book

Linux System Programming, 2nd Edition

by Robert Love

May 2013

Intermediate to advanced

456 pages

11h 56m

English

O'Reilly Media, Inc.

Read now

Unlock full access

Audience and AssumptionsContents of This BookVersions Covered in This BookConventions Used in This BookUsing Code ExamplesSafari® Books OnlineHow to Contact UsAcknowledgments
System ProgrammingWhy Learn System ProgrammingCornerstones of System ProgrammingSystem CallsInvoking system callsThe C LibraryThe C CompilerAPIs and ABIsAPIsABIsStandardsPOSIX and SUS HistoryC Language StandardsLinux and the StandardsThis Book and the StandardsConcepts of Linux ProgrammingFiles and the FilesystemRegular filesDirectories and linksHard linksSymbolic linksSpecial filesFilesystems and namespacesProcessesThreadsProcess hierarchyUsers and GroupsPermissionsSignalsInterprocess CommunicationHeadersError HandlingGetting Started with System Programming
Opening FilesThe open() System CallFlags for open()Owners of New FilesPermissions of New FilesThe creat() FunctionReturn Values and Error CodesReading via read()Return ValuesReading All the BytesNonblocking ReadsOther Error ValuesSize Limits on read()Writing with write()Partial WritesAppend ModeNonblocking WritesOther Error CodesSize Limits on write()Behavior of write()Synchronized I/Ofsync() and fdatasync()Return values and error codessync()The O_SYNC FlagO_DSYNC and O_RSYNCDirect I/OClosing FilesError ValuesSeeking with lseek()Seeking Past the End of a FileError ValuesLimitationsPositional Reads and WritesError ValuesTruncating FilesMultiplexed I/Oselect()Return values and error codesselect() examplePortable sleeping with select()pselect()poll()Return values and error codespoll() exampleppoll()poll() Versus select()Kernel InternalsThe Virtual FilesystemThe Page CachePage WritebackConclusion
User-Buffered I/OBlock SizeStandard I/OFile PointersOpening FilesModesOpening a Stream via File DescriptorClosing StreamsClosing All StreamsReading from a StreamReading a Character at a TimePutting the character backReading an Entire LineReading arbitrary stringsReading Binary DataWriting to a StreamWriting a Single CharacterWriting a String of CharactersWriting Binary DataSample Program Using Buffered I/OSeeking a StreamObtaining the Current Stream PositionFlushing a StreamErrors and End-of-FileObtaining the Associated File DescriptorControlling the BufferingThread SafetyManual File LockingUnlocked Stream OperationsCritiques of Standard I/OConclusion
Scatter/Gather I/Oreadv() and writev()Return valueswritev() examplereadv() exampleImplementationEvent PollCreating a New Epoll InstanceControlling EpollWaiting for Events with EpollEdge- Versus Level-Triggered EventsMapping Files into Memorymmap()The page sizeReturn values and error codesAssociated signalsmunmap()Mapping ExampleAdvantages of mmap()Disadvantages of mmap()Resizing a MappingReturn values and error codesChanging the Protection of a MappingReturn values and error codesSynchronizing a File with a MappingReturn values and error codesGiving Advice on a MappingReturn values and error codesAdvice for Normal File I/OThe posix_fadvise() System CallReturn values and error codesThe readahead() System CallReturn values and error codesAdvice Is CheapSynchronized, Synchronous, and Asynchronous OperationsAsynchronous I/OI/O Schedulers and I/O PerformanceDisk AddressingThe Life of an I/O SchedulerHelping Out ReadsThe Deadline I/O SchedulerThe Anticipatory I/O SchedulerThe CFQ I/O SchedulerThe Noop I/O SchedulerSelecting and Configuring Your I/O SchedulerOptimzing I/O PerformanceScheduling I/O in user spaceSorting by pathSorting by inodeSorting by physical blockConclusion
Programs, Processes, and ThreadsThe Process IDProcess ID AllocationThe Process Hierarchypid_tObtaining the Process ID and Parent Process IDRunning a New ProcessThe Exec Family of CallsThe rest of the familyError valuesThe fork() System CallCopy-on-writevfork()Terminating a ProcessOther Ways to Terminateatexit()on_exit()SIGCHLDWaiting for Terminated Child ProcessesWaiting for a Specific ProcessEven More Waiting VersatilityBSD Wants to Play: wait3() and wait4()Launching and Waiting for a New ProcessZombiesUsers and GroupsReal, Effective, and Saved User and Group IDsChanging the Real or Saved User or Group IDChanging the Effective User or Group IDChanging the User and Group IDs, BSD StyleChanging the User and Group IDs, HP-UX StylePreferred User/Group ID ManipulationsSupport for Saved User IDsObtaining the User and Group IDsSessions and Process GroupsSession System CallsProcess Group System CallsObsolete Process Group FunctionsDaemonsConclusion
Process SchedulingTimeslicesI/O- Versus Processor-Bound ProcessesPreemptive SchedulingThe Completely Fair SchedulerYielding the ProcessorLegitimate UsesProcess Prioritiesnice()getpriority() and setpriority()I/O PrioritiesProcessor Affinitysched_getaffinity() and sched_setaffinity()Real-Time SystemsHard Versus Soft Real-Time SystemsLatency, Jitter, and DeadlinesLinux’s Real-Time SupportLinux Scheduling Policies and PrioritiesThe first in, first out policyThe round-robin policyThe normal policyThe batch scheduling policySetting the Linux scheduling policySetting Scheduling ParametersError codesDetermining the range of valid prioritiessched_rr_get_interval()Error codesPrecautions with Real-Time ProcessesDeterminismPrefaulting data and locking memoryCPU affinity and real-time processesResource LimitsThe LimitsDefault limitsSetting and Retrieving LimitsError codes

Binaries, Processes, and ThreadsMultithreadingCosts of MultithreadingAlternatives to MultithreadingThreading ModelsUser-Level ThreadingHybrid ThreadingCoroutines and FibersThreading PatternsThread-per-ConnectionEvent-Driven ThreadingConcurrency, Parallelism, and RacesRace ConditionsReal-world racesSynchronizationMutexesDeadlocksDeadlock avoidancePthreadsLinux Threading ImplementationsThe Pthread APILinking PthreadsCreating ThreadsThread IDsComparing thread IDsTerminating ThreadsTerminating yourselfTerminating othersJoining and Detaching ThreadsJoining threadsDetaching threadsA Threading ExamplePthread MutexesInitializing mutexesLocking mutexesUnlocking mutexesMutex exampleFurther Study
Files and Their MetadataThe Stat FamilyPermissionsOwnershipExtended AttributesKeys and valuesExtended attribute namespacesExtended Attribute OperationsRetrieving an extended attributeSetting an extended attributeListing the extended attributes on a fileRemoving an extended attributeDirectoriesThe Current Working DirectoryObtaining the current working directoryChanging the current working directoryCreating DirectoriesRemoving DirectoriesReading a Directory’s ContentsReading from a directory streamClosing the directory streamSystem calls for reading directory contentsLinksHard LinksSymbolic LinksUnlinkingCopying and Moving FilesCopyingMovingDevice NodesSpecial Device NodesThe Random Number GeneratorOut-of-Band CommunicationMonitoring File EventsInitializing inotifyWatchesAdding a new watchWatch masksinotify EventsReading inotify eventsAdvanced inotify eventsLinking together move eventsAdvanced Watch OptionsRemoving an inotify WatchObtaining the Size of the Event QueueDestroying an inotify Instance
The Process Address SpacePages and PagingSharing and copy-on-writeMemory RegionsAllocating Dynamic MemoryAllocating ArraysResizing AllocationsFreeing Dynamic MemoryAlignmentAllocating aligned memoryOther alignment concernsManaging the Data SegmentAnonymous Memory MappingsCreating Anonymous Memory MappingsMapping /dev/zeroAdvanced Memory AllocationFine-Tuning with malloc_usable_size() and malloc_trim()Debugging Memory AllocationsObtaining StatisticsStack-Based AllocationsDuplicating Strings on the StackVariable-Length ArraysChoosing a Memory Allocation MechanismManipulating MemorySetting BytesComparing BytesMoving BytesSearching BytesFrobnicating BytesLocking MemoryLocking Part of an Address SpaceLocking All of an Address SpaceUnlocking MemoryLocking LimitsIs a Page in Physical Memory?Opportunistic AllocationOvercommitting and OOM
Signal ConceptsSignal IdentifiersSignals Supported by LinuxBasic Signal ManagementWaiting for a Signal, Any SignalExamplesExecution and InheritanceMapping Signal Numbers to StringsSending a SignalPermissionsExamplesSending a Signal to YourselfSending a Signal to an Entire Process GroupReentrancyGuaranteed-Reentrant FunctionsSignal SetsMore Signal Set FunctionsBlocking SignalsRetrieving Pending SignalsWaiting for a Set of SignalsAdvanced Signal ManagementThe siginfo_t StructureThe Wonderful World of si_codeSending a Signal with a PayloadSignal Payload ExampleA Flaw in Unix?
Time’s Data StructuresThe Original RepresentationAnd Now, Microsecond PrecisionEven Better: Nanosecond PrecisionBreaking Down TimeA Type for Process TimePOSIX ClocksTime Source ResolutionGetting the Current Time of DayA Better InterfaceAn Advanced InterfaceGetting the Process TimeSetting the Current Time of DaySetting Time with PrecisionAn Advanced Interface for Setting the TimePlaying with TimeTuning the System ClockSleeping and WaitingSleeping with Microsecond PrecisionSleeping with Nanosecond ResolutionAn Advanced Approach to SleepA Portable Way to SleepOverrunsAlternatives to SleepingTimersSimple AlarmsInterval TimersAdvanced TimersCreating a timerArming a timerObtaining the expiration of a timerObtaining the overrun of a timerDeleting a timer
GNU CInline FunctionsSuppressing InliningPure FunctionsConstant FunctionsFunctions That Do Not ReturnFunctions That Allocate MemoryForcing Callers to Check the Return ValueMarking Functions as DeprecatedMarking Functions as UsedMarking Functions or Parameters as UnusedPacking a StructureIncreasing the Alignment of a VariablePlacing Global Variables in a RegisterBranch AnnotationGetting the Type of an ExpressionGetting the Alignment of a TypeThe Offset of a Member Within a StructureObtaining the Return Address of a FunctionCase RangesVoid and Function Pointer ArithmeticMore Portable and More Beautiful in One Fell Swoop
Books on the C Programming LanguageBooks on Linux ProgrammingBooks on the Linux KernelBooks on Operating System Design

Overview

Write software that draws directly on services offered by the Linux kernel and core system libraries. With this comprehensive book, Linux kernel contributor Robert Love provides you with a tutorial on Linux system programming, a reference manual on Linux system calls, and an insider’s guide to writing smarter, faster code.

Love clearly distinguishes between POSIX standard functions and special services offered only by Linux. With a new chapter on multithreading, this updated and expanded edition provides an in-depth look at Linux from both a theoretical and applied perspective over a wide range of programming topics, including: