【正文】 oid)* 功 能：定時(shí)查詢TCP連接收發(fā)狀態(tài) ARP表更新， 并響應(yīng)* 入口參數(shù)：無* 出口參數(shù)：無* 說 明：* 調(diào)用方法：無 ****************************************************************************/ void eth_poll(void){ unsigned char i=0。 GPIO_WriteBit(GPIOA, GPIO_Pin_1, Bit_RESET)。 i UIP_CONNS。 //發(fā)送數(shù)據(jù)到以太網(wǎng)（設(shè)備驅(qū)動(dòng)程序） } } if UIP_UDP /* 輪流處理每個(gè)UDP連接, UIP_UDP_CONNS缺省是10個(gè) */ for(i = 0。 //加以太網(wǎng)頭結(jié)構(gòu)，在主動(dòng)連接時(shí)可能要構(gòu)造ARP請(qǐng)求 tapdev_send()。 uip_arp_timer()。 //從網(wǎng)絡(luò)設(shè)備讀取一個(gè)IP包,返回?cái)?shù)據(jù)長(zhǎng)度 if(uip_len 0) //收到數(shù)據(jù) { /* 處理IP數(shù)據(jù)包(只有校驗(yàn)通過的IP包才會(huì)被接收) */ if(BUFtype == htons(UIP_ETHTYPE_IP)) { uip_arp_ipin()。 //發(fā)送數(shù)據(jù)到以太網(wǎng)（設(shè)備驅(qū)動(dòng)程序） } } /* 處理arp報(bào)文 */ else if (BUFtype == htons(UIP_ETHTYPE_ARP)) //是ARP請(qǐng)求包 { uip_arp_arpin()。 tapdev_init()。 if 1 printf(uip ip address : 192,168,1,8\r\n)。 uip_ipaddr(ipaddr, 192,168,1,1)。 //設(shè)置網(wǎng)絡(luò)掩碼 uip_setnetmask(ipaddr)。 //中斷觸發(fā)讀取網(wǎng)絡(luò)接收緩存 eth_poll()。 (led_flag == 0x00) GPIO_WriteBit(GPIOA, GPIO_Pin_1, Bit_SET) 。 /* 輪流處理每個(gè)TCP連接, UIP_CONNS缺省是10個(gè) */ for(i = 0。 //加以太網(wǎng)頭結(jié)構(gòu)，在主動(dòng)連接時(shí)可能要構(gòu)造ARP請(qǐng)求 tapdev_send()。 /*處理UDP通信事件 */ /* 如果上面的函數(shù)調(diào)用導(dǎo)致數(shù)據(jù)應(yīng)該被發(fā)送出去，全局變量uip_len設(shè)定值 0 */ if(uip_len 0) { uip_arp_out()。arp_timer)。 uip_len = tapdev_read()。 //加以太網(wǎng)頭結(jié)構(gòu)，在主動(dòng)連接時(shí)可能要構(gòu)造ARP請(qǐng)求 tapdev_send()。 // }}第四章 系統(tǒng)的程序調(diào)試 系統(tǒng)的主程序調(diào)試 編寫程序根據(jù)各個(gè)程序模塊，編寫所需的程序，下面介紹一下用軟件KeiluVision4編寫過程。單擊，對(duì)所有的程序進(jìn)行編譯和連接，結(jié)果在“Build”里， 編譯和連接 燒寫代碼下載在“project”中選擇“options for target target1”，在界面中選擇“Debug”，設(shè)置“CortexM3 JLink”之后創(chuàng)建，將燒寫速度選擇2MHZ 。 系統(tǒng)調(diào)試結(jié)果 小燈顯示結(jié)果 網(wǎng)絡(luò)助手顯示結(jié)果第五章 畢業(yè)設(shè)計(jì)總結(jié) 在大學(xué)的學(xué)習(xí)過程中，畢業(yè)設(shè)計(jì)是一個(gè)重要的環(huán)節(jié)，和做科研開發(fā)工作一樣，要有嚴(yán)謹(jǐn)求實(shí)的科學(xué)態(tài)度。其間，查找資料，老師指導(dǎo)，與同學(xué)交流，反復(fù)修改，每一個(gè)過程都是對(duì)自己能力的一次檢驗(yàn)和充實(shí)。比如缺乏綜合應(yīng)用專業(yè)知識(shí)的能力，對(duì)設(shè)備的不了解，對(duì)具體設(shè)計(jì)涉及到的規(guī)范要求的不熟悉等等，需要在做的過程中需要去不斷的翻閱相關(guān)的資料和書籍，這降低了自己的速度和設(shè)計(jì)的進(jìn)程，但這個(gè)過程對(duì)我來說是對(duì)自己知識(shí)的不足處的一個(gè)很好的補(bǔ)充和對(duì)已學(xué)過知識(shí)的一個(gè)鞏固，在老師、同學(xué)以及自己的不斷努力下，不怕麻煩，不怕重復(fù)，當(dāng)克服了這些問題之后，我會(huì)感覺到自己的知識(shí)在一點(diǎn)一滴地積累，不知不覺中增加。這次實(shí)踐是對(duì)自己大學(xué)四年所學(xué)的一次大檢閱，使我明白自己知識(shí)還很淺薄，雖然馬上要畢業(yè)了，但是自己的求學(xué)之路還很長(zhǎng)，以后更應(yīng)該在工作中學(xué)習(xí)，努力使自己成為一個(gè)可以參與工作能獨(dú)立完成設(shè)計(jì)的人。陳老師在我畢業(yè)設(shè)計(jì)的過程中給了我充分的指導(dǎo)和幫助，并且在確定研究方案和收集有關(guān)資料時(shí)也給了我很大的幫助。我要感謝大學(xué)期間所有結(jié)識(shí)的朋友，是你們使我擁有了美好的大學(xué)時(shí)光！ 參考文獻(xiàn)[1] ，模擬部分5版[M]. 等教育出版社, [3][J].中國(guó)高新技術(shù)企業(yè), [4] (第4版)[M].北京:清華大學(xué)出版社,[5] (第2版) [M].北京航空航天大學(xué)出版社,[6] :基于8051+Proteus仿真[M].電子工業(yè)出版社,[7][D].青島大學(xué)碩士, [8][J].計(jì)算機(jī)光盤軟件與應(yīng)用,[9][J].電子世界, [10][D].杭州電子科技大學(xué),附錄一：STM32電路原理圖DM9000電路原理圖和驅(qū)動(dòng)電路原理圖附錄二：英文資料和譯文英文資料： Rainbow: ArchitectureBased SelfAdaptation with Reusable InfrastructureHua Liu and Manish ParasharThe Applied Software Systems LaboratoryDept of Electrical and Computer Engineering, Rutgers University, Piscataway, NJ08854, USAEmail:{marialiu, parashar}Salim HaririDepartment of Electrical and Computer Engineering, University of Arizona, AZ, USAEmail:hariri Abstract The Rainbow framework uses software architectures and a reusable infrastructure to support selfadaptation of software systems. The use of external adaptation mechanisms allows the explicit specification of adaptation strategies for multiple system concerns. 1. Introduction Softwarebased systems today increasingly operate in changing environments with variable user needs, resulting in the continued increase in administrative overhead for managing these systems. To reduce these costs, systems are increasingly expected todynamically selfadapt to acmodate resource variability, changing user needs, and system that support selfadaptation currently exist in the form of programming language features such as exceptions and in algorithms such as faulttolerant protocols. But these mechanisms are often highly specific to the application and tightly bound to the code. As a result, selfadaptation in today’s systems is costly to build, difficult to modify, and usually provides only localized treatment of system contrast to these internal mechanisms, recent work uses external models and mechanisms in a closedloop control fashion to achieve various goalsby monitoring and adapting system behavior at runtime. 1,2 As illustrated in Figure 1, control of system adaptation bees the responsibility of ponents outside the system that is being principle, external control mechanisms provide a more effective engineering solution than internal mechanisms for selfadaptation because they localize the concerns of problem detection andre solution in separable modules that can be analyzed, modified, extended, and reused across different systems. Additionally, developers can use this approach to add selfadaptation to legacy systems for which the source code may not be external approach requires using an appropriate model to reason about the system’s dynamic behavior. Several researchers have proposed using architectural models, 3 which represent the systemas a gross position of ponents, their interconnections, and their properties of interest. 4 Such an architecturebased selfadaptation approach offers many benefits. Most significantly, an abstract architectural model can provide a global perspective of the system and expose important systemlevel properties and integrity attractive in principle, architecturebasedselfadaptation raises a number of research and engineering challenges. First, the ability to handle a wide variety of systems must be addressed. Since different systems have radically different architecural styles, properties of interest, and mechanisms supporting dynamic modification, it is critical that the architectural control model and modification strategies be tailored to the specific system. Second,the need to reduce costs in adding external control to a system must be addressed. Creating the monitoring, modeling, and problemdetection mechanisms from scratch for each new system would render the approach prohibitively Rainbow framework attempts to address both problems. By adopting an architectu