The `projection` flag indicates that `index_addr` projects an element address from an array base address, as opposed to being used for general pointer arithmetic.
When this flag is set, the result address can only reach the single element at the given index — it is not possible to chain multiple `index_addr` instructions to reach other array elements from the result.
Without this flag, the result may be used as the base of another `index_addr`, allowing arithmetic across element boundaries (e.g. an `index_addr` with index 1 followed by an `index_addr` with index 2 reaches the element at offset 3).
An `index_addr [projection]` is mandatory to go from an array base address to an element - even if it's the first element, i.e. the index is zero.
This means that the optimizer must not remove `index_addr [projection]` with a zero index.
In Embedded Swift, global variables are generally emitted into client
modules. However, if a global variable is @export(interface), we need
to treat it as something we externally reference in client modules. Do
so, by adjusting linkage on deserialization in the same manner as we do
for SIL functions.
Fixes rdar://178238940.
Just like we do with SILFunction, allow a code generation model to be
specified on a SILGlobalVariable and maintain that through the printed
and serialized forms.
The default code generation model for Embedded Swift is "inlinable".
DeferredCodeGen made the default code generation model
"implementation", and there was no spelling for "interface".
Introduce the experimental feature CodeGenerationModel=<model>, which
can be any of those three options. The default remains "inlinable", but
one can now specify "implementation" (which keeps most everything in
SIL) or "interface" (which only keeps the generic things in SIL). The
"interface" mode is more like non-embedded Swift for non-generic
declarations, emitting them into the IR (only) but not SIL. Generic
declarations would remain in SIL.
Implements rdar://172433062.
We cannot use spare bits or other overlapping storage layout tricks with fundamentally
address-only enums, and we can take advantage of this to do borrowing switches or other
in-place projections without copying the value. However, for resilient enums, the
implementation may use spare bit packing, but the type must be handled address-only
outside of its defining module, and we didn't have a way to express that with
borrowing switch. Optimization passes have also been running into problems with the
complexity that we were using `unchecked_take_enum_data_addr` sometimes as a pure
operation. This patch splits the instruction into three:
- `unchecked_inplace_enum_data_addr` represents a nondestructive in-place enum
projection. It is only allowed for enums whose projection operation is
nondestructive.
- `unchecked_take_enum_data_addr` represents a destructive enum projection,
invalidating the enum and leaving the payload to be further consumed.
This matches the current instruction's semantics.
- `unchecked_borrow_enum_data_addr` represents a borrowing enum projection.
The instruction takes a second operand for "scratch" space, which the
enum representation may be copied into in order to avoid invalidating the
enum value, so the result is dependent on the lifetime of both the
original enum and the scratch buffer. This allows for borrowing switches
over resilient enums.
`unchecked_borrow_enum_data_addr` is implemented by taking advantage of the
"address-only enums can't do spare bit optimization" property at runtime.
We inspect the operand type's bitwise-borrowability from its metadata. If
the type is bitwise-borrowable, then we are allowed to bitwise-copy the
enum to the scratch space and apply the projection to the scratch space,
preserving the original value. If the type is not bitwise-borrowable, then
we cannot use spare bit optimization in its layout, so we apply the
projection in-place.
Fixes rdar://174952822.
The `@export(interface)` and `@export(implementation)` attributes
SE-0497 are queried directly on AST nodes in several places within the
SIL pipeline. However, they don't persist when SIL functions are
serialized, meaning that clients of the original module might make
different assumptions about the availability of a given function's
definition.
Represent these attributes in a SIL function (as an optional
CodeGenerationModel), (de-)serialize them into the module, and add a
textual representation as SIL function attributes `[export_interface]`
and `[export_implementation]`.
If the vtable of a class has conformance entries (at negative offsets), we can do the cast by simply loading the witness table from those entries.
rdar://173916206
Conformance entries are used for fast conformance lookup, which doesn't need to query the runtime's conformance lookup table.
A conformance entry specifies if the class conforms or does not conform to a protocol.
At runtime, a type cast instruction to an existential can directly load the witness table pointer from the VTable.
If null, the class does not conform to the protocol.
The logic here would keep reading vtable entries so long as some other
top-level record wasn't hit, then assert that the entry it found was a
vtable entry. However, the list of "top-level record kinds" was ad hoc
and doesn't regularly get updated. Switch to a loop that reads vtable
entries while there are vtable entries, since there's only one kind
here anyway.
Fixes an Embedded Swift crash where a SIL global variable was after the
vtable entries, rdar://171009492.
Since after address lowering, `Borrow` can remain loadable with a known-
layout address-only referent, we need instructions that handle three
forms:
- borrow and referent are both loadable values
- borrow is a value, but referent is address-only
- borrow and referent are both address-only
This new OSSA invariant simplifies many optimizations because they don't have to take care of the corner case of incomplete lifetimes in dead-end blocks.
The implementation basically consists of these changes:
* add the lifetime completion utility
* add a flag in SILFunction which tells optimization that they need to run the lifetime completion utility
* let all optimizations complete lifetimes if necessary
* enable the ownership verifier to check complete lifetimes
These two new invariants eliminate corner cases which caused bugs if optimization didn't handle them.
Also, it will significantly simplify lifetime completion.
The implementation basically consists of these changes:
* add a flag in SILFunction which tells optimization if they need to take care of infinite loops
* add a utility to break infinite loops
* let all optimizations remove unreachable blocks and break infinite loops if necessary
* add verification to check the new SIL invariants
The new `breakIfniniteLoops` utility breaks infinite loops in the control flow by inserting an "artificial" loop exit to a new dead-end block with an `unreachable`.
It inserts a `cond_br` with a `builtin "infinite_loop_true_condition"`:
```
bb0:
br bb1
bb1:
br bb1 // back-end branch
```
->
```
bb0:
br bb1
bb1:
%1 = builtin "infinite_loop_true_condition"() // always true, but the compiler doesn't know
cond_br %1, bb2, bb3
bb2: // new back-end block
br bb1
bb3: // new dead-end block
unreachable
```
Introduce a new optional flag on the alloc_box SIL instruction to mark boxes as
inferred immutable, indicating that static analysis has proven they are never
written to despite having a mutable type.
The flag is preserved through serialization/deserialization and properly printed/parsed in textual SIL format.
I am doing this to prepare for treating these boxes as being Sendable when they
contain a sendable weak reference.
Introduce a new optional inferred-immutable flag on SILFunctionArgument to mark
closure-captured box parameters that are never written to despite being mutable.
This flag will enable in future commits:
- Marking captured mutable boxes as immutable when interprocedural analysis
proves they are never modified
- Treating these captures as Sendable when they contain Sendable types
- Improving region-based isolation analysis for concurrent code
This complements the inferred-immutable flag on alloc_box by allowing
immutability information to flow through closure boundaries.
Whenever we have a reference to a foreign function/variable in SIL, use
a mangled name at the SIL level with the C name in the asmname
attribute. The expands the use of asmname to three kinds of cases that
it hadn't been used in yet:
* Declarations imported from C headers/modules
* @_cdecl @implementation of C headers/modules
* @_cdecl functions in general
Some code within the SIL pipeline makes assumptions that the C names of
various runtime functions are reflected at the SIL level. For example,
the linking of Embedded Swift runtime functions is done by-name, and
some of those names refer to C functions (like `swift_retain`) and
others refer to Swift functions that use `@_silgen_name` (like
`swift_getDefaultExecutor`). Extend the serialized module format to
include a table that maps from the asmname of functions/variables over
to their mangled names, so we can look up functions by asmname if we
want. These tables could also be used for checking for declarations
that conflict on their asmname in the future. Right now, we leave it
up to LLVM or the linker to do the checking.
`@_silgen_name` is not affected by these changes, nor should it be:
that hidden feature is specifically meant to affect the name at the
SIL level.
The vast majority of test changes are SIL tests where we had expected
to see the C/C++/Objective-C names in the tests for references to
foreign entities, and now we see Swift mangled names (ending in To).
The SIL declarations themselves will have a corresponding asmname.
Notably, the IRGen tests have *not* changed, because we generally the
same IR as before. It's only the modeling at the SIL lever that has
changed.
Another part of rdar://137014448.
This instruction can be used to disable ownership verification on it's result and
will be allowed only in raw SIL.
Sometimes SILGen can produce invalid ownership SSA, that cannot be resolved until
mandatory passes run. We have a few ways to piecewise disable verification.
With unchecked_ownership instruction we can provide a uniform way to disable ownership
verification for a value.
This instruction converts Builtin.ImplicitActor to Optional<any Actor>. In the
process of doing so, it masks out the bits we may have stolen from the witness
table pointer of Builtin.ImplicitActor. The bits that we mask out are the bottom
two bits of the top nibble of the TBI space on platforms that support TBI (that
is bit 60,61 on arm64). On platforms that do not support TBI, we just use the
bottom two tagged pointer bits (0,1).
By using an instruction, we avoid having to represent the bitmasking that we are
performing at the SIL level and can instead just make the emission of the
bitmasking an IRGen detail. It also allows us to move detection if we are
compiling for AArch64 to be an IRGen flag instead of a LangOpts flag.
The instruction is a guaranteed forwarding instruction since we want to treat
its result as a borrowed projection from the Builtin.ImplicitActor.
The asmname attribute allows one to specify the name that will be used
when lowering a given SIL declaration to LLVM IR. It is not currently
exposed in the surface language.
Make sure this attribute round-trips through the parser and
serialization.
Part of rdar://137014448O.
As with SIL functions, track the parent module where a SIL global
variable was originally defined so that we can determine whether we
are outside of its original module for linkage purposes. Use this to
make sure we emit via a weak definition when emitting to a module
other than the originating module.
Fixes rdar://160153163.
use local funcs to implement `defer`, this also fixes several
bugs with that feature, such as it breaking in nonisolated
functions when a default isolation is in effect in the source file.
Change how we compute isolation of local funcs. The rule here is
supposed to be that non-`@Sendable` local funcs are isolated the
same as their enclosing context. Unlike closure expressions, this
is unconditional: in instance-isolated functions, the isolation
does not depend on whether `self` is captured. But the computation
was wrong: it didn't translate global actor isolation between
contexts, it didn't turn parameter isolation into capture isolation,
and it fell through for several other kinds of parent isolation,
causing the compiler to try to apply default isolation instead.
I've extracted the logic from the closure expression path into a
common function and used it for both paths.
The capture computation logic was forcing a capture of the
enclosing isolation in local funcs, but only for async functions.
Presumably this was conditional because async functions need the
isolation for actor hops, but sync functions don't really need it.
However, this was causing crashes with `-enable-actor-data-race-checks`.
(I didn't investigate whether it also failed with the similar
assertion we do with preconcurrency.) For now, I've switched this
to capture the isolated instance unconditionally. If we need to
be more conservative by either only capturing when data-race checks
are enabled or disabling the checks when the isolation isn't captured,
we can look into that.
Fix a bug in capture isolation checking. We were ignoring captures
of nonisolated declarations in order to implement the rule that
permits `nonisolated(unsafe)` variables to be captured in
non-sendable closures. This check needs to only apply to variables!
The isolation of a local func has nothing to do with its sendability
as a capture.
That fix exposed a problem where we were being unnecessarily
restrictive with generic local func declarations because we didn't
consider them to have sendable type. This was true even if the
genericity was purely from being declared in a generic context,
but it doesn't matter, they ought to be sendable regardless.
Finally, fix a handful of bugs where global actor types were not
remapped properly in SILGen.
We are going to need to add more flags to the various checked cast
instructions. Generalize the CastingIsolatedConformances bit in all of
these SIL instructions to an "options" struct that's easier to extend.
Precursor to rdar://152335805.
It derives the address of the first element of a vector, i.e. a `Builtin.FixedArray`, from the address of the vector itself.
Addresses of other vector elements can then be derived with `index_addr`.
Store specialize witness tables in a separate lookup table in the module. This allows that for a normal conformance there can exist the original _and_ a specialized witness table.
Also, add a boolean property `isSpecialized` to `WitnessTable` which indicates whether the witness table is specialized or not.
When performing a dynamic cast to an existential type that satisfies
(Metatype)Sendable, it is unsafe to allow isolated conformances of any
kind to satisfy protocol requirements for the existential. Identify
these cases and mark the corresponding cast instructions with a new flag,
`[prohibit_isolated_conformances]` that will be used to indicate to the
runtime that isolated conformances need to be rejected.
The Protocol field isn't really necessary, because the conformance
stores the protocol. But we do need the substituted subject type
of the requirement, just temporarily, until an abstract conformance
stores its own subject type too.
Introduce a constructor that takes an `llvm::VersionTuple` directly, instead of
needing to spell out `VersionRange::allGTE(<tuple>)` which is unnecessarily
verbose.
This patch adds support for serialization of debug value instructions. Enablement is currently gated behind the -experimental-serialize-debug-info flag.
Previously, debug_value instructions were lost during serialization. This made it harder to debug cross module inlined functions.
The problem with `is_escaping_closure` was that it didn't consume its operand and therefore reference count checks were unreliable.
For example, copy-propagation could break it.
As this instruction was always used together with an immediately following `destroy_value` of the closure, it makes sense to combine both into a `destroy_not_escaped_closure`.
It
1. checks the reference count and returns true if it is 1
2. consumes and destroys the operand
This is used for synthetic uses like _ = x that do not act as a true use but
instead only suppress unused variable warnings. This patch just adds the
instruction.
Eventually, we can use it to move the unused variable warning from Sema to SIL
slimmming the type checker down a little bit... but for now I am using it so
that other diagnostic passes can have a SIL instruction (with SIL location) so
that we can emit diagnostics on code like _ = x. Today we just do not emit
anything at all for that case so a diagnostic SIL pass would not see any
instruction that it could emit a diagnostic upon. In the next patch of this
series, I am going to add SILGen support to do that.
Erik Eckstein
Doug Gregor
Joe Groff
John McCall
Yuta Saito
Alejandro Alonso
Michael Gottesman
Meghana Gupta
Janat Baig
Anthony Latsis
Slava Pestov
Andrew Trick
Nate Chandler
Kuba Mracek
Allan Shortlidge
Usama Hameed