Scheduling and Resource Access Protocols: Basic Aspects

1
Scheduling and Resource Access Protocols Basic
Aspects

• CS 5270 Lecture 3

2
A Classic Policy

• Rate Monotonic Scheduling.
• Task set J1, J2, , Jn
• Each task is periodic. T1, T2,.., Tn
• ?i 0 for each i.
• Di Ti for each i.
• Pre-emption allowed.
• Only one processor
• No precedence constraints
• No shared resources.

3
RMS

• The RMS algorithm
• Assign a static priority to the tasks according
  to their periods.
• Priority of a task does not change during
  execution.
• Tasks with shorter periods have higher
  priorities.
• Preemption policy
• If Ti is executing and Tj arrives which has
  higher priority (shorter period), then preempt Ti
  and start executing Tj.

4
RMS Example
(3, 2)
(5, 1)
5
RMS Example
(3, 1)
(5, 2)
6
RMS Results

• RMS is optimal.
• If a set of of periodic tasks (satisfying the
  assumptions set out previously) is not
  schedulable under RMS then no static priority
  algorithm can schedule this set of tasks.
• RMS requires very little run time processing.

7
Process Utilization Factor

• Task set T1, T2, , Tn
• Process Utilization Factor
• ? Ci / Ti
• C1 / T1 C2 / T2 Cn / Tn
• If this factor is GREATER than 1 then the task
  set can not be scheduled.
• Why?
• If UF 1 it may be RMS-schedulable.

8
RMS Schedulability

• Task set T1, T2, , Tn
• If UF ? Ulub then it is guaranteed to be
  schedulable.
• Ulub - The least upper bound of processor
• utility.
• For RMS, Ulub n( 21/n 1)

9
Process Utilization Factor

• Task set T1, T2, , Tn
• If UF ? Ulub then it is guaranteed to be
  schedulable.
• But if UF is greater than Ulub and not greater
  than 1, we must check explicitly whether the task
  set is RMS-schedulable.

10
RMS Schedulability
This is only a sufficient criterion! This
criterion may fail and yet an RMS may exist.
11
RMS Example
UF 0.66 0.20 0.86 Ulub 0. 828
(3, 2)
(5, 1)
12
RMS Example
UF 0.33 0.40 0.73 Ulub 0. 828
(3, 1)
(5, 2)
13
EDF

• Earliest Deadline First.
• Tasks with earlier deadlines will have higher
  priorities.
• Applies to both periodic and aperiodic tasks.
• EDF is optimal for dynamic priority algorithms.
• A set of periodic tasks is schedulable with EDF
  iff the utilization factor is not greater
• than 1.

14
An Example

• T1, T2
• T1 ( 5, 2)
• T2 (7, 4)
• UF 0.4 0.57 0.97

15
An RMS Schedule?
Time-Overflow
16
The Example

• UF 0.4 0.57 0.97
• Guaranteed to be schedulable under EDF!

17
An EDF Schedule
18
Shared Resources

• Inter-process communication
• Shared resources, critical sections, priority
  inversion, access protocols.
• Material taken from
• Buttazzos Book parts of chapter 7.

19
The Host Computer
DSP
Processor
Timer
ASIC
Memory
20
Tasks
TASK1
TASK2
TASK3
TASK4
DATA SETS
21
Resource Constraints

• resource
• software structure used by a task during its
  execution.
• A data structure, variables, an area of main
  memory, a file, a piece of code, a set of
  registers of a peripheral device.
• Shared resource
• Used by more than one task.
• Exclusive resource
• No simultaneous access.
• Require mutual exclusion.
• Operating must provide a synchronization
  mechanism to ensure sequential access..

22
Critical Section

• Critical section
• A piece of code belonging to task executed under
  mutual exclusion constraints.
• Mutual exclusion enforced by semaphores.
• wait(s)
• Blocked if s 0.
• signal(s)
• s is set to 1 when signal(s) executes.

23
Structure of Critical Sections.
24
Wait State

• A task waiting for an exclusive resource is
  blocked on that resource.
• Tasks blocked on the same resource are kept in a
  wait queue associated with the semaphore
  protecting the resource.
• A task in the running state executing wait(s) on
  a locked semaphore (s 0) enters the waiting
  state.
• When a task currently using the resource executes
  signal(s), the semaphore is released.
• When a task leaves its waiting state (because the
  semaphore has been released) it goes into the
  ready state

25
Waiting State
26
Blocking via Exclusive Resource
J1 has higher priority than J2. Preemption is in
play. Only one processor available.
27
Resource Access Protocols

• Multiple tasks.
• Uniprocessor
• Shared resources.
• Need proper protocols for accessing shared
  resources.
• Resource access protocols.
• Avoid priority inversion!

28
Priority Inversion.
J1
J2
J3
0 1 2 3 4
5 6 7
J1 gt J2 gt J3 3, 6 priority inversion
period. J1 waits for the execution of J2 and the
critical section of J3
29
Priority Inversion

• The Mars pathfinder Mission in 1997 ran into
  serious problem.
• The spacecraft began experiencing total system
  resets with loss of data each time.
• It turned out to be due to priority inversion.
• See the web page and the links there in the IVLE
  area!

30
Avoiding Priority inversion

• Disallow preemption during the execution of a
  critical section.
• Works only if critical sections are short.
• Might unneccesarily block higher priority
  processes that do not even use any shared
  resources!
• Resource access protocols
• Priority inheritance protocol.
• Priority ceiling protocol.

31
Priority Inheritance Protocol

• Tasks have nominal and active priorities.
• Nominal priority
• assigned by the scheduling algorithm (RMS,
  EDF,..)
• Active priority
• assigned by the protocol dynamically- to avoid
  priority inversion.

32
Priority Inheritance Protocol

• Basic idea
• When Ji blocks higher-priority tasks, then its
  active priority is set to the highest of the
  priorities of the tasks it blocks.
• Ji inherits -temporarily the highest priority
  of the blocked tasks.
• This prevents medium priority tasks from
  preempting Ji and prolonging the blocking
  duration of the higher priority tasks.

33
Priority Inheritance Protocol

• The Protocol
• Jobs are scheduled based on their active
  priorities.
• If Ji tries to enter a critical section and the
  corresponding resource is being held by Jj then
  Ji is blocked it is said to be blocked by Jj.
• When a job is blocked on a semaphore, it
  transmits its active priority to the job that
  holds the semaphore in general, a task inherits
  the highest priority of the jobs blocked by it.

34
Priority Inheritance Protocol

• The Protocol
• When Jk exits a critical section, it unlocks the
  semaphore the job with the highest priority that
  is blocked on the semaphore, if any, is awakened.
  The priority of Jk is set to the highest priority
  of the job it is currently blocking. If none, its
  priority is set to its nominal one.

35
Example
36
Nested Critical Sections
37
Priority Inheritance Protocol

• Good news
• If there are m distinct semaphores that can
  block a job J then J can be blocked for at most
  the duration of at most one critical section, one
  for each of the semaphores.
• It can never be as long as the WCET of a lower
  priority task.

38
Priority Inheritance Protocol

• Bad news
• Chained Blocking
• J can get blocked on n critical sections held by
  n distinct lower priority jobs.
• Deadlocks.

39
Chained Blocking
40
Deadlock
41
Priority Ceiling Protocol

• Extension of the Priority Inheritance Protocol.
• Avoids chained blocking and deadlocks.
• Basic Idea
• A task is not allowed to enter a critical section
  if there are already locked semaphores which
  could block it eventually (due to a sub-critical
  section nested within the entering critical
  section).
• Hence, once a task enters a critical section, it
  can not be blocked by lower priority tasks till
  its completion.

42
Priority Ceiling Protocol

• The Protocol
• Each semaphore S is assigned a priority ceiling
  C(S). It is the priority of the highest priority
  task that can lock S. This is a static value.
• Suppose J is currently running and it wants to
  lock the semaphore S. J is allowed to lock S only
  if the priority of J is strictly higher than the
  priority ceiling C(S) of the semaphore S where
• S is the semaphore with the highest priority
  ceiling among all the semaphores which are
  currently locked by jobs other than J.
• In this case, J is said to blocked by the
  semaphore S (and the job currently holding S).

43
Priority Ceiling Protocol

• The Protocol
• When J gets blocked by S then the priority of J
  is transmitted to the job that currently holds
  S.
• When J leaves a critical section guarded by S
  then it unlocks S and the highest priority job,
  if any, which is blocked by S is awakened.
• The priority of J is set to the highest priority
  of the job that is blocked by some semaphore that
  J is still holding. If none, the priority of J
  is set to be its nominal one.

44
Example
45
C (S0) ? C(S1) ? C(S2) ?
46
C (S0) P0 C(S1) P0 C(S2) P1
47
t2 J1 can not lock S2. Currently J2 is holding
S2 and C(S2) P1 and the current priority of J1
is also P1.
48
t5 J0 can not lock S0. Currently J2 is holding
S2 and S1 and C(S1) P0 and the current priority
of J0 is also P0. The (inherited) priority of J2
is now P0.
49
t6 J2 unlocks S1. It awakens J0. But J2s
(inherited) priority is now only P1 while P0 gt
C(S2) P1. So J0 preempts J2 and runs to
completion.
50
t7 J2 resumes execution with priority P1.
51
t8 J2 unlocks S2 and goes back to its nominal
priority P2. So J1 preempts J0 and runs to
completion.
52
Two Key Properties

• Under priority ceiling protocol, a job can be
  blocked for at most the duration of one critical
  section.
• The priority ceiling protocol prevents deadlocks.

53
Blocking Times

• Bi ----- The maximum blocking time of Ji.
• Dj,k ---- The duration the critical section in Jj
  whose access is regulated by the semaphore Sk.
• If Dj,k is 0 this means Jj does not use Sk.
• Then
• Bi max Dj,k Pj lt Pi, C(Sk) ? Pi