Embedded Flash OS Storage With Parity Recovery for Bad Blocks

The patent performs bad block detection and data rewriting operations in advance during the factory initialization process, before the device is shipped. By proactively identifying problematic blocks and pre-rewriting data to alternative locations, the system eliminates the need for time-consuming re-writing operations after deployment, thus resolving the contradiction between writing reliability and time consumption.

The patent creates backup copies of critical data (Nboot and OS information) in multiple locations within Flash memory during initialization. This preparatory cushioning ensures that if a bad block is detected later, the system can immediately switch to backup copies without requiring time-consuming re-writing operations, thereby improving reliability while minimizing time loss.

The patent divides critical data (Nboot and OS information) into multiple separate blocks within Flash memory, rather than storing them in a single block. This segmentation allows the system to distribute data across multiple locations, so that if one block becomes bad, other blocks remain intact and can still provide the necessary information for system initialization, thus maintaining reliability while utilizing increased storage capacity.

The patent applies different quality standards to different blocks: block 0 is designated as the guaranteed fault-free boot block, while other blocks are used for OS information with the understanding that they may contain bad blocks. The system implements localized error handling for each block, ensuring that failures in non-critical blocks do not compromise overall system initialization, thereby enabling reliable operation despite varied block qualities.