Assignment 3, due Feb 15

Always, on every assignment, please write your name legibly as it appears on your University ID and on the class list! All assignments will be due at the start of class on the day indicated (usually a Friday). The only exceptions to this rule will be by advance arrangement unless there is what insurance companies call "an act of God" - something outside your control. Homework must be turned in on paper and in class! Late work may be turned in to the teaching assistant's mailbox, but see the late work policy. Never push late work under someone's door!

1. Background: Position independent machine code is code that can be moved from one range of memory addresses to another range of memory addresses without alteration. If all branches, jumps and calls within a block of code use program-counter relative addressing, and if references to constants embedded in the code use program-counter relative addressing, the control structure of the code is position independent. If, on the other hand, absolute memory addresses are used, these addresses will need modification if the code is relocated.

   a) If you know that the block of W words of code is currently loaded starting in address S (source) and it is to be moved to address D (destination), and you know that addresses S+X, S+Y and S+Z in the original code contain absolute memory addresses, what changes to the data must you make after copying the block of code from S to D. For simplicity, assume word addressing. (0.5 points)

   b) Suppose A and B are variables in C. Normally, to make A point to B, you would use A = &B. Having done this, we would use *A to reference B. This assumes that the type of A is *BT, where BT is the type of B. Give C code to assign a relative pointer to B to A and to follow that pointer. Get the computation right first, then worry about casting the expression. Note, the pointer in question is relative, but it is not PC relative. Rather, it should be relative to the address of A. (0.5 points)

   c) Suppose you wanted the linker to generate PC relative addresses of external symbols when it linked two object files. Suggest a modification to Figure 7.17 that would allow this. For the sake of simplicity, assume 8 and 16-bit values. No need for 32-bit addresses here. (0.5 points)

2. Background: The Unix file system supports the notion of a polymorphic file type. There are files and directories, both of which are conceptually disk files. There are also sequential input files such as the keyboard, and sequential output files such as the (classical output-only) parallel ports. In addition, there are input-output devices such as the serial ports. Ignore other file types for the purpose of this assignment.

   The basic operations on a Unix file are lseek(), read(), and write(). See the man pages for these in section 2 of the programmer's reference manual (available online with the the man 2 lseek command, for example).

   A Question: Informally describe these using the concept of a class hierarchy, identifying all of the subclasses and noting which operations are applicable to each class in this hierarchy. Also note any problems posed by this hierarchy for the notion of class hierarchy as it is present in whatever object oriented programming languages you happen to be familiar with. (0.5 points)

3. Background: One popular Unix variant contains the following text in <stdio.h>:

   #define getc(p) (--(p)->_r < 0 ? _srget(p) : (int)(*(p)->_p++))
   #define getchar() getc(stdin)
   

   a) This is extraordinarily ugly C code! Using the terminology from the notes for lecture 10, explain what it is doing -- in one or two brief English sentences, and then explain why. Do not explain how it is doing it, that is a matter for part b. (0.5 points)

   b) (0.5 points) Parse the definition of getc(), breaking it down by sub-expressions and briefly explaining each of them.