Tentamen i TDDD82 Säkra mobila system (Systemprogramvara)

Size: px

Start display at page:

Download "Tentamen i TDDD82 Säkra mobila system (Systemprogramvara)"

Donald Mills
5 years ago
Views:

1 Tentamen i TDDD82 Säkra mobila system (Systemprogramvara) Inga hjälpmedel är tillåtna. Kom ihåg att svaren på samtliga uppgifter måste MOTIVERAS, och att motiveringarna skall vara uppställda på ett sådant sätt att det går att följa hur Du tänkt. OMOTIVERADE SVAR GER 0 POÄNG OM INGET ANNAT SÄGS. Ansvarig: Mikael Asplund (nåbar på tel ). Maxpoäng är 30 poäng. För betyg 3 krävs minst 15 poäng, för betyg 4 krävs 20 poäng och för betyg 5 krävs 25 poäng. Lycka till!!! 1

2 1. Consider an ordinary hash table T of size M designed for a single-threaded environment. A hash function h that maps data items to the range 0,..., M 1 is given. Data items u, v hashed to the same hash value h(u) = h(v) = i are chained in a singly-linked list, such that list T [i] contains all items with hash value i (see the picture). Each list item contains a data element and a next pointer to the next item in the list (NULL for the last element in a list). Each T [i] points to the first item in its list (or T [i] is NULL if the ith list is empty), i = 0,..., M 1. T 0 M 1 A sequential implementation is given with pseudocode below (assuming garbage collection). The operation insert(t, u) inserts a data item u into hash table T at the beginning of list T [h(u)]: void insert ( hash table T[], data item u ) int i = h(u); e = new list element; e->item = u; e->next = T[i]; T[i] = e; The operation remove(t, u) removes the first element in T that matches u. void remove ( has table T[], data item u ) int i = h(u); e = first element in T[i] while e!= NULL if (e->next!= NULL) and (e->next->item == u) e->next = e->next->next; break; e = e->next; 2

3 Your task will now be to make the hash table implementation thread-safe to allow a hash table stored in shared memory to be accessed properly by multiple threads: (a) Identify the critical section(s) in the pseudocode and suggest how to protect them properly against race conditions by using a single mutex lock. Show the revised (pseudo) code. (b) What is the disadvantage of a single-lock solution if there are many threads accessing T frequently, and why? (c) Describe a multi-lock approach that solves that problem. (1 points) 2. (a) Define the terms process and thread. In particular, what are the main differences between processes and threads, and what do they have in common? Be thorough! (b) Given the following (Unix) C program: #include <stdio.h> #include <unistd.h> int main() printf("hello\n"); return 0; How often is Hello printed to the standard output when executing this program? Explain your answer carefully (just guessing the right number gives no points). (c) How does an operating system prevent a process from monopolizing a processor, and what hardware support is required for that? (1 points) 3

4 3. Consider the following resource allocation problem in a system with 3 resources (R1-R3), and 4 processes (P1-P4). The table indicates the currently allocated resources and in parenthesis the maximum possible demand. R1 R2 R3 P1 0 (0) 1 (1) 0 (1) P2 3 (4) 1 (1) 0 (0) P3 0 (5) 1 (3) 0 (2) P4 4 (6) 3 (3) 1 (5) The currently available resources are: [2, 0, 4]. Use Banker s algorithm to determine if the request [2, 0, 2] from Process P3 should be granted. 4. (a) Exponential averaging can be used to predict the length of the next CPU burst by considering historical data. The following formula is used τ n+1 = αt n + (1 α)τ n, where t n is the actual length of the nth CPU burst and τ n+1 is the predicted value for the next CPU burst. How should the parameter α be interpreted? What possible values can α have and what is the interpretation of the extreme values? (3 points) (b) Applications can have varying QoS requirements. Explain what it means if an application is tolerant or non-tolerant, and give one example of each type of application. 4

5 5. Consider the three timelines for nodes A, B and C in the figure below. The arrows between the lines indicate messages as they are sent and received by the respective nodes. Use Lamport s locgical clock algorithm to set timestamps for the events a-j. P1 a b c P2 d e f g P3 h i j Based on these timestamps create a consistent cut where at least one message has the send and receive on different sides of the cut. Motivate why the cut is consistent. 6. It has been shown that in order to reach majority agreement in a set of replicas in a distributed system where some nodes might crash, one must assume the so called synchronous system model. (a) Explain the meaning of this model (2p) (b) Explain why it can be difficult to fulfil these assumptions for a system which is spread out across a wide area (e.g., multiple cities). (2p) (c) Give an example of when this system model is suitable (1p) 5

TDDB68 Processprogrammering och operativsystem / Concurrent programming and operating systems

TENTAMEN / EXAM TDDB68 Processprogrammering och operativsystem / Concurrent programming and operating systems 2017-06-05 Examiner: Mikael Asplund (0700895827) Hjälpmedel / Admitted material: Engelsk ordbok