These APIs are essential for complete OSSA liveness analysis. The
existing ad-hoc OSSA logic always misses some of the cases handled by
these new utilities. We need to start replacing that ad-hoc logic with
new utilities built on top of these APIs to define away potential
latent bugs.
Add FIXMEs to the inverse API: visitAdjacentBorrowsOfPhi. It should
probably be redesigned in terms of these new APIs.
Factors a mess of code in MemAccessUtils to handle forwarding
instruction types into a simpler utility. This utility is also needed
for ownership APIs, which need to be extended to handle these cases.
I've also fixed this so that it should work on instructions that
define multiple values. Someday we'll change all the open_existential
instructions to produce different values for the type dependency and
the value result; today is not that day, though.
Add TermInst::forwardedOperand.
Add SILArgument::forwardedTerminatorResultOperand. This API will be
moved into a proper TerminatorResult abstraction.
Remove getSingleTerminatorOperand, which could be misused because it's
not necessarilly forwarding ownership.
Remove the isTransformationTerminator API, which is not useful or well
defined.
Rewrite several instances of complex logic to handle block arguments
with the simple terminator result API. This defines away potential
bugs where we don't detect casts that perform implicit conversion.
Replace uses of the SILPhiArgument type and code that explicitly
handle block arguments. Control flow is irrelevant in these
situations. SILPhiArgument needs to be deleted ASAP. Instead, use
simple APIs like SILArgument::isTerminatorResult(). Eventually this
will be replaced by a TerminatorResult type.
Add `deletableInstructions()` and `reverseDeletableInstructions()` in SILBasicBlock.
It allows deleting instructions while iterating over all instructions of the block.
This is a replacement for `InstructionDeleter::updatingRange()`.
It's a simpler implementation than the existing `UpdatingListIterator` and `UpdatingInstructionIteratorRegistry`, because it just needs to keep the prev/next pointers for "deleted" instructions instead of the iterator-registration machinery.
It's also safer, because it doesn't require to delete instructions via a specific instance of an InstructionDeleter (which can be missed easily).
Instead of setting the parent pointer to null, set the `lastInitializedBitfieldID` to -1.
This allows to keep the parent block information, even when an instruction is removed from it's list.
`getValue` -> `value`
`getValueOr` -> `value_or`
`hasValue` -> `has_value`
`map` -> `transform`
The old API will be deprecated in the rebranch.
To avoid merge conflicts, use the new API already in the main branch.
rdar://102362022
Added new C++-to-Swift callback for isDeinitBarrier.
And pass it CalleeAnalysis so it can depend on function effects. For
now, the argument is ignored. And, all callers just pass nullptr.
Promoted to API the mayAccessPointer component predicate of
isDeinitBarrier which needs to remain in C++. That predicate will also
depends on function effects. For that reason, it too is now passed a
BasicCalleeAnalysis and is moved into SILOptimizer.
Also, added more conservative versions of isDeinitBarrier and
maySynchronize which will never consider side-effects.
This lets us write optimizer unit tests and selectively debug the
optimizer in general. We'll be able trace analyses and control
optimization selectively for certain values.
Adding a trace flag to debug_value is the easiest way to start using
it experimentally and develop the rest of the infrastructure. If this
takes off, then we can consider a new `trace_value`
instruction. For now, reusing debug_value is the least intrusive way to
start writing liveness unit tests.
This is a dedicated instruction for incrementing a
profiler counter, which lowers to the
`llvm.instrprof.increment` intrinsic. This
replaces the builtin instruction that was
previously used, and ensures that its arguments
are statically known. This ensures that SIL
optimization passes do not invalidate the
instruction, fixing some code coverage cases in
`-O`.
rdar://39146527
This is exactly like copy_addr except that it is not viewed from the verifiers
perspective as an "invalid" copy of a move only value. It is intended to be used
in two contexts:
1. When the move checker emits a diagnostic since it could not eliminate a copy,
we still need to produce valid SIL without copy_addr on move only types since we
will hit canonical SIL eventually even if we don't actually codegen the SIL. The
pass can just convert said copy_addr to explicit_copy_addr and everyone is
happy.
2. To implement the explicit copy function for address only types.
This is just the very beginning... I still need to implement more parts of
SILGen for this. But all great things start small. I am going to iterate on top
of this and just wanted to get some initial parts of the work in as I go.
Specifically this means that rather than always being owned, we now have owned
and guaranteed versions of copyable_to_moveonlywrapper. Similar to
moveonlywrapper_to_copyable, one chooses which variant one gets by using
specific SILBuilder APIs:
create{Owned,Guaranteed}CopyableToMoveOnlyWrapperValueInst. It is still
forwarding and the rest of the forwarding APIs work as expected except that the
forwarding ownership is fixed (and an assertion will result if one attempts to
do so).
NOTE: It is assumed that trivial operands are always passed to the owned
variant.
In the future, we may not have this requirement, but for now it is a really
convenient way to track down places in SILGen where I have missed converting
from copyable to moveonlywrapped types.
The use of the SWIFT_INLINE_BITFIELD macros in SILNode were a constant source of confusion and bugs.
With this refactoring I tried to simplify the definition of "shared fields" in SILNode, SILValue and SILInstruction classes:
* Move `kind`, `locationKindAndFlags` and the 32-bit fields out of the 64-bitfield into their own member variables. This avoids _a lot_ of manual bit position computations.
* Now we have two separate "shared fields": an 8-bit field (e.g. for boolean flags) and a 32-bit field (e.g. for indices, which can potentially get large). Both fields can be used independently. Also, they are not "bit fields" per se. Instructions can use the field e.g. as a `bool`, `uint32_t`, or - if multiple flags are to be stored - as a packed bit field.
* With these two separate fields, we don't have the need for defining bitfields both in a base class _and_ in a derived value/instruction class. We can get rid of the complex logic which handles such cases. Just keep a check to catch accidental overlaps of fields in base and derived classes.
* Still use preprocessor macros for the implementation, but much simpler ones than before.
* Add documentation.
The main fixes are:
1. MoveOnlyWrapperToCopyableValue needed to be marked as a
FirstArgOwnershipForwardingSingleValueInst instead of just as being an
Ownership mixin. I discovered that in certain cases I was treating it that
way (in the isa check for FirstArgOwnershipForwardingSingleValueInst), but we
were inconsistent. Now we are consistent.
2. MoveOnlyWrapperToCopyableValue is always specified as being initialized as
owned or guaranteed. What is key to understand though is that the
owned/guaranteed property here is more a semantic property around whether the
lifetime of the move only value is ending or if we are allowing it to escape
as an moveonlywrapped unwrapped guaranteed parameter to a function. The main
implication of this is that we can not just use the actual ownership kind to
determine the type of moveonlywrapper_to_copyable we are using. This is b/c
after ownership lowering, the resulting ownership kind will be none, meaning
the instruction will be in an invalid state. Thus the need to represent this
as a separate bit in the instruction. It may make sense to rename the forms
of this instruction to be `[lifetime end]` and `[guaranteed function arg]`
that way it is semantically clear. But I am going to do that change at
another time.
These instructions have the following attributes:
1. copyably_to_moveonlywrapper takes in a 'T' and maps it to a '@moveOnly
T'. This is semantically used when initializing a new moveOnly binding from a
copyable value. It semantically destroys its input @owned value and returns a
brand new independent @owned @moveOnly value. It also is used to convert a
trivial copyable value with type 'Trivial' into an owned non-trivial value of
type '@moveOnly Trivial'. If one thinks of '@moveOnly' as a monad, this is how
one injects a copyable value into the move only space.
2. moveonlywrapper_to_copyable takes in a '@moveOnly T' and produces a new 'T'
value. This is a 'forwarding' instruction where at parse time, we only allow for
one to choose it to be [owned] or [guaranteed].
* moveonlywrapper_to_copyable [owned] is used to signal the end of lifetime of
the '@moveOnly' wrapper. SILGen inserts these when ever a move only value has
its ownership passed to a situation where a copyable value is needed. Since it
is consuming, we know that the no implicit copy checker will ensure that if we
need a copy for it, the program will emit a diagnostic.
* moveonlywrapper_to_copyable [guaranteed] is used to pass a @moveOnly T value
as a copyable guaranteed parameter with type 'T' to a function. In the case of
using no-implicit-copy checking this is always fine since no-implicit-copy is a
local pattern. This would be an error when performing no escape
checking. Importantly, this instruction also is where in the case of an
@moveOnly trivial type, we convert from the non-trivial representation to the
trivial representation.
Some important notes:
1. In a forthcoming commit, I am going to rebase the no implicit copy checker on
top of these instructions. By using '@moveOnly' in the type system, we can
ensure that later in the SIL pipeline, we can have optimizations easily ignore
the code.
2. Be aware of is that due to SILGen only emitting '@moveOnly T' along immediate
accesses to the variable and always converts to a copyable representation when
calling other code, we can simply eliminate from the IR all moveonly-ness from
the IR using a lowering pass (that I am going to upstream). In the evil scheme
we are accomplishing here, we perform lowering of trivial values right after
ownership lowering and before diagnostics to simplify the pipeline.
On another note, I also fixed a few things in SILParsing around getASTType() vs
getRawASTType().
As we do with field indices for struct instructions.
This avoids quadratic behavior in case of enums with lots of cases.
Also: cache field and enum case indices in the SILModule.
