Files
punktfunk/crates/pyrowave-sys/vendor/pyrowave/Granite/util/message_queue.hpp
T
enricobuehler 4c3b11445c feat(host): vendor PyroWave + minimal Granite subset as crates/pyrowave-sys
Phase 0 of design/pyrowave-codec-plan.md — the opt-in wired-LAN ultra-low-
latency codec. Vendored at upstream 509e4f88 (API 0.4.0, Granite 44362775,
volk + vulkan-headers pins in PUNKTFUNK-VENDOR.txt), pruned to the 6.6 MB
the standalone no-renderer build needs; scripts/vendor-pyrowave.sh
reproduces the tree (a pin bump is protocol-affecting, plan §4.2).

build.rs drives the wrapper CMakeLists (static archives incl. a static
C-API lib upstream only ships shared) + bindgen over pyrowave.h; Linux and
Windows only, empty stub elsewhere (Apple gets a native Metal port, §4.7).
Offline-safe by construction: no network, no system lib, vendored Vulkan
headers — same model as the opus dep (flatpak builder has no network).

Phase-0 validation on .21 (RTX 5070 Ti, driver 610.43.03):
- upstream pyrowave-c-test + interop test (incl. dmabuf/DRM-modifier
  Vulkan<->Vulkan) pass, from the pristine AND the pruned tree
- GPU kernel times at ~1.6 bpp noise: encode/decode 0.090/0.042 ms @800p,
  0.146/0.067 @1080p, 0.226/0.103 @1440p, 0.477/0.201 @4K — order of
  magnitude under NVENC's 1-2 ms retrieve, CBR lands within ~100 B of
  target
- cargo test -p pyrowave-sys green (static link + API-version pin check)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-15 00:35:10 +02:00

241 lines
6.4 KiB
C++

/* Copyright (c) 2017-2026 Hans-Kristian Arntzen
*
* Permission is hereby granted, free of charge, to any person obtaining
* a copy of this software and associated documentation files (the
* "Software"), to deal in the Software without restriction, including
* without limitation the rights to use, copy, modify, merge, publish,
* distribute, sublicense, and/or sell copies of the Software, and to
* permit persons to whom the Software is furnished to do so, subject to
* the following conditions:
*
* The above copyright notice and this permission notice shall be
* included in all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
* IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
* CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
* TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
* SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
*/
#pragma once
#include <atomic>
#include <vector>
#include <utility>
#include <stddef.h>
#include <assert.h>
#include <memory>
#include <mutex>
#include <condition_variable>
#include "global_managers_interface.hpp"
#include "logging.hpp"
namespace Util
{
// There can only be one concurrent reader, and one concurrent writer.
// This is useful for lock-less messaging between two threads, e.g. a worker thread and master thread.
template <typename T>
class LockFreeRingBuffer
{
public:
LockFreeRingBuffer()
{
reset(1);
assert(read_count.is_lock_free());
assert(write_count.is_lock_free());
}
void reset(size_t count)
{
//assert((count & (count - 1)) == 0);
ring.resize(count);
read_count.store(0);
write_count.store(0);
}
size_t read_avail() const noexcept
{
return write_count.load(std::memory_order_acquire) -
read_count.load(std::memory_order_relaxed);
}
size_t write_avail() const noexcept
{
return ring.size() -
(write_count.load(std::memory_order_relaxed) -
read_count.load(std::memory_order_acquire));
}
bool write_and_move(T *values, size_t count) noexcept
{
size_t current_written = read_count.load(std::memory_order_relaxed);
size_t current_read = write_count.load(std::memory_order_acquire);
if (count > ring.size() - (current_written - current_read))
return false;
size_t can_write_first = std::min<size_t>(ring.size() - write_offset, count);
size_t can_write_second = count - can_write_first;
std::move(values, values + can_write_first, ring.data() + write_offset);
write_offset += can_write_first;
values += can_write_first;
if (write_offset >= ring.size())
write_offset -= ring.size();
std::move(values, values + can_write_second, ring.data());
write_offset += can_write_second;
// Need release ordering so the reads here can't be ordered after the store.
write_count.store(write_count.load(std::memory_order_relaxed) + count, std::memory_order_release);
return true;
}
bool read_and_move(T *values, size_t count) noexcept
{
size_t current_read = read_count.load(std::memory_order_relaxed);
size_t current_written = write_count.load(std::memory_order_acquire);
if (count > current_written - current_read)
return false;
size_t can_read_first = std::min<size_t>(ring.size() - read_offset, count);
size_t can_read_second = count - can_read_first;
std::move(ring.data() + read_offset, ring.data() + read_offset + can_read_first, values);
read_offset += can_read_first;
values += can_read_first;
if (read_offset >= ring.size())
read_offset -= ring.size();
std::move(ring.data(), ring.data() + can_read_second, values);
read_offset += can_read_second;
// Need release ordering so the reads here can't be ordered after the store.
read_count.store(read_count.load(std::memory_order_relaxed) + count, std::memory_order_release);
return true;
}
bool write_and_move(T value) noexcept
{
return write_and_move(&value, 1);
}
bool read_and_move(T &value) noexcept
{
return read_and_move(&value, 1);
}
private:
std::atomic_size_t read_count;
std::atomic_size_t write_count;
size_t read_offset = 0;
size_t write_offset = 0;
std::vector<T> ring;
};
struct MessageQueuePayloadDeleter
{
void operator()(void *ptr);
};
class MessageQueuePayload
{
public:
template <typename T>
T &as()
{
assert(handle);
return *static_cast<T *>(handle);
}
// The handle might be slightly different from payload if we allocated
// with multiple-inheritance and the base class we care about is not the first one in the inheritance list.
template <typename T>
void set_payload_handle(T *t)
{
handle = t;
}
explicit operator bool() const
{
return bool(payload);
}
size_t get_size() const
{
return payload_size;
}
void set_size(size_t size)
{
assert(size <= payload_capacity);
payload_size = size;
}
void set_payload_data(void *ptr, size_t size)
{
payload.reset(ptr);
payload_capacity = size;
}
void *get_payload_data() const
{
return payload.get();
}
size_t get_capacity() const
{
return payload_capacity;
}
private:
std::unique_ptr<void, MessageQueuePayloadDeleter> payload;
void *handle = nullptr;
size_t payload_size = 0;
size_t payload_capacity = 0;
};
class LockFreeMessageQueue
{
public:
LockFreeMessageQueue();
MessageQueuePayload allocate_write_payload(size_t size) noexcept;
bool push_written_payload(MessageQueuePayload payload) noexcept;
size_t available_read_messages() const noexcept;
MessageQueuePayload read_message() noexcept;
void recycle_payload(MessageQueuePayload payload) noexcept;
private:
LockFreeRingBuffer<MessageQueuePayload> read_ring;
LockFreeRingBuffer<MessageQueuePayload> write_ring[8];
size_t payload_capacity[8] = {};
};
class MessageQueue final : private LockFreeMessageQueue, public MessageQueueInterface
{
public:
MessageQueue();
void cork();
void uncork();
bool is_uncorked() const;
MessageQueuePayload allocate_write_payload(size_t size) noexcept;
bool push_written_payload(MessageQueuePayload payload) noexcept;
size_t available_read_messages() const noexcept;
MessageQueuePayload read_message() noexcept;
void recycle_payload(MessageQueuePayload payload) noexcept;
private:
mutable std::mutex lock;
mutable std::condition_variable cond;
std::atomic_bool corked;
bool log(const char *tag, const char *fmt, va_list va) override;
};
}