With these magic symbols, programs that link against the _Concurrency
module with a deployment target prior to iOS 15 / macOS 12 / watchOS 8
will reference `libswift_Concurrency.dylib` via rpath.
Fixes rdar://81187835.
Some headers switch their inline implementations based on
SWIFT_STDLIB_SINGLE_THREAD_RUNTIME definition.
This fixes linking failure while building runtime unittests
This patch repairs the build failure for cooperative executor.
And also fixed main executor to avoid assertion failure due to
`witnessTable == nullptr`
control swift extended frame information emission
On linux we default to disable the extended frame info (since the system
libraries don't support it).
On darwin the default is to automatically choose based on the deployment target.
The Concurrency library explicitly forces extended frame information and the
back deployment library explicitly disables it.
Older Dispatch versions have width-limited concurrent queues. While not
as good as the newly-introduced cooperative queues, they at least
address the thread-explosion problem, so use them in the back-deployed
concurrency libraries.
Implements rdar://81131034.
If we didn't do this (and we didn't), the tasks get released as we
perform the next() impl, and move the value from the ready task to the
waiting task. Then, the ready task gets destroyed.
But as the task group exists, it performs a cancelAll() and that
iterates over all records. Those records were not removed previously
(!!!) which meant we were pointing at now deallocated tasks.
Previously this worked because we didn't deallocate the tasks, so they
leaked, but we didn't crash. With the memory leak fixed, this began to
crash since we'd attempt to cancel already destroyed tasks.
Solution:
- Remove task records whenever they complete a waiting task.
- This can ONLY be done by the "group owning task" itself, becuause
the contract of ONLY this task being allowed to modify records. o
It MUST NOT be done by the completing tasks as they complete, as it
would race with the owning task modifying this linked list of child
tasks in the group record.
The proper handling of task group child tasks is that:
- if it completes a waiting task immediately, we don't need to retain it
- we just move the value to the waiting task and can destroy the task
- if we need to store the ready task and wait for a waiting task (for a
task that hits `await group.next()`) then we need to retain the ready
task.
- as the waiting task arrives, we move the value from the ready task
to the waiting task, and swift_release the ready task -- it will now
be destroyed safely.
`swift_async_extendedFramePointerFlags` is used in back-deployed
concurrency code as a mask to set the extended async frame info bit on
OS versions that understand it. Set this mask to zero in the
back-deployed libraries.
Darwin OSes support vouchers, which are key/value sets that can be adopted on a thread to influence its execution, or sent to another process. APIs like Dispatch propagate vouchers to worker threads when running async code. This change makes Swift Concurrency do the same.
The change consists of a few different parts:
1. A set of shims (in VoucherShims.h) which provides declarations for the necessary calls when they're not available from the SDK, and stub implementations for non-Darwin platforms.
2. One of Job's reserved fields is now used to store the voucher associated with a job.
3. Jobs grab the current thread's voucher when they're created.
4. A VoucherManager class manages adoption of vouchers when running a Job, and replacing vouchers in suspended tasks.
5. A VoucherManager instance is maintained in ExecutionTrackingInfo, and is updated as necessary throughout a Job/Task's lifecycle.
rdar://76080222
Swift 5.5 does not implement support for structural types that are
Sendable, which breaks type checking of the newer _Concurrency module.
Work around the issue with an `#if` in the offending inline code.
Fixes rdar://82602353.
`SWIFT_STDLIB_SINGLE_THREADED_RUNTIME` mode has been broken for a long time.
This patch guards some includes and use of libdispatch headers so that platforms
that doesn't support libdispatch can build cooperative executor runtime.
And fixed missing implementations for cooperative mode.
The code that saves/restores the exclusivity checks for tasks was
newly introduced into the runtime. Clone that code into the back-
deployed version of the runtime.
Remove the option that explicitly enables concurrency back-deployment,
and instead always enable its support in the compiler. Remove the use
of the extraneous CMake option as well.
When build-script is given `--back-deploy-concurrency`, also use that
to build other parts of Swift with the back-deployed versions:
* The compiler allows async and actors to be defined with the
back-deployed availability, e.g., the same as `-Xfrontend
-enable-experimental-back-deploy-concurrency`. (The latter will go
away soon)
* The standard library unit testing framework and distributed actors
library are build with the older OS versions.
* The tests use the older OS versions, with some adjustments to make
them agnostic to the back-deployment setting.
When enabling the build of the back-deployable concurrency library via
the build-script option `--back-deploy-concurrency`, also build the
compiler and (main) concurrency library to support older deployment targets.
Building the compiler for older deployment targets is effectively the
same as implicitly passing
`-Xfrontend -enable-experimental-back-deploy-concurrency`. That option
should probably go away.
Building the primary _Concurrency library for back-deployment means
setting the "SwiftStdlib 5.5" availability back to the earlier
deployment targets. This should have no effect on how the _Concurrency
library binary is built, but it does ensure that the right
availability annotations are in the _Concurrency module.
Introduce an additional build product to build-script to build
back-deployable concurrency libraries. These libraries would need to
be embedded in apps deployed prior to macOS 12/iOS 15 to support
concurrency.
The built-script option `--back-deploy-concurrency` can be provided to
build these back-deployment libraries. They are built in addition to
the normal concurrency libraries, as a separate product that installs
into `lib/swift-5.5/<platform>` within the toolchain. The macro
`SWIFT_CONCURRENCY_BACK_DEPLOYMENT` is set when building the
concurrency library, so that we can adapt the implementation to older
OS's.
This macro takes the string and parameters directly, and is conditionally defined to either call fprintf or ignore its arguments. This makes the call sites a little more pleasant (no #if scattered about) and ensures every log includes the thread ID and a newline automatically.
If a future task has already completed, then task_future_wait doesn't populate the future context,
but asyncLet_finish was inappropriately relying on it to recover the result buffer pointer to
destroy the value inside. There's room in asyncLet_finish's own context to put this pointer, so
do that instead. Fixes rdar://81481317.
runAsyncAndBlock has long been dead, I'm pulling out the machinery that
lives in the runtime that used to support it. This is the first layer
that can go away.
