Specifically:
1. We error now if one transfers an 'inout sending' parameter and does not
reinitialize it before the end of the function.
2. We error now if one merges an 'inout sending' parameter into an actor
isolated region and do not reinitialize it with a non-actor isolated value
before the end of the function.
rdar://126303739
This asserts only option is an option to make it quicker/easier to triage
unknown pattern match errors by aborting when we emit it (allowing one to
immediately drop into the debugger at that point).
Previously, it only happened for errors in RegionAnalysis not in
TransferNonSendable itself.
From the perspective of the IR, we are changing SILIsolationInfo such that
inferring an actor instance means looking at equivalence classes of values where
we consider operands to look through instructions to be equivalent to their dest
value. The result is that cases where the IR maybe puts in a copy_value or the
like, we consider the copy_value to have the same isolation info as using the
actor directly. This prevents a class of crashes due to merge failings. Example:
```swift
actor MyActor {
init() async {
init(ns: NonSendableKlass) async {
self.k = NonSendableKlass()
self.helper(ns)
}
func helper(_ newK: NonSendableKlass) {}
}
```
Incidently, we already had a failing test case from this behavior rather than
the one that was the original genesis. Specifically:
1. If a function's SILIsolationInfo is the same as the isolation info of a
SILValue, we assume that no transfer actually occurs.
2. Since we were taking too static of a view of actor instances when comparing,
we would think that a SILIsolationInfo of a #isolation parameter to as an
argument would be different than the ambient's function isolation which is also
that same one. So we would emit a transfer non transferrable error if we pass in
any parameters of the ambient function into another isolated function. Example:
```swift
actor Test {
@TaskLocal static var local: Int?
func withTaskLocal(isolation: isolated (any Actor)? = #isolation,
_ body: (consuming NonSendableValue, isolated (any Actor)?) -> Void) async {
Self.$local.withValue(12) {
// We used to get these errors here since we thought that body's isolation
// was different than the body's isolation.
//
// warning: sending 'body' risks causing data races
// note: actor-isolated 'body' is captured by a actor-isolated closure...
body(NonSendableValue(), isolation)
}
}
}
```
rdar://129400019
This is usually the case. Some examples, where they layout is _not_ dependent:
```
struct S<T> {
var x: Int // no members which depend on T
}
struct S<T> {
var c: SomeClass<T> // a class reference does not depend on the layout of the class
}
```
We view the conversion from a Sendable to a non-Sendable function via
convert function to produce a new fresh sendable value. We should
squelch that error.
TLDR:
The reason why I am doing this is it ensures that temporary store_borrow that we
create when materializing a value before were treated as uses. So we would error
on this:
```swift
@MainActor func transferToMain<T>(_ t: T) async {}
func test() async {
let x = NonSendableKlass()
await transferToMain(x)
await transferToMain(x)
}
```
----
store_borrow is an instruction intended to be used to initialize temporary
alloc_stack with borrows. Since it is a temporary, we do not want to error on
the temporaries initialization... instead, we want to error on the use of the
temporary parameter.
This is achieved by making it so that store_borrow still performs an
assign/merge, but does not require that src/dest be alive. So the regions still
merge (yielding diagnostics for later uses).
It also required me to make it so that PartitionOp::{Assign,Merge} do not
require by default. Instead, we want the individual operations to always emit a
PartitionOp::Require explicitly (which they already did).
One thing to be aware of is that when it comes to diagnostics, we already know
how to find a temporaries original value and how to handle that. So this is the
last part of making store_borrow behave nicely.
rdar://129237675
Enhance the utility with the ability to end lifetimes of lexical values
at indicated instructions, overriding the usual behavior of maintaining
such lifetimes' previous endpoints (modulo non-deinit-barrier
instructions).
Parameterized `extendUnconsumedLiveness` on the ends of interest and the
action to take when visiting the extended boundary and named the
resulting function `visitExtendedUnconsumedBoundary`.
The reason why we are doing this is that otherwise, we have that the alloc_stack
formed for the result is disconnected and despite the fact that we merge it into
the actor region of the class method, we do not have that the alloc_stack
specifically is marked when we attempt to squelch Please.
This patch fixes that problem by detecting when an alloc_stack is being used as
a temporary for an out parameter and makes the alloc_stack initially isolated as
appropriate. It only does this in the specific cases where we can pattern match
it which in my limited testing has handled everything.
When merging SILIsolationInfo for regions, we want to drop
nonisolated(unsafe). This is important since nonisolated(unsafe) should only
apply to the specific "value" that it belongs to, not the entire region.
This creates a problem since in a few places in the code base we initialize a
value (producing a disconnected value) and then initialize it by merging in an
actor isolation. This no longer work since we will then always have
nonisolated(unsafe) stripped, so no values would ever be considered to be
nonisolated(unsafe). After analyzing the use case, I realized that these were
just initialization patterns and in this commit, I added a specific
initialization operation called SILIsolationInfo::initializeTrackableValue and
eliminated those calls to SILIsolationInfo::mergeIsolationRegionInfo.
Since SILIsolationInfo no longer has any merge operation on it, I then
eliminated that code in this commit. This completes the behavior split that I
put into the type system in the last commit. Specifically, I defined a
composition type called SILDynamicMergedIsolationInfo. It represents a
SILIsolationInfo that has been merged... that is why I called it the
DynamicMergedIsolationInfo. It could probably use a better name = (.
This fixes one of the last weird test case that I wrote where we were not letting through valid
nonisolated(unsafe) code.
At the same time, I discovered an additional issue (which can be seen in the
TODOs in this commit), where we are being too conservative around a non-Sendable
class var field. I am going to fix that in the next commit.
rdar://128299305
DISCUSSION: The analysis itself is unable to emit errors. So we achieve the same
functionality by in such cases emitting a partition op that signals to our user
that when they process that partition op they should emit an "unknown pattern"
error at the partition op's instructions.
I have wanted this for a long time, but I never got around to it.
Specifically, I introduced a new composition type called
SILDynamicMergedIsolationInfo that just contains a
SILIsolationInfo. Importantly, whenever one merges a SILIsolationInfo with
another SILIsolationInfo, one gets back a SILDynamicMergedIsolationInfo.
The reason why I am doing this is that we drop nonisolated(unsafe) when merging
so I want to ensure that parts of the code that use merging (where the dropping
occurs) and normal SILIsolationInfo where we do not want to merge is
distinguished.
I made sure we match what we get without region isolation by turning off region
isolation in one of the test runs on the test for this.
There is one problem where for non-final classes with nonisolated(unsafe) var
fields, we currently do not properly squelch since I need to do more
infrastructure work. I am going to do that in the next commit.
rdar://128299305
The design change here is that instead of just initializing the regionInfo with
disconnected, we set it as .none and if we see .none, just return a newly
construct disconnected isolation region info when getIsolationRegionInfo() is
called.
This enables us to provide a setIsolationRegionInfo() helper for
RegionAnalysisValueMap::getTrackableValue that does not perform a merge. This is
important since for nonisolated(unsafe), we want to not have nonisolated(unsafe)
propagate through merging. So if we use merging to initialize the internal
regionInfo state of a SILIsolationInfo, we will never have a SILIsolationInfo
with that bit set since it will be lost in the merge. So we need some sort of
other assignment operator. Noting that we should only compute a value's
SILIsolationInfo once in RegionAnalysisValueMap before we cache it in the map,
it made sense to just represent it as an optional that way we can guarantee that
the regionInfo is only ever set exactly once by that routine.
[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 still only parse transferring... but this sets us up for adding the new
'sending' syntax by first validating that this internal change does not mess up
the current transferring impl since we want both to keep working for now.
rdar://128216574
This ensures that we can properly compute isolation for generic types that
conform to AnyActor.
I found this by playing with test cases from the previous commit. We would not
find an actor type for the actor instance isolation and would fall back along an
incorrect path.
rdar://128021548
As part of this I went through how we handled inference and rather than using a
grab-bag getActorIsolation that was confusing to use, I created split APIs for
specific use cases (actor instance, global actor, just an apply expr crossing)
that makes it clearer inside the SILIsolationInfo::get* APIs what we are
actually trying to model. I found a few issues as a result and fixed most of
them if they were small. I also fixed one bigger one around computed property
initializers in the next commit. There is a larger change I didn't fix around allowing function
ref/partial_apply with isolated self parameters have a delayed flow sensitive
actor isolation... this will be fixed in a subsequent commit.
This also fixes a bunch of cases where we were printing actor-isolated instead
of 'self' isolated.
rdar://127295657
The new `swift-driver` seems to enqueue a `wrapmodule` job which uses
the given `module-name` to form the output file name when not doing
optmizations (seems to happen only for `-Onone` in my testing). Since the
CMake functions macros are using the module name also as the explicit output
name, this clashes and ends up in an unhelpful error message from the driver.
```
SwiftDriverExecution/MultiJobExecutor.swift:207: Fatal error: multiple
producers for output ... SwiftCompilerSources/Basic.o: Wrapping Swift
module Basic & Compiling Basic SourceLoc.swift
```
This was reported in https://forums.swift.org/t/debug-swift-build-fails/71380
The changes use a different output object name (by using `.object.o`
suffix) which does not clash with what the `swift-driver` does
automatically. The code around the output objects and the static
libraries have to change slightly to handle this case.
Additionally, the resulting library when in `Debug` is now declaring its
dependency on `swiftSwiftOnoneSupport`, to avoid linking errors when the
libraries are used in the final binaries.
Debug mode seems to enable PURE_BRIDGING_MODE, which seems to skip
transitively including some C headers that files like
`Utilities/Test.swift` depend on. To avoid errors building, add the
missing include in a new `#else` branch.
I think CI will not test the `Debug` mode, so the only thing that it can prove is
that these changes do not break the `Release` mode.
Changes in this CR add part of the, Swift based, Autodiff specific
closure specialization optimization pass. The pass does not modify any
code nor does it even exist in any of the optimization pipelines. The
rationale for pushing this partially complete optimization pass upstream
is to keep up with the breaking changes in the underlying Swift based
compiler infrastructure.
