Later analyses are too conservative to remove a copy, but it should be fairly safe to
elide the copy for noncopyable globals, since accesses are tightly scoped and dynamically checked,
so consumes aren't possible, and borrows and inout accesses of mutable globals are dynamically
guarded and not subject to exclusivity checks. rdar://114329759
* Allow normal function results of @yield_once coroutines
* Address review comments
* Workaround LLVM coroutine codegen problem: it assumes that unwind path never returns.
This is not true to Swift coroutines as unwind path should end with error result.
a closure expression, then don't actually do it. The long term plan is
to actually do this, which should just be a matter of taking some of the
code out of reabstraction thunk emission and using it in prolog/epilog/return
emission. In the short term, the goal is just to get the conversion
information down to the closure emitter so that we can see that we're
erasing into an `@isolated(any)` type and then actually erase the
closure's isolation properly instead of relying on type-based erasure,
which can't handle parameter/capture isolation correctly.
Synthesized stub constructors contain `#line` directives and under certain
circumstances these directives can have invalid source locations. SILGen would
trigger an assert (or invoke UB in non-asserts builds) when attempting to
expand these directives into literal values. Check the source location for
validity before translating the location to a location in the outermost source
file.
Resolves rdar://121971741
The main piece that's still missing here is support for closures;
they actually mostly work, but they infer the wrong isolation for
actor-isolated closures (it's not expressed in the type, so obviously
they're non-isolated), so it's not really functional. We also have
a significant problem where reabstraction thunks collide incorrectly
because we don't mangle (or represent!) formal isolation into
SILFunctionType; that's another follow-up. Otherwise, I think SILGen
is working.
This generates significantly better code directly out of SILGen, at
the cost of having to reimplement a little bit of the argument-emission
logic to handle default arguments. But it also neatly sidesteps the
problems we have with splitting tuple RValues when the tuple contains
a pack expansion, which will require some significant surgery to RValue
to fix. That, in turn, fixes rdar://121489308.
In preparation for inserting mark_dependence instructions for lifetime
dependencies early, immediately after SILGen. That will simplify the
implementation of borrowed arguments.
Marking them unresolved is needed to make OSSA verification
conservative until lifetime dependence diagnostics runs.
This adds SIL-level support and LLVM codegen for normal results of a coroutine.
The main user of this will be autodiff as VJP of a coroutine must be a coroutine itself (in order to produce the yielded result) and return a pullback closure as a normal result.
For now only direct results are supported, but this seems to be enough for autodiff purposes.
When the deployment target is Swift 5.9 aligned or higher, we should not need
to call `_diagnoseUnavailableCodeReached()` via a back-deployment thunk since
the function will always be available in the standard library.
Resolves rdar://121878128
It's better to ask SILType if it is MoveOnly than go to the AST type and
ask if it is noncopyable, because some types in SIL do not have a
well-defined notion of conformance in the AST.
The dependent 'value' may be marked 'nonescaping', which guarantees that the
lifetime dependence is statically enforceable. In this case, the compiler
must be able to follow all values forwarded from the dependent 'value', and
recognize all final (non-forwarded, non-escaping) use points. This implies
that `findPointerEscape` is false. A diagnostic pass checks that the
incoming SIL to verify that these use points are all initially within the
'base' lifetime. Regular 'mark_dependence' semantics ensure that
optimizations cannot violate the lifetime dependence after diagnostics.
The `_diagnoseUnavailableCodeReached()` function was introduced in the Swift
5.9 standard library and employs `@backDeployed` to support compilation of
binaries that target OS releases aligned with earlier Swift releases.
Unfortunately, though, this backdeployment strategy doesn't work well for some
unusual build environments. Specifically, in some configurations code may be
built with a compiler from a recent Swift toolchain and then linked against the
dylibs in an older toolchain. When linking against the older dylibs, the
`_diagnoseUnavailableCodeReached()` function does not exist but the
`@backDeployed` thunks emitted into the binary reference that function and
therefore linking fails.
The idea of building with one toolchain and then linking to the dylibs in a
different, older toolchain is extremely dubious. However, it exists and for now
we need to support it. This PR introduces an alternative
`_diagnoseUnavailableCodeReached()` function that is annotated with
`@_alwaysEmitIntoClient`. Calls to the AEIC variant are now emitted by the
compiler when the deployment target is before Swift 5.9.
Once these unusual build environments upgrade and start linking against a Swift
5.9 toolchain or later we can revert all of this.
Resolves rdar://119046537
For an isolated ObjC function that is not async, we
emit a hops around the call. But if that function
returns an autoreleased pointer, we need to ensure
we're retaining that pointer before hopping back
after the call. We weren't doing that in the case
of an autoreleased NSError:
```
%10 = alloc_stack $@sil_unmanaged Optional<NSError>
%19 = ... a bunch of steps to wrap up %10 ...
%20 = enum $Optional<AutoreleasingUnsafeMutablePointer<Optional<NSError>>>, #Optional.some!enumelt, %19 : $AutoreleasingUnsafeMutablePointer<Optional<NSError>>
hop_to_executor $MainActor
%26 = apply X(Y, %20) : $@convention(objc_method) (NSObject, Optional<AutoreleasingUnsafeMutablePointer<Optional<NSError>>>) -> @autoreleased Optional<NSString>
hop_to_executor $Optional<Builtin.Executor>
// retain the autoreleased pointer written-out.
%28 = load [trivial] %10 : $*@sil_unmanaged Optional<NSError>
%29 = unmanaged_to_ref %28 : $@sil_unmanaged Optional<NSError> to $Optional<NSError>
%30 = copy_value %29 : $Optional<NSError>
assign %31 to %7 : $*Optional<NSError>
```
This patch sinks the hop emission after the call
so it happens after doing that copy.
rdar://114049646
Mark the result of a move-only addressor as unresolved. The pointed-at value
cannot be consumed so ensure that only [read] or [modify] accesses are
performed. Update the move-only checker to recognize code patterns
from addressors.
Following https://github.com/apple/swift/pull/70333, do the same thing for
modify coroutines, marking the result so that we check uses of the result to
ensure it isn't consumed (without being reinitialized).
Mark the result of starting a read coroutine to be checked by the move-only checker, and then
update the pattern matching in the move checker itself so that it recognizes code patterns
involving yielding from and receiving yields from read coroutines. Teach move only diagnostics
to get the property name for an access through a read coroutine from the referenced declaration.
