# US 2008/0307268 A1 · Self-healing cache operations

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2008/0307268 A1
- Document ID: US20080307268A1
- Status: Published application
- Kind: A1
- Application number: 12/135,896
- Filed: 2008-06-09
- Published: 2008-12-11
- Earliest claimed: 2004-07-26
- Assignees: International Business Machines Corporation
- Inventors: Michael J. Azevedo; Carol Spanel; Andrew D. Walls
- Web page: https://patentgrep.com/patent/US20080307268A1

## CPC

- G06F 21/563: {by source code analysis}
- G06F 11/0715: {in a system implementing multitasking (multitasking per se G06F9/46)}
- G06F 11/1064: {in cache or content addressable memories}
- G06F 12/126: with special data handling, e.g. priority of data or instructions, handling errors or pinning
- G06F 21/564: {by virus signature recognition}
- G06F 11/0793: {Remedial or corrective actions (recovery from an exception in an instruction pipeline G06F9/3861; by retry G06F11/1402; for recovering from a failure of a protocol instance or entity H04L69/40)}
- G06F 2212/1032: Reliability improvement, data loss prevention, degraded operation etc
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A method, apparatus and program storage device for performing a self-healing cache process is described. At least one error affecting a cache is detected. The cache may have a matching address tag for a fetching operation. Based on the type of error, a self-healing cache process is performed based.

## Claims

### Claim 1

A method, comprising:
detecting at least one error affecting a cache having a matching address tag for a fetching operation;
determining a type of the at least one error affecting the cache; and
performing a self-healing cache process based on the determined type of the at least one error.

### Claim 2 (depends on 1)

The method of claim 1, wherein determining a type of the at least one error affecting the cache comprises determining the errors affecting the cache are instruction cache errors, data cache errors, unmodified copyback cache errors or modified copyback cache errors.

### Claim 3 (depends on 1)

The method of claim 1, wherein detecting at least one error affecting a cache comprises detecting at least one error affecting an entire cache line.

### Claim 4 (depends on 3)

The method of claim 3, wherein detecting at least one error affecting the entire cache set further comprises setting a deallocate flag for each way when an error is detected.

### Claim 5 (depends on 4)

The method of claim 4 further comprising invalidating the cache set when each way has the deallocate flag set.

### Claim 6 (depends on 4)

The method of claim 4, wherein determining a type of the at least one error affecting the entire cache set comprises determining the errors affecting the cache are parity errors.

### Claim 7 (depends on 3)

The method of claim 3, wherein performing a self-healing cache process based on the determined type of the at least one error comprises, when the determined type of the at least one error is an instruction or copyback cache error, invalidating the entire cache set, continuing with the fetching operation and asserting a deallocate flag for every way in the cache set.

### Claim 8 (depends on 3)

The method of claim 3, wherein performing a self-healing cache process based on the determined type of the at least one error comprises, when the determined type of the at least one error is a data cache error, invoking an error recovery procedure.

### Claim 9 (depends on 1)

The method of claim 1, wherein performing a self-healing cache process based on the determined type of the at least one error further comprises setting a deallocate flag for a way having an address matching the address associated with the fetch operation.

### Claim 10 (depends on 9)

The method of claim 9, wherein performing a self-healing cache process based on the determined type of the at least one error further comprises, when the determined type of the at least one error is one of an instruction cache type, a data cache type, or an unmodified copyback type, fetching data from a next level of cache and allocating data fetched from the next level to a different way in the cache set when the different way is available, else clearing a way in the cache and allocating data fetched from the next level to the cleared way in the cache set when a different way is not available.

### Claim 11 (depends on 9)

The method of claim 9, wherein performing a self-healing cache process based on the determined type of the at least one error further comprises when the determined type of the at least one error is a modified copyback type, generating an interrupt and providing the address for each way having a deallocate flag set.

### Claim 12 (depends on 11)

The method of claim 11, further comprising, when the interrupt is generated, recreating the data when data in the way having the deallocate flag set is capable of being recreated, returning from the interrupt and writing the recreated data in the way having the deallocated flag set when the data in the way having the deallocated flag set is capable of being recreated.

### Claim 13 (depends on 12)

The method of claim 12, further comprising restarting and resetting the deallocate flag of the way when the data in the way having the deallocated flag set is not capable of being recreated.

### Claim 14 (depends on 9)

The method of claim 9, wherein the deallocate flag is a counter and setting the deallocate flag comprises incrementing the counter and wherein performing a self-healing cache process based on the determined type of the at least one error further comprises, when the determined type of the at least one error is one of an instruction cache type, a data cache type, or an unmodified copyback type, determining a value of the counter and when the value of the counter is less than a predetermined threshold fetching data from a next level of cache and allocating data fetched from the next level to the way having the counter of the deallocate flag incremented.

### Claim 15 (depends on 9)

The method of claim 9, wherein the deallocate flag is a counter and setting the deallocate flag comprises incrementing the counter and wherein performing a self-healing cache process based on the determined type of the at least one error further comprises, when the determined type of the at least one error is one of an instruction cache type, a data cache type, or an unmodified copyback type, determining a value of the counter and when the value of the counter is not less than a predetermined threshold fetching data from a next level of cache and allocating data fetched from the next level to a different way in the cache set when the different way is available, else clearing a way in the cache and allocating data fetched from the next level to the cleared way in the cache set when a different way is not available.

### Claim 16 (depends on 15)

The method of claim 15, wherein the way is viewed as invalid when the value of the counter of the deallocate flag is equal to the predetermined threshold.

### Claim 17 (depends on 9)

The method of claim 9, wherein the deallocate flag is a counter and setting the deallocate flag comprises incrementing the counter and wherein performing a self-healing cache process based on the determined type of the at least one error further comprises, when the determined type of the at least one error is a modified copyback type, generating an interrupt and providing the address for each way having a deallocate flag set.

### Claim 18 (depends on 17)

The method of claim 17, further comprising, when the interrupt is generated, recreating the data when data in the way having the deallocate flag counter incremented is capable of being recreated, returning from the interrupt and writing the recreated data in the way having the deallocated flag counter incremented when the data in the way having the deallocated flag counter incremented is capable of being recreated.

### Claim 19 (depends on 18)

The method of claim 18, further comprising restarting and setting the deallocate flag counter to a predetermined threshold when the data in the way having the deallocated flag counter incremented is not capable of being recreated.
