Mohammed Al-Sanabani

This commit introduces a **performance optimization** to the **LED service** by refining how LED status updates are calculated. The `LedManager` component in `fboss/led_service/LedManager.cpp` has been **refactored** to modify the `triggerLedUpdate` function, allowing it to accept and process only specific `PortID`s instead of iterating through all ports. This change, which also updates the `triggerLedUpdate` declaration in `fboss/led_service/LedManager.h`, ensures that LED calculations are performed exclusively for relevant ports, significantly reducing unnecessary computation and improving the efficiency of the **LED status update mechanism**.

This commit performs a **refactoring** within the **LED service**'s `LedManager` to improve code organization and prepare for future functionality. It extracts the core logic responsible for LED state calculation and updates from the existing `updateLedStatus` method into a new, dedicated private method named `triggerLedUpdate`. This change enhances modularity by centralizing the LED update mechanism, which is intended to be reused when subscribing to both `SwitchState` and `TcvrState` changes, streamlining how LED states are managed based on various system events.

This commit **updates the LED state calculation logic** within the **FBOSS LED service** to implement a new, more comprehensive formula. The `calculateLedState` function in `fboss/led_service/BspLedManager.cpp` now determines LED status based on port reachability, Loss of Signal (LOS) status, and drained state. This **feature enhancement** ensures LEDs accurately reflect the current operational state of network ports, with `fboss/led_service/LedManager.cpp` triggering updates when the LOS map changes. The change provides more precise visual feedback for network port conditions, and `fboss/led_service/hw_test/LedServiceTest.cpp` has been updated to validate the new blink patterns.

This commit **refactors** the **LED service** by **removing** the `portActive` state from the **LED state calculation** logic within `fboss/led_service/FsdbSwitchStateSubscriber.cpp`. Previously, port inactivity incorrectly influenced LED color, suggesting a fabric mismatch when it was not. This **cleanup** ensures that **LED indicators** accurately reflect true fabric issues, such as mismatched neighbors, rather than benign port states. By simplifying the determination of **port drainage**, this change improves the reliability and semantic correctness of the LED signaling system.

This commit introduces **new functionality** to the **`fboss/led_service`** by adding methods to detect **Receive Loss of Signal (Rx LOS)** conditions on neighboring ports and verify their reachability. Specifically, the `LedManager` class now includes `areAllNeighborsRxLos` and `portsUpAndCorrectReachability` to enhance LED state determination based on transceiver health. A new **unit test**, `testTcvrLos`, was added to validate LED behavior under various Rx LOS scenarios, ensuring the robustness of these new checks. Additionally, `BspLedManager::calculateLedState` was **refactored** to improve error logging with platform port names, further aiding in debugging the LED service.

This commit **adjusts the slow LED blink duration** within the **LED I/O library** to improve visual clarity. It updates the `kLedBlinkSlow` constant in `fboss/lib/led/LedIO.h` from 1000ms to 2000ms. This **functional adjustment** addresses feedback that previous blink durations (0.5s and 1.0s) were not easily distinguishable. The change enhances the **usability and clarity** of LED indicators, making the slow blink pattern more distinct for operators.

This commit performs a **refactoring** within the **LED service** to generalize its FSDB state subscription mechanism. It renames functions in `FsdbSwitchStateSubscriber` from `subscribeToSwitchState` to `subscribeToStates` and `removeSwitchStateSubscription` to `removeStateSubscriptions`, along with updating their parameters and call sites in `LedManager.cpp`. This **preparatory change** makes the subscription interface more generic and extensible, anticipating future additions like subscribing to `TcvrState` for Line of Sight (LOS) data, thereby improving the maintainability and flexibility of the LED control logic.

This commit introduces a **minor refactoring** within the **LED service** to optimize the `setLedState` function in `fboss/led_service/BspLedManager.cpp`. The changes specifically target improving the efficiency of LED controller retrieval and LED ID lookup operations when performed inside a loop. This **optimization** enhances the performance of LED state management, particularly for scenarios requiring frequent LED updates.

This commit introduces a **new capability** to the **LED service** by enabling it to subscribe to **QSFP transceiver state** from the Fboss State Database (FSDB). The `FsdbSwitchStateSubscriber` now actively subscribes to `qsfp_service().state().tcvrStates()` to retrieve transceiver data, including Loss of Signal (LOS) information. This data is then processed by the `LedManager` through a new `updateLedStatus` overload, allowing the system to accurately reflect the operational status of QSFP ports via their respective LEDs. This enhancement provides more granular visual diagnostics for network administrators regarding transceiver health.

This commit introduces a **new capability** within the **LED service** to provide more granular diagnostic feedback for multi-LED ports. It enables **individual control of LEDs based on Loss of Signal (LOS) status** for 800G ports when they are in a **drained state**. The `BspLedManager` now uses new logic in `setLedState` and a helper function `setLedBasedOnLOS` to differentiate LED colors (e.g., blinking blue for signal, blinking yellow for no signal) for each fiber pair. This enhancement significantly improves troubleshooting by allowing network engineers to quickly identify the status of individual fiber connections. A new test case in `LedServiceTest.cpp` verifies this individual LED color differentiation.

This commit introduces a **feature enhancement** and **refactoring** within the **BSP (Board Support Package) LED management subsystem**. The `getLedController` function in `fboss/lib/bsp/BspSystemContainer` and `getLedContainer` in `fboss/lib/bsp/BspPimContainer` are updated to return a `std::pair` that includes the LED controller or container along with the **set of associated lane IDs**. This change provides callers with a more complete context for LED operations, simplifying client code by directly supplying the relevant lane information. Consequently, the **LED service** (`fboss/led_service/BspLedManager.cpp`) is updated to correctly consume this enriched return type, improving the clarity and efficiency of LED state management.

This commit introduces a **new capability** to detect and report **mismatched neighbor** conditions on network ports within the **LED service**. It enhances the **diagnostic capabilities** by adding a `mismatchedNeighbor` field to the `SwitchStateUpdate` struct and modifying `fboss/led_service/FsdbSwitchStateSubscriber.cpp` to detect these errors. Consequently, the `fboss/led_service/LedManager.cpp`'s `updateLedStatus` logic now incorporates this new status, allowing the system to accurately reflect **cabling errors** via the LED state. This provides more granular insight into network connectivity issues, as mismatched neighbors also contribute to LLDP failures.

This commit introduces a **refactoring** to the **LED service** by implementing a **caching mechanism** for port neighbor state. It defines a new `PortNeighborState` struct within `LedManager.h` and integrates it into `PortDisplayInfo` to store pre-calculated neighbor information. This change **optimizes** the `calculateLedState` function in `BspLedManager.cpp` by leveraging the cached state, specifically improving map lookups within `getCommonLedSwPorts`. The caching helps in logging state changes and is utilized in test diffs, contributing to more efficient and stable LED state management.

This commit **fixes a bug** in the **LED control logic** within `fboss/lib/led/LedIO` where disabling LED blinking would incorrectly set the trigger to "timer" instead of "none". A new `setDelay` function is introduced and the existing `setBlink` function is refactored to ensure the LED trigger is correctly set to 'none' when blinking is disabled. This **bug fix** prevents **unintended LED blinking**, such as when a chassis is undrained, by ensuring the visual status indicators behave as expected. Test cases in `fboss/lib/led/tests/LedTests.cpp` are updated to verify this corrected behavior, specifically checking for the 'none' trigger when blinking is off.

This commit introduces a **new capability** to the **LED service** by enabling direct retrieval of LED states from the hardware. A new virtual method, `getLedStateFromHW`, has been added to the base `LedManager` class in `fboss/led_service/LedManager.h` and subsequently implemented in `fboss/led_service/BspLedManager.cpp`. This implementation in **`BspLedManager`** allows the system to query the actual, current state of LEDs for a given port directly from the underlying hardware via `LedIO`. This enhancement provides a more accurate and verifiable representation of the physical LED status, significantly improving the **LED management** subsystem's ability to reflect real-world conditions.

This commit **enhances the LED service's hardware tests** by introducing direct verification of LED color and blink states at the hardware level. Specifically, the `fboss/led_service/hw_test/LedServiceTest.cpp` file is updated to include new checks within its test helper functions. This **test enhancement** ensures that software modifications to LED states are accurately reflected on the physical hardware. The change improves the reliability of the **LED service** by validating its interaction with underlying hardware components, providing greater confidence in the system's ability to control LEDs as intended.

This commit **enhances the testing framework** for **LPO (Linear Pluggable Optics) transceivers** by introducing **RX LOS (Receive Loss of Signal) checking**. Previously, RX LOS checks were bypassed for LPO modules, but this change now explicitly verifies this condition. The **`verifyTxRxLatches` test** within `AgentEnsembleOpticsTest.cpp` and `OpticsTest.cpp` is updated to include this critical validation. This **feature addition** ensures more comprehensive and robust testing of LPO transceivers, improving their reliability and operational integrity by detecting potential signal loss issues.

This commit performs a **refactoring** within the **FBOSS agent's link tests**, specifically in `OpticsTest.cpp` and `AgentEnsembleOpticsTest.cpp`. It updates the method for retrieving the lane value from the deprecated `get_lane()` to the new `lane().value()` API. This change ensures the **test suite** remains compatible with the latest **optics API** for accessing lane information within the `verifyTxRxLatches` test. The scope is limited to **test code maintenance**, with no direct impact on production functionality.

This commit **re-enables and refines the `verifyTxRxLatches` test** within the **FBOSS agent's optics link test suite**. It introduces conditional logic to **skip Tx/Rx LOL (Loss of Lock) checks specifically for LPO (Linear Pluggable Optics) transceivers**, as these modules do not support such functionalities. This **test maintenance** ensures that the `verifyTxRxLatches` test can now run successfully and provide crucial **quality assurance** for LPO transceivers, improving the overall reliability of optics link functionality. The changes are applied to `AgentEnsembleOpticsTest.cpp` and `OpticsTest.cpp`, expanding test coverage without introducing false failures.

This commit **enhances transceiver state representation** by introducing a new `ModuleTechnology` enum within the `TcvrState` structure. This **feature addition** provides a more granular and extensible way to identify different module technologies, specifically improving the identification of **LPO (Linear Pluggable Optics) modules**, and deprecates the older `lpoModule` boolean field. Concurrently, the **FSDB model** is updated to include the new `moduleTechnology` field in its Thrift path representation, ensuring this detailed information is available for data storage and retrieval. This change standardizes how module technology is categorized, enabling more precise system behavior based on transceiver capabilities.