Lecture No. 02
Dated: 14-03-2025
Single User Systems
These systems are designed to provide most convenience to a single user instead of focusing over maximizing utility of the CPU and other peripheral devices.
Batch Systems
Early computers were controlled by consoles with card readers. These computers would output data on punch cards and also take in punch cards as inputs, which describe the "job" which needs to be performed. These inputs were handed over to computer operators who would feed it to the computer but due to slow mechanical I/O movements, a lot of CPU time was wasted.
In these systems, users would not be able to interact with their jobs and jobs with similar needs were executed in a "batch".
Example
Digital Equipment Corporation’s VMS is an example of a batch operating system.
Memory partitioned into user and system spaces
Multi Programmed Systems
There's a job pool which contains all the jobs ready to be executed but only a fraction of those can be loaded in the memory, at the same time.
The operating system1 arranges them so that CPU executes one job at a time.
This job might require an I/O operation and instead of waiting, CPU switches to the other job execution.
Memory layout for a multi-programmed batch system
Imagine two processes \(P_1\) and \(P_2\).
\(P_1\) completes its CPU burst (unit of time), starts to wait for an event (maybe an I/O operation) to happen and in the meantime, the CPU switches to \(P_2\) and vise versa happens.
Illustration of the multiprogramming concept. Both processes take 6 units instead of 10 to be completed.
Job Pool
The processes which are in the disk and are waiting for memory allocation for the sake of execution, are placed in job pool.
Job Scheduling
The system decides which job to bring in memory if there are competing jobs, through a process called job scheduling.
Time Sharing Systems
Such system is multi-user, multi-process, and interactive, allowing multiple users to run and interact with their multiple jobs at the same time on the computer.
This uses Multi programming and CPU scheduling.
Example
UNIXLinuxWindows NTWindows 2000
Real time Systems
These systems have very tight time constraints and if are failed to be delivered, the system fails.
Example
- Scientific experiments
- Medical imaging
- Industrial control systems
- Display systems
Types
There are 2 types
Hard
These are very time critical and lean less on operating systems and data retrieval from disk etc.
Soft
In these systems, the time critical tasks take the priority and retain that priority until it is finished.
These systems can be mixed with other types of systems unlike hard real time systems because those conflict with operating system kernels of other systems such as time sharing systems and hence cannot be mixed.
Interrupts
It is a signal generated by a hardware device such as an I/O device to get CPU's attention.
These signals transfer the control to Interrupt Service Routine(ISR) generally through the interrupt vector table which contains the addresses of all the service routines.
The address of the interrupted instruction is saved, which is later used to come back to that instruction after the interrupt has been addressed.
During interrupt processing, incoming interrupts are disabled to prevent lost interrupts.
An operating system1 is an interrupt driven software.
Trap
It is also called an exception which is generated by the software as a result of an error such as division by zero or rejection to a user request from the operating system1 service.
Signal
A signal is generated to get attention of a process.
Example
When you press Ctrl + C while a process is running, it sends an interrupt signal (SIGINT).
Possible Actions
Kerneldefined default action which usually terminates theprocessbut sometimes can also generate acorefile which allows the user to know the state of theprocessat termination.- Program can intercept the
signaland ignore it. - Program can intercept the
signaland perform a programmer defined action.
Hardware Protection
Multiprogramming improves system utilization by running multiple programs in memory, but it introduces risks like faulty programs affecting others, overwriting data, or monopolizing the CPU.
This leads to key hardware protection issues
I/OmemoryCPU protection
Dual Mode Operation
To protect the OS1 and programs from faulty processes, shared resources require protection. Modern CPUs support this via privileged and non-privileged instructions. Privileged instructions (e.g., I/O operations) are restricted to the OS.1 If a user process executes them, a trap is triggered and the process is terminated.
To enforce this, CPUs operate in two modes: user mode and monitor (or system/supervisor) mode, distinguished by a mode bit (0 = monitor, 1 = user). This allows the CPU to identify whether code is OS1 or user-level.
System calls let user processes safely request OS-level actions. When a system call, interrupt, or trap occurs, the CPU switches to monitor mode to run kernel code, then switches back to user mode when returning control.
The dual-mode operation of the CPU
I/O Protection
To prevent user processes from disrupting the system, mechanisms are needed to block illegal I/O, memory access, and CPU misuse.
All I/O instructions are privileged, so user processes must access I/O via system calls. To ensure full I/O protection, user programs must never gain monitor mode control.
During execution, the CPU switches to monitor mode on traps or interrupts and jumps to addresses in the interrupt vector. If a user process could modify this vector, it could redirect control to itself in monitor mode, compromising security. Therefore, I/O and system memory modifications must be restricted to monitor mode only.