Introduction to 5143

1
Introduction to 5143

• Key Concepts and Terms
• def. of embedded system
• def. of real-time system
• how the 2 are related
• controlled system
• open/closed-loop control
• hard vs soft real-time

• Class Structure
• topics covered
• schedule
• reading
• assignments
• grading
• lab work

2
Distinctions of Embedded Software Development

• Hardware knowledge
• Efficient code
• Peripheral interfaces
• Robust code
• Minimal resources
• Reusable software
• Development tools

3
What is a real-time system?

• A computer system with strict timing constraints.
• It must provide
• the right answer
• at the right time
• or suffer
• death and destruction!
• (well, at least serious system malfunction)

4
The Real-Time Blanket

• Real-Time is an added constraint
• placed on top of every component of a system
• from the hardware to the application.

5
Overall Class Objective

• Good high-level understanding
• of real-time, embedded systems.
• What are the critical issues during
• design, implementation, and testing?
• Low-level Hardware Control
• Concurrent Programming
• Real-Time Scheduling
• System Modeling
• Worst-Case Execution Time Analysis (WCET)
• Hands-on Experience

6
More specifically .
7
Embedded Systems and UM003
Key Concepts and Terms

• controlled system
• closed and open loop control basic system
  elements
• variations of RAM and ROM
• basics of schematics
• datasheets
• firmware

• cross-compilation
• host machine
• target platform
• ports
• in-circuit emulator
• debug monitor
• on-chip debug

8
Hello World!

• Compiling
• Linking
• Locating
• Downloading
• Executing
• Debugging
• User Interface

9
Debugging Techniques
in-circuit emulator
meta- code
HOST
logic analyzer
TARGET
on-chip debugger
code
Simulator
Registers ALU
code
oscilloscope
UI with comm
Debug Monitor
LEDs
Hello World
10
I/O and Peripherals
Key Concepts and Terms

• control and status registers
• ports
• bitmasking
• bit manipulation
• device drivers
• memory mapped I/O
• bandwidth
• synchronization
• polling
• busy waiting

• serial and parallel interfaces
• analog and digital signals
• analog to digital conversion
• sampling frequency
• active high, active low
• edge-triggered
• level triggered
• timing diagram

11
I/O and UM003

• I/O
• Ports
• Device Drivers and Interfaces
• Control and Status Registers
• Synchronization Techniques
• Signal Processing
• Signal Types
• Sampling
• A/D Conversion

12
Read / Write to Ports

• Memory-Mapped I/O
• Port-Mapped I/O
• DMA

out(PORTA, 0x33) or PORTA 0x33
13
Bit Manipulation

14
Control Status Register Pins/Ports
iom128.h / Port B Data Register - PORTB
/ define PB7 7 define PB6
6 set bit7, bit6, bit0 hi (equivalent to
0xC2) DDRB (1ltltPB7) (1ltltPB6) (1ltltPB0)
read 4 most significant bits data PINA 0xF0
15
Relating to the outside world

• 4 categories of interfaces
• parallel (digital, multiple bits at a time)
• serial (digital, one bit at a time)
• analog (voltage, continuum)
• time (analog period, frequency, )
• Types of Signals (signal processing world)
• continuous-time (analog)
• discrete-time (sampled analog signal or
  inherently discrete signal)
• continuous-value
• discrete-value
• random (noise)

16
Synchronization Techniques(Im just waiting on a
friend)

• blind cycle open door at precisely 204 pm
• busy wait (gadfly) stand at the door
• interrupt keep busy until doorbell rings
• periodic polling doorbell broke, check every 5
  minutes
• DMA friend lets herself in

17
Peripherals and I/O

• Ports
• Device Drivers and Interfaces
• Control and Status Registers
• Synchronization

18
Signal

• a time varying physical phenomenon which is
  intended to convey information.

Nigel Others only go to 10. We can go to 11 1
louder. Interviewer Why dont you just make 10
louder and make 10 be the top number. Nigel
Long pause. These go to 11.
19
Sampling a Sinusoid(How often, is often enough?)

• Shannons Sampling Theorem
• If sampled at a rate more than twice the highest
  frequency (of the component sinusoids)
• it can be exactly reconstructed.
• fm Nyquist Frequency highest frequency
• 2fm Nyquist Rate

20
Sampling a Sinusoid(How often, is often enough?)
21
Sampling a Sinusoid(How often, is often enough?)
RTS Fig. 1.2 p. 4
22
Concurrency, Tasks, and Intertask Communication
Key Concepts and Terms

• semaphore
• context switching
• atomicity
• tasks
• task control block
• threads
• scheduler
• scheduling point
• preemptive scheduler
• nonpreemptive scheduler
• blocked

• synchronization
• intertask communication semaphore
• counting semaphore
• message queue
• event
• signal
• producer-consumer problem
• critical section
• double buffer
• ring buffer

23
Concurrency, Tasks, and Intertask Communication

• Introduction
• Interrupts and Concurrency
• Sharing CPU
• Scheduling
• Role of Scheduler
• Task Control Block
• Preemptive Scheduler
• Scheduling Points
• Tasks
• Defining
• Task dependencies
• Communication and Synchronization
• Semaphores
• Buffers and Queues
• Events
• Signals
• Examples and Constructs in C

24
TCB Task Control Block Data Structure OS uses
to keep track of threads
ReadyQ BlockedQ SpawnedQ
25
Task Dependencies

• data dependencies
• shared data
• computational intensity
• functional dependencies
• sequential cohesion
• functional cohesion
• time dependencies
• critical and urgent activities
• varying periodicity
• temporal cohesion

26
Communication and Synchronization

• Communication (Data Sharing)
• Message Queues
• Mailboxes
• Buffers (semaphores for access control)
• Double Buffer
• Ring Buffer
• Synchronization (Sequencing)
• Semaphores
• Events
• Signals

27
Real-Time Scheduling
Key Concepts and Terms

• round robin
• weighted round robin
• cyclic executive
• job
• task
• periodic task
• sporadic task
• aperiodic task
• periodic task model
• table-driven
• clock-driven
• time slice
• static scheduling
• dynamic scheduling
• frame-based scheduling
• WCET
• jitter

• execution time
• release time
• phase
• period
• deadline
• relative deadline
• feasability
• utilization
• response time
• hyperperiod
• harmonic task set
• optimality
• priority-driven schedule
• state-driven
• polled loop
• slack

28
Not quite a RTS Kernel

• Pseudokernel
• Polled Loop
• Cyclic Executive
• State-Driven Code
• Typical nonRT approaches
• Round Robin
• Weighted Round Robin
• Foreground / Background
• Self-Scheduling

29
Table-Driven Scheduling

• Assumptions
• Known, fixed tasks.
• Known periods and execution times (hard
  real-time).
• Job available at start of period.
• Aperiodic executed during idle times.
• Approach
• Define hyperperiod.
• Define schedule.
• Create table with release times relative to start
  of hyperperiod.
• No error checking.

30
Frame-Based Scheduling

• Scheduler for frame-based is periodic.
• Scheduling decisions made at set intervals (start
  of a frame).
• Schedule table is made up of blocks (all jobs to
  run in a given frame).
• Attempt to simplify and standardize scheduling.
• Error checking performed at start of frame.
• Frame size carefully selected to meet criteria.

31
Aperiodic and Sporadic Jobs

• Aperiodic (soft deadline)
• Scheduled when processor idle as FIFO (no
  acceptance test).
• Slack Stealing (scheduled before periodic tasks).
• Can calculate average response time.
• Sporadic (hard deadline)
• Consider EDF (Earliest Deadline First) Scheduling
• Job with earliest absolute deadline has highest
  priority.
• Acceptance Test
• Enough slack?

32
Priority-Driven Scheduling

• Key Concepts and Terms
• EDF
• RM
• DM
• LST
• schedulability test
• schedulable utilization
• optimality
• critical instant
• time-demand analysis
• busy interval
• general schedulability test

33
Priority Schemes

• Rate Monotonic Assignment (RM)
• Rate 1/p.
• Higher rate Higher priority.
• Deadline Monotonic Assignment (DM)
• Relative deadline.
• Shorter deadline Higher priority
• DM RM when D p.
• Earliest Deadline First (EDF)
• Absolute deadline.
• Closer deadline Higher priority
• Least Slack Time First (LST)
• Slack d t x.
• Smaller slack Higher priority.

34
Schedulability (Breakdown) Utilization

• Schedulability tests when D p
• EDF
• necessary and sufficient condition.
• RM
• sufficient condition

35
Time-Demand Function
36
General Time-Demand Analysis

• Can all jobs in Ti meet their deadline?
• for each task Tk in ?i
• compute wk1(t)
• if all wk1(t) ? pk
• check wi1(t) ? t ? Di
• if ji1 meets deadline
• feasible
• else
• not feasible
• else
• compute length (tl) of in-phase level-?i busy
  interval
• compute max response of ALL ceil(tl/pi) jobs of
  Ti if all meet deadlines
• feasible

37
Scheduling of Aperiodic and Sporadic Tasks

• Key Concepts and Terms
• Periodic Server
• Bandwidth-Preserving Servers
• Deferrable Server
• Consumption Rules
• Replenishment Rules
• Sporadic Server
• Constant Utilization
• Total Bandwidth Server
• Weighted Fair-Queueing Server
• Slack Stealing

38
Objectives of Scheduling Algorithm

• Correct Schedule
• Response Time is minimal
• Scheduling of sporadic is optimal
• Validation is relatively easy

39
Scheduling of Aperiodic ONLY

• Aperiodic Scheduling Schemes
• Background (non-optimal)
• Interrupt-Driven (not correct)
• Slack Stealing Algorithm (high overhead)
• Polling Server (non-optimal)
• Bandwidth Preserving Server

40
Scheduling Aperiodic Sporadic Jobs

• Aperiodic Jobs
• Polling
• Bandwidth Preserving Server
• Deferrable Server
• Sporadic Server
• Constant Utilization and Total Bandwidth Servers
• Weighted Fair-Queueing Server (will not be
  covered)
• Slack-Stealing Algorithms (will not be covered)
• Sporadic Jobs
• EDF
• acceptance based on density
• Fixed Priority Scheduled by Bandwidth Preserving
  Server
• acceptance based on slack

41
Acceptance for Sporadic Tasks in Deadline-Driven
System

• Density of periodic jobs
• Density of sporadic job
• Acceptance of first sporadic job

42
Acceptance for Sporadic Tasks with Fixed-Priority

• Acceptance Test for first sporadic task
  S1(t,d1,e1)
• Acceptance Test for any sporadic task Si(t,di,ei)
• enough slack ?
• make any lower priority sporadic tasks miss
  deadline?

43
Two-Level Scheme for Integrated Scheduling

• Application Servers (level)
• Schedule subset of all tasks
• May use any priority scheme
• Runs on slower virtual processor
• OS Scheduler (level)
• replenishes budgets of application servers
• sets deadlines of application servers
• schedules application servers using EDF

44
Priority Scheduling

• Assigning Priority
• task-level fixed (RM, DM)
• task-level dynamic, job-level fixed (EDF)
• Schedulability Test
• Schedulable (Breakdown) Utilization
• Time-Demand Analysis
• Assumptions
• Preemptable at any time.
• No self-suspension.
• Independence.

45
Consequences of Blocking

• Job waits (self-suspends) until resource
  available,
• thus it is blocked by another job.
• If you use self-suspension exclusively
• to control access to resources
• BAD THINGS CAN HAPPEN
• Unbounded priority inversion wait is
  potentially infinite.
• Deadlock job A waits for job B, job B waits for
  job A.

46
Sharing Resources Control Protocols

• Resource Access Control Protocols
• Nonpreemptive Critical Section (NPCS)
• Basic Priority-Inheritance Protocol
• Basic Priority-Ceiling Protocol
• Stack-Based, Priority-Ceiling (Ceiling-Priority)
  Protocol
• Distinguished by
• Scheduling Rule
• Allocation Rule
• Priority Rule
• Implementation
• Priority Algorithm (fixed or dynamic)
• Simplicity of implementation (a priori
  information)
• Addresses concerns
• Unbounded Priority Inversion
• Deadlock
• Blocking Time

47
Assigning Tasks (Subtasks)

• How many of each processor type?
• What is the function of each processor?
• How are tasks and jobs defined?
• How do you synchronize tasks (precedence
  constraints)?
• How do you control shared resources across
  processors?
• What scheduling algorithm should you use?
• How do you optimally assign tasks to processors?
  (NP-Hard)
• Frame as bin-packing (Ignore communication and
  resources).
• Frame as integer linear-programming problem
  (Ignore resources).
• Frame as graph-partitioning.

48
Local and Remote Resources
49
Scheduling the Tasks

• How do you control shared resources across
  processors?

50
Power Management and Embedded Systems

• Importance of power
• Power Estimation
• Power Management
• Hardware
• Software
• Power Aware Computing, MF Jacome and A
  Ramachandran in Embedded Systems Handbook
• Power Management Techniques for Real-Time
  Embedded Systems, S. Gary. Texas Instruments
  White Paper

51
Why save power?

• Effects of power usage
• Dynamic and Static power
• Power usage

52
WCET

• definition
• upper and lower bounds
• safe and tight

53
Methods and Goals

• Static Analysis
• Measurement-Based
• Goals
• affordable
• sufficiently precise
• tractable
• easy

54
Problems in WCET Analysis

• Path Problem
• infeasible paths
• halting problem
• longest path

55
Problems in WCET Analysis

• State Problem
• intractability
• local vs global analysis
• divide-and-conquer
• context (TRPS timing-relevant processor state)
• control-flow interference
• data memory interference
• timing anomolies

56
References

• Formal Methods
• Ten commandments of formal methods... Ten years
  later (Bowen, Jonathan P. and Hinchey, Michael
  G.) Computer, v 39(1), January, 2006, p 40-48.
• Statecharts
• Describing Behaviors Using Statecharts
  (Jonathan M. Kaye and David Castillo) Flash MX
  for Interactive Simulation (Ch. 7).
  www.flashsim.com, 2003, p. 75-101
• STATECHARTS A Visual Formalism for Complex
  Systems (Harel, David) Science of Computer
  Programming, v 8(3), 1987, p 231-274.

57
Formal Methods

• A formal specification language provides
• a notation (its syntactic domain),
• a universe of objects (its semantic domain)
• a rule defining which objects satisfy each
  specification.
• It provides the means of proving
• a specification is realizable
• a system is correct
• properties (behavior) of a system without running
  it

58
Statecharts

• Statecharts state-diagrams depth
  orthogonality broadcast communication
• state-diagrams
• depth
• orthogonality
• broadcast communication
• AND
• permits rapid and robust coding
• provides excellent documentation and validation
  tools

59
Petri Nets

• Graphical and Mathematical Model of Processes
• places, transitions, tokens, markings
• coverability tree
• colored petri nets
• timed petri nets