R. Wahbe. Efficient data breakpoints. In Proc. of 5th Intl. Conf. on Architectural Support for Programming Languages and Operating Systems, pages 200--212, Oct. 1992.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....suggest changing a machine s microcode to check certain conditions when writing a memory word. This is similar to modifying the memory controller to enforce protection, as are Johnson s and Wahbe s suggestions for various hardware mechanisms to trap the updates of certain memory locations [36] [70]. Hive uses the Flash firewall to protect memory against wild writes by other processors in a multiprocessor [14] Hive preemptively discards pages that are writable by failed processors, an option not available when storing permanent data in memory. Object code modification has been suggested as ....

....memory against wild writes by other processors in a multiprocessor [14] Hive preemptively discards pages that are writable by failed processors, an option not available when storing permanent data in memory. Object code modification has been suggested as a way to provide data breakpoints [40] [70] and fault isolation between software modules [71] One can also isolate memory from the operating system by hiding the memory [19] via a device driver interface or by physically isolating the memory via a hardware interface. The former approach requires user applications and kernel to access ....

R. Wahbe, "Efficient Data Breakpoints," Proc. 1992.

....MMP could catch a program s attempt to jump into writable data. It could also catch the program trying to write off the end of a piece of memory. Bounds checking is useful for program debugging and if implemented by MMP would be available to the kernel. A related functionality, data watchpoints [33], can be easily implemented with our fine grained protection. A data watchpoint generates a trap when a given word in memory is modified. Some processors support a handful of watched memory locations [15, 14] but our fine grained protection scales to thousands of individually protected words. ....

R. Wahbe. Efficient data breakpoints. In Proceedings of ASPLOS-V, Oct 1992.

....ffl code patching replacing each store and or load instructions with an inline check or call to a function that gives control to the debugger if the accessed location is being watched, and subsequently executes or emulates the instruction. The range of implementations is discussed in Wahbe [Wah92], who compared simulations of each type of watchpoint implementation. According to his simulations, code patching is the most efficient method of implementing watchpoints in software. We implemented watchpoints via code patching on the Titan, a Mips R3000 based multiprocessor machine sold by ....

R. Wahbe, "Efficient Data Breakpoints", Proceedings of the Fifth International Conference on Architectural Support for Programming Languages and Systems, SIGPLAN Notices, Vol. 27, No. 9, pp. 200-212, September 1992.

.... Moss 93, Appel Li 91] The majority of these studies have concentrated on the use of write detection to implement efficient garbage collection in object oriented and type safe languages. Recently, software write detection has been investigated for other purposes, such as debugger breakpoints [Wahbe 92] and fault isolation [Wahbe et al. 93] Our experiments, which are based on parallel programs running on a DSM system, differ from previous work in the choice of language used, whether or not the language is compiled or interpreted, style of programming (parallel vs. sequential) the existence of ....

Wahbe, R. Efficient Data Breakpoints. In Proceedings of the Fifth ACM Conference on Architectural Support for Programming Languages and Operating Systems, pages 200--212, September 1992.

....the executor is activated by continue. The director detects a breakpoint by calling waitbreak. A variation of breakpoints are watchpoints which report access to specified variables. Because watchpoints can impose significant time penalty without special hardware support or program modifications [29], there are currently no constructs in the directing interface to support them. Tracing, described next, can provide equivalent functionality. 3.4 Tracing Directors perform tracing by reading events which describe the executor s behavior. Events are generated by the tracing mechanism in the ....

R. Wahbe. Efficient data breakpoints. In Proc. of the 5nd International Conference on Architectural Support for Programming Languages and Operating Systems, pages 200--212. ACM, 1992.

....leisurely, or errors that require the operating system to terminate the program. However, applications and run time systems are using exception mechanisms increasingly and in novel ways; for example, exceptions are being used for garbage collection [Appel et al. 88] conditional watchpoints [Wahbe 92] transaction support [Chang Mergen 88] persistent storage management [Wilson Kakkad 92] and distributed virtual memory [Li Hudak 89] To compound the problem, several studies y Author s current address: DEC Systems Research Center, 130 Lytton Avenue, Palo Alto, CA 94301. This work was ....

R. Wahbe. Efficient data breakpoints. In Proceedings of the 5th International Conference on Architectural Support for Programming Languages and Operating Systems, pages 200--212, October 1992.

.... hardware support for debugging [Miy88] Virtual memory based approaches protect all pages indicated by the debugger and detect when programs attempts to write to those pages [Bea83, Sul91] Finally, software based methods modify the program to check the target location of all write instructions [Kes90, Has92, Wah92, Wah93]. Currently the major emphasis in the software compilation debugging domain is on integration between optimization technique based on transformations or aggressive scheduling techniques such as speculative scheduling and debugging process [Cou88, Coh91, Mil91, Bro92, Shu93] Numerous controllers ....

R. Wahbe, "Efficient Data Breakpoint", ACM SIGPLAN Notices, Vol. 27, No. 9, pp. 200-212, 1992.

....and update the master (recoverable) object state all at once. This approach strikes us as difficult to program correctly, and runs the risk of costing more in execution time than it saves. Another application level approach would be to extend the sandboxing techniques describedby Wahbe et al. [17, 18] to log write operations. Sandboxingwould eliminate the burden on the application programmer of specifying individual modifications to be logged. However, it would increase the execution overheadbecause it would introduce logging code for nearly every write access, not just those that the ....

R. Wahbe. Efficient data breakpoints. In ASPLOS. ACM, October 1992.

....thesis shows that dramatic improvements in the performance of watchpoints can be made in a variety of ways. This thesis also compares these solutions for efficient watchpoints with each other. Having an efficient implementation of watchpoints would help programmers immensely in debugging programs [24, 17, 11]. Without watchpoints it is difficult to find problems with pointer uses that result in inadvertent modification of unrelated data. For example, if there is a bug in the program that incorrectly manipulates a pointer so that it points to memory associated with a different data structure and then ....

....up watchpoints. One is to reduce the number of times the check is made to see if the expression has changed. The other is to make the check faster. There are three ways that have been proposed to reduce the number of checks. The first is to check only write instructions instead of all instructions [24, 17]; the second is to check only the write instructions that write to the same page as the watched memory [24, 17] the third is to checkpoint [17] that is, periodically stop and check the condition with the ability to roll back 4 to a previous state if the stopping condition has been met. There ....

[Article contains additional citation context not shown here]

Robert Wahbe. Efficient data breakpoints. In Proceedings of the Fifth International Symposium on Architectural Support for Programming Languages and Operating Systems (ASPLOS-V), pages 200--212, October 1992.

....enough. For watching an array of several Mbytes, the debugger would have to maintain a copy of this array in order to be able to compare old and new values at each step, involving intolerable memory and computing time overheads. More efficient and general methods are proposed and compared in [Wah92]. They fall into three general categories: 1) specialized hardware support, 2) operating system support, and (3) other software based approaches. Hardware based data breakpoints are by far most efficient, but lack generality. Often only a few words of memory can be watched, and not all ....

R. Wahbe. Efficient data breakpoints. In Proceedings of the Fifth International Conference on Architectural Support for Programming Languages and Operating Systems, pages 200--212, Boston, Massachusetts, 1992.

....of heap objects. To this date, no performance information has been reported for Magpie or for VAX DEBUG. Several authors have speculated that efficient data breakpoints require special purpose hardware [4, 11, 15] To quantify the differences among data breakpoint implementation strategies, Wahbe [19] compared facilities based on specialized processor support, virtual memory page protection, checking the destination address of machine instructions via an operating system trap, and checking the destination address of machine instructions via a procedure call. Virtual memory and the use of traps ....

....call this operation address lookup. We found that the best strategy for implementing efficient address lookup is to mini mize the average number of memory accesses required. The write checks tested in Wahbe s pilot study of data breakpoint implementations used a hash table for address lookup [19]. This data structure uses memory efficiently, consuming space proportional to the number of monitored regions. However, it requires several memory accesses for each address lookup. We tested this data structure using the spec benchmarks and verified that the write check overhead generally matched ....

R. Wahbe. "Efficient Data Breakpoints,". In Fifth International Conference on Architectural Support for Programming Languages and Operating Systems, pages 200--212, October 1992. Appeared as SIGPLAN Notices 27(9).

No context found.

R. Wahbe. Efficient data breakpoints. In Proc. of 5th Intl. Conf. on Architectural Support for Programming Languages and Operating Systems, pages 200--212, Oct. 1992.

No context found.

R. Wahbe. Efficient data breakpoints. In Proceedings of the fifth International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS-V), 1992.

No context found.

R. Wahbe. Efficient Data Breakpoints. In Fifth International Conference on Architectural Support for Programming Languages and Operating Systems, October 1992, pp. 200-212.

No context found.

R. Wahbe, `Efficient data breakpoints', Proc. of the 5nd International Conference on Architectural Support for Programming Languages and Operating Systems. ACM, 1992, pp. 200--212.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC