Electrical computers and digital processing systems: multicomput – Computer-to-computer data routing – Least weight routing
Reexamination Certificate
1997-09-24
2001-06-12
Banankhah, Majid (Department: 2151)
Electrical computers and digital processing systems: multicomput
Computer-to-computer data routing
Least weight routing
C711S202000
Reexamination Certificate
active
06247042
ABSTRACT:
FIELD OF THE INVENTION
The invention relates to memory management in a computer, and more specifically relates to a method for controlling the state of physical memory in a multitasking operating system that uses virtual memory to enable concurrently executing programs to share physical memory.
BACKGROUND OF THE INVENTION
The term “virtual memory” refers to a method for allowing several concurrently running application programs to share the physical memory of a computer. The physical memory refers to the main memory of a computer used to execute computer programs and is typically implemented with Random Access Memory (RAM). Multitasking operating systems typically use virtual memory to expand the memory available to each of the application programs executing in the computer. Virtual memory has the effect of making memory appear much larger to applications. To create this effect, a virtual memory manager (VMM) allocates memory from a virtual memory space that is much larger than the size of physical memory. The VMM uses secondary storage space in the computer such as a hard disk to extend the effective size of physical memory. The VMM only loads code and data from secondary storage to physical memory when an application actually needs it, e.g., to process a read or write request.
When a program makes a read or write request to virtual memory, the virtual memory manager determines whether the code or data requested is either located in physical memory or in secondary storage. If it is in physical memory, the virtual memory manager maps the virtual address into a physical address where it is located in physical memory. On the other hand, if the code or data is not in physical memory, the virtual memory manager fetches it from the secondary storage device and places it in physical memory. Thus, the virtual memory manager makes the physical memory appear larger to the application by swapping program code and data in and out of physical memory as needed to satisfy memory requests.
To illustrate the concept of virtual memory, consider an example of an operating system executing on a personal computer with 4 megabytes of physical memory and a hard drive with additional free memory space. The operating system itself might occupy up to a megabyte of the physical memory. If the user wishes to launch a game program occupying 2 Megabytes from the hard drive, then the total memory occupied in physical memory is about 3 Megabytes. Now assume that the game program attempts to load additional code or data files exceeding 1 Megabyte. Under these circumstances there is insufficient physical memory to hold the code and data for the currently executing programs in the computer.
The VMM solves this problem by swapping code and data needed to run the executing programs back and forth between physical memory and the hard drive. For example, if the instructions of a particular piece of code are to be executed, the piece of code must be loaded into physical memory of the computer. Other pieces of code can stay on disk until they are needed. Whenever a piece of code or data is not held in physical memory, the operating system marks its absence by setting (or clearing) a flag associated with that code or data. Then, if an access to that code or data is attempted, the processor will generate a not present interrupt that notifies the operating system of the problem. The operating system then arranges to load the missing code or data into an available area of physical memory and restarts the program that caused the interrupt. The swapping of code and data to and from the hard drive and the interrupts are transparent to the application programs executing in the computer in the sense that the application programs do not process the interrupt nor manage swapping of data back and forth. Rather, the application program only deals with a virtual address space of virtual memory, and the operating system maps requests for virtual memory to physical memory and swaps data back and forth between physical memory and the hard drive.
In a typical virtual memory system, some operating system components are guaranteed access to a portion of physical memory and several other software components contend for the remainder of physical memory. Operating system components that always occupy physical memory include memory resident components of the operating system kernel and a disk cache. The remainder of the physical memory is shared among other software such as dynamically loaded operating system components (DLLs), application program code and data, and dynamically allocated regions of memory such as Direct Memory Access (DMA) buffers and cache regions for the operating system's file system.
The operating system components that always occupy physical memory have a “lock” on a portion of the physical memory. A “lock” is an attribute of a memory management system that commits or reserves a portion of physical memory to a piece of code or data. In many operating systems, it is typical for a lock to be on a portion of physical memory if that memory contains a piece of code that must be able to run at interrupt time or a piece of data that needs to be accessible at interrupt time or that needs to be accessed asynchronously by hardware devices in the computer.
Initially, the operating system allocates virtual memory to the application programs. However, the operating system will not actually allocate physical memory to an application program until that program attempts to access memory. As code executing in the system attempts to access memory allocated to it, the operating system will allocate physical memory until it is filled, and then start to swap portions of physical memory to the hard drive to accommodate memory accesses.
The virtual memory system typically uses a portion of the hard drive, called a swap file, to swap code and data to and from physical memory. The operating system loads program code such as the executable code of an application program (e.g., a .exe file) directly from the hard drive. As an application requests access to program data, the operating system allocates physical memory, and subsequently, swaps this program data to and from physical memory once physical memory is filled up.
At run time, an application can either implicitly or explicitly request additional memory. An implicit request occurs when an application asks the operating system for a resource such as a new window, and the operating system allocates memory as a side effect to responding to the request for the resource. An explicit request occurs when the application directly invokes a function to specifically ask the operating system to allocate extra memory to it. In both cases, the operating system claims memory for resource allocation from virtual address space.
One form of virtual memory in common use today is referred to as paged virtual memory. In a paged virtual memory scheme, the operating system carries out all memory allocation, de-allocation, and swapping operations in units of memory called pages. In a microprocessor compatible with the 386 architecture from Intel Corporation, for example, a memory page is 4K and each memory segment is made up of one or more 4K pages. The Windows® 95 operating system is one example of an operating system that implements a paged virtual memory system.
Terms commonly used to describe a paged virtual memory scheme include paging, page file, and page fault. The term “paging” refers to the process of swapping code or data between physical memory and secondary storage. The term “page file” refers to the swap file maintained in a secondary storage device to hold pages of code and data swapped to and from the physical memory. Finally, the term “page fault” refers to an interrupt generated by a microprocessor indicating that the memory request cannot be satisfied from physical memory because the page containing the requested code or data is not located in physical memory.
The implementation details of any virtual memory system vary depending on the design and memory addressing scheme of the processor
Eisler Craig G.
Engstrom G. Eric
Banankhah Majid
Klarquist Sparkman Campbell & Leigh & Whinston, LLP
Microsoft Corporation
LandOfFree
Method and system for restoring the state of physical memory... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for restoring the state of physical memory..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for restoring the state of physical memory... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2447186