International Business Machines Corporation

Patrick M Gannon

John Stephen Liptay

Roger E Hough

Ronald M Smith

Lishing Liu

Michael Ignatowski

Kian-Bon K Sy

Wan Lin Leung

Matthew A Krygowski

Gregory Salyer

Peter H Gum

Robert E Murray

Donald W Price

Timothy R Marchini

Yee-Ming Ting

Andrew R Heller

Dale M Junod

William K Rodiger

Henry R Brandt

Julius Dwight Jones

Richard L Partridge

Thomas R Wright

Michael P Witt

A system for reducing purging of a translation lookaside buffer (TLB) to reduce operating system overhead in a system running multiple levels of virtual machines. A system typically must purge TLB entries whenever an underlying page table entry is invalidated due to paging activity on the host machine, or an underlying guest machine. A system for reducing the number of cases in which guest translations are based on host page table entries is provided. Additional logic is supplied to analyze each invalidate page table entry (IPTE) instruction to minimize the extent of purging required as a result of that instruction. When the region relocate facility is in operation, or when no pageable TLB's have been constructed, only the entry corresponding to the page table entry to be invalidated need be purged. This limited purging reduces the overhead by reducing the time spent in purging and the time required in address translation to rebuild the TLB. This time saving results in increased performance in systems with multi-level guests.

Apparatus and method for TLB purge reduction in a multi-level machine

A coherence directory and its methods of operation are disclosed for private processor caches in a multiple processor system to control data coherence in the system. It provides cross-invalidate (XI) controls for the assignment of exclusive and public ownership to data units in the processor caches, including required cross-invalidation of data units among the processor caches to obtain data coherence in the system in an efficient manner. The coherence directory can be used in a multiple processor system with or without any shared second level (L2) cache, shared or private. When a shared L2 cache is used to improve system access time, the coherence directory can also be used as the second level directory for the shared L2 cache and eliminate the need for any additional L2 directory(s).

Coherence control by data invalidation in selected processor caches

A unique high-speed hardware arrangement for generating double-level address translations in combination a translation look-aside buffer (TLB) structure that can store and lookup intermediate translations during a double-level translation. The hardware proceeds to the completion of a double-level translation without having to backup its operation, although an intermediate TLB miss is encountered, without danger of CPU deadlock occurring. The hardware arrangement also performs all single-level address translations required by the system.

Fast two-level dynamic address translation method and means

In the case of a cache miss, the successive fetch requests by the I-unit for sublines (e.g. doublewords) of a variable length field operand are provided by the first through the highest-address fetched sublines in a line being accessed from main storage via a cache bypass. This avoids the time delay for the I-unit caused by waiting until the complete line has been transferred to the cache before all required sublines in the line are obtainable from the cache. Address operand pairs (AOP's) consisting of request and buffer registers are provided in the I-unit to handle the fetched sublines as fast as the cache bypass can provide them from main storage. If there is a cache hit, the sublines are accessed from the cache.

Cache bypass control for operand fetches

This specification relates to performance of partial store operation in a hierarchical memory system which has a buffer store interposed between a processor interrogating the memory system and the main memory of the memory system. Such a partial store operation can be performed on a word of data in the main memory using the buffer store copy of that word of data. The copy of the word of data is read out of the buffer store into a register where it is modified to form a new word by replacing one or more but not all of the bytes in the word of data with bytes supplied by the processor. The new word is then placed in the main memory by performing a full store operation. The problem with performing a partial store operation in this manner is that the copy of the word of data in the buffer store may not be up-to-date. A technique is provided to eliminate the possibility of this old data being rewritten back into the main memory.

Technique for performing partial stores in store-thru memory

The described embodiment provides storage control (PSCF) for overlapping the handling of processor store requests between their generation by an instruction execution means (IPPF) and their presentation to system main storage (MS). The embodiment uses a store counter, an inpointer counter, an outpointer counter, a translator pointer register, an output counter and a plurality of registers sets to process and control the sequencing of all store requests so that the PSCF can output them to MS in the order received from the IPPF. The embodiment uses the counters to coordinate the varying delays in PSCF processing of plural store request contained in different register sets and the translator. The store counter obtains independence between plural IPPF operand address (OA) registers which send the store requests and plural PSCF register sets which handle the store request. The number of OA registers is made independent of the number of register sets. The store counter is also used for serializing instruction control.

High speed store request processing control

The embodiments relate to special controls in a processor which eliminate synonym entries in a translation lookaside buffer (DLAT) for a system which has DLAT entries that can concurrently translate virtual addresses in multiple address spaces into real main storage addresses. The controls provide a common space bit in any segment table entry (STE) or alternatively in any page table entry (PTE) in any private address space to indicate whether the segment or page, respectively, contains programs and data private to the address space or shared by all address spaces. Each DLAT entry contains a common/private storage indictor which is set to the state of the common space bit in the STE or PTE used in an address translation loaded into the DLAT entry. When the entry is read, the private/common storage indicator controls whether the DLAT can only be used by the address space identified in the DLAT, or by all address spaces.

Synonym control means for multiple virtual storage systems

Growing community of inventors

Poughkeepsie, NY, United States of America

Patrick M Gannon