This is a large patch; I couldn't split it up further while still
keeping things working. There are four things being changed at
once here:
- Places that call SILType::isAddressOnly()/isLoadable() now call
the SILFunction overload and not the SILModule one.
- SILFunction's overloads of getTypeLowering() and getLoweredType()
now pass the function's resilience expansion down, instead of
hardcoding ResilienceExpansion::Minimal.
- Various other places with '// FIXME: Expansion' now use a better
resilience expansion.
- A few tests were updated to reflect SILGen's improved code
generation, and some new tests are added to cover more code paths
that previously were uncovered and only manifested themselves as
standard library build failures while I was working on this change.
If the initializer witnesses a protocol requirement, the protocol witness
thunk references the allocating entry point directly, since it has no
vtable entry. For this reason we must use alloc_ref_dynamic and not
alloc_ref to ensure the right type of instance is allocated.
Fixes <rdar://problem/49560721>, <https://bugs.swift.org/browse/SR-10285>.
The initialization of an instance property that has an attached
property delegate involves the initial value written on the property
declaration, the implicit memberwise initializer, and the default
arguments to the implicit memberwise initializer. Implement SILGen
support for each of these cases.
There is a small semantic change to the creation of the implicit
memberwise initializer due to SE-0242 (default arguments for the
memberwise initializer). Specifically, the memberwise initializer will
use the original property type for the parameter to memberwise
initializer when either of the following is true:
- The corresponding property has an initial value specified with the
`=` syntax, e.g., `@Lazy var i = 17`, or
- The corresponding property has no initial value, but the property
delegate type has an `init(initialValue:)`.
The specific case that changed is when a property has an initial value
specified as a direct initialization of the delegate *and* the
property delegate type has an `init(initialValue:)`, e.g.,
```swift
struct X {
@Lazy(closure: { ... })
var i: Int
}
```
Previously, this would have synthesized an initializer:
```swift
init(i: Int = ???) { ... }
```
However, there is no way for the initialization specified within the
declaration of i to be expressed via the default argument. Now, it
synthesizes an initializer:
```swift
init(i: Lazy<Int> = Lazy(closure: { ... }))
```
The initializer associated with a lazy property should not be executed
directly, because it is subsumed by code synthesized into the
getter. Generalize the terminology here so we can re-use this path for
property delegate initialization.
Since 'decl' was already a StructDecl, 'decl->getDeclContext()' was
giving us a non-generic module scope context.
Fixes <https://bugs.swift.org/browse/SR-10075>.
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.
When emitting initialization code for stored properties that have
default values within an initializer in an extension containing a
same-type-to-concrete constraint, we can end up with conformances
involving concrete types. Look for that conformance within the module
of the initializer rather than asserting.
Fixes rdar://problem/47723747.
This undoes some of Joe's work in 8665342 to add a guarantee: if an
@objc convenience initializer only calls other @objc initializers that
eventually call a designated initializer, it won't result in an extra
allocation. While Objective-C /allows/ returning a different object
from an initializer than the allocation you were given, doing so
doesn't play well with some very hairy implementation details of
compiled nib files (or NSCoding archives with cyclic references in
general).
This guarantee only applies to
(1) calling `self.init`
(2) where the delegated-to initializer is @objc
because convenience initializers must do dynamic dispatch when they
delegate, and Swift only stores allocating entry points for
initializers in a class's vtable. To dynamically find an initializing
entry point, ObjC dispatch must be used instead.
(It's worth noting that this patch does NOT check that the calling
initializer is a convenience initializer when deciding whether to use
ObjC dispatch for `self.init`. If we ever add peer delegation to
designated initializers, which is totally a valid feature, that should
use static dispatch and therefore should not go through objc_msgSend.)
This change doesn't /always/ result in fewer allocations; if the
delegated-to initializer ends up returning a different object after
all, the original allocation was wasted. Objective-C has the same
problem (one of the reasons why factory methods exist for things like
NSNumber and NSArray).
We do still get most of the benefits of Joe's original change. In
particular, vtables only ever contain allocating initializer entry
points, never the initializing ones, and never /both/ (which was a
thing that could happen with 'required' before).
rdar://problem/46823518
Dynamic replacements are currently written in extensions as
extension ExtendedType {
@_dynamicReplacement(for: replacedFun())
func replacement() { }
}
The runtime implementation allows an implementation in the future where
dynamic replacements are gather in a scope and can be dynamically
enabled and disabled.
For example:
dynamic_extension_scope CollectionOfReplacements {
extension ExtentedType {
func replacedFun() {}
}
extension ExtentedType2 {
func replacedFun() {}
}
}
CollectionOfReplacements.enable()
CollectionOfReplacements.disable()
This is NFC for now, but I plan to build on this to (1) immediately
remove some unnecessary materialization and loads of the base value
and (2) to allow clients to load a borrowed value.
- getAsDeclOrDeclExtensionContext -> getAsDecl
This is basically the same as a dyn_cast, so it should use a 'getAs'
name like TypeBase does.
- getAsNominalTypeOrNominalTypeExtensionContext -> getSelfNominalTypeDecl
- getAsClassOrClassExtensionContext -> getSelfClassDecl
- getAsEnumOrEnumExtensionContext -> getSelfEnumDecl
- getAsStructOrStructExtensionContext -> getSelfStructDecl
- getAsProtocolOrProtocolExtensionContext -> getSelfProtocolDecl
- getAsTypeOrTypeExtensionContext -> getSelfTypeDecl (private)
These do /not/ return some form of 'this'; instead, they get the
extended types when 'this' is an extension. They started off life with
'is' names, which makes sense, but changed to this at some point. The
names I went with match up with getSelfInterfaceType and
getSelfTypeInContext, even though strictly speaking they're closer to
what getDeclaredInterfaceType does. But it didn't seem right to claim
that an extension "declares" the ClassDecl here.
- getAsProtocolExtensionContext -> getExtendedProtocolDecl
Like the above, this didn't return the ExtensionDecl; it returned its
extended type.
This entire commit is a mechanical change: find-and-replace, followed
by manual reformatted but no code changes.
This initial implementation just delegates from SILGenFunctionBuilder to
SILFunctionBuilder. I was going to start transforming verbose uses of
SILFunctionBuilder into higher level APIs on SILGenFunctionBuilder, but I have
run out of time.
This is a good incremental step forward that will let me hide the constructor of
SILFunctionBuilder after I update the optimizer and thus ensure that
SILFunctionBuilder is only used through appropriate composition APIs.
rdar://42301529
ClassDecl::getSuperclass() produces a complete interface type describing the
superclass of a class, including any generic arguments (for a generic type).
Most callers only need the referenced ClassDecl, which is (now) cheaper
to compute: switch those callers over to ClassDecl::getSuperclassDecl().
Fixes an existing test for SR-5993.
This commit does not modify those APIs or their usage. It just:
1. Moves the APIs onto SILFunctionBuilder and makes SILFunctionBuilder a friend
of SILModule.
2. Hides the APIs on SILModule so all users need to use SILFunctionBuilder to
create/destroy functions.
I am doing this in order to allow for adding/removing function notifications to
be enforced via the type system in the SILOptimizer. In the process of finishing
off CallerAnalysis for FSO, I discovered that we were not doing this everywhere
we need to. After considering various other options such as:
1. Verifying after all passes that the notifications were sent correctly and
asserting. Turned out to be expensive.
2. Putting a callback in SILModule. This would add an unnecessary virtual call.
I realized that by using a builder we can:
1. Enforce that users of SILFunctionBuilder can only construct composed function
builders by making the composed function builder's friends of
SILFunctionBuilder (notice I did not use the word subclass, I am talking
about a pure composition).
2. Refactor a huge amount of code in SILOpt/SILGen that involve function
creation onto a SILGenFunctionBuilder/SILOptFunctionBuilder struct. Many of
the SILFunction creation code in question are straight up copies of each
other with small variations. A builder would be a great way to simplify that
code.
3. Reduce the size of SILModule.cpp by 25% from ~30k -> ~23k making the whole
file easier to read.
NOTE: In this commit, I do not hide the constructor of SILFunctionBuilder since
I have not created the derived builder structs yet. Once I have created those in
a subsequent commit, I will hide that constructor.
rdar://42301529
The storage kind has been replaced with three separate "impl kinds",
one for each of the basic access kinds (read, write, and read/write).
This makes it far easier to mix-and-match implementations of different
accessors, as well as subtleties like implementing both a setter
and an independent read/write operation.
AccessStrategy has become a bit more explicit about how exactly the
access should be implemented. For example, the accessor-based kinds
now carry the exact accessor intended to be used. Also, I've shifted
responsibilities slightly between AccessStrategy and AccessSemantics
so that AccessSemantics::Ordinary can be used except in the sorts of
semantic-bypasses that accessor synthesis wants. This requires
knowing the correct DC of the access when computing the access strategy;
the upshot is that SILGenFunction now needs a DC.
Accessor synthesis has been reworked so that only the declarations are
built immediately; body synthesis can be safely delayed out of the main
decl-checking path. This caused a large number of ramifications,
especially for lazy properties, and greatly inflated the size of this
patch. That is... really regrettable. The impetus for changing this
was necessity: I needed to rework accessor synthesis to end its reliance
on distinctions like Stored vs. StoredWithTrivialAccessors, and those
fixes were exposing serious re-entrancy problems, and fixing that... well.
Breaking the fixes apart at this point would be a serious endeavor.
Replace two prominent uses of SubstitutionList, in ConcreteDeclRef and
Witness, with SubstitutionMap. Deal with the myriad places where we
now have substitution maps and need substitution lists (or vice versa)
caused by this change.
Overall, removes ~50 explicit uses of SubstitutionList (of ~400).
This method was not distinguishing in between in_guaranteed and in
parameters. This would cause the curry thunk where this is used to not copy
in_guaranteed parameters before passing in the parameter to the
partial_apply. This can not affect +1 code since the curry thunk will always
have self at +1.
I also refactored code in:
1. SILGenPoly.
2. SILGenProlog.
3. SILGenConstructor.
to use this function instead of their own reimplementations of the same thing.
This should be NFC for +1 code and is tested by test updates when +0 is enabled.
rdar://34222540
This is really an initializer, so use the initializer conventions. Today this is
NFC given the current convention. With +0 arguments, this is a different story
though.
Found by fixing SILGen tests for +0.
rdar://34222540
This patch both makes debug variable information it optional on
alloc_stack and alloc_box instructions, and forced variable
information on debug_value and debug_value_addr instructions. The
change of the interface uncovered a plethora of bugs in SILGen,
SILTransform, and IRGen's LoadableByAddress pass.
Most importantly this fixes the previously commented part of the
DebugInfo/local-vars.swift.gyb testcase.
rdar://problem/37720555
This was exposed by a test in the SILGen filecheck tests that I am updating out
of tree.
What is basically happening here is that we construct an enum from a guaranteed
argument and then return that enum. We need to ensure in such a case that the
return value is still at +1. This commit fixes the issue by using an
ensurePlusOne, ensuring that in normal cases we do not emit an extra copy, but
in a world where we have +0 arguments, ensures we insert a copy before we
return.
rdar://34222540
If the base class defines a no-argument initializer, we allow the
'super.init()' call to be omitted. However, SILGen would generate
invalid SIL if the base class initializer was throwing or failing.
The problem was that we would attempt to insert instructions
after a try_apply or cond_br, which is invalid, because we would
incorrectly save and restore the insertion point when emitting the
delegation. This is incorrect, because a delegation to a throwing
or failing initializer emits new basic blocks.
Fixes <https://bugs.swift.org/browse/SR-5974>,
<rdar://problem/36256335>.
This patch moves the ownership of profiling state from SILGenProfiling
to SILFunction, where it always belonged. Similarly, it moves ownership
of the profile reader from SILGenModule to SILModule.
The refactor sets us up to fix a few outstanding code coverage bugs and
does away with sad hacks like ProfilerRAII. It also allows us to locally
guarantee that a profile counter increment actually corresponds to the
SILFunction at hand.
That local guarantee causes a bugfix to accidentally fall out of this
refactor: we now set up the profiling state for delayed functions
correctly. Previously, we would set up a ProfilerRAII for the delayed
function, but its counter increment would never be emitted :(. This fix
constitutes the only functional change in this patch -- the rest is NFC.
As a follow-up, I plan on removing some dead code in the profiling
logic and fixing a few naming inconsistencies. I've left that for later
to keep this patch simple.
This patch moves the ownership of profiling state from SILGenProfiling
to SILFunction, where it always belonged. Similarly, it moves ownership
of the profile reader from SILGenModule to SILModule.
The refactor sets us up to fix a few outstanding code coverage bugs and
does away with sad hacks like ProfilerRAII. It also allows us to locally
guarantee that a profile counter increment actually corresponds to the
SILFunction at hand.
That local guarantee causes a bugfix to accidentally fall out of this
refactor: we now set up the profiling state for delayed functions
correctly. Previously, we would set up a ProfilerRAII for the delayed
function, but its counter increment would never be emitted :(. This fix
constitutes the only functional change in this patch -- the rest is NFC.
As a follow-up, I plan on removing some dead code in the profiling
logic and fixing a few naming inconsistencies. I've left that for later
to keep this patch simple.
