D686 Operating Systems for Computer Scientists - Set 1 - Part 2

Correct Answer: Buffer overflow

Buffer overflow occurs when programs write beyond allocated memory boundaries, potentially overwriting return addresses to execute malicious code with elevated privileges.
Option a is incorrect because SQL injection targets database queries through malformed input, not memory buffer boundaries in program execution space directly.
Option c is incorrect because session hijacking steals authenticated session tokens to impersonate users, not exploiting memory allocation vulnerabilities in programs.
Option d is incorrect because logic bombs are malicious code triggered by specific conditions or times, not exploitation of buffer boundary checking failures.

Correct Answer: Shared address space

Threads share the same address space within a process, eliminating memory copying overhead and enabling efficient communication between concurrent execution units.
Option a is incorrect because separate memory spaces require inter-process communication mechanisms, making it less efficient than threads' shared memory model approach.
Option c is incorrect because threads actually share file descriptors with their parent process, not maintaining independent ones which would increase overhead.
Option d is incorrect because isolated execution contexts would prevent efficient resource sharing that makes threads lighter weight than full processes.

Correct Answer: Ring zero

Ring zero has highest privileges where kernel executes with complete hardware access, forming the trusted base for system security and protection.
Option a is incorrect because ring three has least privileges where user applications run with restricted access to protect system integrity and security.
Option b is incorrect because ring two has intermediate privileges historically used for device drivers, less privileged than kernel's ring zero level.
Option c is incorrect because ring one was intended for system services with more than user but less than kernel privileges in design.

Correct Answer: Internal fragmentation

Internal fragmentation wastes space within allocated partitions when process size is smaller than the fixed partition size, creating unusable gaps.
Option a is incorrect because external fragmentation occurs between allocated blocks when free space becomes scattered in small unusable pieces throughout memory.
Option c is incorrect because memory leaks happen when programs fail to release allocated memory after use, different from size mismatch waste.
Option d is incorrect because page faults occur in virtual memory when accessing pages not in physical memory, unrelated to partition size waste.

Correct Answer: Eliminating hold and wait

Requiring all resources upfront eliminates hold and wait condition since processes cannot hold some resources while waiting for others during execution.
Option a is incorrect because eliminating mutual exclusion isn't always possible since some resources like printers inherently require exclusive access for correctness.
Option c is incorrect because allowing preemption means forcibly taking resources from processes, different from requesting all resources initially before starting.
Option d is incorrect because preventing circular wait involves ordering resource requests, not requiring all resources to be obtained before execution begins.

Correct Answer: File management

File management system calls handle file operations including create, delete, open, close, read, and write operations for persistent storage manipulation.
Option a is incorrect because process control system calls manage process lifecycle including creation, termination, and execution control, not file operations directly.
Option c is incorrect because device management system calls handle hardware device requests and releases, not file creation in the file system.
Option d is incorrect because information maintenance system calls get or set system data like time and system configuration, not file operations.

Correct Answer: Direct memory access

DMA controllers transfer data between devices and memory independently after CPU setup, freeing processor for other tasks during transfer operations.
Option a is incorrect because programmed I/O requires CPU to control every data transfer byte explicitly, consuming processor cycles throughout entire operation.
Option b is incorrect because interrupt-driven I/O still needs CPU to handle each data transfer when interrupted, though more efficient than polling.
Option d is incorrect because memory-mapped I/O treats device registers as memory addresses for access but doesn't eliminate CPU involvement in transfers.

Correct Answer: Range of values

Counting semaphores can hold any non-negative integer for resource counting while binary semaphores only hold zero or one for mutual exclusion.
Option a is incorrect because operation speed is similar for both types since they use same wait and signal primitives regardless of value range.
Option b is incorrect because both semaphore types can have multiple processes waiting; the difference lies in resource count representation, not waiters.
Option d is incorrect because memory requirements differ negligibly between integer and boolean storage, not the defining characteristic between these types.

Correct Answer: SCAN algorithm

SCAN moves the arm from one end to the other servicing requests along the way, like an elevator serving all floors before reversing.
Option a is incorrect because FCFS services requests in arrival order without considering arm position, potentially causing excessive seek times between requests.
Option b is incorrect because SSTF always chooses the closest request next, which may cause starvation but doesn't follow directional sweep pattern.
Option d is incorrect because LOOK reverses direction at last request in each direction rather than going to physical end like SCAN does.

Correct Answer: Uniform interface

VFS provides consistent API for applications regardless of underlying file system type, enabling seamless access to NTFS, EXT4, or FAT systems.
Option a is incorrect because VFS adds abstraction layer that may slightly decrease performance rather than improving file access speed directly overall.
Option c is incorrect because security depends on underlying file system implementation and permissions, not the VFS abstraction layer itself primarily.
Option d is incorrect because storage capacity is determined by physical media and file system limits, not affected by VFS abstraction layer.

Correct Answer: Condition variables

Condition variables in monitors allow threads to wait for specific conditions and be signaled when conditions change, enabling synchronized access coordination.
Option a is incorrect because spin locks cause busy waiting by continuously checking conditions, which monitors specifically avoid through blocking wait mechanisms.
Option c is incorrect because message queues facilitate inter-process communication through message passing, not monitor's internal thread synchronization for shared data.
Option d is incorrect because signal handlers respond to asynchronous events from the operating system, not monitor's synchronization conditions between threads.

Correct Answer: Optimal algorithm

Optimal algorithm replaces the page used furthest in future, requiring impossible foreknowledge of access patterns, serving only as theoretical performance benchmark.
Option a is incorrect because FIFO simply replaces oldest page without needing future knowledge, though it may suffer from Belady's anomaly occasionally.
Option b is incorrect because LRU tracks past usage to predict future behavior, implementable through various techniques without requiring future knowledge directly.
Option d is incorrect because clock algorithm approximates LRU using reference bits in circular fashion, practical to implement without future knowledge requirements.

Correct Answer: Heap section

Heap section grows upward for dynamic allocation during runtime when programs explicitly request memory through allocation functions like malloc or new.
Option a is incorrect because text section contains executable code instructions and is typically read-only, not used for dynamic memory allocation requests.
Option b is incorrect because data section stores global and static variables with sizes known at compile time, not runtime dynamic allocations.
Option d is incorrect because stack section handles function calls, local variables, and parameters with automatic allocation, not explicit dynamic memory requests.

Correct Answer: Contiguous allocation

Contiguous allocation places files in consecutive blocks enabling fast sequential access but causes external fragmentation as files are created and deleted.
Option b is incorrect because linked allocation chains blocks with pointers, eliminating external fragmentation but providing poor random access performance for files.
Option c is incorrect because indexed allocation uses index blocks to track file blocks, supporting random access without requiring consecutive block placement.
Option d is incorrect because multilevel indexing extends indexed allocation for large files using multiple index levels, not consecutive block allocation method.

Correct Answer: Master-slave relationship

Asymmetric multiprocessing assigns one processor as master running OS while others execute user tasks, creating hierarchical control unlike symmetric systems' equality.
Option a is incorrect because both asymmetric and symmetric systems can share memory; the distinction lies in processor roles, not memory architecture.
Option c is incorrect because clock synchronization is necessary for all multiprocessor systems regardless of symmetric or asymmetric organization for coordinated operation.
Option d is incorrect because cache coherence protocols maintain data consistency in all multiprocessor systems, not distinguishing asymmetric from symmetric designs.

Correct Answer: Upon first modification

Copy-on-write delays copying shared pages until a process attempts modification, saving memory and time when pages remain unmodified between processes.
Option a is incorrect because fork initially shares pages between parent and child, marking them copy-on-write rather than immediately copying them all.
Option c is incorrect because process termination releases pages rather than copying them, as the process no longer needs the memory pages.
Option d is incorrect because context switches save and restore process state without triggering page copying in copy-on-write mechanism between processes.

Correct Answer: Recently referenced pages

Working set contains pages referenced within recent time window, representing the process's locality of reference and minimum memory needed for efficiency.
Option a is incorrect because total allocated memory includes all pages assigned to process, not just the actively used subset in current phase.
Option c is incorrect because available free frames represent system-wide unallocated memory, not the process-specific set of active pages being referenced.
Option d is incorrect because page fault frequency measures thrashing symptoms, while working set identifies the pages needed to prevent thrashing occurrence.

Correct Answer: Through role assignment

RBAC assigns users to roles and permissions to roles, simplifying management by changing role permissions rather than individual user permissions directly.
Option a is incorrect because direct user permissions characterize DAC where owners control access, becoming unmanageable in large organizations with many users.
Option c is incorrect because access lists are implementation mechanisms for various models, not the distinguishing feature of RBAC versus DAC approaches.
Option d is incorrect because capability tokens are another implementation approach for permissions, not the role-based indirection that characterizes RBAC systems fundamentally.

Correct Answer: Kernel mode only

Device drivers run in kernel mode to access hardware directly through privileged instructions while being trusted components of operating system architecture.
Option a is incorrect because user mode lacks privileges for direct hardware access that device drivers require for controlling physical devices effectively.
Option c is incorrect because drivers don't switch between modes dynamically; they operate consistently in kernel space for hardware access privileges needed.
Option d is incorrect because hypervisor mode relates to virtualization platforms running multiple operating systems, not device driver execution within single OS.

Correct Answer: Queues output data

Spooling queues output in disk files allowing processes to continue while devices process data at their own pace asynchronously from process execution.
Option a is incorrect because spooling doesn't change device speed; it decouples process execution from slow device operation through intermediate storage buffering.
Option c is incorrect because spooling doesn't create device duplicates but uses disk as intermediate storage for device data queueing mechanism.
Option d is incorrect because spooling focuses on asynchronous operation through queueing, not data compression which is separate optimization technique for storage.