Please enable JavaScript.

Coggle requires JavaScript to display documents.

old 471 notes (module 2), process - Coggle Diagram

- - - - no resources allocated to the process yet
      - process has not entered system
      - process enters this state when it is created
      - process in this state take up no space in memory
      - process ID is assigned here by the operating system
        
        process ID
        
        aka
        
        pid
    - - process can enter this state in 5ways
        
        process enters this state as soon as it is admitted to the system
        
        process enters this state after completing I/O or an event while in the waiting state
        
        process enters this state when it receives a system interrupt while in the running state/ is waiting for some other interrupt to occur
        
        interrupt
        
        cpu scheduler takes the
        
        a child process is forked and is executed, taking up the current CPU and placing its parent in the ready queue
        
        child process
        
        when created, is an exact copy of its parent,
        has the same values
        
        although they share code,
        they do not share data, as threads do
        
        once created, exists and runs independent of parent
        
        if parent is terminated, what happens tot he child?
        
        either dies or becomes an orphan
        
        1 more item...
        
        why is it called "forking" a child process?
        
        it runs alongside the parent
        
        a process's time slice for its CPU expired
      - once it enters the system, resources are allocated to the process
        
        which resources
        
        everything needed(memory) except the CPU
      - means
        
        ready to execute
    - - process enters this state when a CPU has been allocated to it by a CPU scheduler
        
        aka
        
        scheduler dispatch
        
        dispatcher module
        
        gives control of the CPU to the process selected by the short-term scheduler while a scheduler is selecting the next process to run
        
        process scheduling
        
        process of selecting one of the processes in
        the ready queue processor allocation and future execution
        
        ready queue
        
        queue of all PCBs belonging to processes in the ready state
        
        stored in operating systems portion of main memory
        
        affected by degree of multiprogramming
        
        degree of multiprogramming
        
        definition
        
        number of processes a single-processor system can handle efficiently
        
        determined by 5 things
        
        amount of memory available
        to be allocated to running processes
        
        program IO needs
        
        program CPU needs
        
        memory access speed
        
        disk access speed
        
        formula
        
        number of ready, waiting, running processes
        
        3 types of schedulers
        
        short-term scheduler
        
        aka
        
        cpu scheduler
        
        1 more item...
        
        long-term scheduler
        
        aka
        
        job scheduler
        
        1 more item...
        
        medium-term scheduler
        
        swaps out processes to reduce degree of mulitprogramming, releasing its memory and increasing available memory to be allocated to processes
        
        swap out
        
        3 more items...
        
        swap in
        
        3 more items...
        
        maintain a good mix of processes in the system,
        prevent cpu over/underuse, io device over/under use do to flood of cpu-bound or io-bound processes
        
        2 types of scheduling
        
        non-preemptive scheduling.
        
        currently executing process is stoppable and put to ready state, with the CPU getting assigned to another process
        
        default, unless specified otherwise
        
        preemptive scheduling.
        
        currently executing process is allowed to continue execution until it either terminates or goes to a wait state
        
        4 cases where scheduling decisions are made
        
        Switches from running to waiting state (non-preemptive)
        
        Switches from running to ready state (preemptive)
        
        Switches from waiting to ready (preemptive)
        
        Terminates (Non-preemptive)
        
        when a process:
        
        3 steps
        
        1 switching context
        
        time to do these things
        
        Dispatch latency
        
        1 more item...
        
        2 switch to user mode
        
        3 jump to location in user program to restart that program
      - means
        
        process is currently running on the CPU
        
        process is being executed
      - what can happen to a running process?
        
        the CPU scheduler can interrupt the process, sending the process back to the ready state and freeing up the CPU it was running on
    - - process enters this state when it receives I/O, event wait
        
        the processor's PCB is then entered into that I/O device's I/O queue
        
        I/O queue
        
        each IO device has one
        
        stored in the operating system's portion of main memory
      - process exits this state when I/O or event is completed
    - - process enters this state when it leaves the system
      - processes in this state take up no space in memory
      - any resources used by the process are freed after terminating
        
        types of resources used by process that are freed on termination
        
        files closed
        
        memory
      - 4 reasons for system to terminate a process automatically
        
        system couldnt deliver required system resources
        
        kill() command, or other unhandled interrupt
        
        parent can kill its child if the task given to it is no longer needed
        
        parent exits and does not want child to go on without it or operating system kills children by default
  - - - so that when execution is resumed, it can start back where it left off
    - - within a process control block
        
        process control block
        
        diagram
        
        consists of
        
        process state
        
        process number
        
        program counter
        
        address of the next instruction to be executed
        
        CPU registers
        
        contents of the registers prior to the last halting of a process
        
        memory-limits
        
        info about the memory allocated to the process
        
        list of open files
        
        files opened by the process
        
        ...
        
        other info needed to resume a process
        
        part of a process
- - - - cpu-based
        
        cpu burst
        
        consecutive time spent executing cpu based instructions
        
        cpu-io burst
        
        one instance of cpu burst followed by an io burst
        
        cpu-io burst cycle
        
        alternating between cpu and io bursts
      - io-based
        
        io burst
        
        consecutive time spent executing io based instructions
- - - - two methods
        
        message passing
        
        types of message passing/communication'
        
        non-blocking
        
        aka
        
        asynchronous
        
        nonblocking send
        
        message sent and next instruction is executed
        
        nonblocking receive
        
        receiver receives a message or a null msg if none have been sent
        
        blocking
        
        blocking send
        
        unless receiving process receives the message,
        the sender that sent the message to the receiving process cannot execute the next instruction until it has been received
        
        blocking receive
        
        process executing receive operation cant execute the next instruction unless it has something to receive
        
        aka
        
        synchronous
        
        direct
        
        asymmetric
        
        only sender must know name of receiver
        
        symmetric
        
        receiver must also know name of sender to receive a message from them
        
        indirect
        
        multiple processes can share mailboxes(ports),
        only one can read a message from a port at a time,
        by default the process creating the mailbox is the owner and only they have read access, but this can be transferred
        
        creating/deleting,sending to/from mailboxes requires
        
        1 more item...
        
        Automatic buffering
        
        msg buf allocated dynamicaly
        
        explicit buffering
        
        message buffer allocated directly
        
        buffering
        
        3 types
        
        zero capacity
        
        1 more item...
        
        bounded capacity
        
        1 more item...
        
        unbounded capacity
        
        1 more item...
        
        two kinds of messages
        
        receive
        
        process receives a message from a sender
        
        send
        
        sender sends a message to a receiving process
        
        requires 2 things
        
        os support
        
        system calls for every message transfer, making it slower
        
        2 advantages
        
        simpler to set up
        
        works well across multiple computers
        
        example
        
        client-server system
        
        can a client-server system use shared memory?
        
        no, because theyre on different machines
        
        3 client-server communication stategies
        
        Sockets
        
        2 more items...
        
        Remote procedure calls
        
        3 more items...
        
        Pipes
        
        3 more items...
        
        shared memory
        
        the memory is stored in the address space of a process but is accessible to other processes
        
        does not require
        
        system calls after setup, access occurs at memory speed, making it faster
        
        2 disadvantages
        
        complicated to set up
        
        doesnt work well across multiple computers
        
        better in what situation
        
        large amounts of information must be shared quickly on the same computer
- - - - modularity
        
        process can be broken up into cooperating modules with each one implemented as a process
      - information sharing
        
        different processes can use the same file
        
        ex
        
        pipelines
      - computation speed
        
        task may broken down into sub-tasks to be solved simultaneously
        (particularly when multiple processors are involved
      - convenience
        
        lets a single user multitask
- - - - heavyweight
      - single-threaded
  - - - process has multiple tasks to perform independently of one another, especially if one of them can block the others and we want them to go on without being blocked
    - - web server being able to fulfil multiple requests at once, without doing one after the other or forking new processes to handle each one
    - - Responsiveness
        
        prevents the blocking or slowness
        of one thread from slowing down the others
      - Resource sharing
        
        By default, threads share common code, data, and other resources, which allows multiple tasks to be performed simultaneously in a single address space.
      - Economy
        
        easier to make,manage, context switch threads compared to processes
      - Scalability
        
        execution of a mt app can be split
        across availabe processors
- - - - unit of cpu utilization
    - - kernal thread
        
        can execute one or more user threads
        
        associated with the operating system
        
        exist within os
        
        a single kernel thread only operate on a single cpu
      - user thread
        
        associated with the applications
        
        exist above the kernel
        
        written by app programmers
        
        user threads
        must be
        mapped to
        kernel threads
        
        3 models of thread mapping
        
        many-to-one model
        
        many user-level threads are all mapped onto a single kernel thread
        
        1 advantage
        
        thread mgmt handled by thread library in user space,
        making it efficient
        
        2 drawbacks
        
        if a blocking system call is made, then the entire process blocks, even if the other user threads would otherwise be able to continue.
        
        a single kernel thread can operate only on a single CPU, therefore individual processes cant be split across multiple CPUs
        
        One-To-One model
        
        separate kernel thread for each user thread
        
        2 advantages
        
        blocking system call doesnt stop entire process
        
        individual processes can be split across cpus
        
        1 drawback
        
        more overhead, slows system
        
        many-to-many
        
        any number of user threads to equal or smaller number of kernels
        
        3 advantages
        
        3 more items...
    - - program couinter
      - stack
      - set of registers
    - - POSIX Pthreads
        
        either user or kernel level,
        extends POSIX standard thread library
      - Win32 threads
        
        kernel level, windows
      - Java threads
        
        uses posix or win32 depending on system
- - - - create a child process
      - create a thread