操作系統(tǒng)課程設計試驗報告格式

1、四 川 大 學操作系統(tǒng)課程設計報告學 院： 軟件 學 院 專 業(yè)： 軟件工程 專 業(yè) 年 級： 07級 組 編 號： 第 X組 組 成 員： 喬心軻（姓名） 0743111340（學號） 何趙平（姓名） XXXXXXXX（學號） 崔蓉 （姓名） XXXXXXXX（學號） 張雯（姓名） XXXXXXXX（學號） 康小芳（姓名） XXXXXXXX（學號）提交時間： 2009年 月 日 指導教師評閱意見： . . . . .指導教師評閱成績：XXX1： XXX1： XXX1： XXX1： XXX1： 實驗項目一項目名稱：實驗目的：實驗時間：人員分工：實驗環(huán)境：實驗環(huán)境的搭建過程、選用的操作系統(tǒng)、機器

2、配置、編譯器等。實驗內(nèi)容：對實踐過程的詳細說明，針對已經(jīng)滿足的實踐要求，采用了何種算法或思想，對Nachos平臺的哪些代碼進行了什么樣的修改。實驗結(jié)果：對實踐要求的滿足程度，代碼是否編寫完成，是否調(diào)試通過，能否正常運行，本項目的要求中共滿足了哪幾項。參考文獻：實驗項目二實驗項目名稱：Nachos中的線程管理實驗項目目的：1.最多能夠同時存在128個用戶線程 2.改變?yōu)樽裱皟?yōu)先級調(diào)度”的搶占式調(diào)度 參與人員及分工：喬心軻，何趙平，康小芳完成主要代碼編寫；崔蓉，張文進行程序的測試及維護。實驗環(huán)境：n Nachos: Not Another Completely Heuristic Operati

3、ng Systemn Linuxn Gcc n Windows實驗內(nèi)容：1.對于最多能夠同時存在128個用戶線程，我們在Thread.cc中聲明了一個static變量numOfThreads；具體代碼如下：static int numOfThreads = 0;/the count of the threads在Thread的構(gòu)造函數(shù)中對其值進行加1；即每創(chuàng)建一個線程時，都會把numOfThreads加1；+numOfThreads; 并在SimpleThread（）中進行了如下修改，完成了最多能夠同時存在128個用戶線程。static voidSimpleThread(int which)

4、if(numOfThreads<128) for(inti=0;i<(kernel->currentThread->Time()-(kernel->currentThread->executeTime)+2;i+) cout << "* thread " <<which<<" looped "<<kernel->currentThread->executeTime<<" times.n"cout<<"pri

5、ority "<<kernel->currentThread->Priority()<<"n"kernel->currentThread->Yield(); elseif(count = 0) printf("The number of the threads can not be larger than 128!n");kernel->currentThread->Yield(); count+; 為了實現(xiàn)遵循“優(yōu)先級調(diào)度”的搶占式調(diào)度策略，首先為Thread增加了三個變量：exec

6、uteTime；time；priority在Thread.h中對它們的聲明：public: Thread(char* debugName);/ initialize a Thread Thread(); / deallocate a Thread/ NOTE - thread being deleted/ must not be running when delete / is called int executeTime;/The time that this thread has executed / basic thread operation void setExTime(int ex

7、T);/set the executeTime of the thread to exT int Time();/return the time that the thread should be served void setTime(int t);/set the time that the thread should be served to t void setPriority(int pri); /set the priority of the thread to priint Priority(); /return the priority of the thread privat

8、e: / some of the private data for this class is listed above int time; / the time that the thread should be served int priority; /the priority of each thread在Thread.cc中的實現(xiàn)：Thread:Thread(char* threadName) name = threadName; stackTop = NULL; stack = NULL; status = JUST_CREATED; for (int i = 0; i <

9、MachineStateSize; i+) machineStatei = NULL;/ not strictly necessary, since / new thread ignores contents / of machine registers space = NULL; executeTime=1; setPriority(0); setTime(2);在Fork（）中對numOfThreads進行了+1操作：void Thread:Fork(VoidFunctionPtr func, void *arg) Interrupt *interrupt = kernel->int

10、errupt;Scheduler *scheduler = kernel->scheduler;IntStatus oldLevel;DEBUG(dbgThread, "Forking thread: " << name << " f(a): " << (int) func << " " << arg); StackAllocate(func, arg); oldLevel = interrupt->SetLevel(IntOff); scheduler->

11、ReadyToRun(this);/ ReadyToRun assumes that interrupts / are disabled! (void) interrupt->SetLevel(oldLevel); +numOfThreads; /-/Thread:setExTime/ set the executeTime of the thread to exT/-voidThread:setExTime(int exT) executeTime = exT;/-/Thread:setPriority/ set the priority of the thread to pri/-v

12、oidThread:setPriority(int pri) priority=pri;/-/Thread:Priority/ return the priority of the thread/-intThread:Priority() return priority;/-/Thread:setPriority/ set the time of the thread to t/-void Thread:setTime(int t) time=t;/-/Thread:Time/ return the time of the thread need to run/-intThread:Time(

13、) return time;并在SimpleThread（）中進行了如下修改：static voidSimpleThread(int which) if(numOfThreads<128) for(int i=0;i<(kernel->currentThread->Time()-(kernel->currentThread->executeTime)+2;i+) cout << "* thread " <<which<<" looped "<<kernel->cur

14、rentThread->executeTime<<" times.n"cout<<"priority "<<kernel->currentThread->Priority()<<"n"kernel->currentThread->Yield(); elseif(count = 0) printf("The number of the threads can not be larger than 128!n");kernel->cur

15、rentThread->Yield(); count+; 對Scheduler.cc中進行了以下修改，實現(xiàn)了按照優(yōu)先級進行線程的調(diào)度，返回最高優(yōu)先級的線程，讓其運行。Thread *Scheduler:FindNextToRun () ASSERT(kernel->interrupt->getLevel() = IntOff); if (readyList->IsEmpty() return NULL; else Thread *highestPri = readyList->RemoveFront(); for (int i =1;i<readyList-

16、>NumInList();i+) if(readyList->IsEmpty() return highestPri; if(readyList->Front()->Priority()<=highestPri->Priority() readyList->Append(highestPri);highestPri = readyList->RemoveFront(); return highestPri; 并在Run（）中進行了以下修改：在Run（）代碼的最后加了以下語句：kernel->currentThread->setPrio

17、rity(kernel->currentThread->Priority()+1);kernel->currentThread->setExTime(kernel->currentThread->executeTime+1);實驗結(jié)果和結(jié)論：對實驗進行的測試，首先編譯通過測試：經(jīng)過編譯測試，程序已完成了以上要求，具體測試如下：對遵循“優(yōu)先級調(diào)度”搶占式調(diào)度的實現(xiàn)：voidThread:SelfTest() DEBUG(dbgThread, "Entering Thread:SelfTest"); for(int i=0;i<=3;i

18、+)Thread *t = new Thread("forked thread");t->setPriority(3-i);t->Fork(VoidFunctionPtr) SimpleThread, (void *) numOfThreads); 對最多只有128個用戶線程存在的實現(xiàn)：參考文獻：實驗項目三實驗項目名稱：Nachos中的文件管理系統(tǒng)實驗項目目的： 實現(xiàn)多個線程同時存在內(nèi)存，虛擬地址到物理地址的轉(zhuǎn)換，限制程序大小，不實現(xiàn)缺頁中斷處理 不實用虛擬內(nèi)存 維護二類頁表 實頁頁表（空閑頁表）：操作系統(tǒng)維護（Machine中創(chuàng)建）nachos現(xiàn)在有多少頁表是

19、可以供用戶使用 線程頁表：各線程維護自己的頁表參與人員及分工：喬心軻，何趙平，康小芳完成主要代碼編寫；崔蓉，張文進行程序的測試及維護。實驗環(huán)境：n Nachos: Not Another Completely Heuristic Operating Systemn Linuxn Gcc n Windows實驗內(nèi)容：首先是在Machine.h中對空閑頁表及其相關(guān)函數(shù)進行了聲明： TranslationEntry *freePageTable; /空閑頁表 TranslationEntry NextFreePage();/下一個空閑的頁 int NumOfFreePage();/剩余空閑頁的數(shù)目其

20、次在Machine.cc中空閑頁表進行初始化以及函數(shù)的實現(xiàn)：Machine:Machine(bool debug) int i; for (i = 0; i < NumTotalRegs; i+) registersi = 0; mainMemory = new charMemorySize; for (i = 0; i < MemorySize; i+) mainMemoryi = 0; freePageTable = new TranslationEntryNumPhysPages; for (i = 0; i < NumPhysPages; i+) freePageTa

21、blei.virtualPage=i;freePageTablei.physicalPage = i;freePageTablei.valid = TRUE;freePageTablei.use = FALSE;freePageTablei.dirty = FALSE;freePageTablei.readOnly = FALSE; #ifdef USE_TLB tlb = new TranslationEntryTLBSize; for (i = 0; i < TLBSize; i+)tlbi.valid = FALSE; pageTable = NULL;#else/ use lin

22、ear page table tlb = NULL; pageTable = NULL;#endif singleStep = debug; CheckEndian(); /-/Machine:NextFreePage()/-TranslationEntry Machine:NextFreePage() for(int i = 0;i<NumPhysPages;i+) if(freePageTablei.valid=TRUE)return freePageTablei; /-/Machine:numOfFreePage()/-intMachine:NumOfFreePage() int

23、num = 0; for(int i = 0;i<NumPhysPages;i+) if(freePageTablei.valid=TRUE)num+; return num;對AddrSpace.cc的修改如下：AddrSpace:AddrSpace() AddrSpace:AddrSpace() if(pageTable!=NULL) for(int i=0;i<numPages;i+)pageTablei.valid = TRUE;pageTablei.use = FALSE;pageTablei.dirty = FALSE;pageTablei.readOnly = FAL

24、SE; delete pageTable; 為了實現(xiàn)限制程序大小以及實現(xiàn)線程頁表和空閑頁表的轉(zhuǎn)換，對代碼進行了如下修改：bool AddrSpace:Load(char *fileName) OpenFile *executable = kernel->fileSystem->Open(fileName); NoffHeader noffH; unsigned int size; if (executable = NULL) cerr << "Unable to open file " << fileName << "

25、;n"return FALSE; executable->ReadAt(char *)&noffH, sizeof(noffH), 0); if (noffH.noffMagic != NOFFMAGIC) && (WordToHost(noffH.noffMagic) = NOFFMAGIC) SwapHeader(&noffH); ASSERT(noffH.noffMagic = NOFFMAGIC);#ifdef RDATA/ how big is address space? size = noffH.code.size + noffH.

26、readonlyData.size + noffH.initData.size + noffH.uninitData.size + UserStackSize; / we need to increase the size/ to leave room for the stack#else/ how big is address space? size = noffH.code.size + noffH.initData.size + noffH.uninitData.size + UserStackSize;/ we need to increase the size/ to leave r

27、oom for the stack#endif numPages = divRoundUp(size, PageSize); size = numPages * PageSize; ASSERT(numPages <= NumPhysPages);/ check we're not trying/ to run anything too big -/ at least until we have/ virtual memory DEBUG(dbgAddr, "Initializing address space: " << numPages <

28、;< ", " << size); pageTable = new TranslationEntrynumPages;if(kernel->machine->NumOfFreePage()>=numPages)for(int i=0;i<numPages;i+)TranslationEntry temp = kernel->machine->NextFreePage();pageTablei.virtualPage=temp.virtualPage;pageTablei.physicalPage=temp.physica

29、lPage;pageTablei.valid=FALSE;pageTablei.use=FALSE;kernel->machine->freePageTabletemp.physicalPage.valid = FALSE;kernel->machine->freePageTabletemp.physicalPage.use = TRUE; /sgeTablei.virtualPage=0;/ then, copy in the code and data segments into memory/ Note: this code assumes that virtua

30、l address = physical address int i=0;if (noffH.code.size > 0) DEBUG(dbgAddr, "Initializing code segment."); DEBUG(dbgAddr, noffH.code.virtualAddr << ", " << noffH.code.size); for(;i<(noffH.code.size/PageSize);i+) executable->ReadAt(&(kernel->machine-&g

31、t;mainMemorypageTablei.physicalPage), PageSize, noffH.code.inFileAddr+i*PageSize);cout<<"Page "<<pageTablei.virtualPage<<" for code segment is used!n" cout<<"Page for code is allocatedn" if (noffH.initData.size > 0) DEBUG(dbgAddr, "Initia

32、lizing data segment.");DEBUG(dbgAddr, noffH.initData.virtualAddr << ", " << noffH.initData.size);for(;i<(noffH.code.size/PageSize)+(noffH.initData.size/PageSize);i+)executable->ReadAt(&(kernel->machine->mainMemorypageTablei.physicalPage), PageSize, noffH.in

33、itData.inFileAddr+i*PageSize);cout<<"Page "<<pageTablei.virtualPage<<" for data segment is used!n" #ifdef RDATA if (noffH.readonlyData.size > 0) DEBUG(dbgAddr, "Initializing read only data segment.");DEBUG(dbgAddr, noffH.readonlyData.virtualAddr <

34、;< ", " << noffH.readonlyData.size);for(;i<(noffH.code.size/PageSize)+(noffH.initData.size/PageSize)+(noffH.readonlyData.size/PageSize);i+) executable->ReadAt( &(kernel->machine->mainMemorypageTablei.physicalPage), PageSize, noffH.readonlyData.inFileAddr+i*PageSize

35、);cout<<"Page "<<pageTablei.virtualPage<<" for readOnlydata segment is used!n" #endif elsecout<<"Not enough space for the program!n"delete pageTable; delete executable;/ close file return TRUE;/ success實驗結(jié)果和結(jié)論：程序已編譯通過，實現(xiàn)了1.多個線程同時存在內(nèi)存，虛擬地址到物理地址的轉(zhuǎn)換

36、，限制程序大小，不實現(xiàn)缺頁中斷處理2.維護二類頁表（1）實頁頁表（空閑頁表）：操作系統(tǒng)維護（Machine中創(chuàng)建）nachos現(xiàn)在有多少頁表是可以供用戶使用（2）線程頁表：各線程維護自己的頁表具體如下：編譯通過：測試：參考文獻：實驗項目四實驗項目名稱：Nachos中的文件管理模塊升級實驗項目目的：l 多個線程不能同時對一個文件進行寫操作,因為會發(fā)生沖突,但是可以同時進行讀操作.l 將簡單的通過信號量進行控制的同步機制修改為讀者/寫者問題的機制. l 實現(xiàn)目錄結(jié)構(gòu)的多級機制參與人員及分工：實驗環(huán)境：n Nachos: Not Another Completely Heuristic Operat

37、ing Systemn Linuxn Gcc n Windows實驗內(nèi)容：實現(xiàn)多個線程不能同時對一個文件進行寫操作,因為會發(fā)生沖突,但是可以同時進行讀操作. 將簡單的通過信號量進行控制的同步機制修改為讀者/寫者問題的機制：在SynchDisk.h中的修改如下： private: int readcount; /count the number of the readers int writecount; /count the number of the writers Disk *disk; / Raw disk device Semaphore *x; Semaphore *y; Semap

38、hore *z; Semaphore *wsem; Semaphore *rsem; 在SynchDisk.cc中的構(gòu)造函數(shù)中對信號量進行初始化。SynchDisk:SynchDisk() x = new Semaphore("synch disk", 1); y= new Semaphore("synch disk", 1); z = new Semaphore("synch disk", 1); wsem= new Semaphore("synch disk", 1); rsem= new Semaphore(

39、"synch disk", 1); lock = new Lock("synch disk lock"); disk = new Disk(this); readcount=0; writecount=0;在SynchDisk.cc中的析構(gòu)函數(shù)中把信號量的指針刪除。SynchDisk:SynchDisk() delete disk; delete lock; delete x; delete y; delete z; delete wsem; delete rsem;在ReadSector函數(shù)中進行的修改如下：voidSynchDisk:ReadSect

40、or(int sectorNumber, char* data)while(TRUE) z->P();/只有一個讀進程可以進入 rsem->P(); x->P(); readcount+; if(readcount=1) wsem->P(); x->V(); rsem->V(); z->V(); x->P(); lock->Acquire(); disk->ReadRequest(sectorNumber, data); lock->Release(); x->V(); readcount-; if(readcount=0

41、) wsem->V(); x->V();voidSynchDisk:WriteSector(int sectorNumber, char* data)while(TRUE)y->P();/多個寫進程可以進行排隊writecount+;if(writecount=1)rsem->P();y->V();wsem->P();lock->Acquire();disk->WriteRequest(sectorNumber, data);lock->Release();wsem->V();y->P();writecount-;if(writ

42、ecount=0)rsem->V();y->V();實現(xiàn)目錄結(jié)構(gòu)的多級機制：創(chuàng)建了一個SubFileHeader類：主要就是把FileHeader的內(nèi)容拷貝過去了；然后在FileHeader中進行了如下修改：boolFileHeader:Allocate(PersistentBitmap *freeMap, int fileSize) int hdr_size;SubFileHeader *subHdr30;numBytes = fileSize;if(fileSize<NumDirect*SectorSize) printf("filehdr.cc 50 ->n"); numHdrSectors = divRoundUp(fileSize, SectorSize); if (freeMap->NumClear() <