第三章 存儲管理（memory management）武漢工業(yè)

1、LibraryModule#1Module#2Module#nLinkerLoadModuleLoaderRAM Maintains the states of partitionsUsing the logical segment no as the index to find the correspondent physics segment（使用邏輯段號作（使用邏輯段號作為索引查段表得到物理段號）；為索引查段表得到物理段號）；Physics address= offset in logical address + start address in segment table entry;

2、 Protection: If the offset in logical address is large than the length in the segment table entry, a protection error will generate and OS will process the error. const int BUFFSIZE=256; int main() char * buff=(char *)malloc(BUFFSIZE);/平臺無關(guān)sprintf(buff,”%s”,”Hello World”);/sleep(10000);/for kswapdfr

3、ee(buff); /平臺無關(guān)return SUCCESS; ./ hello world /實(shí)現(xiàn)映射,使用物理內(nèi)存,運(yùn)行時分配,運(yùn)行完回收v Modify BitP h y s i c a l M e m o r y P a g e T a b l e P a g e D i r e c t o r y T w o L e v e l P a g e T a b l e Virtual Address Page# Offset Page# Entry Chain (Hash) Frame# Hash Table Inverted Page Table Frame# Offset Real A

4、ddress Invert Page Table v Segment Limit Check OS uses special register to maintain the length of segment table and the length of every segment.v Access Control Every segment table entry has several access control bits to define the access limit.v Privilege Ring Every segment table entry has define

5、the privilege for process running in a ring.v A program can access the data which is in the same ring or lower privilege ring.v A program can call the functions which are in the same ring or higher privilege ring.v Paging, which is transparent to the programmer, eliminates external fragmentation and

6、 reduces the internal fragmentation, and thus provides efficient use of main memory.v Segmentation, which is visible to the programmer, has the ability to handle growing data structures, modularity, and support for sharing and protection.v To combine the advantages of both, using segmentation/paging

7、.v The design of memory management portion of an OS depends on three fundamental areas of choice:v The choices made in the first two areas depend on the hardware platform available. All important OS provide virtual memory. Pure segmentation systems are becoming rare.v The choices in the third areas

8、are the domain of OS. Gives a process a fixed number of pages within which to execute. The number is decided at initial load time and may be determined based on the type of process. When a page fault occurs, one of the pages of that process must be replaced by the needed page. The number of page fra

9、mes allocated to a process to be varied over the lifetime of the process. The number of pages allocated to a process can be decided by the page fault rate.v On a nonuniform memory access (NUMA), multiprocessor, the distributed, shared memory of the machine can be referenced by any processor on the m

10、achine, but the time for accessing a particular physical location varies with the distance between the processor and the memory module.v Thus, performance depends on the extent to which data reside close to the processors that use them. A page is brought into main memory only when a reference is mad

11、e to a location on that page. As more and more pages are brought in, page faults should drop to a very low level. Chooses only among the resident pages of the process that generated the page fault in selecting a page to replace.v How to determine resident set size and the timing of changes.-working

12、set strategy Page Address Stream 2 3 2 1 5 2 4 5 3 2 5 2 2 2 2 2 2 2 4 4 4 2 2 2 3 3 3 3 3 3 3 3 3 3 3 1 5 5 5 5 5 5 5 5 F F F F F F O OP PT T a al lg go or ri it th hm m Page Address Stream 2 3 2 1 5 2 4 5 3 2 5 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 5 5 5 5 5 5 5 5 1 1 1 4 4 4 2 2 2 F F F F F F F L LR RU

13、 U a al lg go or ri it th hm m v Examination of large numbers of traces of programs with sequential reference patterns shows that LRU performs poorly. Page Address Stream 2 3 2 1 5 2 4 5 3 2 5 2 2 2 2 2 5 5 5 5 3 3 3 3 3 3 3 3 2 2 2 2 2 5 5 1 1 1 4 4 4 4 4 2 F F F F F F F F F F FI IF FO O a al lg go or ri it th hm m v FIFO is a simple algorithm and is easy to implement.v For a program that has sequential reference patterns, FIFO can get good performanc