home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
The Fred Fish Collection 1.5
/
ffcollection-1-5-1992-11.iso
/
ff_disks
/
100-199
/
ff128.lzh
/
Dis
/
dis.doc
< prev
next >
Wrap
Text File
|
1988-01-02
|
8KB
|
183 lines
Dis 68000 disassembler
Greg Lee, February, 1986
The program Dis, its source code, and
this document are in the public domain.
Commands to Dis are:
<CR>, which repeats the last listing command ('l','b','B','d', or
'a') 21 times
<command-letter><CR>, where command-letter is a single character
from the list below
<command-letter><argument><CR>, where the argument follows
immediately after the command-letter
In the list below, arguments in '[...]' are optional, and 'addr' is
either a hexadecimal number or a symbol. One must enter numbers
beginning with digits A-F with a leading 0, so they can be distinguished
from symbols. Note that there is no provision for ignoring spaces before
or after the command-letter.
command argument what it does
l [addr] list instructions
Instructions are disassembled in a format which
would be acceptable to a standard assembler, except
that program counter relative references are given
as hex addresses without a leading '$'. Such
addresses would have to be assigned a symbolic
value before any reassembly.
b [addr] build symbols
This does the same as 'l' above, except that symbols
are made and added to the symbol table for any
operand references which are PC relative, or
absolute long. Symbols are not constructed for
immediate opereands. Symbols are given types
if they appear to be references to data (see the
note below on symbol types). The operand of LEA
is assumed by Dis to refer to instructions,
although this is usually not the case. The target
of JSR or BSR is given a symbol beginning with 'R',
which will cause a preceding blank line when it is
listed as a label.
B [addr] Build locals
This does the same as 'l' above, except that certain
symbols that were probably made with the 'b' command
are replaced with local labels of the form '1$',
'2$',...,'999$'. The labels susceptible to
replacement are those which begin with 'l' and are
7 characters in length. When a label not of this
form is encountered in the listing, a counter is
reset so that the next local generated will be '1$'.
d [addr] dump in hex
a [addr] ascii dump
/ addr address is this
This assigns a new current address for the next
listing command or symbol assignment, but
nothing is listed. Convenient for assigning a
symbol to an IO location which should not be
accessed.
= <symbol> add symbol to table
The string following the '=' is stored in the
symbol table with a value equal to the address of
the last instruction or data that was listed.
The space available for the table is fixed at
1600 symbols containing 11,200 characters. Space
occupied by no longer used symbols is not
automatically reclaimed, but it may be reclaimed
by saving the table to disk, then reading it back.
r [<name>] read file (symtab)
A file of symbols and the addresses to which they
refer is read from disk into Dis's symbol table.
Any symbols already in the table are lost.
The file will commonly have been created by by
the 's' command below, but it could have been
made or modified with an editor. Each line of
the file should either be empty or consist of
a hex number followed by exactly one space, and
then the string of characters that is to make
up the symbol name.
s [<name>] save to file (symtab)
Any symbols currently in the table are saved in
the file for future use. For both 'r' and 's', if
no name is given, 'symtab' is assumed.
q quit
Exit to AmigaDos.
w <num> where is this word?
W <num> where is this longword?
These two commands start searching memory at
the current listing address, but skip memory
from 80000 to FC0000.
f <name> file to disassemble
The requested file is loaded and its first
instruction is listed. The file must be a load
module. If for some reason the file could not be
loaded, the IO error code is shown in hex.
> next code segment
The segments of a load module are scattered in
memory by the AmigaDos loader. This command
alters the current listing address to the beginning
of the code segment after the current one (which
is the first segment, after a successful use of
the 'f' command). If nothing is listed, this means
there is no next segment (or, in 'p' mode, that the
next segment is empty).
< first code segment
After one or more uses of the '>' command, you
might want to go back and look at the first
segment again.
o [num] offset addresses
All address references are adjusted by subtracting
the number before being displayed. If no number
is given, the last listing address is used. This
command is of little use, given the availability
of the below 'p' command.
p offset by segment toggle
After a file is loaded with the 'f' command, this
will adjust all displayed addresses to offsets
from the beginning of the first segment. Thus
the segments will appear to occur in memory in
the same order they occupied the the load file.
In addition, the listing of memory locations
outside any segment of the loaded file is
suppressed, and after 'b', symbols are not
constructed for references outside the loaded
file. 'p' mode is automatic after loading a file.
k<name> keep output in file
Subsequent listing output to the screen will also
be kept in the named file (possibly for subsequent
reassembly). The saving continues until you exit
with the 'q' command.
t trim toggle
Suppresses or enables the display of addresses at
the left-hand side of the screen.
n<num> print n lines after <CR>
Notes on symbols.
Symbols may be up to 255 characters and may contain any
non-control characters. A symbol may not be associated with the
value zero or a value greater than 7FFFFFFFH. Symbols are case sensitive.
There are some special conventions for symbols. A symbol may
carry information about the type of data being disassembled. If a
symbol starts with '.' and one of the letters 'b','a','w','l','c','i', then
code at this and subsequent locations will be treated as bytes, ascii
characters, words, longwords, a constant block, or instructions,
respectively. Otherwise, it will be treated as instructions. Two character
symbols beginning with '.' are not listed, since they are assumed to be
intended to carry only type information.
When a symbol starts with '.c', Dis displays 'dcb.b <length>,<value>'
where the value is that of the byte found at the address of the symbol
and subsequent memory locations, and the length is the count of the number
of bytes in the block. The block is considered to end when:
a byte with a different value is encountered,
an address with an associated symbol is found,
or in 'p' mode the end of a segment is reached.
Dis does not do the counting of bytes in blocks very efficiently when there
are lots of symbols in the table, so after 'dcb.b' has been displayed,
some patience may be required.
It is possible for a symbol to be concealed and not listed as a
label at the left of the screen in some circumstances. This will happen
when its address comes within an instruction being listed, or when its
address is odd and a preceding address was listed as 'dc.w' or 'dc.l'.
However, if a symbol with a following even address would be concealed by a
'dc.l' listing, this is prevented by listing data as bytes rather than
as a longword. In 'p' mode, this is also done one word away from the
end of a segment so as not to show data outside the segment.
Symbols beginning with asterisks may be used to enter comments,
in which case they are not listed as operands, but only as labels.
If a single address is to have both a comment and an ordinary label
associated with it, the comment should be entered after the ordinary
label (else the comment will be lost). Multiple comments for a single
address are listed in the order in which they were entered. Once a
comment has been entered, it can be eliminated only in the way just
mentioned, or by saving the symbol table and editing it.