(3.235.245.219) 您好!臺灣時間:2021/05/09 23:42
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果

詳目顯示:::

我願授權國圖
: 
twitterline
研究生:王吉生
研究生(外文):Jyi-Sheng Wang
論文名稱:中斷監督器:運用於中央處理器限制程序之即時排程輔助器
論文名稱(外文):Interrupt Monitor: A Real-Time Scheduling Aid for CPU-Bound Processes
指導教授:薛 智文
指導教授(外文):Chih-Wen Hsueh
學位類別:碩士
校院名稱:國立中正大學
系所名稱:資訊工程研究所
學門:工程學門
學類:電資工程學類
論文種類:學術論文
論文出版年:2001
畢業學年度:89
語文別:英文
論文頁數:33
中文關鍵詞:中斷監督器即時排程驗證虛擬裝置驅動程式核心模組
外文關鍵詞:Interrupt Monitorreal-time schedulingpseudo device driverkernel module
相關次數:
  • 被引用被引用:0
  • 點閱點閱:135
  • 評分評分:系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
可預測性對於即時系統是一項很重要的特質。理想上,在設計系統時,即使在負載很高的狀況下,系統也要能夠符合其所規定的時間限制。然而在一般的作業系統環境下,時常會有突發的事件(如硬體中斷)使得可預測性很難達成。當一個正在執行的工作被硬體中斷搶先,中央處理器會儲存部份或全部正在執行工作的本文,然後執行中斷服務程式。因為在硬體中斷完成後,被阻斷的工作可能會重新進入運行佇列,而排程器會從中挑選新的工作去執行,所以中央處理器的控制權未必會回到之前被中斷的工作,因此工作時序行為的預測會更加困難。如果能夠偵測到中斷對工作的影響,設計可預測性系統的發展人員可簡單地定位出由硬體中斷所引起的不正常行為,亦可容易地驗證和改進排程的結果。
為了使這個問題可以更容易解決,我們使用一個即時監督器作為排程輔助工具以驗證中央處理器限制程序系統的排程結果。這個即時監督器稱為中斷監督器,是一個虛擬裝置驅動程式。它提供介面給使用者以存取核心中的資料結構及事件,包含偵查並找出由硬體中斷所引發之不正常行為的機制﹔除此之外,亦可應用於測量系統效能諸如系統呼叫的延遲時間和程序執行時間。為了減少與機器相關的因素,本論文只針對硬體中斷對中央處理器限制程序的影響。我們使用中斷監督器去測量硬體中斷對中央處理器限制程序的影響,同時也測量中斷監督器對系統所帶來的負荷並展示其對於作業系統是常駐且透明(不被使用者察覺)的一部分,幾乎不影響系統效能且不干擾系統本身的執行。

One of the most important properties for a real-time system is to be predictable. Ideally, it should be determined at design time that the system can meet all of its timing constrains, even at peak load. In general operating systems, however, unexpected events make predictable very difficult, for example, interrupts. When a running task is preempted by an interrupt, CPU saves part or all of its context and jumps to an interrupt service routine. After an interrupt is served, the control of CPU does not necessarily return to the same task because a blocked task may enter the ready queue again and the scheduler will select the new ready task. Therefore, it is more difficult to predict the timing behavior of the tasks. If the impact from interrupts on tasks can be detected, system designers who design a predictable system can simply locate abnormal behavior caused by interrupts and the scheduling results also can be verified and improved easily.
To alleviate the problem, we use a real-time monitor as the
scheduling aid to verify the scheduling results of a system with
CPU-Bound processes. The real-time monitor, called Interrupt
Monitor, is a pseudo device driver, which provides an interface
for a user process to access the data structures or events
occurring in the low level kernel layer and contains a mechanism
to detect and locate abnormal behavior caused by interrupts.
Besides, it can be used to measure some system performance such as system call latency, and process computation time. To reduce
machine dependent factors, the paper focuses on CPU-Bound
processes. Using the monitor, we measure the impact from
interrupts on a CPU-Bound process. We also measure the overhead
which Interrupt Monitor adds to a system and show that the monitor can be a permanent and transparent part of an operation system, which degrades little system performance, and hardly interfere with the execution of a host system.

1.Introduction
2.Background
2.1 Kernel Module
2.2 Data Stream Kernel Interface
2.3 Real-Time Hardware Abstraction Layer
3.Implementation of Interrupt Monitor
3.1 Overview of Interrupt Monitor
3.2 MyHAL
3.3 Communication System Call
3.4 Checker
4 The Overhead of Interrupt Monitor
4.1 The Overhead of Intercepting
4.2 The Overhead of Invoking System Call
4.3 The Overhead of Triggering Port
5. Experiment
5.1 CPU-Bound Process
5.2 CPU-Bound Process Invoking System Calls
6 Conclusion and Future Work

[1] Z. Deng anf J.W.-S Liu. Scheduling real-time application in open system enviroment. In Proc IEEE Real-Time System Symposium, December 1997.
[2] K. Arvind. A new probabilistic alogorithm for clock synchronization. In Proc. IEEE Real-Time Systems Symp, December 1989.
[3] Sarah E. Chodrow and Farnam Jahanian. Run-time monitoring of real-time systems. Proc. IEEE Real-Time Systems Symp, 1991.
[4] C.Kilpatrick, K.Schwan, and D.Ogle. Using languages for capture, analysis and display of performance information for parallel or distrubuted systems. In International Conference on Comp. Language '90, March 1990.
[5] D.Haban and K.Shin. Application of real-time monitoring to scheduling tasks with random execution times. In IEEE TSE, December 1989.
[6] D.Haban and D. Wybranietz. A hybrid monitor for behavior and performance analysis of distubuted systems. In IEEE TSE, February 1990.
[7] M. Donner and F. Jahanian. Rtl meets ore. In 7th IEEE Workshop on Real-Time Operating Systems and Software, May 1990.
[8] Marc D. Donner and David H. Jameson. Language and operating system features for real-time programming. In Computering Systems, 1988.
[9] B. Srinivasan et al. Afirm real-time system implementation using commercial on-the shelf hardware and free software. In Proc. IEEE Real-Time Technology and Application Symposium, June 1998.
[10] Data Stream Kernel Interface. http://www.ittc.ukans.edu/datastream/. 2000.
[11] F. Jahanian and A. Goyal. A formalism for monitoring real-time constrains at run-time. June 1990.
[12] Farnam Jahnian and Aloysius Ka-Lau Mok. Safety analysis of timing properties in real-time systems. In IEEE Transactions on Software Engineering, 1986.
[13] Strosnider J.K Katcher D.I., Arakawa H. Engineering and analysis offixed priority schedulers. In IEEE Trans. on Software Eng, 1993.
[14] C. L. Liu and J. Layland. Scheduling algorithms for multiprogramming in a hard real-time environment. Journal of the ACM, 10(1):46{61, 1973.
[15] RTAI. http://www.rtai.org. 2000.
[16] R. Snodgrass. A relational approach to monitoring complex systems. In ACM Tranctions on Computer Systems, May 1988.
[17] J.P. Tasi, K-Y Fang, and H-Y Chen. A noninvasive architecture to monitor real-time distributed systems. In Computer, March 1990.
[18] H. Tokuda, M. Koreta, and C.W. Mercer. A real-time monitor for distributed real-time o.s. In ACM SIGPlan Notices, January 1989.
[19] H. Tokuda, M. Koreta, and C.W. Mercer. A integrated approach to monitoring and scheduling in real-time systems. In IEEE Workshop on Real-Time Operating Systems and Software, May 1991.
[20] Yu-Chung Wang and Kwei-Jay Lin. Enchancing the real-time capability of linux kernel. In Proc. of 5th RTCSA '98, October 1999.
[21] Yu-Chung Wang and Kwei-Jay Lin. Implementing a general real-time scheduling framework in the red-linux real-time kernel. In Proc. of the 20th IEEE Real-Time Systems Symp, 1999.

QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top
系統版面圖檔 系統版面圖檔