Abstract: A split directory-based cache coherency technique utilizes a secondary directory in memory to implement a bit mask used to indicate when more than one processor cache in a multi-processor computer system contains the same line of memory which thereby reduces the searches required to perform the coherency operations and the overall size of the memory needed to support the coherency system. The technique includes the attachment of a “coherency tag” to a line of memory so that its status can be tracked without having to read each processor's cache to see if the line of memory is contained within that cache. In this manner, only relatively short cache coherency commands need be transmitted across the communication network (which may comprise a Sebring ring) instead of across the main data path bus thus freeing the main bus from being slowed down by cache coherency data transmissions while removing the bandwidth limitations inherent in other cache coherency techniques.

Abstract: A virtual memory lock is placed upon a region of physical memory within a computer system in response to an I/O request through the use of a range lock. Each range lock represents pages of virtual memory that are present and locked in the physical memory. The range locks are cached in memory and used subsequently to process a lock or unlock request, thus avoiding constant locking or unlocking. Regions of memory that are locked, but have no outstanding I/O operations may still have a range lock existing corresponding to that region. If no range lock exists for an I/O request, the virtual memory lock function is called and a range lock is created for that region. If a range lock exists, its usage counter is incremented. Upon notification of the completion of an I/O operation upon a particular region, the usage counter for the range lock corresponding to that region is decremented, and the range lock continues to exist even if there are no outstanding I/O requests for that region.

Abstract: A flexible, low cost range locking mechanism allows a process requesting a lock to place a lock upon any requested range within a resource of a computer system. Various processes may hold locks upon different ranges of a resource simultaneously. A particular range may also be locked by different processes that are able to share the range. A sub-lock represents a unique range of the resource and has begin and end points that identify that portion of the requested range to which the sub-lock corresponds. A locked range may include numerous sub-locks. Each sub-lock has a mode indicating whether the sub-lock represents a shared lock, an exclusive lock or other. Sub-locks also have an incremental counter indicating the number of processes that hold a read lock upon a region, a flag to indicate whether a process is waiting to lock the region, a queue for listing waiting processes and other attributes.