Unie HL 3 Linkers and Loaders ‘3a: Introduction 4 What is loader ? ‘Aus: Loader is # wilty program which takes objet code as input repures ft forexestion and Tass the execuable code into th memery ‘Thus loader is actully responsible for inating the exeation process, 2 Discuss four atereat functions of loader. SMP: Ap, nas 4) ‘Ans. The louder i responsible for the activites such as allocation, linking, relocation and lading 1) Wallets she apace for program in the memory, by eating the sizeof the program. This stv is called allocation. 2) resolves te symbolic references (codeldata) beeen the eet modules by assigning all the user subroutine and brary subroutine adresses, Ths activity is elle linking 3) There are some addtess dependent locations in the program, such adéress constants must be adjusted according to allocated space, such setivity done by loader is called relocation, 4) Finally it places all the machine instructions and data of comesponding programs and subroutines into: the memory. Ths program now becomes ready for execution, this activity is called loading. 32 Sates and Haters Loader Schemes! 3 What are types of londers ? Explain compile and go loader scheme with advantages and disadvantages, ca[S0 : Apia, Marks 6] ‘Ans. + Types of Loaders : ‘Based on various functionalities of the loader, there are various types of loaders, which are known as loading schemes. Various loaders sebemes 1, Compile and Go 3, Absolute Loader 5. Relocating Loader: CComplle and go Loader : 2 General Loader Scheme 4. Subroutine Linkage 6. Direct Linkage Loader ‘+n this ype of loader, the insrction is readline by line, its machine code is obtained and itis dveedy put in the main memory at some known address ‘2 That_means the assembler runs in one part of memory and the assembled machine insructions and data i directly put into their signed memory locations . ‘After completion of assembly process loader contains the instruction using. which the location courte if set 10 the start of the newly assembled program 1 The typical example is WATFOR.77, its ¢ FORTRAN compiler which uses such “load and go" scheme, This loading scheme is also called as assemble and g er Sonate sin meray Fig. 0.3.1 Compllo and go losing schome‘sPaos. Hither Long, es ee “Advantage :Itis simple to implement. a serge 1 Asem crocs m=O. HERE Me Sie oS scion of jl the rues code i diety Cenc ensue frm ua renin pana! mning seme (lng LDIC laren) yy tages and deaantge- ae [EB [SPPU + Deci-48, End Sem, March-20, In Sem, Marky tet som ne “anc Tn is fader scheme the sce rogram’ enmvertcd Wey, rogram ty some excl (see) rere de wes thn ejest modules and pus machbe ining Te aa Selon ono ir ined memory. «The loader oceupies some portion of main memory. ‘Advantages 1. The program need not be retanslated each time while rimming it Ti 4s because intially when source program gets executed an objet program gets generated. If program i not modified, then loader can make use ofthis object program fo conver it to executable form, 2. There is no wastage of memory, because assembler is not paced ia the memory, insiead of it, loader occupies some portion of the memory. And size of loader is smaller than assembler, so more memory is available t the user. 3. Ibis possible to write source program with multiple programs and ‘multiple languages, because the source programs are fist converted to sn ts mone Fig. Q.4.1 Gonoral loader scheme “sd fr ninering Sa sas. Linker ond Loaiers ‘object programs alvays, and loader acepts these object module’ to ‘convert i to executable form, (5 What are subroutine linkages ? What are the benefits of using a " EPISPPU: May, Marks 4) ‘Ans. + Subroutine Linkage : It is a loading scheme which makes it possible to call and retum from subroutines. In this mechanism the assembler pseudo-op EXTRN i followed by lst of| symbols. These are basically the symbols which are defined externally but feferenced in the present program. Similarly if a symbol is defined in one program and is referenced in ‘others, then this type of symbol is preceded by pseudo-op ENTRY in the | syibol table }. | | EXTRN 5 //rubrovtine Bis defined extemally| 15.46) BAL 115 END ‘Recomblor Linkage Psoudo-Ops Suppose we have a main program named A and it uses variables |ALA2,A3. which are defined in the same program then the pseudo-op | ENTRY is sed for their declaration ‘A START | ENTRY ALAZA a a A Gale Engage3 ae eudo-Op) ae keno saa lg Se ann as Tg ce ah re ae ng For example EXTRN B1B2 Here B1 and B2 are the extem symbols. {a6 Explain : Absolute load nt ‘ans. « Absolute loader is a kind of loader in which relocated ebje, fee are created, loader acceps these files and places them at spi Tocatons in the memory «This ype of loader i called absolute because no relocation infomation {a needed; rather itis obtained ftom the programmer or assembler, 7 What is the difference between Absolute loader and Compile and Go loader ? ‘Ans. Difference between Absolute and Compile and Go loader Absolute loader Compile and Go teader ‘The assembled code Is place in reloeted object files. Then loader faccepts these files and objet code {place at some assigned memory lection. ‘The assembled code is directly ‘summed into the memory at sme assigned memory location “This echeme allows to load object modules of msltiple programe weten is diferent languages. ‘Tis scheme does not allow obec rmadules of different source program. ‘This is efficent scheme, AS In this type of loadar the fnstruction is readline by line #8 ‘object code is obtained and it is then diretly put ito the memon this scheme i less efficent but simple to implement 7 Gite for Engng | sr40s a6 ‘The drawback ofthis method ts | In this scheme its programmers that assembler an loader occupies | duty to adjust all the inter segment larg block of memory. Hence | addeesees and mancally do the there isa wastage of memory. | linking sii If tal any rmosifction fr done in some opment then starting odes of rent subsequent sogment change. ‘The programmer has to take cae ofall there ieee 8 What are the advantages and disadvantages of relocating foader ? Ans. : Advantages 1. Using. branch instruction to comesponding. subroutine the desired sulouine can be accessed. Thus the four finctions of leader Le alloction, linking, relocation and loading are performed automaticaly by the fader 2, Using relocation bits it can be identified that which instruction needs to be relocated and which instruction needs to be erty placed, 3 Using rnafeter vector linking of external subroutines can be: done swith the min procedue. ‘Thus taster vector of relocating loader helps in solving the extemal references. 4. In this looder scheme, the ‘assembler provides he addtional information such es length of entire program and length of transfer ‘estor to the leader, Using ts information allocation problem gets ralved Disadvantages 1) The transfer vector Hinks are useful for wansering the contol fo Cxtrnl subroutines but this vector tale is not well-suited for leading tr storing external data (Le. data preset in another program sepmen). 2) Dae to transfer vector, th size of object progam inthe memory ges incresed 5) The relocating loader handles the sired procedure segments but it can not handle the shared data segments =Linkers end Loan sraos sing of Binary Symbolic ‘Subroutine(Bss) a8 Extn 0 errr rit a hol age moi is cime en le stoaag : oe fos eet oa en ot wie eee “s See ene SS a 8) nk wt See ee Senta mere ea od eee gee ee Segment independent, This information along with intersegment references is passed on to loader. Say ease en to be changed. Ifit is t be loaded in an arbitrary place in core, 7 ovjact Feccatig program ‘Assemer iescor ‘Inter segment ‘etorenees Fig, 09.4 Relocating load ‘+The assembler takes the source program a input this source program ‘may call some extemal routines, Hence assembler produces the object code as an ouput slong with a table. This table contains entties ofthe fztemal references made inthe program. This table is called teaser For example numbers (SUM) SUM, If a program refers a routine for summing up the then transfer vestor would contain the symbolic name Zhe sistement calling SUM would be translated into a transfer instruction indicating the pointer to the location of the transfer vector ‘associated with SUM. ' “S * = A Gale for Engineering Stes sraos Linkers and Loaders The assembler also provide some addtional information to the loader ‘hie inludes Length of entice program. + Length of transfer vector portion. ‘The loader loads the object program and transfer vector into core. Then it would load each subroutine identified inthe transfer vectr. ‘+The louder then place the transfer instruction to the corresponding subroutine foreach entry in wansfer veetar. Thus execution of the Call for SUM routine will rut in branch to the fst location in transfer vector. Tis locaton will provide tnfer instruction to the location of SUM. ‘The BSS loader scheme is normally used on the computers with a fixed length direct adress nsiniction format. (2.10 Explain the concept of transfer vestor in BSS loader ‘Ans. + + The assembler takes the source program at input, this source program may call some external routines. Hence assembler produces the abject code as an output along with a table. This table contains entries of | the extemal references made in the program. This table is called transfer For example’: If @ program refers a routine for summing up the numbers (SUM) then irnsfer vector would contain the symbolic name SUM. Q.11 Give the direct address instruction format for BSS loader scheme Ans. ¢ The BSS loader scheme is nomlly used on the computers with fixed lenghhdigeet adress instruction format. This format follows - 326 op | rt| x2 a Soo a4 g 6 ‘hiss an adross of opera, hicks dvest acsresinatucton Fig. tt “A One for Enginaning Stentinsmeton 2H 65386 bytes of or ‘ 2 Deine the term ~ Relocation bit ie J isuucdons i the program source need 10 be elo tae oe ny Ben st 1 nad SS ou eros ate, 1 ul fen bat means comesponding address field must be relocated, of hades eld need no be eosated ‘a3 Which information does the assembler submit 1 the dire inking loner ? "aus: The source program is read by assembler and assembler submiy folonag information to loader ~ i) Tae length of promam/seqment. iA list of symbols in the segment which may be referenced by’ ca Segment and rlative addres of these symbols within the spment 48) lit of all the symbols not defined in the segment but refed by coher seznents ie) The infomation about the Tovtions of address constant, in the segment slong withthe description of how 10 revise them, ¥) Te tamed machine code along with the assigned relative adress, 3.3 : Overlay Structure | 244 Explain overlay structure. Ans Refer QS SIMU :Deeot8, Mars 2) Q45 Write short note on - Overlay structure | cane coulion of Overlay : Overlay is part ofthe program tht is me “other parts of the program. | ere ee ta td pans_of the program Requirement’ POE can be loaded “in the memory as per the =e “A Guid for Engicering Sue a0 Linkers and Loaders ‘In asemble, the overlay structure is sed in assembler, + Following Fig, Q15.1 represents the concept = Ste oS} =m) ew we cmc) (ea _ a] = Fig. @.18-1 Overlay structure is clear ftom the figure that some pans of the program can reside in the memory simultaneously. Thus the total memory allocated is 78K with overlay sructae whereas the memory requirement of entire program is 152K. Thus lot of space can be saved using overlay Ssructring, Sle = lee wl of Free “4 ‘Space Fig. 0.15.2 Sogments stored In the memory using overlay structure “A Gute for Enginerng Suter3-1 Lis gos ti f Absolute Loader 3A: Designo! ‘ sign of about Toader with Suitable eg {246 Give compete SIPPY Dee, Hoss 4 scheme the programmer and tn stot Toading shen oma Fin‘ often ein 8 IE i ode program) consists of two vey, object dock (bist as mal Bren ie, omaton i present 00 190 {ype of ea err an card and send one is transfer card et nthe text card the are machin instructions created by aseny Sling ih te nt () sess. Loader simply place | [boctons nthe memory a pei absolute adresses, + On the eranster cad the entry point of the progam i mentioned ‘The format ofthe Text card is as given below As sr4os “eR Linkers wn Loaders + Design of absolute loader canbe illutated by folowing flowchart ~ cum | + [2 | 30s | 6-7 | om | aw anes [pet [com of [adsense | Empy | tseucns] ca cau(=0 | mins ih std dia o| seq + fa te ie ease ee ae ig he i nh | in meme ae se | ae byes 2 L fot) | Spice T ‘The format of Transfer card is as follows - en eas cpars 8-72 eset 2 | as fon) n socinLoc a t c Daa | com=0 | Adsess of] Eapy | cud pete ey point spe Fig. 0.16.1 Flowchart for das as sumber - jj 3.5 : Design of Direct Linking Loader] When the catd is read ‘al itis stored in 80 continuous bytes in memory ‘the formats of ESD, RLD, TXT and END cards with O47 Explai respect to direct linking loader with suitable example. EZ [SPPU + Hy 9, Har 6)ie, fs aia Mitepetas ‘ne Tene mde ae Res. No. \. ‘given below - [tie Relative] Object eva 2 |- man [stat | |S 12,13, 14, 15 t Se fy sch ; im yee, ‘subroutine (entry eo fs Preare useful only in the high priority pr “a 2, Waiting time : The ea Weage period of time s process spends an re dling it single to implement I egures «ine Pose samy day oe sm rasan pm Se, UGS] 4 Tara tena ape Teer Pate ae oem a Tamarind tine = sien melo : te + Wag tne ive and Nowpreemptin] 4 Response tine # in difference between preemptive and in {s the time from the submission of a re 29 Bat until the fist respoaeis eee 5 rags 5. CPU utilization : I is ee ‘sverage fimction of time during wi processor is busy, 3 time ducing which the i : : 6 Rita ath ; reamptve seeding | Noxpreemptive scheduling Doses fom te san. Al te pees aust be given equal opporuniy to exces a preemptive sched 1. Priority + 1 the operat 1. | Preemptive seedling allows a | Non-proemptive scheduling eas] perating system assigns protes to processes, | process fo be interop inthe | tata proces relinquishes con| the scheduling mecbasism should. faver"the higher posnty tr of fs execton, tsking | of the CPU oaly when it Hiss processes, the CPU aay and allocating it | with its current CPU burst. 8. Predictabili ven process 10 another process. ity : A given process always should rm in about the sume anus! of tine une nr ye nods * Depeniig on te nae of operons the li dif. The CPU ution ed nought ‘Parameters. Fairness is affect by user and system, 31 What Is CPU scheduling ? Explain 2 dierent scheduling Fe oes no ats the design of | algorithms with example. 0 Keret Nonpreemptve scheduling does ‘not increase the cost Preemptive scheduling incurs 2 cost associated with acess shaved dt, ‘Ans. : CPU scheduling : + In a multiprogramming environment, usually Italo affects the design ofthe operating system Kernel more programs to be executed than could possibly be run at one ‘ime. In Preemptive scheduling is more | Simple, but very inefficient. CPU scheduling, it switches fom one proces to another process. CPU. complex resource management is commonly known as scheduling. Objective of the multprogramming is to increases the CPU utzon. CPU scheduling is one kindof fundamental operating system funtion. User program contains combination of CPU bust cycle and UO bust cycles. Process stats with CPU burst eye then 11 burst eyele again Example: Round robin method,| Example ; First come first serve 1 Gale for Engncerng States eo aaa A Gnide for Engineering Salese su Sg proces from the 1%) q| Soe a Se cr ack 5 SL ol for multiplexing proses ; ea ae er scheduling: * FIFO is _ fir ream ne) Sear scheduling emaay queve. FIFO is non-preempiy it uses a FIFO queue. systems. FIFO is i process contro block | fe system, its «When new process eaters into she from the fc r= oor ingress ae oad pace around time is unpredictable with the FIFO algorithm. Aver} omen et ihe TOS meee ee ee cite: Tet 10 ee, oo eee ee oe ‘ear pa tuance ‘depends on the scheduling policy and whether th] oe er a Se 2. Priority scheduling : * Priovty scheduling is preemptive ¢| non-preemplive. Priority of the process can be defined either internally extemally,Inerally defined priority considers the time limits, mumbe ‘open files, use of memory and use of VO devices, * External priorities are set by using external parameter of the proces] like importance of process, cost of process et. ‘* In this scheduling algorithm, CPU select hi igh it first. CP selet higher proty process fis the priority of two process is same then FCFS scheduling algorithm i applied for solving the problem. ieialae =.“ srtos “ ‘A non-preemptive prioiy = goes wit ‘ice th Say ereing te, mopentaee es See currently exesuting roses wl not change sees “Neti aoc + came ry ching gti wit pret A see ricning reer! ROC Is ghey aan ce I tt eurenty running proces, nee Drawback of prionty scheduling 1 Waiting time is mote fo CPU burst time is. fee starvation problem. lower pri "ory process Crem if there wy scheduling algorithma faces the « Starvation roblem can bs saved by mag epg tech ‘In aging, the priority of the process ue Tong time in the ready quene "Mil increase which ie waiting for Processes |" Arrival tine | Bure ine | Peary eee eae Pa) @ 2 Bs z 7 3 Pinta AE SRR ae q Ans. + Priority based (Non-preempave) a Pe 2 Pa 5 70 67 SIF preemptive TFs Tr oF * ” “A Gute for Engnerng Studentspret Sy ce a = nt — ts re ‘Waiting me Toranrest fe inn crime aati = 2) been fr oat Fe im E a ma — a sa pee » La Th FP oe " Wastng ime | rrocess] Tursaoeed ine : | _wstet re _| re ae SUF RR SuF RR _ an-preamein| "| ee ea -pecg 7 B a ew ; +101 M457 fess Seah oe rit apm 20 eens et ae : t(a |e alee g+3t0+Q/4=3 + 16+ 15+ 7/4 ‘Average waiting time SIP preempriv Average tumarousd time priority non - preemptive Average waiting time for SIF non preemptive = (0 + 8 +9 +3)/4 aus a Average tumarond tine SIF preemptive =(17+94142)/4" 7.25. | average wating time fr RR = (8 +749 +5)/4+ 33/4 = 938 4233 Draw Gantt chart and caleulate average turnaround time, ‘Average tumaround time for SIF non preemptive = (6 + 12+ 17 + average waiting time for the following proceses using SIF non 2 pecans <(6 = 2212 O14 preemptive and round robin with time quantum 2. = 41/4= 1025 SPY: Oc Ed Sem, Maks 6} | Average tumaround time for RR = (14-411 +17 + 13)/4 Sm =sis= 035 free ete a Era tive 136 What is preemptive scheduling ? Consider following example 7 a é and supply round robin scheduling policy and calculate average Ps 7 7 ‘Turnaround (TT) Time. (consider time slice = 1). % 7 A (SPP: Dec, Ed Sem, Maks 6 Fi 3 3 Process 1D | Arrival ime | Serve time + SIF Non -premeptive : Gant cha [a a 3 z 2 é Fin Py c + 4 g so ca D é 5 21 E @ 2 “Gwe Essien Sade “A Gn for EngnsrngSdetsSPOS : 4-33 Operating Syste Ans. : Gantt chart for Round Robin (Time slice = 1) Aja |e Ja ye Je Je Te fo fe le le jeje ye le 2 pP OE ee eee a tr Precess ID Waiting time Turnaround tim, A 1 4 10 16 esl ufo] s a cS Average turnaround time = SeleH ss = 10.6 END...2‘u Synchronization and Concurrency Control 5.1 Concurrency Q.1 Explain the principle of concurrency, Auns.: © Concurrent access to shared data may result in data inconsistency. Maintaining data consistency requires mechanisms to ensure the orderly execution of cooperating processes. . © Concurrency arises in the same way at different levels of execution streams. Following are the example of concurrency in different types of operating systems : 1. Concurrenrrency in multiprogramming : An interaction between multiple processes running on one CPU. 2. Concucy in multithreading: An interaction between multiple threads running in one process. 3. Concurrency in multiprocessors : An interaction between multiple CPUs running muitiple processes or threads. 4. Concurrency in multi-computers: An interaction between multiple computers running distributed processes or threads. ¢ Java is a concurrent programming language. © Process synchronization is required in uni-processor system, multiprocessor system and network. If more than one thread exists ina system at the same time, then the threads are said to-be ‘concurrent. Synchronization problems can occur whenever two or more concurrent processes use any shared resource. * Cooperating process share the logical address space. Is concurrency is possible without parallelism ? ° Yes, concurrency is possible without parallelism.a aos $2 _Synchromtation and CONCHITENEY Con than one process or thread i9 Progressin ‘it does not imply that the TOCESSCS ap allows for eoncureney ‘with more than on: i ame dees meas fe Sein nome ee er ae canon 7 ) Criteat section H) Race condition Fig. 0.21 Gritcal section * Critieal section means, process may change some common variable, writing files, updating memory location, updating a process table cts When process is accessing shared modifiable dat, it fs said to bei ‘Each process takes permission from operating system to enter into the critical section. Structure of proses ie as follows + 1, Entry setion 2, Remainder ection 3. But section. ‘+ Entry section: 11 is block of code executed in preparation for entering ‘+ Exit setion : The code executed wpon leaving the ettcal sec « Remainder section: Rest ofthe code is remainder section, <= =< sr408 Se ate et orang a * Each process cycles though remainder, entry, erica, exit sections in si) Race Condition : Race condion cei shen two or more operons crit in aa sti mam Whey ome pores eng wag shen, ar eed race condones 7 MH HS ly + Ph ae codon” ems deeds th tof he “interleaving” of the threads. ria Neen ee ee + Race conn should te avoided tease they can ase See fn applications and are difficult to debug, 54 41m banking ssom balance i» shard vibe unt, ope by bank tnployee Balance = Balance + Amount At the same time, some amount is transfer then it becomes Balance = Balance Amouat. These two lav Ins ne we vale tlc. i sro Ot nt tt eal eof +The exoares etoon of prs mt gummed 10 te Etomidate creo hr ae pf a Samay nt pore sane et 23 tapi the flying tras 1 con ein) Spcronietion dhs 1) Mutual xcaion + Mona! exclain mets we wed a cocaveat popenning bold in Stone ef Soon scour ass Ba yr by prof compas cx ele te dom, S"Requemen of muta exchsion i, wen posi PL sein eRe’ ihn eer rr ube abl se aaa cresal pes Pl hm fase! kr apn ih owes Pe + Bumpls of ssh seme ne sale onto ean ce eireintonSyoneton mes trig stem ous Oe yh nen sco Sel dat de Joc ninig te coc onsen et ‘After depositing the VO devices such as printersBe, G4 List the requirements of mutual exclusion. [pry Mey tt, Dres?> ‘Aas. : Requirements of mutusl exclusion "As any time, only one process is allowed to et ‘Solution is implememed purely in software 0” # [A process remains imide ts critical section for # No assumption can be made about relative sets © ‘concern PrOCESSES 5. A process camnct prevent any otber section 16. A process must not be indefinitely postponed fom em ‘section nd Sem, Marks 5) machine. ‘pounded time only, pepe $2: Mutual Exclusion : Hardware and Software Approach {5 How livelock situation is avaolded by using Dekker’s sotution ? ‘Ans. + + Dekker’ solution avoids « livelock snation pekkers algorithm sas the fist correct soliton to the etic section problem for two processesthresds. It uses an amy of Boolean values Sand an integer varsble 1s Dekkers algorithm is purely software solution with no special purp hardware instruction, I uses Dag to indicate a process desire to enter ial section 4 If two processes try 10 coer their cra sections atthe same time ‘variable "um indicates which of the proceses should be allowed t Be anv ale mens sae ‘using only shared memory for communication, nie Fe en ees a ree my oe ig a ig i inte ettcal section sf asynelronous process for entering into critical ering its exten 1 Here, if PI finds <2 = 0, i defers to P2 only if tum = simply silts for €2 to become I before entng it ential section, + Process P2, which is also eying 12 ime, is forced to defer to PL only if sm = I ‘Dekker'salgortim is corect and it satstes mutal exclusion and itis fe from deadlock and sarvatien Limitation of Dekkers Algorithm Dekker’ algoritim des not provide tet akemation, Dekke’salgoritan wil aot work with many moders CPUS. This algorithm wont work on Symmetric Muli-Processors (SMP) cpus. 4. Additionally, Dekkers algorithm can fll regardless of platform due to many optimizing compilers. Compller may remove writs 10 fag since never accessed in Joop. Creating an innit 100, 0.6 Write short note on : Peterson's solution ‘Ams, ¢'+ Peterson's solution is software based solution for ert section problem, This algorithm will nt work coresty on the modem computer Urchitccture, Tei simpler algorithm for two process mum exctusion with busy writing, : om atthe same 1 2 3ssiaes ge Syncroizaton end Concurrency Coy 217 loot ists complied than Dekker’ algorithm. Reson algo solves ciel section probit of 10 proce, iy. does not require any special hardware, TWO POEeSSes alten xscution between tir crcl seton and remainder sctons, *EGTRSK hd indesne possponenent are impossible in. Petey jv long as no processor dead frites unexpectedly," * For indefinite posponement 10 ecu 2c 0 contimally compete ode ther proces bisy wate, 7 Explain inter ABS. + On a Possible Batt ics sulin cone proces would have to ne risa section ‘while >t alsabling concept of mutual exclusion, irocessor system, parallel execution of process is coe onl be inctlesved. To gurantee mutual excuse "9 Prevent a process from bring intemupted. P ee Behe Fiona $A process can pe © vent the operating issuing any supervisor rie + Use interupis, 8) Implement » ©) Inroduce uninterruptible code regions ¥) Think sequentially most ofthe ime ©) Delay handting of external events ‘Econ ag | pjsadvantages Ii was on in he mee rong 2, ier ca lint end pr 4. A process waiting to ener a em ‘to tc cnt eg a Wat ae the sdvntgs and g aT brn of tot aad a ‘Advantage 1. Simple to implement 2, Ieworta well fora small mine of pcs 3, Ikea be used to suppor mute eri sean, Drawbacks : 1. If suffers from starvation. 2, There is posibilty of busy wsitng 3. There may be deadlock. G9 Explain with definition, the concept of general and binary semaphore, ‘Ans. + Concept of general semaphore : * Semaphore is described by Dist Semaphore is a noaneatve integer variable that is used as a fag. Semaphore is an opemtng sysea atbstract datatype. I takes only integer value Its wed to solve erica Section problem, * Dijkstra introduced two operons (P and V) to operate on semaphore to ‘solve process synchronization problem. A proces cally the P fpemtion when it anis to exerts eal seeion and el V ‘operation when it wants to exit it titel sion. The P operation is called as wait operation and V operation is called as signal operon, A wait operation on a semaphore decease its vale by one waits : while $ <0 ao opeSAE EE secre at concen ret 4, Merits eect Rly Se ata se ee ‘ ie ‘ey ac sor sana gper p soar dex i ace iy we Ss SUEY. Ba eg ‘nee oe te pt oy sg # 3 met. Sn Ce Soca os Se pt 57 HO eater aries once Seite fer meng to om ee & mnie sax sco sip ee vaste Ser ealy yostcs ts late Ph canibon sarily tera (contin vate cM eel pi on a condition varisble automatically does an up on the in ot 9 poor an ‘monitor and blocks the calle, 8 Wied sey ewait on Be ms tr tgs 4 5.5 : Reader Writer Problem 146 Last the Arewtwed. of . q ysl pefvvhcstoeahebemea 0.46 spain reader writer problem. a a wat mae tn Aout» When'o peso psc ss ses al es [ihin x eon a tne ine only one posers ag AY HT a such as a file or database, They called these processes reader and q Is the possibility of deadlocks int = satiate als I C886 Of es cae * Multiple readers ean coneurenly read from the data base, But when 2 Spiatng the qc, sas enly eon we ice et sie anno ede te) = * More than one reader may read shared data (no weiter). When @ ses shared data, all other writers and readers must be excluded, Geos “A Guide for Engineering Stdents= coer SM = a net vi 16 os ae Pt @~ & ron Pe Rad Se symm en a i "net eon Tepe led ee Pe EE do Sogo te ee a lane ho any Sse ee Tce We de an al ion shang = “Re Noes a the ming raison pea i Wer tn mad te dans ar mca ares + When writer is active, no other readers or writes may be active, Therefore, st must ea free muta exeusion if tee ae groups of readers and @ er, * Reader writer problem is snilar smong a set of processes. po may shice the Ale with any othe fle 'o one which a fle i t0 be shared ces wantonly to ead the fil, then it proces that also wants 10 read the * We apply rules for access order Lia writer is more readers, tn fer ders 0 be fied ont stow any 2. a reader is waiting for writer 4 JF the writer wants to append the 0 ftsh, he reader pray have acess Yo the file when ihe wi Sis, then no otter ers aes i! PO ould * irae having higher pricy than the wie ty arvaton with writers, For wre he Sher yi ving higher priory gg tll be then starvation with readers, er ory ta eae 47 Write a semaphore solution for readers - writers rotten * When two wypes of process need to 88 0 file oF database, such 0 shoe They called these procesey et ecg dey nd ‘witers may be active, Therefore, it must enforee mutual exchsion if there are groups of readeo nd serie «Reader writer problem is similer to one which a file isto be shared among. set of processes. If a process want nly to read the file, then it may share the fle with any other process that alo wants to read the file, If the writer wants to append the file, then no ether groves should Ihave access tothe file when the writer has access toi Mf) reader; 1 assume multiple instances so verter: 17 assume multiple instances jwoid reader ) © whoa) { wate readcount ++ wold wt) { Dw) ¢ waive) aowting toneading signalweem) waite pay roadcount > fk readcount: sigual(waer); signplix, }gmt COHEN Cony see ee en] [rr cme Poe 56 : Producer Cons et problem sing 37, In Sm, Marks 6) x ee er rags fem by using bing Ans: Solon to prodoesr consumer bl ry Semaphore oid main () pensions {nt consumer () me FemWaita(detey), le re) semsignal(e conmumedata Wp==0) sen } joa Ft Synchronization nd Cncureney Control J What is clea section? Give semaphore ot oamer problem, Intion for producer vee A erdeal settion i a blk of cade that only one process at a {ime cam exeeute; #0, When one process isin its real Section, ne other process may be in its eral section, The erica section problem 10 Epgure that only ONE process ata time is allowed to be opcating in iy ‘atial section ‘Producer consumer problem “is example cassie problems of synchronization. Producer process produce data item th, consumer process consumes ater. ‘Buller is used between producer and consumer. Buller size may be fixed ot variable. The produce portion ofthe application generates data and stores it in a buffer and the consumer reads data ftom the buff. + In order to synchronize these processes, both procedure and ‘consumer are. blocked on some condition. The producer is blocked when the butter is fll and the consumer is blocked when the buffer is empty 4. Code for producer 2, Code for consuiner process |S ome | eocuer i [ensues it i I tn |" ecto we HU) ‘a UB) t t produce tes (ta secu ==) rece tec tun secre tm (tm slaep (i enter itom (tem) counter = countsr + t) i (counter = = 1) ‘wakeup (consumer); if (count = = N=1) veakeup (producerh ‘consume jtom (tem):Somcontaton and ‘sraos sn : Cenenrenee “3.7 : Dining Philosopher Problem 20 Write semaphore solution for dinning Philosophers proj, SFSU D749 4 Som, sg ‘Ans. : There isone semaphore for each chopstick: 10 addition, two-phase pritation scheme is weds mder which philosophers ger thee mighors who have declared themselves "hungry." AI arith module 5 ‘yetom DINING_ PHILOSOPHERS van Me: semaphorn, nial t; for mutual axehseon */ 15k. semaphore e(, nin © 1 for synehrenteston */ fgl6l (THINK, HUNGRY, EAT, inilly THINK: 1 pilosophor fag */ 1 he before enc ps Ae beter, ench pitccopbor i on endlogs cyto of thinking and wr) F ° lake. cropesicet Enming, , ) 110 sobs eesti hci ‘ion going on so Rae procedure take ( Dowty, aoa 2= HUNORY; pene a 59 Seine ee ‘ypten0d: 1 ond erica ssction */ 4 ¢phagtt| == HUNGRY 8 pflagil-t] b= EAT ‘ae pllagil+ 1] = BAT) thon t pflagisl = BAT ‘uP(ett) } ) 11 Once « philosophor fnishos eating, al th: main eto elias thw ronourcon, Ke to chopsticks and teraby rlansa wating nolboss. void drop_chopatcketine 8) { DOWN(me), ——/* een! section */ oat), 7 Lot pl. on lft eat if posible ot) 7 Lot pl on eg oat if possible */ P(e} 1° wp exinl soction 1 > 2.24 Fxplain caslenl problem of synchronization in terms. of it Philosopher problem, £9 (PPV: Hy, End Sm, ks 51 Ans. + Dining philosophers problem one of lasial process Tiere lve philosophers are seated. around & hoking and cating. Five plates nehronization problem. creular table. They spend thee lies in 0 with ive forks are kept on the creular table + ve hsp nk hy oe, Tar Where olson desde o ex then hee must oii 4 fon, one a Teh side fond note fom et ie shoot pep ol a ingle to HE: SSS “a erieyeni and Coney, replace the for and Afr consuming fd the pisos Te18 E, TE 6 ey | Sa ms whe ok oo, SIE Py { ‘eanaot eat while the dinning philosophers 15 ate ‘ed rou. ve philosophers ‘Fig. Q21.1 shows seating amangnent of five philoso seas st 2214 Dinning philosophers + Ga der the following code oid inning Palosoper () | te rue) eid ating () 4 takolehtor ¢ ‘honigtone ¢ esticeddlay) atsighttore (, Putettoce ( } oe Svechronestion end Concurency Cone! ce i (an es wpe Gs He ee a ot mcr 4 it This isnot possible because ofthe fellowieg i = Ee} +e l ze} Fig, 0.24.2 Philosopher states ‘Each philosopher will hold exactly one fork, and 20 available on the table. The philosophers will ll deadlock, To solve this problem, write some extra code in takerightfork (-) procedure, User can specify the philosophers to put down the lel fork if that philosophers cannot obtain the right fork. Problem analysis # Shaved resource : Here each fork i shared by two philosophers so we called i is a shared resource, will remain + Race condition We do not want philosopher to pick up «fork that hha already been picked up by his neighbor. Ths isa race condition. * Solution + We consider each fork as a shared item and protected by mutex lock, Before eating, each philosopher fist lock lef fork and then right fork. If philosopher scquires both locks sucessful, then philosophers have two locks (ovo fork) so he ean eat a fod, Aer finishes'easting, this philosopher releases both for, and thinks Some other alternatives 1. Pick up the left fork, ifthe right fork isnot available fora given time, put the Jet fork down, wait and try again. Even if each philosopher ‘waits different random time, an unlucky pilosopher may starve. 2. Requie all philosophers to acquire a binary semaphore before picking ‘up any forks. This guarantees that no philosopher staves but limits parallelism damatialy “ie for Eaierng SesP| sre0s se Src cy, pstik, cach cho ‘Because we need to lok and unlock a chopstick, tig fot ct ines, ead 0 cat fe ted op ich philosopher. Since each philosopher must have access tq the tye =e Sanit saa annees IMMes locks ae global vale, 5.8 : Deadlocks ; Principle of Deadlock 4s deadlock 9 2.22 Whar ‘State and explain the conditions fy, aeadtock, PIS: Ma nd Sem, Hs OR What are ‘necessary conditions for deadlock ? Explain the cae eadlock ? Explain them with STP: be, Ed Sem, rs 4) pre Desilock can be dtd asthe manent blocking of Processes that, +r: Day Om sharale essary Pfr ste resus. Dealoek ea ecu TA process is in ae ceca Sa Ht ms wating for «penal event aan Sem dae, ono mae pretest * Following four conditions ae neces 2 Mata excision 2 Ho ag vat a for deadlock to exis, 3. No preemption 1. Mutual exch One process at atime 2 Hold and visit + ranted cartier 4. Citeular wait TesouCe may be acquired ‘rehasively by only ty ay HES ye 3 No preemption +o cannot. remove jg aa Som the pce conta. A te system finished using the resounee se ws es ar 4 Glreular wait: A circular ein g Grr nO ld a ing eit in All four of these otis ms be pre fo «ye Occur. If one Of them i “abser 7 ie 4 no source dei deg ong ‘Mince ‘i 8 it inde ¥ 1 A (oct may Vat indefnitly teense aber proces o srr gin and geting the requested ‘koe Tesources before tis process s. Sitch lin, mes & i invol Fa fe ropes. Ina starvation siuation 4 sats mt being allowed exeeae ‘q24 Write short notes on : ‘Resource allocation graph and resource ssiocation graph algorithm, Fig, 264 Process Bal lu eoa| Fig, 242 Revourse “sh 8 instances. etal re base invine of te x has been allocated fo the proc ‘ + Regus ig coed a some ln HS allocated (othe process. When Smet lp econ i oo rtre _sycronition and COUCH Coy, 5-35 Ssctroieion Secon ET Coty nd Concerncy Conta sags st j vlsey_ ns F SMES OEE Ry tl eg | Acie SSE OEE Rs wren roe ee onired oe @==GEs | pesto (6) Claim edge lope a osu 1 lease by Fhe aint sige 8 i (a) Request " cried to a claim edge P; +R. con" ues can be granted only if converting the Pai ng the request ge to an eat edge does not result in the formation ‘fs ce ie ae location graph. re ig, 0243 Edges 0.244 shows «resource allocation graph. SYSIEM consis. gy 5.9 : Deadlock Prevention 28 Explain deadlock prevention techniques with example, ISU ay 6, En Sam, 5) ‘ans: © To prevent a deadlock, the OS must eliminate one ofthe four recessary conditions TeMutual exclusion 2. Hold and wait 3.No preemption 4. Circular wat 1, Mutual exclusion = It is necessary in any computer system becuse some resources (memory, CPU) must be exclusively aloested to one ser st a time, No other process can we = resource while i is ces sp, p, TOMA Rte tector gph | atocsted to process. Re es 2. Hold and walt If a proces holding eorain resources is dened rae RRs further request, i must release its original resources and if requied Resouree allocation graph stgorithm + ee orally requst ede and assignment sage je yee alsaon. graph, | 3. No. preemption : It could be bypassed by allowing the operating In addon 10 the request and s system to dallocate resources fom proces. ssignment ee, new ede called aay 4. Cirewlar wait ¢ Circular wait can be bypassed ifthe oprtng system on ete is added, Foe “SUID prevents the formation ofa eile Sram cali eg indeer tht proces 9 rete rue fone tg inte we, “tie vm doe esi Bae. Pig 0.045 ae idk svete vith eg ‘sllocation graph, a) +A. deadlock is possible only if all four of hese conditions ‘simultaneously holdin the sytem. : + Prevention straegies ensure that at least one ofthe conditions is always false 28 How can the hold-and wale condition be prevented ? ‘Ans. + To prevent holé-and wait condition, prooeses must be prevented fiom holding one or more resources while sinullnesesly wating for one for more others. =S_—.sra0s. Thc ae seer pss fr thi 1 Regi taal proses request al SORES atone ting Ferma of sem Tens 4 CESS needs yg My) enly ints exeoton and doesnt ned some other reg) che ~ 2. Regue that cet Blig rescures must release i ee ee ae ae Along wh he nw ones in 2 single NEW TqUEL, This CH protlen if eros hs pri completed an operating "sare an then flo gt it relocated afer releasing i, "4 ‘Bier of the methods decried above can lead to sarvayj Foss ees on moe ppl eu io 0 : Deadlock Avoidance te adic ie it abe xg z OR Write Banker's ‘etre ad expuin with suitable example TIS: beet, End Sem, Mas OR Write and ex lal Denker lpr forevidance of deadlace Abe: tr c/n ite is ened bese de tn ane ae A, a. = me ‘modeled after banker who ‘ae ; "she pom tae med Algriin is eek a se it ‘state, If it does, the rec Save see fe de ig ens an . Tes lads +The Dist a proposed an algorit tid desde The ba avoidance method, " * By using avoidance m peer ries rt em ta ey dase) reck if @ deadlock is created Sah sa analysis method is used for th » Sri hs. be a reques ‘ ra et tM saan known of the st Syehrenzon en Conary Cot ps0 3 cnte «These 1a stone Sec of resus 1 st et des ot result in a dalek. a safe state only if there exist a safe sequence. A safe sem i ce. A safe state psa site an ona Se. dos mem te ten in 2 locations to deadlock. «ps ong 8 the state Safe, te resource marae cn be gaat 0 void 8 deadlock lly the system i in a safe state, When process equests resource hat resource is available then the system must decide whether the or process must wai allocting resource then only OS i and t fesources can be allocated immedi ‘eae system remains in safe state a allocates tesourees 10 Protss. a Banker algoritim uses following dat sucrures, 4 Allocation : Allocation isa tbl in which row represents process and column represents resources (R) allo {i j] = Number of unit of re sae Ry hed by process Py 2.Max: Max be the maximum number of resources that posts requires during its execution ‘Need (claim) + It is curet js equal to its maximum need minus is Need = Max ~ Allocation ‘Available: There will be sumber of © ftocaion. This is equivalent tothe total cumber of Sham of the allocation 1 all processes inthe sytem, ‘Number of resourses ~ Sum ofthe allocation Jaim of a proses, where process's cai ‘current allocation. sources sill available for resources minus the punter ores - Alon) th total number of resource in fat once alloested & more that ‘aust also guarantee That resource to the sytem within & «Each process cannot rues the system. Bash process resource the process il nie ime, * Deadlock: avoidance Fon eg ianee wes the worst-case anal adc M8 ey os "0 check for. Gone 3