Managing Faulty Memory Pages In A Computing System

Abstract

Managing faulty memory pages in a computing system, including: tracking, by a page management module, a number of errors associated with a memory page; determining, by the page management module, whether the number of errors associated with the memory page exceeds a predetermined threshold; responsive to determining that the number of errors associated with the memory page exceeds the predetermined threshold, attempting, by the page management module, to retire the memory page; determining, by the page management module, whether the memory page has been successfully retired; and responsive to determining that the memory page has not been successfully retired, generating, by the page management module, a predictive failure alert.

1 . A method of managing faulty memory pages in a computing system, the method comprising:
 tracking, by a page management module, a number of errors associated with a memory page;   determining, by the page management module, whether the number of errors associated with the memory page exceeds a predetermined threshold;   responsive to determining that the number of errors associated with the memory page exceeds the predetermined threshold, attempting, by the page management module, to retire the memory page;   determining, by the page management module, whether the memory page has been successfully retired; and   responsive to determining that the memory page has not been successfully retired, generating, by the page management module, a predictive failure alert.   
     
     
         2 . The method of  claim 1  wherein tracking the number errors associated with the memory page further comprises:
 receiving, by the page management module, an indication that an error has occurred at a particular memory page; 
 determining, by the page management module, whether an error count table includes an entry associated with the particular memory page; 
 responsive to determining that the error count table includes an entry associated with the particular memory page, incrementing one or more counters associated with the particular memory page in the error count table; and 
 responsive to determining that the error count table does not include an entry associated with the particular memory page:
 removing, from the error count table, an entry that includes a lowest error count; and 
 adding, to the error count table, an entry for the particular memory page. 
 
 
     
     
         3 . The method of  claim 1  wherein tracking the number of errors associated with the memory page further comprises decrementing, upon expiration of a predetermined period of time, an error count for each entry in the error count table. 
     
     
         4 . The method of  claim 1  wherein determining whether the memory page has been successfully retired further comprises determining whether a new error has occurred at the memory page during a predetermined period of time. 
     
     
         5 . The method of  claim 1  wherein tracking a number of errors associated with a memory page includes:
 tracking a number of short term errors associated with the memory page; and 
 tracking a number of long term errors associated with the memory page. 
 
     
     
         6 . The method of  claim 1  further comprising:
 tracking, by the page management module, a number of retired pages for a memory module; 
 determining, by the page management module, whether the number of retired pages for the memory module exceeds a predetermined retirement threshold; and 
 responsive to determining that the number of retired pages for the memory module exceeds the predetermined retirement threshold, generating, by the page management module, a predictive failure alert. 
 
     
     
         7 - 20 . (canceled)