Look through a call to the ConvertPointerToPointerArgument compiler intrinsic
just like it is a copy of the pointer. At the source level, that's all it is:
Treat this like a direct use of the argument 'p' rather than the result of the
invisible pointer conversion call:
func foo(_: UnsafeRawPointer)
func bar(p: UnsafePointer<T>) {
foo(p)
}
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
With OSSA it can happen more easily that the final release is not immediately located before the related dealloc_stack_ref.
Therefore do a more precise check (using escape analysis) if any instruction between a release and a dealloc_stack_ref might (implicitly) release the allocated object.
This is part of fixing regressions when enabling OSSA modules:
rdar://140229560
This is needed after running the SSAUpdater for an existing OSSA value, because the updater can
insert unnecessary phis in the middle of the original liverange which breaks up the original
liverange into smaller ones:
```
%1 = def_of_owned_value
%2 = begin_borrow %1
...
br bb2(%1)
bb2(%3 : @owned $T): // inserted by SSAUpdater
...
end_borrow %2 // use after end-of-lifetime!
destroy_value %3
```
It's not needed to run this utility if SSAUpdater is used to create a _new_ OSSA liverange.
* 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]`
Replace destructure_tuple with tuple_extract instructions and destructure_struct with struct_extract instructions.
This canonicalization helps other optimizations to e.g. CSE tuple_extract/struct_extract.
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.
Just don't store the begin instruction.
This led to problem if the "begin" was not actually an instruction but a block argument.
Using the first instruction of that block is not correct in case the range ends immediately at the first instruction, e.g.
```
bb0(%0 : @owned $C):
destroy_value %0
```
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".
Sometimes it can happen that a deinit function, which is imported from another module, has shared linkage.
In this case it is important to de-serialize the function body. Otherwise it would be illegal SIL.
Unfortunately I don't have a test case for this.
Some utilities, like Builder.insert(after:) pass a Builder object that
represents an insertion point. That insertion point is sometimes needed to
determine which instructions to create (endApply vs. abortApply). But there is
no way to recover the insertion point. We don't want to simply return an
instruction, because that could be interpreted in different ways. Instead we
provide insertion block, but only in those situations where it makes sense and can't
be used incorrectly.
And simplify enclosingAccessScope.
These two APIs now directly compute what is being asked for. Previously, each
invocation of enclosingAccessScope would recompute the base and all nested
scopes, but throw away that information. Sometimes we do not want to compute
that information, and other times we need all of it. Instead, provide separate
APIs that do exactly what is needed.
These are common APIs used by other utilities. This avoids quadratic traversals in those
cases. It is also easier to understand and debug.
These utilities will be used in LifetimeDependenceScopeFixup.
LifetimeDependence uses the address of the AccessBase to compute liveness. It is
more convenient to switch over all AccessBase kinds here rather than force all
clients to do it.
AccessBase already identifies a base address as either a .global (addressor) or
a .pointer. When initializing an AccessBase from a pointer_to_address, precisely
identify one of these kinds of accesses, rather than simply marking it
.unidentified.
This used to be a special case in the AccessPathWalker. But that's not the right
place to handle it.
Scope-ending instructions, like `end_borrow` are only irrelevant for RLE if the preceding load is not changed.
If it is changed from `load [copy]` -> `load [take]` the memory effects of those scope-ending instructions prevent that the `load [take]` will illegally mutate memory which is protected from mutation by the scope.
Fixes a memory verifier crash
rdar://139824805
