Method, system, and program product of error recovery while decoding cached compressed data

1. A method for caching data, comprising:

loading compressed data into cache from a non-volatile storage device, wherein reentry data sets each include a pointer to one location in the compressed data and decoding information that enables decoding to start from the location;

receiving a request for data, wherein the compressed data in the cache includes the requested data; and

using the reentry data sets to decode only that portion of the compressed data in the cache that includes the requested data.

2. The method of claim 1 , wherein if the requested data is not in uncompressed format in the cache, then further performing:

determining a first location in the compressed data whose decoded output comprises the requested data;

determining a reentry data set whose pointer addresses a second location in the compressed data preceding the first location;

using the decoding information in the determined reentry data set to decode the compressed data from the second location through the first location to output the requested data in uncompressed format; and

returning the requested data in the uncompressed format.

3. The method of claim 2 , wherein determining the reentry data set whose pointer addressed the second location further comprises:

determining the reentry data set from the plurality of reentry data sets whose pointer addresses a closest location in the compressed data to the first location that precedes the first location.

4. The method of claim 1 , wherein the cache comprises a volatile memory that is capable of processing Input/Output (I/O) operations faster than the non-volatile storage device.

5. The method of claim 1 , wherein the decoding information includes decoded data preceding output produced by decoding the compressed data at the location addressed by the pointer and probability estimates used to enable decoding the compressed data at the location addressed by the pointer.