Howdy Class --

Here is the third homework assignment.  If you are taking the
class for elective credit in the CS Major, do all the problems
-- Parts I and II.

If you are taking the class for credit, but not as an elective
in the CS Major, just do the first set of problems -- the ones
for the Nemeth book -- Part I.

The problems are due midnight Friday, April 18.  I prefer that
you just fill out the answers in a copy of this e-mail and send
it to me by e-mail.

Many of these problems can be answered well with just a
sentence or two.  Some require a bit of thought, and some extra
wording.  Use good judgement.  That's a part of what I will be
grading you on.  Answers that offer inadequate information
and/or aid to understanding will receive little credit, even if
technically correct.  

Remember that the point of these exercises is for you to
demonstrate that you have read and understood the material.
Therefore always give the "textbook" answer (Nemeth or Daemon,
as the case may be) along with any individual opinions or
answers that you want to put forward. 

						-- john
						
//////////////////////////////////////////////////
PART I: Homework #3 Problems for Nemeth
          Chapter 14: TCP/IP and Routing
          Chapter 25: Printing and Imaging
//////////////////////////////////////////////////

///////////////////////////////////////////////////////////
Questions on Chapter 14 of Nemeth -- TCP/IP and Routing
///////////////////////////////////////////////////////////

-----------------------------------------------------------------
1.
In the context of TCP/IP networking, what is routing and
what are routers?  For example, what essentially is the kind of
decision that a router has to make, and what essentially
distinguishes a host that needs to do non-trivial routing from
a host that needs only to do trivial routing?
-----------------------------------------------------------------
2.
Explain in simple terms what actually happens when a host
makes an ARP request.
-----------------------------------------------------------------
3.
You do a ping, and this is what happens:

% ping ishi
PING ishi.csustan.edu: 56 data bytes
^C
----ishi.csustan.edu PING Statistics----
11 packets transmitted, 0 packets received, 100% packet loss

What is the probable explanation of this 100% packet loss?  Why
did it happen?
-----------------------------------------------------------------

///////////////////////////////////////////////////////////
Questions on Chapter 25 of Nemeth -- Printing and Imaging
///////////////////////////////////////////////////////////

-----------------------------------------------------------------
4.
Describe how lpr and lpd cooperate to print a file.  Which
process does what, and when and where?  Give details.
-----------------------------------------------------------------

/////////////////////////////////////////////////////////////////
PART II: Homework #2 Problems for The Kernel (Daemon) Book.
           Chapter 04: Process Management
/////////////////////////////////////////////////////////////////

//////////////////////////////////
Chapter 05 -- Memory Management
//////////////////////////////////

-----------------------------------------------------------------
1.
What is copy-on-write?  In most UNIX applications, the fork
system call is followed almost immediately by an exec system
call.  Why does this behavior make it particularly attractive
to use copy-on-write in implementing fork?
-----------------------------------------------------------------
2.
When a 4.4BSD process gets a page fault, what is the data
structure that gets searched, and how does the kernel know
whether the page fault was caused by a valid page not resident
or an invalid page?
-----------------------------------------------------------------
3.
Describe the main sequence of events that occurs when a
4.4BSD process exits.
-----------------------------------------------------------------
4.
A shadow object S used by a process P is typically sparsely
populated: S may cover a large range of pages, yet the backing
store for S contains only pages that have been modified by P
*and* chosen for replacement by the pagedaemon.  Probably there
will be few such pages, but there *could* be a great many.

In addition, the existence of long chains of shadow objects in
P's vmspace structure may mean that numerous pager queries will
have to be done to locate the correct copy of an object page
while servicing a page fault.

The facts above demonstrate that the swap pager for 4.4BSD had
to be designed with special attention to the problem of
providing efficient swap-space allocation for sparsely
populated objects and to providing efficient page lookup.

Describe the aspects of the actual design that solved the
problem.
-----------------------------------------------------------------