Abstract: A data structure used by an operating system to refer to a filesystem object includes information that permits the structure to be associated with filesystem objects having different filesystem implementations. The data structure is associated with a filesystem object that is linked into different directories that have different filesystem implementations. Therefore, the data structure includes respective information within it about each different link. As a result, a filesystem object, such as a directory, having one filesystem implementation may be created in or moved to another directory having a different filesystem implementation; or, a file may be linked in multiple directories having different implementations.

Abstract: A filesystem conversion process converts a filesystem from one implementation type to another type while the filesystem remains active and fully accessible. The conversion process builds a list of directories in the filesystem that are of the old filesystem type and then converts each directory until the entire filesystem is converted. Instead of the filesystem being shutdown to perform the conversion, the conversion process may run as a background thread that does not interfere with normal filesystem operations.

Abstract: A data structure used by an operating system to refer to a filesystem object includes information that permits the structure to be associated with filesystem objects having different filesystem implementations. The data structure is associated with a filesystem object that is linked into different directories that have different filesystem implementations. Therefore, the data structure includes respective information within it about each different link. As a result, a filesystem object, such as a directory, having one filesystem implementation may be created in or moved to another directory having a different filesystem implementation; or, a file may be linked in multiple directories having different implementations.

Abstract: Multiple links to the same object are managed by storing the attributes of the object in only one place. Each directory which links to the object has an object attribute table; each table has a number of entries all capable of storing the attributes of the files linked to the directory. One of the entries in a directory linked to a file stores the attributes of that file, called the master entry; other entries in the same or other directories that also link to the same file have slave entries capable of storing the attributes but not storing the attributes in their respective object attribute table. The file need only address the master entry. The master entry may point to the other slave entries which may point back to the master or to other slave entries as determined by a pointer protocol. When the master entry is to be deallocated, another of the qualified slave entries receives the attribute information and the file is updated to reflect that a new master has been declared.

Abstract: Multiple links to the same object are managed by storing the attributes of the object in only one place. Each directory which links to the object has an object attribute table; each table has a number of entries all capable of storing the attributes of the files linked to the directory. One of the entries in a directory linked to a file stores the attributes of that file, called the master entry; other entries in the same or other directories that also link to the same file have slave entries capable of storing the attributes but not storing the attributes in their respective object attribute table. The file need only address the master entry. The master entry may point to the other slave entries which may point back to the master or to other slave entries as determined by a pointer protocol. When the master entry is to be deallocated, another of the qualified slave entries receives the attribute information and the file is updated to reflect that a new master has been declared.