DRAM Poison Table for Uncorrectable L2P Mapping Errors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory systems, particularly SSDs, face significant challenges with uncorrectable DRAM errors leading to data loss and system assertions, which can result in hours of recovery time and potential loss of critical user data.
Innovation Solution
Implementing a logic-to-physical (L2P) mapping table and a poison table in DRAM to manage data faults, marking uncorrectable errors as 'poisoned' to avoid data corruption and system assertions, allowing hosts to overwrite or trim such data, thereby enhancing reliability and availability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional error handling is used in DRAM, then uncorrectable errors cause system assertions and data loss, but the system requires hours of recovery time and risks critical user data loss
Solution Approach 1:
The patent applies preliminary action by pre-marking mapping entries in the L2P table as poisoned when uncorrectable errors are detected, before any data access occurs. This proactive marking prevents silent data corruption and allows the system to handle errors gracefully without requiring hours of recovery time or risking data loss. The poison table is updated in advance to track these problematic entries.
Solution Approach 2:
The patent introduces a poison table as an intermediary data structure that mediates between the L2P mapping table and the data storage. This poison table stores indices of poisoned mapping entries and acts as a filter, preventing access to corrupted data while allowing the rest of the system to continue operating normally. This intermediary mechanism resolves the contradiction by enabling fast error handling without system-wide disruption.
2Reliability
If uncorrectable DRAM errors are detected, then data integrity is compromised, but traditional handling causes system assertions and complete data loss
Solution Approach 1:
The patent converts the harmful effect of uncorrectable DRAM errors into a beneficial outcome by detecting these errors and marking the affected mapping entries as poisoned in the poison table. Instead of allowing silent data corruption or system assertions, the system transforms the error into a controlled state where the host can be notified and choose to overwrite or trim the affected data. This converts a potentially catastrophic failure into a manageable condition.
Solution Approach 2:
The patent implements feedback by continuously monitoring the L2P mapping table for uncorrectable errors and updating the poison table accordingly. When a poisoned entry is detected, the system provides feedback to the host through poison indication messages, allowing the host to take corrective actions such as overwriting or trimming the affected data. This feedback loop prevents data integrity compromise while avoiding system assertions.
3Reliability
If the poison table is introduced to manage errors, then reliability and availability improve, but the memory system complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the error management functionality into separate components: the L2P mapping table for address translation and the poison table for error tracking. This segmentation allows each component to have a specialized function, improving reliability without significantly increasing overall system complexity. The poison table is a simple data structure that stores only indices of poisoned entries, keeping the added complexity minimal.
Solution Approach 2:
The patent changes the state representation of mapping entries by adding a poison indicator parameter. Instead of fundamentally changing the memory system architecture, the solution modifies the L2P table entries to include a poison flag and uses a separate poison table to store indices of poisoned entries. This parameter-based approach improves reliability while maintaining system simplicity, as it builds upon existing data structures rather than introducing entirely new complex mechanisms.
Data Source
AI summary
This application is directed to managing errors on a controller of a memory device. The memory device includes non-volatile memory (NVM) storing data and dynamic random-access memory storing a logic-to-physical (L2P) table and a poison table. The memory device obtains a data access request to access a data item stored in the NVM, and the data access request includes a logical address of the data item. The memory device identifies, in the L2P table, a mapping entry corresponding to the logical address of the data item, and the mapping entry maps the logical address of the data item to a physical address of the data item within the NVM, and determines that the mapping entry has an uncorrectable error. In accordance with a determination that the mapping entry has the uncorrectable error, the memory device adds, in the poison table, an index identifying the mapping entry in the L2P table.