path: root/man/as86.1

.TH as86 1 "Mar, 1999"
.BY Bruce Evans, Robert de Bath
.nh
.SH NAME
as86 \- Assembler for 8086..80386 processors
.SH SYNOPSIS
.B as86
.RB [ -0123agjuw ]
.RB [ -lm [ list ]]
.RB [ -n\ name ]
.RB [ -o\ obj ]
.RB [ -b [ bin ]]
.RB [ -s\ sym ]
.RB [ -t\ textseg ]
.B src

.B as86_encap\ prog.s\ prog.v
.RB [ prefix_ ]
.RB [ as86_options ]

.SH DESCRIPTION
.B as86
is an assembler for the 8086..80386 processors, it's syntax is closer
to the intel/microsoft form rather than the more normal generic form of
the unix system assembler.

The
.B src
file can be '-' to assemble the standard input.

This assembler can be compiled to support the 6809 cpu and may even work.

.B as86_encap
is a shell script to call as86 and convert the created binary into a C file
.B prog.v
to be included in or linked with programs like boot block installers.
The
.B prefix_
argument is a prefix to be added to all variables defined by the source,
it defaults to the name of the source file. The variables defined include
.B prefix_start
.B prefix_size
and
.B prefix_data
to define and contain the code, plus integers containing the values of all
exported labels.
Either or both the
.B prog.s
and
.B prog.v
arguments can be '-' for standard in/out.

.\" The 'src' file can be '-' for stdin but ONLY on 'big' machines.

.SH OPTIONS

.\"defaults (off or none except for these; no output is produced without a flag):
.\"-03		native
.\"list		stdout (beware of clobbering next arg)
.\"name		basename of the source name

.TP
.B -0
start with 16-bit code segment, warn for all instructions > 8086
.TP
.B -1
start with 16-bit code segment, warn for all instructions > 80186
.TP
.B -2
start with 16-bit code segment, warn for all instructions > 80286
.TP
.B -3
start with 32-bit code segment, don't warn for any instructions. (not even
486 or 586)
.TP
.B -a
enable partial compatibility with Minix asld. This swaps the interpretation
of round brackets and square brackets as well as making alterations to the
code generation and syntax for 16bit jumps and calls. ("jmp @(bx)" is then
a valid instruction)
.TP
.B -g
only put global symbols in object or symbol file
.TP
.B -j
replace all short jumps with similar 16 or 32 bit jumps, the 16 bit
conditional branches are encoded as a short conditional and a long
unconditional branch.
.TP
.B -O 
this causes the assembler to add extra passes to try to use forward
references to reduce the bytes needed for some instructions. 
If the labels move on the last pass the assembler will keep adding passes
until the labels all stabilise (to a maximum of 30 passes)
It's probably not a good idea to use this with hand written assembler 
use the explicit
.B br\ bmi\ bcc
style opcodes for 8086 code or the
.B jmp\ near
style for conditional i386 instructions and make sure all variables are
defined before they are used.
.TP
.B -l
produce list file, filename may follow
.TP
.B -m
print macro expansions in listing
.TP
.B -n
name of module follows (goes in object instead of source name)
.TP
.B -o
produce object file, filename follows
.TP
.B -b
produce a raw binary file, filename may follow.
This is a 'raw' binary file with no header, if there's no
.B -s
option the file starts at location 0.
.TP
.B -s
produce an ASCII symbol file, filename follows.
The format of this table is designed to be easy to parse for encapsulation
and related activities in relation to binary files created with the
.B -b
option.  If a binary file doesn't start at location zero the first two
items in the table are the start and end addresses of the binary file.
.TP
.B -u
assume undefined symbols are imported-with-unspecified segment.
.TP
.B -w-
allow the assembler to print warning messages.
.TP
.B -t n
move all text segment data in segment n+3.
.P
.SH AS86 SOURCE
Special characters
.TP
.B *
Address of the start of the current line.
.TP
.B ; !
Either of these marks the start of a comment. In addition any 'unexpected'
character at the start of a line is assumed to be a comment (but it's also
displayed to the terminal).
.TP
.B $
Prefix for hexadecimal numbers, the 'C' syntax, eg\ 0x1234, is also accepted.
.TP
.B %
Prefix for binary numbers.
.TP
.B #
Prefix for immediate operands.
.TP
.B [ ]
Specifies an indirect operand.
.br
Unlike MASM the assembler has no type information on labels just a
segment and offset. This means that the way this operator and the
immediate prefix work are like traditional assemblers.
.sp
Examples:
.br
	mov     ax,bx
.br
	jmp     bx
.br
Direct register addressing, the jump copies BX into PC.
.sp
.nf
	mov ax,[bx]
	jmp [bx]
.fi
Simple indirect register addressing, the jump moves the contents of the location specified by BX into the PC.
.sp
	mov ax,#1234
.br
Immediate value, ax becomes 1234.
.sp
.nf
	mov ax,1234
	mov ax,_hello
	mov ax,[_hello]
.fi
Absolute addressing, ax is set to contents of location 1234. Note the
third option is not strictly consistant but is in place mainly for asld
compatibility.

.sp
.nf
	mov ax,_table[bx]
	mov ax,_table[bx+si]
	mov eax,_table[ebx*4]

	mov ax,[bx+_table]
	mov ax,[bx+si+_table]
	mov eax,[ebx*4+_table]
.fi
Indexed addressing, both formats are ok, I think the first is more correct
but I tend to used the second. :-)
.br
.P
Conditionals
.TP
.B IF, ELSE, ELSEIF, ENDIF
Numeric condition
.TP
.B IFC, ELSEIFC
String compare (str1,str2)
.TP
.B FAIL .FAIL
Generate user error.
.P
Segment related
.TP
.B .TEXT .ROM .DATA .BSS
Set current segment. These can be preceded by the keyword
.B .SECT
.TP
.B LOC
Set numeric segment 0=TEXT, 3=DATA,ROM,BSS, 14=MAX. 
The segment order set by the linker is now 0,4,5,6,7,8,9,A,B,C,D,E,1,2,3.
Segment 0 and all segments above 3 are assumed to be text segment.
Note the 64k size restrictions are not imposed for segments 3-14.
.P
Label type definition
.TP
.B EXPORT PUBLIC .DEFINE
Export label defined in this object
.TP
.B ENTRY
Force linker to include the specified label in a.out
.TP
.B .GLOBL .GLOBAL
Define label as external and force import even if it isn't used.
.TP
.B EXTRN EXTERN IMPORT .EXTERN
Import list of externally defined labels
.br
NB: It doesn't make sense to use imports for raw binary files.
.TP
.B .ENTER
Mark entry for old binary file (obs)
.P
Data definition
.TP
.B DB .DATA1 .BYTE  FCB
List of 1 byte objects.
.TP
.B DW .DATA2 .SHORT FDB .WORD
List of 2 byte objects.
.TP
.B DD .DATA4 .LONG
List of 4 byte objects.
.TP
.B .ASCII FCC
Ascii string copied to output.
.TP
.B .ASCIZ
Ascii string copied to output with trailing
.B nul
byte.
.P
Space definition
.TP
.B .BLKB RMB .SPACE
Space is counted in bytes.
.TP
.B .BLKW .ZEROW
Space is counted in words. (2 bytes each)
.TP
.B COMM .COMM LCOMM .LCOMM
Common area data definition
.P
Other useful pseudo operations.
.TP
.B .ALIGN .EVEN
Alignment
.TP
.B EQU
Define label
.TP
.B SET
Define re-definable label
.TP
.B ORG .ORG
Set assemble location
.TP
.B BLOCK
Set assemble location and stack old one
.TP
.B ENDB
Return to stacked assemble location
.TP
.B GET INCLUDE
Insert new file (no quotes on name)
.TP
.BR USE16\  [ cpu ]
Define default operand size as 16 bit, argument is cpu type the code is
expected to run on (86, 186, 286, 386, 486, 586) instructions for cpus
later than specified give a warning.
.TP
.BR USE32\  [ cpu ]
Define default operand size as 32 bit, argument is cpu type the code is
expected to run on (86, 186, 286, 386, 486, 586) instructions for cpus
later than specified give a warning. If the cpu is not mentioned the
assembler ensures it is >= 80386.
.TP
.B END
End of compilation for this file.
.TP
.B .WARN
Switch warnings
.TP
.B .LIST
Listings on/off (1,-1)
.TP
.B .MACLIST
Macro listings on/off (1,-1)
.P
Macros, now working, the general form is like this.

    MACRO sax
       mov ax,#?1
    MEND
    sax(1)

.TP
Unimplemented/unused.
.TP
IDENT
Define object identity string.
.TP
SETDP
Set DP value on 6809
.TP
MAP
Set binary symbol table map number.
.TP
Registers
.br
BP BX DI SI
.br
EAX EBP EBX ECX EDI EDX ESI ESP
.br
AX CX DX SP
.br
AH AL BH BL CH CL DH DL
.br
CS DS ES FS GS SS 
.br
CR0 CR2 CR3 DR0 DR1 DR2 DR3 DR6 DR7
.br
TR3 TR4 TR5 TR6 TR7 ST
.TP
Operand type specifiers
BYTE DWORD FWORD FAR PTR PWORD QWORD TBYTE WORD NEAR
.sp
The 'near and 'far' do not allow multi-segment programming, all 'far'
operations are specified explicitly through the use of the instructions:
jmpi, jmpf, callf, retf, etc. The 'Near' operator can be used to force
the use of 80386 16bit conditional branches. The 'Dword' and 'word' 
operators can control the size of operands on far jumps and calls.  
.TP
General instructions.
These are in general the same as the instructions found in any 8086 assembler,
the main exceptions being a few 'Bcc' (BCC, BNE, BGE, etc) instructions
which are shorthands for a short branch plus a long jump and 'BR' which
is the longest unconditional jump (16 or 32 bit).
.TP
Long branches
BCC BCS BEQ BGE BGT BHI BHIS BLE BLO BLOS BLT BMI BNE BPC BPL BPS BVC BVS
BR
.TP
Intersegment
CALLI CALLF JMPI JMPF 
.TP
Segment modifier instructions
ESEG FSEG GSEG SSEG
.TP
Byte operation instructions
ADCB ADDB ANDB CMPB DECB DIVB IDIVB IMULB INB INCB MOVB MULB NEGB NOTB ORB
OUTB RCLB RCRB ROLB RORB SALB SARB SHLB SHRB SBBB SUBB TESTB XCHGB XORB
.TP
Standard instructions
AAA AAD AAM AAS ADC ADD AND ARPL BOUND BSF BSR BSWAP BT BTC BTR BTS CALL
CBW CDQ CLC CLD CLI CLTS CMC CMP CMPS CMPSB CMPSD CMPSW CMPW CMPXCHG
CSEG CWD CWDE DAA DAS DEC DIV DSEG ENTER HLT IDIV IMUL IN INC INS INSB
INSD INSW INT INTO INVD INVLPG INW IRET IRETD J JA JAE JB JBE JC JCXE
JCXZ JE JECXE JECXZ JG JGE JL JLE JMP JNA JNAE JNB JNBE JNC JNE JNG JNGE
JNL JNLE JNO JNP JNS JNZ JO JP JPE JPO JS JZ LAHF LAR LDS LEA LEAVE LES
LFS LGDT LGS LIDT LLDT LMSW LOCK LODB LODS LODSB LODSD LODSW LODW LOOP
LOOPE LOOPNE LOOPNZ LOOPZ LSL LSS LTR MOV MOVS MOVSB MOVSD MOVSW MOVSX
MOVW MOVZX MUL NEG NOP NOT OR OUT OUTS OUTSB OUTSD OUTSW OUTW POP POPA
POPAD POPF POPFD PUSH PUSHA PUSHAD PUSHF PUSHFD RCL RCR RDMSR REP REPE REPNE
REPNZ REPZ RET RETF RETI ROL ROR SAHF SAL SAR SBB SCAB SCAS SCASB SCASD
SCASW SCAW SEG SETA SETAE SETB SETBE SETC SETE SETG SETGE SETL SETLE
SETNA SETNAE SETNB SETNBE SETNC SETNE SETNG SETNGE SETNL SETNLE SETNO
SETNP SETNS SETNZ SETO SETP SETPE SETPO SETS SETZ SGDT SHL SHLD SHR SHRD
SIDT SLDT SMSW STC STD STI STOB STOS STOSB STOSD STOSW STOW STR SUB TEST
VERR VERW WAIT WBINVD WRMSR XADD XCHG XLAT XLATB XOR 
.TP
Floating point
F2XM1 FABS FADD FADDP FBLD FBSTP FCHS FCLEX FCOM FCOMP FCOMPP FCOS
FDECSTP FDISI FDIV FDIVP FDIVR FDIVRP FENI FFREE FIADD FICOM FICOMP
FIDIV FIDIVR FILD FIMUL FINCSTP FINIT FIST FISTP FISUB FISUBR FLD FLD1
FLDL2E FLDL2T FLDCW FLDENV FLDLG2 FLDLN2 FLDPI FLDZ FMUL FMULP FNCLEX
FNDISI FNENI FNINIT FNOP FNSAVE FNSTCW FNSTENV FNSTSW FPATAN FPREM
FPREM1 FPTAN FRNDINT FRSTOR FSAVE FSCALE FSETPM FSIN FSINCOS FSQRT FST
FSTCW FSTENV FSTP FSTSW FSUB FSUBP FSUBR FSUBRP FTST FUCOM FUCOMP
FUCOMPP FWAIT FXAM FXCH FXTRACT FYL2X FYL2XP1
.P
.SH Using GASP

The Gnu assembler preprocessor provides some reasonable implementations
of user biased pseudo opcodes.

It can be invoked in a form similar to:
.TP
.B gasp
.RB [ -a... ]
.B file.s
.RB [ file2.s ]
.B |
.br
.B as86
.RB [ ... ]
.B -
.RB [ -o\ obj ]\ [ -b\ bin ]
.P
Be aware though that Gasp generates an error for 
.B .org
commands, if you're not using alternate syntax you can use 
.B org
instead, otherwise use 
.BR block
and
.BR endb .
The directive
.B export
is translated into
.BR .global ,
which forces an import, if you are making a file using
.B -b
use
.B public
or 
.B .define
instead.

The GASP list options have no support in as86.
.SH SEE ALSO
as(1), ld86(1), bcc(1)
.SH BUGS
The 6809 version does not support -0, -3, -a or -j.

If this assembler is compiled with BCC this is classed as a 'small'
compiler, so there is a maximum input line length of 256 characters
and the instruction to cpu checking is not included.

The checking for instructions that work on specific cpus is probably
not complete, the distinction between 80186 and 80286 is especially
problematic.

The
.B .text
and
.B .data
pseudo operators are not useful for raw binary files.

When using the
.B org
directive the assembler can generate object files that may break ld86(1).