This document defines the Canonical ABI used to convert between the values and functions of components in the Component Model and the values and functions of modules in Core WebAssembly. See the AST explainer for a walkthrough of the static structure of a component and the async explainer for a high-level description of the async model being specified here.
The Canonical ABI specifies, for each component function signature, a
corresponding core function signature and the process for reading
component-level values into and out of linear memory. While a full formal
specification would specify the Canonical ABI in terms of macro-expansion into
Core WebAssembly instructions augmented with a new set of (spec-internal)
administrative instructions, the informal presentation here instead specifies
the process in terms of Python code that would be logically executed at
validation- and run-time by a component model implementation. The Python code
is presented by interleaving definitions with descriptions and eliding some
boilerplate. For a complete listing of all Python definitions in a single
executable file with a small unit test suite, see the
canonical-abi directory.
The convention followed by the Python code below is that all traps are raised
by explicit trap()/trap_if() calls; Python assert() statements should
never fire and are only included as hints to the reader. Similarly, there
should be no uncaught Python exceptions.
While the Python code appears to perform a copy as part of lifting the contents of linear memory into high-level Python values, a normal implementation should never need to make this extra intermediate copy. This claim is expanded upon below.
Lastly, independently of Python, the Canonical ABI defined below assumes that
out-of-memory conditions (such as memory.grow returning -1 from within
realloc) will trap (via unreachable). This significantly simplifies the
Canonical ABI by avoiding the need to support the complicated protocols
necessary to support recovery in the middle of nested allocations. In the MVP,
for large allocations that can OOM, streams would usually
be the appropriate type to use and streams will be able to explicitly express
failure in their type. Post-MVP, adapter functions would allow fully custom
OOM handling for all component-level types, allowing a toolchain to
intentionally propagate OOM into the appropriate explicit return value of the
function's declared return type.
In the explainer, component value types are classified as
either fundamental or specialized, where the specialized value types are
defined by expansion into fundamental value types. In most cases, the canonical
ABI of a specialized value type is the same as its expansion so, to avoid
repetition, the other definitions below use the following despecialize
function to replace specialized value types with their expansion:
def despecialize(t):
match t:
case Tuple(ts) : return Record([ Field(str(i), t) for i,t in enumerate(ts) ])
case Enum(labels) : return Variant([ Case(l, None) for l in labels ])
case Option(t) : return Variant([ Case("none", None), Case("some", t) ])
case Result(ok, error) : return Variant([ Case("ok", ok), Case("error", error) ])
case _ : return tThe specialized value types string and flags are missing from this list
because they are given specialized canonical ABI representations distinct from
their respective expansions.
Each value type is assigned an alignment which is used by subsequent
Canonical ABI definitions. Presenting the definition of alignment piecewise,
we start with the top-level case analysis:
def alignment(t):
match despecialize(t):
case Bool() : return 1
case S8() | U8() : return 1
case S16() | U16() : return 2
case S32() | U32() : return 4
case S64() | U64() : return 8
case F32() : return 4
case F64() : return 8
case Char() : return 4
case String() : return 4
case List(t, l) : return alignment_list(t, l)
case Record(fields) : return alignment_record(fields)
case Variant(cases) : return alignment_variant(cases)
case Flags(labels) : return alignment_flags(labels)
case Own(_) | Borrow(_) : return 4List alignment is the same as tuple alignment when the length is fixed and otherwise uses the alignment of pointers.
def alignment_list(elem_type, maybe_length):
if maybe_length is not None:
return alignment(elem_type)
return 4Record alignment is tuple alignment, with the definitions split for reuse below:
def alignment_record(fields):
a = 1
for f in fields:
a = max(a, alignment(f.t))
return aAs an optimization, variant discriminants are represented by the smallest integer
covering the number of cases in the variant (with cases numbered in order from
0 to len(cases)-1). Depending on the payload type, this can allow more
compact representations of variants in memory. This smallest integer type is
selected by the following function, used above and below:
def alignment_variant(cases):
return max(alignment(discriminant_type(cases)), max_case_alignment(cases))
def discriminant_type(cases):
n = len(cases)
assert(0 < n < (1 << 32))
match math.ceil(math.log2(n)/8):
case 0: return U8()
case 1: return U8()
case 2: return U16()
case 3: return U32()
def max_case_alignment(cases):
a = 1
for c in cases:
if c.t is not None:
a = max(a, alignment(c.t))
return aAs an optimization, flags are represented as packed bit-vectors. Like variant
discriminants, flags use the smallest integer that fits all the bits, falling
back to sequences of i32s when there are more than 32 flags.
def alignment_flags(labels):
n = len(labels)
assert(0 < n <= 32)
if n <= 8: return 1
if n <= 16: return 2
return 4Handle types are passed as i32 indices into the Table[HandleElem]
introduced below.
Each value type is also assigned an elem_size which is the number of bytes
used when values of the type are stored as elements of a list. Having this
byte size be a static property of the type instead of attempting to use a
variable-length element-encoding scheme both simplifies the implementation and
maps well to languages which represent lists as random-access arrays. Empty
types, such as records with no fields, are not permitted, to avoid
complications in source languages.
def elem_size(t):
match despecialize(t):
case Bool() : return 1
case S8() | U8() : return 1
case S16() | U16() : return 2
case S32() | U32() : return 4
case S64() | U64() : return 8
case F32() : return 4
case F64() : return 8
case Char() : return 4
case String() : return 8
case List(t, l) : return elem_size_list(t, l)
case Record(fields) : return elem_size_record(fields)
case Variant(cases) : return elem_size_variant(cases)
case Flags(labels) : return elem_size_flags(labels)
case Own(_) | Borrow(_) : return 4
def elem_size_list(elem_type, maybe_length):
if maybe_length is not None:
return maybe_length * elem_size(elem_type)
return 8
def elem_size_record(fields):
s = 0
for f in fields:
s = align_to(s, alignment(f.t))
s += elem_size(f.t)
assert(s > 0)
return align_to(s, alignment_record(fields))
def align_to(ptr, alignment):
return math.ceil(ptr / alignment) * alignment
def elem_size_variant(cases):
s = elem_size(discriminant_type(cases))
s = align_to(s, max_case_alignment(cases))
cs = 0
for c in cases:
if c.t is not None:
cs = max(cs, elem_size(c.t))
s += cs
return align_to(s, alignment_variant(cases))
def elem_size_flags(labels):
n = len(labels)
assert(0 < n <= 32)
if n <= 8: return 1
if n <= 16: return 2
return 4The subsequent definitions of lifting and lowering depend on three kinds of ambient information:
- static ABI options supplied via
canonopt - dynamic state in the containing component instance
- dynamic state in the current task
These sources of ambient context are stored as the respective opts, inst
and task fields of the Context object:
class Context:
opts: CanonicalOptions
inst: ComponentInstance
task: Task
def __init__(self, opts, inst, task):
self.opts = opts
self.inst = inst
self.task = taskThe cx parameter in functions below refers to the ambient Context. The
Task and Subtask classes derive Context and thus having a task or
subtask also establishes the ambient Context.
The opts field of CallContext contains all the possible canonopt
immediates that can be passed to the canon definition being implemented.
@dataclass
class CanonicalOptions:
memory: Optional[bytearray] = None
string_encoding: Optional[str] = None
realloc: Optional[Callable] = None
post_return: Optional[Callable] = None
sync: bool = True # = !canonopt.async
callback: Optional[Callable] = None(Note that the async canonopt is inverted to sync here for the practical
reason that async is a keyword and most branches below want to start with the
sync = True case.)
The inst field of CallContext points to the component instance which the
canon-generated function is closed over. Component instances contain all the
core wasm instance as well as some extra state that is used exclusively by the
Canonical ABI and introduced below as the fields are used.
class ComponentInstance:
handles: HandleTables
async_subtasks: Table[AsyncSubtask]
may_leave: bool
backpressure: bool
calling_sync_export: bool
num_tasks: int
interruptible: asyncio.Event
pending_tasks: list[tuple[Task, asyncio.Future]]
def __init__(self):
self.handles = HandleTables()
self.async_subtasks = Table[AsyncSubtask]()
self.may_leave = True
self.backpressure = False
self.calling_sync_export = False
self.num_tasks = 0
self.interruptible = asyncio.Event()
self.interruptible.set()
self.pending_tasks = []The HandleTables class stored in ComponentInstance.handles maps
ResourceTypes to Tables of HandleElems (defined next), establishing a
separate i32-indexed array per resource type:
class HandleTables:
rt_to_table: MutableMapping[ResourceType, Table[HandleElem]]
def __init__(self):
self.rt_to_table = dict()
def table(self, rt):
if rt not in self.rt_to_table:
self.rt_to_table[rt] = Table[HandleElem]()
return self.rt_to_table[rt]
def get(self, rt, i):
return self.table(rt).get(i)
def add(self, rt, h):
return self.table(rt).add(h)
def remove(self, rt, i):
return self.table(rt).remove(i)While this Python code performs a dynamic hash-table lookup on each handle
table access, as we'll see below, the rt parameter is always statically known
such that a normal implementation can statically enumerate all Table objects
at compile time and then route the calls to get, add and remove to the
correct Table at the callsite. The net result is that each component instance
will contain one handle table per resource type used by the component, with
each compiled adapter function accessing the correct handle table as-if it were
a global variable.
The ResourceType class represents a concrete resource type that has been
created by the component instance impl. ResourceType objects are used as
keys by HandleTables above and thus we assume that Python object identity
corresponds to resource type equality, as defined by [type checking] rules.
class ResourceType(Type):
impl: ComponentInstance
dtor: Optional[Callable]
dtor_sync: bool
dtor_callback: Optional[Callable]
def __init__(self, impl, dtor = None, dtor_sync = True, dtor_callback = None):
self.impl = impl
self.dtor = dtor
self.dtor_sync = dtor_sync
self.dtor_callback = dtor_callbackThe Table class, used by HandleTables above, encapsulates a single
mutable, growable array of generic elements, indexed by Core WebAssembly
i32s.
ElemT = TypeVar('ElemT')
class Table(Generic[ElemT]):
array: list[Optional[ElemT]]
free: list[int]
def __init__(self):
self.array = [None]
self.free = []
def get(self, i):
trap_if(i >= len(self.array))
trap_if(self.array[i] is None)
return self.array[i]
def add(self, e):
if self.free:
i = self.free.pop()
assert(self.array[i] is None)
self.array[i] = e
else:
i = len(self.array)
trap_if(i >= 2**30)
self.array.append(e)
return i
def remove(self, i):
e = self.get(i)
self.array[i] = None
self.free.append(i)
return eTable maintains a dense array of elements that can contain holes created by
the remove method (defined below). When table elements are accessed (e.g., by
canon_lift and resource.rep, below), there are thus both a bounds check and
hole check necessary. Upon initialization, table element 0 is allocated and
set to None, effectively reserving index 0 which is both useful for
catching null/uninitialized accesses and allowing 0 to serve as a sentinel
value.
The add and remove methods work together to maintain a free list of holes
that are used in preference to growing the table. The free list is represented
as a Python list here, but an optimizing implementation could instead store the
free list in the free elements of array.
The limit of 2**30 ensures that the high 2 bits of table indices are unset
and available for other use in guest code (e.g., for tagging, packed words or
sentinel values).
The HandleElem class defines the elements of the per-resource-type Tables
stored in HandleTables:
class HandleElem:
rep: int
own: bool
scope: Optional[Task]
lend_count: int
def __init__(self, rep, own, scope = None):
self.rep = rep
self.own = own
self.scope = scope
self.lend_count = 0The rep field of HandleElem stores the resource representation (currently
fixed to be an i32) passed to resource.new.
The own field indicates whether this element was created from an own type
(or, if false, a borrow type).
The scope field stores the Task that created the borrowed handle. When a
component only uses sync-lifted exports, due to lack of reentrance, there is at
most one Task alive in a component instance at any time and thus an
optimizing implementation doesn't need to store the Task per HandleElem.
The lend_count field maintains a conservative approximation of the number of
live handles that were lent from this own handle (by calls to borrow-taking
functions). This count is maintained by the ImportCall bookkeeping functions
(above) and is ensured to be zero when an own handle is dropped.
An optimizing implementation can enumerate the canonical definitions present
in a component to statically determine that a given resource type's handle
table only contains own or borrow handles and then, based on this,
statically eliminate the own and the lend_count xor scope fields,
and guards thereof.
Additional runtime state is required to implement the canonical built-ins and
check that callers and callees uphold their respective parts of the call
contract. This additional call state derives from CallContext, adding extra
mutable fields. There are two subclasses of CallContext: Task, which is
created by canon_lift and Subtask, which is created by canon_lower.
Additional sync-/async-specialized mutable state is added by the AsyncTask
and AsyncSubtask subclasses.
The Task class and its subclasses depend on the following type definitions:
class AsyncCallState(IntEnum):
STARTING = 0
STARTED = 1
RETURNED = 2
DONE = 3
class EventCode(IntEnum):
CALL_STARTING = AsyncCallState.STARTING
CALL_STARTED = AsyncCallState.STARTED
CALL_RETURNED = AsyncCallState.RETURNED
CALL_DONE = AsyncCallState.DONE
YIELDED = 4
EventTuple = tuple[EventCode, int]
EventCallback = Callable[[], EventTuple]
class OnBlockResult(IntEnum):
BLOCKED = 0
COMPLETED = 1The AsyncCallState enum describes the linear sequence of states that an async
call necessarily transitions through: STARTING,
STARTED, RETURNING and DONE. The EventCode enum
shares common code values with AsyncCallState to define the set of integer
event codes that are delivered to waiting or polling tasks.
The current_Task global holds an asyncio.Lock that is used to prevent the
Python runtime from arbitrarily switching between Python coroutines (async def functions). Instead, switching between execution can only happen at
specific points where the current_task lock is released and reacquired (viz.,
Task.suspend and Task.exit below). Without this lock, Python's normal async
scheduler would be able to switch between concurrent Tasks at every
spec-internal await, which would end up looking like preemptive
multi-threading to guest code. Alternatively, if Python had standard-library
fibers, fibers could be used instead of asyncio, obviating the need for
this lock.
current_task = asyncio.Lock()A Task object is created for each call to canon_lift and is implicitly
threaded through all core function calls. This implicit Task parameter
represents the "current task". A Task is-a Context, with its opts
corresponding to the canonopt supplied to the canon lift definition that
the Task was created for.
class Task(Context):
caller: Optional[Task]
on_block: asyncio.Future[OnBlockResult]
borrow_count: int
events: list[EventCallback]
has_events: asyncio.Event
num_async_subtasks: int
def __init__(self, opts, inst, caller, on_block):
super().__init__(opts, inst, self)
self.caller = caller
self.on_block = on_block
self.borrow_count = 0
self.events = []
self.has_events = asyncio.Event()
self.num_async_subtasks = 0The fields of Task are introduced in groups of related Task methods next.
Using a conservative syntactic analysis of the component-level definitions of a
linked component DAG, an optimizing implementation can statically eliminate
these fields when the particular feature (borrow handles, async imports) is
not used.
The trap_if_on_stack helper method is called (below) to prevent reentrance.
The definition uses the caller field which points to the task's supertask in
the async call tree defined by Component Model structured concurrency.
def trap_if_on_the_stack(self, inst):
c = self.caller
while c is not None:
trap_if(c.inst is inst)
c = c.callerBy analyzing a linked component DAG, an optimized implementation can avoid the
O(n) loop in trap_if_on_the_stack:
- Reentrance by a child component can (often) be statically ruled out when the parent component doesn't both lift and lower the child's imports and exports (i.e., "donut wrapping").
- Reentrance of the root component by the host can either be asserted not to happen or be tracked in a per-root-component-instance flag.
- When a potentially-reenterable child component only lifts and lowers synchronously, reentrance can be tracked in a per-component-instance flag.
- For the remaining cases, the live instances on the stack can be maintained in a packed bit-vector (assigning each potentially-reenterable async component instance a static bit position) that is passed by copy from caller to callee.
The enter method is called immediately after constructing a Task and, along
with may_enter and may_start_pending_task, implements backpressure. If a
Task tries to enter when may_enter is false, the Task suspends itself
(via suspend, shown next) and goes into a pending_tasks queue, waiting to
be unblocked by another task calling maybe_start_pending_task. One key
property of this backpressure scheme is that pending_tasks are only dequeued
one at a time, ensuring that if an overloaded component instance enables and
then disables backpressure, there will not be an unstoppable thundering herd of
pending tasks started all at once that OOM the component before it can
re-enable backpressure.
async def enter(self):
assert(current_task.locked())
self.trap_if_on_the_stack(self.inst)
if not self.may_enter(self) or self.inst.pending_tasks:
f = asyncio.Future()
self.inst.pending_tasks.append((self, f))
await self.suspend(f)
if self.opts.sync:
self.inst.calling_sync_export = True
self.inst.num_tasks += 1
def may_enter(self, pending_task):
return not self.inst.backpressure and \
not self.inst.calling_sync_export and \
not (pending_task.opts.sync and self.inst.num_tasks > 0) and \
self.inst.interruptible.is_set()
def maybe_start_pending_task(self):
if self.inst.pending_tasks:
pending_task, future = self.inst.pending_tasks[0]
if self.may_enter(pending_task):
self.inst.pending_tasks.pop(0)
future.set_result(None)Considering the 4 conditions in may_enter:
- The
backpressurecheck reflects explicit backpressure triggered by core wasm code (viacanon task.backpressure, defined below). - The next two conditions ensure that synchronous exports do not interleave execution with any other task in the same component instance. The first condition prevents async calls from starting while there is a sync call in flight, while the second condition prevents sync calls from starting while there are any other calls in flight.
- The
interruptiblecheck is automatically set and cleared before and after any synchronouscanonbuilt-in called by core wasm, temporarily preventing any interleaved execution during synchronous call.
The suspend method, called by enter, wait and yield_, takes a future to
await and allows other tasks to make progress in the meantime. Since the
current_task lock is shared by all linked component instances, releasing then
acquiring it allows switching to any task in the same or different component
instance. The final loop in suspend waits until interruptible is set to
ensure that tasks that have been suspended cannot resume execution. Together
with the guard may_enter, this ensures that clearing interruptible ensures
traditional synchronous execution (for use in synchronous canon definitions
below).
async def suspend(self, future):
assert(self.inst.interruptible.is_set())
if not self.on_block.done():
self.on_block.set_result(OnBlockResult.BLOCKED)
else:
current_task.release()
v = await future
await current_task.acquire()
while not self.inst.interruptible.is_set():
current_task.release()
await self.inst.interruptible.wait()
await current_task.acquire()
return vWhen there is an async-lowered caller waiting on the stack, the on_block
field will point to an unresolved future. In this case, suspend sets the
result of on_block and leaves current_task locked so that control flow
transfers deterministically to async-lowered caller (in canon_lower,
defined below).
The suspend method is so named because it uses Python's async machinery to
emulate the suspend instruction of the stack-switching proposal. Similarly,
the handle_blocking method below emulates the combination of the cont.new
and resume instructions where resume handles the on_block effect raised
by suspend. The combined effect of handle_blocking and suspend is that if
the callee blocks, control flow is immediately transferred from the callee to
the caller of handle_blocking, which can then use the return value to
determine whether the call completed eagerly or blocked.
async def handle_blocking(self, callee) -> OnBlockResult:
on_block = asyncio.Future()
async def cont():
await callee(on_block)
if not on_block.done():
on_block.set_result(OnBlockResult.COMPLETED)
else:
current_task.release()
asyncio.create_task(cont())
return await on_blockWhile a task is running, it may call wait (via canon task.wait or, when a
callback is present, by returning to the event loop) to block until there is
an event that will allow the Task to make progress. If there are no
currently-pending events in the events list, wait suspends until the
has_events asyncio.Event is set by notify. Once there is at least one
event in the list, an event is popped non-deterministically to provide the
runtime flexibility in event queueing and delivery.
async def wait(self) -> EventTuple:
self.maybe_start_pending_task()
if not self.events:
await self.suspend(self.has_events.wait())
return self.next_event()
def next_event(self) -> EventTuple:
assert(self.events)
if DETERMINISTIC_PROFILE:
i = 0
else:
i = random.randrange(len(self.events))
event = self.events.pop(i)
if not self.events:
self.has_events.clear()
return event()
def notify(self, event: EventCallback):
self.events.append(event)
self.has_events.set()Note that events are represented as first-class functions that are called by
next_event to produce the tuple of scalar values that are actually delivered
to core wasm. This allows an event source to report the latest status when the
event is handed to the core wasm code instead of the status when the event was
first generated. This allows multiple redundant events to be collapsed into
one, reducing overhead.
Although this Python code represents events as a list of closures, an optimizing implementation should be able to avoid actually allocating these things and instead embed a linked list of "ready" events into the table elements associated with the events.
In addition to wait, the current task can also use poll (via canon task.poll),
to not block and not allow the runtime to switch to another task:
def poll(self) -> Optional[EventTuple]:
if self.events:
return self.next_event()
return NoneLastly, a task may cooperatively yield (via canon task.yield), allowing the
runtime to switch execution to another task without having to wait for any
external I/O (as emulated in the Python code by awaiting sleep(0):
async def yield_(self):
self.maybe_start_pending_task()
await self.suspend(asyncio.sleep(0))All Tasks (whether lifted async or not) are allowed to call async-lowered
imports. Calling an async-lowered import creates an AsyncSubtask (defined
below) which is stored in the current component instance's async_subtasks
table and tracked by the current task's num_async_subtasks counter, which is
guarded to be 0 in Task.exit (below) to ensure structured concurrency.
def add_async_subtask(self, subtask):
assert(subtask.task is self and not subtask.notify_supertask)
subtask.notify_supertask = True
self.num_async_subtasks += 1
return self.inst.async_subtasks.add(subtask)The notify_supertask flag signals to AsyncSubtask methods (defined below)
to notify this Task when the lowered callee makes progress.
The borrow_count field is used by the following methods to track the number
of borrowed handles that were passed as parameters to the export that have not
yet been dropped (and thus might dangle if the caller destroys the resource
after this export call finishes):
def create_borrow(self):
self.borrow_count += 1
def drop_borrow(self):
assert(self.borrow_count > 0)
self.borrow_count -= 1Lastly, when a task exits, the runtime enforces the guard conditions mentioned above and allows pending tasks to start.
def exit(self):
assert(current_task.locked())
assert(not self.events)
assert(self.inst.num_tasks >= 1)
trap_if(self.borrow_count != 0)
trap_if(self.num_async_subtasks != 0)
self.inst.num_tasks -= 1
if self.opts.sync:
self.inst.calling_sync_export = False
self.maybe_start_pending_task()While canon_lift creates Tasks, canon_lower creates Subtask objects.
Like Task, Subtask is a subclass of Context, gets its opts from the
canon lower. Importantly, the Context.task field refers to the current
task at the point of the canon lower call, thereby linking a subtask to its
supertask.
class Subtask(Context):
lenders: list[HandleElem]
def __init__(self, opts, task):
super().__init__(opts, task.inst, task)
self.lenders = []
def track_owning_lend(self, lending_handle):
assert(lending_handle.own)
lending_handle.lend_count += 1
self.lenders.append(lending_handle)
def finish(self):
for h in self.lenders:
h.lend_count -= 1A Subtask tracks the owned handles that have been lent for the duration of
the call, ensuring that the caller doesn't drop them during the call (which
might create a dangling borrowed handle in the callee). Note, the lenders
list usually has a fixed size (in all cases except when a function signature
has borrows in lists) and thus can be stored inline in the native stack
frame.
The following AsyncTask/AsyncSubtask classes extend the preceding
Task/Subtask classes with additional state and methods that apply only to
the async case.
The first 3 fields of AsyncTask are simply immutable copies of
arguments/immediates passed to canon_lift and are used by the task.start
and task.return built-ins below. The last field is used to check the
above-mentioned state machine transitions from methods that are called by
task.start, task.return and canon_lift below.
class AsyncTask(Task):
ft: FuncType
on_start: Callable
on_return: Callable
state: AsyncCallState
def __init__(self, opts, inst, caller, on_block, ft, on_start, on_return):
super().__init__(opts, inst, caller, on_block)
self.ft = ft
self.on_start = on_start
self.on_return = on_return
self.state = AsyncCallState.STARTING
def start(self):
trap_if(self.state != AsyncCallState.STARTING)
self.state = AsyncCallState.STARTED
def return_(self):
trap_if(self.state != AsyncCallState.STARTED)
self.state = AsyncCallState.RETURNED
def exit(self):
trap_if(self.state != AsyncCallState.RETURNED)
self.state = AsyncCallState.DONE
super().exit()Finally, the AsyncSubtask class extends Subtask with fields that are used
by the methods of Task, as shown above. AsyncSubtasks have the same linear
state machine as AsyncTasks, except that the state transitions are guaranteed
by the Canonical ABI to happen in the right order, as asserted by the methods
below.
class AsyncSubtask(Subtask):
state: AsyncCallState
notify_supertask: bool
enqueued: bool
def __init__(self, opts, task):
super().__init__(opts, task)
self.state = AsyncCallState.STARTING
self.notify_supertask = False
self.enqueued = False
def maybe_notify_supertask(self):
if self.notify_supertask and not self.enqueued:
self.enqueued = True
def event():
self.enqueued = False
return (EventCode(self.state), self.inst.async_subtasks.array.index(self))
self.task.notify(event)
def start(self):
assert(self.state == AsyncCallState.STARTING)
self.state = AsyncCallState.STARTED
self.maybe_notify_supertask()
def return_(self):
assert(self.state == AsyncCallState.STARTED)
self.state = AsyncCallState.RETURNED
self.maybe_notify_supertask()
def finish(self):
super().finish()
assert(self.state == AsyncCallState.RETURNED)
self.state = AsyncCallState.DONE
self.maybe_notify_supertask()The maybe_notify_supertask method only sends events to the supertask if this
AsyncSubtask actually blocked and got added to the async_subtasks table
(as signalled by notify_supertask). Additionally, maybe_notify_supertask
uses the enqueued flag and the fact that "events" are first-class functions
to collapse N events down to 1 if a subtask advances state multiple times
before the supertask receives the event.
The load function defines how to read a value of a given value type t
out of linear memory starting at offset ptr, returning the value represented
as a Python value. Presenting the definition of load piecewise, we start with
the top-level case analysis:
def load(cx, ptr, t):
assert(ptr == align_to(ptr, alignment(t)))
assert(ptr + elem_size(t) <= len(cx.opts.memory))
match despecialize(t):
case Bool() : return convert_int_to_bool(load_int(cx, ptr, 1))
case U8() : return load_int(cx, ptr, 1)
case U16() : return load_int(cx, ptr, 2)
case U32() : return load_int(cx, ptr, 4)
case U64() : return load_int(cx, ptr, 8)
case S8() : return load_int(cx, ptr, 1, signed=True)
case S16() : return load_int(cx, ptr, 2, signed=True)
case S32() : return load_int(cx, ptr, 4, signed=True)
case S64() : return load_int(cx, ptr, 8, signed=True)
case F32() : return decode_i32_as_float(load_int(cx, ptr, 4))
case F64() : return decode_i64_as_float(load_int(cx, ptr, 8))
case Char() : return convert_i32_to_char(cx, load_int(cx, ptr, 4))
case String() : return load_string(cx, ptr)
case List(t, l) : return load_list(cx, ptr, t, l)
case Record(fields) : return load_record(cx, ptr, fields)
case Variant(cases) : return load_variant(cx, ptr, cases)
case Flags(labels) : return load_flags(cx, ptr, labels)
case Own() : return lift_own(cx, load_int(cx, ptr, 4), t)
case Borrow() : return lift_borrow(cx, load_int(cx, ptr, 4), t)Integers are loaded directly from memory, with their high-order bit interpreted according to the signedness of the type.
def load_int(cx, ptr, nbytes, signed = False):
return int.from_bytes(cx.opts.memory[ptr : ptr+nbytes], 'little', signed=signed)Integer-to-boolean conversions treats 0 as false and all other bit-patterns
as true:
def convert_int_to_bool(i):
assert(i >= 0)
return bool(i)Floats are loaded directly from memory, with the sign and payload information of NaN values discarded. Consequently, there is only one unique NaN value per floating-point type. This reflects the practical reality that some languages and protocols do not preserve these bits. In the Python code below, this is expressed as canonicalizing NaNs to a particular bit pattern.
See the comments about lowering of float values for a discussion of possible optimizations.
DETERMINISTIC_PROFILE = False # or True
CANONICAL_FLOAT32_NAN = 0x7fc00000
CANONICAL_FLOAT64_NAN = 0x7ff8000000000000
def canonicalize_nan32(f):
if math.isnan(f):
f = core_f32_reinterpret_i32(CANONICAL_FLOAT32_NAN)
assert(math.isnan(f))
return f
def canonicalize_nan64(f):
if math.isnan(f):
f = core_f64_reinterpret_i64(CANONICAL_FLOAT64_NAN)
assert(math.isnan(f))
return f
def decode_i32_as_float(i):
return canonicalize_nan32(core_f32_reinterpret_i32(i))
def decode_i64_as_float(i):
return canonicalize_nan64(core_f64_reinterpret_i64(i))
def core_f32_reinterpret_i32(i):
return struct.unpack('<f', struct.pack('<I', i))[0] # f32.reinterpret_i32
def core_f64_reinterpret_i64(i):
return struct.unpack('<d', struct.pack('<Q', i))[0] # f64.reinterpret_i64An i32 is converted to a char (a Unicode Scalar Value) by dynamically
testing that its unsigned integral value is in the valid Unicode Code Point
range and not a Surrogate:
def convert_i32_to_char(cx, i):
assert(i >= 0)
trap_if(i >= 0x110000)
trap_if(0xD800 <= i <= 0xDFFF)
return chr(i)Strings are loaded from two i32 values: a pointer (offset in linear memory)
and a number of bytes. There are three supported string encodings in canonopt:
UTF-8, UTF-16 and latin1+utf16. This last options allows a dynamic
choice between Latin-1 and UTF-16, indicated by the high bit of the second
i32. String values include their original encoding and byte length as a
"hint" that enables store_string (defined below) to make better up-front
allocation size choices in many cases. Thus, the value produced by
load_string isn't simply a Python str, but a tuple containing a str,
the original encoding and the original byte length.
def load_string(cx, ptr):
begin = load_int(cx, ptr, 4)
tagged_code_units = load_int(cx, ptr + 4, 4)
return load_string_from_range(cx, begin, tagged_code_units)
UTF16_TAG = 1 << 31
def load_string_from_range(cx, ptr, tagged_code_units):
match cx.opts.string_encoding:
case 'utf8':
alignment = 1
byte_length = tagged_code_units
encoding = 'utf-8'
case 'utf16':
alignment = 2
byte_length = 2 * tagged_code_units
encoding = 'utf-16-le'
case 'latin1+utf16':
alignment = 2
if bool(tagged_code_units & UTF16_TAG):
byte_length = 2 * (tagged_code_units ^ UTF16_TAG)
encoding = 'utf-16-le'
else:
byte_length = tagged_code_units
encoding = 'latin-1'
trap_if(ptr != align_to(ptr, alignment))
trap_if(ptr + byte_length > len(cx.opts.memory))
try:
s = cx.opts.memory[ptr : ptr+byte_length].decode(encoding)
except UnicodeError:
trap()
return (s, cx.opts.string_encoding, tagged_code_units)Lists and records are loaded by recursively loading their elements/fields:
def load_list(cx, ptr, elem_type, maybe_length):
if maybe_length is not None:
return load_list_from_valid_range(cx, ptr, maybe_length, elem_type)
begin = load_int(cx, ptr, 4)
length = load_int(cx, ptr + 4, 4)
return load_list_from_range(cx, begin, length, elem_type)
def load_list_from_range(cx, ptr, length, elem_type):
trap_if(ptr != align_to(ptr, alignment(elem_type)))
trap_if(ptr + length * elem_size(elem_type) > len(cx.opts.memory))
return load_list_from_valid_range(cx, ptr, length, elem_type)
def load_list_from_valid_range(cx, ptr, length, elem_type):
a = []
for i in range(length):
a.append(load(cx, ptr + i * elem_size(elem_type), elem_type))
return a
def load_record(cx, ptr, fields):
record = {}
for field in fields:
ptr = align_to(ptr, alignment(field.t))
record[field.label] = load(cx, ptr, field.t)
ptr += elem_size(field.t)
return recordAs a technical detail: the align_to in the loop in load_record is
guaranteed to be a no-op on the first iteration because the record as
a whole starts out aligned (as asserted at the top of load).
Variants are loaded using the order of the cases in the type to determine the
case index, assigning 0 to the first case, 1 to the next case, etc. To
support the subtyping allowed by refines, a lifted variant value semantically
includes a full ordered list of its refines case labels so that the lowering
code (defined below) can search this list to find a case label it knows about.
While the code below appears to perform case-label lookup at runtime, a normal
implementation can build the appropriate index tables at compile-time so that
variant-passing is always O(1) and not involving string operations.
def load_variant(cx, ptr, cases):
disc_size = elem_size(discriminant_type(cases))
case_index = load_int(cx, ptr, disc_size)
ptr += disc_size
trap_if(case_index >= len(cases))
c = cases[case_index]
ptr = align_to(ptr, max_case_alignment(cases))
case_label = case_label_with_refinements(c, cases)
if c.t is None:
return { case_label: None }
return { case_label: load(cx, ptr, c.t) }
def case_label_with_refinements(c, cases):
label = c.label
while c.refines is not None:
c = cases[find_case(c.refines, cases)]
label += '|' + c.label
return label
def find_case(label, cases):
matches = [i for i,c in enumerate(cases) if c.label == label]
assert(len(matches) <= 1)
if len(matches) == 1:
return matches[0]
return -1Flags are converted from a bit-vector to a dictionary whose keys are
derived from the ordered labels of the flags type. The code here takes
advantage of Python's support for integers of arbitrary width.
def load_flags(cx, ptr, labels):
i = load_int(cx, ptr, elem_size_flags(labels))
return unpack_flags_from_int(i, labels)
def unpack_flags_from_int(i, labels):
record = {}
for l in labels:
record[l] = bool(i & 1)
i >>= 1
return recordown handles are lifted by removing the handle from the current component
instance's handle table, so that ownership is transferred to the lowering
component. The lifting operation fails if unique ownership of the handle isn't
possible, for example if the index was actually a borrow or if the own
handle is currently being lent out as borrows.
def lift_own(cx, i, t):
h = cx.inst.handles.remove(t.rt, i)
trap_if(h.lend_count != 0)
trap_if(not h.own)
return h.repThe abstract lifted value for handle types is currently just the internal
resource representation i32, which is kept opaque from the receiving
component (it's stored in the handle table and only accessed indirectly via
index). (This assumes that resource representations are immutable. If
representations were to become mutable, the address of the mutable cell would
be passed as the lifted value instead.)
In contrast to own, borrow handles are lifted by reading the representation
from the source handle, leaving the source handle intact in the current
component instance's handle table:
def lift_borrow(cx, i, t):
assert(isinstance(cx, Subtask))
h = cx.inst.handles.get(t.rt, i)
if h.own:
cx.track_owning_lend(h)
return h.repThe track_owning_lend call to CallContext participates in the enforcement
of the dynamic borrow rules, which keep the source own handle alive until the
end of the call (as an intentionally-conservative upper bound on how long the
borrow handle can be held). This tracking is only required when h is an
own handle because, when h is a borrow handle, this tracking has already
happened (when the originating own handle was lifted) for a strictly longer
call scope than the current call.
The store function defines how to write a value v of a given value type
t into linear memory starting at offset ptr. Presenting the definition of
store piecewise, we start with the top-level case analysis:
def store(cx, v, t, ptr):
assert(ptr == align_to(ptr, alignment(t)))
assert(ptr + elem_size(t) <= len(cx.opts.memory))
match despecialize(t):
case Bool() : store_int(cx, int(bool(v)), ptr, 1)
case U8() : store_int(cx, v, ptr, 1)
case U16() : store_int(cx, v, ptr, 2)
case U32() : store_int(cx, v, ptr, 4)
case U64() : store_int(cx, v, ptr, 8)
case S8() : store_int(cx, v, ptr, 1, signed=True)
case S16() : store_int(cx, v, ptr, 2, signed=True)
case S32() : store_int(cx, v, ptr, 4, signed=True)
case S64() : store_int(cx, v, ptr, 8, signed=True)
case F32() : store_int(cx, encode_float_as_i32(v), ptr, 4)
case F64() : store_int(cx, encode_float_as_i64(v), ptr, 8)
case Char() : store_int(cx, char_to_i32(v), ptr, 4)
case String() : store_string(cx, v, ptr)
case List(t, l) : store_list(cx, v, ptr, t, l)
case Record(fields) : store_record(cx, v, ptr, fields)
case Variant(cases) : store_variant(cx, v, ptr, cases)
case Flags(labels) : store_flags(cx, v, ptr, labels)
case Own() : store_int(cx, lower_own(cx.opts, v, t), ptr, 4)
case Borrow() : store_int(cx, lower_borrow(cx.opts, v, t), ptr, 4)Integers are stored directly into memory. Because the input domain is exactly
the integers in range for the given type, no extra range checks are necessary;
the signed parameter is only present to ensure that the internal range checks
of int.to_bytes are satisfied.
def store_int(cx, v, ptr, nbytes, signed = False):
cx.opts.memory[ptr : ptr+nbytes] = int.to_bytes(v, nbytes, 'little', signed=signed)Floats are stored directly into memory, with the sign and payload bits of NaN values modified non-deterministically. This reflects the practical reality that different languages, protocols and CPUs have different effects on NaNs.
Although this non-determinism is expressed in the Python code below as generating a "random" NaN bit-pattern, native implementations do not need to use the same "random" algorithm, or even any random algorithm at all. Hosts may instead chose to canonicalize to an arbitrary fixed NaN value, or even to the original value of the NaN before lifting, allowing them to optimize away both the canonicalization of lifting and the randomization of lowering.
When a host implements the deterministic profile, NaNs are canonicalized to a particular NaN bit-pattern.
def maybe_scramble_nan32(f):
if math.isnan(f):
if DETERMINISTIC_PROFILE:
f = core_f32_reinterpret_i32(CANONICAL_FLOAT32_NAN)
else:
f = core_f32_reinterpret_i32(random_nan_bits(32, 8))
assert(math.isnan(f))
return f
def maybe_scramble_nan64(f):
if math.isnan(f):
if DETERMINISTIC_PROFILE:
f = core_f64_reinterpret_i64(CANONICAL_FLOAT64_NAN)
else:
f = core_f64_reinterpret_i64(random_nan_bits(64, 11))
assert(math.isnan(f))
return f
def random_nan_bits(total_bits, exponent_bits):
fraction_bits = total_bits - exponent_bits - 1
bits = random.getrandbits(total_bits)
bits |= ((1 << exponent_bits) - 1) << fraction_bits
bits |= 1 << random.randrange(fraction_bits - 1)
return bits
def encode_float_as_i32(f):
return core_i32_reinterpret_f32(maybe_scramble_nan32(f))
def encode_float_as_i64(f):
return core_i64_reinterpret_f64(maybe_scramble_nan64(f))
def core_i32_reinterpret_f32(f):
return struct.unpack('<I', struct.pack('<f', f))[0] # i32.reinterpret_f32
def core_i64_reinterpret_f64(f):
return struct.unpack('<Q', struct.pack('<d', f))[0] # i64.reinterpret_f64The integral value of a char (a Unicode Scalar Value) is a valid unsigned
i32 and thus no runtime conversion or checking is necessary:
def char_to_i32(c):
i = ord(c)
assert(0 <= i <= 0xD7FF or 0xD800 <= i <= 0x10FFFF)
return iStoring strings is complicated by the goal of attempting to optimize the
different transcoding cases. In particular, one challenge is choosing the
linear memory allocation size before examining the contents of the string.
The reason for this constraint is that, in some settings where single-pass
iterators are involved (host calls and post-MVP adapter functions), examining
the contents of a string more than once would require making an engine-internal
temporary copy of the whole string, which the component model specifically aims
not to do. To avoid multiple passes, the canonical ABI instead uses a realloc
approach to update the allocation size during the single copy. A blind
realloc approach would normally suffer from multiple reallocations per string
(e.g., using the standard doubling-growth strategy). However, as already shown
in load_string above, string values come with two useful hints: their
original encoding and byte length. From this hint data, store_string can do a
much better job minimizing the number of reallocations.
We start with a case analysis to enumerate all the meaningful encoding
combinations, subdividing the latin1+utf16 encoding into either latin1 or
utf16 based on the UTF16_BIT flag set by load_string:
def store_string(cx, v, ptr):
begin, tagged_code_units = store_string_into_range(cx, v)
store_int(cx, begin, ptr, 4)
store_int(cx, tagged_code_units, ptr + 4, 4)
def store_string_into_range(cx, v):
src, src_encoding, src_tagged_code_units = v
if src_encoding == 'latin1+utf16':
if bool(src_tagged_code_units & UTF16_TAG):
src_simple_encoding = 'utf16'
src_code_units = src_tagged_code_units ^ UTF16_TAG
else:
src_simple_encoding = 'latin1'
src_code_units = src_tagged_code_units
else:
src_simple_encoding = src_encoding
src_code_units = src_tagged_code_units
match cx.opts.string_encoding:
case 'utf8':
match src_simple_encoding:
case 'utf8' : return store_string_copy(cx, src, src_code_units, 1, 1, 'utf-8')
case 'utf16' : return store_utf16_to_utf8(cx, src, src_code_units)
case 'latin1' : return store_latin1_to_utf8(cx, src, src_code_units)
case 'utf16':
match src_simple_encoding:
case 'utf8' : return store_utf8_to_utf16(cx, src, src_code_units)
case 'utf16' : return store_string_copy(cx, src, src_code_units, 2, 2, 'utf-16-le')
case 'latin1' : return store_string_copy(cx, src, src_code_units, 2, 2, 'utf-16-le')
case 'latin1+utf16':
match src_encoding:
case 'utf8' : return store_string_to_latin1_or_utf16(cx, src, src_code_units)
case 'utf16' : return store_string_to_latin1_or_utf16(cx, src, src_code_units)
case 'latin1+utf16' :
match src_simple_encoding:
case 'latin1' : return store_string_copy(cx, src, src_code_units, 1, 2, 'latin-1')
case 'utf16' : return store_probably_utf16_to_latin1_or_utf16(cx, src, src_code_units)The simplest 4 cases above can compute the exact destination size and then copy with a simply loop (that possibly inflates Latin-1 to UTF-16 by injecting a 0 byte after every Latin-1 byte).
MAX_STRING_BYTE_LENGTH = (1 << 31) - 1
def store_string_copy(cx, src, src_code_units, dst_code_unit_size, dst_alignment, dst_encoding):
dst_byte_length = dst_code_unit_size * src_code_units
trap_if(dst_byte_length > MAX_STRING_BYTE_LENGTH)
ptr = cx.opts.realloc(0, 0, dst_alignment, dst_byte_length)
trap_if(ptr != align_to(ptr, dst_alignment))
trap_if(ptr + dst_byte_length > len(cx.opts.memory))
encoded = src.encode(dst_encoding)
assert(dst_byte_length == len(encoded))
cx.opts.memory[ptr : ptr+len(encoded)] = encoded
return (ptr, src_code_units)The choice of MAX_STRING_BYTE_LENGTH constant ensures that the high bit of a
string's byte length is never set, keeping it clear for UTF16_BIT.
The 2 cases of transcoding into UTF-8 share an algorithm that starts by optimistically assuming that each code unit of the source string fits in a single UTF-8 byte and then, failing that, reallocates to a worst-case size, finishes the copy, and then finishes with a shrinking reallocation.
def store_utf16_to_utf8(cx, src, src_code_units):
worst_case_size = src_code_units * 3
return store_string_to_utf8(cx, src, src_code_units, worst_case_size)
def store_latin1_to_utf8(cx, src, src_code_units):
worst_case_size = src_code_units * 2
return store_string_to_utf8(cx, src, src_code_units, worst_case_size)
def store_string_to_utf8(cx, src, src_code_units, worst_case_size):
assert(src_code_units <= MAX_STRING_BYTE_LENGTH)
ptr = cx.opts.realloc(0, 0, 1, src_code_units)
trap_if(ptr + src_code_units > len(cx.opts.memory))
for i,code_point in enumerate(src):
if ord(code_point) < 2**7:
cx.opts.memory[ptr + i] = ord(code_point)
else:
trap_if(worst_case_size > MAX_STRING_BYTE_LENGTH)
ptr = cx.opts.realloc(ptr, src_code_units, 1, worst_case_size)
trap_if(ptr + worst_case_size > len(cx.opts.memory))
encoded = src.encode('utf-8')
cx.opts.memory[ptr+i : ptr+len(encoded)] = encoded[i : ]
if worst_case_size > len(encoded):
ptr = cx.opts.realloc(ptr, worst_case_size, 1, len(encoded))
trap_if(ptr + len(encoded) > len(cx.opts.memory))
return (ptr, len(encoded))
return (ptr, src_code_units)Converting from UTF-8 to UTF-16 performs an initial worst-case size allocation (assuming each UTF-8 byte encodes a whole code point that inflates into a two-byte UTF-16 code unit) and then does a shrinking reallocation at the end if multiple UTF-8 bytes were collapsed into a single 2-byte UTF-16 code unit:
def store_utf8_to_utf16(cx, src, src_code_units):
worst_case_size = 2 * src_code_units
trap_if(worst_case_size > MAX_STRING_BYTE_LENGTH)
ptr = cx.opts.realloc(0, 0, 2, worst_case_size)
trap_if(ptr != align_to(ptr, 2))
trap_if(ptr + worst_case_size > len(cx.opts.memory))
encoded = src.encode('utf-16-le')
cx.opts.memory[ptr : ptr+len(encoded)] = encoded
if len(encoded) < worst_case_size:
ptr = cx.opts.realloc(ptr, worst_case_size, 2, len(encoded))
trap_if(ptr != align_to(ptr, 2))
trap_if(ptr + len(encoded) > len(cx.opts.memory))
code_units = int(len(encoded) / 2)
return (ptr, code_units)The next transcoding case handles latin1+utf16 encoding, where there general
goal is to fit the incoming string into Latin-1 if possible based on the code
points of the incoming string. The algorithm speculates that all code points
do fit into Latin-1 and then falls back to a worst-case allocation size when
a code point is found outside Latin-1. In this fallback case, the
previously-copied Latin-1 bytes are inflated in place, inserting a 0 byte
after every Latin-1 byte (iterating in reverse to avoid clobbering later
bytes):
def store_string_to_latin1_or_utf16(cx, src, src_code_units):
assert(src_code_units <= MAX_STRING_BYTE_LENGTH)
ptr = cx.opts.realloc(0, 0, 2, src_code_units)
trap_if(ptr != align_to(ptr, 2))
trap_if(ptr + src_code_units > len(cx.opts.memory))
dst_byte_length = 0
for usv in src:
if ord(usv) < (1 << 8):
cx.opts.memory[ptr + dst_byte_length] = ord(usv)
dst_byte_length += 1
else:
worst_case_size = 2 * src_code_units
trap_if(worst_case_size > MAX_STRING_BYTE_LENGTH)
ptr = cx.opts.realloc(ptr, src_code_units, 2, worst_case_size)
trap_if(ptr != align_to(ptr, 2))
trap_if(ptr + worst_case_size > len(cx.opts.memory))
for j in range(dst_byte_length-1, -1, -1):
cx.opts.memory[ptr + 2*j] = cx.opts.memory[ptr + j]
cx.opts.memory[ptr + 2*j + 1] = 0
encoded = src.encode('utf-16-le')
cx.opts.memory[ptr+2*dst_byte_length : ptr+len(encoded)] = encoded[2*dst_byte_length : ]
if worst_case_size > len(encoded):
ptr = cx.opts.realloc(ptr, worst_case_size, 2, len(encoded))
trap_if(ptr != align_to(ptr, 2))
trap_if(ptr + len(encoded) > len(cx.opts.memory))
tagged_code_units = int(len(encoded) / 2) | UTF16_TAG
return (ptr, tagged_code_units)
if dst_byte_length < src_code_units:
ptr = cx.opts.realloc(ptr, src_code_units, 2, dst_byte_length)
trap_if(ptr != align_to(ptr, 2))
trap_if(ptr + dst_byte_length > len(cx.opts.memory))
return (ptr, dst_byte_length)The final transcoding case takes advantage of the extra heuristic
information that the incoming UTF-16 bytes were intentionally chosen over
Latin-1 by the producer, indicating that they probably contain code points
outside Latin-1 and thus probably require inflation. Based on this
information, the transcoding algorithm pessimistically allocates storage for
UTF-16, deflating at the end if indeed no non-Latin-1 code points were
encountered. This Latin-1 deflation ensures that if a group of components
are all using latin1+utf16 and one component over-uses UTF-16, other
components can recover the Latin-1 compression. (The Latin-1 check can be
inexpensively fused with the UTF-16 validate+copy loop.)
def store_probably_utf16_to_latin1_or_utf16(cx, src, src_code_units):
src_byte_length = 2 * src_code_units
trap_if(src_byte_length > MAX_STRING_BYTE_LENGTH)
ptr = cx.opts.realloc(0, 0, 2, src_byte_length)
trap_if(ptr != align_to(ptr, 2))
trap_if(ptr + src_byte_length > len(cx.opts.memory))
encoded = src.encode('utf-16-le')
cx.opts.memory[ptr : ptr+len(encoded)] = encoded
if any(ord(c) >= (1 << 8) for c in src):
tagged_code_units = int(len(encoded) / 2) | UTF16_TAG
return (ptr, tagged_code_units)
latin1_size = int(len(encoded) / 2)
for i in range(latin1_size):
cx.opts.memory[ptr + i] = cx.opts.memory[ptr + 2*i]
ptr = cx.opts.realloc(ptr, src_byte_length, 1, latin1_size)
trap_if(ptr + latin1_size > len(cx.opts.memory))
return (ptr, latin1_size)Lists and records are stored by recursively storing their elements and are symmetric to the loading functions. Unlike strings, lists can simply allocate based on the up-front knowledge of length and static element size.
def store_list(cx, v, ptr, elem_type, maybe_length):
if maybe_length is not None:
assert(maybe_length == len(v))
store_list_into_valid_range(cx, v, ptr, elem_type)
return
begin, length = store_list_into_range(cx, v, elem_type)
store_int(cx, begin, ptr, 4)
store_int(cx, length, ptr + 4, 4)
def store_list_into_range(cx, v, elem_type):
byte_length = len(v) * elem_size(elem_type)
trap_if(byte_length >= (1 << 32))
ptr = cx.opts.realloc(0, 0, alignment(elem_type), byte_length)
trap_if(ptr != align_to(ptr, alignment(elem_type)))
trap_if(ptr + byte_length > len(cx.opts.memory))
store_list_into_valid_range(cx, v, ptr, elem_type)
return (ptr, len(v))
def store_list_into_valid_range(cx, v, ptr, elem_type):
for i,e in enumerate(v):
store(cx, e, elem_type, ptr + i * elem_size(elem_type))
def store_record(cx, v, ptr, fields):
for f in fields:
ptr = align_to(ptr, alignment(f.t))
store(cx, v[f.label], f.t, ptr)
ptr += elem_size(f.t)Variants are stored using the |-separated list of refines cases built
by case_label_with_refinements (above) to iteratively find a matching case (which
validation guarantees will succeed). While this code appears to do O(n) string
matching, a normal implementation can statically fuse store_variant with its
matching load_variant to ultimately build a dense array that maps producer's
case indices to the consumer's case indices.
def store_variant(cx, v, ptr, cases):
case_index, case_value = match_case(v, cases)
disc_size = elem_size(discriminant_type(cases))
store_int(cx, case_index, ptr, disc_size)
ptr += disc_size
ptr = align_to(ptr, max_case_alignment(cases))
c = cases[case_index]
if c.t is not None:
store(cx, case_value, c.t, ptr)
def match_case(v, cases):
assert(len(v.keys()) == 1)
key = list(v.keys())[0]
value = list(v.values())[0]
for label in key.split('|'):
case_index = find_case(label, cases)
if case_index != -1:
return (case_index, value)Flags are converted from a dictionary to a bit-vector by iterating through the case-labels of the variant in the order they were listed in the type definition and OR-ing all the bits together. Flag lifting/lowering can be statically fused into array/integer operations (with a simple byte copy when the case lists are the same) to avoid any string operations in a similar manner to variants.
def store_flags(cx, v, ptr, labels):
i = pack_flags_into_int(v, labels)
store_int(cx, i, ptr, elem_size_flags(labels))
def pack_flags_into_int(v, labels):
i = 0
shift = 0
for l in labels:
i |= (int(bool(v[l])) << shift)
shift += 1
return iFinally, own and borrow handles are lowered by initializing new handle
elements in the current component instance's handle table:
def lower_own(cx, rep, t):
h = HandleElem(rep, own=True)
return cx.inst.handles.add(t.rt, h)
def lower_borrow(cx, rep, t):
assert(isinstance(cx, Task))
if cx.inst is t.rt.impl:
return rep
h = HandleElem(rep, own=False, scope=cx)
cx.create_borrow()
return cx.inst.handles.add(t.rt, h)The special case in lower_borrow is an optimization, recognizing that, when
a borrowed handle is passed to the component that implemented the resource
type, the only thing the borrowed handle is good for is calling
resource.rep, so lowering might as well avoid the overhead of creating an
intermediate borrow handle.
With only the definitions above, the Canonical ABI would be forced to place all
parameters and results in linear memory. While this is necessary in the general
case, in many cases performance can be improved by passing small-enough values
in registers by using core function parameters and results. To support this
optimization, the Canonical ABI defines flatten_functype to map component
function types to core function types by attempting to decompose all the
non-dynamically-sized component value types into core value types.
For a variety of practical reasons, we need to limit the total number of flattened parameters and results, falling back to storing everything in linear memory. The number of flattened results is currently limited to 1 due to various parts of the toolchain (notably the C ABI) not yet being able to express multi-value returns. Hopefully this limitation is temporary and can be lifted before the Component Model is fully standardized.
When there are too many flat values, in general, a single i32 pointer can be
passed instead (pointing to a tuple in linear memory). When lowering into
linear memory, this requires the Canonical ABI to call realloc (in lower
below) to allocate space to put the tuple. As an optimization, when lowering
the return value of an imported function (via canon lower), the caller can
have already allocated space for the return value (e.g., efficiently on the
stack), passing in an i32 pointer as an parameter instead of returning an
i32 as a return value.
Given all this, the top-level definition of flatten_functype is:
MAX_FLAT_PARAMS = 16
MAX_FLAT_RESULTS = 1
def flatten_functype(opts, ft, context):
flat_params = flatten_types(ft.param_types())
flat_results = flatten_types(ft.result_types())
if opts.sync:
if len(flat_params) > MAX_FLAT_PARAMS:
flat_params = ['i32']
if len(flat_results) > MAX_FLAT_RESULTS:
match context:
case 'lift':
flat_results = ['i32']
case 'lower':
flat_params += ['i32']
flat_results = []
return CoreFuncType(flat_params, flat_results)
else:
match context:
case 'lift':
flat_params = []
flat_results = []
case 'lower':
if len(flat_params) > 1:
flat_params = ['i32']
if len(flat_results) > 0:
flat_params += ['i32']
flat_results = ['i32']
return CoreFuncType(flat_params, flat_results)
def flatten_types(ts):
return [ft for t in ts for ft in flatten_type(t)]As shown here, the core signatures async functions use a lower limit on the
maximum number of parameters (1) and results (0) passed as scalars before
falling back to passing through memory.
Presenting the definition of flatten_type piecewise, we start with the
top-level case analysis:
def flatten_type(t):
match despecialize(t):
case Bool() : return ['i32']
case U8() | U16() | U32() : return ['i32']
case S8() | S16() | S32() : return ['i32']
case S64() | U64() : return ['i64']
case F32() : return ['f32']
case F64() : return ['f64']
case Char() : return ['i32']
case String() : return ['i32', 'i32']
case List(t, l) : return flatten_list(t, l)
case Record(fields) : return flatten_record(fields)
case Variant(cases) : return flatten_variant(cases)
case Flags(labels) : return ['i32']
case Own(_) | Borrow(_) : return ['i32']List flattening of a fixed-length list uses the same flattening as a tuple
(via flatten_record below).
def flatten_list(elem_type, maybe_length):
if maybe_length is not None:
return flatten_type(elem_type) * maybe_length
return ['i32', 'i32']Record flattening simply flattens each field in sequence.
def flatten_record(fields):
flat = []
for f in fields:
flat += flatten_type(f.t)
return flatVariant flattening is more involved due to the fact that each case payload can
have a totally different flattening. Rather than giving up when there is a type
mismatch, the Canonical ABI relies on the fact that the 4 core value types can
be easily bit-cast between each other and defines a join operator to pick the
tightest approximation. What this means is that, regardless of the dynamic
case, all flattened variants are passed with the same static set of core types,
which may involve, e.g., reinterpreting an f32 as an i32 or zero-extending
an i32 into an i64.
def flatten_variant(cases):
flat = []
for c in cases:
if c.t is not None:
for i,ft in enumerate(flatten_type(c.t)):
if i < len(flat):
flat[i] = join(flat[i], ft)
else:
flat.append(ft)
return flatten_type(discriminant_type(cases)) + flat
def join(a, b):
if a == b: return a
if (a == 'i32' and b == 'f32') or (a == 'f32' and b == 'i32'): return 'i32'
return 'i64'Values are lifted by iterating over a list of parameter or result Core WebAssembly values:
@dataclass
class CoreValueIter:
values: list[int|float]
i = 0
def next(self, t):
v = self.values[self.i]
self.i += 1
match t:
case 'i32': assert(isinstance(v, int) and 0 <= v < 2**32)
case 'i64': assert(isinstance(v, int) and 0 <= v < 2**64)
case 'f32': assert(isinstance(v, (int,float)))
case 'f64': assert(isinstance(v, (int,float)))
case _ : assert(False)
return vThe match is only used for spec-level assertions; no runtime typecase is
required.
The lift_flat function defines how to convert a list of core values into a
single high-level value of type t. Presenting the definition of lift_flat
piecewise, we start with the top-level case analysis:
def lift_flat(cx, vi, t):
match despecialize(t):
case Bool() : return convert_int_to_bool(vi.next('i32'))
case U8() : return lift_flat_unsigned(vi, 32, 8)
case U16() : return lift_flat_unsigned(vi, 32, 16)
case U32() : return lift_flat_unsigned(vi, 32, 32)
case U64() : return lift_flat_unsigned(vi, 64, 64)
case S8() : return lift_flat_signed(vi, 32, 8)
case S16() : return lift_flat_signed(vi, 32, 16)
case S32() : return lift_flat_signed(vi, 32, 32)
case S64() : return lift_flat_signed(vi, 64, 64)
case F32() : return canonicalize_nan32(vi.next('f32'))
case F64() : return canonicalize_nan64(vi.next('f64'))
case Char() : return convert_i32_to_char(cx, vi.next('i32'))
case String() : return lift_flat_string(cx, vi)
case List(t, l) : return lift_flat_list(cx, vi, t, l)
case Record(fields) : return lift_flat_record(cx, vi, fields)
case Variant(cases) : return lift_flat_variant(cx, vi, cases)
case Flags(labels) : return lift_flat_flags(vi, labels)
case Own() : return lift_own(cx, vi.next('i32'), t)
case Borrow() : return lift_borrow(cx, vi.next('i32'), t)Integers are lifted from core i32 or i64 values using the signedness of the
target type to interpret the high-order bit. When the target type is narrower
than an i32, the Canonical ABI ignores the unused high bits (like load_int).
The conversion logic here assumes that i32 values are always represented as
unsigned Python ints and thus lifting to a signed type performs a manual 2s
complement conversion in the Python (which would be a no-op in hardware).
def lift_flat_unsigned(vi, core_width, t_width):
i = vi.next('i' + str(core_width))
assert(0 <= i < (1 << core_width))
return i % (1 << t_width)
def lift_flat_signed(vi, core_width, t_width):
i = vi.next('i' + str(core_width))
assert(0 <= i < (1 << core_width))
i %= (1 << t_width)
if i >= (1 << (t_width - 1)):
return i - (1 << t_width)
return iThe contents of strings and variable-length lists are stored in memory so
lifting these types is essentially the same as loading them from memory; the
only difference is that the pointer and length come from i32 values instead
of from linear memory. Fixed-length lists are lifted the same way as a
tuple (via lift_flat_record below).
def lift_flat_string(cx, vi):
ptr = vi.next('i32')
packed_length = vi.next('i32')
return load_string_from_range(cx, ptr, packed_length)
def lift_flat_list(cx, vi, elem_type, maybe_length):
if maybe_length is not None:
a = []
for i in range(maybe_length):
a.append(lift_flat(cx, vi, elem_type))
return a
ptr = vi.next('i32')
length = vi.next('i32')
return load_list_from_range(cx, ptr, length, elem_type)Records are lifted by recursively lifting their fields:
def lift_flat_record(cx, vi, fields):
record = {}
for f in fields:
record[f.label] = lift_flat(cx, vi, f.t)
return recordVariants are also lifted recursively. Lifting a variant must carefully follow
the definition of flatten_variant above, consuming the exact same core types
regardless of the dynamic case payload being lifted. Because of the join
performed by flatten_variant, we need a more-permissive value iterator that
reinterprets between the different types appropriately and also traps if the
high bits of an i64 are set for a 32-bit type:
def lift_flat_variant(cx, vi, cases):
flat_types = flatten_variant(cases)
assert(flat_types.pop(0) == 'i32')
case_index = vi.next('i32')
trap_if(case_index >= len(cases))
class CoerceValueIter:
def next(self, want):
have = flat_types.pop(0)
x = vi.next(have)
match (have, want):
case ('i32', 'f32') : return decode_i32_as_float(x)
case ('i64', 'i32') : return wrap_i64_to_i32(x)
case ('i64', 'f32') : return decode_i32_as_float(wrap_i64_to_i32(x))
case ('i64', 'f64') : return decode_i64_as_float(x)
case _ : assert(have == want); return x
c = cases[case_index]
if c.t is None:
v = None
else:
v = lift_flat(cx, CoerceValueIter(), c.t)
for have in flat_types:
_ = vi.next(have)
return { case_label_with_refinements(c, cases): v }
def wrap_i64_to_i32(i):
assert(0 <= i < (1 << 64))
return i % (1 << 32)Finally, flags are lifted by lifting to a record the same way as when loading flags from linear memory.
def lift_flat_flags(vi, labels):
assert(0 < len(labels) <= 32)
i = vi.next('i32')
return unpack_flags_from_int(i, labels)The lower_flat function defines how to convert a value v of a given type
t into zero or more core values. Presenting the definition of lower_flat
piecewise, we start with the top-level case analysis:
def lower_flat(cx, v, t):
match despecialize(t):
case Bool() : return [int(v)]
case U8() : return [v]
case U16() : return [v]
case U32() : return [v]
case U64() : return [v]
case S8() : return lower_flat_signed(v, 32)
case S16() : return lower_flat_signed(v, 32)
case S32() : return lower_flat_signed(v, 32)
case S64() : return lower_flat_signed(v, 64)
case F32() : return [maybe_scramble_nan32(v)]
case F64() : return [maybe_scramble_nan64(v)]
case Char() : return [char_to_i32(v)]
case String() : return lower_flat_string(cx, v)
case List(t, l) : return lower_flat_list(cx, v, t, l)
case Record(fields) : return lower_flat_record(cx, v, fields)
case Variant(cases) : return lower_flat_variant(cx, v, cases)
case Flags(labels) : return lower_flat_flags(v, labels)
case Own() : return [lower_own(cx, v, t)]
case Borrow() : return [lower_borrow(cx, v, t)]Since component-level values are assumed in-range and, as previously stated,
core i32 values are always internally represented as unsigned ints,
unsigned integer values need no extra conversion. Signed integer values are
converted to unsigned core i32s by 2s complement arithmetic (which again
would be a no-op in hardware):
def lower_flat_signed(i, core_bits):
if i < 0:
i += (1 << core_bits)
return [i]Since strings and variable-length lists are stored in linear memory, lifting
can reuse the previous definitions; only the resulting pointers are returned
differently (as i32 values instead of as a pair in linear memory).
Fixed-length lists are lowered the same way as tuples (via lower_flat_record
below).
def lower_flat_string(cx, v):
ptr, packed_length = store_string_into_range(cx, v)
return [ptr, packed_length]
def lower_flat_list(cx, v, elem_type, maybe_length):
if maybe_length is not None:
assert(maybe_length == len(v))
flat = []
for e in v:
flat += lower_flat(cx, e, elem_type)
return flat
(ptr, length) = store_list_into_range(cx, v, elem_type)
return [ptr, length]Records are lowered by recursively lowering their fields:
def lower_flat_record(cx, v, fields):
flat = []
for f in fields:
flat += lower_flat(cx, v[f.label], f.t)
return flatVariants are also lowered recursively. Symmetric to lift_flat_variant above,
lower_flat_variant must consume all flattened types of flatten_variant,
manually coercing the otherwise-incompatible type pairings allowed by join:
def lower_flat_variant(cx, v, cases):
case_index, case_value = match_case(v, cases)
flat_types = flatten_variant(cases)
assert(flat_types.pop(0) == 'i32')
c = cases[case_index]
if c.t is None:
payload = []
else:
payload = lower_flat(cx, case_value, c.t)
for i,(fv,have) in enumerate(zip(payload, flatten_type(c.t))):
want = flat_types.pop(0)
match (have, want):
case ('f32', 'i32') : payload[i] = encode_float_as_i32(fv)
case ('i32', 'i64') : payload[i] = fv
case ('f32', 'i64') : payload[i] = encode_float_as_i32(fv)
case ('f64', 'i64') : payload[i] = encode_float_as_i64(fv)
case _ : assert(have == want)
for _ in flat_types:
payload.append(0)
return [case_index] + payloadFinally, flags are lowered by packing the flags into an i32 bitvector.
def lower_flat_flags(v, labels):
assert(0 < len(labels) <= 32)
return [pack_flags_into_int(v, labels)]The lift_flat_values function defines how to lift a list of core
parameters or results (given by the CoreValueIter vi) into a tuple
of component-level values with types ts.
def lift_flat_values(cx, max_flat, vi, ts):
flat_types = flatten_types(ts)
if len(flat_types) > max_flat:
return lift_heap_values(cx, vi, ts)
else:
return [ lift_flat(cx, vi, t) for t in ts ]
def lift_heap_values(cx, vi, ts):
ptr = vi.next('i32')
tuple_type = Tuple(ts)
trap_if(ptr != align_to(ptr, alignment(tuple_type)))
trap_if(ptr + elem_size(tuple_type) > len(cx.opts.memory))
return list(load(cx, ptr, tuple_type).values())Symmetrically, the lower_flat_values function defines how to lower a
list of component-level values vs of types ts into a list of core
values. As already described for flatten_functype above,
lowering handles the greater-than-max_flat case by either allocating
storage with realloc or accepting a caller-allocated buffer as an
out-param:
def lower_flat_values(cx, max_flat, vs, ts, out_param = None):
cx.inst.may_leave = False
flat_types = flatten_types(ts)
if len(flat_types) > max_flat:
flat_vals = lower_heap_values(cx, vs, ts, out_param)
else:
flat_vals = []
for i in range(len(vs)):
flat_vals += lower_flat(cx, vs[i], ts[i])
cx.inst.may_leave = True
return flat_vals
def lower_heap_values(cx, vs, ts, out_param):
tuple_type = Tuple(ts)
tuple_value = {str(i): v for i,v in enumerate(vs)}
if out_param is None:
ptr = cx.opts.realloc(0, 0, alignment(tuple_type), elem_size(tuple_type))
flat_vals = [ptr]
else:
ptr = out_param.next('i32')
flat_vals = []
trap_if(ptr != align_to(ptr, alignment(tuple_type)))
trap_if(ptr + elem_size(tuple_type) > len(cx.opts.memory))
store(cx, tuple_value, tuple_type, ptr)
return flat_valsThe may_leave flag is guarded by canon_lower below to prevent a component
from calling out of the component while in the middle of lowering, ensuring
that the relative ordering of the side effects of lifting followed by lowering
cannot be observed and thus an implementation may reliably fuse lifting with
lowering when making a cross-component call to avoid the intermediate copy.
Using the above supporting definitions, we can describe the static and dynamic
semantics of component-level canon definitions. The following subsections
cover each of these canon cases.
For a canonical definition:
(canon lift $callee:<funcidx> $opts:<canonopt>* (func $f (type $ft)))validation specifies:
$calleemust have typeflatten_functype($opts, $ft, 'lift')$fis given type$ft- a
memoryis present if required by lifting and is a subtype of(memory 1) - a
reallocis present if required by lifting and has type(func (param i32 i32 i32 i32) (result i32)) - if a
post-returnis present, it has type(func (param flatten_functype({}, $ft, 'lift').results))
When instantiating component instance $inst:
- Define
$fto be the partially-bound closurecanon_lift($opts, $inst, $callee, $ft)
The resulting function $f takes 4 runtime arguments:
caller: the caller'sTaskor, if this lifted function is being called by the host,Noneon_block: a possibly-already-resolvedasyncio.Futurethat the callee must resolve (if not already) withOnBlockResult.BLOCKEDif the callee blockson_start: a nullary function that must be called to return the caller's arguments as a list of component-level valueson_return: a unary function that must be called afteron_start, passing the list of component-level return values
The indirection of on_start and on_return are used to model the
interleaving of reading arguments out of the caller's stack and memory and
writing results back into the caller's stack and memory, which will vary in
async calls.
If $f ends up being called by the host, the host is responsible for, in a
host-defined manner, conjuring up component-level values suitable for passing
into lower and, conversely, consuming the component values produced by
lift. For example, if the host is a native JS runtime, the JavaScript
embedding would specify how native JavaScript values are converted to and from
component values. Alternatively, if the host is a Unix CLI that invokes
component exports directly from the command line, the CLI could choose to
automatically parse argv into component-level values according to the
declared types of the export. In any case, canon lift specifies how these
variously-produced values are consumed as parameters (and produced as results)
by a single host-agnostic component.
Based on this, canon_lift is defined:
async def canon_lift(opts, inst, callee, ft, caller, on_block, on_start, on_return):
if opts.sync:
task = Task(opts, inst, caller, on_block)
await task.enter()
flat_args = lower_flat_values(task, MAX_FLAT_PARAMS, on_start(), ft.param_types())
flat_results = await call_and_trap_on_throw(callee, task, flat_args)
on_return(lift_flat_values(task, MAX_FLAT_RESULTS, CoreValueIter(flat_results), ft.result_types()))
if opts.post_return is not None:
[] = await call_and_trap_on_throw(opts.post_return, task, flat_results)
task.exit()
else:
task = AsyncTask(opts, inst, caller, on_block, ft, on_start, on_return)
await task.enter()
if not opts.callback:
[] = await call_and_trap_on_throw(callee, task, [])
else:
[packed_ctx] = await call_and_trap_on_throw(callee, task, [])
while packed_ctx != 0:
is_yield = bool(packed_ctx & 1)
ctx = packed_ctx & ~1
if is_yield:
await task.yield_()
event, payload = (EventCode.YIELDED, 0)
else:
event, payload = await task.wait()
[packed_ctx] = await call_and_trap_on_throw(opts.callback, task, [ctx, event, payload])
assert(opts.post_return is None)
task.exit()
async def call_and_trap_on_throw(callee, task, args):
try:
return await callee(task, args)
except CoreWebAssemblyException:
trap()The only fundamental difference between sync and async lifting is whether
parameters/results are automatically lowered/lifted (with canon_lift calling
on_start and on_return) or whether the callee explicitly triggers
on_start/on_return via task.start/task.return (defined below).
In a sync call, after the results have been copied from the callee's memory
into the caller's memory, the callee's post_return function is called to
allow the callee to reclaim any memory. An async call doesn't need a
post_return function, since the callee can keep running after calling
task.return.
Within the async case, there are two sub-cases depending on whether the
callback canonopt was set. When callback is present, waiting happens in
an "event loop" inside canon_lift which also allows yielding (i.e., allowing
other tasks to run without blocking) by setting the LSB of the returned i32.
Otherwise, waiting must happen by calling task.wait (defined below), which
potentially requires the runtime implementation to use a fiber (aka. stackful
coroutine) to switch to another task. Thus, callback is an optimization for
avoiding fiber creation for async languages that don't need it (e.g., JS,
Python, C# and Rust).
Uncaught Core WebAssembly exceptions result in a trap at component
boundaries. Thus, if a component wishes to signal an error, it must use some
sort of explicit type such as result (whose error case particular language
bindings may choose to map to and from exceptions).
For a canonical definition:
(canon lower $callee:<funcidx> $opts:<canonopt>* (core func $f))where $callee has type $ft, validation specifies:
$fis given typeflatten_functype($opts, $ft, 'lower')- a
memoryis present if required by lifting and is a subtype of(memory 1) - a
reallocis present if required by lifting and has type(func (param i32 i32 i32 i32) (result i32)) - there is no
post-returnin$opts
When instantiating component instance $inst:
- Define
$fto be the partially-bound closure:canon_lower($opts, $callee, $ft)
The resulting function $f takes 2 runtime arguments:
task: theTaskthat was created bycanon_liftwhen entering the current component instanceflat_args: the list of core values passed by the core function calller
Given this, canon_lower is defined:
async def canon_lower(opts, callee, ft, task, flat_args):
assert(current_task.locked())
trap_if(not task.inst.may_leave)
flat_args = CoreValueIter(flat_args)
flat_results = None
if opts.sync:
subtask = Subtask(opts, task)
def on_start():
return lift_flat_values(subtask, MAX_FLAT_PARAMS, flat_args, ft.param_types())
def on_return(results):
nonlocal flat_results
flat_results = lower_flat_values(subtask, MAX_FLAT_RESULTS, results, ft.result_types(), flat_args)
task.inst.interruptible.clear()
await callee(task, task.on_block, on_start, on_return)
task.inst.interruptible.set()
subtask.finish()
else:
subtask = AsyncSubtask(opts, task)
async def do_call(on_block):
def on_start():
subtask.start()
return lift_flat_values(subtask, 1, flat_args, ft.param_types())
def on_return(results):
subtask.return_()
lower_flat_values(subtask, 0, results, ft.result_types(), flat_args)
await callee(task, on_block, on_start, on_return)
subtask.finish()
match await task.handle_blocking(do_call):
case OnBlockResult.BLOCKED:
i = task.add_async_subtask(subtask)
flat_results = [pack_async_result(i, subtask.state)]
case OnBlockResult.COMPLETED:
flat_results = [0]
assert(current_task.locked())
return flat_resultsIn the synchronous case, clearing interruptible prevents new tasks from
starting (via Task.may_enter) and prevents suspended tasks from resuming
execution (via Task.suspend). Propagating the caller's on_block future
ensures that an async-lowered caller anywhere higher up on the stack still
gets control flow if callee blocks.
In the asynchronous case, Task.handle_blocking immediately returns control
flow from canon lift if callee blocks (in the host or in another
component in Task.suspend).
Whether or not the callee blocks, the on_start and on_return handlers
will be called before the callee completes. Note that, in an async-lowered
call, on_start and on_return can happen after canon_lower returns and
thus the caller must task.wait for progress events to know when the supplied
linear memory pointers can be reclaimed by the caller.
If an async-lowered call blocks, the AsyncSubtask is added to the component
instance's async_subtasks table, and the index and state are returned to the
caller packed into a single i32 as follows:
def pack_async_result(i, state):
assert(0 < i < 2**30)
assert(0 <= int(state) < 2**2)
return i | (int(state) << 30)The above definitions of sync/async canon_lift/canon_lower ensure that a
sync-or-async canon_lift may call a sync-or-async canon_lower, with all
combinations working. This is why the Task class, which is used for both sync
and async canon_lift calls, contains the code for handling async-lowered
subtasks. As mentioned above, conservative syntactic analysis of all canon
definitions in a component can statically rule out combinations so that, e.g.,
a DAG of all-sync components use a plain synchronous callstack and a DAG of all
async callback components use only an event loop without fibers. It's only
when async (without a callback) or various compositions of async and sync
components are used that fibers (or Asyncify) are required to implement the
above async rules.
Since any cross-component call necessarily transits through a statically-known
canon_lower+canon_lift call pair, an AOT compiler can fuse canon_lift and
canon_lower into a single, efficient trampoline. In the future this may allow
efficient compilation of permissive subtyping between components (including the
elimination of string operations on the labels of records and variants) as well
as post-MVP adapter functions.
For a canonical definition:
(canon resource.new $rt (core func $f))validation specifies:
$rtmust refer to locally-defined (not imported) resource type$fis given type(func (param $rt.rep) (result i32)), where$rt.repis currently fixed to bei32.
Calling $f invokes the following function, which adds an owning handle
containing the given resource representation in the current component
instance's handle table:
async def canon_resource_new(rt, task, rep):
trap_if(not task.inst.may_leave)
h = HandleElem(rep, own=True)
i = task.inst.handles.add(rt, h)
return [i]For a canonical definition:
(canon resource.drop $rt (core func $f))validation specifies:
$rtmust refer to resource type$fis given type(func (param i32))
Calling $f invokes the following function, which removes the handle from the
current component instance's handle table and, if the handle was owning, calls
the resource's destructor.
async def canon_resource_drop(rt, sync, task, i):
trap_if(not task.inst.may_leave)
inst = task.inst
h = inst.handles.remove(rt, i)
flat_results = [] if sync else [0]
if h.own:
assert(h.scope is None)
trap_if(h.lend_count != 0)
if inst is rt.impl:
if rt.dtor:
await rt.dtor(h.rep)
else:
if rt.dtor:
caller_opts = CanonicalOptions(sync = sync)
callee_opts = CanonicalOptions(sync = rt.dtor_sync, callback = rt.dtor_callback)
ft = FuncType([U32()],[])
callee = partial(canon_lift, callee_opts, rt.impl, rt.dtor, ft)
flat_results = await canon_lower(caller_opts, callee, ft, task, [h.rep, 0])
else:
task.trap_if_on_the_stack(rt.impl)
else:
h.scope.drop_borrow()
return flat_resultsIn general, the call to a resource's destructor is treated like a
cross-component call (as-if the destructor was exported by the component
defining the resource type). This means that cross-component destructor calls
follow the same concurrency rules as normal exports. However, since there are
valid reasons to call resource.drop in the same component instance that
defined the resource, which would otherwise trap at the reentrance guard of
Task.enter, an exception is made when the resource type's
implementation-instance is the same as the current instance (which is
statically known for any given canon resource.drop).
When a destructor isn't present, the rules still perform a reentrance check since this is the caller's responsibility and the presence or absence of a destructor is an encapsualted implementation detail of the resource type.
For a canonical definition:
(canon resource.rep $rt (core func $f))validation specifies:
$rtmust refer to a locally-defined (not imported) resource type$fis given type(func (param i32) (result $rt.rep)), where$rt.repis currently fixed to bei32.
Calling $f invokes the following function, which extracts the resource
representation from the handle.
async def canon_resource_rep(rt, task, i):
h = task.inst.handles.get(rt, i)
return [h.rep]Note that the "locally-defined" requirement above ensures that only the component instance defining a resource can access its representation.
For a canonical definition:
(canon task.backpressure (core func $f))validation specifies:
$fis given type[i32] -> []
Calling $f invokes the following function, which sets the backpressure
flag on the current ComponentInstance:
async def canon_task_backpressure(task, flat_args):
trap_if(task.opts.sync)
task.inst.backpressure = bool(flat_args[0])
return []The backpressure flag is read by Task.enter (defined above) to prevent new
tasks from entering the component instance and forcing the guest code to
consume resources.
For a canonical definition:
(canon task.start $ft (core func $f))validation specifies:
$fis given type$ft, which validation requires to be a (core) function type
Calling $f invokes the following function which extracts the arguments from the
caller and lowers them into the current instance:
async def canon_task_start(task, core_ft, flat_args):
assert(len(core_ft.params) == len(flat_args))
trap_if(not task.inst.may_leave)
trap_if(task.opts.sync)
trap_if(core_ft != flatten_functype(CanonicalOptions(), FuncType([], task.ft.params), 'lower'))
task.start()
args = task.on_start()
flat_results = lower_flat_values(task, MAX_FLAT_RESULTS, args, task.ft.param_types(), CoreValueIter(flat_args))
assert(len(core_ft.results) == len(flat_results))
return flat_resultsAn expected implementation of task.start would generate a core wasm function
for each lowering of an async-lifted export that performs the fused copy of
the arguments into the caller, storing the index of this function in the Task
structure and using call_indirect to perform the function-type-equality check
required here. The call to Task.start (defined above) ensures that canon task.start is called exactly once, before canon task.return, before an async
call finishes.
For a canonical definition:
(canon task.return $ft (core func $f))validation specifies:
$fis given type$ft, which validation requires to be a (core) function type
Calling $f invokes the following function which lifts the results from the
current instance and passes them to the caller:
async def canon_task_return(task, core_ft, flat_args):
assert(len(core_ft.params) == len(flat_args))
trap_if(not task.inst.may_leave)
trap_if(task.opts.sync)
trap_if(core_ft != flatten_functype(CanonicalOptions(), FuncType(task.ft.results, []), 'lower'))
task.return_()
results = lift_flat_values(task, MAX_FLAT_PARAMS, CoreValueIter(flat_args), task.ft.result_types())
task.on_return(results)
assert(len(core_ft.results) == 0)
return []An expected implementation of task.return would generate a core wasm function
for each lowering of an async-lifted export that performs the fused copy of
the results into the caller, storing the index of this function in the Task
structure and using call_indirect to perform the function-type-equality check
required here. The call to Task.return_ (defined above) ensures that canon task.return is called exactly once, after canon task.start, before an async
call finishes.
For a canonical definition:
(canon task.wait (core func $f))validation specifies:
$fis given type(func (param i32) (result i32))
Calling $f waits for progress to be made in a subtask of the current task,
returning the event (which is currently simply an AsyncCallState value)
and writing the subtask index as an outparam:
async def canon_task_wait(task, ptr):
trap_if(not task.inst.may_leave)
trap_if(task.opts.callback is not None)
event, payload = await task.wait()
store(task, payload, U32(), ptr)
return [event]The trap_if ensures that, when a component uses a callback all events flow
through the event loop at the base of the stack.
Note that task.wait will suspend the current Task, allowing other tasks
to run. Note also that task.wait can be called from a synchronously-lifted
export so that even synchronous code can make concurrent import calls. In these
synchronous cases, though, the automatic backpressure (applied by Task.enter)
will ensure there is only ever at most once synchronously-lifted task executing
in a component instance at a time.
For a canonical definition:
(canon task.poll (core func $f))validation specifies:
$fis given type(func (param i32) (result i32))
Calling $f does a non-blocking check for whether an event is already
available, returning whether or not there was such an event as a boolean and,
if there was an event, storing the i32 event+payload pair as an outparam.
async def canon_task_poll(task, ptr):
trap_if(not task.inst.may_leave)
ret = task.poll()
if ret is None:
return [0]
store(task, ret, Tuple([U32(), U32()]), ptr)
return [1]Note that there is no await of poll and thus no possible task switching.
For a canonical definition:
(canon task.yield (core func $f))validation specifies:
$fis given type(func)
Calling $f calls Task.yield_, trapping if called when there is a callback.
(When there is a callback, yielding is achieved by returning with the LSB set.)
async def canon_task_yield(task):
trap_if(not task.inst.may_leave)
trap_if(task.opts.callback is not None)
await task.yield_()
return []For a canonical definition:
(canon subtask.drop (core func $f))validation specifies:
$fis given type(func (param i32))
Calling $f removes the indicated subtask from the instance's table, trapping
if the subtask isn't done or isn't a subtask of the current task. The guard
on enqueued ensures that supertasks can only drop subtasks once they've been
officially notified of their completion (via task.wait or callback).
async def canon_subtask_drop(task, i):
trap_if(not task.inst.may_leave)
subtask = task.inst.async_subtasks.remove(i)
trap_if(subtask.enqueued)
trap_if(subtask.state != AsyncCallState.DONE)
subtask.task.num_async_subtasks -= 1
return []For a canonical definition:
(canon thread.spawn (type $ft) (core func $st))validation specifies:
$ftmust refer to asharedfunction type; initially, only the type(func shared (param $c i32))is allowed (see explanation below)$stis given type(func (param $f (ref null $ft)) (param $c i32) (result $e i32)).
Note: ideally, a thread could be spawned with arbitrary thread parameters. Currently, that would require additional work in the toolchain to support so, for simplicity, the current proposal simply fixes a single
i32parameter type. However,thread.spawncould be extended to allow arbitrary thread parameters in the future, once it's concretely beneficial to the toolchain. The inclusion of$ftensures backwards compatibility for when arbitrary parameters are allowed.
Calling $st checks that the reference $f is not null. Then, it spawns a
thread which:
- invokes
$fwith$c - executes
$funtil completion or trap in asharedcontext as described by the shared-everything threads proposal.
In pseudocode, $st looks like:
def canon_thread_spawn(f, c):
trap_if(f is None)
if DETERMINISTIC_PROFILE:
return [-1]
def thread_start():
try:
f(c)
except CoreWebAssemblyException:
trap()
if spawn(thread_start):
return [0]
else:
return [-1]For a canonical definition:
(canon thread.hw_concurrency (core func $f))validation specifies:
$fis given type(func shared (result i32)).
Calling $f returns the number of threads the underlying hardware can be
expected to execute concurrently. This value can be artificially limited by
engine configuration and is not allowed to change over the lifetime of a
component instance.
def canon_thread_hw_concurrency():
if DETERMINISTIC_PROFILE:
return [1]
else:
return [NUM_ALLOWED_THREADS]