- The DeclContext versions of these methods have equivalents
on the DeclContext class; use them instead.
- The GenericEnvironment versions of these methods are now
static methods on the GenericEnvironment class. Note that
these are not made redundant by the instance methods on
GenericEnvironment, since the static methods can also be
called with a null GenericEnvironment, in which case they
just assert that the type is fully concrete.
- Remove some unnecessary #includes of ArchetypeBuilder.h
and GenericEnvironment.h. Now changes to these files
result in a lot less recompilation.
Changes:
* Terminate all namespaces with the correct closing comment.
* Make sure argument names in comments match the corresponding parameter name.
* Remove redundant get() calls on smart pointers.
* Prefer using "override" or "final" instead of "virtual". Remove "virtual" where appropriate.
The substitution only replaces archetypes with abstract generic parameters, so no conformance lookup is necessary, and we can provide a "lookup" callback now that just vends abstract conformances.
(Ideally, we'd be able to do this for mapTypeIntoContext too, but we run into problems with generic signatures with same-type constraints on associated types with protocol requirements. Mapping `t_0_0.AssocType` into such a context will require conformance lookup for the concrete type replacement, since same-type Requirements don't preserve the conformances that satisfy the protocol requirements for the same-type relationship.)
Rather than directly using the ArchetypeBuilder associated with a
canonical generic signature, use a canonical GenericEnvironment
associated with that canonical generic signature. This has a few
benefits:
* It's cleaner to not have IRGen working with archetype builders;
GenericEnvironment is the right abstraction for mapping between
dependent types and archetypes for a specific context.
* It helps us separate the archetype builder from a *specific*
set of archetypes. This is an ongoing refactor that is intended to
allow us to re-use archetype builders across different generic
environments.
As part of this, ArchetypeBuilder::substDependentType() has gone away
in favor of GenericEnvironment::mapTypeIntoContext().
Before this commit all code relating to handling arguments in SILBasicBlock had
somewhere in the name BB. This is redundant given that the class's name is
already SILBasicBlock. This commit drops those names.
Some examples:
getBBArg() => getArgument()
BBArgList => ArgumentList
bbarg_begin() => args_begin()
This eliminates all inline creation of SILBasicBlock via placement new.
There are a few reasons to do this:
1. A SILBasicBlock is always created with a parent function. This commit
formalizes this into the SILBasicBlock API by only allowing for SILFunctions to
create SILBasicBlocks. This is implemented via the type system by making all
SILBasicBlock constructors private. Since SILFunction is a friend of
SILBasicBlock, SILFunction can still create a SILBasicBlock without issue.
2. Since all SILBasicBlocks will be created in only a few functions, it becomes
very easy to determine using instruments the amount of memory being allocated
for SILBasicBlocks by simply inverting the call tree in Allocations.
With LTO+PGO, normal inlining can occur if profitable so there shouldn't be
overhead that we care about in shipping compilers.
This is a NFC change, since verification still will be behind the flag. But this
will allow me to move copy_value, destroy_value in front of the
EnableSILOwnership flag and verify via SILGen that we are always using those
instructions.
rdar://28851920
Over the past day or so I have been thinking about how we are going to need to
manage verification of semantic ARC semantics in the pass pipeline. Specifically
the Eliminator pass really needs to be a function pass to ensure that we can
transparently put it at any stage of the optimization pipeline. This means that
just having a flag on the SILVerifier that states whether or not ownership is
enabled is not sufficient for our purposes. Instead, while staging in the SIL
ownership model, we need a bit on all SILFunctions to state whether the function
has been run through the ownership model eliminator so that the verifier can
ensure that we are in a world with "SIL ownership" or in a world without "SIL
ownership", never in a world with only some "SIL ownership" instructions. We
embed this distinction in SIL by creating the concept of a function with
"qualified ownership" and a function with "unqualified ownership".
Define a function with "qualified ownership" as a function that contains no
instructions with "unqualified ownership" (i.e. unqualified load) and a function
with "unqualified ownership" as a function containing such no "ownership
qualified" instructions (i.e. load [copy]) and at least 1 unqualified ownership
instruction.
This commit embeds this distinction into SILFunction in a manner that is
transparently ignored when compiling with SIL ownership disabled. This is done
by representing qualified or unqualified ownership via an optional Boolean on
SILFunction. If the Boolean is None, then SILOwnership is not enabled and the
verifier/passes can work as appropriate. If the Boolean is not None, then it
states whether or not the function has been run through the Ownership Model
Eliminator and thus what invariants the verifier should enforce.
How does this concept flow through the compilation pipeline for functions in a
given module? When SIL Ownership is enabled, all SILFunctions that are produced
in a given module start with "qualified ownership" allowing them to contain SIL
ownership instructions. After the Ownership Model eliminator has run, the
Ownership Model sets the "unqualified" ownership flag on the SILFunction stating
that no more ownership qualified instructions are allowed to be seen in the
given function.
But what about functions that are parsed or are deserialized from another
module? Luckily, given the manner in which we have categories our functions, we
can categorize functions directly without needing to add anything to the parser
or to the deserializer. This is done by enforcing that it is illegal to have a
function with qualified ownership and unqualified ownership instructions and
asserting that functions without either are considered qualified.
rdar://28685236
There's a bit of a hack to deal with generic typealiases, but
overall this makes things more logical.
This is the last big refactoring before we can allow constrained
extensions to make generic parameters concrete. All that remains
is a small set of changes to SIL type lowering, and retooling
some diagnostics in Sema.
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 is the first, and most trivial, usage of the new
GenericSignature::getSubstitutions() method.
Note that getForwardingSubstitutions() now takes a
GenericSignature, which is slightly awkward.
However, this is in line with our goal of 'hollowing out'
GenericParamList by removing knowledge of the finalized
generic requirements.
Also, there is now a new getForwardingSubstitutionMap()
function, which returns an interface type substitution
mapping. This is used in the new getForwardingSubstitutions()
implementation, and all also be used elsewhere later.
Finally, in the SILFunction we now cache the forwarding
substitutions, instead of re-computing them every time.
I doubt this makes a big difference in performance, but
it's a simple enhancement and every little bit helps.
Mostly NFC, this is just plumbing for the next patch.
Note that isNever() returns true for any uninhabited
enum.
It should be generalized so that stuff like (Never, Int)
is also known to be uninhabited, or even to support
generic substitutions that yield uninhabited types,
but for now I really see no reason to go that far, and
the current check for an enum with no cases seems
perfectly adequate.
Two fixes to optimization passes to maintain restrictions about what
[fragile] functions can reference:
- When devirtualizing witness methods, don't devirtualize if the caller
is fragile and the callee is not. This matches existing logic in
class devirtualization.
- When performing generic or function signature specialization, don't
specialize non-fragile functions referenced from fragile functions.
Since @_transparent functions are allowed to call 'static inline'
imported functions, also be sure to mark the foreign-to-native thunk
for such a function as [fragile].
With this patch, the standard library and performance test suite
now build with -enable-resilience.
No new tests for this stuff here -- the existing tests together
with an -enable-resilience build provide coverage.
Closes out <https://bugs.swift.org/browse/SR-267> and
<https://bugs.swift.org/browse/SR-268>.
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)
Similarly to how we've always handled parameter types, we
now recursively expand tuples in result types and separately
determine a result convention for each result.
The most important code-generation change here is that
indirect results are now returned separately from each
other and from any direct results. It is generally far
better, when receiving an indirect result, to receive it
as an independent result; the caller is much more likely
to be able to directly receive the result in the address
they want to initialize, rather than having to receive it
in temporary memory and then copy parts of it into the
target.
The most important conceptual change here that clients and
producers of SIL must be aware of is the new distinction
between a SILFunctionType's *parameters* and its *argument
list*. The former is just the formal parameters, derived
purely from the parameter types of the original function;
indirect results are no longer in this list. The latter
includes the indirect result arguments; as always, all
the indirect results strictly precede the parameters.
Apply instructions and entry block arguments follow the
argument list, not the parameter list.
A relatively minor change is that there can now be multiple
direct results, each with its own result convention.
This is a minor change because I've chosen to leave
return instructions as taking a single operand and
apply instructions as producing a single result; when
the type describes multiple results, they are implicitly
bound up in a tuple. It might make sense to split these
up and allow e.g. return instructions to take a list
of operands; however, it's not clear what to do on the
caller side, and this would be a major change that can
be separated out from this already over-large patch.
Unsurprisingly, the most invasive changes here are in
SILGen; this requires substantial reworking of both call
emission and reabstraction. It also proved important
to switch several SILGen operations over to work with
RValue instead of ManagedValue, since otherwise they
would be forced to spuriously "implode" buffers.
Dead function elimination deletes functions making them zombies. A zombie is a
function whose name only exists because debug info might still refer to it.
However, later passes of specialization might create a new function by the same
name again. We now have a zombie function (which is just an alias to a deleted
method stub) and a function definition.
No test case since I could not reduce this to a small test case.
rdar://24659988
This is something that we have wanted for a long time and will enable us to
remove some hacks from the compiler (i.e. how we determine in the ARC optimizer
that we have "fatalError" like function) and also express new things like
"noarc".
Use malloc/free for allocating/freeing SIL instructions instead of using the BumpPtrAllocator. This allows for memory reuse and significantly reduces the memory footprint of the compiler.
For example, a peak memory usage during a compilation of the standard library and StdlibUnitTest is reduced by 25%-30%. The performance of the compiler seems to be not affected by this change, i.e. no slowdown is measured.
The use-after-free issues reported by build bots are fixed now.
rdar://23303031
The drivers for this change are providing a simpler API to SIL pass
authors, having a more efficient of the in-memory representation,
and ruling out an entire class of common bugs that usually result
in hard-to-debug backend crashes.
Summary
-------
SILInstruction
Old New
+---------------+ +------------------+ +-----------------+
|SILInstruction | |SILInstruction | |SILDebugLocation |
+---------------+ +------------------+ +-----------------+
| ... | | ... | | ... |
|SILLocation | |SILDebugLocation *| -> |SILLocation |
|SILDebugScope *| +------------------+ |SILDebugScope * |
+---------------+ +-----------------+
We’re introducing a new class SILDebugLocation which represents the
combination of a SILLocation and a SILDebugScope.
Instead of storing an inline SILLocation and a SILDebugScope pointer,
SILInstruction now only has one SILDebugLocation pointer. The APIs of
SILBuilder and SILDebugLocation guarantees that every SILInstruction
has a nonempty SILDebugScope.
Developer-visible changes include:
SILBuilder
----------
In the old design SILBuilder populated the InsertedInstrs list to
allow setting the debug scopes of all built instructions in bulk
at the very end (as the responsibility of the user). In the new design,
SILBuilder now carries a "current debug scope" state and immediately
sets the debug scope when an instruction is inserted.
This fixes a use-after-free issue with with SIL passes that delete
instructions before destroying the SILBuilder that created them.
Because of this, SILBuilderWithScopes no longer needs to be a template,
which simplifies its call sites.
SILInstruction
--------------
It is neither possible or necessary to manually call setDebugScope()
on a SILInstruction any more. The function still exists as a private
method, but is only used when splicing instructions from one function
to another.
Efficiency
----------
In addition to dropping 20 bytes from each SILInstruction,
SILDebugLocations are now allocated in the SILModule's bump pointer
allocator and are uniqued by SILBuilder. Unfortunately repeat compiles
of the standard library already vary by about 5% so I couldn’t yet
produce reliable numbers for how much this saves overall.
rdar://problem/22017421
This re-applies commit r31763 with a change to the predicate we
use for determining the linkage of a definition. It turns out we
could have definitions with a Clang body that were still public,
so instead of checking for a Clang body just check if the Clang
declaration is externally visible or not.
Swift SVN r31777
If an external SIL function has a Clang-generated body, I think this
means we have a static function, and we want to use Shared linkage,
not Public.
Add a new flag to SILFunction for this and plumb it through to
appease assertions from SILVerifier.
Swift SVN r31763
This patch implements the pre-specialization for the most popular generic types from the standard library. If there are invocations of generic functions from the standard library in the user-code and the compiler can find the specialized, optimized versions of these functions, then calls of generic functions are simply replaced by the calls of the specialized functions.
This feature is supposed to be used with -Onone to produce much faster (e.g. 5x-10x faster) executables in debug builds without impacting the compile time. In fact, the compile-time is even improved, because IRGen has less work to do. The feature can be considered a light-weight version of the -Odebug, because pre-specialization is limited in scope, but does not have a potentially negative compile-time impact compared to -Odebug. It is planned to enable it by default in the future.
This feature is disabled by default for the time being. It can be enabled by using a hidden flag: -Xllvm -use-prespecialized.
The implementation consists of two logical steps:
- When the standard library is being built, we force a creation of specializations for the most popular generic types from the stdlib, e.g. Arrays of integer and floating point types, Range<Int>, etc. The list of specializations is not fixed and can be easily altered by editing the Prespecialized.swift file, which is responsible for forcing the specialization of generic types (this is simple solution for now, until we have a proper annotation to indicate which specializations of a given generic type or function we want to generate by means of the pre-specialization). These specializations are then optimized and preserved in the stdlib dylib and in the Swift SIL module. The size increase of the stdlib due to creation of pre-specializations is currently about 3%-7%.
- When a user-code is being compiled with -Onone, the compiler would run a generic specializer over the user-code. If there are calls of generic functions from the standard library, the specializer would check if there is an existing specialization matching these invocations. If such a specialization is found, the original call is replaced by the call of this more efficient specialized version.
Swift SVN r30309
The two ways functions are created currently is via the two
SILModule::getOrCreateFunction(). One of the methods, takes in a raw mangled
name and uses that to create the function. The other takes in a SILDeclRef to
generate the mangled name. Most function emission (besides some thunk creation
functions) goes through the latter. For now we update the map there. This is ok,
since this map will only be used to provide extra verification that guaranteed
self is occuring everywhere that it is supposed to (since constructors and
destructors still have @owned self).
Swift SVN r27240
Previously some parts of the compiler referred to them as "fields",
and most referred to them as "elements". Use the more generic 'elements'
nomenclature because that's what we refer to other things in the compiler
(e.g. the elements of a bracestmt).
At the same time, make the API better by providing "getElement" consistently
and using it, instead of getElements()[i].
NFC.
Swift SVN r26894
