type of os:
- batch
- time sharing
- single user
- multi user
- distributed
- real time
- general purpose
interrupt
- System Calls
- System Programs
- OS Design
- policy vs mechanisms
critical modules that require high performance
- interrupt handler
- i/o manager
- memory manager
- cpu scheduler
two common models of ipc (inter process communication)
- message passing
- shared memory
virtual connections between processes, users, computer systems
Background Services are called services / subsystems / daemons
Operating Systems Structure
- MS-DOS
- Unix
Kernel Provides
- File System
- CPU Scheduling
- Memory Management
Layered OS
- One Problem - Overhead system calls
- Another Problem - Designing Layers Order
- Fewer Layers with More Functionality
MicroKernel - All Non Essential Parts are moved to system programs and User Space
Mac OSX Darwin is also partly based on the Mach MicroKernel
Loadable Kernel Modules - Linking service dynamically
- Architecture of Mac OSX
- Architecture of iOS
- Architecture of Android
OS Debugging
- Log File
- Core Dump
- Crash
- Crash Dump
What Kernel Do in unrecoverable situation?
Save Kernel's Memory state to a section of disk that does not have any FileSystem. When Kernel encounter a unrecoverable error, it take it's entire content of memory or kernel owned part and dump it into disk section. when booting, a process runs and tries to gather the memory dump from the disk and save it as file in filesystem
Trace Listing
System Generation
- CPU
- Boot Disk
- Memory
- Device
- OS Options
System Boot
- initial bootstrap program
job == process == cpu_activities
-
processs in memory
- text (code/instructions)
- data (global vars)
- heap (dynamically allocated)
- stack (temporary data - function params, return addresses, local vars)
-
a process itself can be execution environment
- jvm is a process (which takes java code and take actions via native machine instructions on behalf of that code) job == process == cpu_activities
Process State
- new - created process
- ready - ready to be assigned to processor
- running - instructions are being executed
- waiting - waiting for an event to occur (I/O Completion or reception of a Signal)
- terminated - finished execution
Scheduling Queues
- job queue
- ready queue
- device queue
once the process is allocated to CPU one of the following events may occur
- process issuing I/O request -> I/O Queue -> I/O Happens -> then it placed into the Ready Queue Again (does it stay all the time in CPU if not is it immediately moved to Ready Queue?)
- Time Slice Expired -> Moved to Ready Queue
- Fork A Child -> wait for child's termination
- Wait for an interrupt -> Interrupt Occurs -> then it process from Ready Queue
Process Termination
-
removed from all queues
-
PCB(process control block) and resources deallocated
-
single threaded vs multithreaded process
When a request is made, rather than creating another process, the server creates a new thread to service the request and resume listening for additional requests.