Link Layer Register Persistent Error Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Hardware controllers face challenges in error recovery due to the overwrite of error information, making it difficult to assess and correct errors, which can lead to downtime and increased costs, as the conditions under which the error occurred need to be replicated to determine the cause.
Innovation Solution
A hardware controller is configured with a link layer register that persistently stores errors, allowing a processor to address them after they have been overwritten, enabling the differentiation between critical and noncritical errors and pausing I/O requests until critical errors are corrected.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If error information is stored in the link layer, then error detection capability is improved, but error information is overwritten when another error is detected, leading to loss of diagnostic data
Solution Approach 1:
The patent creates a copy of the error information from the link layer error register to a persistent storage location (system management controller or non-volatile memory). This copy mechanism ensures that even when the original error information is overwritten in the link layer, a preserved copy remains available for diagnostic purposes, thereby preventing information loss while maintaining reliable error recovery capability
Solution Approach 2:
The patent introduces an intermediary component (system management controller or error logging mechanism) that acts as a mediator between the link layer error register and persistent storage. This intermediary captures and stores error information, preventing its loss when overwritten, while allowing the link layer to continue functioning with its overwrite behavior, thus resolving the contradiction between information preservation and system reliability
2Ease of operation
If error information is persistently stored, then error recovery is simplified, but storage footprint on the hardware controller increases
Solution Approach 1:
The patent extracts the persistent error storage function from the link layer hardware and relocates it to a separate system management controller or external non-volatile memory. This extraction allows the link layer to maintain its compact form factor while error information is preserved externally, simplifying error recovery without significantly increasing the storage footprint of the hardware controller itself
Solution Approach 2:
The patent moves error storage from the horizontal dimension (link layer register within the controller) to a vertical dimension (external system management controller or non-volatile memory). This dimensional transition enables persistent error storage while maintaining the hardware controller's compact footprint, as the storage resource is utilized in a different architectural layer
3Measurement precision
If the hardware controller monitors errors continuously, then error detection capability is improved, but downtime increases when errors need to be replicated for analysis
Solution Approach 1:
The patent implements preliminary action by continuously capturing and persistently storing error information as it occurs, rather than waiting for errors to be replicated for analysis. The error information is preserved in real-time in the link layer register and copied to persistent storage, so when errors need to be analyzed, the data is already available, eliminating the need for time-consuming condition replication and reducing controller downtime
Data Source
AI summary
Methods and structure for providing methods and structure for recovering errors in a hardware controller after an overwrite event, such as the detection of another error. In this regard, a link layer of the hardware controller is configured with a register that persistently stores errors until a processor can address them. The link layer is adapted to establish a connection between an initiator and a target and detect errors associated with the connection. As each detected error is overwritten by a subsequently detected error, the link layer register persistently stores the detected errors associated with the connection for recovery after the detected error has been overwritten in the link layer at least until the error can be handled.