only has a
single `Arc` type, which includes the driver handles, keeping all
"shared state" together.
This is the next step in the internal runtime refactoring. This patch
updates `CurrentThread` to use `Arc<Handle>` for all of its cross-thread
needs instead of `Arc<Shared>`.
The effect of this change is that the current-thread scheduler only has
a single `Arc` type, which includes the driver handles, keeping all
"shared state" together.
The `runtime::Handle` struct is part of the public API but is also used
internally. This has created a bit of tension. An earlier patch made
defined Handle as a private struct in some cases when `rt` is not
enabled.
This patch splits out internal handle concerns into a new
`scheduler::Handle` type, which will only be internal. This also defines
a `Handle` type for each scheduler variant. Eventually, the
per-scheduler `Handle` types will replace the per-scheduler `Spawner`
types, but more work is needed before we can make that change.
The `tokio::select!` macro polls branches in a random order. While this
is desirable in production, for testing purposes a more deterministic
approach can be useul.
This change adds an additional parameter `rng_seed` to the runtime
`Builder` to set the random number generator seed. This value is then
used to reset the seed on the current thread when the runtime is entered
into (restoring the previous value when the thread leaves the runtime). All
threads created explicitly by the runtime also have a seed set as the
runtime is built. Each thread is set with a seed from a deterministic
sequence.
This guarantees that calls to the `tokio::select!` macro which are
performed in the same order on the same thread will poll branches in the
same order.
Additionally, the peer chosen to attempt to steal work from also uses a
deterministic sequence if `rng_seed` is set.
Both the builder parameter as well as the `RngSeed` struct are marked
unstable initially.
