summaryrefslogtreecommitdiff
path: root/src/common/bounded_threadsafe_queue.h
diff options
context:
space:
mode:
Diffstat (limited to 'src/common/bounded_threadsafe_queue.h')
-rw-r--r--src/common/bounded_threadsafe_queue.h167
1 files changed, 167 insertions, 0 deletions
diff --git a/src/common/bounded_threadsafe_queue.h b/src/common/bounded_threadsafe_queue.h
new file mode 100644
index 000000000..7e465549b
--- /dev/null
+++ b/src/common/bounded_threadsafe_queue.h
@@ -0,0 +1,167 @@
+// SPDX-FileCopyrightText: Copyright (c) 2020 Erik Rigtorp <erik@rigtorp.se>
+// SPDX-License-Identifier: MIT
+
+#pragma once
+
+#include <atomic>
+#include <bit>
+#include <condition_variable>
+#include <memory>
+#include <mutex>
+#include <new>
+#include <stop_token>
+#include <type_traits>
+#include <utility>
+
+namespace Common {
+
+#if defined(__cpp_lib_hardware_interference_size)
+constexpr size_t hardware_interference_size = std::hardware_destructive_interference_size;
+#else
+constexpr size_t hardware_interference_size = 64;
+#endif
+
+#ifdef _MSC_VER
+#pragma warning(push)
+#pragma warning(disable : 4324)
+#endif
+
+template <typename T, size_t capacity = 0x400>
+class MPSCQueue {
+public:
+ explicit MPSCQueue() : allocator{std::allocator<Slot<T>>()} {
+ // Allocate one extra slot to prevent false sharing on the last slot
+ slots = allocator.allocate(capacity + 1);
+ // Allocators are not required to honor alignment for over-aligned types
+ // (see http://eel.is/c++draft/allocator.requirements#10) so we verify
+ // alignment here
+ if (reinterpret_cast<uintptr_t>(slots) % alignof(Slot<T>) != 0) {
+ allocator.deallocate(slots, capacity + 1);
+ throw std::bad_alloc();
+ }
+ for (size_t i = 0; i < capacity; ++i) {
+ std::construct_at(&slots[i]);
+ }
+ static_assert(std::has_single_bit(capacity), "capacity must be an integer power of 2");
+ static_assert(alignof(Slot<T>) == hardware_interference_size,
+ "Slot must be aligned to cache line boundary to prevent false sharing");
+ static_assert(sizeof(Slot<T>) % hardware_interference_size == 0,
+ "Slot size must be a multiple of cache line size to prevent "
+ "false sharing between adjacent slots");
+ static_assert(sizeof(MPSCQueue) % hardware_interference_size == 0,
+ "Queue size must be a multiple of cache line size to "
+ "prevent false sharing between adjacent queues");
+ }
+
+ ~MPSCQueue() noexcept {
+ for (size_t i = 0; i < capacity; ++i) {
+ std::destroy_at(&slots[i]);
+ }
+ allocator.deallocate(slots, capacity + 1);
+ }
+
+ // The queue must be both non-copyable and non-movable
+ MPSCQueue(const MPSCQueue&) = delete;
+ MPSCQueue& operator=(const MPSCQueue&) = delete;
+
+ MPSCQueue(MPSCQueue&&) = delete;
+ MPSCQueue& operator=(MPSCQueue&&) = delete;
+
+ void Push(const T& v) noexcept {
+ static_assert(std::is_nothrow_copy_constructible_v<T>,
+ "T must be nothrow copy constructible");
+ emplace(v);
+ }
+
+ template <typename P, typename = std::enable_if_t<std::is_nothrow_constructible_v<T, P&&>>>
+ void Push(P&& v) noexcept {
+ emplace(std::forward<P>(v));
+ }
+
+ void Pop(T& v, std::stop_token stop) noexcept {
+ auto const tail = tail_.fetch_add(1);
+ auto& slot = slots[idx(tail)];
+ if (!slot.turn.test()) {
+ std::unique_lock lock{cv_mutex};
+ cv.wait(lock, stop, [&slot] { return slot.turn.test(); });
+ }
+ v = slot.move();
+ slot.destroy();
+ slot.turn.clear();
+ slot.turn.notify_one();
+ }
+
+private:
+ template <typename U = T>
+ struct Slot {
+ ~Slot() noexcept {
+ if (turn.test()) {
+ destroy();
+ }
+ }
+
+ template <typename... Args>
+ void construct(Args&&... args) noexcept {
+ static_assert(std::is_nothrow_constructible_v<U, Args&&...>,
+ "T must be nothrow constructible with Args&&...");
+ std::construct_at(reinterpret_cast<U*>(&storage), std::forward<Args>(args)...);
+ }
+
+ void destroy() noexcept {
+ static_assert(std::is_nothrow_destructible_v<U>, "T must be nothrow destructible");
+ std::destroy_at(reinterpret_cast<U*>(&storage));
+ }
+
+ U&& move() noexcept {
+ return reinterpret_cast<U&&>(storage);
+ }
+
+ // Align to avoid false sharing between adjacent slots
+ alignas(hardware_interference_size) std::atomic_flag turn{};
+ struct aligned_store {
+ struct type {
+ alignas(U) unsigned char data[sizeof(U)];
+ };
+ };
+ typename aligned_store::type storage;
+ };
+
+ template <typename... Args>
+ void emplace(Args&&... args) noexcept {
+ static_assert(std::is_nothrow_constructible_v<T, Args&&...>,
+ "T must be nothrow constructible with Args&&...");
+ auto const head = head_.fetch_add(1);
+ auto& slot = slots[idx(head)];
+ slot.turn.wait(true);
+ slot.construct(std::forward<Args>(args)...);
+ slot.turn.test_and_set();
+ cv.notify_one();
+ }
+
+ constexpr size_t idx(size_t i) const noexcept {
+ return i & mask;
+ }
+
+ static constexpr size_t mask = capacity - 1;
+
+ // Align to avoid false sharing between head_ and tail_
+ alignas(hardware_interference_size) std::atomic<size_t> head_{0};
+ alignas(hardware_interference_size) std::atomic<size_t> tail_{0};
+
+ std::mutex cv_mutex;
+ std::condition_variable_any cv;
+
+ Slot<T>* slots;
+ [[no_unique_address]] std::allocator<Slot<T>> allocator;
+
+ static_assert(std::is_nothrow_copy_assignable_v<T> || std::is_nothrow_move_assignable_v<T>,
+ "T must be nothrow copy or move assignable");
+
+ static_assert(std::is_nothrow_destructible_v<T>, "T must be nothrow destructible");
+};
+
+#ifdef _MSC_VER
+#pragma warning(pop)
+#endif
+
+} // namespace Common