Sie sind auf Seite 1von 5

exclusion and synchronize concurrent

Operating Systems processes.


Following are a few basic questions 4. What is thrashing?
that cover the essentials of OS: It is a phenomenon in virtual
memory schemes when the processor
1. Explain the concept of spends most of its time swapping
Reentrancy. pages, rather than executing
It is a useful, memory-saving instructions. This is due to an
technique for multiprogrammed inordinate number of page faults.
timesharing systems. A Reentrant 5. List the Coffman's conditions that
Procedure is one in which multiple lead to a deadlock.
users can share a single copy of a  Mutual Exclusion: Only one
program during the same period. process may use a critical resource
Reentrancy has 2 key aspects: The at a time.
program code cannot modify itself,
 Hold & Wait: A process may be
and the local data for each user process
allocated some resources while
must be stored separately. Thus, the
waiting for others.
permanent part is the code, and the
temporary part is the pointer back to  No Pre-emption: No resource can
the calling program and local variables be forcible removed from a process
used by that program. Each execution holding it.
instance is called activation. It  Circular Wait: A closed chain of
executes the code in the permanent processes exist such that each
part, but has its own copy of local process holds at least one resource
variables/parameters. The temporary needed by another process in the
part associated with each activation is chain.
the activation record. Generally, the 6. What are short-, long- and
activation record is kept on the stack. medium-term scheduling?
Note: A reentrant procedure Long term scheduler
can be interrupted and called by an determines which programs are
interrupting program, and still execute admitted to the system for processing.
correctly on returning to the procedure. It controls the degree of
2. Explain Belady's Anomaly. multiprogramming. Once admitted, a
Also called FIFO anomaly. job becomes a process.
Usually, on increasing the number of Medium term scheduling is
frames allocated to a process' virtual part of the swapping function. This
memory, the process execution is relates to processes that are in a
faster, because fewer page faults occur. blocked or suspended state. They are
Sometimes, the reverse happens, i.e., swapped out of real-memory until they
the execution time increases even are ready to execute. The swapping-in
when more frames are allocated to the decision is based on memory-
process. This is Belady's Anomaly. management criteria.
This is true for certain page reference Short term scheduler, also
patterns. know as a dispatcher executes most
3. What is a binary semaphore? frequently, and makes the finest-
What is its use? grained decision of which process
A binary semaphore is one, should execute next. This scheduler is
which takes only 0 and 1 as values. invoked whenever an event occurs. It
They are used to implement mutual may lead to interruption of one process
by preemption.
7. What are turnaround time and 11. When is a system in safe state?
response time? The set of dispatchable
Turnaround time is the interval processes is in a safe state if there
between the submission of a job and its exists at least one temporal order in
completion. Response time is the which all processes can be run to
interval between submission of a completion without resulting in a
request, and the first response to that deadlock.
request. 12. What is cycle stealing?
8. What are the typical elements of a We encounter cycle stealing in
process image? the context of Direct Memory Access
 User data: Modifiable part of user (DMA). Either the DMA controller can
space. May include program data, use the data bus when the CPU does
user stack area, and programs that not need it, or it may force the CPU to
may be modified. temporarily suspend operation. The
 User program: The instructions to latter technique is called cycle stealing.
be executed. Note that cycle stealing can be done
 System Stack: Each process has only at specific break points in an
one or more LIFO stacks instruction cycle.
associated with it. Used to store 13. What is meant by arm-stickiness?
parameters and calling addresses If one or a few processes have
for procedure and system calls. a high access rate to data on one track
 Process control Block (PCB): Info of a storage disk, then they may
needed by the OS to control monopolize the device by repeated
processes. requests to that track. This generally
9. What is the Translation happens with most common device
Lookaside Buffer (TLB)? scheduling algorithms (LIFO, SSTF,
In a cached system, the base C-SCAN, etc). High-density
addresses of the last few referenced multisurface disks are more likely to
pages is maintained in registers called be affected by this than low density
the TLB that aids in faster lookup. ones.
TLB contains those page-table entries 14. What are the stipulations of C2
that have been most recently used. level security?
Normally, each virtual memory C2 level security provides for:
reference causes 2 physical memory  Discretionary Access Control
accesses-- one to fetch appropriate  Identification and Authentication
page-table entry, and one to fetch the  Auditing
desired data. Using TLB in-between,  Resource reuse
this is reduced to just one physical 15. What is busy waiting?
memory access in cases of TLB-hit. The repeated execution of a
10. What is the resident set and loop of code while waiting for an event
working set of a process? to occur is called busy-waiting. The
Resident set is that portion of CPU is not engaged in any real
the process image that is actually in productive activity during this period,
real-memory at a particular instant. and the process does not progress
Working set is that subset of resident toward completion.
set that is actually needed for 16. Explain the popular
execution. (Relate this to the variable- multiprocessor thread-scheduling
window size method for swapping strategies.
techniques.)
 Load Sharing: Processes are not the relevant process' address space.
assigned to a particular processor. Global replacement policy allows any
A global queue of threads is page frame from any process to be
maintained. Each processor, when replaced. The latter is applicable to
idle, selects a thread from this variable partitions model only.
queue. Note that load balancing 20. Define latency, transfer and seek
refers to a scheme where work is time with respect to disk I/O.
allocated to processors on a more Seek time is the time required to
permanent basis. move the disk arm to the required
 Gang Scheduling: A set of related track. Rotational delay or latency is the
threads is scheduled to run on a set time it takes for the beginning of the
of processors at the same time, on required sector to reach the head. Sum
a 1-to-1 basis. Closely related of seek time (if any) and latency is the
threads / processes may be access time. Time taken to actually
scheduled this way to reduce transfer a span of data is transfer time.
synchronization blocking, and 21. Describe the Buddy system of
minimize process switching. Group memory allocation.
scheduling predated this strategy. Free memory is maintained in
 Dedicated processor assignment: linked lists, each of equal sized blocks.
Provides implicit scheduling Any such block is of size 2^k. When
defined by assignment of threads to some memory is required by a process,
processors. For the duration of the block size of next higher order is
program execution, each program chosen, and broken into two. Note that
is allocated a set of processors the two such pieces differ in address
equal in number to the number of only in their kth bit. Such pieces are
threads in the program. Processors called buddies. When any used block
are chosen from the available pool. is freed, the OS checks to see if its
 Dynamic scheduling: The number buddy is also free. If so, it is rejoined,
of thread in a program can be and put into the original free-block
altered during the course of linked-list.
execution. 22. What is time-stamping?
17. When does the condition It is a technique proposed by
'rendezvous' arise? Lamport, used to order events in a
In message passing, it is the distributed system without the use of
condition in which, both, the sender clocks. This scheme is intended to
and receiver are blocked until the order events consisting of the
message is delivered. transmission of messages. Each system
18. What is a trap and trapdoor? 'i' in the network maintains a counter
Trapdoor is a secret Ci. Every time a system transmits a
undocumented entry point into a message, it increments its counter by 1
program used to grant access without and attaches the time-stamp Ti to the
normal methods of access message. When a message is received,
authentication. A trap is a software the receiving system 'j' sets its counter
interrupt, usually the result of an error Cj to 1 more than the maximum of its
condition. current value and the incoming time-
19. What are local and global page stamp Ti. At each site, the ordering of
replacements? messages is determined by the
Local replacement means that following rules: For messages x from
an incoming page is brought in only to site i and y from site j, x precedes y if
one of the following conditions location on that page is actually
holds....(a) if Ti<Tj or (b) if Ti=Tj and referenced during execution. With pre-
i<j. paging, pages other than the one
23. How are the wait/signal operations demanded by a page fault are brought
for monitor different from those for in. The selection of such pages is done
semaphores? based on common access patterns,
If a process in a monitor signal especially for secondary memory
and no task is waiting on the condition devices.
variable, the signal is lost. So this 27. Paging a memory management
allows easier program design. Whereas function, while
in semaphores, every operation affects multiprogramming a processor
the value of the semaphore, so the wait management function, are the two
and signal operations should be interdependent?
perfectly balanced in the program. Yes.
24. In the context of memory 28. What is page cannibalizing?
management, what are placement Page swapping or page
and replacement algorithms? replacements are called page
Placement algorithms cannibalizing.
determine where in available real- 29. What has triggered the need for
memory to load a program. Common multitasking in PCs?
methods are first-fit, next-fit, best-fit.  Increased speed and memory
Replacement algorithms are used when capacity of microprocessors
memory is full, and one process (or together with the support fir virtual
part of a process) needs to be swapped memory and
out to accommodate a new program.  Growth of client server computing
The replacement algorithm determines 30. What are the four layers that
which are the partitions to be swapped Windows NT have in order to
out. achieve independence?
25. In loading programs into memory,  Hardware abstraction layer
what is the difference between  Kernel
load-time dynamic linking and  Subsystems
run-time dynamic linking?
 System Services.
For load-time dynamic linking:
31. What is SMP?
Load module to be loaded is read into
To achieve maximum
memory. Any reference to a target
efficiency and reliability a mode of
external module causes that module to
operation known as symmetric
be loaded and the references are
multiprocessing is used. In essence,
updated to a relative address from the
with SMP any process or threads can
start base address of the application
be assigned to any processor.
module.
32. What are the key object oriented
With run-time dynamic
concepts used by Windows NT?
loading: Some of the linking is
postponed until actual reference during  Encapsulation
execution. Then the correct module is  Object class and instance
loaded and linked. 33. Is Windows NT a full blown object
26. What are demand- and pre- oriented operating system? Give
paging? reasons.
With demand paging, a page is No Windows NT is not so,
brought into memory only when a because its not implemented in object
oriented language and the data In Windows NT a mutant
structures reside within one executive provides kernel mode or user mode
component and are not represented as mutual exclusion with the notion of
objects and it does not support object ownership.
oriented capabilities . 41. What is an idle thread?
34. What is a drawback of MVT? The special thread a dispatcher
It does not have the features will execute when no ready thread is
like found.
 ability to support multiple 42. What is FtDisk?
processors It is a fault tolerance disk
 virtual storage driver for Windows NT.
 source level debugging 43. What are the possible threads a
35. What is process spawning? thread can have?
When the OS at the explicit  Ready
request of another process creates a  Standby
process, this action is called process  Running
spawning.  Waiting
36. How many jobs can be run  Transition
concurrently on MVT?  Terminated.
15 jobs
37. List out some reasons for process 44. What are rings in Windows NT?
termination. Windows NT uses protection
 Normal completion mechanism called rings provides by
 Time limit exceeded the process to implement separation
 Memory unavailable between the user mode and kernel
 Bounds violation mode.
 Protection error 45. What is Executive in Windows
 Arithmetic error NT?
 Time overrun In Windows NT, executive
 I/O failure refers to the operating system code that
runs in kernel mode.
 Invalid instruction
46. What are the sub-components of
 Privileged instruction I/O manager in Windows NT?
 Data misuse  Network redirector/ Server
 Operator or OS intervention  Cache manager.
 Parent termination.  File systems
38. What are the reasons for process
 Network driver
suspension?
 Device driver
 swapping
47. What are DDks? Name an
 interactive user request operating system that includes
 timing this feature.
 parent process request DDks are device driver kits,
39. What is process migration? which are equivalent to SDKs for
It is the transfer of sufficient writing device drivers. Windows NT
amount of the state of process from includes DDks.
one machine to the target machine 48. What level of security does
Windows NT meets?
40. What is mutant? C2 level security.

Das könnte Ihnen auch gefallen