Ham Radio 2000

home *** CD-ROM | disk | FTP | other *** search

/ Ham Radio 2000 / Ham Radio 2000.iso / ham2000 / misc / tasm / tasm.doc < prev next >

Wrap

Text File | 1996-09-05 | 71KB | 1,697 lines

TASM USER'S MANUAL TASM - A Table Driven Cross Assembler for the MSDOS* Environment Thomas N. Anderson Speech Technology Incorporated 837 Front Street South, Issaquah, WA 98027 August, 1990 Version 2.8 [Speech Technology Incorporated manufactures electronic devices to aid the visually impaired employing digital speech synthesis technology.] TASM software is Copyright (C) 1985-1990 by Speech Technology Incorporated. All rights reserved. This document is Copyright (C) 1985-1990 by Speech Technology Incorporated. All rights reserved. Permission is granted to copy this document and related software except for the source code. The source code, distributed to registered users, may be copied for the sole use of the registered user. Portions of TASM.EXE (C runtime library) are Copyright 1987 by Microsoft Corporation. * MSDOS is a trademark of Microsoft Corporation. TASM - Table Driven Assembler Version 2.8 Page 2 TABLE OF CONTENTS SECTION PAGE ______________________________________________________________________ INTRODUCTION 4 INVOCATION 5 File Names 5 Option: a - Assembly Control 6 Option: b - Binary Object Format 7 Option: c - Contiguous Block Output 7 Option: d - Define a Macro 7 Option: e - Expand Source 7 Option: f - Fill Memory 7 Option: g - Select Object File Format 8 Option: h - Hex Object Code Table 8 Option: i - Ignore Case in Labels 8 Option: l - Label Table 8 Option: m - MOS Technology Object Format 8 Option: o - Set Number of Bytes per Object Record 8 Option: p - Page Listing File 8 Option: q - Disable Listing File 8 Option: r - Set Read Buffer Size 8 Option: s - Enable Symbol File Generation 9 Option: t - Table Name 9 Option: x - Enable Extended Instruction Set 10 Option: y - Enable Assembly Timing 10 ENVIRONMENT VARIABLES 11 TASMTABS 11 TASMOPTS 10 EXIT CODES 10 SOURCE FILE FORMAT 12 EXPRESSIONS 13 Labels 13 Numeric Constants 13 Character Constants 14 String Constants 14 Location Counter Symbol 14 Operators 14 ASSEMBLER DIRECTIVES 17 ADDINSTR 17 AVSYM 17 BLOCK 17 BSEG/CSEG/DSEG/NSEG/XSEG 17 BYTE 18 CODES/NOCODES 18 DB 18 DW 18 DEFINE 18 DEFCONT 19 EJECT 19 ELSE 20 END 20 ENDIF 20 EQU 20 EXPORT 21 IFDEF 21 TASM - Table Driven Assembler Version 2.8 Page 3 IFNDEF 21 IF 22 INCLUDE 22 LIST/NOLIST 22 LSFIRST/MSFIRST 22 ORG 22 PAGE/NOPAGE 23 SET 23 SYM/AVSYM 23 TEXT 24 TITLE 25 WORD 25 OBJECT FILE FORMATS 26 Intel Hex Object Format 26 MOS Technology Hex Object Format 27 Motorola Hex Object Format 27 Binary Object Format 28 LISTING FILE FORMAT 29 PROM PROGRAMMING 30 ERROR MESSAGES 31 BUGS AND LIMITATIONS 34 APPENDIX A - ORDERING INFORMATION 35 TASM - Table Driven Assembler Version 2.8 Page 4 INTRODUCTION TASM is a table driven cross assembler for the MS-DOS environment. Assembly source code, written in the appropriate dialect (generally very close to the manufacturers assembly language), can be assembled with TASM, and the resulting object code transferred to the target microprocessor system via PROM or other mechanisms. The current microprocessor families supported by TASM are: 6502 8048 6800/6801 8051 6805 8080/8085 TMS32010 Z80 TMS7000 The user so inclined may build tables for other microprocessors. The descriptions of the various existing tables and instructions on building new tables are not in this document but can be found in the TASMTABS.DOC file on the TASM distribution disk. TASM characteristics include: 1. Powerful expression parsing (17 operators). 2. Supports a subset of the 'C' preprocessor commands. 3. Macro capability (through use of DEFINE directive). 4. Multiple statements per line. 5. Four object file formats: Intel hex, MOS Technology hex, Motorola hex, binary. 6. Absolute code generation only. 7. Source code available (in C). 8. Uniform syntax across versions for different target machines. 9. Features in support of PROM programming (preset memory, contiguous block). 10. Supports extended instructions for many of the supported microprocessor families. 11. Tables read at run time - single TASM executable for all table versions. 12. Symbol table export for inclusion in subsequent assemblies (.EXP format). 13. Symbol table export file compatible with the Avocet 8051 simulator (.SYM format). SHAREWARE TASM is distributed as shareware. TASM is not public domain. The TASM distribution files may be freely copied (excluding the source code files) and freely used for the purpose of evaluating the suitability of TASM for a given purpose. Use of TASM beyond a reasonable evaluation period requires registration. Prolonged use without registration is unethical. TASM - Table Driven Assembler Version 2.8 Page 5 INVOCATION TASM can be invoked as follows (optional fields shown in brackets, symbolic fields in <>): tasm -<pn> [-<option_flag ...>] <src_file> [<obj_file> [<lst_file> [<exp_file> [<sym_file>]]]] Where <option_flag> can be one or more of the following: -<table> Specify version (<table> = table designation) -t<table> Table (alternate form of above) -a<amask> Assembly control (optional error checking) -b Produce object in binary (.COM) format -c Object file written as a contiguous block -d<macro> Define a macro (or just a macro label) -e Show source lines after macro expansion -f<fillbyte> Fill entire memory space with <fillbyte> (hex) -g<objtype> Object file (0=Intel, 1=MOS Tech, 2=Motorola, 3=binary) -h Produce hex table of the assembled code (in list file) -i Ignore case for labels -l Produce a label table in the listing -m Produce object in MOS Technology format -o<obytes> Bytes per object record (for hex obj formats) -p[<lines>] Page the listing file (<lines per page>. default=60) -q Quiet, disable the listing file -r<kb> Set read buffer size in Kbytes (default 2 Kbytes) -s Write a symbol table file -x[<xmask>] Enable extended instruction set (if any) -y Time the assembly The file parameters are defined as follows: <src_file> Source file name <obj_file> Object code file name <lst_file> Listing file name <exp_file> Symbol export file (only if the EXPORT directive is used). <sym_file> Symbol table file (only if the "-s" option or the SYM/AVSYM directives are used). The source file must be specified. If not, some usage information is displayed. Default file names for all the other files are generated if they are not explicitly provided. The filename is formed by taking the source filename and changing the extension to one of the following: Extension File type ________________________________________________________ .OBJ Object file .LST Listing file .EXP Symbol export file .SYM Symbol table file TASM has no built-in instruction set tables. Instruction set definition files are read at run time. TASM determines which table to use based on the '-table' field shown above. For example, to TASM - Table Driven Assembler Version 2.8 Page 6 assemble the code in a file called source.asm, one would enter: tasm -48 source.asm for an 8048 assembly tasm -65 source.asm for a 6502 assembly tasm -51 source.asm for an 8051 assembly. tasm -85 source.asm for an 8085 assembly. tasm -80 source.asm for a Z80 assembly. tasm -05 source.asm for a 6805 assembly. tasm -68 source.asm for a 6800/6801 assembly. tasm -70 source.asm for a TMS7000 assembly. tasm -32 source.asm for a TMS32010 assembly. The file name that the tables are read from is formed by taking the digits specified after the '-' and appending it to 'TASM' then appending the '.TAB' extension. Thus, the '-48' flag would cause the tables to be read from the file 'TASM48.TAB'. It is possible to designate tables by non numeric part numbers if the '-t' flag is used. For example, if a user built a table called TASMF8.TAB then TASM could be invoked as follows: tasm -tf8 source.asm Each option flag must be preceded by a dash. Options need not precede the file names, however. The various options are described below: a - Assembly Control. TASM can provide additional error checking by specifying the '-a' option at the time of execution. If the '-a' is provided without a digit following, then all the available error checking is done. If a digit follows, then it is used as a mask to determine the error checks to be made. The bits of the mask are defined as follows: Bit Option Description __________________________________________________________________ 0 -a1 Check for apparent illegal use of indirection 1 -a2 Check for unused data in the arguments 2 -a4 Check for duplicate labels 3 -a8 Check for non-unary operators at start of expression. 4 -a10 Check for off-page jumps for 8051 ACALL/AJMP. Combinations of the above bits can also be used. For example, '-a5' would enable the checking for illegal indirection and duplicate labels. Illegal indirection applies to micros that use parenthesis around an argument to indicate indirection. Since it is always legal to put an extra pair of parenthesis around any expression (as far as the expression parser is concerned), the user may think that he/she is indicating indirection for an instruction that has no indirection and TASM would not complain. Enabling this checking will result in an error message (warning) whenever an outer pair of parenthesis is used and the instruction set definition table does not explicitly indicate that to be a valid form of addressing. TASM - Table Driven Assembler Version 2.8 Page 7 Unused data in arguments applies to cases where a single byte of data is needed from an argument, but the argument contains more than one byte of data. If a full sixteen bit address is used in a 'Load Immediate' type instruction that needs only a single byte, for example, an error message would be generated. Here is an example (6502 code): 0001 1234 .org $1234 tasm: Unused data in MS byte of argument. Line 0002 in test.asm 0002 1234 A9 34 start lda #start To make the above checks occur whenever you do an assembly, add a line like this to your AUTOEXEC.BAT file: SET TASMOPTS=-a b - Binary Object Format. This option causes the object file to be written in binary - one byte for each byte of code/data. Note that no address information is included in the object file in this format. The contiguous block (-c) output mode is forced when this option is invoked. This flag is equivalent to '-g3'. c - Contiguous Block Output. If this option is specified, then all bytes in the range from the lowest used byte to the highest will be defined in the object file. Normally, with the default Intel Hex object format enabled, if the Program Counter (PC) jumps forward because of an .ORG directive, the bytes skipped over will not have any value assigned them in the object file. With this option enabled, no output to the object file occurs until the end of the assembly at which time the whole block is written. This is useful when using TASM to generate code that will be put into a PROM so that all locations will have a known value. This option is often used in conjunction with the -f option to ensure all unused bytes will have a known value. d - Define a Macro. Macros are defined on the command line generally to control the assembly of various IFDEF's that are in the source file. This is a convenient way to generate various versions of object code from a single source file. e - Expand Source. Normally TASM shows lines in the listing file just as they are in the source file. If macros are in use (via the DEFINE directive) it is sometimes desirable to see the source lines after expansion. Use the '-e' flag to accomplish this. f - Fill Memory. This option causes the memory image that TASM maintains to be initialized to the value specified by the two hex characters immediately following the 'f'. TASM maintains a memory image that is a full 64K bytes in size (even if the target processor cannot utilize that memory space). Invocation of this option introduces a delay at start up of up to 2 seconds (time required to initialize all 64K bytes). g - Object File Format. TASM can generate object code in four different formats as indicated below: TASM - Table Driven Assembler Version 2.8 Page 8 -g0 for Intel hex (default) -g1 for MOS Technology hex (same as -m) -g2 for Motorola hex -g3 for binary (same as -b) The '-m' and '-b' flags may also be used, as indicated above. See the section on OBJECT FILE FORMATS for descriptions of each of the above. h - Hex Object Code Table. This option causes a hex table of the produced object code to appear in the listing file. Each line of the table shows sixteen bytes of code. i - Ignore Case in Labels. TASM is normally case sensitive when dealing with labels. For those that prefer case insensitivity, the '-i' command line option can be employed. l - Label Table. This option causes a label table to appear in the listing file. Each label is shown with its corresponding value. Macro labels (as established via the DEFINE directives) do not appear. m - MOS Technology Object Format. This option causes the object file to be written in MOS Technology hex format rather than the default Intel hex format. See section on OBJECT FILE FORMATS for a description of the format. o - Set Number of Bytes per Object Record. When generating object code in either the MOS Technology format or the Intel hex format, a default of 24 (decimal) bytes of object are defined on each record. This can be altered by invoking the '-o' option immediately followed by two hex digits defining the number of bytes per record desired. For example, if 32 bytes per record are desired, one might invoke TASM as: TASM -48 -o20 source.asm p - Page Listing File. This option causes the listing file to have top of page headers and form feeds inserted at appropriate intervals (every sixty lines of output). To override the default of sixty lines per page, indicate the desired number of lines per page as a decimal number immediatly following the '-p'. Here is an example: TASM -48 -p56 source.asm q - Disable Listing File. This option causes all output to the listing file to be suppressed, unless a .LIST directive is encountered in the source file (see LIST/NOLIST directives). r - Set Read Buffer Size. This option overrides the default read buffer size of 2 Kbytes. The first hexadecimal digit immediately after the 'r' is taken as the number of K bytes to allocate for the read buffer (.e.g. '-r8' indicates an 8K byte buffer, '-rf' TASM - Table Driven Assembler Version 2.8 Page 9 indicates a 15K byte buffer). Note that that read buffers are taken from the same memory pool as labels and macro storage, and that additional read buffers are needed if "includes" are used. Thus, using 8K byte buffers may be suitable for most assemblies, but programs with large numbers of symbols may not allow such a value. Also, reducing the buffer size to 1 Kbyte can increase the memory pool available for label storage, if such is needed. s - Enable Symbol File Generation. If this flag is set, a symbol file is generated at the end of the assembly. The format of the file is one line per label, each label starts in the first column and is followed by white space and then four hexadecimal digits representing the value of the label. The following illustrates the format: label1 FFFE label2 FFFF label3 1000 The symbol file name can be provided as the fifth file name on the the command line, or the name will be generated from the source file name with a '.SYM' extension. The symbol table file can also be generated by invoking the SYM directive. The AVSYM directive also generates the symbol file but in a different format (see section on ASSEMBLER DIRECTIVES). t - Table Name. As an alternative to specifying the instruction set table as two decimal digits, the table indication may be preceeded by the '-t' option. This is useful if the desired table name starts with a non-numeric. Thus, a table for an F8 might be selected as: TASM -tf8 source.asm TASM would expect to read the instruction set definition tables from a file named TASMF8.TAB. TASM - Table Driven Assembler Version 2.8 Page 10 x - Enable Extended Instruction Set. If a processor family has instructions that are valid for only certain members, this option can be used to enable those beyond the basic standard instruction set. A hex digit may follow the 'x' to indicate a mask value used in selecting the appropriate instruction set. Bit 0 of the mask selects the basic instruction set, thus a '-x1' would have no effect. A '-x3' would enable the basic set plus whatever instructions have bit 1 set in their class mask. A '-x' without a digit following is equivalent to a '-xf' which sets all four of the mask bits. The following table indicates the current extended instruction sets available in the TASM tables: Base Extension 1 Extension 2 Extension 3 Base Table Family (-x3) (-x5) (-x9) ____________________________________________________________________ TASM48.TAB 8048 8041A 8022 8021 TASM65.TAB 6502 R65C02 R65C00/21 TASM05.TAB 6805 M146805 CMOS HC05C4 TASM80.TAB Z80 HD64180 TASM68.TAB 6800 6801/6803 TASM51.TAB 8051 TASM85.TAB 8080 TASM32.TAB TMS32010 TASM70.TAB TMS7000 The above table does not attempt to show the many microprocessor family members that may apply under a given column. y - Enable Assembly Timing. If this option is enabled TASM will generate a statement of elapsed time and assembled lines per second at the end of the assembly. TASM - Table Driven Assembler Version 2.8 Page 11 ENVIRONMENT VARIABLES The TASM environment can be customized by using the enviroment variables listed below: TASMTABS. This variable specifies the path to be searched for TASM instruction set definition tables. If it is not defined then the table(s) must exist in the current working directory. If it was desired to put the instruction set definition tables in a directory called 'TASM' on a drive called 'C:', the following statement would be appropriate in the AUTOEXEC.BAT file: set TASMTABS=C:\TASM TASMOPTS. This variable specifies TASM command line options that are to be invoked every time TASM is executed. For example, if TASM is being used for 8048 assemblies with binary object file output desired, the following statement would be appropriate in the AUTOEXEC.BAT file: set TASMOPTS=-48 -b EXIT CODES When TASM terminates, it will return to DOS the following exit codes: Exit Code Meaning ___________________________________________________________ 0 Normal completion, no assembly errors 1 Normal completion, with assembly errors 2 Abnormal completion, insufficient memory 3 Abnormal completion, file access error 4 Abnormal completion, general error Exit codes 2 and above will also be accompanied by messages to the console concerning the error. TASM - Table Driven Assembler Version 2.8 Page 12 SOURCE FILE FORMAT Statements in the source file must conform to a format as follows (except for assembler directive statements which are described in a subsequent section): label operation operand comment All of the fields are optional, under appropriate circumstances. An arbitrary amount of white space (space and tabs) can separate each field (as long as the maximum line length of 255 characters is not exceeded). Each of the fields are described below: Label Field. If the first character of the line is alphabetic, it is assumed to be the start of a label. Subsequent characters are accepted as part of that label until a space, tab, or ':' is encountered. The assembler assigns a value to the label corresponding to the current location counter. Labels can be a maximum of 13 characters long. Labels can contain upper and lower case letters, digits, underscores, and periods (the first character must be alphabetic). Labels are case sensitive - the label 'START' is a different label from 'start' - unless the '-i' (ignore case) option is enabled. Operation Field. The operation field contains an instruction mnemonic which specifies the action to be carried out by the target processor when this instruction is executed. The interpretation of each mnemonic is dependent on the target microprocessor (as indicated by the selected TASM table). The operation field may begin in any column except the first. The operation field is case insensitive. Operand Field. The operand field specifies the data to be operated on by the instruction. It may include expressions and/or special symbols describing the addressing mode to be used. The actual format and interpretation is dependent on the target processor. For a description of the format for currently supported processors, see the TASMTABS.DOC file on the TASM distribution disk. Comment Field. The comment field always begins with a semicolon. The rest of the line from the semicolon to the end of the line is ignored by TASM, but passed on to the listing file for annotation purposes. The comment field must be the last field on a line, but it may be the only field, starting in column one, if desired. Multiple Statement Lines. If the backslash character is encountered on a source line, it is treated as a newline. The remainder of the line following the backslash will be processed as an independent line of source code. This allows one to put multiple statements on a line. This facility is not so useful of itself, but when coupled with the capability of the DEFINE directive, powerful multiple statement macros can be constructed (see section on ASSEMBLER DIRECTIVES). Note that when using the statement separator, the character immediately following it should be considered the first character of a new line, and thus must either be a start of a label or white space TASM - Table Driven Assembler Version 2.8 Page 13 (not an instruction). As the examples show, a space is put between the backslash and the start of the next instruction. Some examples of valid source statements follow (6502 mnemonics shown): label1 lda byte1 ;get the first byte dec byte1 jne label1 ; label2 sta byte2,X ; a multiple statement line follows lda byte1\ sta byte1+4\ lda byte2\ sta byte2+4 EXPRESSIONS Expressions are made up of various syntactic elements (tokens) combined according to a set of syntactical rules. The tokens are summarized as follows: 1. Labels 2. Constants 3. Location Counter Symbol 4. Operators 5. Parenthesis Labels. Labels are strings of characters that have a numeric value associated with them, generally representing an address. Labels can contain upper and lower case letters, digits, underscores, and periods. The first character must be a letter (to distinguish it from a numeric constant). The value of a label is limited to 32 bit precision. Labels can contain up to 13 characters, all of which are significant (none are ignored when looking at a label's value, as in some assemblers). Case is significant unless the '-i' command line option is invoked. Numeric Constants. Numeric constants must always begin with a decimal digit (thus hexadecimal constants that start with a letter must be prefixed by a '0' unless the '$' prefix is used). The radix is determined by a letter immediately following the digit string according to the following table: Radix Suffix Prefix ______________________________________________________ 2 B or b % 8 O or o @ 10 D or d (or nothing) 16 H or h $ Decimal is the default radix, so decimal constants need no suffix or prefix. TASM - Table Driven Assembler Version 2.8 Page 14 The following representations are equivalent: 1234H or $1234 100d or 100 177400O or @177400 01011000b or %01011000 The prefixes are provided for compatibility with some other source code formats but introduce a problem of ambiguity. Both '%' and '$' have alternate uses ('%' for modulo, '$' for location counter symbol). The ambiguity is resolved by examining the context. The '%' character is interpreted as the modulo operator only if it is in a postion suitable for a binary operator. Similarly, if the first character following a '$' is a valid hexadecimal digit, it is assumed to be a radix specifier and not the location counter. Character Constants. Character constants are single characters surrounded by single quotes (following quote is optional). The ASCII value of the character in the quotes is returned. No escape provision exists to represent non-printable characters within the quotes, but this is not necessary since these can be just as easily represented as numeric constants (or using the TEXT directive which does allow escapes). String Constants. String constants are one or more characters surrounded by double quotes. Note that string constants are not allowed in expressions. They are only allowable following the TITLE, BYTE and TEXT assembler directives. The quoted strings may also contain escape sequences to put in unprintable values. The following escape sequences are supported: Escape Sequence Description ___________________________________________________________ \n Line Feed \r Carriage return \b Backspace \t Tab \f Formfeed \\ Backslash \" Quote \000 Octal value of character Location Counter Symbol. The current value of the location counter (PC) can be used in expressions by placing a '$' in the desired place. The Location Counter Symbol is allowable anywhere a numeric constant is. (Note that if the '$' is followed by a decimal digit then it is taken to be the hexadecimal radix indicator instead of the Location Counter symbol, as mentioned above). The '*' may also be used to represent the location counter, but is less preferred because of its ambiguity with the multiplicative operator. Operators. Expressions can optionally contain operators to perform some alterations or calculations on particular values. The operators are summarized as follows: TASM - Table Driven Assembler Version 2.8 Page 15 Operator Type Description __________________________________________ + Additive addition - subtraction * Multiplicative multiplication / division % modulo << logical shift left >> logical shift right ~ Unary bit inversion (one's complement) - unary negation = Relational equal == equal != not equal < less than > greater than <= less than or equal >= greater than or equal & Binary binary 'and' | binary 'or' ^ binary 'exclusive or' The syntax is much the same as in 'C' with the following notes: 1. No operator precedence is in effect. Evaluation is from left to right unless grouped by parenthesis (see example below). 2. All evaluations are done with 32 bit signed precision. 3. Both '=' and '==' are allowable equality checkers. This is allowed since the syntax does not provide assignment capability (as '=' would normally imply). The relational operators return a value of 1 if the relation is true and 0 if it is false. Thirtytwo bit signed arithmetic is used. It is always a good idea to explicitly indicate the desired order of evaluation with parenthesis, especially to maintain portability since TASM does not evaluate expressions in the same manner as many other assemblers. To understand how it does arrive at the values for expressions, consider the following example: 1 + 2*3 + 4 TASM would evaluate this as: (((1 + 2) * 3) + 4) = 13 Typical rules of precedence would cause the (2*3) to be evaluated first, such as: 1 + (2*3) + 4 = 11 TASM - Table Driven Assembler Version 2.8 Page 16 To make sure you get the desired order of evaluation, use parenthesis liberally. Note that TASM versions earlier than 2.7.8 employed a somewhat different method of evaluating expressions (also without precedence) that essentially resulted in a right to left evaluation. Here are some examples of valid expressions: (0f800H + tab) (label_2 >> 8) (label_3 << 8) & $f000 $ + 4 010010000100100b + 'a' (base + ((label_4 >> 5) & (mask << 2)) TASM - Table Driven Assembler Version 2.8 Page 17 ASSEMBLER DIRECTIVES Most of the assembler directives have a format similar to the machine instruction format. However, instead of specifying operations for the processor to carry out, the directives cause the assembler to perform some function related to the assembly process. TASM has two types of assembler directives - those that mimic the 'C' preprocessor functions, and those that resemble the more traditional assembler directive functions. Each of these will be discussed. The 'C' preprocessor style directives are invoked with a '#' as the first character of the line followed by the appropriate directive (just as in 'C'). Thus, these directives cannot have a label preceding them (on the same line). Note that in the examples directives are shown in upper case, however, either upper or lower case is acceptable. ADDINSTR. This directive can be used to define additional instructions for TASM to use in this assembly. The format is: [<label>] .ADDINSTR inst args opcode nbytes modop class shift binor The fields are separated by white space just as they would appear in an instruction definition file. See the TASMTABS.DOC file on the TASM distribution disk for more detail. AVSYM. See SYM/AVSYM. BLOCK. This directive causes the Instruction Pointer to advance the specified number of bytes without assigning values to the skipped over locations. The format is: [label] .BLOCK expr Some valid examples are: word1 .BLOCK 2 byte1 .block 1 buffer .block 80 BSEG/CSEG/DSEG/NSEG/XSEG. These directives can be invoked to indicate the appropriate address space for symbols and labels defined in the subsequent code. The invocation of these directives in no way affects the code generated, only provides more information in the symbol table file if the AVSYM directive is employed. Segment control directives such as these are generally supported by assemblers that generate relocatable object code. TASM does not generate relocatable object code and does not support a link phase, so these directives have no direct effect on the resulting object code. The segments are defined as follows: TASM - Table Driven Assembler Version 2.8 Page 18 Directive Segment Description _________________________________________________________ BSEG Bit address CSEG Code address DSEG Data address (internal RAM) NSEG Number or constant (EQU) XSEG External data address (external RAM) BYTE. This directive allows a value assignment to the byte pointed to by the current Instruction Pointer. The format is: [label] .BYTE expr [, expr ...] Only the lower eight bits of expr are used. Multiple bytes may be assigned by separating them with commas or (for printable strings) enclosed in double quotes. Here are some examples: label1 .BYTE 10010110B .byte 'a' .byte 0 .byte 100010110b,'a',0 .byte "Hello", 10, 13, "World" CODES/NOCODES. These directives can be used to alternately turn on or off the generation of formatted listing output with line numbers, opcodes, data, etc. With NOCODES in effect, the source lines are sent to the listing file untouched. This is useful around blocks of comments that need a full 80 columns of width for clarity. DB. This is alternate form of the BYTE directive. DW. This is alternate form of the WORD directive. DEFINE. The DEFINE directive is one of the most powerful of the directives and allows string substitution with optional arguments (macros). The format is as follows: #DEFINE <macro_label>[(<arg_list>)] [<macro_def>] <macro_label> := character string to be expanded when found in the source file. <arg_list> := optional argument list for variable substitution in macro expansion. <macro_def> := character string to replace the occurrences of macro_label in the source file. The simplest form of the DEFINE directive might look like this: #DEFINE MLABEL Notice that no substitutionary string is specified. The purpose of TASM - Table Driven Assembler Version 2.8 Page 19 a statement like this would typically be to define a label for the purpose of controlling some subsequent conditional assembly (IFDEF or IFNDEF). A more complicated example, performing simple substitution, might look like this: #DEFINE VAR1_LO (VAR1 & 255) This statement would cause all occurrences of the string 'VAR1_LO' in the source to be substituted with '(VAR1 & 255)'. As a more complicated example, using the argument expansion capability, consider this: #DEFINE ADD(xx,yy) clc\ lda xx\ adc yy\ sta xx If the source file then contained a line like this: ADD(VARX,VARY) It would be expanded to: clc\ lda VARX\ adc VARY\ sta VARX The above example shows the use of the backslash ('\') character as a multiple instruction statement delimiter. This approach allows the definition of fairly powerful, multiple statement macros. The example shown generates 6502 instructions to add one memory location to another. Some rules associated with the argument list: 1. Use a maximum of 10 arguments. 2. Each argument should be a maximum of 15 characters. Note that macros can be defined on the TASM command line, also, with the '-d' option flag. DEFCONT. This directive can be used to add to the last macro started with a DEFINE directive. This provides a convenient way to define long macros without running off the edge of the page. The ADD macro shown above could be defined as follows: #DEFINE ADD(xx,yy) clc #DEFCONT \ lda xx #DEFCONT \ adc yy #DEFCONT \ sta xx EJECT. This directive can be used to force a top of form and the generation of a page header on the list file. It has no effect if the paging mode is off (see PAGE/NOPAGE). The format is: .EJECT TASM - Table Driven Assembler Version 2.8 Page 20 ELSE. This directive can optionally be used with IFDEF, IFNDEF and IF to delineate an alternate block of code to be assembled if the block immediately following the IFDEF, IFNDEF or IF is not assembled. Here are some examples of the use of IFDEF, IFNDEF, IF, ELSE, and ENDIF: #IFDEF label1 lda byte1 sta byte2 #ENDIF #ifdef label1 lda byte1 #else lda byte2 #endif #ifndef label1 lda byte2 #else lda byte1 #endif #if ($ >= 1000h) ; generate an invalid statement to cause an error ; when we go over the 4K boundary. !!! PROM bounds exceeded. #endif END. This directive should follow all code/data generating statements in the source file. It forces the last record to be written to the object file. The format is: [<label>] .END ENDIF. This directive must always follow an IFDEF, IFNDEF, or IF directive and signifies the end of the conditional block. EQU. This directive can be used to assign values to labels. The labels can then be used in expressions in place of the literal constant. The format is: <label> .EQU <expr> Here is an example: MASK .EQU 0F0H ; lda IN_BYTE and MASK sta OUT_BYTE An alternate form of 'EQU' is '='. The previous example is equivalent to: TASM - Table Driven Assembler Version 2.8 Page 21 MASK = 0F0H or MASK =0FOH MASK =$FO White space must exist after the label, but none is required after the '='. EXPORT. This directive can be used to define labels (symbols) that are to be written to the export symbol file. The symbols are written as equates (using the .EQU directive) so that the resulting file can be included in a subsequent assembly. This feature can help overcome some of the deficiencies of TASM due to its lack of a relocating linker. The format is: [<label>] .EXPORT <label > The following example illustrates the use of the EXPORT directive and the format of the resulting export file: Source file: .EXPORT read_byte .EXPORT write_byte .EXPORT open_file Resulting export file: read_byte .EQU $1243 write_byte .EQU $12AF open_file .EQU $1301 IFDEF. This directive can be used to optionally assemble a block of code. It has the following form: #IFDEF <macro_label> When invoked, the list of macro labels (established via DEFINE directives) is searched. If the label is found, the following lines of code are assembled. If not found, the input file is skipped until an ENDIF or ELSE directive is found. Lines that are skipped over still appear in the listing file, but a '~' will appear immediately after the current PC and no object code will be generated (this is applicable to IFDEF, IFNDEF, and IF). IFNDEF. This directive is the opposite of the IFDEF directive. The block of code following is assembled only if the specified macro_label is undefined. It has the following form: #IFNDEF <macro_label> When invoked, the list of macro labels (established via DEFINE TASM - Table Driven Assembler Version 2.8 Page 22 directives) is searched. If the label is not found, the following lines of code are assembled. If it is found, the input file is skipped until an ENDIF or ELSE directive is found. IF. This directive can be used to optionally assemble a block of code dependent on the value of a given expression. The format is as follows: #IF <expr> If the expression <expr> evaluates to non-zero, the following block of code is assembled (until an ENDIF or ELSE is encountered). INCLUDE. The INCLUDE directive reads in and assembles the indicated source file. INCLUDEs can be nested up to six levels. This allows a convenient means to keep common definitions, declarations, or subroutines in files to be included as needed. The format is as follows: #INCLUDE "<filename>" The <filename> must be enclosed in double quotes. Here are some examples: #INCLUDE "macros.h" #include "equates" #include "subs.asm" LIST/NOLIST. These directives can be used to alternately turn the output to the list file on (LIST) or off (NOLIST). The format is: .LIST .NOLIST LSFIRST/MSFIRST. These directives determine the byte order rule to be employed for the WORD directive. The default (whether correct or not) for all TASM versions is the least significant byte first (LSFIRST). The following illustrates its effect: 0000 34 12 .word $1234 0002 .msfirst 0002 12 34 .word $1234 0004 .lsfirst 0004 34 12 .word $1234 ORG. This directive provides the means to set the Instruction Pointer (a.k.a. Program Counter) to the desired value. The format is: [<label>] .ORG <expr> The <label> is optional. The Instruction pointer is assigned the value of the expression, <expr>. For example, to generate code starting at address 1000H, the following could be done: TASM - Table Driven Assembler Version 2.8 Page 23 start .ORG 1000H The expression (<expr>) may contain references to the current Instruction Pointer, thus allowing various manipulations to be done. For example, to align the Instruction Pointer on the next 256 byte boundary, the following could be done: .ORG (($ + 0FFH) & 0FF00H) ORG can also be used to reserve space without assigning values: .ORG $+8 An alternate form of ORG is '*=' or '$='. Thus the following two examples are exactly equivalent to the previous example: *=*+8 $=$+8 PAGE/NOPAGE. These directives can be used to alternately turn the paging mode on (PAGE) or off (NOPAGE). If paging is in effect, then every sixty lines of output will be followed by a Top of Form character and a two line header containing page number, filename, and the title. The format is: .PAGE .NOPAGE The number of lines per page can be set with the '-p' command line option. SET. This directive allows the value of an existing label to be changed. The format is: <label> .SET <expr> The use of the SET directive should be avoided since changing the value of a label can sometimes cause phase errors between pass 1 and pass 2 of the assembly. SYM/AVSYM. These directives can be used to cause a symbol table file to be generated. The format is: .SYM ["<symbol_filename>"] .AVSYM ["<symbol_filename>"] For example: .SYM "symbol.map" .SYM .AVSYM "prog.sym" .AVSYM The two directives are similar, but result in a different format of the symbol table file. The format of the SYM file is one line per TASM - Table Driven Assembler Version 2.8 Page 24 symbol, each symbol starts in the first column and is followed by white space and then four hexadecimal digits representing the value of the symbol. The following illustrates the format: label1 FFFE label2 FFFF label3 1000 The AVSYM is provided to generate symbol tables compatible with the Avocet 8051 simulator. The format is similar, but each line is prefixed by an 'AS' and each symbol value is prefixed by a segment indicator: AS start C:1000 AS read_byte C:1243 AS write_byte C:1280 AS low_nib_mask N:000F AS buffer X:0080 The segment prefixes are determined by the most recent segment directive invoked (see BSEG/CSEG/DSEG/NSEG/XSEG directives). TEXT. This directive allows an ASCII string to be used to assign values to a sequence of locations starting at the current Instruction Pointer. The format is: [<label>] .TEXT "<string>" The ASCII value of each character in string is taken and assigned to the next sequential location. Some escape sequences are supported as follows: Escape Sequence Description ___________________________________________________ \n Line Feed \r Carriage return \b Backspace \t Tab \f Formfeed \\ Backslash \" Quote \000 Octal value of character Here are some examples: message1 .TEXT "Disk I/O error" message2 .text "Enter file name " .text "abcdefg\n\r" .text "I said \"NO\"" TITLE. This directive allows the user to define a title string that appears at the top of each page of the list file (assuming the PAGE mode is on). The format is: TASM - Table Driven Assembler Version 2.8 Page 25 .TITLE "<string>" The string should not exceed 80 characters. Here are some examples: .TITLE "Controller version 1.1" .title "This is the title of the assembly" .title "" WORD. This directive allows a value assignment to the next two bytes pointed to by the current Instruction Pointer. The format is: [<label>] .WORD <expr> [, <expr> ...] The least significant byte of <expr> is put at the current Instruction Pointer with the most significant byte at the next sequential location (unless the MSFIRST directive has been invoked). Here are some examples: data_table .WORD (data_table + 1) .word $1234 .Word (('x' - 'a') << 2) .Word 12, 55, 32 TASM - Table Driven Assembler Version 2.8 Page 26 OBJECT FILE FORMATS TASM supports four object file formats: 1. Intel Hex (default). 2. MOS Technology Hex. 3. Motorola Hex. 4. Binary Each are described below: Intel Hex Object Format. This is the default format. This format is line oriented and uses only printable ASCII characters except for the carriage return/line feed at the end of each line. Each line in the file assumes the following format: :NNAAAARRHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHCCTT Where: All fields marked 'hex' consist of two or four ASCII hexadecimal digits (0-9, A-F). A maximum of 24 data bytes will be represented on each line (override the 24 byte default with the '-o' command line option). : = Record Start Character NN = Byte Count (hex) AAAA = Address of first byte (hex) RR = Record Type (hex, 00 except for last record which is 01) HH = Data Bytes (hex) CC = Check Sum (hex) TT = Line Terminator (carriage return, line feed) The last line of the file will be a record conforming to the above format with a byte count of zero: :00000001FF The checksum is defined as: sum = byte_cnt+addr_hi+addr_lo+rec_type + (sum of all data bytes) checksum = ((-sum) & ffh) TASM - Table Driven Assembler Version 2.8 Page 27 MOS Technology Hex Object Format. This format is line oriented and uses only printable ASCII characters except for the carriage return/line feed at the end of each line. Each line in the file assumes the following format: ;NNAAAAHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHCCCCTT All fields marked 'hex' consist of two or four ASCII hexadecimal digits (0-9, A-F). A maximum of 24 data bytes will be represented on each line (override the 24 byte default with the '-o' command line option). ; = Record Start Character NN = Byte Count (hex) AAAA = Address of first byte (hex) HH = Data Bytes (hex) CCCC = Check Sum (hex) TT = Line Terminator (carriage return, line feed) The last line of the file will be a record with a byte count of zero (';00'). The checksum is defined as: sum = byte_cnt+addr_hi+addr_lo+rec_type + (sum of all data bytes) checksum = (sum & ffffh) Motorola Hex Object Format. This format is line oriented and uses only printable ASCII characters except for the carriage return/line feed at the end of each line. Each line in the file assumes the following format: S1NNAAAAHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHCCCCTT All fields marked 'hex' consist of two or four ASCII hexadecimal digits (0-9, A-F). A maximum of 24 data bytes will be represented on each line (override the 24 byte default with the '-o' command line option). S1 = Record Start Prefix NN = Byte Count (hex) (data byte count + 3) AAAA = Address of first byte (hex) HH = Data Bytes (hex) CC = Check Sum (hex) TT = Line Terminator (carriage return, line feed) The last line of the file will be a record with a byte count of zero ('S903AAAACCTT'). The checksum is defined as: sum = byte_cnt+addr_hi+addr_lo+rec_type + (sum of all data bytes) checksum = (~sum & ffh) Binary Object Format. This file format has only a binary representation of each data byte with no address, checksum or format TASM - Table Driven Assembler Version 2.8 Page 28 description, whatsoever. It is often a convenient format to use to pass the data to other programs on your PC (like a PROM programmer package) but because of the non-printability and lack of address information, it is not often used to transmit the code to other systems. Note that when this object format is selected (-b option), the -c option is forced. This is done so that no ambiguity arises as a result of the lack of address information in the file. Without the -c option, discontinuous blocks of object code would appear contiguous. TASM - Table Driven Assembler Version 2.8 Page 29 LISTING FILE FORMAT Each line of source code generates one (or more) lines of output in the listing file. The fields of the output line are as follows: 1. Current source file line number (4 decimal digits). 2. An optional '+' appears if this is an 'INCLUDE' file. (One '+' for each level of INCLUDE invoked). 3. Current Instruction Pointer (4 hex digits). An optional '~' follows the Instruction Pointer if the line of source code is not being assembled because of an IFDEF, IFNDEF, or IF directive. 4. Resulting code/data generated from this source line (two hex digits per byte, each byte separated by a space, up to six bytes per line). 5. The source line exactly as it appears in the source file. If paging is enabled (by either the '-p' option flag or the .PAGE directive) some additional fields will be inserted into the listing file every 60 lines. These fields are: 1. Top of Form (form feed). 2. Assembler identifier (e.g. "TASM 6502 Assembler"). 3. Initial source file name. 4. Page number. 5. Title. TASM - Table Driven Assembler Version 2.8 Page 30 PROM PROGRAMMING A wide variety of PROM programming equipment is available that can use object code in one or more of the formats TASM supports. Here are some notes concerning the generation of code to be put in PROMs: 1. PRESET MEMORY. It is often desirable to have all bytes in the PROM programmed even if not explicitly assigned a value in the source code (e.g. the bytes are skipped over with a .ORG statement). This can be accomplished by using the -c (contiguous block) and the -f (fill) command line option flags. The -c will ensure that every byte from the lowest byte assigned a value to the the highest byte assigned a value will be in the object file with no gaps. The -f flag will assign the specified value to all bytes before the assembly begins so that when the object file is written, all bytes not assigned a value in the source code will have a known value. As an example, the following command line will generate object code in the default Intel Hex format with all bytes not assigned a value in the source set to EA (hex, 6502 NOP): tasm -65 -c -fEA test.asm 2. CONTIGUOUS BLOCKS. To ensure that TASM generates object code to cover the full address range of the target PROM, put a .ORG statement at the end of the source file set to the last address desired. For example, to generate code to be put in a 2716 EPROM (2 Kbytes) from hex address $1000 to $17ff, do something like this in the source file: ;start of the file .ORG $1000 ;rest of the source code follows <source code> ;end of the source code .ORG $17ff .BYTE 0 .END Now, to invoke TASM to generate the code in the binary format with all unassigned bytes set to 00 (6502 BRK), do the following: tasm -65 -b -f00 test.asm Note that -b forces the -c option. TASM - Table Driven Assembler Version 2.8 Page 31 ERROR MESSAGES ERROR MESSAGE DESCRIPTION _______________________________________________________________________ Unrecognized directive A statement starting with a '.' or '#' has a mnemonic that is not defined as a directive. Unrecognized instruction A statement has an opcode mnemonic that is not defined. Unrecognized argument A statement has an operand format that is not defined. Label value misaligned The value of a label appears to have a different value on the second pass then it was computed to have on the first pass. This is generally due to Zero Page Addressing mode problems with the 6502 version of TASM. Labels that are used in operands for statements that could utilize Zero Page addressing mode should always be defined before used as an operand. Label table overflow To many labels have been encountered. No END directive before EOF The source file did not have an END directive in it. This is not fatal, but may cause the last object file record to be lost. No files specified TASM was invoked with no source file specified. Unknown option flag TASM was invoked with an undefined option flag on the command line. Source file open error TASM was not able to open the specified source file. List file open error TASM was not able to open the specified list file. Object file open error TASM was not able to open the specified object file. Unknown token Unexpected characters were TASM - Table Driven Assembler Version 2.8 Page 32 encountered while parsing an expression. maximum number of macros exceeded To many macros (DEFINEs) have been encountered. macro buffer overflow The buffer from which space is allocated for macro definitions is exhausted. range of relative branch exceeded A branch instruction exceeds the maximum range (6502 Version). macro expansion too long. The expansion of a macro resulted in a line that exceeded the maximum length. Heap overflow on label definition. TASM was unable to allocate memory to store the label. No such label yet defined. A SET directive was encountered for a label not yet defined. The value of labels that are modified by the SET directive must already exist. No indirection for this instruction. A parenthesis was found around the operand expression. This may indicate an attempt to use indirection where it is inappropriate. Non-unary operator at start of expression A non-unary operator at the beginning of an expression is invalid. Binary operator where value expected. Two binary operators in a row indicate a missing value. Invalid token where value expected. Two binary operators in a row are not allowed. Duplicate label. Duplicate label checks are optionally enabled by the '- a' option. Label must pre-exist for SET. The SET directive can only be applied to an existing label. Unused data in MS byte of argument. An instruction or directive used the least significant byte of an argurment and left the most significant byte unused, but it was non-zero. TASM - Table Driven Assembler Version 2.8 Page 33 File name too short A file name on the command line is fewer than 3 characters. This is done to prevent a garbled option flag from being taken as a source file, which in turn can result in the source file taken as an object file (which are truncated at start-up time). Unkown option Flag. Invalid option flag has been specified on the command line. invoke TASM with nothing on the command line to see a list of valid options. TASM - Table Driven Assembler Version 2.8 Page 34 BUGS AND LIMITATIONS Limitations and Specifications _____________________________________________________________________ Maximum number of labels 2000 Maximum length of labels 13 characters Maximum address space 64 Kbytes (65536 bytes) Maximum number of nested INCLUDES 4 Maximum length of TITLE string 79 characters Maximum source line length 255 characters Maximum length after macro expansion 255 characters Maximum length of expressions 255 characters Maximum length of pathnames 79 characters Maximum length of command line 127 characters Maximum number of instructions (per table) 600 Maximum number of macros 1000 Maximum number of macro arguments 10 Maximum length of macro argument 16 characters Heap size (for labels, macros, & buffers) 20000 bytes Memory requirements 160K Bugs 1. The 8048 version of TASM does not check for use of memory beyond any reasonable bounds (e.g. an 8048 has a maximum address space of 4 Kbytes but TASM will let you pretend that you have 64 Kbytes). 2. Expression evaluation has no operator precedence in effect which can make for unexpected results if not explicitly grouped with parenthesis. 3. First page of listing file will not show a user defined title (defined via TITLE directive). 4. TASM sometimes does not generate error messages for improperly formed expressions. 5. TASM does not generate an error message when a EQU directive has an undefined label on the right hand side. TASM - Table Driven Assembler Version 2.8 Page 35 APPENDIX A - ORDERING INFORMATION TASM is distributed as shareware. The shareware portion of the product may be freely copied and used for evaluation purposes. Use of TASM beyond an evaluation period of 90 days requires registration. Registered users receive the following benefits: 1. The recent version of TASM. 2. TASM source code (in C). 3. Bound TASM manual. 4. Telephone support. 5. Knowledge that they are supporting the development of useful but inexpensive software. DESCRIPTION UNIT PRICE PRICE ____________________________________________________________________ TASM Registration (TASM disk, manual, & source) $40.00 _______ TASM Site Registration (for sites with multiple 90.00 _______ users. Includes same materials as above.) TASM User's Manual (included above) 10.00 _______ TASM update for registered users 10.00 _______ (latest disk (with source), and manual) Subtotal _______ Tax (Washington state residents add 8.1%) _______ Billing fee (for orders not accompanied by check) 10.00 _______ Foreign postage (outside North America) add $10.00 _______ (Foreign orders must be in US funds drawn on a US bank) TOTAL (post paid) _______ Which processors are of primary interest to you? __________________ (This is for information only. You will receive all current tables). Shipping Address: ______________________________________________ ______________________________________________ ______________________________________________ Send check or money order to: Speech Technology Incorporated Software Division 837 Front Street South Issaquah, WA 98027 USA