International Business Machines Corporation

Wade Byron Ouren

Stuart Zachary Jacobs

Kenneth Charles Vossen

David Anthony Larson

Edward Curtis Prosser

Naresh Nayar

Richard Karl Kirkman

Bruce Gerard Mealey

Peter Joseph Heyrman

David Alan Hepkin

Michael Joseph Corrigan

Patrick Joseph McCarthy

Troy David Armstrong

Daniel Charles Birkestrand

Kenneth Roger Allen

William Joseph Armstrong

Richard Loise Arndt

Florian Alexander Auernhammer

Kyle Alan Lucke

Alongkorn Kitamorn

Andrew T Koch

Daniel James Henderson

Wayne L Lemmon

Paul LuVerne Godtland

Naresh Nayer

George David Timms, Jr

Christopher Brian Wallis

A throttling engine throttles access to a high latency hybrid memory. A request is received for partition mapping of a virtual address for an R/W memory page. An entry is added to a partition page table that maps a virtual address to a physical address and comprises access information that is R/W. A throttled flag is set in an entry of a partition page extension table. The throttle entry corresponds to the entry. The access information is saved in an original access part of the partition page extension table, and the access information is replaced with an R value. Upon application fault receipt, a throttling test is performed on an address of the application fault. If the throttling test is false, the fault is passed through to an operating system fault handler and the throttling fault stage is ended, otherwise, a delay is implemented for slowing access to the memory.

Throttling access to high latency hybrid memory DIMMs

Within a hashed page table maintained by a hypervisor, a special host real address configured to access a hardware device using access data intended to be invalidated is identified as part of a firmware assisted dump process for a virtual machine executing on the hypervisor. The special host real address is translated to a corresponding special guest real address. Within the hashed page table, the special host real address is replaced with the corresponding special guest real address, and the corresponding special guest real address is marked as invalid and requiring special handling. Subsequent to the replacing, the special host real address is invalidated. The special guest real address and translated address information are provided to the virtual machine subsequent to the invalidating, the translated address information comprising a set of host real addresses translated to corresponding guest real addresses.

Preserving transient address translation information across a firmware assisted dump procedure

A technique for handling interrupts includes receiving an event notification message (ENM) that specifies an event target number (ETN) and a number of bits to ignore (NBI). The ETN identifies a specific virtual processor thread (VPT) and the NBI identifies the number of lower-order bits of the specific VPT to ignore when determining a group of VPTs that may be potentially interrupted. In response to two or more VPTs within the group of VPTs being dispatched and operating on an associated physical processor, whether multiple of the two or more VPTs do not have a pending interrupt is determined. In response to determining that multiple of the two or more VPTs do not have a pending interrupt, one of the two or more VPTs is selected to service an interrupt associated with the ENM based, at least in part, on respective preferred bits for the two or more VPTs.

Techniques for indicating a preferred virtual processor thread to service an interrupt in a data processing system

Memory page de-duplication in a computer system that includes a plurality of virtual machine partitions managed by a hypervisor, where each virtual machine is assigned a different dedicated memory partition, may include: identifying, by the hypervisor, a plurality of identical memory pages in memory of one or more dedicated memory partitions; assigning, by the hypervisor, one of the identical memory pages as a master page; mapping, for each virtual machine having an identical memory page, each of the identical memory pages to the master page; and directing, by the hypervisor, reads of the memory page to the master page.

Memory page de-duplication in a computer system that includes a plurality of virtual machines

Relocating data in a virtualized environment maintained by a hypervisor administering access to memory with a Cache Page Table ('CPT') and a Physical Page Table ('PPT'), the CPT and PPT including virtual to physical mappings. Relocating data includes converting the virtual to physical mappings of the CPT to virtual to logical mappings; establishing a Logical Memory Block (&#x2018;LMB&#x2019;) relocation tracker that includes logical addresses of an LMB, source physical addresses of the LMB, target physical addresses of the LMB, a translation block indicator for each relocation granule, and a pin count associated with each relocation granule; establishing a PPT entry tracker including PPT entries corresponding to the LMB to be relocated; relocating the LMB in a number of relocation granules including blocking translations to the relocation granules during relocation; and removing the logical addresses from the LMB relocation tracker.

Relocating page tables and data amongst memory modules in a virtualized environment

In a computer system that includes multiple nodes and multiple logical partitions, a dynamic partition manager computes current memory affinity and potential memory affinity to help determine whether a reallocation of resources between nodes may improve memory affinity for a logical partition or for the computer system. If so, the reallocation of resources is performed so memory affinity for the logical partition or computer system is improved. Memory affinity is computed relative to the physical layout of the resources according to a hardware domain hierarchy that includes a plurality of primary domains and a plurality of secondary domains.

Dynamically improving memory affinity of logical partitions

Dynamic control of memory affinity is provided for a shared memory logical partition within a shared memory partition data processing system having a plurality of nodes. The memory affinity control approach includes: determining one or more home node assignments for the shared memory logical partition, with each assigned home node being one node of the plurality of nodes of the system; determining a desired physical page level per node for the shared memory logical partition; and allowing the shared memory partition to run and using the home node assignment(s) and its desired physical page level(s) in the dispatching of tasks to physical processors in the nodes and in hypervisor page memory management to dynamically control memory affinity of the shared memory logical partition in the data processing system.

Dynamic control of partition memory affinity in a shared memory partition data processing system

Hypervisor managed memory paging is provided in a data processing system having multiple logical partitions. The data processing system includes a shared memory pool defined within physical memory. The shared memory pool includes a volume of physical memory with dynamically adjustable sub-volumes or sets of physical pages associated with the multiple logical partitions. Each sub-volume or set is associated with a particular logical partition and includes mapped logical memory pages for that logical partition. A hypervisor memory manager interfaces the multiple logical partitions and the shared memory pool, and manages access to logical memory pages within the shared memory pool. The hypervisor memory manager further manages page-out and page-in of logical memory pages from the shared memory pool to one or more external paging devices. This page-out and page-in managing by the hypervisor memory manager is transparent to the multiple logical partitions.

Shared memory partition data processing system with hypervisor managed paging

In an embodiment, a target number of discretionary pages is calculated for a first partition. If the target number of discretionary pages for the first partition is less than a number of the discretionary pages that are allocated to the first partition, a result page is found that is allocated to the first partition and the result page is deallocated from the first partition. If the target number of discretionary pages for the first partition is greater than the number of the discretionary pages that are allocated to the first partition, a free page is allocated to the first partition.

Selecting a target number of pages for allocation to a partition

Migration management is provided for a shared memory logical partition migrating from a source system to a target system. The management approach includes managing migration of the logical partition from the source system to the target system by: transferring a portion of logical partition state information for the migrating logical partition from the source system to the target system by copying at the source system contents of a logical page of the migrating logical partition into a state record buffer for forwarding to the target system; forwarding the state record buffer to the target system; and determining whether the migrating logical partition is suspended at the source system, and if not, copying at the target system contents of the state record buffer to paging storage of the target system, the paging storage being external to physical memory managed by a hypervisor of the target system.

Managing migration of a shared memory logical partition from a source system to a target system

Transparent hypervisor pinning of critical memory areas is provided for a shared memory partition data processing system. The transparent hypervisor pinning includes receiving at a hypervisor a hypervisor call initiated by a logical partition to register a logical memory area of the logical partition with the hypervisor. Responsive to this hypervisor call, the hypervisor transparently determines whether the logical memory is a critical memory area for access by the hypervisor. If the logical memory area is a critical memory area, then the hypervisor automatically pins the logical memory area to physical memory of the shared memory partition data processing system, thereby ensuring that the memory area will not be paged-out from physical memory to external storage, and thus ensuring availability of the logic memory area to the hypervisor.

Growing community of inventors

Rochester, MN, United States of America

Wade Byron Ouren