The memory manager components include a set of executive system services for allocating, deallocating, and managing virtual memory, a translation-not-valid and access fault trapper for hardware exceptions, and a set of kernel mode system thread routines as listed below:
The working set manager to handle global policies such as trimming, aging, and modified page writing.
The process/stack swapper that handles inswapping and outswapping.
The modified page writer that writes dirty pages in mapped files to disk.
The dereference segment thread which is responsible for system cache and page file growth shrinkage.
The zero page thread that zeros out pages on the free list/
The memory manager provides reserving and committing pages, locking memory, allocation granularity, shared memory and mapped files, protecting memory, copy on write, heap functions and address windowing extensions. A look at these in brief:
reserving and committing: reserved pages are not mapped to any storage. This is useful for large contiguous buffer. For example, when a thread is created, a stack is reserved. The committed pages are pages that when accessed, ultimately translate to valid pages in physical memory. Committed pages are either private and not share-able or mapped to a view of the section.
locking memory: Pages can be locked in memory in two ways:
1) Many pages can be requested by say a device driver and locked until it releases.
2) User mode applications can lock pages in their process working set.
Allocation Granularity This defines the system page size so that the risks associated with the assumptions about allocation alignment would be reduced.
Shared memory and mapped files Each process maintains its private memory area in which to store private data, but the program instruction and unmodified data pages could be shared. Shared memory is implemented via section objects
Memory is protected first by accessing all systemwide data structures and pools using kernel mode system components which user threads can't access.
Second, each process has a separate private address space which others can't access.
Third, in addition ti implicit protection, some form of hardware supported protection is also done.
And finally, shared memory section objects are protected via the standard access-control lists.
Copy on write page protection is an optimization the memory manager uses to conserve physical memory. Each new process that writes to a page will also get its own private copy.
A heap is a region of one or more pages of reserved address space that can be subdivided and allocated in smaller chunks by the heap manager.
Address windowing extensions is done by allocating the physical memory to be used, creating a region of virtual address space to act as a window to map views and mapping views of the physical memory into the window.
The working set manager to handle global policies such as trimming, aging, and modified page writing.
The process/stack swapper that handles inswapping and outswapping.
The modified page writer that writes dirty pages in mapped files to disk.
The dereference segment thread which is responsible for system cache and page file growth shrinkage.
The zero page thread that zeros out pages on the free list/
The memory manager provides reserving and committing pages, locking memory, allocation granularity, shared memory and mapped files, protecting memory, copy on write, heap functions and address windowing extensions. A look at these in brief:
reserving and committing: reserved pages are not mapped to any storage. This is useful for large contiguous buffer. For example, when a thread is created, a stack is reserved. The committed pages are pages that when accessed, ultimately translate to valid pages in physical memory. Committed pages are either private and not share-able or mapped to a view of the section.
locking memory: Pages can be locked in memory in two ways:
1) Many pages can be requested by say a device driver and locked until it releases.
2) User mode applications can lock pages in their process working set.
Allocation Granularity This defines the system page size so that the risks associated with the assumptions about allocation alignment would be reduced.
Shared memory and mapped files Each process maintains its private memory area in which to store private data, but the program instruction and unmodified data pages could be shared. Shared memory is implemented via section objects
Memory is protected first by accessing all systemwide data structures and pools using kernel mode system components which user threads can't access.
Second, each process has a separate private address space which others can't access.
Third, in addition ti implicit protection, some form of hardware supported protection is also done.
And finally, shared memory section objects are protected via the standard access-control lists.
Copy on write page protection is an optimization the memory manager uses to conserve physical memory. Each new process that writes to a page will also get its own private copy.
A heap is a region of one or more pages of reserved address space that can be subdivided and allocated in smaller chunks by the heap manager.
Address windowing extensions is done by allocating the physical memory to be used, creating a region of virtual address space to act as a window to map views and mapping views of the physical memory into the window.
No comments:
Post a Comment