So far, argument effects were printed in square brackets before the function name, e.g.
```
sil [escapes !%0.**, !%1, %1.c*.v** => %0.v**] @foo : $@convention(thin) (@guaranteed T) -> @out S {
bb0(%0 : $*S, %1 : @guaranteed $T):
...
```
As we are adding more argument effects, this becomes unreadable.
To make it more readable, print the effects after the opening curly brace, and print a separate line for each argument. E.g.
```
sil [ossa] @foo : $@convention(thin) (@guaranteed T) -> @out S {
[%0: noescape **]
[%1: noescape, escape c*.v** => %0.v**]
bb0(%0 : $*S, %1 : @guaranteed $T):
...
```
This is a dedicated instruction for incrementing a
profiler counter, which lowers to the
`llvm.instrprof.increment` intrinsic. This
replaces the builtin instruction that was
previously used, and ensures that its arguments
are statically known. This ensures that SIL
optimization passes do not invalidate the
instruction, fixing some code coverage cases in
`-O`.
rdar://39146527
Include the parent `ModuleDecl` when serializing a `SILFunction` so that it is available on deserialized functions even though the full `DeclContext` is not present. With the parent module always available we can reliably compute whether the `SILFunction` comes from a module that was imported `@_weakLinked`.
Serialize the `DeclContext` member of `SILFunction` so that it can be used to look up the module that a function belongs to in order to compute weak import status.
Resolves rdar://98521248
This is exactly like copy_addr except that it is not viewed from the verifiers
perspective as an "invalid" copy of a move only value. It is intended to be used
in two contexts:
1. When the move checker emits a diagnostic since it could not eliminate a copy,
we still need to produce valid SIL without copy_addr on move only types since we
will hit canonical SIL eventually even if we don't actually codegen the SIL. The
pass can just convert said copy_addr to explicit_copy_addr and everyone is
happy.
2. To implement the explicit copy function for address only types.
This attribute will, in the near future, be used to inform IRGen that a nominal type that conforms to such protocol must have its type metadata always emitted into the binary, regardless of whether it is used/public.
Local types are not ABI, and the only time we care about the mangling here is
when we look them up using the DWARF mangling in debug info, which doesn't
respect @_originallyDefinedIn either.
Fixes https://github.com/apple/swift/issues/59773.
Since I am beginning to prepare for adding real move only types to the language,
I am renaming everything that has to do with copyable types "move only wrapped"
values instead of move only. The hope is this reduces/prevents any confusion in
between the two.
The ObjCMethodLookupTable for protocols was not being serialized and rebuilt on load, so NominalTypeDecl::lookupDirect() on selectors was not working correctly for deserialized types. Correct this oversight.
When a synchronous, actor-isolated declaration witnesses an
asynchronous, not-similarly-isolated requirement, emit an actor hop
within the witness thunk to ensure that we properly enter the context
of the actor.
Fixes#58517 / rdar://92881539.
This class is a perennial source of bugs when we add new type nodes because there is no signal when you forget to update it since it's just comparing record IDs in the switch. Break this class down and add new macros that enforce structural checks and require exhaustiveness at compile time.
This TYPE_LAYOUT(...) macro replaces the usual \c BCRecordLayout coding structure by enforcing structural checks for the definition of deserialization members. If you forget to define a TYPE_LAYOUT(...) for a TYPE(...) there will be a gnarly SFINAE error pointing at it in DeclTypeRecordNodes.def.
This macro pairs with \c DESERIALIZE_TYPE(...) in Deserialization.cpp such that if you forget \c DESERIALIZE_TYPE(...) you will come up with a linker error.
We didn't serialize the ownership kind, which resulted in a miscompile causing an over-release.
The unchecked_ref_cast is the only instruction for which we change the ownership in the optimizer, so that it doesn't match the operand's ownership. Therefore this fix is sufficient for now - we don't need to serialize the ownership of other instructions.
But this is really a design flaw of the `OwnershipForwardingMixin`. It should not allow to set the ownership to arbitrary values.
rdar://92696202
- Add a `[reflection]` bit to `alloc_box` instructions, to indicate that a box
should be allocated with reflection metadata attached.
- Add a `@captures_generics` attribute to SILLayouts, to indicate a type layout
that captures the generic arguments it's substituted with, meaning it can
recreate the generic environment without additional ABI-level arguments, like
a generic partial application can.
This attribute is designed for let-bound variables whose initializing
assignment is synthesized by the compiler. This assignment is
expected to happen at some point before DefiniteInitialization has
run, which is the pass that verifies whether the compiler truly
initialized the variable.
I generally expect that this will never be a user-facing feature, and
that the synthesized assignment happens in SILGen.
Write the real module name for XRefs in swiftmodule files instead of the
exported module name, from `export_as` declarations in module maps.
Swiftmodule files are internal details now, they should represent the
truth.
We keep using the exported module name for the extensions lookup table
as clients should still use the exported name. However we may need to
write both alternatives in the lookup table in the future if extensions
can't be found from clients not importing the exported as module.
rdar://90272035
The main point of this change is to make sure that a shared function always has a body: both, in the optimizer pipeline and in the swiftmodule file.
This is important because the compiler always needs to emit code for a shared function. Shared functions cannot be referenced from outside the module.
In several corner cases we missed to maintain this invariant which resulted in unresolved-symbol linker errors.
As side-effect of this change we can drop the shared_external SIL linkage and the IsSerializable flag, which simplifies the serialization and linkage concept.
This ensures that opened archetypes always inherit any outer generic parameters from the context in which they reside. This matters because class bounds may bind generic parameters from these outer contexts, and without the outer context you can wind up with ill-formed generic environments like
<τ_0_0, where τ_0_0 : C<T>, τ_0_0 : P>
Where T is otherwise unbound because there is no entry for it among the generic parameters of the environment's associated generic signature.
We now schedule conformance emissions in basically the same way
we do for types and declarations, which means that we'll emit them
uniquely in the module file instead of redundantly at every use.
This should produce substantially smaller module files overall,
especially for modules that heavily use generics. It also means
that we can remove all the unfortunate code to support using
different abbrev codes for them in different bitcode blocks.
Requirement lists are now emitted inline in the records that need
them instead of as trailing records. I think this will improve
space usage, but mostly it assists in eliminating the problem
where abbrev codes are shared between blocks.
The main effect of this will be that in IRGen we will use llvm.dbg.addr instead
of llvm.dbg.declare. We must do this since llvm.dbg.declare implies that the
given address is valid throughout the program.
This just adds the instructions/printing/parsing/serialization/deserialization.
rdar://85020571
