Fixes a correctness issue with unsafe addressors: `unsafeAddress` and
`unsafeMutableAddress`. Previously, the resulting `Unsafe[Mutable]Pointer` did
not depend on `self`, meaning that the compiler is allowed to destroy `self`
before any uses of the pointer. This happens to be valid for
`UnsafePointer.pointee` because, in that case, `self` does not have a lifetime
anyway; the correctness burden was on the programmer to use
`withExtendedLifetime` around all uses of `self`.
Now, unsafe addressors can be used for arbitrary `Self` types.
This also enables lifetime dependence diagnostics when the addressor points to a
`~Escapable` type.
Addressors can now be used as an implementation of borrowed properties.
Type annotations for instruction operands are omitted, e.g.
```
%3 = struct $S(%1, %2)
```
Operand types are redundant anyway and were only used for sanity checking in the SIL parser.
But: operand types _are_ printed if the definition of the operand value was not printed yet.
This happens:
* if the block with the definition appears after the block where the operand's instruction is located
* if a block or instruction is printed in isolation, e.g. in a debugger
The old behavior can be restored with `-Xllvm -sil-print-types`.
This option is added to many existing test files which check for operand types in their check-lines.
This generates significantly better code directly out of SILGen, at
the cost of having to reimplement a little bit of the argument-emission
logic to handle default arguments. But it also neatly sidesteps the
problems we have with splitting tuple RValues when the tuple contains
a pack expansion, which will require some significant surgery to RValue
to fix. That, in turn, fixes rdar://121489308.
This is largely a matter of changing the main loop over subst
params in TranslateArguments to use the generators I added,
then plugging back into the general reabstraction infrastructure.
Because we don't have pack coroutines, we're kind of stuck in
the code generation for pack reabstraction: we have to write
+1 r-values into a temporary tuple and then write those tuple
element addresses into the output pack. It's not great. We
also have lifetime problems with things like non-escaping
closures --- we have that problem outside of reabstraction
thunks, too.
Other than that glaring problem, I'm feeling relatively good
about the code here. It's missing some peepholes, but it should
work. But that that's not to say that arity reabstraction works
in general; my attempts to test it have been exposing some
problems elsewhere, and in particular the closure case crashes,
which is really bad. But this gets a few more things working,
and this PR is quite large already.
Although nonescaping closures are representationally trivial pointers to their
on-stack context, it is useful to model them as borrowing their captures, which
allows for checking correct use of move-only values across the closure, and
lets us model the lifetime dependence between a closure and its captures without
an ad-hoc web of `mark_dependence` instructions.
During ownership elimination, We eliminate copy/destroy_value instructions and
end the partial_apply's lifetime with an explicit dealloc_stack as before,
for compatibility with existing IRGen and non-OSSA aware passes.
The old syntax was
@opened("UUID") constraintType
Where constraintType was the right hand side of a conformance requirement.
This would always create an archetype where the interface type was `Self`,
so it couldn't cope with member types of opened existential types.
Member types of opened existential types is now a thing with SE-0309, so
this lack of support prevented writing SIL test cases using this feature.
The new syntax is
@opened("UUID", constraintType) interfaceType
The interfaceType is a type parameter rooted in an implicit `Self`
generic parameter, which is understood to be the underlying type of the
existential.
Fixes rdar://problem/93771238.
`Builtin.once` has type `(Builtin.RawPointer, (Builtin.RawPointer) -> ())`
at Swift level, but lazy global init passes its initializer as `() -> ()`,
so their callee and caller signatures doesn't match.
It was originally convenient for exclusivity optimization to treat
boxes specially. We wanted to know that the 'Box' kind was always
uniquely identified. But that's not really important. And now that
AccessedStorage is being used more generally, the inconsistency is
problematic.
A consistent model is also must easier to understand and explain.
This also make the implementation of the utility simpler and more powerful.
Functional changes:
isRCIdentical will look through mark_dependence and mark_uninitialized.
findReferenceRoot is used consistently everywhere increasing analysis precision.
AccessPath was treating init_enum_data_addr as an address base, which
is not ideal. It should be able to identify the underlying enum object
as the base. This issue was caught by LoadBorrowImmutabilityChecker
during SIL verification.
Instead handle init_enum_data_addr as a access projection that does
not affect the access path. I expect this SIL pattern to disappear
with SIL opaque values, but it still needs to be handled properly
after lowering addresses.
Functionality changes:
- any user of AccessPath now sees enum initialization stores as writes
to the underlying enum object
- SILGen now generates begin/end access markers for enum
initialization patterns. (Originally, we did not "see through"
init_enum_data_addr because we didn't want to generate these
markers, but that behavior was inconsistent and problematic).
Fixes rdar://70725514 fatal error encountered during compilation;
Unknown instruction: init_enum_data_addr)
Fixes <rdar://60046018> assert: (v->getType().isAddress())
in getAccessedAddress.
MemBehavior compares known memory accesses with some arbitrary value,
which may not be an address. However, we should not call utilities
that work on accessed addresses with an arbitrary value.
This assert is a result of very recent changes to gradually make
memory access utilities more type safe and introduce the concept of a
canonical accessed address. In the future, we may even have a wrapper
type for such a thing. In the SIL optimizer, there are several
different notions of what constitutes the base of a memory
access. Mismatches can lead to subtle bugs.
In order to allow this, I've had to rework the syntax of substituted function types; what was previously spelled `<T> in () -> T for <X>` is now spelled `@substituted <T> () -> T for <X>`. I think this is a nice improvement for readability, but it did require me to churn a lot of test cases.
Distinguishing the substitutions has two chief advantages over the existing representation. First, the semantics seem quite a bit clearer at use points; the `implicit` bit was very subtle and not always obvious how to use. More importantly, it allows the expression of generic function types that must satisfy a particular generic abstraction pattern, which was otherwise impossible to express.
As an example of the latter, consider the following protocol conformance:
```
protocol P { func foo() }
struct A<T> : P { func foo() {} }
```
The lowered signature of `P.foo` is `<Self: P> (@in_guaranteed Self) -> ()`. Without this change, the lowered signature of `A.foo`'s witness would be `<T> (@in_guaranteed A<T>) -> ()`, which does not preserve information about the conformance substitution in any useful way. With this change, the lowered signature of this witness could be `<T> @substituted <Self: P> (@in_guaranteed Self) -> () for <A<T>>`, which nicely preserves the exact substitutions which relate the witness to the requirement.
When we adopt this, it will both obviate the need for the special witness-table conformance field in SILFunctionType and make it far simpler for the SILOptimizer to devirtualize witness methods. This patch does not actually take that step, however; it merely makes it possible to do so.
As another piece of unfinished business, while `SILFunctionType::substGenericArgs()` conceptually ought to simply set the given substitutions as the invocation substitutions, that would disturb a number of places that expect that method to produce an unsubstituted type. This patch only set invocation arguments when the generic type is a substituted type, which we currently never produce in type-lowering.
My plan is to start by producing substituted function types for accessors. Accessors are an important case because the coroutine continuation function is essentially an implicit component of the function type which the current substitution rules simply erase the intended abstraction of. They're also used in narrower ways that should exercise less of the optimizer.
This just eliminates -enable-sil-ownership from all target-swift-frontend and
target-swift-emit-silgen RUN lines. Both of those now include
enable-sil-ownership in their expansion.
NFC unless -enable-verify-exclusivity is on.
Boxed (indirect) enum loads are not formal accesses so they should not
have access markers. For a while, we used UnenforcedAccess in this
case to signal to the verifier that this is actually a safe load. But
that isn't needed. The verifier just ignores loads from unidentified
storage instead.
There are two code paths that had this UnenforcedAccess marker. One
was cleaned up, and the other was not. So finish the cleanup and add a
unit test for both SILGen code paths.
The bulk of the changes are to SILGenApply. As we must now evaluate the
payload ArgumentSource to an RValue, we follow the example of subscripts
and lie to the argument emitter. This evaluates arguments at +1 which
can lead to slightly worse codegen at -Onone.
In a previous commit, I banned in the verifier any SILValue from producing
ValueOwnershipKind::Any in preparation for this.
This change arises out of discussions in between John, Andy, and I around
ValueOwnershipKind::Trivial. The specific realization was that this ownership
kind was an unnecessary conflation of the a type system idea (triviality) with
an ownership idea (@any, an ownership kind that is compatible with any other
ownership kind at value merge points and can only create). This caused the
ownership model to have to contort to handle the non-payloaded or trivial cases
of non-trivial enums. This is unnecessary if we just eliminate the any case and
in the verifier separately verify that trivial => @any (notice that we do not
verify that @any => trivial).
NOTE: This is technically an NFC intended change since I am just replacing
Trivial with Any. That is why if you look at the tests you will see that I
actually did not need to update anything except removing some @trivial ownership
since @any ownership is represented without writing @any in the parsed sil.
rdar://46294760
This commit allows the initial switch subject value to be emitted at +0 if we
can emit it that way. As you can imagine since we have +0 normal function
arguments this should tighten up a lot of code around switches on arguments. So
I got to delete a bunch of code in the tests. = ).
Some things to note:
1. If the switch is given a +1 value, we will still try to let it through at +1
until we hit a part of the decision tree where previously we would need to use
TakeOnSuccess. This means that +1 values that go through irrefutable patterns
like tuple splitting should still be emitted at +1.
2. If we are returned an address only type without a cleanup, we copy it and
pass it down at +1 like the old code.
3. I also elided the last ownership check in SILGenPattern in this commit. After
this, there is only 1 specialization for ownership in the swift compiler that is
in Apply emission. Thats my next target.
rdar://29791263
This does not eliminate the entrypoints on SILBuilder yet. I want to do this in
two parts so that it is functionally easier to disentangle changing the APIs
above SILBuilder and changing the underlying instruction itself.
rdar://33440767
Delay allocating the result buffer for an opened Self return until right before it's needed. When a mutating method is invoked on an existential, the Self type won't be opened until late, when the formal access to the mutable value begins. Fixes rdar://problem/43507711.
Most of this patch is just removing special cases for materializeForSet
or other fairly mechanical replacements. Unfortunately, the rest is
still a fairly big change, and not one that can be easily split apart
because of the quite reasonable reliance on metaprogramming throughout
the compiler. And, of course, there are a bunch of test updates that
have to be sync'ed with the actual change to code-generation.
This is SR-7134.
* Make _sanityCheck internal
* Make _debugPrecondition internal
* Make Optional._unsafelyUnwrappedUnchecked internal.
* Make _precondition internal
* Switch Foundation _sanityChecks to assertions
* Update file check tests
* Remove one more _debugPrecondition
* Update Optimization-with-check tests
Now that access marker verification is strict and exhaustive, adjust some code
to handle the extra markers and extra checks produced by -enable-verify-exclusivity.
This causes default arguments to get emitted after formal evaluation of `self`, allowing protocol methods to access the opened `Self` type of an existential receiver. Fixes rdar://problem/39524104
