AccessPathWithBase::compute can return a valid access path with unidentified base.
In such cases, we cannot LICM stores, because there is no base address to check if it is invariant
AccessPathDefUseTraversal accumulates the offsets that it has seen on
the def-use walk while visiting index_addr and casts.
This is quite tricky because either the AccessPath being matched or
the def-use chain could contain unknown offsets. We need to correctly
classify all cases as either an exact, inner, or overlapping match.
For SIL like this
%91 = integer_literal $Builtin.Int64, 0
%113 = builtin "truncOrBitCast_Int64_Word"(%91 : $Builtin.Int64) : $Builtin.Word
%115 = index_addr %114 : $*MyStruct, %113 : $Builtin.Word
%119 = struct_element_addr %115 : $*MyStruct, #MyStruct.klass
We have Path: (@Unknown,#0)
There are two issues
(1) When AccessPathDefUseTraversal::checkAndUpdateOffset visits the
struct_element_addr, it fails to pop the unknown offset from the
expected path but continues trying to handle the struct_element_addr
and hits an assert. This PR fixes this.
(2) If the builtin "truncOrBitCast_Int64_Word" comes from a literal,
we should not treat it as an unknown offset (given that the literal is
within the size of a word). Added a test for this case, we should be able
to identify the access path accurately in this.
Change the code generation patterns for `async let` bindings to use an ABI based on the following
functions:
- `swift_asyncLet_begin`, which starts an `async let` child task, but which additionally
now associates the `async let` with a caller-owned buffer to receive the result of the task.
This is intended to allow the task to emplace its result in caller-owned memory, allowing the
child task to be deallocated after completion without invalidating the result buffer.
- `swift_asyncLet_get[_throwing]`, which replaces `swift_asyncLet_wait[_throwing]`. Instead of
returning a copy of the value, this entry point concerns itself with populating the local buffer.
If the buffer hasn't been populated, then it awaits completion of the task and emplaces the
result in the buffer; otherwise, it simply returns. The caller can then read the result out of
its owned memory. These entry points are intended to be used before every read from the
`async let` binding, after which point the local buffer is guaranteed to contain an initialized
value.
- `swift_asyncLet_finish`, which replaces `swift_asyncLet_end`. Unlike `_end`, this variant
is async and will suspend the parent task after cancelling the child to ensure it finishes
before cleaning up. The local buffer will also be deinitialized if necessary. This is intended
to be used on exit from an `async let` scope, to handle cleaning up the local buffer if necessary
as well as cancelling, awaiting, and deallocating the child task.
- `swift_asyncLet_consume[_throwing]`, which combines `get` and `finish`. This will await completion
of the task, leaving the result value in the result buffer (or propagating the error, if it
throws), while destroying and deallocating the child task. This is intended as an optimization
for reading `async let` variables that are read exactly once by their parent task.
To avoid an epoch break with existing swiftinterfaces and ABI clients, the old builtins and entry
points are kept intact for now, but SILGen now only generates code using the new interface.
This new interface fixes several issues with the old async let codegen, including use-after-free
crashes if the `async let` was never awaited, and the inability to read from an `async let` variable
more than once.
rdar://77855176
Rather than using group task options constructed from the Swift parts
of the _Concurrency library and passed through `createAsyncTask`'s
options, introduce a separate builtin that always takes a group. Move
the responsibility for creating the options structure into IRGen, so
we don't need to expose the TaskGroupTaskOptionRecord type in Swift.
Introduce a builtin `createAsyncTask` that maps to `swift_task_create`,
and use that for the non-group task creation operations based on the
task-creation flags. `swift_task_create` and the thin function version
`swift_task_create_f` go through the dynamically-replaceable
`swift_task_create_common`, where all of the task creation logic is
present.
While here, move copying of task locals and the initial scheduling of
the task into `swift_task_create_common`, enabling by separate flags.
When an instruction is "deleted" from the SIL, it is put into the SILModule::scheduledForDeletion list.
The instructions in this list are eventually deleted for real in SILModule::flushDeletedInsts(), which is called by the pass manager after each pass run.
In other words: instruction deletion is deferred to the end of a pass.
This avoids dangling instruction pointers within the run of a pass and in analysis caches.
Note that the analysis invalidation mechanism ensures that analysis caches are invalidated before flushDeletedInsts().
Through various means, it is possible for a synchronous actor-isolated
function to escape to another concurrency domain and be called from
outside the actor. The problem existed previously, but has become far
easier to trigger now that `@escaping` closures and local functions
can be actor-isolated.
Introduce runtime detection of such data races, where a synchronous
actor-isolated function ends up being called from the wrong executor.
Do this by emitting an executor check in actor-isolated synchronous
functions, where we query the executor in thread-local storage and
ensure that it is what we expect. If it isn't, the runtime complains.
The runtime's complaints can be controlled with the environment
variable `SWIFT_UNEXPECTED_EXECUTOR_LOG_LEVEL`:
0 - disable checking
1 - warn when a data race is detected
2 - error and abort when a data race is detected
At an implementation level, this introduces a new concurrency runtime
entry point `_checkExpectedExecutor` that checks the given executor
(on which the function should always have been called) against the
executor on which is called (which is in thread-local storage). There
is a special carve-out here for `@MainActor` code, where we check
against the OS's notion of "main thread" as well, so that `@MainActor`
code can be called via (e.g.) the Dispatch library's
`DispatchQueue.main.async`.
The new SIL instruction `extract_executor` performs the lowering of an
actor down to its executor, which is implicit in the `hop_to_executor`
instruction. Extend the LowerHopToExecutor pass to perform said
lowering.
- stop storing the parent task in the TaskGroup at the .swift level
- make sure that swift_taskGroup_isCancelled is implied by the parent
task being cancelled
- make the TaskGroup structs frozen
- make the withTaskGroup functions inlinable
- remove swift_taskGroup_create
- teach IRGen to allocate memory for the task group
- don't deallocate the task group in swift_taskGroup_destroy
To achieve the allocation change, introduce paired create/destroy builtins.
Furthermore, remove the _swiftRetain and _swiftRelease functions and
several calls to them. Replace them with uses of the appropriate builtins.
I should probably change the builtins to return retained, since they're
working with a managed type, but I'll do that in a separate commit.
Since it is an always take, we know that the original value will always be
invalidated by the checked_cast_addr_br.
This also lets me use this to recognize simple cases of checked casts in
opt-remark-gen.
The immediate desire is to minimize the set of ABI dependencies
on the layout of an ExecutorRef. In addition to that, however,
I wanted to generally reduce the code size impact of an unsafe
continuation since it now requires accessing thread-local state,
and I wanted resumption to not have to create unnecessary type
metadata for the value type just to do the initialization.
Therefore, I've introduced a swift_continuation_init function
which handles the default initialization of a continuation
and returns a reference to the current task. I've also moved
the initialization of the normal continuation result into the
caller (out of the runtime), and I've moved the resumption-side
cmpxchg into the runtime (and prior to the task being enqueued).
In their previous form, the non-`_f` variants of these entry points were unused, and IRGen
lowered the `createAsyncTask` builtins to use the `_f` variants with a large amount of caller-side
codegen to manually unpack closure values. Amid all this, it also failed to make anyone responsible
for releasing the closure context after the task completed, causing every task creation to leak.
Redo the `swift_task_create_*` entry points to accept the two words of an async closure value
directly, and unpack the closure to get its invocation entry point and initial context size
inside the runtime. (Also get rid of the non-future `swift_task_create` variant, since it's unused
and it's subtly different in a lot of hairy ways from the future forms. Better to add it later
when it's needed than to have a broken unexercised version now.)
Ignore end_borrow as a user of the access path. I don't think there's
any value in viewing it as a use of the access path because we really
care about address reads and writes in this context, not object
lifetime extension. Treating it as an access path use clutters the analysis,
and I'm afraid it will inhibit optimization.
It would be more abstractly correct if this got DI support so
that we destroy the member if the constructor terminates
abnormally, but we can get to that later.
In derivatives of loops, no longer allocate boxes for indirect case payloads. Instead, use a custom pullback context in the runtime which contains a bump-pointer allocator.
When a function contains a differentiated loop, the closure context is a `Builtin.NativeObject`, which contains a `swift::AutoDiffLinearMapContext` and a tail-allocated top-level linear map struct (which represents the linear map struct that was previously directly partial-applied into the pullback). In branching trace enums, the payloads of previously indirect cases will be allocated by `swift::AutoDiffLinearMapContext::allocate` and stored as a `Builtin.RawPointer`.
AccessPath was treating init_enum_data_addr as an address base, which
is not ideal. It should be able to identify the underlying enum object
as the base. This issue was caught by LoadBorrowImmutabilityChecker
during SIL verification.
Instead handle init_enum_data_addr as a access projection that does
not affect the access path. I expect this SIL pattern to disappear
with SIL opaque values, but it still needs to be handled properly
after lowering addresses.
Functionality changes:
- any user of AccessPath now sees enum initialization stores as writes
to the underlying enum object
- SILGen now generates begin/end access markers for enum
initialization patterns. (Originally, we did not "see through"
init_enum_data_addr because we didn't want to generate these
markers, but that behavior was inconsistent and problematic).
Fixes rdar://70725514 fatal error encountered during compilation;
Unknown instruction: init_enum_data_addr)
`Builtin.createAsyncTask` takes flags, an optional parent task, and an
async/throwing function to execute, and passes it along to the
`swift_task_create_f` entry point to create a new (potentially child)
task, returning the new task and its initial context.
Implement a new builtin, `cancelAsyncTask()`, to cancel the given
asynchronous task. This lowers down to a call into the runtime
operation `swift_task_cancel()`.
Use this builtin to implement Task.Handle.cancel().
When casting from existentials to class - and vice versa - it can happen that a cast is not RC identity preserving (because of potential bridging).
This also affects mayRelease() of such cast instructions.
For details see the comments in SILDynamicCastInst::isRCIdentityPreserving().
This change also includes some refactoring: I centralized the logic in SILDynamicCastInst::isRCIdentityPreserving().
rdar://problem/70454804
For class storage AccessedStorage is now close to what some passes use
for RC identity, but it still does not look past wrapping references
in an Optional.
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.
Add AccesssedStorage::compute and computeInScope to mirror AccessPath.
Allow recovering the begin_access for Nested storage.
Adds AccessedStorage.visitRoots().
Things that have come up recently but are somewhat blocked on this:
- Moving AccessMarkerElimination down in the pipeline
- SemanticARCOpts correctness and improvements
- AliasAnalysis improvements
- LICM performance regressions
- RLE/DSE improvements
Begin to formalize the model for valid memory access in SIL. Ignoring
ownership, every access is a def-use chain in three parts:
object root -> formal access base -> memory operation address
AccessPath abstracts over this path and standardizes the identity of a
memory access throughout the optimizer. This abstraction is the basis
for a new AccessPathVerification.
With that verification, we now have all the properties we need for the
type of analysis requires for exclusivity enforcement, but now
generalized for any memory analysis. This is suitable for an extremely
lightweight analysis with no side data structures. We currently have a
massive amount of ad-hoc memory analysis throughout SIL, which is
incredibly unmaintainable, bug-prone, and not performance-robust. We
can begin taking advantage of this verifably complete model to solve
that problem.
The properties this gives us are:
Access analysis must be complete over memory operations: every memory
operation needs a recognizable valid access. An access can be
unidentified only to the extent that it is rooted in some non-address
type and we can prove that it is at least *not* part of an access to a
nominal class or global property. Pointer provenance is also required
for future IRGen-level bitfield optimizations.
Access analysis must be complete over address users: for an identified
object root all memory accesses including subobjects must be
discoverable.
Access analysis must be symmetric: use-def and def-use analysis must
be consistent.
AccessPath is merely a wrapper around the existing accessed-storage
utilities and IndexTrieNode. Existing passes already very succesfully
use this approach, but in an ad-hoc way. With a general utility we
can:
- update passes to use this approach to identify memory access,
reducing the space and time complexity of those algorithms.
- implement an inexpensive on-the-fly, debug mode address lifetime analysis
- implement a lightweight debug mode alias analysis
- ultimately improve the power, efficiency, and maintainability of
full alias analysis
- make our type-based alias analysis sensistive to the access path
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.
