L-3.1: Process Synchronization | Process Types | Race Condition | Operating System-1

Main Ideas

Process synchronization coordinates processes that run concurrently and share data or resources. Without coordination, their operations can interfere and produce inconsistent results.

In serial execution, processes run one at a time: the next process starts after the current one finishes. The lecture explains that this avoids the interference shown in the example.

In time-sharing or multiprogrammed systems, processes may take turns using the CPU. Even on a single-CPU system, preemption and context switching can interleave their instructions.

Process Types

• Cooperative processes: One process’s execution can affect another because they share something, such as a variable, memory, buffer, code, or hardware resource.
• Independent processes: One process’s execution does not affect another because they share nothing relevant. The lecturer gives transactions handled by separate banks’ servers as an example.

Synchronization is especially important for cooperative processes, because the order in which they execute can affect shared data or resources.

Race-Condition Example

The lecture demonstrates how two processes can update the same shared variable:

• The shared variable starts at 5.
• Process P1 reads the value into a local variable, increments it, pauses, and later writes it back.
• Process P2 reads the value into another local variable, decrements it, pauses, and later writes it back.
• During a pause, the operating system can preempt one process and run the other. When a context switch occurs, the process’s state is saved in its PCB (Process Control Block), allowing it to resume later.

One possible interleaving is:

1. P1 reads 5 and increments its local copy to 6.
2. P1 pauses. P2 reads the still-shared value 5 and decrements its copy to 4.
3. P2 pauses. P1 resumes and writes 6 to the shared variable.
4. P2 resumes and writes 4, leaving the final value 4.

If the operations were completed in the opposite order, the final value could be 6. The expected result of incrementing and decrementing 5 is 5, so either outcome is incorrect.

This is a race condition: the final result depends on the timing and order in which concurrent processes access and update shared data.

Key Lesson

Shared operations need to be synchronized so concurrent processes do not overwrite one another’s work. The lecturer names semaphores, Peterson’s solution, and locking as methods to be covered later.

Speakers and Sources

• Speaker: A lecturer or instructor from the Gate Smashers video. The subtitles do not identify the speaker by name.
• Source: The lecturer’s spoken explanation and examples.