Operating Systems Flashcards

Question

What are some of the common features of Modern Desktop Operating Systems

Answer 1

Modern Operating Systems are more reliable, this is because 'unreliable' interfaces can run in user mode and normally implement preemptive multitasking. Rare Failures for modern systems are: Hardware Failure CPU stuck in device driver Fatal exception in device driver/kernel

Answer 2

Virtual memory can be defined as a technique that uses main memory as a 'cache' for secondary storage. Or a feature of an OS, which allows the system to compensate for shortages in physical memory by temporarily transferring pages of data from RAM to disk storage. It allows a program to access memory via a virtual address

Answer 3

A virtual address can be translated into a physical address and is used to access main memory. It allows for efficient sharing of memory between programs, although one must keep in mind how this is structured. It removes the programming burdens of limited memory

Answer 4

A page is a contiguous block of virtual memory, contiguous meaning solid block of data.

Answer 5

A page fault is when an accessed page is not found in the main memory. This means the page must be retrieved from the disk which is more than 1000x slower than retrieving from main memory. This severely slows down program execution speed.

Answer 6

A super loop is an alternative to OS. It is simply a loop in which other tasks are run. The loop is not meant to end.

Answer 7

Simple, easy to build, debug, test and maintain Efficient as minimal hardware requirement Highly portable, can easily be implemented on many kinds of devices. However, Not accurate/ flexible enough for precision timing. Operates at full power at all times, the device never sleeps.

Answer 8

Delays can be used to run a program at specific times, however, this approach is flawed as it requires precise knowledge of how long a task takes and for said task o never vary in execution time. Alternatively, hardware timers can be used to generate interrupts. As long as the timer counts continuously the program can be designed to generate ticks at precise intervals.

Answer 9

A tick is an arbitrary unit for measuring internal system time. What exactly is a tick varies based on the OS and may vary by installation. ## Footnote -Super User

Answer 10

The rate of tick generation depends on the crystal frequency and the counter configuration

Answer 11

This can be achieved by putting the system to a kind of sleep. Here interrupts can be used to wake the system up at next processor overflow.

Answer 12

Running- actually using the processor Ready- Able to run but a higher priority task is currently running Blocked- waiting for an external event or time, but is timeout protected Suspended - Not available for scheduling and is not timeout protected

Answer 13

Multitasking and intertask communication allow for complex applications to be partitioned into smaller, more manageable tasks. This makes software testing easier as well as work breakdown between teams and code reuse. It means that timing and sequencing details can be removed from the application code and laid into the hands of the OS.

Answer 14

Context switching is used as tasks do not have an awareness of when they are suspended, so the OS saves the context of the system to be restored via the kernel near execution time. This increases performance as it means that higher priority tasks can be run when ready and allow for scheduling.

Answer 15

It is an algorithm that a scheduler uses to determine which task to execute at a particular time

Answer 16

**Task Scheduling:** Tasks are called when needed, ensuring better program flow and event response times **Multitasking:.** Task scheduling gives the illusion of multiple simultaneous task executions **Deterministic Behaviour:** Events and interrupts led are handled within a defined time **Shorter ISRs:** enables more deterministic interrupt behaviour **InterTask Communication:** manages sharing of data, memory and hardware resources between multiple tasks **System Management:** Allows focusing on app development rather than resource management **Defined Stack Usage:** Each task has a defined stack space allowing for predictable memory usage.

Answer 17

Unlike desktop Os, programs cannot be loaded after compilation or easily exchange components. The entire system must be completed together at compile-time, whilst this is efficient as optimisation is applied on the entire program, this is also restrictive as any changes require a firmware update.

Answer 18

**Strict Timing:** Events must happen on a strict schedule e.g. every 10ms **Flexible timing**: Wide range of acceptable times **D****eadline only timing**:Only Upper Limit is specified

Answer 19

Ensure that the highest priority task that is able to execute is the task given processing time.

Answer 20

This means processing time may have to be shared, 'fairly' between tasks of equal priority if the tasks are able to run simultaneously.

Answer 21

When sleeping the task specifies the wake time When blocked, the task specifies the maximum wait time.

Answer 22

The RTOS measures time using a tick 'count' variable, a tick is implemented using timer interrupts, the resolution of a tick is dependent on the interrupt frequency. At each tick the kernel checks whether to unblock/wake tasks

Answer 23

When sleeping a task gives up its access to the CPU. It specifies its wake time.

Answer 24

When blocked a task specifies its maximum wait time.

Answer 25

Each task as a priority The idle task has a priority of zero Low priority numbers are for low priority tasks

Answer 26

Tasks have no knowledge of the scheduler Whilst the scheduler's job is to ensure the context around a task is saved whilst the task may be blocked, suspended etc.

Answer 27

**Benefits:** Simple, no restriction on use, supports full preemption, fully prioritised **Flaws:** High usage from individual stacks, restrictions on re-entrancy are required, i.e. on subroutines calling themselves.

Answer 28

The Idle task runs at lowest priority, it is used to clean up resources allocated by the kernel to deleted tasks It can call a 'hook' function when no higher priority tasks are available Hook function puts the processor into low power mode

Answer 29

1. TaskA is executing, TaskB previously suspended, so context already saved on TaskB stack. 2. RTOS ‘tick’ occurs as TaskA is about to execute LDI. When tick occurs, places PC on stack before jumping to the RTOS tick ISR 3. To save context, pushes context of processor onto TaskA stack. Kernel stores copy of stack pointer for each task 4. After context has been saved, increments tick count. In this example, incremented tick count means that TaskB is now ready to run. TaskB has a higher priority than TaskA, so kernel selects TaskB to be given processor time. 5. Task B stack pointer is retrieved. 6. Restores TaskB context from the stack into the processor registers 7. The stack pointer was adjusted to point to TaskB’s stack, Program Counter restored for TaskB. So, when e.g. returning from interrupt, executes next TaskB instruction

Answer 30

Ported to at least 34 Processors Open Source free software for rtos system

Answer 31

Linux kernel operates as a thread Interrupts for deterministic processing are processed by the real-time core; other interrupts forwarded to Linux (running at a lower priority) Linux drivers handle general I/O Can use FIFO to transfer between OS and RTLinux Example capabilities (on x86 hardware): 15 µs to begin an ISR after an interrupt event Periodic tasks run within 25 µs of schedule Compared against ‘standard’ Linux Up to millisecond level scheduling precision •Not designed for guaranteed timing

Answer 32

RTLinux : modified Linux kernel Adds hard real time capabilities to ‘standard’ Linux OS Runs whole Linux OS as fully pre emptive process Can use for real time control Capability of running real time tasks and interrupt handlers: execute regardless of Linux OS operations http://

Answer 33

Linux drivers handle general I/O Can use FIFO to transfer between OS and RTLinux Example capabilities (on x86 hardware): 15 µs to begin an ISR after an interrupt event Periodic tasks run within 25 µs of schedule Compared against ‘standard’ Linux Up to millisecond level scheduling precision Not designed for guaranteed timing

Answer 34

Linux kernel operates as a thread Interrupts for deterministic processing are processed by the real time core; other interrupts forwarded to Linux (running at a lower priority)

Answer 35

The timed interrupt is declared early on in the program. The task is executed using functions called from the ISR

Answer 36

Interrupt Service Routine

Answer 37

Events must happen on a strict schedule e.g. every 15ms.

Answer 38

Events have a wide range of Acceptable times

Answer 39

Only the upper limit of event duration is specified.

Answer 40

TaskA is executing TaskB previously suspended, so context already saved on TaskB stack

Answer 41

RTOS ‘tick’ occurs as TaskA is about to execute LDI When tick occurs, places PC on stack before jumping to the RTOS tick ISR

Answer 42

To save context, pushes context of processor onto TaskA stack Kernel stores copy of stack pointer for each task

Answer 43

After context has been saved, increments tick count In this example, incremented tick count means that TaskB is now ready to run TaskB has higher priority than TaskA , so kernel selects TaskB to be given processor time

Answer 44

Task B stack pointer is retrieved

Answer 45

Restores TaskB context from the stack into the processor registers

Answer 46

Stack pointer was adjusted to point to TaskB’s stack, Program Counter restored for TaskB So, when e.g. returning from the interrupt, executes next TaskB instruction