List of questions (ITI8510)

Real-Time Operating Systems and Systems Programming

Basic Linux command-line and utilities for accessing it Standard Streams Stack Exploration Exploring Debugger Signals Threads Processes File Attributes Reading Binary Files LAB: Hufmann Encoder Network Sockets Optimization (mmap() helps here)

[ @ ]

This list of questions serves as a pool of questions that might be present at the exam. Your mileage may vary. The image is illustrative. Items in mirror may be closer than they appear.

Lecture 1: Introduction

• What are real-time systems?
• What are the main characteristics of real-time systems?
• How did priority inversion affect Pathfinder rover in 1997?
• What is priority inheritance?
• Describe the difference between synchronous and asynchronous events
• How does state based system differ from event based system?
• What is deterministic real-time system?
• What is the difference between soft, firm and hard real-time systems?
• When workload changes but resources remain constants: is it possible to create statically predictable real-time system?
• What are the possibilities for certifying a dynamic real-time system?
• What is latency?
• What system constraints can create design conflicts apart from real-time requirement?

Lecture 2: Real-Time Operating Systems

• Why are operating systems created?
• Describe the possible services for operating systems.
• What services are usually not included in real-time operating systems?
• Why is predictablility, for RTOS, more important than speed?
• What is re-entrancy and why is it important for RTOS libraries?
• Why is buying a RTOS considered better than writing one?
• What is POSIX?
• What are the problems for Linux as RTOS? Benefits?
• How does RTLinux run RT tasks?
• Describe possible scheduling strategies. Which are feasible for RTOS?

Lecture 3: Input / Output

• What is memory mapped hardware access?
• For memory mapping: how do you write into a device's registers?
• What is IO mapped hardware access?
• What does port polling mean?
• Explain the difference between blocking and nonblocking IO operations
• Why is IO blocking needed at all?
• How is select() function used?
• How do interrupts help IO?
• What are the architectural requirements for CPU to support interrupts?
• What happens when interrupt line activates on CPU?
• Why can interrupts be generated internally in CPU?
• Interrupt types: Polling, Interrupt Vector Table (IVT), Programmable Interrupt Controller (PIC)
• How vector interrupts work? What is Interrupt Vector Table (IVT)
• Explain the first and last actions of an Interrupt Service Routine.
• How can interrupt priorities be handled in software?

Lecture 4: Understanding memory (Stack)

• What kind of memory is faster to access than RAM?
• What are the possible operations of a stack data structure?
• Write a short possible stack implementation in C
• What are the hardware requirements for stack implementation?
• What characteristics of assembly language make it unpopular?
• What is the purpose of registers %esp and %ebp on IA32 architecture?
• Can you refer to memory addresses in assembly?
• What do the assembly commands "push %eax"; "pop %edx" do?
• Explain the contents of a non-active stack frame
• How are call, leave and ret operators used?
• In general: while it's good to know the stack on operator level, general understanding of its workings should be sufficient. See slides "Recap" and "Stack frame"
• Why are some registers Caller save and some Callee save registers?
• Explain some of the ways in which call stack affects the behaviour of C programs.
• Explain how overflowing buffers are exploited by hackers.

Lecture 5: Understanding memory (Heap)

• What does "brk" pointer hold?
• Why is dynamic allocation avoided for Real-Time Systems?
• How can fragmentation reduce the amount of free memory?
• In theory you could solve fragmentation by moving memory contents. Why it is not done?
• How are block headers hidden from modification?
• Why can we use 3 lower bits of block size in header to store data?
• Describe first fit; next fit; best fit allocation strategies
• Sometimes it is necessary to join memory blocks, when is it done?
• Why are free memory blocks sometimes segregated?
• What is Garbage Collection?

Lecture 6: Interrupts, Signals, Threads

• How is protected mode related to interrupts?
• Why are interrupts considered synchronous and asynchronous?
• Why are interrupts sometimes blocked?
• Explain access serialization as a method for data protection
• What are system calls and how are they related to interrupts?
• Explain how signals are different from and similar to (hardware) interrupts?
• Do blocked signals wait or are they discarded?
• What are the possible registered default actions for a signal?
• Is it possible for several signals of the same type to arrive while blocked or pending?
• What are the common patterns for writing signal handlers?
• Threads: should only know what is a thread, since they were used as a tool for explaining data sharing and critical sections for concurrent programming.
• What is a race condition?
• What is a mutex?
• How can a mutex solve race conditions?

Lecture 7: Scheduling

• Are the goals for scheduling - processor usage maximization and response time maximization mutually exclusive?
• What does it mean for scheduling to be preemptive; non-preemptive?
• What is static scheduling?
• What are the problems of round-robin scheduling?
• How does rate monotonic scheduling work?
• Can you find cases in which Earliest Deadline First is better than Rate Monotonic Scheduling?
• How does a simple cyclic executive work?
• What are the three intuitive priority levels for cyclic executive?
• What can be done if you have a task which takes 100ms and time-slot is only 60ms and the task exeeds its time slot?
• What happens if a POSIX process is set to be scheduled with SCHED_RR; SCHED_FIFO?
• Can you set priorities for tasks which are scheduled with SCHED_OTHER
• How can paging affect real time processes and what can be done to avoid it?

Lecture 8: Programming an Operating System

• What would be the minimal data that is stored for a suspended task?
• What are the three common states for tasks?
• What is a scheduling point?
• What is the purpose of OS clock tick?
• What is stored in a ready list?
• Why is there an idle task?
• How does context switch work?
• Why one cannot write context switching without assembly code?
• Describe locking/setting of a mutex which uses a waiting list.
• What is a deadlock?

Lecture 9: Optimization

• Why is optimization not the default option for compilers?
• Describe some of the limitations for optimization.
• Why is cycles per element better way of measuring optimization than running time?
• Explain the reasoning behind following optimizations:
  • Move calculations out of loop
  • Reduce function calls
  • Extra storage variables in loops
  • Loop unrolling
  • Pointer access for arrays
• How can knowledge of CPU pipelining help optimization?
• How can cache and branch prediction affect optimization.
• What is code profiling?
• Why is C99 standard suggesting use of inline keyword in stead of a macro?
• What methods can you use to reduce:
  • code size?
  • memory usage?
  • power consumption?
• Describe problems which optimization can introduce.

Lecture 10: Networking

• Describe the difference between Unix IO and Standard IO functions.
• What are the main functions for Unix IO?
• There are 3 structures for storing open files in a system kernel: what are they?
• How many bytes are there in an IP-address?
• What is the purpose of DNS services?
• Is network byte order little- or big-endian? What does it mean?
• What function is used for domain name retrieval?
• What is a socket?
• What does a socket address store?
• What is the sequence of functions for:
  • creating and connecting a socket to a computer?
  • creating and listening on a socket to create a server?
  • Why is it necessary to create a handler for EPIPE signal when working with sockets?

Lecture 11: Getting to know the hardware

• What general steps are needed for undestanding the hardware?
• Why are hardware component datasheets interesting for a programmer?
• Describe some of the things an assembly programmer should know about the processor.
• Why is it necessary to use assembly in initialization code?

Lecture 12: Real-Time Operating Systems

• Which of the following are very popular RTOS-es? (w/ some list)
• What is the background for OSEK/VDX standard?
• What is a priority ceiling?
• Nucleus has a very uncharacteristic feature for an RTOS, what is it?
• VxWorks is a pun on another operating system, which?
• How does VxWorks ease development of RT-systems?
• How did uC/OS-II start and why is it popular?