It is valid to leak a value on paths into dead-end regions.
Specifically, it is valid to leak an `alloc_box`. Thus, "final
releases" in dead-end regions may not destroy the box and consequently
may not release its contents. Therefore it's invalid to lower such final
releases to `dealloc_stack`s, let alone `destroy_addr`s. The in-general
invalidity of that transformation results in miscompiling whenever a box
is leaked and its projected address is used after such final releases.
Fix this by not treating final releases as boundary markers of the
`alloc_box` and not lowering them to `destroy_addr`s and
`dealloc_stack`s.
rdar://158149082
The rewrite was missing the intentional omission of `dealloc_stack`s
corresponding to `[dead_end]` `dealloc_box`es. Add the necessary
bridging to get to parity with the original.
Without this check, `dealloc_box [dead_end]`s are promoted to
`dealloc_stack`s but the memory projected out of such `alloc_box`s need
not be valid.
rdar://159271158
After computing side effects, we also remove any global or argument effects that
are computed to happen, but are defined not to, based on the effect attribute.
This allows us to compute the deinit_barrier effect for such functions, fixing
the test case here: rdar://155870190
This supercedes #38324.
Extend temporary allocations (sink dealloc_stacks) initialized by a store_borrow
across lifetime dependent uses.
Fixes rdar://143159873 ([nonescapable] extend rvalue lifetimes when they are the source of a dependency)
Look through `@_rawLayout` projections, which "type casts" a raw-layout struct to it's content, which must match the like-type of the raw-layout, e.g.
```
@_rawLayout(like: T)
struct S {}
%2 = builtin "addressOfRawLayout"<S>(%1 : $*S) : $Builtin.RawPointer
%3 = pointer_to_address %2 to $*T
```
Previously, AutoDiff closure specialization pass was triggered only on
VJPs containing single basic block. However, the pass logic allows
running on arbitrary VJPs. This PR enables the pass for all VJPs
unconditionally. So, if the pullback corresponding to multiple-BB VJP
accepts some closures directly as arguments, these closures might become
specialized by the pass. Closures passed via payload of branch tracing
enum are not specialized - this is subject for future changes.
The PR contains several commits.
1. The thing named "call site" in the code is partial_apply of pullback
corresponding to the VJP. This might appear only once, so we drop
support for multiple "call sites".
2. Enhance existing SILOptimizer tests for the pass.
3. Add validation-tests for single basic block case.
4. The change itself - delete check against single basic block.
5. Add validation-tests for multiple basic block case.
6. Add SILOptimizer tests for multiple basic block case.
Bypess lifetime dependence diagnostics completely for immortal values. We did
not do this initially because we wanted to potentially consider a value with a missing
dependency to mean that it could not escape the current function. But now we use
`Void` as a stand-in for immortal values.
This is needed for reassigning a Span/MutableSpan to an empty, immortal
Span:
func inoutToImmortal(_ s: inout RawSpan) {
let tmp = RawSpan(_unsafeBytes: UnsafeRawBufferPointer(start: nil, count: 0))
s = _overrideLifetime(tmp, borrowing: ())
}
Fixes rdar://152572002 ([GH:#81976] Cannot reinitialize inout parameter of type
`MutableSpan<T>?`)
This is needed after running the SSAUpdater, because the updater can insert unnecessary phis in the middle of the original liverange of a value.
Fixes an ownership error.
rdar://153229472
switch_enum_addr was being treated like a store instruction, which killed
the local enum's liveness. This could result local variable analysis reporting a
shorter lifetime for the local.
This showed up as a missing exclusivity diagnostic because an access scope was
not fully extended across a dependent local variable of Optional type.
This prevents the following pattern from miscompiling. It should report an exclusivity violation:
var mutableView = getOpaqueOptionalView(holder: &holder)!
mutate(&holder)
mutableView.modify()
Fixes rdar://151231236 ([~Escapable] Missing 'overlapping acceses' error when
called from client code, but exact same code produces error in same module)
This peephole optimization also combined the instructions if the `destroy_addr` appears before the `copy_addr` in the same basic block.
https://github.com/swiftlang/swift/issues/82466
rdar://154236276
This fix enables exclusive access to a MutableSpan created from an UnsafeMutablePointer.
The compiler has a special case that allows MutableSpan to depend on a mutable
pointer *without* extending that pointer's access scope. That lets us implement
standard library code like this:
mutating public func extracting(droppingLast k: Int) -> Self {
//...
let newSpan = unsafe Self(_unchecked: _pointer, byteCount: newCount)
return unsafe _overrideLifetime(newSpan, mutating: &self)
Refine this special case so that is does not apply to inout parameters where the
programmer has an expectation that the unsafe pointer is not copied when being
passed as an argument. Now, we safely get an exclusivity violation when creating
two mutable spans from the same pointer field:
@lifetime(&self)
mutating func getSpan() -> MutableSpan<T> {
let span1 = makeMutableSpan(&self.pointer)
let span2 = makeMutableSpan(&self.pointer) // ERROR: overlapping access
return span1
}
If we don't fix this now, it will likely be source breaking in the future.
Fixes rdar://153745332 (Swift allows constructing two MutableSpans to the same underlying pointer)
This pass replaces `alloc_box` with `alloc_stack` if the box is not escaping.
The original implementation had some limitations. It could not handle cases of local functions which are called multiple times or even recursively, e.g.
```
public func foo() -> Int {
var i = 1
func localFunction() { i += 1 }
localFunction()
localFunction()
return i
}
```
The new implementation (done in Swift) fixes this problem with a new algorithm.
It's not only more powerful, but also simpler: the new pass has less than half lines of code than the old pass.
The pass is invoked in the mandatory pipeline and later in the optimizer pipeline.
The new implementation provides a module-pass for the mandatory pipeline (whereas the "regular" pass is a function pass).
This is required because the mandatory pass needs to remove originals of specialized closures, which cannot be done from a function-pass.
In the old implementation this was done with a hack by adding a semantic attribute and deleting the function later in the pipeline.
I still kept the sources of the old pass for being able to bootstrap the compiler without a host compiler.
rdar://142756547
Originally this was a "private" utility for the ClosureSpecialization pass.
Now, make it a general utility which can be used for all kind of function specializations.
This is a safer API than using
```
let argIdx = applySite.calleeArgumentIndex(of: op)
let arg = callee.arguments[argIdx]
```
because there is no potential misuse of the index.
When extending a coroutine, handle the end_borrow instruction used to end a
coroutine lifetime at a dead-end block.
Fixes rdar://153479358 (Compiler crash when force-unwrapping optional ~Copyable type)
LifetimeDependenceInsertion inserts mark_dependence on token result of a begin_apply
when it yields a lifetime dependent value. When such a begin_apply gets inlined,
the inliner can crash because of the remaining uses of the token result.
Fix this by inserting mark_dependence on parameter operands that are lifetime dependence sources
and deleting the mark_dependence on token results in the inliner.
Fixes rdar://151568816
When extending an access scope over a coroutines, instead of simply
considering the lifetime of the coroutine scope, recurse through all
uses of yielded values. They may be copyable, non-Escapable values
that depend on the coroutine operand.
Fixes rdar://152693622 (Extend coroutines over copied yields)
* re-implement the pass in swift
* support alloc_stack liveranges which span over multiple basic blocks
* support `load`-`store` pairs, copying from the alloc_stack (in addition to `copy_addr`)
Those improvements help to reduce temporary stack allocations, especially for InlineArrays.
rdar://151606382
Add support for diagnosing calls to closures that return a generic
non-Escapable result.
Closures do not yet model lifetime dependencies. The diagnostics have
a special case for handling nonescaple result with no lifetime
dependence, but it previously only handled direct results. This fix handles
cases like the following:
func callIndirectClosure<T>(f: () -> NE<T>) -> NE<T> {
f()
}
Fixes rdar://134318846 ([nonescapable] diagnose function types with nonescapable results)
This mostly makes it easier to test dependency corner cases. The analysis still
doesn't recognize UnsafeRawPointer.init(), so regular users still need to use
_overrideLifetime.
Fixes rdar://137608270 ([borrows] Add Builtin.addressof() support
for @addressable arguments)
Introduce a new pass MandatoryTempRValueElimination, which works as the original TempRValueElimination, except that it does not remove any alloc_stack instruction which are associated with source variables.
Running this pass at Onone helps to reduce copies of large structs, e.g. InlineArrays or structs containing InlineArrays.
Copying large structs can be a performance problem, even at Onone.
rdar://151629149
* Move the mutating APIs into Context.swift, because SIL can only be mutated through a MutatingContext
* move the `baseOperand` and `base` properties from the instruction classes to the `MarkDependenceInstruction` protocol
* add `valueOrAddressOperand` and `valueOrAddress` in the `MarkDependenceInstruction` protocol
