| 词条 | Readers–writers problem |
| 释义 |
}} In computer science, the readers-writers problems are examples of a common computing problem in concurrency. There are at least three variations of the problems, which deal with situations in which many threads (small processes which share data) try to access the same shared resource at one time. Some threads may read and some may write, with the constraint that no process may access the shared resource for either reading or writing while another process is in the act of writing to it. (In particular, we want to prevent more than one thread modify the shared resource simultaneity and allowed for two or more readers to access the shared resource at the same time). A readers-writer lock is a data structure that solves one or more of the readers-writers problems. The basic reader-writers problem was first formulated and solved by Courtois et al.{{r|courtois}}[1] For ex: Suppose in a class there is one board, where only one person can write at a time, but at the same time many readers can read the board First readers-writers problemSuppose we have a shared memory area (critical section) with the basic constraints detailed above. It is possible to protect the shared data behind a mutual exclusion mutex, in which case no two threads can access the data at the same time. However, this solution is sub-optimal, because it is possible that a reader R1 might have the lock, and then another reader R2 requests access. It would be foolish for R2 to wait until R1 was done before starting its own read operation; instead, R2 should be allowed to read the resource alongside R1 because reads don't modify data, so concurrent reads are safe. This is the motivation for the first readers-writers problem, in which the constraint is added that no reader shall be kept waiting if the share is currently opened for reading. This is also called readers-preference, with its solution: semaphore resource=1; semaphore rmutex=1; readcount=0; /* resource.P() is equivalent to wait(resource) resource.V() is equivalent to signal(resource) rmutex.P() is equivalent to wait(rmutex) rmutex.V() is equivalent to signal(rmutex)
} reader() { readcount++; //Indicate that you are a reader trying to enter the Critical Section if (readcount == 1) //Checks if you are the first reader trying to enter CS resource.P(); //If you are the first reader, lock the resource from writers. Resource stays reserved for subsequent readers readcount--; //Indicate that you are no longer needing the shared resource. One fewer reader if (readcount == 0) //Checks if you are the last (only) reader who is reading the shared file resource.V(); //If you are last reader, then you can unlock the resource. This makes it available to writers. } In this solution of the readers/writers problem, the first reader must lock the resource (shared file) if such is available. Once the file is locked from writers, it may be used by many subsequent readers without having them to re-lock it again. Before entering the critical section, every new reader must go through the entry section. However, there may only be a single reader in the entry section at a time. This is done to avoid race conditions (in computer-science race condition is a condition which two or more threads is waking up simultaneity and try to enter the critical section, in this case the behavior is unexpected) on the readers (e.g. two readers increment the readcount at the same time, and both try to lock the resource, causing one reader to block). To accomplish this, every reader which enters the Once the first reader is in the entry section, it will lock the resource. Doing this will prevent any writers from accessing it. Subsequent readers can just utilize the locked (from writers) resource. The very last reader (indicated by the readcount variable) must unlock the resource, thus making it available to writers. In this solution, every writer must claim the resource individually. This means that a stream of readers can subsequently lock all potential writers out and starve them. This is so, because after the first reader locks the resource, no writer can lock it, before it gets released. And it will only be released by the very last reader. Hence, this solution does not satisfy fairness. Second readers-writers problemThe first solution is suboptimal, because it is possible that a reader R1 might have the lock, a writer W be waiting for the lock, and then a reader R2 requests access. It would be unfair for R2 to jump in immediately, ahead of W; if that happened often enough, W would starve. Instead, W should start as soon as possible. This is the motivation for the second readers-writers problem, in which the constraint is added that no writer, once added to the queue, shall be kept waiting longer than absolutely necessary. This is also called writers-preference. A solution to the writers-preference scenario is:[2] int readcount, writecount; //(initial value = 0) semaphore rmutex, wmutex, readTry, resource; //(initial value = 1) //READER reader() { readTry.P(); //Indicate a reader is trying to enter rmutex.P(); //lock entry section to avoid race condition with other readers readcount++; //report yourself as a reader if (readcount == 1) //checks if you are first reader resource.P(); //if you are first reader, lock the resource rmutex.V(); //release entry section for other readers readTry.V(); //indicate you are done trying to access the resource //reading is performed rmutex.P(); //reserve exit section - avoids race condition with readers readcount--; //indicate you're leaving if (readcount == 0) //checks if you are last reader leaving resource.V(); //if last, you must release the locked resource rmutex.V(); //release exit section for other readers } //WRITER writer() { wmutex.P(); //reserve entry section for writers - avoids race conditions writecount++; //report yourself as a writer entering if (writecount == 1) //checks if you're first writer readTry.P(); //if you're first, then you must lock the readers out. Prevent them from trying to enter CS wmutex.V(); //release entry section resource.P(); //reserve the resource for yourself - prevents other writers from simultaneously editing the shared resource //writing is performed resource.V(); //release file wmutex.P(); //reserve exit section writecount--; //indicate you're leaving if (writecount == 0) //checks if you're the last writer readTry.V(); //if you're last writer, you must unlock the readers. Allows them to try enter CS for reading wmutex.V(); //release exit section } In this solution, preference is given to the writers. This is accomplished by forcing every reader to lock and release the readtry semaphore individually. The writers on the other hand don't need to lock it individually. Only the first writer will lock the readtry and then all subsequent writers can simply use the resource as it gets freed by the previous writer. The very last writer must release the readtry semaphore, thus opening the gate for readers to try reading. No reader can engage in the entry section if the readtry semaphore has been set by a writer previously. The reader must wait for the last writer to unlock the resource and readtry semaphores. On the other hand, if a particular reader has locked the readtry semaphore, this will indicate to any potential concurrent writer that there is a reader in the entry section. So the writer will wait for the reader to release the readtry and then the writer will immediately lock it for itself and all subsequent writers. However, the writer will not be able to access the resource until the current reader has released the resource, which only occurs after the reader is finished with the resource in the critical section. The resource semaphore can be locked by both the writer and the reader in their entry section. They are only able to do so after first locking the readtry semaphore, which can only be done by one of them at a time. If there are no writers wishing to get to the resource, as indicated to the reader by the status of the readtry semaphore, then the readers will not lock the resource. This is done to allow a writer to immediately take control over the resource as soon as the current reader is finished reading. Otherwise, the writer would need to wait for a queue of readers to be done before the last one can unlock the readtry semaphore. As soon as a writer shows up, it will try to set the readtry and hang up there waiting for the current reader to release the readtry. It will then take control over the resource as soon as the current reader is done reading and lock all future readers out. All subsequent readers will hang up at the readtry semaphore waiting for the writers to be finished with the resource and to open the gate by releasing readtry. The rmutex and wmutex are used in exactly the same way as in the first solution. Their sole purpose is to avoid race conditions on the readers and writers while they are in their entry or exit sections. Third readers-writers problemIn fact, the solutions implied by both problem statements can result in starvation — the first one may starve writers in the queue, and the second one may starve readers. Therefore, the third readers-writers problem is sometimes proposed, which adds the constraint that no thread shall be allowed to starve; that is, the operation of obtaining a lock on the shared data will always terminate in a bounded amount of time. A solution with fairness for both readers and writers might be as follows: int readCount; // init to 0; number of readers currently accessing resource // all semaphores initialised to 1 Semaphore resourceAccess; // controls access (read/write) to the resource Semaphore readCountAccess; // for syncing changes to shared variable readCount Semaphore serviceQueue; // FAIRNESS: preserves ordering of requests (signaling must be FIFO) void writer() {serviceQueue.P(); // wait in line to be serviced // } void reader() {serviceQueue.P(); // wait in line to be serviced readCountAccess.P(); // request exclusive access to readCount // } Note also that this solution can only satisfy the condition that "no thread shall be allowed to starve" if and only if semaphores preserve first-in first-out ordering when blocking and releasing threads. Otherwise, a blocked writer, for example, may remain blocked indefinitely with a cycle of other writers decrementing the semaphore before it can. Simplest Reader Writer ProblemThe simplest reader writer problem which uses only two semaphores and doesn't need an array of readers to read the data in buffer. Please notice that this solution gets simpler than the general case because it is made equivalent to the Bounded buffer problem, and therefore only N readers are allowed to enter in parallel, N being the size of the buffer. Readerdo { wait(read) ............ reading data ............ signal(write) } while(TRUE); Writerdo { wait(write) ............. writing data ............. signal(read) } while(TRUE); Algorithm1. Reader will run after Writer because of read semaphore. 2. Writer will stop writing when the write semaphore has reached 0. 3. Reader will stop reading when the read semaphore has reached 0. In writer, the value of write semaphore is given to read semaphore and in reader, the value of read is given to write on completion of the loop. See also
References1. ^{{Cite book |title=Synchronization Algorithms and Concurrent Programming |first=Gadi |last=Taubenfeld |publisher=Pearson Education |year=2006 |page=301}} 2. ^{{cite journal |journal=Communications of the ACM |title=Concurrent Control with "Readers" and "Writers" |first1=P. J. |last1=Courtois |first2=F. |last2=Heymans |first3=D. L. |last3=Parnas |year=1971 |url=http://cs.nyu.edu/~lerner/spring10/MCP-S10-Read04-ReadersWriters.pdf}}
External links
1 : Concurrency (computer science) |
| 随便看 |
|
开放百科全书收录14589846条英语、德语、日语等多语种百科知识，基本涵盖了大多数领域的百科知识，是一部内容自由、开放的电子版国际百科全书。