This is done by splitting the `begin_borrow` of the whole struct into individual borrows of the fields (for trivial fields no borrow is needed).
And then sinking the `struct` to it's consuming use(s).
```
%3 = struct $S(%nonTrivialField, %trivialField) // owned
...
%4 = begin_borrow %3
%5 = struct_extract %4, #S.nonTrivialField
%6 = struct_extract %4, #S.trivialField
use %5, %6
end_borrow %4
...
end_of_lifetime %3
```
->
```
...
%5 = begin_borrow %nonTrivialField
use %5, %trivialField
end_borrow %5
...
%3 = struct $S(%nonTrivialField, %trivialField)
end_of_lifetime %3
```
This optimization is important for Array code where the Array buffer is constantly wrapped into structs and then extracted again to access the buffer.
Beside supporting OSSA, this change significantly simplifies the pass.
The main change is that instead of starting at a closure (e.g. `partial_apply`) and finding all call sites, we now start at a call site and look for closures for all arguments. This makes a lot of things much simpler, e.g. not so many intermediate data structures are required to track all the states.
I needed to remove the 3 unit tests because the things those tests were testing are not there anymore. However, the pass is tested with a lot of sil tests (and I added quite a few), which should give good test coverage.
The old ClosureSpecializer pass is still kept in place, because at that point in the pipeline we don't have OSSA, yet. Once we have that, we can replace the old pass withe the new one.
However, the autodiff closure specializer already runs in the OSSA pipeline and there the new changes take effect.
(old name: CapturePropagation)
The pass is now rewritten in swift which makes the code smaller and simpler.
Compared to the old pass it has two improvements:
* It can constant propagate whole structs (and not only builtin literals). This is important for propagating "real" Swift constants which have a struct type of e.g. `Int`.
* It constant propagates keypaths even if there are other non-constant closure captures which are not propagated. This is something the old pass didn't do.
rdar://151185177
This allows to move many SIL APIs and utilities, which require a context, to the SIL module.
The SIL-part of SwiftPassInvocation is extracted into a base class SILContext which now lives in SIL.
Also: simplify the begin/end-pass functions of the SwiftPassInvocation.
We insert end_cow_mutation_addr for lifetime dependent values dependent on mutable addresses.
end_cow_mutation_addr can be simplified to end_cow_mutation after other optimizations like inlining, specialization etc
This PR adds an instruction simplification to transform end_cow_mutation_addr to end_cow_mutation.
This can enable array optimizations which look for end_cow_mutation.
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
* 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
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
Reimplement the simplification in swift and add a new transformation:
```
%1 = unchecked_addr_cast %0 : $*Builtin.FixedArray<N, Element> to $*Element
```
->
```
%1 = vector_base_addr %0 : $*Builtin.FixedArray<N, Element>
```
Beside cleaning up the source code, the motivation for the translation into Swift is to make it easier to improve the pass for some InlineArray specific optimizations (though I'm not sure, yet if we really need those).
Also, the new implementation doesn't contain the optimize-store-into-temp optimization anymore, because this is covered by redundant load elimination.
1. move embedded diagnostics out of the PerformanceDiagnostics pass. It was completely separated from the other logic in this pass, anyway.
2. rewrite it in swift
3. fix several bugs, that means: missed diagnostics, which led to IRGen crashes
* look at all methods in witness tables, including base protocols and associated conformances
* visit all functions in the call tree, including generic functions with class bound generic arguments
* handle all instructions, e.g. concurrency builtins
4. improve error messages by adding meaningful call-site information. For example:
* if the error is in a specialized function, report where the generic function is originally specialized with concrete types
* if the error is in a protocol witness method, report where the existential is created
Fixes a false alarm in case of recursive calls with different type parameters.
For example:
```
protocol P {
associatedtype E: P
}
func noRecursionMismatchingTypeArgs1<T: P>(_ t: T.Type) {
if T.self == Int.self {
return
}
noRecursionMismatchingTypeArgs1(T.E.self)
}
```
* Reimplement most of the logic in Swift as an Instruction simplification and remove the old code from SILCombine
* support more cases of existential archetype replacements:
For example:
```
%0 = alloc_stack $any P
%1 = init_existential_addr %0, $T
use %1
```
is transformed to
```
%0 = alloc_stack $T
use %0
```
Also, if the alloc_stack is already an opened existential and the concrete type is known,
replace it as well:
```
%0 = metatype $@thick T.Type
%1 = init_existential_metatype %0, $@thick any P.Type
%2 = open_existential_metatype %1 : $@thick any P.Type to $@thick (@opened("X", P) Self).Type
...
%3 = alloc_stack $@opened("X", any P) Self
use %3
```
is transformed to
```
...
%3 = alloc_stack $T
use %3
```
If an apply uses an existential archetype (`@opened("...")`) and the concrete type is known, replace the existential archetype with the concrete type
1. in the apply's substitution map
2. in the arguments, e.g. by inserting address casts
For example:
```
%5 = apply %1<@opend("...")>(%2) : <τ_0_0> (τ_0_0) -> ()
```
->
```
%4 = unchecked_addr_cast %2 to $*ConcreteType
%5 = apply %1<ConcreteType>(%4) : <τ_0_0> (τ_0_0) -> ()
```
Replace `unconditional_checked_cast` to an existential metatype with an `init_existential_metatype`, it the source is a conforming type.
Note that init_existential_metatype is better than unconditional_checked_cast because it does not need to do any runtime casting.
Which consists of
* removing redundant `address_to_pointer`-`pointer_to_address` pairs
* optimize `index_raw_pointer` of a manually computed stride to `index_addr`
* remove or increase the alignment based on a "assumeAlignment" builtin
This is a big code cleanup but also has some functional differences for the `address_to_pointer`-`pointer_to_address` pair removal:
* It's not done if the resulting SIL would result in a (detectable) use-after-dealloc_stack memory lifetime failure.
* It's not done if `copy_value`s must be inserted or borrow-scopes must be extended to comply with ownership rules (this was the task of the OwnershipRAUWHelper).
Inserting copies is bad anyway.
Extending borrow-scopes would only be required if the original lifetime of the pointer extends a borrow scope - which shouldn't happen in save code. Therefore this is a very rare case which is not worth handling.
`String(describing:)` does a bunch of dynamic casts
that can be pretty slow. Use interpolation instead,
which bypasses them.
For `swift-frontend`, this brings the time taken
for type-checking an empty file down from ~100ms
to ~70ms.
For `swift build`, this brings the time taken for
a null build down from ~600ms to ~450ms (the
larger delta is presumably due to the fact that
there's much more Swift code in `swift-package`).
Canonicalize a `fix_lifetime` from an address to a `load` + `fix_lifetime`:
```
%1 = alloc_stack $T
...
fix_lifetime %1
```
->
```
%1 = alloc_stack $T
...
%2 = load %1
fix_lifetime %2
```
This transformation is done for `alloc_stack` and `store_borrow` (which always has an `alloc_stack` operand).
The benefit of this transformation is that it enables other optimizations, like mem2reg.
This peephole optimization was already done in SILCombine, but it didn't handle store_borrow.
A good opportunity to make an instruction simplification out of it.
This is part of fixing regressions when enabling OSSA modules:
rdar://140229560
* Remove dead `load_borrow` instructions (replaces the old peephole optimization in SILCombine)
* If the `load_borrow` is followed by a `copy_value`, combine both into a `load [copy]`
It hoists `destroy_value` instructions without shrinking an object's lifetime.
This is done if it can be proved that another copy of a value (either in an SSA value or in memory) keeps the referenced object(s) alive until the original position of the `destroy_value`.
```
%1 = copy_value %0
...
last_use_of %0
// other instructions
destroy_value %0 // %1 is still alive here
```
->
```
%1 = copy_value %0
...
last_use_of %0
destroy_value %0
// other instructions
```
The benefit of this optimization is that it can enable copy-propagation by moving destroys above deinit barries and access scopes.
It removes a `copy_value` where the source is a guaranteed value, if possible:
```
%1 = copy_value %0 // %0 = a guaranteed value
// uses of %1
destroy_value %1 // borrow scope of %0 is still valid here
```
->
```
// uses of %0
```
This optimization is very similar to the LoadCopyToBorrow optimization.
Therefore I merged both optimizations into a single file and renamed it to "CopyToBorrowOptimization".
The optimization replaces a `load [copy]` with a `load_borrow` if possible.
```
%1 = load [copy] %0
// no writes to %0
destroy_value %1
```
->
```
%1 = load_borrow %0
// no writes to %0
end_borrow %1
```
The new implementation uses alias-analysis (instead of a simple def-use walk), which is much more powerful.
rdar://115315849
As the optimizer uses more and more AST stuff, it's now time to create an "AST" module.
Initially it defines following AST datastructures:
* declarations: `Decl` + derived classes
* `Conformance`
* `SubstitutionMap`
* `Type` and `CanonicalType`
Some of those were already defined in the SIL module and are now moved to the AST module.
This change also cleans up a few things:
* proper definition of `NominalTypeDecl`-related APIs in `SIL.Type`
* rename `ProtocolConformance` to `Conformance`
* use `AST.Type`/`AST.CanonicalType` instead of `BridgedASTType` in SIL and the Optimizer
The main changes are:
*) Rewrite everything in swift. So far, parts of memory-behavior analysis were already implemented in swift. Now everything is done in swift and lives in `AliasAnalysis.swift`. This is a big code simplification.
*) Support many more instructions in the memory-behavior analysis - especially OSSA instructions, like `begin_borrow`, `end_borrow`, `store_borrow`, `load_borrow`. The computation of end_borrow effects is now much more precise. Also, partial_apply is now handled more precisely.
*) Simplify and reduce type-based alias analysis (TBAA). The complexity of the old TBAA comes from old days where the language and SIL didn't have strict aliasing and exclusivity rules (e.g. for inout arguments). Now TBAA is only needed for code using unsafe pointers. The new TBAA handles this - and not more. Note that TBAA for classes is already done in `AccessBase.isDistinct`.
*) Handle aliasing in `begin_access [modify]` scopes. We already supported truly immutable scopes like `begin_access [read]` or `ref_element_addr [immutable]`. For `begin_access [modify]` we know that there are no other reads or writes to the access-address within the scope.
*) Don't cache memory-behavior results. It turned out that the hit-miss rate was pretty bad (~ 1:7). The overhead of the cache lookup took as long as recomputing the memory behavior.
