Sugared GenericTypeParamTypes point to GenericTypeParamDecls,
allowing the name of the parameter as written by the user to be
recovered. Canonical GenericTypeParamTypes on the other hand
only store a depth and index, without referencing the original
declaration.
When printing SIL, we wish to output the original generic parameter
names, even though SIL only uses canonical types. Previously,
we used to accomplish this by mapping the generic parameter to an
archetype and printing the name of the archetype. This was not
adequate if multiple generic parameters mapped to the same
archetype, or if a generic parameter was mapped to a concrete type.
The new approach preserves the original sugared types in the
GenericEnvironment, adding a new GenericEnvironment::getSugaredType()
method.
There are also some other assorted simplifications made possible
by this.
Unfortunately this makes GenericEnvironments use a bit more memory,
however I have more improvements coming that will offset the gains,
in addition to making substitution lists smaller also.
This lets us get to the goal of +0 guaranteed closure contexts. NFC yet, just add the under-the-hood ability for partial_apply instructions producing callee-guaranteed closures to be parsed, printed, and serialized.
It's the same thing as for alloc_ref: the optional [tail_elems ...] attribute specify the tail elements to allocate.
For details see docs/SIL.rst
This feature is needed so that we can allocate a MangedBuffer with alloc_ref_dynamic.
The ManagedBuffer.create() function uses the dynamic self type to create the buffer instance.
The new instructions are: ref_tail_addr, tail_addr and a new attribute [ tail_elems ] for alloc_ref.
For details see docs/SIL.rst
As these new instructions are not generated so far, this is a NFC.
We didn't serialize the their var-info so after de-serializing they were useless anyway.
By not serializing them we get rid of useless debug_value instructions coming from the stdlib.
In future we might change this and fully serialize them.
The new instructions are: ref_tail_addr, tail_addr and a new attribute [ tail_elems ] for alloc_ref.
For details see docs/SIL.rst
As these new instructions are not generated so far, this is a NFC.
This patch is rather large, since it was hard to make this change
incrementally, but most of the changes are mechanical.
Now that we have a lighter-weight data structure in the AST for mapping
interface types to archetypes and vice versa, use that in SIL instead of
a GenericParamList.
This means that when serializing a SILFunction body, we no longer need to
serialize references to archetypes from other modules.
Several methods used for forming substitutions can now be moved from
GenericParamList to GenericEnvironment.
Also, GenericParamList::cloneWithOuterParameters() and
GenericParamList::getEmpty() can now go away, since they were only used
when SILGen-ing witness thunks.
Finally, when printing generic parameters with identical names, the
SIL printer used to number them from highest depth to lowest, by
walking generic parameter lists starting with the innermost one.
Now, ambiguous generic parameters are numbered from lowest depth
to highest, by walking the generic signature, which means test
output in one of the SILGen tests has changed.
This establishes a real def-use relation from the self-parameter to any instruction which uses the dynamic-self type.
This is an addition to what was already done for opened archetypes.
The biggest part of this commit is to rename "OpenedArchetypeOperands" to "TypeDependentOperands" as this name is now more appropriate.
Other than that the change includes:
*) type-dependent operands are now printed after a SIL instruction in a comment as "type-defs:" (for debugging)
*) FuncationSignatureOpts doesn't need to explicitly check if a function doesn't bind dynamic self to remove a dead self metadata argument
*) the check if a function binds dynamic self (used in the inliner) is much simpler now
*) also collect type-dependent operands for ApplyInstBase::SubstCalleeType and not only in the substitution list
*) with this SILInstruction::mayHaveOpenedArchetypeOperands (used in CSE) is not needed anymore and removed
*) add type dependent operands to dynamic_method instruction
Regarding the generated code it should be a NFC.
My earlier patch started serializing SIL basic blocks using the RPOT order. While it works, changing the existing order of BBs during the serialization may be very surprising for users. After all, serialization is not supposed to transform the code.
Therefore, this patch follows a different approach. It uses the existing order of BBs during the serialization. When it deserializes SIL and detects a use of an opened archetype before its definition, it basically introduced a forward definition of this opened archetype. Later on, when the actual definition of the opened archetype is found, it replaces the forward definition. There is a correctness check at the end of a SIL function deserialization, which verifies that there are no forward definitions of opened archetypes left unresoved.
Strict aliasing only applies to memory operations that use strict
addresses. The optimizer needs to be aware of this flag. Uses of raw
addresses should not have their address substituted with a strict
address.
Also add Builtin.LoadRaw which will be used by raw pointer loads.
Serialize SIL basic blocks in the RPOT order to make sure that instructions defining open archetypes are serialized before instructions using those opened archetypes.
Till now there was no way in SIL to explicitly express a dependency of an instruction on any opened archetypes used by it. This was a cause of many errors and correctness issues. In many cases the code was moved around without taking into account these dependencies, which resulted in breaking the invariant that any uses of an opened archetype should be dominated by the definition of this archetype.
This patch does the following:
- Map opened archetypes to the instructions defining them, i.e. to open_existential instructions.
- Introduce a helper class SILOpenedArchetypesTracker for creating and maintaining such mappings.
- Introduce a helper class SILOpenedArchetypesState for providing a read-only API for looking up available opened archetypes.
- Each SIL instruction which uses an opened archetype as a type gets an additional opened archetype operand representing a dependency of the instruction on this archetype. These opened archetypes operands are an in-memory representation. They are not serialized. Instead, they are re-constructed when reading binary or textual SIL files.
- SILVerifier was extended to conduct more thorough checks related to the usage of opened archetypes.
Till now there was no way in SIL to explicitly express a dependency of an instruction on any opened archetypes used by it. This was a cause of many errors and correctness issues. In many cases the code was moved around without taking into account these dependencies, which resulted in breaking the invariant that any uses of an opened archetype should be dominated by the definition of this archetype.
This patch does the following:
- Map opened archetypes to the instructions defining them, i.e. to open_existential instructions.
- Introduce a helper class SILOpenedArchetypesTracker for creating and maintaining such mappings.
- Introduce a helper class SILOpenedArchetypesState for providing a read-only API for looking up available opened archetypes.
- Each SIL instruction which uses an opened archetype as a type gets an additional opened archetype operand representing a dependency of the instruction on this archetype. These opened archetypes operands are an in-memory representation. They are not serialized. Instead, they are re-constructed when reading binary or textual SIL files.
- SILVerifier was extended to conduct more thorough checks related to the usage of opened archetypes.
Till now there was no way in SIL to explicitly express a dependency of an instruction on any opened archetypes used by it. This was a cause of many errors and correctness issues. In many cases the code was moved around without taking into account these dependencies, which resulted in breaking the invariant that any uses of an opened archetype should be dominated by the definition of this archetype.
This patch does the following:
- Map opened archetypes to the instructions defining them, i.e. to open_existential instructions.
- Introduce a helper class SILOpenedArchetypesTracker for creating and maintaining such mappings.
- Introduce a helper class SILOpenedArchetypesState for providing a read-only API for looking up available opened archetypes.
- Each SIL instruction which uses an opened archetype as a type gets an additional opened archetype operand representing a dependency of the instruction on this archetype. These opened archetypes operands are an in-memory representation. They are not serialized. Instead, they are re-constructed when reading binary or textual SIL files.
- SILVerifier was extended to conduct more thorough checks related to the usage of opened archetypes.
Now that ObjC types can be generic, we need to satisfy the type system by plumbing pseudogeneric parameters through func-to-block invocation thunks. Fixes rdar://problem/26524763.
...in code that I wrote. The integrated REPL, deprecated though it may
be, does not have an associated DeclContext because its SourceFile is
not considered complete. (The proper LLDB REPL does not suffer from
this problem because they use a new SourceFile for every block of
input.)
Elsewhere, tighten up code that may have hit similar bugs, though we
haven't seen anything hit these yet.
rdar://problem/26476281
Also, mark witness thunks for [fragile] witnesses as [fragile].
This allows us to serialize the witness table, as well as any thunks
for witnesses declared @_transparent and @inline(__always).
If a thunk is referenced from two different functions, the thunk inherits
the fragile attribute from the first function that forced it to be emitted.
This is wrong, in case the first function might not be fragile, while
the second one is. Copying the fragile attribute to an existing thunk when
checking if it has already been emitted is also wrong, because the thunk
might reference another thunk, and so on.
The correct fix is to have SIL serialization serialize the transitive
closure of all fragile functions and thunks referenced from fragile
functions. Re-work SIL function serialization to use a worklist so that
we can do this.
Part of https://bugs.swift.org/browse/SR-267.
This was mistakenly reverted in an attempt to fix buildbots.
Unfortunately it's now smashed into one commit.
---
Introduce @_specialize(<type list>) internal attribute.
This attribute can be attached to generic functions. The attribute's
arguments must be a list of concrete types to be substituted in the
function's generic signature. Any number of specializations may be
associated with a generic function.
This attribute provides a hint to the compiler. At -O, the compiler
will generate the specified specializations and emit calls to the
specialized code in the original generic function guarded by type
checks.
The current attribute is designed to be an internal tool for
performance experimentation. It does not affect the language or
API. This work may be extended in the future to add user-visible
attributes that do provide API guarantees and/or direct dispatch to
specialized code.
This attribute works on any generic function: a freestanding function
with generic type parameters, a nongeneric method declared in a
generic class, a generic method in a nongeneric class or a generic
method in a generic class. A function's generic signature is a
concatenation of the generic context and the function's own generic
type parameters.
e.g.
struct S<T> {
var x: T
@_specialize(Int, Float)
mutating func exchangeSecond<U>(u: U, _ t: T) -> (U, T) {
x = t
return (u, x)
}
}
// Substitutes: <T, U> with <Int, Float> producing:
// S<Int>::exchangeSecond<Float>(u: Float, t: Int) -> (Float, Int)
---
[SILOptimizer] Introduce an eager-specializer pass.
This pass finds generic functions with @_specialized attributes and
generates specialized code for the attribute's concrete types. It
inserts type checks and guarded dispatch at the beginning of the
generic function for each specialization. Since we don't currently
expose this attribute as API and don't specialize vtables and witness
tables yet, the only way to reach the specialized code is by calling
the generic function which performs the guarded dispatch.
In the future, we can build on this work in several ways:
- cross module dispatch directly to specialized code
- dynamic dispatch directly to specialized code
- automated specialization based on less specific hints
- partial specialization
- and so on...
I reorganized and refactored the optimizer's generic utilities to
support direct function specialization as opposed to apply
specialization.
Temporarily reverting @_specialize because stdlib unit tests are
failing on an internal branch during deserialization.
This reverts commit e2c43cfe14, reversing
changes made to 9078011f93.
This attribute can be attached to generic functions. The attribute's
arguments must be a list of concrete types to be substituted in the
function's generic signature. Any number of specializations may be
associated with a generic function.
This attribute provides a hint to the compiler. At -O, the compiler
will generate the specified specializations and emit calls to the
specialized code in the original generic function guarded by type
checks.
The current attribute is designed to be an internal tool for
performance experimentation. It does not affect the language or
API. This work may be extended in the future to add user-visible
attributes that do provide API guarantees and/or direct dispatch to
specialized code.
This attribute works on any generic function: a freestanding function
with generic type parameters, a nongeneric method declared in a
generic class, a generic method in a nongeneric class or a generic
method in a generic class. A function's generic signature is a
concatenation of the generic context and the function's own generic
type parameters.
e.g.
struct S<T> {
var x: T
@_specialize(Int, Float)
mutating func exchangeSecond<U>(u: U, _ t: T) -> (U, T) {
x = t
return (u, x)
}
}
// Substitutes: <T, U> with <Int, Float> producing:
// S<Int>::exchangeSecond<Float>(u: Float, t: Int) -> (Float, Int)
This instruction creates a "virtual" address to represent a property with a behavior that supports definite initialization. The instruction holds references to functions that perform the initialization and 'set' logic for the property. It will be DI's job to rewrite assignments into this virtual address into calls to the initializer or setter based on the initialization state of the property at the time of assignment.
Only declarations of whitelisted pre-specializations from with public linkage need
to be serialized as they will be used by UsePrespecializations pass during -Onone
compilations to check for availability of concrete pre-specializations.
The bodies of these functions are not required as they cannot be used anyways,
because they may refer to symbols with non-public linkage.
Pre-specializations were only used by Onone builds, but were kept inside the standard library dylyb anyways. This commit moves all the pre-specializations into a dedicated Swift module and a dynamic library, which are only used by Onone builds.
This reduces the code size of libswiftCore.dylib by 5%.
This is effectively fallout from 36a44cf3 where we switched representations of DeclID
and friends, but when I went to add ++ and -- to llvm::PointerEmbeddedInt I realized
it wasn't really intended to be mutated in place, i.e. it doesn't make sense to
increment a DeclID. Represent counters as plain integers instead that get converted
to DeclID when they are first used.
The two types are nearly identical, and Fixnum is only in the Swift branches of LLVM,
not in mainline LLVM.
I do want to add ++ to PointerEmbeddedInt and fix some of this ugliness, but that'll
have to go through LLVM review, so it might take a bit.
As part of SE-0022, introduce an 'objc_selector' encoding for string
literals that places the UTF-8 string literal into the appropriate
segment for uniquing of Objective-C selector names.
As there are no instructions left which produce multiple result values, this is a NFC regarding the generated SIL and generated code.
Although this commit is large, most changes are straightforward adoptions to the changes in the ValueBase and SILValue classes.
