Method and apparatus for substituting spare memory chip for malfunctioning memory chip with scrubbing

A data storage system, comprising:

• a memory array;
• fetch control means for fetching an uncorrected data word from said memory array;
• buffer means, coupled to said memory array, for temporarily storing said uncorrected data word;
• complementing control means, coupled to said buffer means, for complementing said uncorrected data so as to form a complemented uncorrected data word;
• write back control means, coupled to said complementing means and said memory array, for storing said complemented uncorrected data word back into said memory array at the location from which said uncorrected data word was fetched;
• retrieval control means, coupled to said memory array, for retrieving said complemented uncorrected data word from said memory array;
• recomplementing control means, coupled to said retrieval control means, for recomplementing said complemented uncorrected data word retrieved from said memory array so as to form a recomplemented uncorrected data word;
• comparing and count means, coupled to said buffer means and said recomplementing means for comparing each bit position in said recomplemented uncorrected data word with a like bit position in said uncorrected data word stored in said buffer and for counting for each bit position, a number of miscompares between said recomplemented uncorrected data word and said uncorrected data word stored in said buffer means;
• testing means, coupled to said compare and count means, for determining that said number of miscompares for any bit position has reached or exceeded a predetermined threshold;
• sparing vector generation means, coupled to said testing means, for determining which particular memory chip in said memory array stores data for said bit position reaching said threshold, selecting a spare memory chip to replace said particular memory chip for data reads, for initializing said spare memory chip, and copying all contents of said particular memory chip to said spare memory chip;
• and
• memory write control means, coupled to said sparing vector means, for continuing to perform scrubbing all memory write operations to both said spare memory chip and said particular memory chip.

A computer system, comprising:

• a) a central processor;
• b) a memory coupled to said central processor for retrieval and storage of data in said memory by memory read and memory write commands provided from said central processor, said memory comprising a plurality of normally on-line memory chips and a spare memory chip;
• said memory further comprising substituting means for causing said spare memory chip to be substituted for a malfunctioning one of said normally on-line memory chips;
• c) memory maintenance means coupled to said memory, said memory maintenance means comprising:
• means for testing and identifying a malfunctioning memory chip from among said plurality of normally on-line memory chips;
• and
• means for providing an identity of said malfunctioning memory chip to said memory;
• wherein in response to receiving said identity, said memory will substitute said spare memory chip for said malfunctioning memory chip for all substitution means read commands and will cause all memory write commands to be performed for both said spare memory chip and said malfunctioning memory chip.