ikt-102_operativsystemer_20231214230226 Flashcards

Question

Talk about the states a process can be in.

Answer 1

There are three possible states for a process to be in: - Running- Ready- Blocked. It is common for processes to be interrupted many times during their lives, but from the process' point of view, it would be as if they never stopped.When a process is running, it is using the CPU. When a process becomes blocked, that means it is waiting for an external event to occur. This is known as an interrupt and are common for many processes. When a processes are ready, they are waiting for their turn to use the CPU. For example, a process that was interrupted by an I/O will become blocked, when the input is available it will then enter the ready state and wait for its turn to use the CPU.| running, ready, blocked, interrupt, ready state, never stopped at all

Answer 2

- Process table (alt. Process Control Table = PCB)- One entry per processIn order for states to be saved, the system has an array of structures called the process table, with one entry per process. With these parameters, it is possible for a process to change between the three aforementioned states.| Process table, one entry per process

Answer 3

Threads may avail themselves of their host process' items, such as: - Address space, - memory, - global variables, - **open files,**- child processes, - pending alarms, - signals and signal handlers, - and accounting information.

Answer 4

- A way of viewing a process is that of a thread of execution, usually shortened to thread. - Each process has an address space, and if multithreading is possible, there can be two or more threads of control within one process.| thread of execution, address space, memory, open files

Answer 5

In each thread's stack there is a **frame** for each **procedure** that has been *called but not returned*, and the frame's **local variables** and *return address* are there as well for when the procedure has finished.In the stack are frames, with local variables, that haven't been returned, and return address.

Answer 6

Threads themselves have items of their own, such as: - Program counter- Registers- Stack- State.| Program counter, registers, stack, state

Answer 7

We mentioned earlier that threads share address space, open files and memory with the process of which they are part. This is important for some applications where you cannot share all resources among multiple processes. An example of such an application is Microsoft Word, where**one thread can interact with the user**, **another can reformat the document as changes are being made**, **and a third can occasionally write to RAM**. This is only possible because the threads all share memory, otherwise they could not have access to the same document being edited.| address space, open files, memory, microsoft word (user, reformat, RAM)

Answer 8

Threads are reminiscent of processes, where multiprogramming is the rapid switching between processes. Multithreading is the same thing, **except now we have threads who share memory**. Creating threads also happens very fast, about 10-100 times faster than processes, resulting in less overhead from creating them.Furthermore, in systems with high computing requirements and significant I/O interrupts, threads will yield better CPU performance.| switch between process (equivalent) multithreading (threads share memory

Answer 9

One important difference between processes and threads is the absence of a clock interrupt within processes themselves to force threads to share CPU time. Without such a call to enforce resource sharing among threads, a thread call **thread_yield** is used to make threads share the CPU, giving other threads a chance to make progress on their work.| absence of clock interrupt, thread_yield

Answer 10

[memory] instruction word -> [register] -> instruction word [memory]Registers typically hold **instruction words from memory**, or **words to send to memory**, such as - **variables** and - **temporary results**, as in temporary variables.

Answer 11

The program counter is used to**fetch instructions**, and it **points to the memory address of where the instructions are**.

Answer 12

The stack pointer **points to the stack in memory that is currently being used**, and holds **local variables**, **temporary variables**, and **input parameters** that are **not held in registers**.

Answer 13

UNIX processes have their memory divided into **text**-, (program code)**data**-, (variables)and the **stack**. (grows as needed)| text, data, stack

Answer 14

File management is about organizing the**location** of **files** or **processes**.

Answer 15

**Root directory** deals with the **location of files**, and **working directory** deals with the **location of processes conducting work**.

Answer 16

File descriptor deals with**access permission** to **read/write to files**.

Answer 17

One problem that could occur with user-level threads is related to **page faults**. If **page fault** -> **process is blocked** while **system looks for missing instruction**,regardless if other threads are runnable| page fault, blocked, missing instruction

Answer 18

thread about to block -> run-time system procedure -> switch to another threadswitch to another thread =kind of like= multiprogrammingrun-time system procedure allows for rapid change between threads| run-time system procedure

Answer 19

kernel thread =/= run-time system procedurekernel thread switch, must make system callrun-time system procedure = fast and cheapsystem call = costlyOne problem with kernel-level threads is that they do not have this run-time system procedure, meaning kernel threads must make system calls when a kernel thread is about to be blocked. This comes at a much greater cost than user-level thread switching.| run-time system procedure, system call, blocked, greater cost

Answer 20

**Processes** and **threads** have *states*, or *contexts*, which are recorded in the process control blocks (PCB). The PCB can also be called the process-table, and is the key to multiprocessing.| process control block (PCB), process table

Answer 21

Information about the process is stored at all time in the PCB, or procecss table, and it all is managed by the OPERATING SYSTEM.It records:- **program counter**- **registers**- **schedules next process**Changing from **running state** to either **ready** or **blocked** state is possible because of the information in the PCB.From the process POV, it would be as if it had never stopped running, from start to completion.| state change, PCB, schedule

Answer 22

The operating system assigns processes **unique identifiers**, usually **indices**, in order to *keep track of them and what they are doing in cooperating with other processes. *

Answer 23

While the process is running, information about it is kept in **registers**, and in the case of interrupts this information is saved.

Answer 24

- page table- state- scheduling- program counter

Answer 25

This information is recorded in the PCB, and may be broadly grouped into three categories:- Process identification- Processor state information- Process control information

Answer 26

The condition for initiating a process switch is usually an interrupt. Normal interrupts include **clock**-, interrupts**I/O**- interrupts and**memory fault** interrupts. Whenever these occur, an **interrupt handler** receives control of the process for some work to be done, depending on the nature of the interrupt **before it gives back control to the operating system**.| interrupt, clock, I/O, memory fault

Answer 27

There are four possible changes between the three process states, and they are:running -> ready (stopped by scheduler)running -> blocked (stopped by interrupt)ready -> running (scheduler picked this process to run)blocked -> ready (I/O work is complete)| running, ready, blocked

Answer 28

compute bound = few I/O requestsI/O bound = many I/O requests, short bursts of computingcomputer systems alternate a lot between actively computing and waiting for I/O requests to finish| cpu bound, I/O bound

Answer 29

In order to improve performance of the system, you can design the operating system with I/O bound processes in mind. A feature that could be of benefit, would be to focus on **allowing I/O bound processes to run as soon as possible**,thus beginning I/O work quickly and **keeping I/O busy instead of idle**.

Answer 30

The round robin scheduling algorithm is based on the round-robin principle, where everyone gets a fair and equal share of something.This is categorized as a **pre-emptive scheduling algorithm**. It is an old and simple scheduling algorithm, and is used mostly for**multitasking**.In computing we share time slices of the CPU's time, giving processes a tiny and equal piece. Processes are preempted and are moved to the back of the queue when they have spent their alloted time.| equal share of something, preemptive scheduling

Answer 31

Priority scheduling is a non-preemptive algorithm and one of the most common scheduling algorithms in **batch systems**. **Each process is assigned first arrival time**, and if two processes have the same arrival time, OS compares priority -> highest priority winsif priority is the same, process number is compared,lowest number wins| batch systems, first arrival time > priority > lowest number

Answer 32

Inter process communication (IPC) happens through:**Shared Memory** (some shared variable)**Message Passing** (direct or indirect link)Using these techinques processes can **communicate** and,**synchronize** their actions

Answer 33

Independent processes are not affected by the execution of other processes, whereas a co-operating process certainly can be affected by other executing processes.

Answer 34

A co-operating process is dependent upon other processes. In many situations, co-operating processes can**increase modularity**,**computational speed**, **convencience** and**modularity**

Answer 35

A process can largely be defined as two types:- Independent process- Co-operating process| independent, co-operating, shared memory, message passing, IPC

Answer 36

The **page table** is a data structure used by the **virtual memory system** to store the mapping between **logical addresses and physical addresses**. These addresses are managed by the **MMU**.| page table, data structure, mapping, MMU

Answer 37

In order to make use of virtual addresses we need a memory manager to map the virtual address generated by the program onto physical memory.## Footnotec

Answer 38

A physical address is as the name suggests a physical location in memory, and **virtual addresses are generated by programs** and make up the **virtual address space**.## Footnotec

Answer 39

Since virtual address space >> physical memory:- can run programs much larger than available memory- can run multiple programs simultaneously## Footnotec

Answer 40

It is still possible to run multiple programs on systems without a virtual address space if you make use of **swapping**. Swapping entails swapping out the entire program from memory, in order to bring in the new one so that can be run instead.## Footnotec

Answer 41

Page replacement algorithms are used to determine what page to **evict** from memory when a **page fault** occurs. Algorithms are useful in this regard since they hopefully will improve **system performance** by using a set of criteria to evict pages, rather than just selecting at random.## Footnotec

Answer 42

The clock algorithm will assign what is known as an **R bit**, which is either 1 or 0, to pages when they are referenced. Whenever a page fault occurs, the algorithm looks at the page it is **currently pointing to**, **if the bit is 1 it is cleared to 0** **and the pointer moves to the next page**, this repeats **until it encounters a page that is initially 0**, at which point it is evicted from memory.## Footnotec

Answer 43

The memory hierarchy refers to an aspect of different types of memory related to- size, - speed, and - cost basically,smaller = faster, expensivelarger = slower, cheaperregisters > cache > main memory (RAM) > magnetic disk > magnetic tape## Footnotec

Answer 44

MMU handles:- **efficient allocation of memory** to processes- **keeping track of what locations in memory are available**- **mapping** virtual addresses to physical addresses, this is known as paging- **swapping**## Footnotec

Answer 45

relocation solutionmapping each process' virtual address space to physical memory**generate addresses automatically** with **base and limit registers** before sending to memory (SEGMENTATION)A disadvantage of segmentation is that a comparison must be made for ever reference made.## Footnotec

Answer 46

protection solutionis solved by the concept of an address space. This way, **every process has its own address** space where it can address memory.## Footnotec

Answer 47

Segmentation is a memory management technique in which the memory is divided into the **variable size parts**, the details of which are stored in the **segment table**. The **segment table contains** mainly two items of information about the segment:SEGMENT TABLE INFO:1. Base: The **base address** of the segment.2. Limit: The **length of the segment**.## Footnotec

Answer 48

- **No internal fragmentation**- Average Segment Size is larger than the actual page size.- **Less overhead**- It is easier to relocate segments than entire address space.- The segment table is of lesser size as compared to the page table in paging.## Footnotec

Answer 49

- It can have **external fragmentation**.- it is **difficult to allocate contiguous memory** to variable sized partition.- **Costly memory management algorithms**.## Footnotec

Answer 50

The MMU manages virtual memory, and the the virtual address space consists of units called pages, and correspondingly the unit in physical memory is called page frames. virtual address space = **pages**physical = **page frame**This is in essence what paging is, it is what the MMU does when it maps a page from the virtual address space onto a page frame of the physical memory.PAGING IS:**virtual -> physical = pages -> page frame**## Footnotec

Answer 51

Free lists and bitmaps## Footnotec

Answer 52

Bitmaps can use to track memory usage.Bitmaps divide memory into allocation unitseach AU is assigned a bit, 1 or 01 = occupied0 = freebitmaps allow us to find holes in memory a process can occupy## Footnotec

Answer 53

NFU uses a **counter** to count how often a page has been used. The count is tracked by the R bit, which is 0 for no reference made, and 1 for a reference made. 0 = no reference made1 = reference madeThese bits are added to the **counter**. When more space is needed, the page with the lowest count is evicted.## Footnotec

Answer 54

page fault -> page replacement algorithm -> evict some page- aging algorithm is based in software, - and is a variation on NFU (not freq. used)- **NFU never forgets counts**, regardless how they are gained- some process from the boot sequence might **linger in memory**- **aging algorithm limits how far back you count references by the R bit**- NFU = count references made- Aging = count sequence of bits, maximum 8 (or other number)In aging, on every clock tick:-> add R bit to leftmost place of sequence,-> historical R bits are shifted to the right-> evict lowest count## Footnotec

Answer 55

The drawback of paging is the need to make at least **two memory references to the page**, **effectively halving performance**, since without paging you would only need to make one reference.## Footnotec

Answer 56

Since paging effectively halves performance, you can use a hardware component: **Translation Lookaside Buffer (TLB),** implemented in the MMU, **to carry the most used pages**, improving performance substantially.## Footnotec

Answer 57

In the **software solution of TLB the OS finds the missing page itself**,(-)not as fast as the hardware solution,(+)software solution frees up space on the CPU chip## Footnotec

Answer 58

Access to printer/file only for one process at a timeto protect against two processes writing to the sameresource simultaneouslyif this safeguard fails, race conditions can happen## Footnotec

Answer 59

Deadlocks may occur then when two or more resources wish to use the same resources, which end up not being available. In this manner both processes wait for the other to release the resource, which will not happen.## Footnotec

Answer 60

For a deadlock to occur you need two or more processes who want to access the same resources, but each process being programmed to do so in different ways. We can imagine two processes, **A** and **B**, who want to use the resources **x** and **y**, but they ask for permission in an order such that the following happens:- Process A requests permission to use resource x and is granted it.- Process B requests permission to use resource y and is granted it.The problem that arises then is apparent when process A asks to use resource y, and process B asks to use resource x. Since both processes cannot do anything but wait until the resource is available, the outcome is a deadlock that cannot resolve itself.## Footnotec

Answer 61

A race condition can occur in certain circumstances where two or more processes are writing to the same shared data, such as an address in a spooler directory, in the case where they are sending input to a printer.## Footnotec

Answer 62

This can happen if an ill-timed interrupt occurs, resulting in a process B thinking it is writing to an available slot, instead it is overwriting the input of another process A. This sequence of events continues, such that process B is shadowing process A, overwriting every input. Thereby, process B ensures that that unfortunate process A never receives any output.| writing to shared data, interrupt -> Process B overwrite pr. A## Footnotec

Answer 63

Race condition and deadlock are serious bugs that can cause unpredictable behavior in a computer system. Race conditions are especially nasty since it can be very confusing for users what the underlying problem is.## Footnotec

Answer 64

Since race conditions a dependent upon such **strict conditions** for them to occur, they are **not guaranteed** to happen. Thus, during test runs it is unlikely for someone to replicate a race condition. Similarly, deadlocks are dependent upon a specific sequence of events, and thus can also be very tricky to debug.| strict conditions, hard to replicate## Footnotec

Answer 65

There two operations for the semaphore, **up and down**. For every iteration of a loop the semaphore **checks the count of wakeups**, **if it is greater than zero it will reduce them by one.**If there are processes on the semaphore that still haven’t completed a down operation, they will be selected at random until all processes on the semaphore are down.## Footnotec

Answer 66

A semaphore is an integer variable that represents available wakeups, it can be zero or higher, its aim is to achieve mutual exclusion between processes who wish to enter critical regions of shared resources.## Footnotec

Answer 67

There is another kind of semaphore called the binary semaphore. This is useful for ensuring that processes cannot enter into a critical region if another is already there. The **critical region** being a **shared resource** processes can write to.**critical region = shared resource**Processes in this case will do a 1 **down** operation **when entering** this space, 2 and an **up** operation **when leaving**. Thus, ensuring mutual exclusion.| protect critical region, shared resource, enter = down, leave = up## Footnotec

Answer 68

A mutex is similar, but simpler, than a binary semaphore,it holds and **integer value** representing **locked** or **unlocked**, the mutex is a shared variable between processes,that is intended to protect some shared resource,mutex is useful in user level thread packagesWhen a **mutex is unlocked**, a process can call mutex_lock and enter the region protected by the mutex. **Processes calling while a mutex is locked** must wait until the process calls mutex_unlock and vacates the space. If multiple processes are calling mutex_lock, one is chosen at random and may enter the space| protects shared resource, locked, unlocked, calls until it can enter## Footnotec

Answer 69

- Semaphore- Binary semaphore- Mutex- Busy waiting## Footnotec

Answer 70

There are various implementations of busy waiting, such as: - **Peterson’s solution** and the - **TSL instruction**.## Footnotec

Answer 71

The four basic criteria for successful solution to mutual exclusion challenge- No two processes may be simultaneously inside their critical regions.- No assumptions may be made about speeds or the number of CPU’s.- No process running outside its critical region may block any process. - No process should have to wait forever to enter its critical region.## Footnotec

Answer 72

1. **Disabling interrupts** might conflict with the 4th condition, also it is not feasible to disable interrupts in a multiprocessor environment, since there is nothing stopping the other cores from causing interference. **cannot work in multiprocessor environments!**2. **Lock variables** (MUTEX) have the same underlying problem we discussed with race conditions, thus **failing on the 1st** condition. Mutex fail 1st condition (**no two processes inside crit region simultaneously**)3. **Peterson’s** and the **TSL** solutions are valid and fulfil all the conditions.## Footnotec

Answer 73

Busy waiting is another method of achieving mutual exclusion in critical regions in shared memory. However, busy waiting, is in essence the wasting of CPU time, and is in and of itself inherently undesirable. Therefore, busy waiting is not desired in operating systems.| busy waiting, conditions for successful solution, 4 points## Footnotec

Answer 74

1 Direct Memory Access, DMA, is a hardware component that aids the CPU in managing I/O requests, holding I/O requests in registers accessible by the CPU. 2 Most DMA controllers make use of **physical memory addresses**, and for that to be possible the **operating system has to convert the virtual address of the intended memory buffer into a physical address**, and then write this physical address to the DMA’s registers.| hardware component, I/O requests, registers## Footnotec

Answer 75

To explain how the DMA works, consider the case of transferring data to memory from a disk. 1. Initially, the DMA receives instructions from the CPU to transfer data to memory, 2. the DMA then makes a request of the disk and transfers the data to memory. 3. When the task is **complete**, the DMA **interrupts** the CPU to make it known.| transfer data from memory -> disk, CPU instructions, interrupt## Footnotec

Answer 76

(+) The obvious benefit of the DMA is that it **saves the CPU valuable time** by handling time consuming I/O requests. However, this argument is relative and dependent on the use case. If you have a system where the **I/O is fast**, it makes little sense for a slow DMA controller to handle transfer requests rather than having the fast CPU do it instead. THUS,(-) IF I/O is fast -> let fast CPU handle I/O requests**You do not want your CPU to wait on the DMA controller.**(-)Furthermore, not having a DMA at all saves money, which is a benefit in and of itself.| Save CPU time for I/O bound systems, no DMA = save money## Footnotec

Answer 77

**1. signal goes to -> interrupt controller**Interrupts are signals that go to the interrupt controller, they tell the controller that the work assigned to the I/O device by the CPU is now complete. **2. interrupt controller decides next action -> CPU interrupted**From there, the interrupt controller determines what happens next. Depending on whether there are other interrupt requests from pending, it might have to wait and continue to assert its interrupt signal until it is serviced by the CPU. i.e., **I/O must wait until it is serviced by the CPU****3. CPU interrupted -> process is blocked**1 When the CPU gets interrupted it stops its current task to address the request. Thereby, blocking the process it was working on. 2 The interrupt vector fetches a new program counter, which points to the start of the corresponding interrupt service procedure.## Footnotec

Answer 78

- Interrupts- busy waiting- DMA## Footnotec

Answer 79

ADVANTAGES* Interrupts enable I/O to work while CPU can do something else* Save costs on DMA by using interrupts* Lets higher priority tasks block less important processesDISADVANTAGES- Can happen at inconvenient times- CPU can disable interrupts if it wants to## Footnotec

Answer 80

The **operating system is responsible** for maintaining this abstraction for programs to interface with, and to handle all the complexities of the underlying hardware. There are ways to aid in this, a couple of examples are:- **Uniform naming**. Names of files or devices are independent of the device itself, in favor of addressing them by a path name.PATH NAME - **Error handling**. Whereby, errors that occur are handled at lower layers, and often resolved there, without the operating system being aware of it.OS NOT AWARE OF ERRORS HANDELED## Footnotec

Answer 81

Device independence means it **should not matter what type of device is being read from or written to in programs**, regardless of how different they actually are behind the abstraction.## Footnotec

Answer 82

1. HDD must undergo **low-level format**, they are empty from factory2. format consists of three fields, **preamble**, **data** and **ECC** (Sizes may vary dependant on device)3. **Disk partition**1 each partition is logical disk2 sector 0 holds **Master Boot Record** (MBR) 3 MBR holds partition table and boot code4 partition example: C: drive4. **High-level format** of each partition that establishes:1 root directory2 empty file system5. after high-level format system is boot ready6. **BIOS reads MBR**, checking boot sector to runIf successful -> launch bootstrap program to find OS kernel -> load into memory -> execute| Error Correcting Code## Footnotec

Answer 83

To understand what this means, firstly understand that the file system in UNIX **grows like a tree**, starting from the root directory. In order to locate a file in a given directory we use path names. directories use path names, the **file system** looks for the **relevant i-node associated with the file path name opened within the root directory**,when it has the i-node, **the file system looks for the i-number to the next i-node***and so it goes *i-node (in root) -> i-number -> i-node -> i-number et cetera.thusly the file system propagates through i-nodes until it has reached the file## Footnotec

Answer 84

**opening a file retrieves an i-number (index)**i-number -> **table of i-nodes** (index nodes)i-nodes contain information such as:- **file size,**- creation- last access- last modification- owner, group- **protection information**## Footnotec

Answer 85

**General explanation what OS is**:- Common platform- Software that acts as an Interface between hardware and softwareTwo main roles of an OS:**Extended Machine****Resource Manager**- Process- Memory- I/O- Files

Answer 86

- Hardware is complex- Programmers don't want to worry about hardware- OS hides complexity behind abstractions- OS acts as a common platform

Answer 87

OS manages:- **Open files** - **I/O**- **Memory**- **Processes****Multiplexing**OS shares resources in two ways:- Time (CPU)- Space (Memory)OS Responsibility:- **Control all activity in the system**, I/O devices, programs, storage

Answer 88

:: **Old solutions might suit new technology better**It may be useful to know older solutions when innovating new technology that might be better if implemented with an old "outdated" solution than with a new one.:: **New solutions solve old problems**Newer solutions naturally attempt to solve some problem with older systems. Those problems might not be relevant at all for what you are trying to achieve. By knowing more one may be better able to discern what concepts to include or exclude from a technology.:: **Multiprogramming and hardware protection**An example:Multiprogramming is a given in modern computer systems, but not all technologies can make use of running multiple programs simultaneously. If you have multiprogramming available to you, you probably have costly protection hardware as well. Perhaps your new technology doesn't need to run multiple programs, so it wont need protection from different processes interfering with each other.

Answer 89

1. **Running -> Blocked**The process in this case was running, then either due to a memory fault, or an I/O request, it was blocked.2. **Running -> Ready**In this case, the process is likely subject to a clock interrupt and made to go back to queue and wait for its turn on the CPU.3. **Ready -> Running**In this case, the CPU was chosen by the scheduler to use the CPU4. **Blocked -> Ready**In this case, the process was initially blocked by an I/O request. When the I/O request has been completed the process changes to Ready state.5. **-> Ready**A process has been initiated and it enters an initial state of Ready, waiting in queue for its turn on the CPU.6. **Running -> Exit**The process has completed its work, it Exits and is terminated by the operating system.

Answer 90

:: **Generell forklaring på process og tråd**- A process is an instance of a running program, where the program is simply the recipe.- A thread can be seen as a program counter within a program counter- A process can be seen as a thread of execution

Answer 91

:: **Ressursgruppering for prosesser og tråder**- *Processes* have an address space, text, data, stack- *Threads* **share all memory** with their host process as well as **address space**, and they have their **own stack** and **program counter**

Answer 92

- **Process Control Block** (PCB) tracks all information about processes and records their state when a context switch happens- Processes and threads have three states, **running**, **blocked**, **ready**.- **Text** holds the program source code- **Data** holds variables and grows in memory- **Stack** grows in memory as the program runs- **Registers** hold words to memory and words from memory that have not yet been returned, as well as their address to return to

Answer 93

The process and thread is the keystone of modern computing. It enables:- **Multiprogramming**- **Multiprocessing** - **Multithreading**Leading to vast improvement to system performance. As the limit of how many atoms can fit together on a computer chip is getting closer, optimizing programs to make use of multiprocessor and multithread environments helps push computing power and efficiency forward.- System performance - Limit on how many transistors on CPU gjetting nearer

Answer 94

You may want as small a kernel as possible to minimize numbers of bugs within kernel space where access to hardware is. Bugs here may cause more harm than bugs in user space.The most basic functions to include in an such a microkernel are:- **Process management**, **Scheduler**- **Interprocess communication (IPC)**- Calls to functions, like **interrupt handlers**And ones to exclude are:- **Memory management**- **I/O**, you don't want buggy audio driver to crash your system

Answer 95

- Paging is simply put: **the act of mapping virtual addresses (virtual pages) onto physical addresses (page frames)**- **Paging is done by the MMU**- The virtual address space may contain many more addresses than available physical memory addresses.

Answer 96

- **The address space is an abstraction of physical memory, broken up into pages**.- **Initially, programmers had to split their programs into overlays**. With the virtual address space, the operating system can do it for you. - Therefore, even if you don't need an abundance of address spaces, paging can still provide a tremendous benefit for programmers by saving them from creating overlays, and letting the OS do the work instead.

Answer 97

As opposed to **programmed I/O** where the processor checks if whether the I/O device is ready to transmit data or not, interrupt driven I/O makes the I/O device itself emit a signal to the CPU that it is ready to be serviced.- Programmed I/O... **CPU to I/O: "HELLO ARE YOU READY?"**- Interrupt I/O: **I/O to CPU: "I AM READY!"**(+)Interrupt driven I/O is:- **Fast**- **Efficient**(-)It does have some problems:- Interrupts can come at inconvenient times, resulting in **race condition**.**DMA** is slow, works in heavy I/O environmentDMA handles transfer of data to a buffer location instead of the CPU, however, it is slow. Thus, if there is not enough jobs for the DMA, the fast CPU really should be handling the requests instead.

Answer 98

Process uses its CPU time to ask an I/O device if it is ready or not. In this way, it makes sure other processes cannot enter the critical region of this shared process.

Answer 99

Process will ask for access to the semaphore. This action is an atomic operation. If it is available, the process will be allowed to enter the critical region. The **semaphore has a simple queue**, first come first served.Processes waiting in the queue are in the **blocked state**.

Answer 100

Drivers are programs that operate and control devices in computer systems. The driver provides a **software interface to hardware devices**, enabling the OS and other programs to **access hardware functions** without needing to know the details about the hardware in the device itself.

Answer 101

Cons:- **Storing data in i-nodes increases overhead**, making the directory larger, - **it also takes longer to traverse the file path to the directory**.Pros- It is possible to **delete/rename files that are currently being opened by a process**. The file will continue to exist until no process has it open anymore.

Answer 102

Cons:- **Storing data in i-nodes increases overhead**, making the directory larger, - **it also takes longer to traverse the file path to the directory**.Pros- It is possible to **delete/rename files that are currently being opened by a process**. The file will continue to exist until no process has it open anymore.i-nodes are for FILE MANAGEMENT, not memory management.Every file gets an i-nodepro:- i-node trenger kun være i minnet når fil er åpen

Answer 103

New process -> must be managed -> information stored in the PCBMajor PCB categories are:PCB -> process **identification**PCB -> process **state** (status of suspended process)PCB -> process **control**process **identification** -> **index** (unique identifier)process **state** -> **save registers** before suspending processprocess **control** -> **page table**process **control** -> **state**process **control** -> **scheduling**process **control** -> **program counter**

Answer 104

1. 64-bit1. fast, small and cheap1. 32-bit1. **fast, large and cheap**

Answer 105

1. larger1. slower1. **faster**1. **smaller**

Answer 106

1. **Integrated Circuit**1. Intel Computer1. Intern CPU

Answer 107

1. No answer text provided. [ ]1. When the HDD is switching between different plates [ ]1. When booting(turning on) the machine [ ]1. **When switching from one process to another [X]**

Answer 108

1. True1. **False**

Answer 109

1. A system call only happens when a program requests a service from the kernel and the request is denied.1. **A system call is always needed when a program requests a service from the kernel, such as reading data from a file.**1. A system call only happens when a program requests a service from the kernel and the request is granted.

Answer 110

1. **Grouping files together in a system to make it more accessible for users and programs.**1. Making it possible to automatically place new files when they are created.1. Keeping track of all available system calls for creating, removing, reading and writing files.

Answer 111

1. **containing the memory address of the next instruction to be fetched**1. telling the CPU when to execute the next instruction of the program code1. keeping track of how many programs that are running

Answer 112

1. **kernel mode**1. user mode1. private mode1. windows mode

Answer 113

1. **POSIX**1. MINIX1. LINUX1. UNIX

Answer 114

1. **multiplexing**1. multiallocating1. multispacing

Answer 115

1. **True**1. False

Answer 116

1. GPU1. SSD1. HDD1. **MMU**

Answer 117

1. process table1. file descriptor1. **address space**1. process control block

Answer 118

1. private mode1. **kernel mode**1. user mode

Answer 119

1. VOLATILE, meaning that it WILL NOT lose its content when the power is switched off. [ ]1. **NON-VOLATILE, meaning that it WILL NOT lose its content when the power is switched off. [X]**1. **VOLATILE, meaning that it WILL lose its content when the power is switched off. [X]**1. NON-VOLATILE, meaning that it WILL lose its content when the power is switched off. [ ]

Answer 120

1. containing the memory address of the next instruction to be fetched1. containing the memory address of the top of the current stack in memory1. **containing essential information about the process, such as priority and mode(user/kernel).**

Answer 121

1. **Keyboard/Mouse [X]**1. **Printer [X]**1. CPU [ ]1. **HDD (hard disk drive) [X]**

Answer 122

1. **fetching and executing instructions**1. storing data1. making sure of a stable internet-connection

Answer 123

1. Information about which processes that are currently running, and each process’ address space.1. Information about which process that are currently existing, and each process’ core image.1. **Information about which process that are currently existing, and all the information about each process other than the content of its own address space.**

Answer 124

1. The process-concept exists only in personal computer operating systems [ ]1. **A process is fundamentally a container that holds all the information needed to run a program. [X]**1. **A process can be seen as a program in execution [X]**1. All processes share a common memory-location called the address space [ ]

Answer 125

1. That the program is currently not running1. **That the program is currently running**1. That the program is about to exit

Answer 126

1. Lese og skrive til hardware1. **Håndtere maskinens ressurser og tilby abstraksjoner som skjuler kompleksiteten til systemet**1. Håndtere maskinens ressurser og vise GUI(Graphical User Interface) til brukeren

Answer 127

1. **cache hit**1. cache miss1. cache line1. memory hit

Answer 128

1. Når en prosess har én stor oppgave som for eksempel å regne ut et stort tall.1. **Når en prosess har flere oppgaver som kan utføres uavhengig av hverandre.**1. Når flere forskjellige prosesser skal kommunisere med hverandre.

Answer 129

1. Impossible to answer, it depends on the size of the process. 1. At least two. 1. **One.**

Answer 130

1. True1. **False**

Answer 131

1. When two different processes are competing about creating new files as fast as possible. 1. **When two or more processes are simultaneously trying to queue a file for printing in the same printer queue.** 1. When two or more processes are simultaneously trying to read from a file they don’t have permission to write to.

Answer 132

1. **True**1. False

Answer 133

1. True1. **False**

Answer 134

1. **En I/O-bound prosess bruker mesteparten av tiden sin på å vente på ekstern informasjon mens en compute-bound prosess kun er begrenset av prosessoren sin hastighet.**1. I/O-bound prosesser eksisterer kun på UNIX-baserte operativsystemer, mens compute-bound prosesser kun eksisterer på Windows.1. I/O-bound prosesser eksisterer kun på Windows mens compute-bound prosesser kun eksisterer på UNIX-baserte operativsystemer.1. En I/O-bound prosess er begrenset av hvor raskt prosessoren kan arbeide, mens en compute-bound prosess er begrenset av hvor raskt den kan motta informasjon fra en ekstern kilde.

Answer 135

1. True1. **False**

Answer 136

1. **True**1. False

Answer 137

1. **Det er den minste handlingen som kan utføres av prosessoren. Enten er den helt utført eller ikke påbegynt i det hele tatt.**1. Det er en stor oppgave som prosessoren bruker lang tid på å utføre. 1. Det er en farlig handling, f.eks. når en prosess forsøker å dele på null. 1. Det er et begrep som brukes om oppgavene til prosesser som kjører på kjernekraft-anlegg

Answer 138

1. **True**1. False

Answer 139

1. **True**1. False

Answer 140

1. True1. **False**

Answer 141

1. **Ready, running, blocked**1. Ready, runnable, running, blocked1. Running, blocked1. Ready, not ready, blocked

Answer 142

1. It’s a trick question. It’s not possible to have multiple programs in memory at the same time.1. **The different programs are assigned different timeslices to run, and the CPU is switching back and forth between them.**1. Multiple cores, one core for each program.

Answer 143

1. **True**1. False

Answer 144

* Recipe [**Program**]* Person baking the cake [**CPU**]* Cake ingredients [**Input data**]* Activity consisting of the baker reading the recipe, fetching the ingredients and baking the cake [**Process**]

Answer 145

1. **processes that stay in the background to handle some activity such as email, web pages, printing, and so on.**1. processes that handle interaction with (human) users and perform work for them.1. processes that create evil viruses that will try to steal your soul.1. processes that executes their program code in the kernel.

Answer 146

1. True1. **False**

Answer 147

1. Et sett av tråder er i en “race condition” når hver tråd i settet venter på en ressurs som er holdt av en annen tråd i settet.1. En prosess som prøver å få tilgang til en ressurs som er avslått.1. En prosess som står i ventekø, men ikke får tilgang på CPUen fordi andre prosesser har høyere prioritet.1. **To eller flere tråder som kjemper om en ressurs og utfallet avhenger av synkroniseringen av trådene.**

Answer 148

1. mellom 1 - 10 ganger raskere1. mellom 1 000 - 10 000 ganger raskere1. **mellom 10 - 100 ganger raskere**1. mellom 100 - 1 000 ganger raskere

Answer 149

1. **Multiprocessor systems have two or more CPUs sharing the same physical memory.**1. Multiprocessor systems have three or more processes running at the same time, because the OS is also a process1. Multiprocessor systems have two or more processes running at the same time.

Answer 150

1. True1. **False**

Answer 151

1. **True**1. False

Answer 152

1. At sider(pages) kun fjernes fra minnet når det er fullt og en ny side trenger plass.1. At brukeren selv får mulighet til å velge hvilken paging-algoritme systemet skal ta i bruk.1. **At sider(pages) kun lastes inn i minnet når det er bruk for dem.**

Answer 153

1. Removing pages from memory before the memory is filled up is called prepaging.1. **Loading the pages before letting processes run is called prepaging.**1. Loading the pages before the user can interact with the computer is called prepaging.

Answer 154

1. address bus1. **address space**1. address book1. home address

Answer 155

1. **En datamaskin kan ha flere virtuelle adresser enn det finnes fysiske adresser tilgjengelig på maskinen.**1. En datamaskin kan kun ha like mange virtuelle adresser som fysiske adresser.1. En datamaskin kan kun ha like mange eller færre virtuelle adresser enn det finnes fysiske adresser tilgjengelig på maskinen.

Answer 156

1. En sidefeil/pagefault er når det den prosessen som skal kjøres ikke er klar og hoppes over.1. En sidefeil/pagefault er når det skjer noe galt med svitsjingen av prosesser i minnet.1. En sidefeil/pagefault er når det skjer en feil ved avlesing av en fil fra harddisken.1. **En sidefeil/pagefault er når den siden som blir referert ikke befinner seg i det fysiske minnet og må bli hentet.**

Answer 157

1. Manage allocation of memory to processes.1. Manage pages such as defining page-sizes, allocate pages to processes and deleting pages that are not necessary.1. **Map virtual addresses onto the physical memory addresses.**

Answer 158

1. The Optimal page replacement algorithm.1. **The Second Chance Page Replacement Algorithm.**1. The Not Recently Used page replacement algorithm.1. The Clock Page Replacement Algorithm.1. The First-In,First-Out page replacement algorithm.

Answer 159

1. **True**1. False

Answer 160

1. True1. **False**______________________<__________________

Answer 161

1. True1. **False**

Answer 162

1. A bit telling if the file has caused a fault since it was loaded into memory.1. **A bit telling if the page has been modified since it was loaded into memory.**1. A bit telling if the page has been modified since the program was started.

Answer 163

1. **A page table is a table containing information about all pages belonging to a process and mapping between these virtual pages and the physical page frames. Every process has its own page tabl**e.1. A page table is one (often large) table containing information about all pages existing in the entire system, and mapping between these virtual pages and the physical page frames.1. A page table is a table containing information about only the pages that are currently in main memory, and mapping between these virtual pages and the physical page frames.

Answer 164

1. virtual addresses.1. **physical addresses.**1. No answer text provided.1. relative addresses.

Answer 165

1. **True**1. False

Answer 166

* I/O bit [ ]* Oversize bit [ ]* **Referenced bit [X]*** **Modified bit [X]*** Running bit [ ]* **Present/absent bit [X]**

Answer 167

* **to keep track of which part of memory that are in use. [X]*** to map virtual addresses into physical addresses. [ ]* **to deallocate memory when a process doesn’t need it anymore. [X]*** **to allocate memory to processes when they need it. [X]**

Answer 168

1. True1. **False**

Answer 169

1. Second Chance1. **FIFO (First In First Out)**1. LRU (Least Recently Used)1. NRU (Not Recently Used)

Answer 170

1. 301. 201. **50**1. 10_________________________

Answer 171

1. Swapping is when parts of the process is loaded into main memory temporarily to run before it is moved back on disk so another process can be in memory.1. Swapping is when a process is partially on one RAM-module and partially on one or more other ram-modules.1. **Swapping is when the entire process is loaded into main memory temporarily to run before it is moved back on disk so another process can be in memory.**

Answer 172

1. **True**1. False________________________________________

Answer 173

1. A special book for programmers that contains difficult computer terminology and definitions.1. An algorithm deciding which page to evict from the memory when a page fault occur.1. A protection-mechanism for assuring that a process cannot reference addresses outside its own address-space.1. **A hardware device for mapping virtual addresses to physical addresses without going through the page table so that heavily referenced pages can be accessed faster.**

Answer 174

1. **WSClock**1. Aging1. Working Set1. Second Chance

Answer 175

1. The amount of time that the CPU has actually been in use in total.1. The amount of actual time since the OS started.1. The amount of CPU time a process has actually used since it started.1. **The amount of actual time since the process started.**

Answer 176

1. The Not Recently Used page replacement algorithm.1. The Clock Page Replacement Algorithm.1. The First-In,First-Out page replacement algorithm.1. The Second Chance Page Replacement Algorithm.1. **The Optimal page replacement algorithm.**

Answer 177

1. A file used to model block-devices such as disks.1. **A file related to input/output devices that is used to model serial I/O devices such as printers and networks.**1. A file containing user information in binary.1. A file containing user information in ASCII.

Answer 178

1. “Physical dump” tar kun med seg filene som har blitt forandret siden sist og som også ligger i spesifiserte mapper.1. “Physical dump” tar backup av hele disken, men tar ikke med ødelagte og tomme blokker.1. **“Physical dump” tar backup av hele disken og skriver dermed også blokker som er ødelagte eller tomme/ledige.**1. “Physical dump” tar kun med spesifiserte mapper, men tar med seg alt innholdet i disse mappene.

Answer 179

1. **Hele stinavnet fra root og til filnavnet**1. Bare stinavnet fra den arbeidsmappen vi står i til filnavnet1. Bare filnavnet1. Hele stinavnet fra arbeidsmappen og til filnavnet

Answer 180

1. Charles strukturen1. **Hierarkisk struktur**1. Hickney strukturen1. To level struktur

Answer 181

1. It makes the files more resistant against viruses [ ]1. **It is simple to implement [X]**1. **The read performance is excellent [X]**1. It prevents disks of becoming fragmented [ ]

Answer 182

1. **That all the instructions up to the program counter (of the program issuing the interrupt request) have completed and none of those instructions beyond it has started.**1. That different instructions near the program counter(of the program issuing the interrupt request) are in different stages of completion.1. That the interrupt tells exactly what it want instead of asking the OS to determine what to do.

Answer 183

1. smart device1. **block device**1. character device1. hard device

Answer 184

1. **Programmert I/O (Programmed I/O)**1. I/O ved bruk av DMA1. Avbrudds-dreven I/O(interruptdriven I/O)

Answer 185

1. 58 sylindere = 348 msek1. **60 sylindere = 360 msek**1. 146 sylindere = 876 msek1. 87 sylindere = 522 msek

Answer 186

1. Avbrudds-dreven I/O(interruptdriven I/O)1. Programmert I/O (Programmed I/O)1. **I/O ved bruk av DMA**

Answer 187

1. A field is a local address-space belonging to one single process, while the record is a global address-space shared among all running processes.1. A record is just a basic element of data containing one value, but the field is a unit containing a collection of related records.1. **A field is just a basic element of data containing one value, but the record is a unit containing a collection of related fields.**1. A field contains information about all kind of files, but a record contains only information about audio-files.

Answer 188

1. **Sant**1. Usant

Answer 189

* **linked lists [X]*** single level directories [ ]* i-nodes [ ]* path names [ ]* **bitmaps [X]**

Answer 190

1. Interrupt mode1. Cycle stealing1. **Burst mode**1. Acknowledgement mode

Answer 191

* **DMAen har tilgang til bus’en uavhengig av CPUen. [X]*** **DMAen brukes for å redusere antall interrupts som går via CPUen. [X]*** Ved bruk av DMA trenger ikke CPUen å håndtere interrupts i forbindelse med I/O. [ ]* Ved overføring av data fra I/O til minnet jobber DMAen mye mer effektivt enn en CPU ville klart å få til. [ ]

Answer 192

1. Trolling1. Tracing1. Lagring1. **Spooling**

Answer 193

1. In the \dev\partition on UNIX systems and in the C://Windows/System32 on WIndows systems.1. In the end of the super block1. **In the end of the MBR(Master Boot Record)**1. In the end of the boot block

Answer 194

1. I-nodes1. Kontinuerlig tildeling(Contiguous allocation)1. **Lenket liste tildeling (Linked list allocation)**1. Lenket liste tildeling med bruk av minne-tabell

Answer 195

1. Smart devices and stupid devices1. **Block devices and character devices**1. Soft devices and Hard devices1. Bit devices and byte devices

Answer 196

Svaret er Sant1. Usant1. **Sant**

Answer 197

1. The device itself1. A Cathode Ray Tube1. A preamble1. **A device controller / an adapter**

Answer 198

1. Sant1. **Usant**____________________________________

Answer 199

1. Liten blokk-størrelse(block size) fører til at man sløser med både tiden og plassen.1. Stor blokk-størrelse(block size) fører til at man sløser med tiden, men ikke med plassen.1. Stor blokk-størrelse(block size) fører til at man sløser med både tiden og plassen.1. **Stor blokk-størrelse(block size) fører til at man sløser med plassen, men ikke med tiden.**

Answer 200

1. Hode(preamble), data, TCC (time check code)1. **Hode (preamble), data, ECC (error correcting code)**1. Data, adresse, ECC (error correcting code)1. Hode, Data, Adresse

Answer 201

1. **Programmert I/O (Programmed I/O)**1. I/O ved bruk av DMA1. Avbrudds-dreven I/O(interruptdriven I/O)

Answer 202

1. Ved å defragmentere harddisken så snart filer slettes for å komprimere området som skrives på.1. Ved å lagre alle filer tilhørende én prosess etter hverandre.1. Ved å lagre hele filer som fortløpende bytes (altså alt innhold i én fil etter hverandre).1. **Ved å å dele opp i blokker på bestemte størrelser og lagre dem der det er mest hensiktsmessig.**

Answer 203

1. 1 - kunne lagre store mengder informasjon, 2 - informasjonen må kunne overleve selv om man forsøker å slette den(papirkurv), 3 - bare prosessen som eier filen må kunne endre informasjonen.1. 1 - kunne lagre flere typer informasjon, 2 - informasjonen må overleve selv om den flyttes til andre lagringsmedier, 3 - informasjonen må kunne aksesseres på tvers av operativsystemer.1. **1 - kunne lagre store mengder informasjon, 2 - informasjonen må overleve selv om prosessen terminerer, 3 - flere prosesser må ha mulighet til å aksessere informasjonen samtidig.**1. 1 - kunne skille på filtyper for å avgjøre hvilket program som skal åpne filen 2 - kunne lagre informasjonen permanent, 3 - kunne lagre filer i et hierarkisk filtre.

Answer 204

1. I operativsystemet1. I harddisken1. I applikasjonen1. **I kontrolleren**

Answer 205

1. Faktisk overføringstid1. Rotasjonsforsinkelsen1. Hoppetiden1. **Søketiden**

Answer 206

1. Med memory-mapped I/O får kontroll-registerne til enhetskontrollerne(device controllers) tildelt en egen I/O port.1. Med memory-mapped I/O kan man aktivere caching av kontroll-registerne til enhetskontrollerne(device controllers) for å kommunisere mer effektivt med disse registerne.1. **Med memory-mapped I/O får kontroll-registerne til enhetskontrollerne(device controllers) tildelt en egen minneadresse.**

Answer 207

1. Det opprettes kun én i-node per prosess.1. **i-noden trenger kun å være i minnet når den korresponderende filen er åpen.**1. i-noden kan kun kjøres på Apple-produkter.1. alle i-nodene må alltid være i minnet.

Answer 208

1. That the information must be possible to access from all types of Operating Systems.1. **That the information must not be affected by process creation and termination.**1. That the information must be stored in a tidy order to ensure fast reading and writing.1. That the information must be possible to access from all UNIX-based Operating Systems.

ikt-102_operativsystemer_20231214230226 Flashcards

(232 cards)