Consider an `@_alwaysEmitIntoClient` function and a custom derivative
defined
for it. Previously, such a combination resulted different errors under
different
circumstances.
Sometimes, there were linker errors due to missing derivative function
symbol -
these occurred when we tried to find the derivative in a module, while
it
should have been emitted into client's code (and it did not happen).
Sometimes, there were SIL verification failures like this:
```
SIL verification failed: internal/private function cannot be serialized or serializable: !F->isAnySerialized() || embedded
```
Linkage and serialization options for the derivative were not handled
properly,
and, instead of PublicNonABI linkage, we had Private one which is
unsupported
for serialization - but we need to serialize `@_alwaysEmitIntoClient`
functions
so the client's code is able to see them.
This patch resolves the issue and adds proper handling of custom
derivatives
of `@_alwaysEmitIntoClient` functions. Note that either both the
function and
its custom derivative or none of them should have
`@_alwaysEmitIntoClient`
attribute, mismatch in this attribute is not supported.
The following cases are handled (assume that in each case client's code
uses
the derivative).
1. Both the function and its derivative are defined in a single file in
one module.
2. Both the function and its derivative are defined in different files
which
are compiled to a single module.
3. The function is defined in one module, its derivative is defined in
another
module.
4. The function and the derivative are defined as members of a protocol
extension in two separate modules - one for the function and one for the
derivative. A struct conforming the protocol is defined in the third
module.
5. The function and the derivative are defined as members of a struct
extension in two separate modules - one for the function and one for the
derivative.
The changes allow to define derivatives for methods of `SIMD`.
Fixes#54445
<!--
If this pull request is targeting a release branch, please fill out the
following form:
https://github.com/swiftlang/.github/blob/main/PULL_REQUEST_TEMPLATE/release.md?plain=1
Otherwise, replace this comment with a description of your changes and
rationale. Provide links to external references/discussions if
appropriate.
If this pull request resolves any GitHub issues, link them like so:
Resolves <link to issue>, resolves <link to another issue>.
For more information about linking a pull request to an issue, see:
https://docs.github.com/issues/tracking-your-work-with-issues/linking-a-pull-request-to-an-issue
-->
<!--
Before merging this pull request, you must run the Swift continuous
integration tests.
For information about triggering CI builds via @swift-ci, see:
https://github.com/apple/swift/blob/main/docs/ContinuousIntegration.md#swift-ci
Thank you for your contribution to Swift!
-->
This ensures that when we generate the vtable thunk for a
nonisolated(nonsending) override (or vis-a-versa), we get the ABI correct. I
also added tests for all of the relevant cases for vtables that we check for
protocols.
rdar://151394209
Now that coroutine kind (and consequently ABI) for the accessors is
keyed off a SIL option, it's no longer possible to read whether a given
SILFunction arose from a read/modify coroutine just by checking its
coroutine kind. Regardless of ABI, read/modify coroutines may only
unwind (i.e. are only permitted not to "run to completion") if the
relevant experimental (soon to be deleted) feature is enabled.
This involved cleaning up and generalizing the work from
f245389bb3. I followed the same approach: hide the
implicit parameter while we run things through translateArgument and friends and
then put it back in manually.
As an additional benefit, I think I found a good place to put FunctionIsolation
onto lowered AnyFunctionTypes that will not cause cycles in the evaluator since
it is done in TypeLowering.
rdar://148785846
Specifically:
1. I made it so that thunks from caller -> concurrent properly ignore the
isolated parameter of the thunk when calling the concurrent function.
rdar://148112362
2. I made it so that thunks from concurrent -> caller properly create a
Optional<any Actor>.none and pass that into the caller function.
rdar://148112384
3. I made it so that in cases where we are assigning an @Sendable caller to a
non-sendable caller variable, we allow for the conversion as long as the
parameters/results are sendable as well.
rdar://148112532
4. I made it so that when we generate a thunk from @execution(caller) ->
@GlobalActor, we mangle in @GlobalActor into the thunk.
rdar://148112569
5. I discovered that due to the way we handle function conversion expr/decl ref
expr, we were emitted two thunks when we assigned a global @caller function to a
local @caller variable. The result is that we would first cast from @caller ->
@concurrent and then back to @caller. The result of this would be that the
@caller function would always be called on the global queue.
rdar://148112646
I also added a bunch of basic tests as well that showed that this behavior was
broken.
This just ensures that all thunks are verified right when they are created. This
is necessary since some thunks do not go through function post processing.
This would make sure that async function types marked as `@execution(caller)`
have correct isolation.
Also defines all of the possible conversions to and from `caller`
isolated function types.
If the conformance generic signature fixes all generic parameters,
F.getForwardingSubstitutionMap() is empty. Instead, map the
replacement types of the substitution map into the generic
environment earlier, before we strip off a fully-concrete generic
signature.
When the CoroutineAccessors feature is enabled, `begin_apply`
instructions produce an additional result representing the allocation
done by the callee. Fix a couple of cases where this additional result
was not being handled.
This corresponds to the parameter-passing convention of the Itanium C++
ABI, in which the argument is passed indirectly and possibly modified,
but not destroyed, by the callee.
@in_cxx is handled the same way as @in in callers and @in_guaranteed in
callees. OwnershipModelEliminator emits the call to destroy_addr that is
needed to destroy the argument in the caller.
rdar://122707697
The reason why I am fixing this is that otherwise, we get a warning when one
creates an actor isolated closure and pass it into a task, e.x.:
```swift
@MainActor func test() {
// We would get a warning on the closure below saying that we are sending
// a closure that is MainActor isolated.
Task {
...
}
}
```
Create two versions of the following functions:
isConsumedParameter
isGuaranteedParameter
SILParameterInfo::isConsumed
SILParameterInfo::isGuaranteed
SILArgumentConvention::isOwnedConvention
SILArgumentConvention::isGuaranteedConvention
These changes will be needed when we add a new convention for
non-trivial C++ types as the functions will return different answers
depending on whether they are called for the caller or the callee. This
commit doesn't change any functionality.
This reverts commit aa5dddb952.
Fixes `preset=buildbot,tools=RA,stdlib=DA` CI job, which without this revert fails on `AutoDiff/SILGen/nil_coalescing.swift` test.
Although I don't plan to bring over new assertions wholesale
into the current qualification branch, it's entirely possible
that various minor changes in main will use the new assertions;
having this basic support in the release branch will simplify that.
(This is why I'm adding the includes as a separate pass from
rewriting the individual assertions)
[serialized_for_package] if Package CMO is enabled. The latter kind
allows a function to be serialized even if it contains loadable types,
if Package CMO is enabled. Renamed IsSerialized_t as SerializedKind_t.
The tri-state serialization kind requires validating inlinability
depending on the serialization kinds of callee vs caller; e.g. if the
callee is [serialized_for_package], the caller must be _not_ [serialized].
Renamed `hasValidLinkageForFragileInline` as `canBeInlinedIntoCaller`
that takes in its caller's SerializedKind as an argument. Another argument
`assumeFragileCaller` is also added to ensure that the calle sites of
this function know the caller is serialized unless it's called for SIL
inlining optimization passes.
The [serialized_for_package] attribute is allowed for SIL function, global var,
v-table, and witness-table.
Resolves rdar://128406520
We do this by pushing the conversion down to the emission of the
closure expression, then teaching closure emission to apply the isolation
to the closure. Ideally, we combine the isolation along with the rest of
the conversion peephole, but if necessary, we make sure we emit the
isolation.
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.
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.
The thunk is a reabstraction thunk with a custom prolog that
has a runtime precondition check that makes sure that concurrent
environment where the thunk is run matches that of a global
actor associated with the thunked type.
For `@preconcurrency` conformance witness thunks replace hop to
executor with a precondition to make sure that the thunk is always
called in the expected context.
There are a bunch of static `collectExistentialConformances` copied
around Sema and SILGen that are almost the same, save for whether they
want to permit missing conformances and/or check conditional
conformances.
This commit requestifies and combines all but one of these functions
into a `ModuleDecl::collectExistentialConformances`. The motivation for
this clean-up is another place that will need this procedure.
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.
patterns) in argument positions in reabstraction thunks.
Most of the difficulty in this work continues to center around
(1) trying to reuse as much code as possible between the parameter
and result paths and (2) propagating ownership information as
necessary throughout the code. I did my best to assert the preconditions
and postconditions here, but undoubtedly I'm missing cases. Some
simplicity here is still eluding me here.
This patch necessarily changes quite a bit of the code used in
non-variadic paths. I tried to avoid doing things that I knew would
be risky, like optimizing copies. I did fail in a few places: e.g.
we should now generate significantly better code when erasing to
Optional<Any>, just because the code was oddly poorly-factored before.
You can see the effect on the function_conversion test case.
