Memory Refresh Control Using Adjacent-Row Recharge After Interrupts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory devices fail to effectively protect against row hammer damage when an interrupt command is received during a sequential refresh operation, leading to potential damage of memory rows that have not yet reached their access count threshold.
Innovation Solution
A control method and memory device that includes a control circuit to manage access count values and perform refresh operations, where the refresh operation is interrupted on an interrupt memory row, and adjacent memory rows are recharged to prevent row hammer damage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the refresh operation is interrupted by an interrupt command during sequential refresh, then the interrupt memory row can be accessed immediately, but adjacent memory rows are not protected against row hammer damage
Solution Approach 1:
The patent applies preliminary action by recharging adjacent memory rows in advance after an interrupt occurs. When an interrupt command is received during refresh operation, the system identifies the interrupt memory row and proactively recharges its adjacent rows (n-1 and n+1) before they can be vulnerable to row hammer attacks. This preliminary protective action ensures that even though the refresh was interrupted, the adjacent rows are restored to a safe state.
Solution Approach 2:
The patent implements preliminary anti-action by counteracting the potential row hammer effect before it can cause damage. After detecting an interrupt and identifying the interrupt memory row, the system immediately applies a counter-measure by recharging the adjacent memory rows. This anti-action neutralizes the vulnerability that would otherwise exist, preventing the cumulative charge effect that leads to row hammer damage.
2Reliability
If the refresh operation continues sequentially without interruption, then all memory rows are protected, but the access time is increased
Solution Approach 1:
The patent applies the taking out principle by extracting the essential protective action from the complete sequential refresh process. Instead of requiring the full sequential refresh of all memory rows, the system identifies only the critical adjacent rows (n-1 and n+1) that need recharging after an interrupt. This extracted approach maintains adequate protection while significantly reducing the time loss compared to completing the entire sequential refresh cycle.
3Ease of operation
If access count values are reset sequentially during refresh, then the refresh operation is systematic, but some rows may be missed when interrupted
Solution Approach 1:
The patent implements feedback by using the interrupt command itself as a signal to trigger a protective response. When an interrupt is detected, the system receives feedback about which row is being accessed (the interrupt memory row) and uses this information to determine which adjacent rows need recharging. This feedback mechanism ensures that the protection is targeted and adaptive, maintaining reliability even when the systematic sequential refresh is interrupted.
Data Source
AI summary
A memory device and a control method of the memory device are provided. The control method includes: providing access count values to memory rows of the memory device respectively when the memory row are accessed; performing a refresh operation on the memory rows sequentially and resetting the access count values sequentially in response to a refresh command; receiving an interrupt command to interrupt the refresh operation, wherein the refresh operation is interrupted when performing on an interrupt memory row; completing the refresh operation on the interrupt memory row; and recharging at least one adjacent memory row next to the interrupt memory row.