mirror of
https://github.com/mofeng-git/One-KVM.git
synced 2025-12-12 01:00:29 +08:00
162 lines
6.5 KiB
Python
162 lines
6.5 KiB
Python
# ========================================================================== #
|
||
# #
|
||
# KVMD - The main PiKVM daemon. #
|
||
# #
|
||
# Copyright (C) 2018-2024 Maxim Devaev <mdevaev@gmail.com> #
|
||
# #
|
||
# This program is free software: you can redistribute it and/or modify #
|
||
# it under the terms of the GNU General Public License as published by #
|
||
# the Free Software Foundation, either version 3 of the License, or #
|
||
# (at your option) any later version. #
|
||
# #
|
||
# This program is distributed in the hope that it will be useful, #
|
||
# but WITHOUT ANY WARRANTY; without even the implied warranty of #
|
||
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the #
|
||
# GNU General Public License for more details. #
|
||
# #
|
||
# You should have received a copy of the GNU General Public License #
|
||
# along with this program. If not, see <https://www.gnu.org/licenses/>. #
|
||
# #
|
||
# ========================================================================== #
|
||
|
||
|
||
import asyncio
|
||
import threading
|
||
import dataclasses
|
||
|
||
import gpiod
|
||
|
||
from . import aiotools
|
||
|
||
|
||
# =====
|
||
@dataclasses.dataclass(frozen=True)
|
||
class AioReaderPinParams:
|
||
inverted: bool
|
||
debounce: float
|
||
|
||
|
||
class AioReader: # pylint: disable=too-many-instance-attributes
|
||
def __init__(
|
||
self,
|
||
path: str,
|
||
consumer: str,
|
||
pins: dict[int, AioReaderPinParams],
|
||
notifier: aiotools.AioNotifier,
|
||
) -> None:
|
||
|
||
self.__path = path
|
||
self.__consumer = consumer
|
||
self.__pins = dict(pins)
|
||
self.__notifier = notifier
|
||
|
||
self.__values: (dict[int, _DebouncedValue] | None) = None
|
||
|
||
self.__thread = threading.Thread(target=self.__run, daemon=True)
|
||
self.__stop_event = threading.Event()
|
||
|
||
self.__loop: (asyncio.AbstractEventLoop | None) = None
|
||
|
||
def get(self, pin: int) -> bool:
|
||
value = (self.__values[pin].get() if self.__values is not None else False)
|
||
return (value ^ self.__pins[pin].inverted)
|
||
|
||
async def poll(self) -> None:
|
||
if not self.__pins:
|
||
await aiotools.wait_infinite()
|
||
else:
|
||
assert self.__loop is None
|
||
self.__loop = asyncio.get_running_loop()
|
||
self.__thread.start()
|
||
try:
|
||
await aiotools.run_async(self.__thread.join)
|
||
finally:
|
||
self.__stop_event.set()
|
||
await aiotools.run_async(self.__thread.join)
|
||
|
||
def __run(self) -> None:
|
||
assert self.__values is None
|
||
assert self.__loop
|
||
|
||
pins = sorted(self.__pins)
|
||
with gpiod.request_lines(
|
||
self.__path,
|
||
consumer=self.__consumer,
|
||
config={tuple(pins): gpiod.LineSettings(edge_detection=gpiod.line.Edge.BOTH)},
|
||
) as line_request:
|
||
|
||
line_request.wait_edge_events(0.1)
|
||
self.__values = {
|
||
pin: _DebouncedValue(
|
||
initial=bool(value.value),
|
||
debounce=self.__pins[pin].debounce,
|
||
notifier=self.__notifier,
|
||
loop=self.__loop,
|
||
)
|
||
for (pin, value) in zip(pins, line_request.get_values(pins))
|
||
}
|
||
self.__loop.call_soon_threadsafe(self.__notifier.notify)
|
||
|
||
while not self.__stop_event.is_set():
|
||
if line_request.wait_edge_events(1):
|
||
new: dict[int, bool] = {}
|
||
for event in line_request.read_edge_events():
|
||
(pin, value) = self.__parse_event(event)
|
||
new[pin] = value
|
||
for (pin, value) in new.items():
|
||
self.__values[pin].set(value)
|
||
else: # Timeout
|
||
# XXX: Лимит был актуален для 1.6. Надо проверить, поменялось ли это в 2.x.
|
||
# Размер буфера ядра - 16 эвентов на линии. При превышении этого числа,
|
||
# новые эвенты потеряются. Это не баг, это фича, как мне объяснили в LKML.
|
||
# Штош. Будем с этим жить и синхронизировать состояния при таймауте.
|
||
for (pin, value) in zip(pins, line_request.get_values(pins)):
|
||
self.__values[pin].set(bool(value.value)) # type: ignore
|
||
|
||
def __parse_event(self, event: gpiod.EdgeEvent) -> tuple[int, bool]:
|
||
if event.event_type == event.Type.RISING_EDGE:
|
||
return (event.line_offset, True)
|
||
elif event.event_type == event.Type.FALLING_EDGE:
|
||
return (event.line_offset, False)
|
||
raise RuntimeError(f"Invalid event {event} type: {event.type}")
|
||
|
||
|
||
class _DebouncedValue:
|
||
def __init__(
|
||
self,
|
||
initial: bool,
|
||
debounce: float,
|
||
notifier: aiotools.AioNotifier,
|
||
loop: asyncio.AbstractEventLoop,
|
||
) -> None:
|
||
|
||
self.__value = initial
|
||
self.__debounce = debounce
|
||
self.__notifier = notifier
|
||
self.__loop = loop
|
||
|
||
self.__queue: "asyncio.Queue[bool]" = asyncio.Queue() # type: ignore
|
||
self.__task = loop.create_task(self.__consumer_task_loop())
|
||
|
||
def set(self, value: bool) -> None:
|
||
if self.__loop.is_running():
|
||
self.__check_alive()
|
||
self.__loop.call_soon_threadsafe(self.__queue.put_nowait, value)
|
||
|
||
def get(self) -> bool:
|
||
return self.__value
|
||
|
||
def __check_alive(self) -> None:
|
||
if self.__task.done() and not self.__task.cancelled():
|
||
raise RuntimeError("Dead debounce consumer")
|
||
|
||
async def __consumer_task_loop(self) -> None:
|
||
while True:
|
||
value = await self.__queue.get()
|
||
while not self.__queue.empty():
|
||
value = await self.__queue.get()
|
||
if self.__value != value:
|
||
self.__value = value
|
||
self.__notifier.notify()
|
||
await asyncio.sleep(self.__debounce)
|