Lecture 07

Relocation tool: cs241.MERL

mips.twoints and mips.array - optional second argument = address at which to load the mips file.


myobj.MERL to be loaded at 0x10000

java cs241.MERL 0x10000 < myobj.MERL > myobj.mips
java mips.twoints myobj.mips 0x10000

Loader Relocation Algorithm

read()                                  //skip cookie
endMod = read()                         //end of MERL file
codeLen = read()                        //length of code only
a = FindFreeRAM(codeLen + stack space)
for(i = 0; i < codeLen; i += 4)
    MEM[a + i] = read();
i = codeLen + 12                        //start of relocation table
while(i < endMod)
    format = read()
    if(format == 1)
        rel = read()                    //address to be relocated, relative to start of header (not code)
        Mem[a + rel - 12] += a - 12     //actual location (header not loaded), adjust forward by a, backwards by header length since header not loaded
    else ERROR
    i += 8


Convenient to split large MIPS programs into smaller units.

How can the assembler resolve a reference to a label if it is in a different file?

Solution 1: cat all asm files together and then assemble

cat a.asm b.asm c.asm | java cs241.binasm > result.mips

This will work, but why should we have to assemble these files more than once (i.e. every time we cat them into a larger project).

Cannot we assemble first and then cat?


 ______          ______
|header|        |header|
|body  |        |body  |
|footer|\       |footer|
|______| \      |header|
 ______   cat-> |body  |
|header| /      |footer|
|body  |/       |______|

Solution 2: Need a tool that understands MERL files and puts them together intelligently -- a linker

But still

Where label id: not found, it fills in 0 temporarily, and it must indicate that the program requires the value of id before it can run.


lis $3
.word a


a: ...

one.asm cannot be executed without being linked to code that contains babel a

How does the assembler notify us?

Make an entry in the MERL file, but

lis $3
.word abc
abd: ...

How can the assembler know what is an error and what is intentional?

New assembler directive:

.import id

So when the assembler encounters .word abc, iF label abc does not occur in current file and no .import abc directive, then error

Notifying the system: make an entry in MERL table

Formal code 0x11 means External Symbol Reference (ESR)

What info is needed?

Format of an ESR entry:

word 1                  - 0x11
word 2                  - location where the symbol is used
word 3                  - length of the name in chars (n)
word 4 ... word (3 + n) - ASCII chars in the symbol name (each char in its own word)

The other side:

a.asm (external reference to abc)

lis $3
.word abc

b.asm (abc is the beginning of a procedure)

abc: sw $3, -4($30)
jr $31

c.asm (just an internal label)

abc: add $1, $2, $3
beq $2, $0, abc

How can the linker know which abc to link to?

Cannot assume labels will not be duplicated.

How can we make abc in b.asm accessible, and abc in c.asm inaccessible?

Solution: another assembler directive and MERL entry.

Directive: .export abc * make abc available for linking * does NOT assemble to a wor of MIPS * tells the assembler to make an entry in the MERL symbol table

MERL entry: External Symbol Definition (ESD)


word 1              - 0x05 format code for ESD
word 2              - address the symbol represents
word 3              - length of name (n)
word 4 ~ word (3+n) - Then name in ASCII one word per char