Understanding Operating Systems
Fifth Edition
Chapter 5
Process Management
Deadlock
• Resource sharing: Memory and CPU sharing
• Many programs competing for limited resources
• Lack of process synchronization consequences
– Deadlock: “deadly embrace”
•
•
•
•
Two or more jobs placed in HOLD state
Jobs waiting for unavailable vital resource
System comes to standstill
OS can’t resolve the deadlock; external intervention bys
users and/or sys admin needed to terminate jobs
– Starvation
• Infinite postponement of job
Understanding Operating Systems, Fifth Edition
2
Deadlock (continued)
• More serious than starvation
• Affects entire system
– Affects more than one job
• Not just a few programs
– All system resources become unavailable
• Example: traffic jam (Figure 5.1)
• More prevalent in interactive systems
• Real-time systems
– Deadlocks quickly become critical situations
• No simple and immediate solution
Understanding Operating Systems, Fifth Edition
3
Deadlock (continued)
Understanding Operating Systems, Fifth Edition
4
Seven Cases of Deadlock
• Deadlocks usually happen when non-sharable and nonpreemptable resources (such as files, printers, and scanners)
are allocated to jobs that eventually require other such
resources which have been locked by other jobs.
• Resource types locked by competing jobs
–
–
–
–
–
–
–
File requests
Databases
Dedicated device allocation
Multiple device allocation
Spooling
Disk sharing
Network
Understanding Operating Systems, Fifth Edition
5
Case 1: Deadlocks on File Requests
• Jobs request and hold files for execution duration
• Example (Figure 5.2)
– Two programs (P1, P2) and two files (F1, F2)
– Deadlock sequence
• P1 has access to F2 and also requires F1
• P2 has access to F1 and also requires F2
– Deadlock remains
• Until one program withdrawn or
• Until one program forcibly removed and file released
– Other programs requiring F1 or F2
• Put on hold for duration of situation
Understanding Operating Systems, Fifth Edition
6
Case 1: Deadlocks on File Requests
(continued)
Understanding Operating Systems, Fifth Edition
7
Case 2: Deadlocks in Databases
• A deadlock can also occur if two processes access
and lock records in a database.
• Locking: Technique to guarantee the integrity of
the data through which the user locks out all other
users while working with the database.
– Three locking levels
• Entire database for duration of request
• Subsection of database
• Individual record until request completed
Understanding Operating Systems, Fifth Edition
8
Case 2: Deadlocks in Databases
(continued)
• Example: two processes (P1 and P2)
– Each needs to update two records (R1 and R2)
– Deadlock sequence
•
•
•
•
P1 accesses R1 and locks it
P2 accesses R2 and locks it
P1 requests R2 but locked by P2
P2 requests R1 but locked by P1
• Race between processes
– Results when locking not used
– Causes incorrect final version of data
– Depends on process execution order
Understanding Operating Systems, Fifth Edition
9
Case 2: Deadlocks in Databases
(continued)
Understanding Operating Systems, Fifth Edition
10
Case 3: Deadlocks in Dedicated
Device Allocation
• Limited number of dedicated devices (devices that
can be assigned to a single process at a time)
• Example
– Two programs (P1, P2)
• Need two tape drives each
• Only two tape drives in system
– Deadlock sequence
•
•
•
•
P1 requests tape drive 1 and gets it
P2 requests tape drive 2 and gets it
P1 requests tape drive 2 but blocked
P2 requests tape drive 1 but blocked
Understanding Operating Systems, Fifth Edition
11
Case 4: Deadlocks in Multiple
Device Allocation
• Several processes request and hold dedicated
devices
• Example (Figure 5.4)
– Three programs (P1, P2, P3)
– Three dedicated devices (scanner, printer, plotter)
– Deadlock sequence
•
•
•
•
•
•
P1 requests and gets scanner
P2 requests and gets printer
P3 requests and gets the plotter
P1 requests printer but blocked
P2 requests plotter but blocked
P3 requests scanner but blocked
Understanding Operating Systems, Fifth Edition
12
Case 4: Deadlocks in Multiple
Device Allocation (continued)
Understanding Operating Systems, Fifth Edition
13
Case 5: Deadlocks in Spooling
• Virtual device
– Dedicated device made sharable
– Example: The printer has been transformed into a
sharable device by installing a high-speed disk
between the printer and the CPU.
• Spooling
– Process
• The disk accepts output from several users
• Acts as temporary storage for output
• Output resides in disk area until printer accepts job
data
Understanding Operating Systems, Fifth Edition
14
Case 5: Deadlocks in Spooling
(continued)
• Deadlock sequence
– Printer needs all job output before printing begins
•
•
•
•
Spooling system fills disk space area
No one job has entire print output in spool area
Results in partially completed output for all jobs
Results in deadlock
Understanding Operating Systems, Fifth Edition
15
Case 6: Deadlocks in a Network
• A network that’s congested or has filled a large
percentage of its I/O queues can become
deadlocked if it doesn’t have protocols controlling
the network message flow.
• Example (Figure 5.5)
– Seven computers on network
• Each on different nodes
– Direction of arrows
• Indicates message flow
– Deadlock sequence
• All available buffer space fills
Understanding Operating Systems, Fifth Edition
16
Case 6: Deadlocks in a Network
(continued)
Understanding Operating Systems, Fifth Edition
17
Case 7: Deadlocks in Disk Sharing
• Competing processes send conflicting commands
– Scenario: disk access
• Example (Figure 5.6)
– Two processes
– Each process waiting for I/O request
• One at cylinder 20 and one at cylinder 310
– Deadlock sequence
• Neither I/O request satisfied
• Device puts request on hold while attempting to fulfill
other request for each request
– Livelock results
Understanding Operating Systems, Fifth Edition
18
Case 7: Deadlocks in Disk Sharing
(continued)
Understanding Operating Systems, Fifth Edition
19
Conditions for Deadlock
• Four conditions simultaneously occurring prior to deadlock or
livelock
– Mutual exclusion
– Resource holding
– No preemption
– Circular wait
• All needed by operating system
– Must recognize simultaneous occurrence of four conditions
• Resolving deadlock
– Removal of one condition resolves the deadlock. Although this
concept is obvious, it isn’t easy to implement.
Understanding Operating Systems, Fifth Edition
20
Conditions for Deadlock (continued)
• Mutual exclusion
– Allowing only one process access to dedicated
resource
• Resource holding
– A process holds on to a resource, does not release it
and waits for the other job to retreat
• No preemption
– A resource is allocated to a process for as long as
needed. Lack of temporary reallocation of resources.
Understanding Operating Systems, Fifth Edition
21
Conditions for Deadlock (continued)
• Circular wait
– Each process involved in the impasse is waiting for
another process to voluntarily release a resource so
at least one process can continue
• All four conditions are required for deadlock
occurrence but if one condition is removed then the
deadlock is broken.
Understanding Operating Systems, Fifth Edition
22
Modeling Deadlocks
• Directed graphs
–
–
–
–
Used to model deadlocks
Circles represent processes
Squares represent resources
Solid arrow from a resource to a process
• Process is holding the resource
– Solid arrow from a process to a resource
• Process is waiting for the resource
– Arrow direction indicates flow
– If there’s a cycle in the graph then there’s a deadlock
involving the processes and the resources in the
cycle
Understanding Operating Systems, Fifth Edition
23
Modeling Deadlocks (continued)
Understanding Operating Systems, Fifth Edition
24
Modeling Deadlocks (continued)
• Three graph scenarios to help detect deadlocks
– System has three processes (P1, P2, P3)
– System has three resources (R1, R2, R3)
• Scenario one: no deadlock
– Resources released before next process request
• Scenario two: deadlock
– Processes waiting for resource held by another
• Scenario three: no deadlock
– Resources released before deadlock
Understanding Operating Systems, Fifth Edition
25
Modeling Deadlocks (continued)
• No deadlock
– Resources released before next process request
Understanding Operating Systems, Fifth Edition
26
Modeling Deadlocks (continued)
Understanding Operating Systems, Fifth Edition
27
Modeling Deadlocks (continued)
• Deadlock
– Processes waiting for resource held by another
Understanding Operating Systems, Fifth Edition
28
Modeling Deadlocks (continued)
Understanding Operating Systems, Fifth Edition
29
Modeling Deadlocks (continued)
• No deadlock
– Resources released before deadlock
Understanding Operating Systems, Fifth Edition
30
Modeling Deadlocks (continued)
Understanding Operating Systems, Fifth Edition
31
Modeling Deadlocks (continued)
• Examples studied so far involved the allocation of
one or more resources of different types (e.g. a
printers and a disk drive) to a process.
• The directed graphs can be expanded to include
several resources of the same type (e.g. a tape
drive) that can be allocated individually or in groups
to the same process.
• These graph clusters devices of the same type into
one entity (shown in Fig. 5.11 as a rectangle).
Understanding Operating Systems, Fifth Edition
32
Modeling Deadlocks (continued)
Understanding Operating Systems, Fifth Edition
33
Strategies for Handling Deadlocks
• The requests and releases of resources are received in an
unpredictable order making it very difficult to design a
foolproof preventive policy. Operating systems use one of
these strategies to deal with deadlocks:
• Prevention
– Prevent occurrence of one condition
• Mutual exclusion, resource holding, no preemption, circular
wait
• Avoidance
– Avoid deadlock if it becomes probable
• Detection and Recovery
– Detect deadlock when it occurs
– Recover gracefully
Understanding Operating Systems, Fifth Edition
34
Strategies for Handling Deadlocks
(continued)
• Prevention eliminates one of four conditions
– Complication: same condition can’t be eliminated from every resource
– Mutual exclusion
• Some resources must allocated exclusively to one user (e.g.
memory, CPU)
• Bypassed if I/O device uses spooling (e.g. using a print spooler on a
printer). However, we may be trading one type of deadlock for
another (Case 3: Deadlocks in Dedicated Device Allocation for
Case 5: Deadlocks in Spooling)
– Resource holding
• Job holds on to one resource while waiting for another that’s not yet
available.
• Sidestepped if jobs request every necessary resource at creation
time
• Multiprogramming degree significantly decreased
• Idle peripheral devices
• Can be successfully used in batch systems but doesn’t work well in
interactive systems
Understanding Operating Systems, Fifth Edition
35
Strategies for Handling Deadlocks
(continued)
• Prevention (continued)
– No preemption
• Bypassed if operating system allowed to deallocate
resources from jobs
• Okay if job state easily saved and restored
• Not accepted to preempt dedicated I/O devices or files
during the modification process
– Circular wait
• Bypassed if operating system prevents circle formation
• Requires jobs to anticipate resource request order
• Difficult to satisfy all users
Understanding Operating Systems, Fifth Edition
36
Strategies for Handling Deadlocks
(continued)
• Avoidance: even if one of the 4 conditions to prevent
deadlock cannot be removed, the OS can avoid deadlock if
the OS knows ahead of time the sequence of requests
associated with each active process.
• Dijkstra’s Bankers Algorithm
– Algorithm to regulate resource allocation to avoid deadlock. It is
based on a bank with a fixed amount of capital that operates on
the following principles:
• No customer granted loan exceeding bank’s total capital
• All customers given maximum credit limit
• No customer allowed to borrow over limit
• Sum of all loans will not exceed bank’s total capital
Understanding Operating Systems, Fifth Edition
37
Strategies for Handling Deadlocks
(continued)
Understanding Operating Systems, Fifth Edition
38
Strategies for Handling Deadlocks
(continued)
Understanding Operating Systems, Fifth Edition
39
Strategies for Handling Deadlocks
(continued)
Understanding Operating Systems, Fifth Edition
40
Strategies for Handling Deadlocks
(continued)
Understanding Operating Systems, Fifth Edition
41
Strategies for Handling Deadlocks
(continued)
• Operating systems deadlock avoidance assurances
– Never satisfy request if job state moves from safe to
unsafe
• Identify job with smallest number of remaining
resources
• Number of available resources >= number needed for
selected job to complete
• Block request jeopardizing safe state
Understanding Operating Systems, Fifth Edition
42
Strategies for Handling Deadlocks
(continued)
• Problems with the Banker’s Algorithm
– Jobs must predict maximum number needed resources. This is
impractical for interactive systems.
– Requires constant number of total resources for each class.
– Number of jobs must remain fixed. This is not possible in an
interactive system.
– Possible high overhead cost incurred by running the avoidance
algorithm because it must be invoked for every request.
– Resources not well utilized
• Algorithm assumes worst case
– Scheduling suffers
• Result of poor resource utilization
• Jobs kept waiting for resource allocation
Understanding Operating Systems, Fifth Edition
43
Strategies for Handling Deadlocks
(continued)
• Detection: build directed resource graphs
– Look for cycles
• Algorithm detecting circularity
– Executed whenever appropriate
• Detection algorithm used to reduce directed graphs.
– Remove process using current resource and not
waiting for one
– Repeat step until all connecting lines removed
Understanding Operating Systems, Fifth Edition
44
Strategies for Handling Deadlocks
(continued)
Understanding Operating Systems, Fifth Edition
45
Strategies for Handling Deadlocks
(continued)
Understanding Operating Systems, Fifth Edition
46
Strategies for Handling Deadlocks
(continued)
• Recovery
– Deadlock untangled once detected
– System returns to normal quickly
• All recovery algorithms require at least one victim (an
expendable job)
• Recovery methods
– Terminate every job active in system
• Restart all jobs from beginning
– Terminate only jobs involved in deadlock
• Ask users to resubmit jobs
– Identify jobs involved in deadlock
• Terminate jobs one at a time, checking to see if the deadlock
is broken after each removal, until the deadlock is resolved.
Understanding Operating Systems, Fifth Edition
47
Strategies for Handling Deadlocks
(continued)
• Recovery methods (continued)
– Select a non-deadlocked job
• Preempt its resources
• Allocate resources to a deadlocked process
– Stop new jobs from entering system
•
•
•
•
Allow non-deadlocked jobs to complete
Releases resources when complete
No victim
Will work only if the number of available resources
exceeds that needed by at least one of the deadlocked
jobs to run.
Understanding Operating Systems, Fifth Edition
48
Strategies for Handling Deadlocks
(continued)
• Factors to consider
– Select victim with least-negative effect on the system
– Most common
• Job priority under consideration: high-priority jobs
usually untouched
• CPU time used by job: jobs close to completion usually
left alone
• Number of other jobs affected if job selected as victim
• Jobs modifying data: usually not selected for
termination (a database issue) because a database
that is only partially updated is only partially correct
(consistency and validity of the database must not be
compromised).
Understanding Operating Systems, Fifth Edition
49
Starvation
• Job execution prevented
– Waiting for resources that never become available
– Results from conservative resource allocation
• Example
– “The dining philosophers” by Dijkstra
• Starvation avoidance
– Implement algorithm tracking how long each job
waiting for resources (aging)
– Block new jobs until starving jobs satisfied
Understanding Operating Systems, Fifth Edition
50
Starvation (continued)
Understanding Operating Systems, Fifth Edition
51
Starvation (continued)
Understanding Operating Systems, Fifth Edition
52