Dynamic_Data_Race_Detection.pdf

发布时间：2022-06-12 发布人：admin 分类：说明书资料大小：0.37M 资料格式：pdf 举报版权申诉

57f9cb96-a624-4674-a12e-9639984d027f.pdf-第1页.png

第1页 / 共27页

57f9cb96-a624-4674-a12e-9639984d027f.pdf-第2页.png

第2页 / 共27页

57f9cb96-a624-4674-a12e-9639984d027f.pdf-第3页.png

第3页 / 共27页

57f9cb96-a624-4674-a12e-9639984d027f.pdf-第4页.png

第4页 / 共27页

57f9cb96-a624-4674-a12e-9639984d027f.pdf-第5页.png

第5页 / 共27页

57f9cb96-a624-4674-a12e-9639984d027f.pdf-第6页.png

第6页 / 共27页

57f9cb96-a624-4674-a12e-9639984d027f.pdf-第7页.png

第7页 / 共27页

57f9cb96-a624-4674-a12e-9639984d027f.pdf-第8页.png

第8页 / 共27页

文本预览

Dynamic Data Race Detection Based Pure Locksets 杨振 2015/4/6

Data Race • “A data race occurs when two concurrent threads access a shared variable and … – at least one access is a write and – the threads use no explicit mechanism to prevent the accesses from being simultaneous” • In other words, a data race can lead to a potential violation of mutual exclusion.

Data Race Example: Threads with unsynchronized access to a shared array Thread 1 int i; … Thread 2 int i; … for (i = 1; for (i = 1; i < MAX; i < MAX; i++) i++) { { cin >> x; if (A[i] < B[i]) A[i] = 2*x; B[i] = A[i]; } … } …

Approaches to Data Race Detection • Static: performed at compile time or earlier • Dynamic: runtime analysis • Most current researches focus on dynamic detection • The environment is a single multithreaded program rather than separate programs.

Static Data Race Detection • Type-based analysis – Language type system augmented to “express common synchronization relationships”: correct typing→no data races – Difficult to do & restricts the type of synchronization primitives • Language features – e.g., use of monitors – Only works for static data – not dynamic data • Path analysis; – Doesn’t scale well – Too many false positives

Dynamic Data Race Detection • Dynamic detection: monitor programs during execution and look for problems – The program may be “instrumented” with additional instructions – The additions don’t change program functionality but are used to monitor conditions of interest - in this case, access to shared variables and synchronization operations.

Dynamic Detection • Problems: – Can only check paths that are actually executed – Adds overhead at runtime • Techniques – Happens-before – Lockset analysis ( Eraser ) – Hybrids of happen-before and lockset

Basic Idea of Eraser • Observe all instances where a shared variable is accessed by a thread. • If there is a chance that a data race can occur, be sure the shared variable is protected by a lock. – Simple algorithm – basic locks – Advanced algorithm – reader/writer locks • If variable isn’t protected, issue a warning.

分享到：

赞收藏

资料库

Dynamic_Data_Race_Detection.pdf

相关推荐

课程资源

热门标签

最新资料