摘要:
A system and method for dynamically altering a Virtual Memory Manager (VMM) Sequential-Access Read Ahead settings based upon current system memory conditions is provided. Normal VMM operations are performed using the Sequential-Access Read Ahead values set by the user. When low memory is detected, the system either turns off Sequential-Access Read Ahead operations or decreases the maximum page ahead (maxpgahead) value based upon whether the amount of free space is simply low or has reached a critically low level. The altered VMM Sequential-Access Read Ahead state remains in effect until enough free space is available so that normal VMM Sequential-Access Read Ahead operations can be performed (at which point the altered Sequential-Access Read Ahead values are reset to their original levels).
摘要:
A system and method for dynamically altering a Virtual Memory Manager (VMM) Sequential-Access Read Ahead settings based upon current system memory conditions is provided. Normal VMM operations are performed using the Sequential-Access Read Ahead values set by the user. When low memory is detected, the system either turns off Sequential-Access Read Ahead operations or decreases the maximum page ahead (maxpgahead) value based upon whether the amount of free space is simply low or has reached a critically low level. The altered VMM Sequential-Access Read Ahead state remains in effect until enough free space is available so that normal VMM Sequential-Access Read Ahead operations can be performed (at which point the altered Sequential-Access Read Ahead values are reset to their original levels).
摘要:
A fork system call by a first process is detected. A second process is created as a replication of the first process with a second affinity. If a replication of the replicated shared library is present in the second affinity domain, effective addresses of the replication of the replicated shared library are mapped using a mapping mechanism of the present invention to physical addresses in the second affinity domain.
摘要:
A method, an apparatus, and a computer program product are presented for memory page initialization operations. After an application thread attempts to reference a memory page, an exception or fault may be generated, and a physical memory page is allocated. The application thread is put to sleep, and a page initialization request is given to a kernel off-level worker thread, after which the interrupt-level processing is concluded. During the normal course of execution for the worker thread, the worker thread recognizes the page initialization request, and the worker thread initializes the newly allocated page by zeroing the page or by copying the contents of a source page to the newly allocated page, as appropriate. The worker thread then puts the application thread into a runnable state.
摘要:
A method, system and computer program product for allocating real memory to virtual memory page sizes when all real memory is in use is disclosed. In response to a page fault, a page frame for a virtual page is selected. In response to determining that said page does not represent a new page, a page is paged-in into said page frame a repaging rate for a page size of the page is modified in a repaging rates data structure.
摘要:
A method and system for efficiently migrating in-use small pages to enable promotion of contiguous small pages into large pages in a memory environment that includes small pages pinned to real memory and/or and small pages mapped to direct memory access (DMA) within real memory. The operating system is designed with a two-phase page promotion engine/utility that enables coalescing contiguous small virtual memory pages to create large virtual memory pages by migrating in-use small memory pages including those that are pinned and/or mapped to DMA.
摘要:
In a data processing system utilizing multiple page sizes for virtual memory paging, a system, method, and article of manufacture for managing page replacement. In one embodiment, the page replacement method begins with a page frame allocation request, such as may be generated following a page fault. A page replacement procedure is invoked to select one or more pages to be replaced by the requested page(s). In a preferred embodiment, the page replacement includes a step of selecting, in accordance with a page type allocation of at least one of the multiple page sizes, a page size to be utilized for page replacement for the page frame allocation request.
摘要:
A system and method for improving dynamic memory removals by reducing the file cache size prior to the dynamic memory removal operation initiating are provided. In one exemplary embodiment, the maximum amount of physical memory that can be used to cache files is reduced prior to performing a dynamic memory removal operation. Reducing the maximum amount of physical memory that can be used to cache files causes the page replacement algorithm to aggressively target file pages to bring the size of the file cache below the new maximum limit on the file cache size. This results in more file pages, rather than working storage pages, being paged-out.