/
githubmirror
/
julia
Обзор
Документация
Войти
/
githubmirror
/
julia
Код
Запросы
0
Пакеты
0
Релизы
0
Аналитика
Безопасность
master
base/channels.jl
794 строки
22 KB
Keno Fischer
cancellation: Hook up libuv to cancellation (#62557)
05 авг 2026, 07:46
Не верифицирован
05 авг 2026, 07:46
35b7e12
Код
Авторство
О чём код?
# This file is a part of Julia. License is MIT: https://julialang.org/license """ AbstractChannel{T} Representation of a channel passing objects of type `T`. """ abstract type AbstractChannel{T} end push!(c::AbstractChannel, v) = (put!(c, v); c) popfirst!(c::AbstractChannel) = take!(c) """ Channel{T=Any}(size::Int=0) Construct a `Channel` with an internal buffer that can hold a maximum of `size` objects of type `T`. [`put!`](@ref) calls on a full channel block until an object is removed with [`take!`](@ref). `Channel(0)` constructs an unbuffered channel. `put!` blocks until a matching `take!` is called. And vice-versa. Other constructors: * `Channel()`: default constructor, equivalent to `Channel{Any}(0)` * `Channel(Inf)`: equivalent to `Channel{Any}(typemax(Int))` * `Channel(sz)`: equivalent to `Channel{Any}(sz)` !!! compat "Julia 1.3" The default constructor `Channel()` and default `size=0` were added in Julia 1.3. """ mutable struct Channel{T} <: AbstractChannel{T} cond_take::Threads.Condition # waiting for data to become available cond_wait::Threads.Condition # waiting for data to become maybe available cond_put::Threads.Condition # waiting for a writeable slot @atomic state::Symbol excp::Union{Exception, Nothing} # exception to be thrown when state !== :open data::Vector{T} @atomic n_avail_items::Int # Available items for taking, can be read without lock sz_max::Int # maximum size of channel function Channel{T}(sz::Integer = 0) where T if sz < 0 throw(ArgumentError("Channel size must be either 0, a positive integer or Inf")) end lock = ReentrantLock() cond_put, cond_take = Threads.Condition(lock), Threads.Condition(lock) cond_wait = (sz == 0 ? Threads.Condition(lock) : cond_take) # wait is distinct from take iff unbuffered return new(cond_take, cond_wait, cond_put, :open, nothing, Vector{T}(), 0, sz) end end function Channel{T}(sz::Float64) where T sz = (sz == Inf ? typemax(Int) : convert(Int, sz)) return Channel{T}(sz) end Channel(sz=0) = Channel{Any}(sz) # special constructors """ Channel{T=Any}(func::Function, size=0; taskref=nothing, spawn=false, threadpool=nothing) Create a new task from `func`, [`bind`](@ref) it to a new channel of type `T` and size `size`, and schedule the task, all in a single call. The channel is automatically closed when the task terminates. `func` must accept the bound channel as its only argument. If you need a reference to the created task, pass a `Ref{Task}` object via the keyword argument `taskref`. If `spawn=true`, the `Task` created for `func` may be scheduled on another thread in parallel, equivalent to creating a task via [`Threads.@spawn`](@ref). If `spawn=true` and the `threadpool` argument is not set, it defaults to `:default`. If the `threadpool` argument is set (to `:default` or `:interactive`), this implies that `spawn=true` and the new Task is spawned to the specified threadpool. Return a `Channel`. # Examples ```jldoctest julia> chnl = Channel() do ch foreach(i -> put!(ch, i), 1:4) end; julia> typeof(chnl) Channel{Any} julia> for i in chnl @show i end; i = 1 i = 2 i = 3 i = 4 ``` Referencing the created task: ```jldoctest julia> taskref = Ref{Task}(); julia> chnl = Channel(taskref=taskref) do ch println(take!(ch)) end; julia> istaskdone(taskref[]) false julia> put!(chnl, "Hello"); Hello julia> istaskdone(taskref[]) true ``` !!! compat "Julia 1.3" The `spawn=` parameter was added in Julia 1.3. This constructor was added in Julia 1.3. In earlier versions of Julia, Channel used keyword arguments to set `size` and `T`, but those constructors are deprecated. !!! compat "Julia 1.9" The `threadpool=` argument was added in Julia 1.9. ```jldoctest julia> chnl = Channel{Char}(1, spawn=true) do ch for c in "hello world" put!(ch, c) end end; julia> String(collect(chnl)) "hello world" ``` """ function Channel{T}(func::Function, size=0; taskref=nothing, spawn=false, threadpool=nothing) where T chnl = Channel{T}(size) task = Task(() -> func(chnl)) if threadpool === nothing threadpool = :default else spawn = true end task.sticky = !spawn bind(chnl, task) if spawn Threads._spawn_set_thrpool(task, threadpool) schedule(task) # start it on (potentially) another thread else yield(task) # immediately start it, yielding the current thread end isa(taskref, Ref{Task}) && (taskref[] = task) return chnl end Channel(func::Function, args...; kwargs...) = Channel{Any}(func, args...; kwargs...) # This constructor is deprecated as of Julia v1.3, and should not be used. # (Note that this constructor also matches `Channel(::Function)` w/out any kwargs, which is # of course not deprecated.) # We use `nothing` default values to check which arguments were set in order to throw the # deprecation warning if users try to use `spawn=` with `ctype=` or `csize=`. function Channel(func::Function; ctype=nothing, csize=nothing, taskref=nothing, spawn=nothing, threadpool=nothing) # The spawn= keyword argument was added in Julia v1.3, and cannot be used with the # deprecated keyword arguments `ctype=` or `csize=`. if (ctype !== nothing || csize !== nothing) && (spawn !== nothing || threadpool !== nothing) throw(ArgumentError("Cannot set `spawn=` or `threadpool=` in the deprecated constructor `Channel(f; ctype=Any, csize=0)`. Please use `Channel{T=Any}(f, size=0; taskref=nothing, spawn=false, threadpool=nothing)` instead!")) end # Set the actual default values for the arguments. ctype === nothing && (ctype = Any) csize === nothing && (csize = 0) spawn === nothing && (spawn = false) return Channel{ctype}(func, csize; taskref=taskref, spawn=spawn, threadpool=threadpool) end closed_exception() = InvalidStateException("Channel is closed.", :closed) isbuffered(c::Channel) = c.sz_max==0 ? false : true function check_channel_state(c::Channel) if !isopen(c) # if the monotonic load succeed, now do an acquire fence (@atomic :acquire c.state) === :open && concurrency_violation() excp = c.excp excp !== nothing && throw(excp) throw(closed_exception()) end end """ close(c::Channel[, excp::Exception]) Close a channel. An exception (optionally given by `excp`), is thrown by: * [`put!`](@ref) on a closed channel. * [`take!`](@ref) and [`fetch`](@ref) on an empty, closed channel. """ close(c::Channel) = close(c, closed_exception()) # nospecialize on default arg seems to confuse makedocs function close(c::Channel, @nospecialize(excp::Exception)) lock(c) try c.excp = excp @atomic :release c.state = :closed notify_error(c.cond_take, excp) notify_error(c.cond_wait, excp) notify_error(c.cond_put, excp) finally unlock(c) end nothing end """ isopen(c::Channel) Determine whether a [`Channel`](@ref) is open for new [`put!`](@ref) operations. Notice that a `Channel` can be closed and still have buffered elements which can be consumed with [`take!`](@ref). # Examples ## Buffered channel with task ```jldoctest julia> c = Channel(ch -> put!(ch, 1), 1); julia> isopen(c) # The channel is closed to new `put!`s false julia> isready(c) # The channel is closed but still contains elements true julia> take!(c) 1 julia> isready(c) false ``` ## Unbuffered channel ```jldoctest julia> c = Channel{Int}(); julia> isopen(c) true julia> close(c) julia> isopen(c) false ``` """ function isopen(c::Channel) # Use acquire here to pair with release store in `close`, so that subsequent `isready` calls # are forced to see `isready == true` if they see `isopen == false`. This means users must # call `isopen` before `isready` if you are using the race-y APIs (or call `iterate`, which # does this right for you). return ((@atomic :acquire c.state) === :open) end """ empty!(c::Channel) Empty a Channel `c` by calling `empty!` on the internal buffer. Return the empty channel. """ function Base.empty!(c::Channel) @lock c begin ndrop = length(c.data) empty!(c.data) _increment_n_avail(c, -ndrop) notify(c.cond_put) end return c end """ bind(chnl::Channel, task::Task) Associate the lifetime of `chnl` with a task. `Channel` `chnl` is automatically closed when the task terminates. Any uncaught exception in the task is propagated to all waiters on `chnl`. The `chnl` object can be explicitly closed independent of task termination. Terminating tasks have no effect on already closed `Channel` objects. When a channel is bound to multiple tasks, the first task to terminate will close the channel. When multiple channels are bound to the same task, termination of the task will close all of the bound channels. # Examples ```jldoctest julia> c = Channel(0); julia> task = @async foreach(i->put!(c, i), 1:4); julia> bind(c,task); julia> for i in c @show i end; i = 1 i = 2 i = 3 i = 4 julia> isopen(c) false ``` ```jldoctest julia> c = Channel(0); julia> task = @async (put!(c, 1); error("foo")); julia> bind(c, task); julia> take!(c) 1 julia> put!(c, 1); ERROR: TaskFailedException Stacktrace: [...] nested task error: foo [...] ``` """ function bind(c::Channel, task::Task) # the close hook is cleanup: shield it from the constructing scope's # cancellation, so a bound channel is closed (and its blocked users # released) even when the scope that bound it is cancelled T = ScopedValues.with(CANCEL_TOKEN => nothing) do Task(() -> close_chnl_on_taskdone(task, c)) end T.sticky = false schedule_on_notify!(task, T) return c end """ channeled_tasks(n::Int, funcs...; ctypes=fill(Any,n), csizes=fill(0,n)) A convenience method to create `n` channels and bind them to tasks started from the provided functions in a single call. Each `func` must accept `n` arguments which are the created channels. Channel types and sizes may be specified via keyword arguments `ctypes` and `csizes` respectively. If unspecified, all channels are of type `Channel{Any}(0)`. Returns a tuple, `(Array{Channel}, Array{Task})`, of the created channels and tasks. """ function channeled_tasks(n::Int, funcs...; ctypes=fill(Any,n), csizes=fill(0,n)) @assert length(csizes) == n "length(csizes) != n" @assert length(ctypes) == n "length(ctypes) != n" chnls = map(i -> Channel{ctypes[i]}(csizes[i]), 1:n) tasks = Task[ Task(() -> f(chnls...)) for f in funcs ] # bind all tasks to all channels and schedule them foreach(t -> foreach(c -> bind(c, t), chnls), tasks) foreach(schedule, tasks) yield() # Allow scheduled tasks to run return (chnls, tasks) end function close_chnl_on_taskdone(t::Task, c::Channel) isopen(c) || return lock(c) try isopen(c) || return if istaskfailed(t) close(c, TaskFailedException(t)) return end close(c) finally unlock(c) end nothing end struct InvalidStateException <: Exception msg::String state::Symbol end showerror(io::IO, ex::InvalidStateException) = print(io, "InvalidStateException: ", ex.msg) """ put!(c::Channel, v) Append an item `v` to the channel `c`. Blocks if the channel is full. For unbuffered channels, blocks until a [`take!`](@ref) is performed by a different task. !!! compat "Julia 1.1" `v` now gets converted to the channel's type with [`convert`](@ref) as `put!` is called. """ function put!(c::Channel{T}, v; cancel::CancelTokenArg=DEFAULT_CANCEL) where T check_channel_state(c) v = convert(T, v) cancel = precheck_cancel_arg(cancel) return isbuffered(c) ? put_buffered(c, v, cancel) : put_unbuffered(c, v, cancel) end # Atomically update channel n_avail, *assuming* we hold the channel lock. function _increment_n_avail(c, inc) # We hold the channel lock so it's safe to non-atomically read and # increment c.n_avail_items newlen = c.n_avail_items + inc # Atomically store c.n_avail_items to prevent data races with other threads # reading this outside the lock. @atomic :monotonic c.n_avail_items = newlen end function put_buffered(c::Channel, v, cancel::CancelTokenArg=DEFAULT_CANCEL) lock(c; cancel) locked = true did_buffer = false try # Increment channel n_avail eagerly (before push!) to count data in the # buffer as well as offers from tasks which are blocked in wait(). _increment_n_avail(c, 1) if length(c.data) == c.sz_max tok = resolve_cancel_token(cancel) while length(c.data) == c.sz_max check_channel_state(c) locked = false wait(c.cond_put, tok) locked = true end end check_channel_state(c) push!(c.data, v) did_buffer = true # notify all, since some of the waiters may be on a "fetch" call. notify(c.cond_take, nothing, true, false) finally # Decrement the available items if this task had an exception before # pushing the item to the buffer (e.g., during `wait(c.cond_put)`). # The fixup needs the lock: reacquire (shielded) when a wait-throw # released this frame's level. locked || lock(c; cancel=nothing) did_buffer || _increment_n_avail(c, -1) unlock(c) end return v end function put_unbuffered(c::Channel, v, cancel::CancelTokenArg=DEFAULT_CANCEL) lock(c; cancel) locked = true taker = try _increment_n_avail(c, 1) tok = resolve_cancel_token(cancel) local taker while true while isempty(c.cond_take.waitq) check_channel_state(c) notify(c.cond_wait) locked = false wait(c.cond_put, tok) locked = true end check_channel_state(c) # unfair scheduled version of: notify(c.cond_take, v, false, false); yield() w = popfirst!(waitqueue(c.cond_take)) t = @atomic :monotonic w.task if t isa Task && claim_wait(t, w) taker = t break end # stale registration: this waiter's wake was already claimed # by an interrupter - drop it and look for the next taker end taker finally # the decrement needs the lock (see put_buffered's finally) locked || lock(c; cancel=nothing) _increment_n_avail(c, -1) unlock(c) end schedule(taker, v) yield() # immediately give taker a chance to run, but don't block the current task return v end """ fetch(c::Channel) Waits for and returns (without removing) the first available item from the `Channel`. Note: `fetch` is unsupported on an unbuffered (0-size) `Channel`. # Examples ## Buffered channel ```jldoctest julia> c = Channel(3) do ch foreach(i -> put!(ch, i), 1:3) end; julia> fetch(c) 1 julia> collect(c) # item is not removed 3-element Vector{Any}: 1 2 3 ``` """ function fetch(c::Channel; cancel::CancelTokenArg=DEFAULT_CANCEL) cancel = precheck_cancel_arg(cancel) return isbuffered(c) ? fetch_buffered(c, cancel) : fetch_unbuffered(c) end function fetch_buffered(c::Channel, cancel::CancelTokenArg=DEFAULT_CANCEL) lock(c; cancel) locked = true try if isempty(c.data) tok = resolve_cancel_token(cancel) while isempty(c.data) check_channel_state(c) locked = false wait(c.cond_take, tok) locked = true end end return c.data[1] finally locked && unlock(c) end end fetch_unbuffered(c::Channel) = throw(ErrorException("`fetch` is not supported on an unbuffered Channel.")) """ take!(c::Channel) Removes and returns a value from a [`Channel`](@ref) in order. Blocks until data is available. For unbuffered channels, blocks until a [`put!`](@ref) is performed by a different task. # Examples ## Buffered channel ```jldoctest julia> c = Channel(1); julia> put!(c, 1); julia> take!(c) 1 ``` ## Unbuffered channel ```jldoctest julia> c = Channel(0); julia> task = Task(() -> put!(c, 1)); julia> schedule(task); julia> take!(c) 1 ``` """ function take!(c::Channel; cancel::CancelTokenArg=DEFAULT_CANCEL) cancel = precheck_cancel_arg(cancel) return isbuffered(c) ? take_buffered(c, cancel) : take_unbuffered(c, cancel) end function take_buffered(c::Channel, cancel::CancelTokenArg=DEFAULT_CANCEL) lock(c; cancel) locked = true try if isempty(c.data) tok = resolve_cancel_token(cancel) while isempty(c.data) check_channel_state(c) locked = false wait(c.cond_take, tok) locked = true end end v = popfirst!(c.data) _increment_n_avail(c, -1) notify(c.cond_put, nothing, false, false) # notify only one, since only one slot has become available for a put!. return v finally locked && unlock(c) end end # 0-size channel function take_unbuffered(c::Channel{T}, cancel::CancelTokenArg=DEFAULT_CANCEL) where T lock(c; cancel) locked = true try check_channel_state(c) notify(c.cond_put, nothing, false, false) locked = false v = wait(c.cond_take, resolve_cancel_token(cancel))::T locked = true return v finally locked && unlock(c) end end """ isready(c::Channel) Determine whether a [`Channel`](@ref) has a value stored in it. Returns immediately, does not block. For unbuffered channels, return `true` if there are tasks waiting on a [`put!`](@ref). # Examples ## Buffered channel ```jldoctest julia> c = Channel(1); julia> isready(c) false julia> put!(c, 1); julia> isready(c) true ``` ## Unbuffered channel ```jldoctest julia> c = Channel(); julia> isready(c) # no tasks waiting to put! false julia> task = Task(() -> put!(c, 1)); julia> schedule(task); # schedule a put! task julia> isready(c) true ``` """ isready(c::Channel) = n_avail(c) > 0 isempty(c::Channel) = n_avail(c) == 0 function n_avail(c::Channel) # Lock-free equivalent to `length(c.data) + length(c.cond_put.waitq)` @atomic :monotonic c.n_avail_items end """ isfull(c::Channel) Determine if a [`Channel`](@ref) is full, in the sense that calling `put!(c, some_value)` would have blocked. Returns immediately, does not block. Note that it may frequently be the case that `put!` will not block after this returns `true`. Users must take precautions not to accidentally create live-lock bugs in their code by calling this method, as these are generally harder to debug than deadlocks. It is also possible that `put!` will block after this call returns `false`, if there are multiple producer tasks calling `put!` in parallel. # Examples ## Buffered channel ```jldoctest julia> c = Channel(1); # capacity = 1 julia> isfull(c) false julia> put!(c, 1); julia> isfull(c) true ``` ## Unbuffered channel ```jldoctest julia> c = Channel(); # capacity = 0 julia> isfull(c) # unbuffered channel is always full true ``` """ isfull(c::Channel) = n_avail(c) ≥ c.sz_max # the `cancel` forward makes the channel operations' preliminary lock # acquisitions honor their resolved token (or explicit shield); see # lock(::GenericCondition{ReentrantLock}) lock(c::Channel; cancel::CancelTokenArg=DEFAULT_CANCEL) = lock(c.cond_take; cancel) lock(f, c::Channel) = lock(f, c.cond_take) unlock(c::Channel) = unlock(c.cond_take) trylock(c::Channel) = trylock(c.cond_take) """ wait(c::Channel) Block until the `Channel` [`isready`](@ref). ```jldoctest julia> c = Channel(1); julia> isready(c) false julia> task = Task(() -> wait(c)); julia> schedule(task); julia> istaskdone(task) # task is blocked because channel is not ready false julia> put!(c, 1); julia> istaskdone(task) # task is now unblocked true ``` """ function wait(c::Channel; cancel::CancelTokenArg=DEFAULT_CANCEL) cancel = precheck_cancel_arg(cancel) isready(c) && return lock(c; cancel) locked = true try if !isready(c) tok = resolve_cancel_token(cancel) while !isready(c) check_channel_state(c) locked = false wait(c.cond_wait, tok) locked = true end end finally locked && unlock(c) end nothing end eltype(::Type{Channel{T}}) where {T} = T show(io::IO, c::Channel) = print(io, typeof(c), "(", c.sz_max, ")") function show(io::IO, ::MIME"text/plain", c::Channel) show(io, c) if !(get(io, :compact, false)::Bool) if !isopen(c) print(io, " (closed)") else n = n_avail(c) if n == 0 print(io, " (empty)") else s = n == 1 ? "" : "s" print(io, " (", n, " item$s available)") end end end end function iterate(c::Channel, state=nothing) if isopen(c) || isready(c) try return (take!(c), nothing) catch e if isa(e, InvalidStateException) && e.state === :closed return nothing else rethrow() end end else # If the channel was closed with an exception, it needs to be thrown if (@atomic :acquire c.state) === :closed e = c.excp if isa(e, InvalidStateException) && e.state === :closed nothing else throw(e) end end return nothing end end IteratorSize(::Type{<:Channel}) = SizeUnknown()