Program address space overview

The Base Operating System provides a number of services for programming application program memory use.

Tools are available to assist in allocating memory, mapping memory and files, and profiling application memory usage. As background, this section describes the system's memory management architecture and memory management policy.

System memory architecture introduction

The system employs a memory management scheme that uses software to extend the capabilities of the physical hardware. Because the address space does not correspond one-to-one with real memory, the address space (and the way the system makes it correspond to real memory) is called virtual memory.

The subsystems of the kernel and the hardware that cooperate to translate the virtual address to physical addresses make up the memory management subsystem. The actions the kernel takes to ensure that processes share main memory fairly comprise the memory management policy. The following sections describe the characteristics of the memory management subsystem in greater detail.

The physical address space of 64-bit systems

The hardware provides a continuous range of virtual memory addresses, from 0x00000000000000000000 to 0xFFFFFFFFFFFFFFFFFFFF, for accessing data. The total addressable space is more than 1000000000000 terabytes. Memory access instructions generate an address of 64 bits: 36 bits to select a segment register and 28 bits to give an offset within the segment. This addressing scheme provides access to more than 64 million segments of up to 256M bytes each. Each segment register contains a 52-bit segment ID that becomes a prefix to the 28-bit offset, which together form the virtual memory address. The resulting 80-bit virtual address refers to a single, large, systemwide virtual memory space.

The process space is a 64-bit address space; that is, programs use 64-bit pointers. However, each process or interrupt handler can address only the systemwide virtual memory space (segment) whose segment IDs are in the segment register.

Segment register addressing

The system kernel loads some segment registers in the conventional way for all processes, implicitly providing the memory addressability needed by most processes. These registers include two kernel segments, and a shared-library segment, and an I/O device segment, that are shared by all processes and whose contents are read-only to non-kernel programs. There is also a segment for the exec system call of a process, which is shared on a read-only basis with other processes executing the same program, a private shared-library data segment that contains read-write library data, and a read-write segment that is private to the process. The remaining segment registers may be loaded using memory mapping techniques to provide more memory, or through memory access to files according to access permissions imposed by the kernel.

The system's 32-bit addressing and the access provided through indirection capabilities gives each process an interface that does not depend on the actual size of the systemwide virtual memory space. Some segment registers are shared by all processes, others by a subset of processes, and yet others are accessible to only one process. Sharing is achieved by allowing two or more processes to load the same segment ID.

Paging space

To accommodate the large virtual memory space with a limited real memory space, the system uses real memory as a work space and keeps inactive data and programs that are not mapped on disk. The area of disk that contains this data is called the paging space. A page is a unit of virtual memory that holds 4K bytes of data and can be transferred between real and auxiliary storage. When the system needs data or a program in the page space, it:

1. Finds an area of memory that is not currently active.
2. Ensures that an up-to-date copy of the data or program from that area of memory is in the paging space on disk.
3. Reads the new program or data from the paging space on disk into the newly freed area of memory.

Memory management policy

The real-to-virtual address translation and most other virtual memory facilities are provided to the system transparently by the Virtual Memory Manager (VMM). The VMM implements virtual memory, allowing the creation of segments larger than the physical memory available in the system. It accomplishes this by maintaining a list of free pages of real memory that it uses to retrieve pages that need to be brought into memory.

The VMM occasionally must replenish the pages on the free list by removing some of the current page data from real memory. The process of moving data between memory and disk as the data is needed is called "paging." To accomplish paging, the VMM uses page-stealing algorithms that categorize pages into three classes, each with unique entry and exit criteria:

• working storage pages
• local file pages
• remote file pages

In general, working pages have highest priority, followed by local file pages, and then remote file pages.

In addition, the VMM uses a technique known as the clock algorithm to select pages to be replaced. This technique takes advantage of a referenced bit for each page as an indication of what pages have been recently used (referenced). When a page-stealer routine is called, it cycles through a page frame table, examining each page's referenced bit. If the page was unreferenced and is stealable (that is, not pinned and meets other page-stealing criteria), it is stolen and placed on the free list. Referenced pages may not be stolen, but their reference bit is reset, effectively "aging" the reference so that the page may be stolen the next time a page-stealing algorithm is issued.

Memory allocation

Version 3 of the operating system uses a delayed paging slot technique for storage allocated to applications. This means that when storage is allocated to an application with a subroutine such as malloc, no paging space is assigned to that storage until the storage is referenced.