Concurrency in Ada

1
Concurrency in Ada

• Programming Languages 1
• Robert Dewar

2
Concurrency in Ada

• What concurrency is all about
• Relation to operating systems
• Language facilities vs library packages
• POSIX threads
• Ada concurrency
• Real time support
• Distributed programming

3
What Concurrency is all About

• Multiple threads of control within one program
• Fairly closely coupled (single address space)
• Each thread is independent
• But can syncrhonize with other threads
• One program has many threads

4
Relation to Operating Systems

• Typical Unix systems provide
• Multiple processes
• separate address spaces, separate scheduling
• Light weight processes/kernel threads
• shared address space, separate scheduling
• User level threads
• shared address space, no separate scheduling

5
Language Feature vs Libraries

• Library approach takes a standard sequential
  language, e.g. C
• And provides a set of packages that provide
  concurrency
• The C program makes calls to the library to
  create tasks etc.

6
Problems with Library Method

• Libraries may not be completely well defined and
  may not be portable
• Language was not defined with concurrency in mind
• e.g. are library routines thread-safe
• are constructs well defined
• e.g. what is rule with shared variables?

7
Thread Safety

• Suppose two threads of control both call a
  routine such as malloc.
• In the middle of one malloc call, the thread is
  interrupted by a higher priority thread, or
  reaches end of time slice.
• Another task calls malloc
• Chaos???

8
Shared Variables

• Suppose we have a global variable a
• One task does VAR 1
• Another task does VAR 256
• Again we have interrupts causing statements to
  get intermingled
• Is result well defined? or could we end up with
  VAR having the value 257?

9
POSIX Threads

• A standardized package of thread primitives
• create thread
• timer functions
• synchronization mechanisms
• etc.
• Several versions
• Lots left undefined

10
Add Concurrency to Language

• Algol-68 had simple semaphores and the notion of
  separate tasks.
• CSP, not really a PL (although consider OCCAM
  derived from CSP) had a simple channel mechanism
• Simula-67 identified tasks with objects

11
More on CSP (OCCAM)

• Program consists of processes and channels
• Process is code containing channel operations
• Channel is a data object
• All synchronization is via channels

12
Channel Operations in CSP

• Read data item D from channel C
• D ? C
• Write data item Q to channel C
• Q ! C
• If reader accesses channel first, wait for
  writer, and then both proceed after transfer.
• If writer accesses channel first, wait for
  reader, and both proceed after transfer.

13
Tasking in Ada

• Declare a task type
• The specification gives the entries
• task type T is entry Put (data in Integer)
  entry Get (result out Integer)end T
• The entries are used to access the task

14
Declaring Task Body

• Task body gives actual code of task
• task body T is x integer -- local per
  thread declarationbegin accept Put (M
  Integer) do end Put end T

15
Creating an Instance of a Task

• Declare a single task
• X T
• or an array of tasks
• P array (1 .. 50) of T
• or a dynamically allocated task
• type AT is access T
• P ATP new T

16
Task execution

• Each task executes independently, until
• an accept call
• wait for someone to call entry, then proceed with
  rendezvous code, then both tasks go on their way
• an entry call
• wait for addressed task to reach corresponding
  accept statement, then proceed with rendezvous,
  then both tasks go on their way.

17
More on the Rendezvous

• During the Rendezvous, only the called task
  executes, and data can be safely exchanged via
  the entry parameters
• If accept does a simple assignment, we have the
  equivalent of a simple CSP channel operation, but
  there is no restriction on what can be done
  within a rendezvous

18
Termination of Tasks

• A task terminates when it reaches the end of the
  begin-end code of its body.
• Tasks may either be very static (create at start
  of execution and never terminate)
• Or very dynamic, e.g. create a new task for each
  new radar trace in a radar system.

19
The Delay Statement

• Delay statements temporarily pause a task
• Delay xyz
• where xyz is an expression of type duration
  causes execution of the thread to be delayed for
  (at least) the given amount of time
• Delay until tim
• where tim is an expression of type time, causes
  execution of the thread to be delayed until (at
  the earliest) the given tim

20
Selective Accept

• Select statement allows a choice of actions
• select entry1 () do .. endor when
  bla entry2 ()or delay ...ddd...end select
• Take whichever open entry arrives first, or if
  none arrives by end of delay, do dddstmts.

21
Timed Entry Call

• Timed Entry call allows timeout to be set
• select entry-call-statement or delay
  xxx end select
• We try to do the entry call, but if the task
  wont accept in xxx time, then do the delay stmts.

22
Conditional Entry Call

• Make a call only if it will be accepted
• select entry-call ..else statementsend
  select
• If entry call is accepted immediately, fine,
  otherwise execute the else statements.

23
Task Abort

• Unconditionally terminate a task
• abort taskname
• task is immediately terminated
• (it is allowed to do finalization actions)
• but whatever it was doing remains incomplete
• code that can be aborted must be careful to leave
  things in a coherent state if that is important!

24
Asynchronous Transfer of Control

• Execute section of code, aborting after specified
  time or event
• select delay or accept statementthen
  abort statementsend select
• Statements start executing and are immediately
  aborted if delay or accept completes.

25
Shared Variables

• A shared variable is one accessed by more than
  one task
• if variable is declared atomic, no restrictions
• otherwise, we cannot have two tasks access the
  same variable without synchronizing (e.g. doing a
  rendezvous).
• model is that variables can normally be in
  registers

26
Tasking Is Completely General

• Any possible synchronization problem can be
  solved using the rendezvous
• We know this because it is more powerfulthan
  CSP/Occam which is itself general
• That means that any synchronization primitive can
  be simulated using the RV

27
An Example, the Semaphore

• The Idea of a (binary) semaphore
• Two operations, p and v
• p grabs semaphore or waits if not available
• v releases the semaphore
• Monitor is
• p (sem) statementsv (sem)

28
A Semaphore using a Task, RV

• The specification
• task type Semaphore is entry p entry vend
  Semaphore

29
A Semaphore using RV

• The body of semaphore is very simple
• task body Semaphore isbegin loop accept
  p accept v end loopend Semaphore

30
Using the Semaphore Abstraction

• Declare an instance of a semaphore
• Lock Semaphore
• Now we can use this semaphore to create a
  monitor, using
• Lock.P code to be protected in monitorLock.V

31
The RV Semaphore

• Very neat expression
• Nice high level semantics
• But awfully heavy if a real task is involved
• A case of abstraction inversion
• We expect to see tasks implemented in terms of
  low level stuff like semaphores, not the other
  way round.

32
Protected Types and Objects

• A protected type is a data object with locks
• specification provides data, like a record, and
  the locked access routines
• functions (read the data with read lock)
• procedures (read/write the data with write lock)
• entries (wait till some condition is met, then
  read/write the data with write lock)

33
Protected Types and Objects

• There is conceptually no separate thread of
  control.
• Body provides code for the functions, procedures
  and entries
• These are executed in the calling thread after
  obtaining necessary locks

34
Semaphore Using Protected Type

• Specification has the data and entries
• protected type Semaphore is entry P procedure
  Vprivate Grabbed Boolean Falseend
  Semaphore
• P is an entry since we may have to wait

35
Protected Type Semaphore

• The body provides the code of P and V
• protected body Semaphore is entry P when not
  Grabbed is begin Grabbed True end
  P procedure V is begin Grabbed
  False endend Semaphore

36
Using the Protected Type Semaphore

• Declare an instance of a semaphore
• Lock Semaphore
• Now we can use this semaphore to create a
  monitor, using
• Lock.P code to be protected in monitorLock.V
• Note this was cut and paste from the task slide

37
Requirements for Real Time

• Eliminate non-determinism
• pragma Dispatching_Policy (FIFO_Within_Priorities)
  
• means run till blocked, no time slicing
• reduces non-determinism
• typical of real time threads, e.g. in NT
• Define priorities of tasks
• exact specs for how priorities are respected
• Define queuing protocols
• first in, first out, or by priority of caller

38
Importance of Priorities

• Proper priority assignment important
• Used to ensure important tasks are completed
• Used to ensure system is schedulable
• Example Monotonic Scheduling
• Collection of cyclic tasks
• Require servicing at fixed interval
• Assign highest priority to shortest cycle
• Regardless of importance of tasks
• Ensures schedulability if possible

39
Priority Inheritance

• Guard against priority inversion
• low priority task grabs resource X
• high priority task needs resource X, waits
• medium priority task preempts low priority task,
  and runs for a long time, holding up high
  priority task
• Solution, while high priority task is waiting,
  lend high priority to low priority task