Just making this easier to debug. Previously, one would have to go into the
debugger to figure out what the issue is. Now upon failure, one just gets the
exact instruction that needs to be added.
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.
access level for optimization: `public`. It requires an extra check for
the actual access level that was declared when determining serialization
since the behavior should be different.
This PR sets its effective access level to `package` as originally defined,
updates call sites to make appropriate acces level comparisons, and removes
`package` specific checks.
Even if the final pattern ends up consuming the value, the match itself
must be nondestructive, because any match condition could fail and cause
us to have to go back to the original aggregate. For copyable values,
we can always copy our way out of consuming operations, but we don't
have that luxury for noncopyable types, so the entire match operation
has to be done as a borrow.
For address-only enums, this requires codifying part of our tag layout
algorithm in SIL, namely that an address-only enum will never use
spare bits or other overlapping storage for the enum tag. This allows
us to assume that `unchecked_take_enum_data_addr` is safely non-side-
effecting and match an address-only noncopyable enum as a borrow.
I put TODOs to remove defensive copies from various parts of our
copyable enum codegen, as well as to have the instruction report
its memory behavior as `None` when the projection is nondestructive,
but this disturbs SILGen for existing code in ways SIL passes aren't
yet ready for, so I'll leave those as is for now.
This patch is enough to get simple examples of noncopyable enum switches
to SILGen correctly. Additional work is necessary to stage in the binding
step of the pattern match; for a consuming switch, we'll need to end
the borrow(s) and then reproject the matched components so we can
consume them moving them into the owned bindings. The move-only checker
also needs to be updated because it currently always tries to convert
a switch into a consuming operation.
Optionally, the dependency to the initialization of the global can be specified with a dependency token `depends_on <token>`.
This is usually a `builtin "once"` which calls the initializer for the global variable.
The instruction exists to express the atomic initialization of a
tuple-typed address in address-lowered mode. In opaque-values mode,
this is expressed as an assign/store.
Previously, mayRequirePackMetadata only considered whether a type
involved a pack. That failed to account for the case of outlined value
functions that require pack metadata when the type involves a pack in
its layout. Here, mayRequirePackMetadata now considers also whether the
layout corresponding to a type involves a pack.
rdar://119829826
Some notes:
This is not emitted by SILGen. This is just intended to be used so I can write
SIL test cases for transfer non sendable. I did this by adding an
ActorIsolationCrossing field to all FullApplySites rather than adding it into
the type system on a callee. The reason that this makes sense from a modeling
perspective is that an actor isolation crossing is a caller concept since it is
describing a difference in between the caller's and callee's isolation. As a
bonus it makes this a less viral change.
For simplicity, I made it so that the isolation is represented as an optional
modifier on the instructions:
apply [callee_isolation=XXXX] [caller_isolation=XXXX]
where XXXX is a printed representation of the actor isolation.
When neither callee or caller isolation is specified then the
ApplyIsolationCrossing is std::nullopt. If only one is specified, we make the
other one ActorIsolation::Unspecified.
This required me to move ActorIsolationCrossing from AST/Expr.h ->
AST/ActorIsolation.h to work around compilation issues... Arguably that is where
it should exist anyways so it made sense.
rdar://118521597
* `alloc_vector`: allocates an uninitialized vector of elements on the stack or in a statically initialized global
* `vector`: creates an initialized vector in a statically initialized global
SILGen optimizes creation of closure down a thin function when it has
no captures:
%1 = thin_to_thick_function %0
: $@convention(thin) () -> () to $@callee_guaranteed () -> ()
ValueOwnership.cpp has a sketchy optimization that treat the result
like a trivial type, even though it is not:
> CONSTANT_OWNERSHIP_INST(None, ThinToThickFunction)
commit 8c5737d1d5
Date: Fri Oct 23 15:12:18 2020 -0700
[ownership] Change thin_to_thick function to always produce a none value.
This creates a mismatch between the SILType and the SILValue
ownership. This is not a coherent design--we have a similar problem
with enums which is endlessly buggy--but reverting the decision will
be hard. Instead, I'll hack the memory verifier to silence this case.
Fixes rdar://115735132 (Lifetime verifier error with opaque return
type and closure)
To verify if a function may read from an indirect argument, don't use AliasAnalysis.
Instead use the CalleeCache to get the list of callees of an apply instruction.
Then use a simple call-back into the swift Function to check if a callee has any relevant memory effect set.
This avoids a dependency from SIL to the Optimizer.
It fixes a linker error when building some unit tests in debug.
`getRootLocalArchetypeDef()` may return a `PlaceholderValue` which makes the
`cast<SingleValueInstruction>` invalid. Use `dyn_cast` and then update the
callers of `getRootLocalArchetypeDefInst()` to handle `nullptr`.
I also included changes to the rest of the SIL optimizer pipeline to ensure that
the part of the optimizer pipeline before we lower tuple_addr_constructor (which
is right after we run TransferNonSendable) work as before.
The reason why I am doing this is that this ensures that diagnostic passes can
tell the difference in between:
```
x = (a, b, c)
```
and
```
x.0 = a
x.1 = b
x.2 = c
```
This is important for things like TransferNonSendable where assigning over the
entire tuple element is treated differently from if one were to initialize it in
pieces using projections.
rdar://117880194
Introduce SILGen support for reabstractions thunks that change the
error, between indirect and direct errors as well as conversions
amongst error types (e.g., from concrete to `any Error`).
This commit just introduces the instruction. In a subsequent commit, I am going
to add support to SILGen to emit this. This ensures that when we assign into a
tuple var we initialize it with one instruction instead of doing it in pieces.
The problem with doing it in pieces is that when one is emitting diagnostics it
looks semantically like SILGen actually is emitting code for initializing in
pieces which could be an error.
Unlike in regular swift, The class_method instruction references the specialized version of a class method.
This must be handled in ReabstractionInfo: it needs to work without a concrete callee SIL function.
Also, the SILVerifier must handle the case that a class_method instruction references a specialized method.
This instruction was given forwarding ownership in the original OSSA
implementation. That will obviously lead to memory leaks. Remove
ownership from this instruction and verify that it is never used for
non-trivial types.
I think from SIL's perspective, it should only worry about whether the
type is move-only. That includes MoveOnlyWrapped SILTypes and regular
types that cannot be copied.
Most of the code querying `SILType::isPureMoveOnly` is in SILGen, where
it's very likely that the original AST type is sitting around already.
In such cases, I think it's fine to ask the AST type if it is
noncopyable. The clarity of only asking the ASTType if it's noncopyable
is beneficial, I think.
KeyPath's getter/setter/hash/equals functions have their own calling
convention, which receives generic arguments and embedded indices from a
given KeyPath argument buffer.
The convention was previously implemented by:
1. Accepting an argument buffer as an UnsafeRawPointer and casting it to
indices tuple pointer in SIL.
2. Bind generic arguments info from the given argument buffer while emitting
prologue in IRGen by creating a new forwarding thunk.
This 2-phase lowering approach was not ideal, as it blocked KeyPath
projection optimization [^1], and also required having a target arch
specific signature lowering logic in SIL-level [^2].
This patch centralizes the KeyPath accessor calling convention logic to
IRGen, by introducing `@convention(keypath_accessor_XXX)` convention in
SIL and lowering it in IRGen. This change unblocks the KeyPath projection
optimization while capturing subscript indices, and also makes it easier
to support WebAssembly target.
[^1]: https://github.com/apple/swift/pull/28799
[^2]: https://forums.swift.org/t/wasm-support/16087/21
This instructions marks the point where all let-fields of a class are initialized.
This is important to ensure the correctness of ``ref_element_addr [immutable]`` for let-fields,
because in the initializer of a class, its let-fields are not immutable, yet.
Codegen is the same, but `begin_dealloc_ref` consumes the operand and produces a new SSA value.
This cleanly splits the liferange to the region before and within the destructor of a class.
