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.
Changes in this CR add part of the, Swift based, Autodiff specific
closure specialization optimization pass. The pass does not modify any
code nor does it even exist in any of the optimization pipelines. The
rationale for pushing this partially complete optimization pass upstream
is to keep up with the breaking changes in the underlying Swift based
compiler infrastructure.
The reason why I am doing this is that I am going to be adding support for
preconcurrency imports to TransferNonSendable. That implies that we can have
preconcurrency import suppression in the SIL pipeline and thus that emitting the
diagnostic in Sema is too early.
To do this, I introduced a new module pass called
DiagnoseUnnecessaryPreconcurrencyImports that runs after the SILFunction pass
TransferNonSendable. The reason why I use a module pass is to ensure that
TransferNonSendable has run on all functions before we attempt to emit these
diagnostics. Then in that pass, we iterate over all of the modules functions and
construct a uniqued array of SourceFiles for these functions. Then we iterate
over the uniqued SourceFiles and use the already constructed Sema machinery to
emit the diagnostic using the source files.
rdar://126928265
Compute, update and handle borrowed-from instruction in various utilities and passes.
Also, used borrowed-from to simplify `gatherBorrowIntroducers` and `gatherEnclosingValues`.
Replace those utilities by `Value.getBorrowIntroducers` and `Value.getEnclosingValues`, which return a lazily computed Sequence of borrowed/enclosing values.
I am doing this since region based isolation hit the same issue that the move
checker did. So it makes sense to refactor the functionality into its own pass
and move it into a helper pass that runs before both.
It is very conservative and only stubifies functions that the specialization
passes explicitly mark as this being ok to be done to.
Over time I am going to be using RegionAnalysis for a series of passes that all
use that same information since I am worried about RegionAnalysis computation
time. With that being said, we want to make sure to eliminate the memory that
RegionAnalysis uses once this series of passes have completed. What this commit
does is create a pass that explicitly invalidates region analysis and explicitly
places it in the pass pipeline after the series of passes. This will ensure that
even if we add an additional pass, there is a strong "rattlesnake" signal to the
new code author that the code needs to be placed before the region analysis
invalidation and will prevent mistakes such as having to recompute the region
analysis in that later pass or the later pass forgeting to invalidate the
analysis.
Add a new mandatory BooleanLiteralFolding pass which constant folds conditional branches with boolean literals as operands.
```
%1 = integer_literal -1
%2 = apply %bool_init(%1) // Bool.init(_builtinBooleanLiteral:)
%3 = struct_extract %2, #Bool._value
cond_br %3, bb1, bb2
```
->
```
...
br bb1
```
This pass is intended to run before DefiniteInitialization, where mandatory inlining and constant folding didn't run, yet (which would perform this kind of optimization).
This optimization is required to let DefiniteInitialization handle boolean literals correctly.
For example in infinite loops:
```
init() {
while true { // DI need to know that there is no loop exit from this while-statement
if some_condition {
member_field = init_value
break
}
}
}
```
By default it lowers the builtin to an `alloc_vector` with a paired `dealloc_stack`.
If the builtin appears in the initializer of a global variable and the vector elements are initialized,
a statically initialized global is created where the initializer is a `vector` instruction.
In regular swift this is a nice optimization. In embedded swift it's a requirement, because the compiler needs to be able to specialize generic deinits of non-copyable types.
The new de-virtualization utilities are called from two places:
* from the new DeinitDevirtualizer pass. It replaces the old MoveOnlyDeinitDevirtualization, which is very basic and does not fulfill the needs for embedded swift.
* from MandatoryPerformanceOptimizations for embedded swift
For chains of async functions where suspensions can be statically
proven to never be required, this pass removes all suspensions and
turns the functions into synchronous functions.
For example, this function does not actually require any suspensions,
once the correct executor is acquired upon initial entry:
```
func fib(_ n: Int) async -> Int {
if n <= 1 { return n }
return await fib(n-1) + fib(n-2)
}
```
So we can turn the above into this for better performance:
```
func fib() async -> Int {
return fib_sync()
}
func fib_sync(_ n: Int) -> Int {
if n <= 1 { return n }
return fib(n-1) + fib(n-2)
}
```
while rewriting callers of `fib` to use the `sync` entry-point
when we can prove that it will be invoked on a compatible executor.
This pass is currently experimental and under development. Thus, it
is disabled by default and you must use
`-enable-experimental-async-demotion` to try it.
It lowers let property accesses of classes.
Lowering consists of two tasks:
* In class initializers, insert `end_init_let_ref` instructions at places where all let-fields are initialized.
This strictly separates the life-range of the class into a region where let fields are still written during
initialization and a region where let fields are truly immutable.
* Add the `[immutable]` flag to all `ref_element_addr` instructions (for let-fields) which are in the "immutable"
region. This includes the region after an inserted `end_init_let_ref` in an class initializer, but also all
let-field accesses in other functions than the initializer and the destructor.
This pass should run after DefiniteInitialization but before RawSILInstLowering (because it relies on `mark_uninitialized` still present in the class initializer).
Note that it's not mandatory to run this pass. If it doesn't run, SIL is still correct.
Simplified example (after lowering):
bb0(%0 : @owned C): // = self of the class initializer
%1 = mark_uninitialized %0
%2 = ref_element_addr %1, #C.l // a let-field
store %init_value to %2
%3 = end_init_let_ref %1 // inserted by lowering
%4 = ref_element_addr [immutable] %3, #C.l // set to immutable by lowering
%5 = load %4
- VTableSpecializer, a new pass that synthesizes a new vtable per each observed concrete type used
- Don't use full type metadata refs in embedded Swift
- Lazily emit specialized class metadata (LazySpecializedClassMetadata) in IRGen
- Don't emit regular class metadata for a class decl if it's generic (only emit the specialized metadata)
The new implementation has several benefits compared to the old C++ implementation:
* It is significantly simpler. It optimizes each load separately instead of all at once with bit-field based dataflow.
* It's using alias analysis more accurately which enables more loads to be optimized
* It avoids inserting additional copies in OSSA
The algorithm is a data flow analysis which starts at the original load and searches for preceding stores or loads by following the control flow in backward direction.
The preceding stores and loads provide the "available values" with which the original load can be replaced.
The old C++ pass didn't catch a few cases.
Also:
* The new pass is significantly simpler: it doesn't perform dataflow for _all_ memory locations at once using bitfields, but handles each store separately. (In both implementations there is a complexity limit in place to avoid quadratic complexity)
* The new pass works with OSSA
Renamed UnitTest to FunctionTest.
FunctionTests are now instantiated once as global objects--with their
names and the code they are to run--at which time they are stored by
name in a global registry.
Moved the types to the SIL library.
Together, these changes enable defining unit tests in the source file
containing the code to be tested.
It sets the `[bare]` attribute for `alloc_ref` and `global_value` instructions if their header (reference count and metatype) is not used throughout the lifetime of the object.
In the next commit, I am modifying the move only checker to ensure that we
always error when partially invalidating. While doing this I discovered that
these copies were getting in the way of emitting good diagnostics in that case.
Deallocate dynamic allocas done for metadata/wtable packs. These
stackrestore calls are inserted on the dominance frontier and then stack
nesting is fixed up. That was achieved as follows:
Added a new IRGen pass PackMetadataMarkerInserter; it
- determines if there are any instructions which might allocate on-stack
pack metadata
- if there aren't, no changes are made
- if there are, alloc_pack_metadata just before instructions that could
allocate pack metadata on the stack and dealloc_pack_metadata on the
dominance frontier of those instructions
- fixup stack nesting
During IRGen, the allocations done for metadata/wtable packs are
recorded and IRGenSILFunction associates them with the instruction that
lowered. It must be the instruction after some alloc_pack_metadata
instruction. Then, when visiting the dealloc_pack_metadata instructions
corresponding to that alloc_pack_metadata, deallocate those packs.
This reverts commit 224674cad1.
Originally, I made this change since we were going to follow the AST in a strict
way in terms of what closures are considered escaping or not from a diagnostics
perspective. Upon further investigation I found that we actually do something
different for inout escaping semantics and by treating the AST as the one point
of truth, we are being inconsistent with the rest of the compiler. As an
example, the following code is considered by the compiler to not be an invalid
escaping use of an inout implying that we do not consider the closure to be
escaping:
```
func f(_ x: inout Int) {
let g = {
_ = x
}
}
```
in contrast, a var is always considered to be an escape:
```
func f(_ x: inout Int) {
var g = {
_ = x
}
}
test2.swift:3:13: error: escaping closure captures 'inout' parameter 'x'
var g = {
^
test2.swift:2:10: note: parameter 'x' is declared 'inout'
func f(_ x: inout Int) {
^
test2.swift:4:11: note: captured here
_ = x
^
```
Of course, if we store the let into memory, we get the error one would expect:
```
var global: () -> () = {}
func f(_ x: inout Int) {
let g = {
_ = x
}
global = g
}
test2.swift:4:11: error: escaping closure captures 'inout' parameter 'x'
let g = {
^
test2.swift:3:10: note: parameter 'x' is declared 'inout'
func f(_ x: inout Int) {
^
test2.swift:5:7: note: captured here
_ = x
^
```
By reverting to the old behavior where allocbox to stack ran early, noncopyable
types now have the same sort of semantics: let closures that capture a
noncopyable type that do not on the face of it escape are considered
non-escaping, while if the closure is ever stored into memory (e.x.: store into
a global, into a local var) or escapes, we get the appropriate escaping
diagnostics. E.x.:
```
public struct E : ~Copyable {}
public func borrowVal(_ e: borrowing E) {}
public func consumeVal(_ e: consuming E) {}
func f1() {
var e = E()
// Mutable borrowing use of e. We can consume e as long as we reinit at end
// of function. We don't here, so we get an error.
let c1: () -> () = {
borrowVal(e)
consumeVal(e)
}
// Mutable borrowing use of e. We can consume e as long as we reinit at end
// of function. We do do that here, so no error.
let c2: () -> () = {
borrowVal(e)
consumeVal(e)
e = E()
}
}
```
This allows to run the NamedReturnValueOptimization only late in the pipeline.
The optimization shouldn't be done before serialization, because it might prevent predictable memory optimizations in the caller after inlining.
It converts a lazily initialized global to a statically initialized global variable.
When this pass runs on a global initializer `[global_init_once_fn]` it tries to create a static initializer for the initialized global.
```
sil [global_init_once_fn] @globalinit {
alloc_global @the_global
%a = global_addr @the_global
%i = some_const_initializer_insts
store %i to %a
}
```
The pass creates a static initializer for the global:
```
sil_global @the_global = {
%initval = some_const_initializer_insts
}
```
and removes the allocation and store instructions from the initializer function:
```
sil [global_init_once_fn] @globalinit {
%a = global_addr @the_global
%i = some_const_initializer_insts
}
```
The initializer then becomes a side-effect free function which let's the builtin-simplification remove the `builtin "once"` which calls the initializer.
Now that we handle inlined global initializers in LICM, CSE and the StringOptimization, we don't need to have a separate mid-level inliner pass, which treats global accessors specially.
Specifically, we already have the appropriate semantics for arguments captured
by escaping closures but in certain cases allocbox to stack is able to prove
that the closure doesn’t actually escape. This results in the capture being
converted into a non-escaping SIL form. This then causes the move checker to
emit the wrong kind of error.
The solution is to create an early allocbox to stack that doesn’t promote move
only types in boxes from heap -> stack if it is captured by an escaping closure
but does everything else normally. Then once the move checking is completed, we
run alloc box to stack an additional time to ensure that we keep the guarantee
that heap -> stack is performed in those cases.
rdar://108905586
