This removes the ambiguity when casting from a SingleValueInstruction to SILNode, which makes the code simpler. E.g. the "isRepresentativeSILNode" logic is not needed anymore.
Also, it reduces the size of the most used instruction class - SingleValueInstruction - by one pointer.
Conceptually, SILInstruction is still a SILNode. But implementation-wise SILNode is not a base class of SILInstruction anymore.
Only the two sub-classes of SILInstruction - SingleValueInstruction and NonSingleValueInstruction - inherit from SILNode. SingleValueInstruction's SILNode is embedded into a ValueBase and its relative offset in the class is the same as in NonSingleValueInstruction (see SILNodeOffsetChecker).
This makes it possible to cast from a SILInstruction to a SILNode without knowing which SILInstruction sub-class it is.
Casting to SILNode cannot be done implicitly, but only with an LLVM `cast` or with SILInstruction::asSILNode(). But this is a rare case anyway.
Instead, I just added some static helper methods that perform the same
operations without needing to deal with generics/etc on OwnershipForwardingMixin
itself. The reason why I did this is that this Mixin is not part of the SILNode
heirarchy so we shouldn't use utilities tied to the SILNode hierarchy.
b644c80f90fb7099ec956bb44065b50e432c5146 caused all owned forwarding
instructions to be sunk to their uses in the exact cases where we could
eliminate the parent forwarding inst (namely that the value we want to fold has
no non-debug, non-consuming users). So I was able to implement this just by
implementing a single basic block algorithm that works via a planner struct
using said canonicalization. One initializes the planner struct with the
instruction that is going to either be eliminated or have its forwarding operand
set. Then one adds each of the individual chains that lead to the use that we
wish to fold, each time checking that we can eliminate the instruction.
Once the user has added all of the intermediate forwarding instructions, by
construction (see paragraph above), we know we can optimize. So we eliminate all
intermediate values and then depending on whether the user called the set value
or replace value method we either set front's operand to be the passed in value
or we RAUW/erase front with that value. It is important to note that before we
do one of those two operations, front's operand is undef, so we need to perform
one of these two operations.
Additional handling of copy_value/destroy_value/load[copy]/
begin_borrow/end_borrow is needed to support OSSA.
TODO: Support handling of 2d array in the pass for OSSA.
Currently hoisting of loads and borrows are not supported in OSSA
It's important that fundamental APIs don't lie to their users.
Make it clear that this API always returns true for deinitialization,
even if we could for example analyze the destructor and determine that
there aren't any actual writes!
`differentiability_function_extract` instruction has an optional explicit
extractee type. This is currently used by TypeSubstCloner and the
LoadableByAddress transform to rewrite `differentiability_function_extract`
instructions while preserving `@differentiable` function type invariants.
There is an assertion that `differentiability_function_extract` instructions do
not have explicit extractee types outside of canonical/lowered SIL. However,
this does not handle the SIL deserialization case above: when a function
containing a `differentiable_function_extract` instruction with an explicit type
is deserialized into a raw SIL module (which happens when optimizations are
enabled).
Removing the assertion unblocks this encountered use case.
A more robust longer-term solution may be to change SIL `@differentiable`
function types to explicitly store component original/JVP/VJP function types.
Also fix `differentiable_function_extract` extractee type serialization.
Resolves SR-14004.
This commit is doing a few things:
1. It is centralizing all decisions about whether an operand's owner instruction
or a value's parent instruction is forwarding in each SILInstruction
itself. This will prevent this information from getting out of sync.
2. This allowed me to hide the low level queries in OwnershipUtils.h that
determined if a SILNodeKind was "forwarding". I tried to minimize the amount of
churn in this PR and thus didn't remove the
is{Owned,Ownership,Guaranteed}Forwarding{Use,Value} checks. Instead I left them
alone but added in asserts to make sure that if the old impl ever returns true,
the neew impl does as well. In a subsequent commit, I am going to remove the old
impl in favor of isa queries.
3. I also in the process discovered that there were some instructions that were
being inconsistently marked as forwarding. All of the asserts in the PR caught
these and I fixed these inconsistencies.
Previously FieldIndexCacheBase only had a parent class of
SingleValueInstruction. I need to be able to in certain cases shim in a
SingleValueInstruction subclass as a parent class instead. In my case it is to
imbue ownership forwarding on StructExtractInst.
This commit itself doesn't make that change and instead just always templatizes
using SingleValueInstruction.
I think what was happening here was that we were using one of the superclass
classofs and were getting lucky since in the place I was using this I was
guaranteed to have single value instructions and that is what I wrote as my
first case X ).
I also added more robust checks tieing the older isGuaranteed...* APIs to the
ForwardingOperand API. I also eliminated the notion of Branch being an owned
forwarding instruction. We only used this in one place in the compiler (when
finding owned value introducers), yet we treat a phi as an introducer, so we
would never hit a branch in our search since we would stop at the phi argument.
The bigger picture here is that this means that all "forwarding instructions"
either forward ownership for everything or for everything but owned/unowned.
And for those listening in, I did find one instruction that was from an
ownership forwarding subclass but was not marked as forwarding:
DifferentiableFunctionInst. With this change, we can no longer by mistake have
such errors enter the code base.
Interestingly this problem can only occur if one invokes
MarkUninitializedInst::getKind() directly. Once our instruction is just a
SILInstruction, we call the appropriate method so we didn't notice it.
I used Xcode's refactoring functionality to find all of the invocation
locations.
This makes it easier to understand conceptually why a ValueOwnershipKind with
Any ownership is invalid and also allowed me to explicitly document the lattice
that relates ownership constraints/value ownership kinds.
This allows us to hoist the error case of having a function signature with
conflicting ownership requirements into the creation of the return inst instead
of at the time of computing Operand Constraints.
This is the last part of the Operand Constraint computation that can fail that
once removed will let me use fail to mean any constriant is allowed.
Previously, we always inferred the ownership of the switch_enum from its phi
operands. This forced us to need to model a failure to find a good
OperandOwnershipKindMap in OperandOwnership.cpp. We want to eliminate such
conditions so that we can use failing to find a constraint to mean that a value
can accept any value rather than showing a failure.
This instructions ensures that all instructions, which need to run on the specified executor actually run on that executor.
For details see the description in SIL.rst.
It can already only accept values with none ownership and the merging of
ownership around ownership phis ensure that if we phi this with a partial_apply
or the like, we get the appropriate ownership on any such ownership phi values.
We are now out of SILGen emitting fewer destroy_value unnecessarily on
thin_to_thick functions. This changed some codegen and also forced me to update
some tests/fix AutoDiff.
I also deleted the DebugInfo test mandatoryinlining-wrongdebugscope.swift since:
1. It was depending on these destroys being there.
2. Given the need to improve the test @aprantl suggested I just eliminate it
solving the test failure for me.
Compute 'isLet' from the VarDecl that is available when constructing
AccessedStorage so we don't need to recover the VarDecl for the base
later.
This generally makes more sense and is more efficient, but it will be
necessary when we look past class casts when finding the reference root.
`get_async_continuation[_addr]` begins a suspend operation by accessing the continuation value that can resume
the task, which can then be used in a callback or event handler before executing `await_async_continuation` to
suspend the task.
I don't have a test case for this bug based on the current code. But
the fix is clearly needed to have a unique AccessStorage object for
each property. The AccessPath commits will contain test cases for this
functionality.
Today unchecked_bitwise_cast returns a value with ObjCUnowned ownership. This is
important to do since the instruction can truncate memory meaning we want to
treat it as a new object that must be copied before use.
This means that in OSSA we do not have a purely ossa forwarding unchecked
layout-compatible assuming cast. This role is filled by unchecked_value_cast.
The ``base_addr_for_offset`` instruction creates a base address for offset calculations.
The result can be used by address projections, like ``struct_element_addr``, which themselves return the offset of the projected fields.
IR generation simply creates a null pointer for ``base_addr_for_offset``.
Use TangentStoredPropertyRequest in differentiation transform.
Improve non-differentiability diagnostics regarding invalid stored
property projection instructions:
`struct_extract`, `struct_element_addr`, `ref_element_addr`.
Diagnose the following cases:
- Original property's type does not conform to `Differentiable`.
- Base type's `TangentVector` is not a struct.
- Tangent property not found: base type's `TangentVector` does not have a
stored property with the same name as the original property.
- Tangent property's type is not equal to the original property's
`TangentVector` type.
- Tangent property is not a stored property.
Resolves TF-969 and TF-970.
Currently we only have load [take] in OSSA which needed to be changed.
(copy_addr is not handled in MemBehavior at all, yet)
Even if the memory is physically not modified, conceptually it's "destroyed" when the value is taken.
Optimizations, like TempRValueOpt rely on this behavior when the check for may-writes.
This fixes a MemoryLifetime failure in TempRValueOpt.
`DifferentiableFunctionInst` now stores result indices.
`SILAutoDiffIndices` now stores result indices instead of a source index.
`@differentiable` SIL function types may now have multiple differentiability
result indices and `@noDerivative` resutls.
`@differentiable` AST function types do not have `@noDerivative` results (yet),
so this functionality is not exposed to users.
Resolves TF-689 and TF-1256.
Infrastructural support for TF-983: supporting differentiation of `apply`
instructions with multiple active semantic results.
* a new [immutable] attribute on ref_element_addr and ref_tail_addr
* new instructions: begin_cow_mutation and end_cow_mutation
These new instructions are intended to be used for the stdlib's COW containers, e.g. Array.
They allow more aggressive optimizations, especially for Array.
This became necessary after recent function type changes that keep
substituted generic function types abstract even after substitution to
correctly handle automatic opaque result type substitution.
Instead of performing the opaque result type substitution as part of
substituting the generic args the underlying type will now be reified as
part of looking at the parameter/return types which happens as part of
the function convention apis.
rdar://62560867
