Method of and architecture for controlling system data with automatic wear leveling in a semiconductor non-volatile mass storage memory

A memory system for storing data comprising:

• a semiconductor non-volatile mass storage memory partitioned into at least two areas, one of which includes sector storage locations used to store system files, a predetermined plurality of said sector storage locations of said one area being grouped in clusters, each said cluster being uniquely identifiable by a predetermined logical block address, a first sector storage location of a particular cluster being used for storing system sector information upon initial access of the particular cluster, and upon re-accesses of the particular cluster, prior to erasure of the particular cluster's sector information, subsequent sector storage locations of the particular cluster being used for storing replacement system sector information;
• and
• a control circuit coupled to the semiconductor non-volatile mass storage memory for controlling the partitioning thereof, for storing system sector information within the cluster, and for storing replacement system sector information within the cluster while leaving intact the system sector information previously stored in the particular cluster, said control circuit further being operative to generate an indication of the location of the most recent sector information stored within the particular cluster,
• thereby avoiding the need to erase the non-volatile mass storage memory every time a cluster identified by a logical block address is re-accessed.

A method of storing information in a semiconductor nonvolatile mass storage memory adapted to store information in sector storage spaces, comprising:

• partitioning the memory into a first group of sector storage spaces for storing a plurality of user files and a second group of sector storage spaces for storing a plurality of system files;
• combining the sector storage spaces of the second such group into clusters of storage spaces, each such cluster being uniquely identifiable by a logical block address and used for storing a current system sector within one of the sector storage spaces in that cluster;
• storing a system file in a particular one of said clusters;
• determining which sector storage space in the particular cluster is an empty sector storage space;
• writing a replacement system file for the stored system file into the empty sector storage space;
• providing an indication for identifying the location of the most recently written sector storage space of the particular cluster;
• and
• marking the particular cluster as old when a last sector storage space within the particular cluster is replaced and there is no remaining empty sector storage space within the particular cluster.