Reader Writer Problem Solution using semaphore
Duration: 12 min
This video lesson is available to enrolled students.
AI Summary
An AI-generated summary of this video lecture.
The video provides a detailed lecture on solving the Readers-Writers synchronization problem using semaphores. The instructor defines shared variables (mutex, wrt, readcount) and constructs the logic for Writer and Reader processes. He explains the necessity of protecting the readcount variable with a mutex and using the wrt semaphore to block writers when readers are active. The lecture culminates in a complete code structure that implements a readers-preference solution, ensuring writers wait if readers are present.
Chapters
0:00 – 2:00 00:00-02:00
The video begins with the instructor introducing the 'Solution using Semaphores' for the Readers-Writers problem. On the slide, he lists the shared data structures: semaphore mutex = 1, semaphore wrt = 1, and int readcount = 0. He explains that these variables are shared among all reader processes. He clarifies the role of mutex as a binary semaphore used to protect the readcount variable from race conditions, ensuring that only one process updates the count at a time. He also defines wrt as a semaphore used to manage access to the shared resource, specifically to block writers when readers are active. He points to the text on the screen to emphasize the initialization values.
2:00 – 5:00 02:00-05:00
The instructor moves to the whiteboard to outline the logic for the Writer process. He writes Wait(wrt) before the Critical Section (CS) for writing and Signal(wrt) after. He explains that this simple structure ensures that only one writer can access the resource at a time and that writers are blocked if readers are currently reading. He then shifts focus to the Reader process, writing Readcount++ to increment the count when a reader enters and Readcount-- when it leaves. He emphasizes that these operations must be atomic to prevent data corruption and that a separate mechanism is needed to protect the counter.
5:00 – 10:00 05:00-10:00
The instructor refines the Reader process logic by adding synchronization primitives. He writes Wait(mutex) before Readcount++ and Signal(mutex) after, ensuring the counter update is safe. He then introduces the critical logic for the first and last readers. He writes If(readcount == 1) followed by wait(wrt), explaining that the first reader must block writers. Conversely, he writes If(readcount == 0) followed by signal(wrt), explaining that the last reader must release the resource for waiting writers. He draws arrows to show the flow of execution and the conditions under which writers are blocked or unblocked. He also writes R1 and R2 to denote different readers.
10:00 – 12:11 10:00-12:11
The instructor completes the solution by filling in the full table for both Writer and Reader processes. He writes the complete code sequence: Writer waits on wrt, enters CS, signals wrt. Reader waits on mutex, increments readcount, checks if it is the first reader (waits on wrt), signals mutex, enters CS, waits on mutex, decrements readcount, checks if it is the last reader (signals wrt), and signals mutex. He reviews the entire flow, pointing out how the wrt semaphore acts as a gatekeeper for writers, ensuring they wait if any readers are present, thus implementing a readers-preference solution. He highlights the specific lines of code on the board to reinforce the logic.
The lecture systematically builds a solution to the Readers-Writers problem. It starts with defining shared resources, then constructs the Writer process to ensure exclusive access. It then develops the Reader process, introducing a counter (readcount) protected by a mutex. The core logic involves the first reader blocking writers and the last reader releasing them. The final output is a complete, synchronized code structure that prioritizes readers.