understanding operating systems fifth edition chapter 5 process management

Understanding Operating SystemsFifth Edition

Chapter 5Process Management

Understanding Operating Systems, Fifth Edition 2

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


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


Deadlock (continued)


Seven Cases of Deadlock

• Deadlocks usually happen when non-sharable and non-preemptable 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


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– Deadlock remains • Until one program withdrawn or• Until one program forcibly removed and file released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


Case 1: Deadlocks on File Requests (continued)


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


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


Case 2: Deadlocks in Databases (continued)


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


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


Case 4: Deadlocks in Multiple Device Allocation (continued)


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


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


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


Case 6: Deadlocks in a Network (continued)


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


Case 7: Deadlocks in Disk Sharing (continued)


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.


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.


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.


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


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



• No deadlock– Resources released before next process request



• Deadlock– Processes waiting for resource held by another



• No deadlock– Resources released before deadlock



• 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).


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


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



• 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



• 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



• 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



• 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



• 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



• 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.



• 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.



• 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).


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


Starvation (continued)

understanding operating systems fifth edition chapter 5 process management

Documents