Mike Choi

This commit **updates the sensor service configuration** for the **Janga BIC platform** to **correctly set alarm and critical thresholds** for the `XP1R2V_SW_A` and `XP1R2V_SW_B` power rails. Previously, these thresholds were based on an incorrect 1.20V assumption, but the rails actually operate at 1.26V. This **maintenance update**, applied to `fboss/platform/configs/janga800bic/sensor_service.json`, ensures **accurate voltage monitoring** and **prevents erroneous alerts** on Janga devices. The change directly impacts the **platform's health monitoring system** by aligning sensor expectations with actual hardware specifications.

This commit introduces a **safety enhancement** to the **`showtech` utility** within the **FBOSS platform**, making its default operation non-disruptive. Potentially disruptive diagnostic functions now require an explicit `--disruptive` CLI flag and root privileges to execute, preventing accidental system interruptions. This **feature enhancement** modifies the execution logic in `fboss/platform/showtech/Main.cpp`, specifically impacting how details are requested and executed via `executeSingleDetail` and `executeRequestedDetails`. The change ensures that diagnostic data collection is safer by default, improving operational reliability and preventing unintended service disruptions.

This commit introduces a **security enhancement** to the **`fboss/platform/showtech`** utility by implementing a **root user check**. The `main` function within `fboss/platform/showtech/Main.cpp` is modified to ensure that the utility can only be executed by the root user. This **maintenance** change prevents non-root users from running the potentially disruptive `showtech` utility, thereby safeguarding system stability and preventing unintended operations.

This commit **disables** the execution of `i2cdetect` and `i2cdump` commands within the **FBOSS `showtech` utility** by modifying `fboss/platform/showtech/Main.cpp`. This is a **critical preventative fix** to address an issue where these I2C detection routines could cause **SiLab clock generators** on specific platforms, such as **Kodiak3**, to enter a bad state. Such a state could lead to **ASIC failure**, making this change essential for platform stability. The modification prevents potential hardware instability and includes code comments to explain the rationale for future maintainers.

This commit **updates the Meta Switch Architecture documentation** by adding a high-level entry to the **BMC-Lite Specification**. It introduces the concept of **hardware-assisted ASIC overtemperature protection and monitoring**, outlining its planned inclusion in the architectural design. This is a **documentation update** and an initial **architectural specification**, laying the groundwork for future detailed discussions and potential implementation of this critical safety feature. The change specifically affects the `docs/docs/architecture/meta_switch_architecture.md` file, providing an overview of this protection mechanism.

This commit **enhances the testing coverage** for the **`Fan_Service`** by adding new **unit tests** for its **optics data management** capabilities. The tests, located in `fboss/platform/fan_service/tests/SensorDataTests.cpp`, validate various aspects of optics entry management, including getting, updating, resetting, and timestamp processing. Furthermore, they cover critical edge cases such as non-existent entries, empty maps, and multiple concurrent updates. This **testing enhancement** ensures the robustness of the optics data fetch and processing logic, improving the overall reliability and maintainability of the `Fan_Service`'s sensor data handling.

This commit **adds new unit tests** for the **`SensorData` class** within the `fan_service` module, significantly enhancing its test coverage. This **testing improvement** covers comprehensive scenarios, including **CRUD operations** (create, retrieve, update, delete) for sensor entries, various **edge cases** like non-existent data and multiple updates, and **const correctness**. By expanding the test suite in `fboss/platform/fan_service/tests/SensorDataTests.cpp`, this work **improves the robustness and reliability** of sensor data management within the `fan_service` component.

This commit **adds comprehensive unit tests** for the `FsdbSensorSubscriber` component within the **`fan_service`**. The new tests, located in `FsdbSensorSubscriberTests.cpp`, validate critical functionalities such as static path generation methods, initial state verification, and the const-correctness of getter methods. This **enhances the reliability and robustness** of the `FsdbSensorSubscriber` module by ensuring its core behaviors function as expected. The work is a **testing enhancement** that improves the overall quality and maintainability of the `fan_service` by providing better coverage for its sensor data subscription logic.

This commit **removes the software overtemperature configuration** from the **Fan Service** for the **Tahan platform**. This **disables the software-based overtemperature protection** for the **TH5 ASIC**, which was causing issues due to inaccurate readings and erroneous shutdowns in Minerva testbeds. The change is a **maintenance fix** to prevent these problematic software-driven thermal responses. While the software protection is removed, the system still relies on the existing **hardware overprotection logic** for thermal safety, ensuring more stable and reliable operation.

This commit **revises the fan sensor thresholds** for **Janga** and **Tahan** platforms, specifically updating the `upperCriticalVal` and `maxAlarmVal` in their respective `sensor_service.json` configurations. This **maintenance update** is a direct result of thermal experts' analysis, which identified the existing thresholds as being too low. By adjusting these values in `fboss/platform/configs/janga800bic/sensor_service.json` and `fboss/platform/configs/tahan800bc/sensor_service.json`, the change aims to **improve the accuracy of thermal monitoring** and **reduce false positive fan warnings** on these devices. This ensures more robust and reliable operation of the **FBOSS platform's fan monitoring subsystem**.

This commit **refactors** the **`Sensor_service`** to adopt a new, dedicated **`asic_temp` Thrift data structure**, replacing the older `platform::sensor_data` for ASIC temperature readings. Specifically, the **`fan_service`** module's data fetching and subscription logic is updated, modifying methods like `getAsicTempThroughThrift` and `getAgentData` to utilize the new `AsicTempResponse` and `AsicTempData` types. Build configurations in `cmake` and `BUCK` files are also updated to link against the necessary `asic_temp_cpp2` libraries and dependencies. This **data model migration** ensures the **`fan_service`** correctly processes and utilizes the standardized ASIC temperature data format.

This commit **introduces a new Thrift data structure** specifically for **ASIC temperature sensor data** within the `fboss/agent/if` module. It defines dedicated structs such as `AsicTempData` and `AsicTempResponse`, along with an `AsicTempThrift` service and `getAsicTemp` function in `fboss/agent/if/asic_temp.thrift`. This **new capability** serves as a **refactoring** effort to replace the existing, more generic `platform::sensor_data` mechanism for ASIC temperature reporting. By providing a standardized and dedicated interface, this change improves the clarity and maintainability of sensor data handling for **ASIC temperature monitoring**.

This commit **refactors** the **FBOSS agent's statistics export** for **ASIC temperature sensor data** by adopting a new **Thrift data structure**, `AsicTempData`. The `agent_stats.thrift` definition is updated to include and utilize this new structure, and the `fillFsdbStats` function in `SwSwitch.cpp` is modified to populate it. Furthermore, the **FSDB key** used for exporting ASIC temperature is **renamed** to align with a new RFC, ensuring consistency. This **schema update** impacts any downstream systems consuming **FBOSS agent statistics** that rely on the previous ASIC temperature data format or key.

This commit **fixes a critical thermal management issue** on the **Whistler (Meru800bfa) platform** by updating the `fan_service` configuration. It **enables a boost mode** to increase fan speeds when QSFP optics temperature data is missing, which previously led to dangerously low fan Pulse Width Modulation (PWM). Additionally, the **minimum PWM threshold is increased to 40** to ensure adequate cooling, preventing potential overheating and hardware damage. This **platform-specific configuration update** improves system stability by maintaining appropriate fan speeds even when optics data is temporarily unavailable, such as during `fboss` stack updates.

This commit introduces a **new test capability** for the **fan_service** component, specifically targeting the `IncrementalPidLogic`. It adds `IncrementalPIDLogicTest.Basic` to `fboss/platform/fan_service/tests/PidLogicTests.cpp`, which verifies the behavior of the incremental PWM control. The tests cover positive and negative hysteresis, as well as PWM increase and decrease scenarios, ensuring the fan control system responds correctly to changing conditions. This work enhances the reliability of the platform's thermal management by validating its core fan speed adjustment logic.

This commit introduces **new functionality** to the **`fboss/platform/fan_service`** by adding **glue logic** within the **Board Support Package (BSP)** to retrieve **ASIC temperature data**. Specifically, it enables the `Bsp` class to fetch temperature readings from an agent using either **FSDB** (File System Database) or **Thrift** protocols. This **enhancement** involves new virtual functions like `getAsicTempData` and `getAsicTempDataOverThrift`, and updates the `Bsp` constructor to subscribe to agent stats from FSDB. This provides critical thermal monitoring capabilities for fan control by allowing the fan service to dynamically acquire ASIC temperature information through flexible communication mechanisms.

This commit **integrates ASIC temperature reading** into the main control logic of the **`fan_service`** module. A new helper function, `getAllTemperatureData`, is introduced within the **`ControlLogic`** component to consolidate all temperature data fetching, now including the critical ASIC temperature. The core `controlFan` method is updated to utilize this unified temperature data source, enabling more comprehensive and accurate thermal management. This **new capability** allows the fan service to make more informed decisions based on the ASIC's thermal state, enhancing system cooling and stability.

This commit introduces a **new capability** to the **FBOSS Fan Service** by enabling it to subscribe to and read **ASIC temperature data** from the `statsRoot.agent.asicTemp` FSDB entry. The `FsdbSensorSubscriber` component is extended with new methods, such as `subscribeToAgentStat` and `getAgentData`, to facilitate the retrieval of this critical thermal information. Additionally, this change includes **refactoring** of data freshness checks into a common helper function, `checkDataFreshness`, improving code maintainability within the subscriber. This enhancement allows the **Fan Service** to leverage ASIC temperature for more informed thermal management decisions.

This commit introduces a **new capability** to the **`fan_service`** by adding functions to its **`DataFetcher`** component. Specifically, it implements `getAsicTempThroughThrift` to perform **Thrift calls to the agent** for fetching **ASIC temperature** readings. This enhancement allows the `fan_service` to acquire critical thermal data directly from the ASIC, improving its ability to monitor and manage device cooling effectively. The change primarily impacts the data acquisition pipeline within the `fan_service`, enabling more comprehensive thermal management.

This commit **refactors sensor naming conventions** within the **FBOSS platform's Tahan800BC configuration files**. It updates `fan_service.json` and `sensor_service.json` to introduce more descriptive prefixes, such as `PDB_` for power brick sensors and `SMB_` for system management bus sensors, while also removing a redundant `TAHAN_` prefix from one entry. This **maintenance task** standardizes sensor identifiers, significantly improving clarity and consistency for **platform monitoring and fan control services**. The changes ensure that sensor data is reported with a more logical and consistent naming scheme across the **Tahan800BC hardware platform**.