Abstract: An extended attribute data page includes extended attribute information about a number of files. The page is divided into areas, each area corresponding to a different file. A given area is further divided into a number of lines containing information about the extended attribute data. Areas include bitmaps indicating available space in the data space and address information for addressing off-page data areas. A directory is included of data types stored in the data area, and additional lines store extended attribute data. The extended attribute directory maintains a list of data types currently stored and related information. Including availability information about each line along with addresses to areas storing extended attribute data enables the extended attribute manager to efficiently determine where to place extended attribute data, mark deleted lines as being available, retrieve data being requested by a process, and modify the contents of a given extended attribute data type.

Abstract: A file system maintains information about files in data areas called dinodes. Dinodes are grouped into dinode pages wherein each dinode page includes information about one or more dinodes. Each dinode may use extended attribute data. Each dinode page having at least one dinode that uses extended attribute data has a corresponding dinodex page. The dinodex page includes extended attribute information about any of the dinodes in the dinode page that use extended attribute information. The page is divided into areas, each area corresponding to a different dinode. A lock is established for each dinode page so that any dinode adding or deleting its extended attribute address obtains the lock before proceeding to maintain cooperation between the dinodes. In this manner multiple extended attribute areas are prevented from corresponding with a particular dinode page and unused extended attribute pages are removed without unduly impacting system performance.

Abstract: A system and method for releasing storage space at the beginning of a byte stream while maintaining sequential byte stream semantics is provided. A ghost offset is initialized when a file is instantiated or opened. When information is deleted at the beginning of the sequential byte stream, the ghost offset is incremented in the amount of the number of bytes being deleted. The ghost offset continues to increment while the file is opened and information is being deleted at the beginning of the data stream. The virtual offset is the ghost offset added to the real offset of a particular entry in the data stream which maintains sequential byte stream semantics. When the file is closed, storage space is released in the amount of the ghost offset. Applications are provided with virtual offsets corresponding to data locations which are converted to real offsets for accessing data in the data stream.

Abstract: A method, computer program product and data processing system for accessing extended attributes. An extended attribute descriptor in a dinode may be read to determine if there exists extended attributes associated with a file system object that is associated with the dinode. An extended attribute descriptor points to a dinodex associated with the dinode if there exists an extended attribute associated with the file system object. The dinodex may then be read to retrieve the extended attritbutes stored either inside or outside the dinodex. However, if there is no extended attribute associated with that particular file system object, then the extended attribute descriptor comprises a null value and does not point to a dinodex associated with the dinode.

Abstract: A file system maintains information about files in data areas called dinodes. Dinodes are grouped into dinode pages wherein each dinode page includes information about one or more dinodes. Each dinode may use extended attribute data. Each dinode page having at least one dinode that uses extended attribute data has a corresponding dinodex page. The dinodex page includes extended attribute information about any of the dinodes in the dinode page that use extended attribute information. The page is divided into areas, each area corresponding to a different dinode. A lock is established for each dinode page so that any dinode adding or deleting its extended attribute address obtains the lock before proceeding to maintain cooperation between the dinodes. In this manner multiple extended attribute areas are prevented from corresponding with a particular dinode page and unused extended attribute pages are removed without unduly impacting system performance.

Abstract: A system and method for releasing storage space at the beginning of a byte stream while maintaining sequential byte stream semantics is provided. A ghost offset is initialized when a file is instantiated or opened. When information is deleted at the beginning of the sequential byte stream, the ghost offset is incremented in the amount of the number of bytes being deleted. The ghost offset continues to increment while the file is opened and information is being deleted at the beginning of the data stream. The virtual offset is the ghost offset added to the real offset of a particular entry in the data stream which maintains sequential byte stream semantics. When the file is closed, storage space is released in the amount of the ghost offset. Applications are provided with virtual offsets corresponding to data locations which are converted to real offsets for accessing data in the data stream.

Abstract: An extended attribute data page includes extended attribute information about a number of files. The page is divided into areas, each area corresponding to a different file. A given area is further divided into a number of lines containing information about the extended attribute data. Areas include bitmaps indicating available space in the data space and address information for addressing off-page data areas. A directory is included of data types stored in the data area, and additional lines store extended attribute data. The extended attribute directory maintains a list of data types currently stored and related information. Including availability information about each line along with addresses to areas storing extended attribute data enables the extended attribute manager to efficiently determine where to place extended attribute data, mark deleted lines as being available, retrieve data being requested by a process, and modify the contents of a given extended attribute data type.