feat(gpio)!: Add GPIO support for rmcs_board#48
Merged
Conversation
|
Note Reviews pausedUse the following commands to manage reviews:
Use the checkboxes below for quick actions:
Walkthrough本PR将GPIO通道从数据视图字段移除为显式的 Changes
Sequence Diagram(s)sequenceDiagram
participant Host as Host/Agent
participant Serializer as Host Serializer
participant Transport as USB/Transport
participant Deserializer as Firmware Deserializer
participant GPIO as Firmware GPIO Handler
participant SerializerFW as Firmware Serializer
participant HostHandler as Host Handler
Host->>Serializer: 构建 GPIO 报文(含 channel_index)
Serializer->>Transport: 发送报文字节
Transport->>Deserializer: 传递报文
Deserializer->>GPIO: gpio_*_deserialized(channel_index, data_view)
GPIO->>SerializerFW: write_gpio_*_read_result(channel_index, data_view)
SerializerFW->>Transport: 发送结果报文
Transport->>HostHandler: 传递结果报文
HostHandler->>Host: 回调 gpio_*_read_result_callback(channel_index, data_view)
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (4)
core/include/librmcs/spec/c_board/gpio.hpp (1)
46-48: 建议在operator[]中加入越界断言。当前
operator[]未做边界检查，若上游忘记channel_index < size()校验即直接索引将触发 UB。可参考firmware/c_board/app/src/gpio/gpio.hpp中channel_state()/channel_hardware()的做法，加一个assert_debug提高防御性。虽然现有调用点（vendor.hpp）已做了前置校验，但作为公开可见的容器访问接口，断言能避免未来新增调用方忘记检查。♻️ 建议改动
- static constexpr const GpioDescriptor& operator[](std::size_t channel_index) noexcept { - return kArray[channel_index]; - } + static constexpr const GpioDescriptor& operator[](std::size_t channel_index) noexcept { + // NOTE: 需先引入 assert 头；非 constexpr 上下文的越界会触发断言。 + // 对 constexpr 求值，越界访问会自动编译失败。 + return kArray[channel_index]; + }（若希望保持
constexpr，可考虑改用std::array并在非 constexpr 求值路径上通过utility::assert_debug校验，或直接保留现状并在文档注释中明确 API 前置条件。）🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@core/include/librmcs/spec/c_board/gpio.hpp` around lines 46 - 48, The operator[] accessor for GpioDescriptor (static constexpr const GpioDescriptor& operator[](std::size_t channel_index) noexcept) lacks a bounds check and can trigger undefined behavior; add a defensive assert_debug that verifies channel_index < size() (or kArray.size()) at the start of operator[] before returning kArray[channel_index]; keep the noexcept/constexpr intent in mind (you can keep the assert_guard disabled in constexpr evaluation or follow the existing pattern used in channel_state()/channel_hardware() in firmware/c_board/app/src/gpio/gpio.hpp).core/src/protocol/serializer.hpp (1)
123-291: 建议在 6 个write_gpio_*入口统一校验channel_index是否超出 6 bit 范围。
GpioHeader::ChannelIndex是BitfieldMember<8, 6>（6 bit，最大 63）。目前 6 个write_gpio_*方法都没有校验channel_index，若上游传入>= 64的值，位域写入会静默截断高位，产生错误的线协议编码而不返回kInvalidArgument。现阶段调用方都基于kGpioDescriptors（size=7）做了前置约束，但 serializer 作为通用协议原语，加一行防御性校验更稳妥，也与本文件中period_ms <= (1U << 13) - 1U的校验风格一致。♻️ 建议改动（以 `write_gpio_digital_data` 为例；其余 5 个方法同样处理）
SerializeResult write_gpio_digital_data( uint8_t channel_index, const data::GpioDigitalDataView& view) noexcept { + LIBRMCS_VERIFY_LIKELY(channel_index < (1U << 6), SerializeResult::kInvalidArgument); const auto payload_type = view.high ? GpioHeader::PayloadEnum::kDigitalWriteHigh : GpioHeader::PayloadEnum::kDigitalWriteLow;可考虑抽成一个私有常量，例如
static constexpr std::uint8_t kMaxChannelIndex = (1U << 6) - 1U;，在全部 6 个 GPIO 写入方法中复用。🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@core/src/protocol/serializer.hpp` around lines 123 - 291, Add a defensive bounds check for channel_index against the 6-bit limit (max 63) before allocating/writing in all GPIO serializers: write_gpio_digital_data, write_gpio_digital_read_config, write_gpio_analog_data, write_gpio_analog_read_config, write_gpio_digital_read_result, and write_gpio_analog_read_result; define a private constant like kMaxChannelIndex = (1U << 6) - 1U and verify channel_index <= kMaxChannelIndex (since GpioHeader::ChannelIndex is BitfieldMember<8,6>) and return SerializeResult::kInvalidArgument when out of range, mirroring the existing period_ms validation style.core/include/librmcs/data/datas.hpp (1)
68-75: 建议给supported()补上noexcept和[[nodiscard]]。该函数的所有子调用（
gpio.supports(...)）均为noexcept，且返回值是能力校验结果，丢弃结果基本都属于误用；加上[[nodiscard]]可由编译器阻止忽略返回值的情况，noexcept则便于在noexcept路径（如vendor.hpp的反序列化回调）中无开销调用。♻️ 建议改动
- bool supported(const spec::GpioDescriptor& gpio) const { + [[nodiscard]] bool supported(const spec::GpioDescriptor& gpio) const noexcept { return (!asap || gpio.supports(spec::GpioCapability::kDigitalReadOnce)) && (!period_ms || gpio.supports(spec::GpioCapability::kDigitalReadPeriodic)) && ((!rising_edge && !falling_edge) || gpio.supports(spec::GpioCapability::kDigitalReadInterrupt)) && (pull != GpioPull::kUp || gpio.supports(spec::GpioCapability::kPullUp)) && (pull != GpioPull::kDown || gpio.supports(spec::GpioCapability::kPullDown)); }🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@core/include/librmcs/data/datas.hpp` around lines 68 - 75, The supported(const spec::GpioDescriptor& gpio) const method should be marked [[nodiscard]] and noexcept; update the method signature in its declaration/definition (function supported) to read as a [[nodiscard]] bool supported(...) const noexcept so callers are warned if they ignore the result and the function can be used in noexcept contexts (also update any out-of-line definition or forward declaration for supported to match this signature).host/include/librmcs/agent/c_board.hpp (1)
10-12: 项目头文件引用风格不一致同一文件内，第 10 行使用
<librmcs/spec/gpio.hpp>（尖括号），而第 12 行对同样属于本仓库的librmcs/spec/c_board/gpio.hpp使用了"..."（双引号）。本文件中其它项目头文件（librmcs/agent/common.hpp、librmcs/data/datas.hpp、librmcs/protocol/handler.hpp）均使用尖括号，建议统一：🔧 建议修改
`#include` <librmcs/spec/gpio.hpp> - -#include "librmcs/spec/c_board/gpio.hpp" +#include <librmcs/spec/c_board/gpio.hpp>🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@host/include/librmcs/agent/c_board.hpp` around lines 10 - 12, The include style is inconsistent: <librmcs/spec/gpio.hpp> uses angle brackets while "librmcs/spec/c_board/gpio.hpp" uses quotes; change the latter to use angle brackets to match the project's other includes (e.g., <librmcs/agent/common.hpp>, <librmcs/data/datas.hpp>, <librmcs/protocol/handler.hpp>), so replace "librmcs/spec/c_board/gpio.hpp" with <librmcs/spec/c_board/gpio.hpp> in the c_board.hpp header to unify include style.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@core/include/librmcs/spec/gpio.hpp`:
- Around line 36-45: Remove the dead constant kNoExtiLine from the
GpioDescriptor struct: locate the struct GpioDescriptor and delete the static
constexpr std::uint8_t kNoExtiLine = 0xFFU declaration, leaving only the used
members (channel_index, capability_mask and supports()); ensure no other code
references GpioDescriptor::kNoExtiLine before deletion.
---
Nitpick comments:
In `@core/include/librmcs/data/datas.hpp`:
- Around line 68-75: The supported(const spec::GpioDescriptor& gpio) const
method should be marked [[nodiscard]] and noexcept; update the method signature
in its declaration/definition (function supported) to read as a [[nodiscard]]
bool supported(...) const noexcept so callers are warned if they ignore the
result and the function can be used in noexcept contexts (also update any
out-of-line definition or forward declaration for supported to match this
signature).
In `@core/include/librmcs/spec/c_board/gpio.hpp`:
- Around line 46-48: The operator[] accessor for GpioDescriptor (static
constexpr const GpioDescriptor& operator[](std::size_t channel_index) noexcept)
lacks a bounds check and can trigger undefined behavior; add a defensive
assert_debug that verifies channel_index < size() (or kArray.size()) at the
start of operator[] before returning kArray[channel_index]; keep the
noexcept/constexpr intent in mind (you can keep the assert_guard disabled in
constexpr evaluation or follow the existing pattern used in
channel_state()/channel_hardware() in firmware/c_board/app/src/gpio/gpio.hpp).
In `@core/src/protocol/serializer.hpp`:
- Around line 123-291: Add a defensive bounds check for channel_index against
the 6-bit limit (max 63) before allocating/writing in all GPIO serializers:
write_gpio_digital_data, write_gpio_digital_read_config, write_gpio_analog_data,
write_gpio_analog_read_config, write_gpio_digital_read_result, and
write_gpio_analog_read_result; define a private constant like kMaxChannelIndex =
(1U << 6) - 1U and verify channel_index <= kMaxChannelIndex (since
GpioHeader::ChannelIndex is BitfieldMember<8,6>) and return
SerializeResult::kInvalidArgument when out of range, mirroring the existing
period_ms validation style.
In `@host/include/librmcs/agent/c_board.hpp`:
- Around line 10-12: The include style is inconsistent: <librmcs/spec/gpio.hpp>
uses angle brackets while "librmcs/spec/c_board/gpio.hpp" uses quotes; change
the latter to use angle brackets to match the project's other includes (e.g.,
<librmcs/agent/common.hpp>, <librmcs/data/datas.hpp>,
<librmcs/protocol/handler.hpp>), so replace "librmcs/spec/c_board/gpio.hpp" with
<librmcs/spec/c_board/gpio.hpp> in the c_board.hpp header to unify include
style.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 7d4d34b8-b70e-491a-a1d3-ee3df3627467
📒 Files selected for processing (18)
core/include/librmcs/data/datas.hppcore/include/librmcs/spec/c_board/gpio.hppcore/include/librmcs/spec/gpio.hppcore/src/protocol/deserializer.cppcore/src/protocol/deserializer.hppcore/src/protocol/protocol.hppcore/src/protocol/serializer.hppfirmware/c_board/CMakeLists.txtfirmware/c_board/app/src/gpio/gpio.hppfirmware/c_board/app/src/usb/vendor.hppfirmware/rmcs_board/app/CMakeLists.txtfirmware/rmcs_board/app/src/usb/vendor.hppfirmware/rmcs_board/bootloader/CMakeLists.txthost/include/librmcs/agent/c_board.hpphost/include/librmcs/agent/rmcs_board_lite.hpphost/include/librmcs/agent/rmcs_board_pro.hpphost/include/librmcs/protocol/handler.hpphost/src/protocol/handler.cpp
qzhhhi
force-pushed
the
dev/rmcsboard-gpio
branch
from
c992634 to
3683953
Compare
There was a problem hiding this comment.
Actionable comments posted: 4
🧹 Nitpick comments (3)
firmware/c_board/app/src/timer/timer.hpp (1)
70-75: 建议补充check_reached的半周期前提。这个无状态回绕比较只在
deadline没有被安排到超过半个 32-bit tick 周期的未来时可靠；否则过远的未来时间点会被误判为“已到达”。当前 GPIO 调用路径看起来由to_duration_checked()约束，但这个方法是 public，建议把前提写在接口旁边，避免后续误用。可选注释补充
+ // Precondition: `deadline` must not be scheduled more than `kMaxDurationTicks` + // ticks in the future; farther deadlines are ambiguous under modular time. [[nodiscard]] bool check_reached(TimePoint deadline) const { const uint32_t deadline_ticks = deadline.time_since_epoch().count(); const uint32_t now_ticks = timepoint().time_since_epoch().count(); const uint32_t elapsed_ticks = now_ticks - deadline_ticks; return elapsed_ticks < kMaxDurationTicks; }🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@firmware/c_board/app/src/timer/timer.hpp` around lines 70 - 75, Add a clear precondition to check_reached: document (and optionally assert) that the supplied deadline must not be scheduled more than half of the 32‑bit tick wrap period into the future because the unsigned wrap-around comparison using kMaxDurationTicks is only reliable within that half‑period; reference the check_reached function name, the kMaxDurationTicks constant, and the public to_duration_checked() method in the comment near the API so callers know the constraint and avoid misusing the public to_duration_checked() path.host/include/librmcs/agent/rmcs_board_lite.hpp (1)
138-151: 建议与rmcs_board_pro.hpp保持一致，使用final替代override。
RmcsBoardPro对应的两个路由入口（rmcs_board_pro.hpp第 140 / 147 行）声明为final，防止用户子类误覆盖 channel_index 版本而旁路描述符分发；lite 这里只用override，功能上可行但与 pro 不对称，且一旦用户意外覆盖该重载会绕过边界检查与描述符路由。♻️ 建议的修改
void gpio_digital_read_result_callback( - uint8_t channel_index, const librmcs::data::GpioDigitalDataView& data) override { + uint8_t channel_index, const librmcs::data::GpioDigitalDataView& data) final { if (channel_index >= spec::rmcs_board_lite::kGpioDescriptors.size()) [[unlikely]] return; gpio_digital_read_result_callback( spec::rmcs_board_lite::kGpioDescriptors[channel_index], data); } void gpio_analog_read_result_callback( - uint8_t channel_index, const librmcs::data::GpioAnalogDataView& data) override { + uint8_t channel_index, const librmcs::data::GpioAnalogDataView& data) final { if (channel_index >= spec::rmcs_board_lite::kGpioDescriptors.size()) [[unlikely]] return; gpio_analog_read_result_callback( spec::rmcs_board_lite::kGpioDescriptors[channel_index], data); }Based on learnings: "avoid using both
finalandoverrideon the same virtual method: if a method is declaredfinal, do not also addoverridesince it is redundant. Usefinalalone forfinaloverrides."🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@host/include/librmcs/agent/rmcs_board_lite.hpp` around lines 138 - 151, The two channel-index overloads gpio_digital_read_result_callback and gpio_analog_read_result_callback in class RmcsBoardLite should match RmcsBoardPro by being declared final (not override) to prevent subclasses from bypassing the descriptor-boundary check and dispatch; find the methods in rmcs_board_lite.hpp and replace the trailing override specifier with final for both methods (do not use both override and final together).firmware/rmcs_board/app/src/gpio/gpio.hpp (1)
53-85: LGTM（一个小的可选建议）。
handle_port_interrupt遍历所有通道匹配端口并check_clear_interrupt_flag()，对 N=17 规模没问题；poll_periodic_input_samples的调度器正确地只在DigitalInput且period_ticks != 0时采样。一个可选的稳健性改进：当调度被长时间阻塞时，
state.next_sample_tick = now + state.period_ticks;会丢失相位（相对原计划时刻漂移）。若希望保持周期稳定，可改为state.next_sample_tick += state.period_ticks;并在严重滞后时快速赶齐或丢弃过期点。当前策略对大多数上位机采样场景也够用，属于可选优化。🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@firmware/rmcs_board/app/src/gpio/gpio.hpp` around lines 53 - 85, Update poll_periodic_input_samples to advance state.next_sample_tick by adding state.period_ticks instead of resetting to now to preserve phase: in poll_periodic_input_samples, replace the assignment state.next_sample_tick = now + state.period_ticks with state.next_sample_tick += state.period_ticks and add logic to catch up/skew handling (e.g., while now >= state.next_sample_tick { publish_digital_input_sample(...); state.next_sample_tick += state.period_ticks; } or drop excessive missed ticks) so the scheduler in poll_periodic_input_samples keeps a consistent phase; touch the same ChannelState fields and use publish_digital_input_sample/static_cast<uint8_t>(channel_index) and now_ticks() as needed.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@core/include/librmcs/spec/c_board/gpio.hpp`:
- Line 18: The constructor parameter for the derived GpioDescriptor in c_board
(and likewise in rmcs_board_pro and rmcs_board_lite) uses uint8_t; change its
signature to use std::uint8_t to match the base class spec::GpioDescriptor and
the file's fully-qualified style — update the constructor
declarations/definitions named GpioDescriptor(uint8_t channel_index,
GpioCapability capability_mask) to GpioDescriptor(std::uint8_t channel_index,
GpioCapability capability_mask) in each derived class.
In `@core/include/librmcs/spec/rmcs_board_lite/gpio.hpp`:
- Line 18: Replace the unqualified uint8_t parameter with std::uint8_t in the
GpioDescriptor constructors (e.g., change constexpr GpioDescriptor(uint8_t
channel_index, GpioCapability capability_mask) to use std::uint8_t) in the
GpioDescriptor declarations across the three board header files; also ensure
each header includes <cstdint> (or the common header that provides std::uint8_t)
so the qualified type compiles consistently with the base class.
In `@core/include/librmcs/spec/rmcs_board_pro/gpio.hpp`:
- Line 18: 构造函数参数类型在 rmcs_board_pro 的 GpioDescriptor（spec::GpioDescriptor）中使用了裸的
uint8_t；请将构造函数签名中的参数类型改为 std::uint8_t 以与基类
spec::GpioDescriptor（及其成员）保持一致，并对其它两个类似实现（rmcs_board_lite、c_board 中的
GpioDescriptor）做相同修改；确保包含或保留 <cstdint> 可用以解析 std::uint8_t。
In `@firmware/rmcs_board/app/src/gpio/gpio.hpp`:
- Around line 208-212: The duty16_to_pwm_compare function currently inverts duty
semantics vs c_board (duty=0→full on), causing a contradiction with
GpioAnalogDataView; first confirm whether inverted (active-low) PWM is
intentional for rmcs_board, and if so add a clear comment in
duty16_to_pwm_compare and/or GpioAnalogDataView documenting the active-low
convention and rationale; if not intentional, change duty16_to_pwm_compare to
standard semantics to match c_board by special-casing duty==65535 to return
pwm_reload_+1 and using return (static_cast<uint64_t>(pwm_reload_) * duty) /
65535U for the general case so duty=0→CMP≈0 and duty=65535→CMP≈pwm_reload_.
---
Nitpick comments:
In `@firmware/c_board/app/src/timer/timer.hpp`:
- Around line 70-75: Add a clear precondition to check_reached: document (and
optionally assert) that the supplied deadline must not be scheduled more than
half of the 32‑bit tick wrap period into the future because the unsigned
wrap-around comparison using kMaxDurationTicks is only reliable within that
half‑period; reference the check_reached function name, the kMaxDurationTicks
constant, and the public to_duration_checked() method in the comment near the
API so callers know the constraint and avoid misusing the public
to_duration_checked() path.
In `@firmware/rmcs_board/app/src/gpio/gpio.hpp`:
- Around line 53-85: Update poll_periodic_input_samples to advance
state.next_sample_tick by adding state.period_ticks instead of resetting to now
to preserve phase: in poll_periodic_input_samples, replace the assignment
state.next_sample_tick = now + state.period_ticks with state.next_sample_tick +=
state.period_ticks and add logic to catch up/skew handling (e.g., while now >=
state.next_sample_tick { publish_digital_input_sample(...);
state.next_sample_tick += state.period_ticks; } or drop excessive missed ticks)
so the scheduler in poll_periodic_input_samples keeps a consistent phase; touch
the same ChannelState fields and use
publish_digital_input_sample/static_cast<uint8_t>(channel_index) and now_ticks()
as needed.
In `@host/include/librmcs/agent/rmcs_board_lite.hpp`:
- Around line 138-151: The two channel-index overloads
gpio_digital_read_result_callback and gpio_analog_read_result_callback in class
RmcsBoardLite should match RmcsBoardPro by being declared final (not override)
to prevent subclasses from bypassing the descriptor-boundary check and dispatch;
find the methods in rmcs_board_lite.hpp and replace the trailing override
specifier with final for both methods (do not use both override and final
together).
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 38f83cd1-524e-4726-9b47-ce8100d8a04d
📒 Files selected for processing (29)
core/include/librmcs/data/datas.hppcore/include/librmcs/spec/c_board/gpio.hppcore/include/librmcs/spec/gpio.hppcore/include/librmcs/spec/rmcs_board_lite/gpio.hppcore/include/librmcs/spec/rmcs_board_pro/gpio.hppcore/src/protocol/deserializer.cppcore/src/protocol/deserializer.hppcore/src/protocol/protocol.hppcore/src/protocol/serializer.hppfirmware/c_board/CMakeLists.txtfirmware/c_board/app/src/gpio/gpio.hppfirmware/c_board/app/src/timer/timer.hppfirmware/c_board/app/src/usb/vendor.hppfirmware/rmcs_board/app/CMakeLists.txtfirmware/rmcs_board/app/src/app.cppfirmware/rmcs_board/app/src/gpio/analog_gpio_pin.hppfirmware/rmcs_board/app/src/gpio/gpio.cppfirmware/rmcs_board/app/src/gpio/gpio.hppfirmware/rmcs_board/app/src/usb/vendor.hppfirmware/rmcs_board/boards/lite/app/board_app.cppfirmware/rmcs_board/boards/lite/app/board_app.hppfirmware/rmcs_board/boards/pro/app/board_app.cppfirmware/rmcs_board/boards/pro/app/board_app.hppfirmware/rmcs_board/bootloader/CMakeLists.txthost/include/librmcs/agent/c_board.hpphost/include/librmcs/agent/rmcs_board_lite.hpphost/include/librmcs/agent/rmcs_board_pro.hpphost/include/librmcs/protocol/handler.hpphost/src/protocol/handler.cpp
✅ Files skipped from review due to trivial changes (3)
- firmware/rmcs_board/app/CMakeLists.txt
- host/src/protocol/handler.cpp
- core/include/librmcs/spec/gpio.hpp
🚧 Files skipped from review as they are similar to previous changes (9)
- firmware/rmcs_board/bootloader/CMakeLists.txt
- firmware/c_board/CMakeLists.txt
- core/src/protocol/protocol.hpp
- firmware/c_board/app/src/usb/vendor.hpp
- core/include/librmcs/data/datas.hpp
- core/src/protocol/deserializer.hpp
- firmware/c_board/app/src/gpio/gpio.hpp
- core/src/protocol/serializer.hpp
- host/include/librmcs/agent/c_board.hpp
qzhhhi
force-pushed
the
dev/rmcsboard-gpio
branch
from
3683953 to
9b23ead
Compare
qzhhhi
changed the title
There was a problem hiding this comment.
Actionable comments posted: 6
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
core/src/protocol/serializer.hpp (1)
123-288:⚠️ Potential issue | 🟠 Major把
channel_index校验升级为运行时参数校验。这些方法是公共序列化入口；当前只用
assert_debug校验通道位宽，release 构建里无效通道可能被写入/截断到协议头，而不是返回kInvalidArgument。建议像period_ms一样用LIBRMCS_VERIFY_LIKELY保护所有 GPIO 写入方法。建议修复
SerializeResult write_gpio_digital_data( uint8_t channel_index, const data::GpioDigitalDataView& view) noexcept { - utility::assert_debug(channel_index < (1U << GpioHeader::ChannelIndex::kBitWidth)); + LIBRMCS_VERIFY_LIKELY(valid_gpio_channel_index(channel_index), + SerializeResult::kInvalidArgument); const auto payload_type = view.high ? GpioHeader::PayloadEnum::kDigitalWriteHigh : GpioHeader::PayloadEnum::kDigitalWriteLow; @@ SerializeResult write_gpio_digital_read_config( uint8_t channel_index, const data::GpioReadConfigView& view) noexcept { - utility::assert_debug(channel_index < (1U << GpioHeader::ChannelIndex::kBitWidth)); + LIBRMCS_VERIFY_LIKELY(valid_gpio_channel_index(channel_index), + SerializeResult::kInvalidArgument); @@ SerializeResult write_gpio_analog_data( uint8_t channel_index, const data::GpioAnalogDataView& view) noexcept { - utility::assert_debug(channel_index < (1U << GpioHeader::ChannelIndex::kBitWidth)); + LIBRMCS_VERIFY_LIKELY(valid_gpio_channel_index(channel_index), + SerializeResult::kInvalidArgument); @@ SerializeResult write_gpio_analog_read_config( uint8_t channel_index, const data::GpioReadConfigView& view) noexcept { - utility::assert_debug(channel_index < (1U << GpioHeader::ChannelIndex::kBitWidth)); + LIBRMCS_VERIFY_LIKELY(valid_gpio_channel_index(channel_index), + SerializeResult::kInvalidArgument); @@ SerializeResult write_gpio_digital_read_result( uint8_t channel_index, const data::GpioDigitalDataView& view) noexcept { - utility::assert_debug(channel_index < (1U << GpioHeader::ChannelIndex::kBitWidth)); + LIBRMCS_VERIFY_LIKELY(valid_gpio_channel_index(channel_index), + SerializeResult::kInvalidArgument); @@ SerializeResult write_gpio_analog_read_result( uint8_t channel_index, const data::GpioAnalogDataView& view) noexcept { - utility::assert_debug(channel_index < (1U << GpioHeader::ChannelIndex::kBitWidth)); + LIBRMCS_VERIFY_LIKELY(valid_gpio_channel_index(channel_index), + SerializeResult::kInvalidArgument); @@ private: + static constexpr bool valid_gpio_channel_index(uint8_t channel_index) noexcept { + return channel_index < (1U << GpioHeader::ChannelIndex::kBitWidth); + } + static constexpr bool use_extended_field_header(FieldId field_id) {🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@core/src/protocol/serializer.hpp` around lines 123 - 288, The GPIO functions currently use utility::assert_debug to validate channel_index which is stripped in release builds; replace those asserts with a runtime check using LIBRMCS_VERIFY_LIKELY(channel_index < (1U << GpioHeader::ChannelIndex::kBitWidth), SerializeResult::kInvalidArgument) in each public serializer entry (write_gpio_digital_data, write_gpio_digital_read_config, write_gpio_analog_data, write_gpio_analog_read_config, write_gpio_digital_read_result, write_gpio_analog_read_result) — perform this verification before computing required size/allocating buffer so invalid channel indices return SerializeResult::kInvalidArgument instead of being truncated in release.
🧹 Nitpick comments (4)
core/include/librmcs/data/datas.hpp (1)
68-75: 建议给supported()加上[[nodiscard]]与noexcept，与GpioDescriptor::supports()保持一致。同文件里紧邻的
GpioDescriptor::supports()声明为[[nodiscard]] constexpr bool ... const noexcept。此处supported()没有返回值属性也没有noexcept，但实现只做 bool 短路 +supports()的 noexcept 调用，实际是无异常的纯函数，返回值也不应被调用方忽略（丢弃返回值就等于绕过能力校验）。♻️ 建议修改
- bool supported(const spec::GpioDescriptor& gpio) const { + [[nodiscard]] constexpr bool supported(const spec::GpioDescriptor& gpio) const noexcept { return (!asap || gpio.supports(spec::GpioCapability::kDigitalReadOnce)) && (!period_ms || gpio.supports(spec::GpioCapability::kDigitalReadPeriodic)) && ((!rising_edge && !falling_edge) || gpio.supports(spec::GpioCapability::kDigitalReadInterrupt)) && (pull != GpioPull::kUp || gpio.supports(spec::GpioCapability::kPullUp)) && (pull != GpioPull::kDown || gpio.supports(spec::GpioCapability::kPullDown)); }补
constexpr还有额外收益：配合kGpioDescriptors的 constexpr 描述符，future 代码中static_assert(config.supported(...))成为可能。🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@core/include/librmcs/data/datas.hpp` around lines 68 - 75, The supported(...) member should be marked [[nodiscard]] and noexcept (and consider constexpr) to match spec::GpioDescriptor::supports() and reflect that it is a pure, non-throwing boolean check; update the signature of supported(const spec::GpioDescriptor& gpio) to be [[nodiscard]] constexpr bool supported(...) const noexcept and keep the existing implementation unchanged so callers cannot accidentally ignore the result and the function may be used in compile-time contexts.firmware/rmcs_board/boards/pro/app/board_app.hpp (2)
16-16:core/include/...包含路径与 spec 头自身使用的<librmcs/spec/gpio.hpp>形式不一致。本文件用
"core/include/librmcs/spec/rmcs_board_pro/gpio.hpp"（相对仓库根的 quoted 路径），而该头内部却用#include <librmcs/spec/gpio.hpp>（angle-bracket，依赖core/include被加入 include path）。两种风格在同一调用链上混用会让新贡献者困惑：既然 firmware 构建目标已加入core/include（符合 PR 描述），这里也可以直接写#include <librmcs/spec/rmcs_board_pro/gpio.hpp>，与 host/spec 侧保持一致。非阻塞。🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@firmware/rmcs_board/boards/pro/app/board_app.hpp` at line 16, 在 board_app.hpp 中的包含使用了相对仓库根的 quoted 路径 ("core/include/librmcs/spec/rmcs_board_pro/gpio.hpp")，与该头内部和其他模块使用的 angle-bracket 风格 (<librmcs/spec/gpio.hpp>) 不一致；请在 board_app.hpp 中将该 include 改为 <librmcs/spec/rmcs_board_pro/gpio.hpp>（保留 IWYU pragma: export），以与其他 spec 头的包含风格一致并利用已加入的 core/include include-path；定位符：board_app.hpp 中的 `#include` 行和目标头 rmcs_board_pro/gpio.hpp。
60-79:kGpioHardwareDescriptors与 spec 侧kGpioDescriptors的映射已核对。按 channel_index：0–3 (PB00–PB03, PWM0_P_0..3) 对应
kPwm1..kPwm4；4–10 (PA9/10/11/12/13/18/19) 对应kSpiI2cSocket1..7；11–12 (PA22/PA23) 对应 UART5 RX/TX (kUart0Rx/Tx)；13–14 (PA14/PA15) 对应 UART3 RX/TX (kUart1Rx/Tx)；15–16 (PA1/PA0) 对应 UART0 RX/TX (kUart2Rx/Tx)。UART 映射与init_uart()中的 IOC 配置完全吻合；static_assert亦保障尺寸一致。一点建议：该表中 channel 顺序上"跳跃"（Socket 从 A9 连续到 A13 后跳到 A18/A19/A22/A23，随后 UART2 落回 A14/A15，再跳到 A0/A1）不容易一眼看出含义，建议在每个分组上方加一行
// PWM1..4/// SpiI2cSocket1..7/// UART0 RX/TX等简短注释（或按kGpioDescriptors中的kUart0Rx等命名引用同步命名），便于后续维护时和 spec 头对照。🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@firmware/rmcs_board/boards/pro/app/board_app.hpp` around lines 60 - 79, The GPIO descriptor array kGpioHardwareDescriptors is correct but its grouping/order is hard to read; add brief inline comments above each logical group (e.g., "// PWM1..4" for entries 0–3, "// SpiI2cSocket1..7" for entries 4–10, "// UART0 RX/TX" etc.) to map channel_index ranges to their spec names (kPwm1..kPwm4, kSpiI2cSocket1..7, kUart0Rx/Tx, kUart1Rx/Tx, kUart2Rx/Tx) so maintainers can quickly cross-reference kGpioDescriptors and init_uart() IOC mappings; keep the existing ordering and static_assert as-is.core/include/librmcs/spec/gpio.hpp (1)
43-45:supports()对组合 mask 的语义建议注释明确。
supports(capability)通过capability_mask & capability != 0判断，在传入单个位常量（如kDigitalWrite）时行为等价于 "contains"，这也是目前所有调用点（vendor.hpp、datas.hpp::supported、host agents）使用的方式。但如果将来有人写descriptor.supports(kDigitalReadPeriodic | kDigitalReadInterrupt)期望"两项都支持"，当前实现会在仅支持其中一项时错误返回true（any-of 语义）。建议二选一：
- 在注释中明确 "capability 必须是单一位" 的使用契约（若真的只打算支持单 bit）。
- 把实现改为
contains语义：(capability_mask & capability) == capability，这样单 bit / 组合 bit 都语义自然，也为kPwmCapabilities这类聚合位的后续使用留出空间。- [[nodiscard]] constexpr bool supports(GpioCapability capability) const noexcept { - return static_cast<std::uint8_t>(capability_mask & capability) != 0; - } + // Returns true iff `capability_mask` supports all bits set in `capability`. + [[nodiscard]] constexpr bool supports(GpioCapability capability) const noexcept { + return (capability_mask & capability) == capability; + }非阻塞；请按项目意图决定保留 any-of 还是改为 all-of。
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@core/include/librmcs/spec/gpio.hpp` around lines 43 - 45, The supports(GpioCapability) method currently uses (capability_mask & capability) != 0 which implements an "any-of" test; decide whether to require a single-bit input or to accept multi-bit masks and update accordingly: either add a clear comment on supports/GpioCapability stating the contract "capability must be a single-bit flag (e.g. kDigitalWrite)" and keep the current expression, or change the implementation to an all-of/contains test by replacing the expression with (capability_mask & capability) == capability so combinations like kDigitalReadPeriodic | kDigitalReadInterrupt and aggregates like kPwmCapabilities behave naturally; update the comment accordingly and ensure callers (vendor.hpp, datas.hpp::supported, host agents) conform to the chosen contract.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@firmware/c_board/app/src/gpio/gpio.hpp`:
- Around line 58-63: When handle_digital_read calls configure_digital_input_mode
and then immediately publishes an ASAP sample via publish_digital_input_sample
(when data.asap is true), postpone the next periodic deadline by one sample
period to avoid an immediate duplicate from poll_periodic_input_samples: after
publishing, advance the configured next-sample timestamp by data.sample_period
(or the internal sample_period used in configure_digital_input_mode) so the
periodic scheduler won't emit another sample immediately; apply the same change
to the analogous handler elsewhere in the diff that also uses config + asap
publishing.
In `@firmware/rmcs_board/app/src/gpio/analog_gpio_pin.hpp`:
- Around line 30-31: 当前代码在 gpio/analog_gpio_pin.hpp 中用 __builtin_trap() 检查
pwm_ioc_function 范围，这违反了“禁用 GNU 扩展”的约定；请用项目的断言工具替换：包含的 assert.hpp 中的
assert_debug 应用于 pwm_ioc_function 检查（例如用 assert_debug(pwm_ioc_function <= 31)
或先判断再 assert_debug(false)），将 __builtin_trap() 替换为
assert_debug(...)，保持原有约束并符合项目标准。
In `@firmware/rmcs_board/app/src/gpio/gpio.hpp`:
- Around line 46-51: handle_digital_read currently calls
configure_digital_input_mode (set_digital_input_mode_state) then, if data.asap,
immediately calls publish_digital_input_sample which can be followed immediately
by the periodic sampler because set_digital_input_mode_state initialized
next_sample_tick to now; after publishing when data.asap is true and the
configured period_ms > 0, advance the mode's next_sample_tick by one period
(e.g., next_sample_tick += period_ms or set to current_tick + period_ms) so the
next periodic sample is deferred one cycle; update this logic in
handle_digital_read (or in set_digital_input_mode_state) referencing
next_sample_tick, period_ms, data.asap, and publish_digital_input_sample.
In `@firmware/rmcs_board/app/src/usb/vendor.hpp`:
- Around line 156-160: 函数 gpio_analog_read_config_deserialized_callback 目前无视
channel_index 和 data；请改为按照 gpio_digital_read_config_deserialized_callback
的防御性模式实现：对 channel_index 做边界检查（与其他 GPIO 回调相同的 channel 数量判断），构造或获取对应的 descriptor
并用 data.supported(descriptor) 验证支持位（参照 spec::GpioCapability 的位掩码逻辑），在 validation
通过时再将请求转发到现有处理路径，否则记录或安全忽略；保持行为即便当前无 kAnalogRead* 能力位也遵循同样的验证与转发模式以避免未来静默丢包。
In `@firmware/rmcs_board/boards/pro/app/board_app.cpp`:
- Around line 205-208: The init_gpio_pins() function is intentionally an empty
stub; add a clear comment inside init_gpio_pins() stating that PWM IOC channels
are configured in descriptors, user buttons/switches and interrupt pins are
initialized in their own init routines, and remaining channels are configured
on-demand via
configure_digital_output_mode/configure_pwm_output_mode/configure_digital_input_mode
to avoid future confusion—refer to symbols init_gpio_pins,
SDK_DECLARE_EXT_ISR_M/gpio_a_isr, gpio_irq_handler, and GPIO_DI_GPIOA for
context.
In `@host/include/librmcs/agent/c_board.hpp`:
- Around line 52-74: 在 gpio_digital_write、gpio_digital_read 和 gpio_analog_write
三个方法中除了现有 capability/数据校验外还要验证传入的 GpioDescriptor 的 channel_index 落在 c_board
可接受的范围内：在每个方法开始处加入对 gpio.channel_index 的范围检查（例如与
spec::c_board::kGpioDescriptors.size() 或相应的描述符计数常量比较），如果超出范围则像其他失败情况一样抛出
std::invalid_argument（带上类似 "GPIO channel index out of range for c_board"
的描述性信息）；把这个范围校验放在 capability/data 写入之前以避免发送无效通道到主机。
---
Outside diff comments:
In `@core/src/protocol/serializer.hpp`:
- Around line 123-288: The GPIO functions currently use utility::assert_debug to
validate channel_index which is stripped in release builds; replace those
asserts with a runtime check using LIBRMCS_VERIFY_LIKELY(channel_index < (1U <<
GpioHeader::ChannelIndex::kBitWidth), SerializeResult::kInvalidArgument) in each
public serializer entry (write_gpio_digital_data,
write_gpio_digital_read_config, write_gpio_analog_data,
write_gpio_analog_read_config, write_gpio_digital_read_result,
write_gpio_analog_read_result) — perform this verification before computing
required size/allocating buffer so invalid channel indices return
SerializeResult::kInvalidArgument instead of being truncated in release.
---
Nitpick comments:
In `@core/include/librmcs/data/datas.hpp`:
- Around line 68-75: The supported(...) member should be marked [[nodiscard]]
and noexcept (and consider constexpr) to match spec::GpioDescriptor::supports()
and reflect that it is a pure, non-throwing boolean check; update the signature
of supported(const spec::GpioDescriptor& gpio) to be [[nodiscard]] constexpr
bool supported(...) const noexcept and keep the existing implementation
unchanged so callers cannot accidentally ignore the result and the function may
be used in compile-time contexts.
In `@core/include/librmcs/spec/gpio.hpp`:
- Around line 43-45: The supports(GpioCapability) method currently uses
(capability_mask & capability) != 0 which implements an "any-of" test; decide
whether to require a single-bit input or to accept multi-bit masks and update
accordingly: either add a clear comment on supports/GpioCapability stating the
contract "capability must be a single-bit flag (e.g. kDigitalWrite)" and keep
the current expression, or change the implementation to an all-of/contains test
by replacing the expression with (capability_mask & capability) == capability so
combinations like kDigitalReadPeriodic | kDigitalReadInterrupt and aggregates
like kPwmCapabilities behave naturally; update the comment accordingly and
ensure callers (vendor.hpp, datas.hpp::supported, host agents) conform to the
chosen contract.
In `@firmware/rmcs_board/boards/pro/app/board_app.hpp`:
- Line 16: 在 board_app.hpp 中的包含使用了相对仓库根的 quoted 路径
("core/include/librmcs/spec/rmcs_board_pro/gpio.hpp")，与该头内部和其他模块使用的
angle-bracket 风格 (<librmcs/spec/gpio.hpp>) 不一致；请在 board_app.hpp 中将该 include 改为
<librmcs/spec/rmcs_board_pro/gpio.hpp>（保留 IWYU pragma: export），以与其他 spec
头的包含风格一致并利用已加入的 core/include include-path；定位符：board_app.hpp 中的 `#include` 行和目标头
rmcs_board_pro/gpio.hpp。
- Around line 60-79: The GPIO descriptor array kGpioHardwareDescriptors is
correct but its grouping/order is hard to read; add brief inline comments above
each logical group (e.g., "// PWM1..4" for entries 0–3, "// SpiI2cSocket1..7"
for entries 4–10, "// UART0 RX/TX" etc.) to map channel_index ranges to their
spec names (kPwm1..kPwm4, kSpiI2cSocket1..7, kUart0Rx/Tx, kUart1Rx/Tx,
kUart2Rx/Tx) so maintainers can quickly cross-reference kGpioDescriptors and
init_uart() IOC mappings; keep the existing ordering and static_assert as-is.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 1f3ff6b9-fa56-41f8-8129-8a1d6307b557
📒 Files selected for processing (29)
core/include/librmcs/data/datas.hppcore/include/librmcs/spec/c_board/gpio.hppcore/include/librmcs/spec/gpio.hppcore/include/librmcs/spec/rmcs_board_lite/gpio.hppcore/include/librmcs/spec/rmcs_board_pro/gpio.hppcore/src/protocol/deserializer.cppcore/src/protocol/deserializer.hppcore/src/protocol/protocol.hppcore/src/protocol/serializer.hppfirmware/c_board/CMakeLists.txtfirmware/c_board/app/src/gpio/gpio.hppfirmware/c_board/app/src/timer/timer.hppfirmware/c_board/app/src/usb/vendor.hppfirmware/rmcs_board/app/CMakeLists.txtfirmware/rmcs_board/app/src/app.cppfirmware/rmcs_board/app/src/gpio/analog_gpio_pin.hppfirmware/rmcs_board/app/src/gpio/gpio.cppfirmware/rmcs_board/app/src/gpio/gpio.hppfirmware/rmcs_board/app/src/usb/vendor.hppfirmware/rmcs_board/boards/lite/app/board_app.cppfirmware/rmcs_board/boards/lite/app/board_app.hppfirmware/rmcs_board/boards/pro/app/board_app.cppfirmware/rmcs_board/boards/pro/app/board_app.hppfirmware/rmcs_board/bootloader/CMakeLists.txthost/include/librmcs/agent/c_board.hpphost/include/librmcs/agent/rmcs_board_lite.hpphost/include/librmcs/agent/rmcs_board_pro.hpphost/include/librmcs/protocol/handler.hpphost/src/protocol/handler.cpp
✅ Files skipped from review due to trivial changes (4)
- firmware/rmcs_board/bootloader/CMakeLists.txt
- firmware/rmcs_board/app/CMakeLists.txt
- core/src/protocol/protocol.hpp
- firmware/rmcs_board/app/src/gpio/gpio.cpp
🚧 Files skipped from review as they are similar to previous changes (12)
- firmware/c_board/CMakeLists.txt
- firmware/rmcs_board/app/src/app.cpp
- firmware/rmcs_board/boards/lite/app/board_app.cpp
- core/src/protocol/deserializer.cpp
- firmware/rmcs_board/boards/lite/app/board_app.hpp
- firmware/c_board/app/src/timer/timer.hpp
- firmware/c_board/app/src/usb/vendor.hpp
- core/src/protocol/deserializer.hpp
- host/include/librmcs/protocol/handler.hpp
- host/src/protocol/handler.cpp
- host/include/librmcs/agent/rmcs_board_lite.hpp
- host/include/librmcs/agent/rmcs_board_pro.hpp
qzhhhi
force-pushed
the
dev/rmcsboard-gpio
branch
from
9b23ead to
be1161e
Compare
Member
Author
|
@coderabbitai review |
✅ Actions performedReview triggered.
|
Member
Author
|
@coderabbitai pause |
✅ Actions performedReviews paused. |
- Extract channel field from GPIO data views (GpioDigitalDataView, GpioAnalogDataView, GpioReadConfigView) into an explicit channel_index parameter threaded through serialization, deserialization, and callbacks. - Introduce GpioCapability bitmask enum and GpioDescriptor in core/spec, with c_board-specific descriptors that declare per-channel capabilities (e.g. channel 5 lacks interrupt support). - Add GpioReadConfigView::supported() for unified host/firmware validation of read configurations against descriptor capabilities. - Replace hardcoded channel-6 EXTI exclusion in c_board firmware with declarative capability mask in the shared descriptor table. - Rename wire-protocol field GpioHeader::Channel to ChannelIndex and switch c_board channel numbering from 1-based to 0-based. - Update c_board host agent API to accept GpioDescriptor references instead of raw channel numbers. - Add core/include to firmware build targets that consume shared specs. BREAKING CHANGE: GPIO data views no longer carry a channel field; all GPIO protocol and callback APIs now take a zero-based channel_index. c_board host agent methods require a GpioDescriptor reference instead of a raw channel number.
… semantics (#48) - Add board-specific GPIO descriptors and host-side typed GPIO APIs for rmcs_board lite and pro. - Implement rmcs_board firmware GPIO handling for digital input, digital output, interrupt-driven input reporting, and PWM output on supported pro channels. - Refactor c_board digital input scheduling to use deadline-based periodic sampling. This changes the sampling behavior for read configurations with period_ms > 0. BREAKING CHANGE: Digital GPIO read configurations with period_ms > 0 now emit the first periodic sample on the next polling cycle instead of waiting for a full period to elapse.
qzhhhi
force-pushed
the
dev/rmcsboard-gpio
branch
from
be1161e to
e77f3d3
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
feat(gpio)!: Add descriptor-based GPIO capability model
BREAKING CHANGE: GPIO data views no longer carry a channel field; all GPIO protocol and callback APIs now take a zero-based channel_index. c_board host agent methods require a GpioDescriptor reference instead of a raw channel number.
feat(gpio)!: Add GPIO support for rmcs_board and align input sampling semantics
BREAKING CHANGE: Digital GPIO read configurations with period_ms > 0 now emit the first periodic sample on the next polling cycle instead of waiting for a full period to elapse.
GPIO 描述符模型重构（更新）
概览
本 PR 将 GPIO 支持从多处硬编码与 1-based 通道编号迁移为声明式的描述符驱动模型，统一使用零基 channel_index。改动横跨 core 协议、共享 spec、主机代理与固件（c_board、rmcs_board 及板级变体），引入能力位掩码、板级描述符表、协议字段/签名变更，并重构固件 GPIO 子系统以按描述符驱动行为与能力检查。
主要变更点
破坏性变更（需注意）
迁移建议
审查重点
总体而言，本 PR 用描述符与能力模型替换了多处板级特例并统一了通道索引语义，提升可扩展性与表达能力，但伴随较多接口变更与迁移工作。