Home

M6800 Family Assembler User's Manual

1. e 121 8 3 Macro Model Statements e 123 8 4 String Concatenation ee 123 8 5 Automatic Label Field Generation o e 124 8 0 MEND v miog a A AR Me As 124 Su MENTE cot a A Soe A AA A AA AA 124 8 8 Macro Call ciao a a a a 125 8 9 Nested Macro CallSicicios iio E tn SP a ER edo ee A iA 127 8 10 Hints for Using Macros e 127 9 Relocatable Assembly 129 9 1 Preliminary Discussion e 129 9 2 Invoking Relocatable Assembly 0 o e 130 9 3 Features of M6800 Relocatable Assembly o 131 9 4 Implications of Relocatable Assembly o o 133 9 5 Example Usage of Relocatable Assembly 136 9 6 A Few Words About the S Record Format o o 144 10 Assembly Errors 147 11 Summary of Nonstandard Features 161 vil ASSEMBLER USER S GUIDE viii Chapter 1 Introduction This manual is a user s guide for the Western Microsystems M6800 Family Assembler For the sake of brevity the rest of this manual refers to the Western Microsystems M6800 Family Assembler simply as the assembler or as the M6800 family assembler whenever one of the simpler terms can be used without ambiguity The assembler runs on any VAX computer system that uses the VMS operating system The reader is expected to be somewhat familiar with the VAX VMS environment so this manual does not include explanations of simple VM
2. lt expression gt HI lt term gt lt expression gt higher than lt term gt lt expression gt HS lt term gt lt expression gt higher than or same as lt term gt lt expression gt EQ lt term gt lt expression gt equal to lt term gt lt expression gt NE lt term gt lt expression gt not equal to lt term gt lt expression gt LS lt term gt lt expression gt lower than or same as lt term gt lt expression gt LO lt term gt lt expression gt lower than lt term gt For example the value of the expression 5 GT 3 is 1 since 5 is in fact greater than 3 but the value of the expression 5 4 GE 2 is zero since 5 4 1 which is not greater than or equal to 2 Also the value of the expression 5 LE 3 is 1 since the relationship is true and the value of the expression 5 LS 3 is also 1 since that relationship is also true However while the value of the expression 7 LT 5 is 1 the value of the expression 7 LO 5 is zero The LO operator specifies an unsigned comparison and 7 equals FFF9 which is definitely not lower than 5 as an unsigned number For complete examples that illustrate good uses for the relational operators please refer to the section that explains conditional assembly 39 ASSEMBLER USER S GUIDE 4 2 5 Selection Operators The assembler s selection operators return either the expression on the left hand side or the term on the right hand side
3. The MAX operator selects the larger of two signed values as the value of an expression and the MIN operator selects the smaller of two signed values as the value of an expression For example the value of the expression 5 MAX 10 is 10 and the value of the expression 5 MIN 10 is 5 Similarly the value of the expression 5 MAX 10 is 5 since 5 is larger than 10 and the value of the expression 5 MIN 10 is 10 The MAXUNS operator selects the larger of two unsigned values as the value of an expression and the MINUNS operator selects the smaller of two unsigned values as the value of an expression For example the value of the expression 5 MAXUNS 10 is 10 and the value of the expression 5 MINUNS 10 is 5 The value of the expression 5 MAXUNS 10 is FFFB because 5 equals FFFB 10 equals FFF6 and FFFB is greater than FFF6 when both numbers are considered to be unsigned Similarly the value of the expression 5 MINUNS 10 is FFF6 since FFF6 is smaller than FFFB when the numbers are considered to be unsigned As a final example the value of the expression 1 MAXUNS 1 is FFFF 4 2 6 Unary Operators The assembler allows one unary operator immediately before any term Since a paren thesized subexpression is a term however you can apply a unary operator to anything by using parentheses Also you can apply multiple unary operators to a single term by using parentheses For example NOT
4. e Relocatable symbols or expressions cannot be multiplied divided or operated on with the logical operators This is because the linker LK6800 can only adjust values by adding or subtracting the load address of the program section i e addresses are defined to within an arbitrary additive constant Improper operators will cause the assembler to generate an UO error e A relocation count is maintained for each program section represented within an ex pression Adding a relocatable symbol causes the relocation count to be incremented 133 ASSEMBLER USER S GUIDE subtracting a relocatable symbol decrements the relocation count After an expression has been evaluated the following criteria must be met All section counts except for one must be zero This is because the linker can only adjust the object value by one load address All but one section s load addresses must cancel The exceptional section must have a count of either zero or one or minus one This is because the linker can only add or subtract the load address to adjust the output value When an expression is used within the one byte immediate addressing mode the indexed addressing mode or with the FCB directive all section counts except the BSCT count must be zero This is necessary to be able to guarantee that the adjusted value will fit in one byte While absolute section ASCT symbols can be evaluated fully at assembly time relocatable symbols
5. value_n The FDB statement requires at least one operand expression and it allows any number of operand expressions after the first one An FDB statement with multiple operand expressions is effectively a shorthand notation for a series of individual FDB statements so a single FDB statement with multiple operand expressions is equivalent to a series of individual FDB statements with the same set of operand expressions The assembler advances the location counter after it processes each expression in an FDB statement s operand field so the location counter symbol in an FDB statement always refers to the memory address of the first object data byte that is being generated by the expression that contains the location counter symbol If a label appears in the label field of an FDB statement the assembler defines that label s value to be the memory address of the first data byte that is generated by the FDB statement 5 26 GOIF The GOIF statement allows a user to perform conditional assembly by skipping a condi tional section of source code or including the conditional section of source code as part of the program during the assembly process The format for the GOIF statement is as follows GOIF conditional_expression target If the conditional expression is false i e equal to zero the assembler simply continues the assembly with the next line and the GOIF statement essentially has no effect If on the other hand the condi
6. ASSEMBLER USER S GUIDE IP A parameter name must begin with an alphabetic character Invalid Parameter name error A nonalphabetic character occurred where a formal parameter name should begin The macro call statement is malformed in some way Perhaps the user attempted to list positional parameter names of the form n in a macro call model statement IS An invalid symbol appears in the symbol list Invalid Symbol error An invalid symbol appears in a symbol list and the assembler skips the invalid symbol but continues to process the rest of the symbol list This error occurs with the XDEF and XREF statements IT A tab stop cannot be less than 1 or greater than 80 Invalid Tab stop A tab stop in a column less than 1 or greater than 80 is specified in an OPT TABS n1 n2 nj statement IX An improper base address or index register has been specified IndeXing error A negative base address is specified with indexed addressing and or X is not specified correctly as the index register LL The local label in the label field is malformed Local Label definition error Column 1 contains a decimal digit thus indicating the presence of a local label in the label field but the syntax for the local label is incorrect The correct syntax is jH where 0 lt j lt 9 LO The location counter exceeds FF in the base section Location counter Overflow error The base section s location counter overflow
7. LBLB LBLC LBLA 4H RMB 2B 3B 1B The coding sequences in the next example illustrate the fact that the same local label can be defined over and over and they also illustrate the fact that a jB or jF local label in an expression never refers to the jH label in the same statement The two sequences of instructions are functionally identical but the sequence on the left uses symbolic labels while the sequence on the right uses local labels LBLA EQU 29 3H EQU 29 ORG LBLA 39 ORG 3B 39 LBLB FDB LBLA LBLC LBLD 3H FDB 3B 3F 1F LBLC FCB LBLD LBLB 3H FCB 1F 3B LBLD RMB LBLC 1H RMB 3B The examples that are shown above are probably somewhat confusing because they are deliberately contrived to illustrate the subtle points of the rules regarding local labels In actual practice however local labels aren t confusing at all The following example illustrates a typical application of local labels The code on the left uses symbolic labels of dubious value and the analogous code on the right makes appropriate use of local labels LDX LIST LDX LIST LDAA LENGTH LDAA LENGTH LDAB KEY LDAB KEY CLR COUNT CLR COUNT BACK CMPB 0 X 1H CMPB 0 X BNE NEXT BNE 2F INC COUNT INC COUNT NEXT INX 2H INX DECA DECA BGT BACK BGT 1B Local labels spare the programmer from the tiresome and unrewarding chore of inventing symbolic names for nearby locations when the programmer merely wants to transfer control to an instruction a few lines away Pr
8. MEXIT Macro exit Stop expanding the current macro The MEXIT statement allows a user to stop expanding a macro before reaching the end of the macro MON Terminate the assembly The MON state ment is equivalent to the more familiar END statement The assembler recognizes the MON statement to maintain compatibil ity with Motorola s definition of the assem bly language NAME string The NAME statement is equivalent to the TITLE statement The TITLE statement is preferred OPT optionl The OPT statement controls the assembler s various options Continued AT ASSEMBLER USER S GUIDE Table 5 1 Continued Directive Parameter Remarks ORG address Set the assembler s location counter equal to the value of the address expression thus specifying the memory location for the next object value PAGE The PAGE statement is equivalent to the EJECT statement The EJECT statement is preferred PRINT value comment Print the value of the expression and the text of the comment as a line of the listing and as a message to the user s terminal PSCT For relocatable assembly only start using the location counter for the program section RCB count valuel Repeat constant byte The RCB statement is similar to the FCB statement The count ex pression defines a count that tells how many times the assembler should perform the FCB function
9. STX SAVE_X Save the current value of the XR and LDX TAB1 load the address of the first location in the block LOOP_A CLR 0 X Clear each memory location and advance 120 CHAPTER 8 MACRO ASSEMBLY INX the index register TAB1 TABSIZ 1 address th CPX TAB1 TABSIZ 1 1 location Loop until the 1 BNE LOOP_A LDX SAVE_X Restore the XR Notice that wherever FROM appears in the model the assembler substituted TAB1 and wherever TO appears in the model the assembler substituted TAB1 TABSIZ 1 The as sembler even used the substitution string where the word to appears in the comment field The user should choose parameter names carefully to avoid such unintentional substitutions Expansion overflow occurred on two lines because the parameter string was much longer than the parameter name In both cases only the comments were truncated but the user should check expansion overflow errors to ensure nothing of value has been lost Refer to the discussion in Chapter 10 regarding the EO error message and in Chapters 5 and 6 regarding suppression of the expression overflow error message There is nothing magical about the use of underscore _ in the names SAVE X and LOOP_ in the example but it does illustrate a technique the user may want to consider Underscore is one of the special characters that the assembler treats as alphabetic The use of some special character on labels associated with macros may help avoid conflicts with oth
10. 10 If XX is false i e zero then the assembler includes the intervening source lines in the assembly The statement GOUNLS YY 200 means If the expression YY is false i e zero then 112 CHAPTER 7 CONDITIONAL ASSEMBLY skip over the intervening source lines until reaching the target label 200 If YY is true i e nonzero then the assembler includes the intervening source lines in the current assembly Notice that the assembler does not require XX and YY to use either explicitly or implicitly the special run time conditional assembly parameter token CAMASK the fact that XX and YY can use CAMASK however makes conditional assembly a dynamic feature that the user can control at assembly time The GOTO statement is an unconditional skip statement and the assembler s action does not depend upon the evaluation of any expression For example GOTO 50 will always result in skipping source lines until a statement with the system label 50 occurs Similarly we can explain GOIFZ and GOULZ with some examples GOIFZ XX 10 means If the expression XX ANDed with the value of the conditional assembly parameter yields zero go to the statement labeled 10 i e skip all source lines until finding the system label 10 If the expression XX ANDed with the conditional assembly parameter does not yield zero the GOIFZ statement is effectively ignored Similarly the statement GOULZ YY 200 means If the ex
11. 5 61 TITLE The TITLE statement defines a title that is to be printed at the top of each page of any symbolic listing that is produced An example of a TITLE statement is as follows TITLE Gas Pump Controller Program Version 3 7 If this statement occurs in a program the title Gas Pump Controller Program Version 3 7 will be printed at the top of every page of the symbolic listing until another TITLE statement is encountered The string that appears in a TITLE statement can contain as many characters as the size of a source line will allow Furthermore the user can change titles as often as desired in a program so various sections of a program can have different titles The assembler automatically performs an EJECT function when it processes a TITLE statement so the new section of the program automatically starts on a new page If a TITLE statement is included as the first statement of a program or at least before any statements that generate lines in the listing that title will be effective on the first page of the listing Otherwise the last title that was used during pass 1 will be effective at the beginning of pass 2 The leading period feature as described in the detailed discussion of the SBTTLE directive above is also active in the TITLE statement 5 62 VERIFY The VERIFY statement provides a convenient method for a programmer to tell the assem bler to verify some assumption that the programmer is making in the pro
12. DEBUG DIRECT NODIRECT EOERR NOEOERR FORMAT NOFORMAT IGNORE NOIGNORE INCLIST INCNOLIST LBLLIMIT n LINENUMBER n LIST NOLIST LLRANGE n LONG SHORT LPAGE n MPERR NOMPERR Any statement that begins with its first significant character more than n columns from the beginning is considered to be a comment line Select the display base for the object code in the assembly listing as octal decimal or hexadecimal respectively Include the symbol table in the object file for debugging purposes Enable or disable the prerogative of the assembler to generate object code with the direct addressing mode Report or do not report macro expansion overflow errors to the assembly listing Select or deselect automatic source line formatting This option is provided for backward compatibility only Suppress or list respectively source lines that start with a pe riod in column one Include or suppress respectively the INCLUDE statements and files from the listing Set the number of significant characters in a symbolic label equal to n Set the current line number of the assembly listing to n Sub sequent source lines are then incremented from this value Increment or decrement respectively the listing level in the as sembly listing Set the local label range equal to n Any references to a local label outside of this range of line numbers generates an assembly error List or suppress respectivel
13. Formatting commands might be necessary for example to set the margins to turn justification on and off and to control other options that the text formatter might have Formatting commands are commonly known as dot commands because nearly every text formatter in widespread use recognizes a formatting command by the fact that it starts with a period in column 1 The use of a period in column 1 to indicate a formatting command has become a de facto standard because a period seldom occurs at the left margin in ordinary text A programmer commonly includes dot commands in a source program with the intention of using a text formatter to justify margins and to provide other enhancements before printing a final listing of a frozen or released version of the program However the user might assemble the program several perhaps even many times during program development before the program is ready to be frozen and the text formatter provides little or no benefit for these intermediate assemblies In fact a programmer who uses a text formatter to justify margins for every intermediate assembly is clearly wasting valuable computer time that could be better used for more worthwhile purposes such as processing a payroll or playing Adventure 18 CHAPTER 2 SOURCE SYNTAX The assembler has a time saving feature that allows a user to bypass the formatting step and to assemble a program that contains dot commands The assembler recognizes a dot comman
14. Listing 105 ASSEMBLER USER S GUIDE files and object files once they have served their immediate purposes should no longer be kept in the directory since they take up valuable space and can easily be generated again from the source file if necessary As another possibly useful suggestion the user is forewarned against the potentially vicious cycle of edit assemble debug during the development of a source program Most users find that careful preparation and documentation away from the machine can greatly improve software productivity The subconscious mind out on an evening walk can be an amazingly effective debugger While printing of listing files has already been covered the discussion in Section 6 1 promised detailed instructions in downloading object files from VAX VMS to a microcom puter and this information follows 6 5 Downloading Object Files from VAX VMS The TYPE command in the VAX can be used to send an object file to the terminal for downloading into a microcomputer To download an object file named LAB5 OBJ from the VAX into a microcomputer in the Digital Systems Lab at Arizona State University proceed as follows 1 Put the communications control switch on the microcomputer into the CPU position and sign onto the VAX Use the same sign on procedures that are used for any ordinary VAX run 2 After the program has been assembled without any errors and the object file has been created move the communic
15. X X or X new additions e The A or B opcode extension must be adjacent to the opcode when it is used No intervening spaces are permitted For example LDA A is illegal but LDAA is recog nized As a result of this rule users may define A and B as ordinary labels in their programs With the Motorola version of the assembler these are special characters and may not be used as labels e The Motorola assembler recognizes LDAA X as being equivalent to a LDAA 0 X instruction but this assembler requires the second form As a result this assembler permits X to be used as a label whereas Motorola does not e The assembler doesn t permit the use of suffixes on numbers to specify bases By eliminating the number suffixes and recognizing prefixes only the assembler eliminates several exceptions and special cases that the user must consider when using suffixes The allowed prefixes are for binary numbers or O for octal numbers the digits 1 through 9 for decimal numbers and for hexadecimal numbers We provide the for octal numbers to be consistent with the Motorola definition 161 ASSEMBLER USER S GUIDE e If the first character in the operand field is an exclamation point e g LDAA SAM this assembler generates an instruction that uses direct addressing This feature allows the user to dictate direct addressing when desired instead of letting the assembler choose whether to use direct or ext
16. and the first nonblank character on the next line will begin the string Examples X ALPHA BRAVO CHARLIE MOD DELTA ALPHA BETA 1 e Any group of characters or a null string preceded by lt and and followed by gt on the same line Examples lt This is a string gt lt gt e Any group of characters or a null string delimited by a nonblank character of the user s choice The opening delimiter must be preceded by an up arrow or caret 7 and a matching delimiter must close the string on the same line Examples So is this Even this is 0 K 7 HIN A ODIN 7 gt lt __ Notice that strings fall into two general categories delimited and nondelimited Non delimited strings may be continued over more than one line and the assembler will copy the string onto successive lines For example we could have used continuation in the last example to minimize expansion overflow If the call had been CLRBLK FROM TAB1 TO TAB1 1 TABSIZ 122 CHAPTER 8 MACRO ASSEMBLY the 8th line of the macro model would have been expanded as CPX TAB1 1 TABSIZ 1 The assembler supports continuing an expression string over any number of lines 8 3 Macro Model Statements Each line of the macro model can contain characters that are to be copied verbatim parameter names or parameter indices for positional parameters concatenation operators 1 and automatic label fields 8 4 Strin
17. and the largest acceptable binary number is 41111111111111111 27 ASSEMBLER USER S GUIDE Table 4 1 Summary of Terms Type of Term Example Remarks Binary Number 01000101 Starts with Octal Number 0123456 Starts with zero Decimal Number 53147 Starts with any nonzero digit Hexadecimal Number FEED Starts with Symbolic Label RESULT Any valid label Local Label Forward 3F jF forO lt j lt 9 Local Label Backward 5B jBfor0 lt j lt 9 Quoted ASCII Character np ASCII value of character Present Location Symbol Current memory address RAD 40 Hash String 7 ABC 16 bit RAD 40 hash code Conditional Assembly Mask CAMASK Run time CAMASK value Subexpression ENTRY LINK 1 Parenthesized subexpression 4 1 2 Octal Numbers Any series of one or more octets 0 7 starting with a zero is an octal number The assembler accepts octal numbers with as many as 16 significant bits because users sometimes find it convenient to code 16 bit constants as octal values The smallest acceptable octal number is 0 and the largest acceptable octal number is 0177777 4 1 3 Decimal Numbers Any series of one or more decimal digits 0 9 starting with a nonzero digit is a decimal number The assembler accepts decimal numbers in the range from 1 through 65 535 The number zero is technically an octal number but users commonly think of it as a decimal number In any case the value zero is the same for any base 4 1 4 Hexadecim
18. assembly to produce different object versions of their software to match the various hardware configurations that exist in the field They code one master source program with blocks of conditionally assembled code in it and when they assemble the program for a specific computer configuration they simply specify a conditional assembly parameter in accordance with the configuration output desired The assembler automatically assembles the proper blocks of code and the resulting program runs on the desired hardware configuration As a specific instance of this situation consider the M6800 assembler itself written for the VAX VMS operating system using the MACRO 32 assembler Because the 6800 and 6809 microprocessors are very similar a M6809 assembler was essentially contained within the existing M6800 assembler Any changes with respect to the slightly different opcodes and addressing modes between the two microprocessors are compensated for with judicious use of conditional assembly This selective inclusion of opcode tables legitimate expression tokens and program segments relevant to specific addressing modes can easily be made by toggling the value of two constants that appear at the very beginning of the assembler source 111 ASSEMBLER USER S GUIDE code In this respect the MACRO 32 assembler is not as flexible as the M6800 assembler since no dynamic parameter can be specified The fact that the value of the conditional assembly para
19. lt cr gt 90 CHAPTER 6 RUNNING THE ASSEMBLER At times the user may only be interested in finding assembly errors in a source file currently being developed In such a case and for other situations as well it may not be necessary to generate an object file The assembler recognizes the special file name NONE or for that matter noNE none and all such permutations of case to indicate that the corresponding file is not to be created It is a good idea to specify NONE for a file when it will not be needed because it prevents wasting both computer time and disk allocation especially for large files The user can use this technique to suppress the listing file the object file or both files for any given assembly This example also demonstrates that the three file names are fixed as positional parame ters The first position always belongs to the source file and must always be separated from the second position with a comma whenever action different from the default is specified Similarly the second position always corresponds to the listing file and a comma must be used to separate it from the third parameter whenever other than the default action is specified After the assembly in this example given that LAB1 ASM exists the listing file LAB1 LST will be created but no object file will be generated Because the action specified by NONE is not the default action for the object file it was necessary to skip over to the third pos
20. syntax error When the M6800 family assembler detects an error however it generates a specific understandable error message that pinpoints the cause of the error for the user In most cases the error message itself contains enough detail to eliminate any need for the user to refer to the user s manual for additional information Besides generating understandable error messages the assembler also links error lines together into an error list Below each statement with an error the assembler generates a linking message that specifies the line number of the previous statement with an error The summary at the end of the listing specifies the line number of the last statement if any with an error so a user can easily start at the end of an assembly and check each erroneous statement without bothering to examine the entire listing The assembler has one other notable feature that will quickly become apparent to the user The assembler is fast In benchmarks against other assemblers that run on VAX VMS systems the M6800 family assembler has consistently shown itself to be from 5 to 10 times as fast as the other assemblers The speed differential varies according to the specific application of course but every test case has shown that the M6800 family assembler is significantly faster than other assemblers The chapters that follow describe the assembler s features in detail The chapters and the assembler are organized in a way that all
21. the assembler can expediently assemble an intermediate version of a source file that will eventually be formatted by a text formatter in its final form The assembler accepts symbolic labels with as many as 80 characters per label so programmers can employ meaningful labels that are reasonably descriptive The assembler also accepts local labels A local label can be defined over and over for many different local applications in a single program Since local labels can be redefined they are especially suitable for local uses in which normal symbolic labels aren t appropriate Local labels are commonly used as destinations of local branches within a routine and they serve other local purposes as well The assembler optionally generates an alphabetized concordance listing that lists each nonlocal label along with the line number of the statement that defined the label the value of the label and the line number of each reference to the label The concordance listing is especially valuable for programmers who are developing or maintaining large programs with many variables and subroutines The assembler has a rich repertoire of operators that can be used in expressions For example besides recognizing the traditional arithmetic operators the assembler also recog nizes logical operators shift and rotate operators signed and unsigned relational operators and a few miscellaneous operators For users who require involved expressions the assem
22. DEST 2 where the symbol represents the memory address of the first byte of the branch class instruction A programmer who wants to use a branch class instruction simply specifies the desired destination address and the assembler automatically computes the self relative offset value that will result in a branch to the indicated destination address at execution time Therefore the operand field for a branch class instruction should contain an expression that directly specifies the desired destination address For example a user might say BGT LOOP to branch to location LOOP when the result of some operation is greater than zero Similarly a user might say BNE ERROR to branch to location ERROR if the condition codes don t indicate equality Note carefully that the programmer always specifies the desired destination address and that the assembler performs the menial computations that are necessary to generate the self relative offset that the microprocessor uses The programmer never needs to compute the offset value manually since the assembler automatically handles that chore 25 ASSEMBLER USER S GUIDE 26 Chapter 4 Expressions A typical instruction such as LDAA KEY or STX POINTER has an operand field that contains an expression Most expressions that programmers use are extremely simple ex pressions In fact in the large majority of cases an expression contains only one term such as the lab
23. For example suppose that a person is writing about Elizabeth I who was the queen of England from 1558 until 1603 Imagine the momentary confusion that might result for a reader if a text formatter happened to change the space after Elizabeth into a line break Generally speaking a person s name e g Thomas Jefferson or a date e g 4 July 1776 should not be split across a line or even expanded Similarly a range such as A Z or 1981 84 shouldn t be split in a way that causes a hyphen to be the first character of a line As another example consider the blank spaces in a simple equation such as a b c A text formatter that indiscriminately changed the space after the a into a line break would effectively mutilate the equation s readability Nearly every text formatter allows a user to specify some seldom used special character as a nonexpandable space i e as an indication of a single blank space that can t be changed or expanded For example a grave accent almost never occurs in English text so a person who is writing in English often includes a dot command to tell the text formatter that it should consider a grave accent to be a nonexpandable space Then the user can safely type Elizabeth I or a b c knowing that each grave accent will be translated into a single space in the text formatter s output The assembler conveniently classifies grave accents along with tabs and blank spa
24. Location Symbol When an asterisk is used as a term in an expression it represents the value of the assembler s location counter which contains the 16 bit memory address of the instruction that is currently being assembled The assembler updates the location counter as the last order of business when it assembles an instruction so the value of an asterisk for any executable instruction is always the memory address of the first byte of the instruction The value of an asterisk is similarly the address of the first byte of the object value for any 30 CHAPTER 4 EXPRESSIONS nonexecutable assembler directive that has a single operand field and generates an object value If the instruction that is currently being assembled is an assembler directive that doesn t generate an object value then the value of the location counter is the memory address that would have been used if the instruction had been an executable instruction The value of an asterisk for a compound assembler directive such as FCB or FDB is the memory address of the first byte of the field that is currently being processed Although the asterisk certainly has valid uses it can also be badly abused by poor programmers For example anyone who codes an instruction such as BRA 5 is making his or her program hard to read and hard to maintain First of all instruction lengths aren t necessarily one byte per instruction so the destination of the branch isn t immedi
25. The ABS and REL switches to this OPT statement select absolute or relocatable assembly respectively Absolute assembly is the default mode of object code generation When the user selects relocatable assembly a number of assembler directives become active to enable part of the assembly work to be done by the linker The user is referred to Chapter 9 on relocatable assembly and to the coverage in this chapter of such directives as IMPORT EXPORT ASCT BSCT CSCT DSCT and PSCT for more information The analogous run time options for determining mode of object code generation are ABS and REL 5 40 2 OPT CALIST CANOLIST By default the assembler lists all statements that are associated with conditional assem bly but the user can change between the list mode and the no list mode for conditional assembly statements as often as desired by selecting the CALIST or CANOLIST option respec tively There are levels associated with the list mode for conditional assembly statements The CANOLIST option decrements the list level and the CALIST option increments the list level The list level is initially zero and a negative list level will result in the suppression of conditional assembly statements Therefore the user can bracket a program segment 70 CHAPTER 5 ASSEMBLER DIRECTIVES with the CANOLIST CALIST pair with no net change to the current list level for conditional assembly statements i e CANOLIST and CALIST options at earl
26. The assembler ignores the extra right parenthesis ES A null string is not a valid operand for this statement Empty String error An FCC statement has no string at all or it has an empty string such as or Since there is no string to process the assembler processes the statement as if it were an EQU instruction FR A forward reference appears where it is not allowed Forward Reference error An expression contains a forward reference but the assembler must evaluate the expression accurately during pass 1 The assembler obviously can t eval uate the expression accurately during pass 1 if the expression contains a forward reference IA A macro parameter assignment statement is missing the equal sign Invalid macro parameter Assignment error The next non white space character fol lowing a formal parameter name in a macro call statement was not the equal sign as required IC A continued macro parameter string is followed by a blank line Invalid string Continuation error A blank line follows a continued macro parameter substitution string The valid portion of the string will be used ICH An invalid character appears in an expression Invalid CHaracter A character that is not valid in an expression appears in an expres sion The assembler ignores the invalid character and continues to evaluate the expression as if the invalid character weren t there II Invalid macro parameter index val
27. The default for the List NList option is List 6 3 15 LOC The LOC option tells the assembler to include the location counter field in the assembler output listing Conversely the NLOC option instructs the assembler to suppress the location counter field from the assembly listing Suppression of this field has questionable merit other than to narrow the listing width The default for the LOC NLOC option is LOC 6 3 16 LPP The LPP n option tells the assembler that the line printer or terminal as the case may be has n lines per physical page where n is a decimal integer greater than or equal to 32 The number of lines per physical page includes all lines from perforation to perforation The top and bottom margins therefore are included in the physical page line count The assembler uses the physical page size to format the listing file correctly for its intended output device If the listing file is printed on a device with a physical page size that is different from the physical page size that was specified in the LPP parameter the listing pages will not match the physical pages and if the NFFeed option is selected see above will eventually walk across the perforations The user can avoid this undesirable situation by simply specifying the correct value for the LPP parameter The default value for the LPP parameter is 66 which corresponds to standard 11 paper with printing at 6 lines per inch Since the assembler automati
28. as the ASCIT statement but the assembler still recognizes the FCC statement for purposes of backward compatibility Continued 45 ASSEMBLER USER S GUIDE Table 5 1 Continued Directive Parameter Remarks FDB valuel Form double byte Generate two object bytes for each expression The first object byte of a pair contains the most significant byte of the corresponding 16 bit signed or unsigned expression value and the second object byte of a pair contains the least significant byte of the expression value De fine the label if any from the label field to be the address of the first object byte GOIF expression target If the expression is true i e nonzero go forward to the target conditional label thus skipping the assembly of all interven ing lines The GOIF statement provides con ditional assembly GOIFZ expression target If the expression ANDed with the CAMASK parameter equals zero go forward to the target conditional label thus skipping the assembly of all intervening lines The GOIFZ statement can be used with conditional as sembly but it is no longer recommended GOTO target Go forward to the target conditional label thus skipping the assembly of all intervening lines The GOTO statement is used with con ditional assembly GOULZ expression target Go unless zero If the expression ANDed with the CAMASK parameter is nonzero go f
29. as they are in blocks of code The assembler also recognizes the ETEXT statement in text mode for rather obvious reasons and it recognizes the BTEXT statement in text mode so it can warn a user about a probable missing ETEXT statement Aside from these few statements the assembler treats all source lines in a text block as comment lines so a programmer can include any desired documentation in a block of text with no need to type any special comment character at the beginning of each line The BTEXT statement like its companion ETEXT statement is suppressed from the sym bolic listing so a text mode comment block appears in the listing without the distracting clutter of the BTEXT and ETEXT statements that bracket it If a programmer uses a BTEXT statement and forgets to include the matching ETEXT statement the assembler eventually reports at least one error In the worst case the assembler processes the rest of the program including the END statement as a text block and then reports that the END statement is missing More commonly the assembler reports a redundant BTEXT statement the next time it finds a BTEXT directive thus alerting the user to the fact that an ETEXT statement was probably omitted somewhere 55 ASSEMBLER USER S GUIDE 5 10 CHKSUM The CHKSUM statement generates a checksum data byte with a unique 8 bit value that can be added to the 8 bit sum of all object bytes since the previous CHKSUM statement if any to yi
30. cannot be fully evaluated since the offset applied at load time can drastically change the symbol s value However since base section BSCT variables cannot be relocated past memory page zero 00 to FF we can allow a BSCT variable in a one byte field e Any number of relocatable symbols including the asterisk denoted current location counter value and symbols defined in BSCT CSCT DSCT and PSCT are accepted by the expression evaluator but only six external references in an expression are accepted This is due to a restriction of the linker and violation of this limit will be marked with the TR assembler error Notice that while several relocatable symbols can and must cancel by alternate addition and subtraction of their load offsets the same does not apply to externally referenced symbols A final notion related to the use of imported symbols or relocatable labels concerns what is effectively a forward reference The assembler must evaluate an expression in a statement such as RMB or ORG or any of the conditional assembly directives during pass one of the assembly An imported label or a relocatable label is not fully defined until link time so an imported label or a relocatable label is not allowed in an expression that the assembler must evaluate during pass one of the assembly Using an imported label or a relocatable label in an expression that the assembler must evaluate during pass one causes the assembler to report an IMP or RLV
31. code values to the assembly listing file or device and NValue tells the assembler not to output the generated code values to the assembly listing The NValue option may be selected to reduce the assembly listing width when the code values are not important The default for the Value NValue option is Value 6 3 30 WIDTH The Width n option tells the assembler that the maximum number of columns in an output line for the assembly listing is n where n is a decimal integer in the range INdent 29 lt n lt 132 Output lines are truncated when they exceed the specified num ber of columns The Width specification applies only to the listing output and does not have any effect on the processing of the source line nor on the object module The maximum number of columns allowed in any one input source line is fixed at 126 columns Refer to the INdent option above for further information The default value for the Width option is 132 6 4 After The Assembly What happens to the source listing and object files after the assembly is over is not really the province of a user s guide on the M6800 assembler however a few more paragraphs of suggestions may be appreciated by some users Generally it is a good idea to prevent possible confusion once a project or source program has been finalized to remove multiple copies and previous versions of the source file and perhaps even to move the final version to another directory for safekeeping
32. command is specific to the VAX VMS installation for the Digital Computer Laboratory at Arizona State University More generally the command RUN M6800ASSEMBLER EXE must be issued to execute the assembler M6800ASSEMBLER EXE 89 ASSEMBLER USER S GUIDE is a generic name and it is the responsibility of the user at any given installation to discover the actual name in use there Using the assembler as described in Section 6 1 often meets the needs of the user however the assembler is by no means limited to this rigid usage Adopting the adage that one learns best by example the following paragraphs show sample interactions with the assembler along with the resultant outcomes To keep the examples simple it is assumed all files are in the home directory so that only file names and file extensions need to be considered The examples are not exhaustive and the user who still feels limited can refer to Section 6 6 in order to examine the parsing routine that processes the user s files and options specification input string The symbol lt cr gt denotes striking the carriage return key of the keyboard FILES SRC LST OBJ LAB1 lt cr gt This example repeats the context of the discussion in Section 6 1 The assembler at tempts to access a source file named LAB1 ASM in the current directory If this source file does not exist or if the assembler is unable to read this file a message alluding to a source file error is generated and the
33. communications channel between itself and MAIN Return from Interrupt NMI2 is the short one Again register and stack usage are not explicitly shown Store some value in the communications channel before returning Return from Interrupt 139 ASSEMBLER USER S GUIDE END End of subordinate module INTRPT ASM The file MAIN LST begins here oN on BS W 11 12 13 14 15 16 a 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 25 Oct 1985 OKOuvU Uv vVVV UY 0000 FFFC 0000 0000 0000 0000 0001 0002 0003 0004 0005 0006 0007 0000 0001 4F 50 54 49 4F 4E 53 3F 19 08 M6800ASM 1 1 MAIN ASM This is a contrived but hopefully informa assembly techniques The program consists of t in this file is a main routine which obtains i user regarding special interrupt handler routin INTRPT ASM contains a series of these interrup OPT REL IMPORT NMI1 NMI2 ASCT NMIVCT EQU FFFC CSCT PARAM RMB 1 DSCT PROMPT FCC OPTIONS PSCT 140 Since reloc with this sou is preferred Bring in th points of the global inform The 6800 mi the NMI throu absolute loca The main ro which is set is global inf The data se constants H with the FCC CHAPTER 9 RELOCATABLE ASSEMBLY 34 P 0000 86 00 MAIN LDAA 0 The main ro 35 no ORG statem 36 code above th 37 This feature 38 linker The 39 of initializa 40 only one or t 41
34. conditional assembly mask and the run time option that sets the conditional assembly mask for the current assembly and the chapter that explains conditional assembly also contains information about the CAMASK feature The CAMASK term is frequently used with conditional assembly because the dynamic CAMASK value allows a programmer to change the assembly without editing and changing the program s source file For example a program that is written with conditional assembly might use the CAMASK value as a code that indicates something about the environment of the program s target machine thus allowing a single source program to generate object programs for multiple target environments Although the CAMASK term is normally used with conditional assembly the assembler accepts the CAMASK term in any expression A programmer can therefore use the CAMASK term in any desired application For example a programmer might do something as simple as using the CAMASK value to specify the beginning memory address of the ROM area for which a program is being assembled or a programmer could use the CAMASK term to specify the length of a list that must be allocated in a program s RAM area The CAMASK term is a special token not a label so references to the CAMASK term don t appear in the concordance A programmer who wants the concordance to include 32 CHAPTER 4 EXPRESSIONS references to the conditional assembly mask can eq
35. continuing an instruction that contains EXPR3 multiple subfields in its operand field 16 CHAPTER 2 SOURCE SYNTAX LABEL2 FDB EXPR1 This example illustrates the fact that a EXPR2 user can include any amount of white space EXPR3 including none before a hyphen that EXPR4 appears between two fields or between two EXPRS subfields of the operand field UNUSUALLY_LONG_LABEL_STRING This example illustrates the use of line LDAA SAM continuation to attach a long label to an instruction MACRO This example shows how line continuation MOVE SOURCE can be useful when a person wants to define DESTINATION a macro that has multiple arguments The PSHA macro in the example is named MOVE and it LDAA SOURCE has two arguments SOURCE and DESTINATION STAA DESTINATION Each argument is expected to be a string PULA that represents a valid operand field for MEND the instruction with which it is used MOVE lt BASE X gt This example uses line continuation in a TABLE OFFSET statement that calls the MOVE macro that was defined in the previous example The simple examples that are shown above are not by any means intended to be comprehensive Indeed they barely begin to illustrate some of the flexibility that is possible with the assembler s continuation feature but they should be enough to give the reader a good idea of how to use statement continuation 2 8 Assembler Support for Text Formatters Programmers and other people w
36. cr gt SOURCE FILES OBJ gt main intrpt lt cr gt OUTPUT FILE OBJ MAP gt final lt cr gt START OF PSCT 2000 gt lt cr gt START OF CSCT gt 3000 lt cr gt No error in linkage Some notes on this transcript are appropriate First of all LK68 is a command to run the LK6800 linkage editor specifically set up to run on the VAX VMS at the Arizona State University ECS installation The actual command name of the linker is site specific The source files in this example are the two relocatable object files produced by as sembling MAIN ASM and INTRPT ASM with the M6800 assembler in relocatable mode The output file FINAL OBJ will then contain the S Record equivalent of the joined relocatable files The linker defaults the PSCT base address to 2000 though any other value could be chosen For the CSCT an input value is required and it was arbitrarily chosen as 3000 for this example The DSCT value is hard coded to 0100 by the LK6800 linker The No error in linkage message reports a successful linking The file ABMAIN ASM begins here 142 CHAPTER 9 RELOCATABLE ASSEMBLY BTEXT This is a comment stripped combined version of the MAIN and INTRPT modules using absolute object code generation Hopefully the resulting object files will be identical It may be useful to point out that the LK6800 linker will default the data sections to 100 and the PSCT to 2000 and therefore these specific values are introduced
37. default assembly mode is absolute 6 3 2 CAMASK The conditional assembly parameter may be set by the user to any 16 bit hexadecimal value with the CAmask hhhh option Only the low order 16 bits of an input hexadecimal value are retained by the assembler and overflow is not indicated to the user The con ditional assembly mask parameter is utilized in two ways The primary use is through the special assembler token CAMASK which is set to the 16 bit value of the run time CAmask parameter This token is then used explicitly in the expression field of the GOIF and GOUNLS conditional assembly statements The conditional assembly parameter is also used 96 CHAPTER 6 RUNNING THE ASSEMBLER implicitly in the less preferred mode of conditional assembly involving the GOIFZ and GOULZ assembler directives Refer to Chapter 7 for more information on conditional assembly The default value for this option is zero 6 3 3 COMMENT The COMment option tells the assembler to include comment lines in the output listing The NCOMment option suppresses comment lines from the assembly listing This option is useful when comments are not needed or an abbreviated output listing is desired In an academic or training environment students may be given listings of code to figure out and later given a commented version of the same listing Note that this option does not affect comments that appear on an instruction source line The default is the COMment op
38. delimited the final parameter with a comma The user can turn off this error message by specifying the NOTMERR as an OPT or with NTMerr as a run time option TO Two consecutive operators appear in an expression Two consecutive operators appear in an expression The assembler uses the second operator and discards the first operator A user might have written an expression like 3 for example TR An expression contains too many imported symbols Too many externally defined symbols Referenced error Six imported symbols may be referenced in an expression UL A local label that appears in the operand field is undefined Undefined Local label A local label that is used in the operand field is not defined anywhere in the program 7U0 Only addition and subtraction are allowed with relocatable values The assembler allows a user to add and subtract relocatable or imported values but it doesn t allow any other operations such as multiplication division or logical operations with relocatable or imported values The linker can handle only addition and subtraction so the assembler must not allow any other operations UP A macro call statement contains an unrecognized parameter name Unrecognized Parameter error A string that is parsed as a parameter name does not match any defined parameter name Perhaps the user misspelled a parameter name or the user attempted to list the substitution string as though the macro uses p
39. discarded MOD Modulo function for signed values MODUNS Modulo function for unsigned values AND Bitwise logical AND OR Bitwise logical inclusive OR XOR Bitwise logical exclusive OR ASL Arithmetic shift left 16 bit arithmetic shift zero fill ASR Arithmetic shift right 16 bit arithmetic shift sign fill LSL Logical shift left 16 bit logical shift zero fill LSR Logical shift right 16 bit logical shift zero fill ROL Rotate left 16 bit rotate with no carry bit involved ROR Rotate right 16 bit rotate with no carry bit involved GT If signed expression gt signed term then 1 otherwise 0 GE If signed expression gt signed term then 1 otherwise 0 EQ If signed expression signed term then 1 otherwise 0 NE If signed expression signed term then 1 otherwise 0 LE If signed expression lt signed term then 1 otherwise 0 LT If signed expression lt signed term then 1 otherwise 0 HI If unsigned expression gt unsigned term then 1 otherwise 0 HS If unsigned expression gt unsigned term then 1 otherwise 0 EQ If unsigned expression unsigned term then 1 otherwise 0 NE If unsigned expression unsigned term then 1 otherwise 0 LS If unsigned expression lt unsigned term then 1 otherwise 0 EO If unsigned expression lt unsigned term then 1 otherwise 0 MIN Select the smaller of two signed values MAX Sele
40. error respectively While the statement of the above rules is intended to convey the rigor of the expression evaluator with respect to relocatable expressions it perhaps is not so very clear to the user what is implied in the rules The following listing represents an example assembly file with nothing but symbol definitions and expressions It comes complete with both errors and examples of correct usage The examples are far from being exhaustive so the best policy as usual is to actually try a questionable statement with the assembler itself The file EXPRESS ASM begins here 134 CHAPTER 9 RELOCATABLE ASSEMBLY al a2 c1 c2 d1 d2 onea oneb onec oned onee two OPT ASCT EQU EQU CSCT RMB RMB DSCT RMB EQU IMPORT PSCT EQU EQU EQU RMB EQU LDX REL 10 20 di imi im2 di d2 ci c2 di c2 di imi 1 d1 Select relocatable assembly so that the directives are permitted Start an absolute section This effect can also be achieved with an ORG statement Start a common section Only allocation of variables is allowed in a CSCT so no initializations can be made Start a relocatable data section Import values from an external module Begin a program section where various expressions will be tried These two expressions have relocation counts of zero since two values with the same relocatable base are subtracted from each other However this express
41. forward references almost never appear in operand ex pressions for instructions that require the assembler to choose between direct and extended addressing The kinds of instructions that force the assembler to choose between direct and extended addressing usually address either variables or constants and good programmers don t make any forward references to variables or constants because they always define all of their variables and constants before they write any code that refers to those variables and constants Many compiler languages actually require programmers to define their variables and constants before they write any code so programmers who have used compiler languages naturally define their variables and constants first by force of habit Besides eliminating unwanted forward references the practice of defining variables and constants at the begin ning of a program enhances program readability by putting the foundation for the code at the front of the listing where it belongs A perverse programmer who insists on using direct addressing with an instruction that contains a forward reference can force the assembler to select direct addressing by including an exclamation point as the first character of the operand field The exclamation point tells the assembler that the user wants direct addressing regardless of any forward references For example STAB TOTAL uses direct addressing even if TOTAL is a forward reference 3 6
42. is defined by the TITLE statement The user can naturally change subtitles as often as she or he wishes so various sections of the same program can have different subtitles If a SBTTLE statement is included as the first statement of a program or at least before any statements that generate lines in the symbolic listing the subtitle from that SBTTLE statement will be effective on the first page of the pass 2 symbolic listing Otherwise no subtitle line will appear on the first page of the pass 2 symbolic listing Since a new subtitle signifies the beginning of a new section of a program the assembler automatically performs an EJECT function after it processes a SBTTLE statement This feature eliminates the need for the user to include an EJECT statement immediately after every SBTTLE statement Recall that an EJECT at the very beginning of a program has no effect since the listing is already at the top of a page then Also recall that two or more consecutive EJECTs perform only one EJECT function so the user who does include an EJECT immediately after a SBTTLE statement won t get a blank page Similarly a TITLE statement and a SBTTLE statement can be used together with no ill effects If the user specifies an empty subtitle string the assembler eliminates the subtitle line entirely instead of printing a blank line in place of the subtitle line Therefore the listing maintains a uniform spacing style throughout an entire program regardless o
43. is meant to provide the rudiments of assembly as directly and simply as possible Assuming that one is logged onto a VAX VMS system and that a source file has been readied for assembly then the command ASM executed at the VMS prompt level will provoke the following response from the M6800 assembler FILES SRC LST OBJ here you type sourcefilename then lt return gt OPTIONS here you type lt return gt Now relax but don t get too comfortable because the assembler is speedily processing the source file cataloging and displaying at the terminal any errors the message no errors in pass 2 can be a welcome positive feedback and generating the corresponding object and listing files When the prompt returns the assembly is finished and the new files have been saved in the current directory as a hybrid between the sourcefilename and the extensions obj and 1st The listing file can now be printed with the VMS command PRINT though it is a good idea to examine it with an editor first to verify its correctness and thereby conserve paper resources The object file is ready to be downloaded to a M6800 system a possibly involved task which is best taught by example though a subsequent section of this chapter attempts to describe in detail what is required 87 ASSEMBLER USER S GUIDE For the tyro staring at the VMS prompt with despair without even the source file assumed in the above discussion cons
44. last null string TEST Unspecified strings are assumed to be null String syntax is discussed in a previous section of this chapter 126 CHAPTER 8 MACRO ASSEMBLY 8 9 Nested Macro Calls Macros can call other macros If the assembler is expanding a macro and an expanded line contains a macro call for another macro the assembler will expand the second macro and resume expansion of the first macro when the second macro expansion is complete Calling another macro from within a macro is called nesting Macros may be nested to any level but macro recursion is not supported at any level A macro must not call itself and a macro must not call another macro that directly or indirectly calls the macro being defined If macro recursion is attempted the assembler reports an error and refuses to expand the illegal macro call 8 10 Hints for Using Macros The user must exercise caution in the use of local labels in macros Notice that none of the previous examples in this chapter uses local labels to avoid giving the impression that it is a good idea The main problem with using local labels within a macro definition is that it is easy to forget that the label is there For example the user might refer to 1B meaning the local label that is clearly visible in the source code and the assembler would reference the closer 1B if one were inside a macro For example consider the following code excerpt 1H TEST BNE 1B If TEST contains lo
45. level is initially zero and the LIST option increments the listing level while the NOLIST option decrements the listing level Therefore the user may bracket a program segment with a LIST NOLIST pair with no net change to the current listing level for program statements That is LIST and NOLIST options at earlier points in the program remain in effect after the assembler has processed a subordinate LIST NOLIST pair A similar option may be selected at run time but the run time option refers to the entire listing and cannot be alternately changed within the program the way the assembly time options can see Chapter refrunning assembler chapter for the corresponding run time option discussion The default for the LIST NOLIST option is LIST 5 40 14 OPT LLRANGE The LLRANGE n option tells the assembler that n is the maximum number of lines that may separate a local label from any reference to the label A local label can be used only within n where it is defined or a local label range error LR is generated The default value of the local label range parameter is 104 lines or approximately two pages but the user may change the unsigned value of this parameter as often as desired Notice that the value zero effectively outlaws the use of local labels since a local label can not be accessed from the same line that defines it see Chapter 2 for a discussion of local labels The local label range limit is a 16 bit unsigned value so the assem
46. lt expression gt MODUNS lt term gt is the unsigned remainder that results from dividing the unsigned value of the expression by the unsigned value of the term For example FFFF MODUNS 2 equals 1 and 2 MODUNS FFFF equals 2 Besides being used as binary operators the addition and subtraction operators can also be used as unary operators as shown in the examples above for signed division A programmer can include a unary operator at the beginning of an expression or parenthe sized subexpression and the unary operator applies to the first term of the expression or parenthesized subexpression Since an entire parenthesized subexpression is a term a pro grammer can use parentheses to make a unary operator apply to any desired part of an expression For example the expression 10 2 evaluates to 12 An asterisk has two possible meanings in an expression but the syntax of an ex pression unambiguously indicates whether an asterisk is being used as the multiplication operator or as the present location symbol For example the rather unlikely expression represents the value of the location counter multiplied by the value of the location counter 35 ASSEMBLER USER S GUIDE 4 2 2 Logical Operators The assembler recognizes three logical operators The AND logical operator specifies a 16 bit logical AND operation the OR logical operator specifies a 16 bit logical inclusive OR operation and the XOR logical operator
47. many other ways The GTE 65150 microprocessor which was designed primarily for use in telephone systems is essentially an extension of the 6502 architecture Please refer to the 6502 or 65150 microprocessor manual for detailed information regarding these processors ASSEMBLER USER S GUIDE By default the assembler runs as a 6800 assembler but a programmer can also tell the assembler to assemble for one of the other processors A conditionally assembled sister version of the M6800 family assembler is available to assemble programs for the M6809 microprocessor The M6809 assembler is completely compatible with the M6800 family assembler except for minor differences that are dictated by the unique characteristics of the M6809 assembly language and its source syntax The M6800 family assembler includes several features that make source preparation and program documentation easier for the user For example the assembler has a text mode that can be used for an introductory comment block at the beginning of each logical section of a program Also the assembler has a variable comment column that allows a programmer to type comments without the need for any special introductory comment characters starting at or beyond the specified comment column The assembler also has a useful feature that allows it to assemble source files that contain dot commands which are commonly used by text formatters such as runoff compose prose nroff and troff so
48. more parameters than the macro supports Since the programmer may deliberately use a comma to delimit the final string as a shortcut to avoid using lt gt syntax to specify a null string the assembler allows the user to suppress reporting of the too many parameters error The user can change between the report too many parameters mode and the suppress too many parameters mode as often as desired by selecting the TMERR or NOTMERR option respectively There are also run time TM and NTM options and the OPT source code options override the run time options The default for the TMERR NOTMERR option is TMERR 77 ASSEMBLER USER S GUIDE 5 41 ORG The ORG Origin statement sets the assembler s location counter equal to the value of the expression in the ORG statement s operand field For example ORG 10 sets the location counter equal to 10 16 decimal so the next word to be assembled will be assembled for location 10 If a label is present in an ORG statement s label field the label is equated to the value of the expression in the operand field However most good programmers don t include a label with an ORG statement because they can achieve more clarity by including the label with the statement that follows the ORG statement The assembler initializes the location counter to zero by default so the location counter starts at zero if a user neglects to include an ORG statement at the beginning of a program A go
49. of signed relational operators for making signed comparisons and the assembler also supports a complete set of unsigned relational opera tors for making unsigned comparisons The relational operators are primarily useful with conditional assembly but they can of course be used in any expression A relational operator tells the assembler to make the specified signed or unsigned com parison between the value of the expression on the left of the operator and the value of the term on the right of the operator The resulting expression value is 1 if the designated 38 CHAPTER 4 EXPRESSIONS relationship is true and the resulting expression value is zero if the specified relationship is false Table 4 5 summarizes the signed relational operators and Table 4 6 summarizes the unsigned relational operators Table 4 5 Signed Relational Operators Usage Meaning lt expression gt GT lt term gt lt expression gt greater than lt term gt lt expression gt GE lt term gt lt expression gt greater than or equal to lt term gt lt expression gt EQ lt term gt lt expression gt equal to lt term gt lt expression gt NE lt term gt lt expression gt not equal to lt term gt lt expression gt LE lt term gt lt expression gt less than or equal to lt term gt lt expression gt LT lt term gt lt expression gt less than lt term gt Table 4 6 Unsigned Relational Operators Usage Meaning
50. often as desired Zero parity is assumed at the start of each assembly pass 5 40 22 OPT RV8 NORV8 When NORV8 is in effect the assembler issues a warning message for any 8 bit expression that could possibly exceed the 8 bit range at link time because of relocation and or imports When RV8 is in effect the assembler doesn t issue such warnings but the user will still get error messages as appropriate at link time NORV8 is the default 5 40 23 OPT TABS The TABS n1 n2 n3 nk option tells the assembler to clear all previous tab stops and to set tabs in each specified column Tab stops may be specified in any order but an invalid tab error 1T is reported unless 0 lt nj lt 81 for each nj value specified Note that all tabs can be cleared by setting a tab stop just in column 1 since a tab stop in column 1 is meaningless anyway Unlike other OPT assembly options no other options may follow a tabs option on the same line because the syntax assumes that all fields after a TABS option are expressions that define tab stop columns Default tabs are set in columns 9 17 25 33 41 49 57 65 and 73 5 40 24 OPT TMERR NOTMERR By default the assembler reports an error if the macro call contains too many parameter substitution strings This error occurs when the macro call processor detects a comma following the last parameter substitution string of a positional parameter macro call The programmer may have tried to input
51. opcode mnemonic conveniently pro vides a logical extension of the instruction set When a microprocessor in the 6800 family subtracts one unsigned number the subtrahend or bottom number from another unsigned number the minuend or top number the carry borrow bit in the condition codes is set if and only if the unsigned value of the minuend or top number is lower than the unsigned value of the subtrahend or bottom number The BLO instruction which branches if the 53 ASSEMBLER USER S GUIDE carry borrow bit is set is therefore useful for comparing unsigned values and the function that the BLO instruction performs for unsigned numbers is analogous to the function that the BLT instruction performs for signed values While the BCS mnemonic is usually appropriate after a shift a rotation or an addition the BLO mnemonic is normally appropriate after a subtraction or a comparison The avail ability of the BLO mnemonic allows a programmer to use the mnemonic BLO or BCS that most clearly describes the function that the programmer intends to perform 5 7 BSCT The BSCT statement tells the assembler to start using the relocatable location counter that applies to the base section of memory so the BSCT statement is allowed only in a relocatable assembly not in an absolute assembly The primary purpose of the base sec tion is to allow programmers to define relocatable items that can be accessed with direct addressing The base section
52. ots GS e EE ta hk 29 4 1 6 Local Labels ooa siora d a rare eR hae a ee 29 4 1 7 Quoted ASCII Characters e 02 00 0080 200048 30 4 1 8 Present Location Symbol 02 0 0 0020004 30 4 1 9 RAD 40 Hash String 0 0 2 2 02 2004 31 4 1 10 Conditional Assembly Mask o o e e 32 4 1 11 Parenthesized Subexpression e e 33 4 2 QPerators de o A a A ee a ees 33 4 2 1 Arithmetic Operators 0 00000 eee eee 33 4 2 2 Logical Operators e 36 4 2 3 Shift and Rotate Operators 0 2 2 000004 36 4 2 4 Relational Operators 0 0 0 020002 eee eee 38 4 2 5 Selection Operators 2 0 0 0 a 40 4 2 0 Unary Operators sima mip aop a Ee ee 2 Se gh 40 did Parentheses lot is A A A De a ds 41 5 Assembler Directives 43 Bal ASG iene ea ate ee A Se 50 Di dy SARS CTICY ose gcse Goines Gente Speke a O 52 5 3 CASCUZ Aaa e hae hao ae Eh es BA ER RA 52 DAs TAS CW ty cake 8s A A e kA eh e e dls ok amp 53 Bibs BIS A io we hae A Re Ree ea Oh A eee 53 BiG BGO es ea fee ow fy elated Wi Fete ae th aaah anh ek ot Qala Uh BL eal pa Oe ah 53 BY BSCE 2 oly deal a E Beh og a A ale aed ee aod 54 DiS BSS ah 2th Bg iced ee Be hohe eg Pk ig BON ain WE ee Beng eg aaa Pot eg 54 CONTENTS 5 10 5 11 5 12 5 13 5 14 5 15 5 16 5 17 5 18 5 19 5 20 5 21 5 22 5 23 5 24 5 25 5 26 5 27 5 28 5 29 5 30 5 31 5 32 5 33 5 34 5 35 5 36 5 37
53. relocatable assemblies 5 14 EJECT The EJECT statement tells the assembler to advance the pass 2 symbolic listing if one is being produced to the top of a new page If the listing already happens to be at the top of a page when the assembler encounters an EJECT statement the assembler doesn t advance to the top of another page so the assembler never wastes paper by putting unnecessary blank pages into the middle of a program listing The assembler doesn t print the EJECT statement itself because that would clutter the listing The EJECT statement doesn t have an operand field of course and it must not have a label in its label field A user normally puts an EJECT statement at the beginning of every logical section of his or her program to improve readability 58 CHAPTER 5 ASSEMBLER DIRECTIVES 5 15 END The END statement terminates the assembly so it must be the last statement in every program The assembler reports an error if the END statement is missing and it simply ignores any lines that occur after the END statement in the source program The END statement doesn t have an operand field and its label field must be empty If an END statement appears in an included file see INCLUDE it terminates the entire assembly just as it would if it appeared in the main source file 5 16 ENTRY The ENTRY pseudo instruction allows the user to specify a starting point in the user program If a program contains an ENTRY statem
54. substituted The assembler provides unique label field strings for the first 65536 calls The example in the section on formal parameters illustrated the use of automatic label field generation The first macro call generated the label LOOP_A and subsequent calls will generate LOOP_B LOOP_C etc The user must ensure that the generated labels do not cause multiply defined errors If two macros for example have the label LOOP in their models then both macros will generate label LOOPA in the first expansion Also if a macro uses the label LOOP then the user should not define any labels that concatenate LOOP with any combination of four letters in the range A through P 8 6 MEND The MEND pseudo op which marks the end of a macro definition is explained in Chap ter 5 Note that the MEND statement must not be labeled 8 7 MEXIT The MEXIT pseudo op which signals the end of macro expansion but not the end of the macro definition is explained in Chapter 5 MEXIT is used in conjunction with conditional assembly see Chapter 7 The user must exercise caution however to ensure that the MEXIT statement does not defeat the conditional assembly In general MEXIT avoids making the assembler expand lines that will be skipped anyway The following example illustrates the use of conditional assembly and MEXIT Suppose that we commonly test a loop condition at the bottom and branch conditionally to con tinue the loop The range of a branc
55. the assembler inserts zero bits at the left end of the 16 bit expression value and it discards bits as they pass out of bit zero to the right Any shift count greater than 15 therefore produces an expression value of zero 37 ASSEMBLER USER S GUIDE The ROL operator rotates the 16 bit expression value to the left by the number of bit positions specified by the term following the operator When a bit rotates out of bit 15 to the left that same bit rotates into bit zero at the same time The 16 bit expression value is therefore effectively circular with bit 15 logically adjacent to bit zero Any rotate count greater than 15 is effectively the same as COUNT MOD 16 because each 16 bit rotation of the 16 bit expression value returns the expression to its original value Although shift counts are definitely unsigned a user can conveniently think of rotate counts as being either signed or unsigned For example VALUE ROL 5 yields the same result as a rotate to the right by 5 The ROR operator specifies a rotation to the right by the number of bit positions specified by the term that follows the operator When a bit rotates out of bit zero to the right that same bit rotates into bit 15 at the same time so the 16 bit expression value is effectively circular with bit zero logically adjacent to bit 15 Any rotate count greater than 15 is effectively the same as COUNT MOD 16 because every 16 bit rotation of the 16 bit expression value leaves th
56. the assembler is likely to find a 3H label in some subsequent routine perhaps much later in the program and then the 3F reference is unfortunately satisfied by the wrong label To guard against this kind of devastating error the assembler issues a warning message whenever a reference to a local label is too far away from the line that defines the local label The assembler considers a reference to a local label to be too far away from the local label s definition when the reference is more than 104 lines about two full pages away from the definition The number 104 is merely the assembler s standard default limit and a programmer can redefine the local label separation limit as often as desired in a single program by using the OPT LLRANGE expression directive which is described in detail in another section The distance between two source lines is defined to be the absolute value of the difference between their line numbers We can summarize the assembler s syntax rules for the definition of a local label as follows If the first column of a statement contains a decimal digit the assembler assumes that the programmer is using the label field to define a local label The format for the definition of a local label requires an H or an h in column two and column three must contain white space or a carriage return to terminate the label field If columns two and three aren t formatted correctly when column one contains a decimal dig
57. the final obj 42 43 P 0002 CE 0000 LDX PROMPT This repres 44 displaying th 45 Actually thi 46 of i o routin 47 48 P 0005 81 01 CMPA 1 This compar 2 25 Oct 1985 19 08 M6800ASM 1 1 MAIN ASM 49 P 0007 26 00 BNE 1F a routine whi 50 and left it 51 the AR 52 53 P 0009 CE 0000 1H LDX NM12 Use the rel 54 P 000C FF FFFC STX NMIVCT second NMI in 55 selected by t 56 an absolute a 57 58 P OOOF F1 0000 WAIT CMPB PARAM Wait in thi 59 P 0012 26 FB BNE WAIT handler has s 60 61 END No errors in pass 2 7 non local labels in the symbol table 1 local label in the symbol table 3 25 Oct 1985 19 08 M6800ASM 1 1 MAIN ASM Concordance M Multiply Defined Label U Unused Label U I Imported E Exported B BSCT C CSCT D DSCT P PSCT Blank Absolut Line Value Symbol Reference List 141 ASSEMBLER USER S GUIDE 34 U 0000 P MAIN 19 FFFC NMIVCT 54 14 IU 0000 NMI1 14 I 0000 NMI2 53 24 0000 C PARAM 58 29 0000 D PROMPT 43 58 OOOF P WAIT 59 4 25 Oct 1985 19 08 M6800ASM 1 1 MAIN ASM Unused Labels M Multiply Defined Label I Imported E Exported B BSCT C CSCT D DSCT P PSCT Blank Absolut Line Value Symbol 34 0000 P MAIN 14 I 0000 NMI1 The monitor session with LK6800 begins here Prompts from the linker are given in CAPITAL letters and the user s responses are given in lower case letters The notation lt cr gt represents a carriage return LK68 lt
58. to some available block of memory 9 2 Invoking Relocatable Assembly As with most features in the M6800 assembler the option to select relocatable object code generation can either be done at run time or with an assembly time option included in the source code Of course if the user does not specify a file or device for the object code none will be generated though errors specific to relocatable assembly will still be reported should they occur At run time relocatable assembly may be selected by specifying the RELOCATABLE option The default option ABSOLUTE for absolute code generation is otherwise in effect Refer 130 CHAPTER 9 RELOCATABLE ASSEMBLY to Chapter 6 for more information about using the RELOCATABLE and ABSOLUTE run time options At assembly time relocatable assembly may be specified with the OPT REL pseudo op Its absolute mode counterpart is OPT ABS Chapter 5 describes the use of assembly time options in detail Note that assembly time source options have precedence over the run time variants It is also important to note that after the first significant statement i e one that defines a label or generates object code is processed a request to change object generation mode is flagged as an AR error and the current mode is left unchanged 9 3 Features of M6800 Relocatable Assembly Relocation is implemented by defining a number of abstract partitions of the micro processor address space called se
59. type a source line with the assembler accepted syntax of an END statement a tab the word END and a carriage return will work Other methods of quitting this mode suffer from undesirable side effects Stopping by entering a lt CONTROL Z gt on a new line will always result in a missing end statement error and a lt CONTROL C gt exit will also abort the entire assembly process though the file TEMP ASM will still reside on the disk Users who terminate this mode with a lt CONTROL C gt are employing the assembler as a very crude editor and should in preference consider the COPY SYS INPUT FILENAME technique outlined for the inexperienced user in Section 6 1 A different situation occurs when the special symbol TT or its equivalent is specified in the input string as the listing or object files In this case the contents of the created file marked by the TT symbol will be displayed on the user s terminal While this may appear to be a useful technique for quickly examining the results of a short trial assembly the user should be cautioned about several side effects of this mode of operation First if the file is displayed at the terminal it will not be saved as a disk file in the current directory Secondly the lines of the displayed file will be interleaved with the messages that the assembler normally generates e g errors in the source file If both the listing and the object files are sent to the terminal the result will be a
60. which it never modifies and as a function of its contents as well as some run time options produces two files The list or listing file contains a formatted representation of the source file more suitable for communicating the ideas of the source program The list file also records any errors in the immediate proximity of the source line that caused them The object file contains the ASCII hex packed machine code in S record format that corresponds to the semantics of the assembly source lines The object file is suitable for downloading into a M6800 microcomputer for actual execution For software to access files they must have unique names or identifiers In VMS a complete specification of a file can consist of up to six parts A full path name then is described by a node a device a directory with possible sublevels a filename written without a space to signify a specific part of what in common parlance is a file name a file extension and a version number Suffice it to say that the M6800 assembler supports 88 CHAPTER 6 RUNNING THE ASSEMBLER full file name specifications however most users will never need to be concerned with more than the filename and the file extension parts of an identifier For the file concept to be useful the user must be able to create and rearrange them Most files like the assembly source files will be created with an editor like EDT on VMS Though the initial contact with so many options and comman
61. with ORG statements The value for the CSCT was selected as 3000 at LK6800 run time as described above ETEXT NMIVCT EQU FFFC ORG 100 PROMPT FCC OPTIONS TRANS1 FCC qu TRANS2 FCC gr ORG 0000 SHORT RMB 1 AND RMB 2 FAST RMB 1 ORG 3000 PARAM RMB 1 GLOBAL EQU PARAM ORG 2000 MAIN LDAA 0 LDX PROMPT CMPA 1 BNE 1F 1H LDX NMI2 STX NMIVCT WAIT CMPB PARAM BNE WAIT NMT1 LDAA FAST STAB GLOBAL RTI NMI2 LDAA SHORT STAB GLOBAL 143 ASSEMBLER USER S GUIDE RTI END Object file from linkage of MAIN and INTRPT to form FINAL OBJ 510B01004F5054494F4E533F88 11320008600CE010081012600CE201 AFFFFFCF1DC 51072010300026FB77 105010841426E S10F20149603F730003B9600F730003BC9 S9030000FC Object file from ABMAIN ASM 510D01004F5054494F4E533F414203 511320008600CE010081012600CE201 AFFFFFCF1DC s1132010300026FB9603F730003B9600F730003B78 S9030000FC The claim is made here that both of these object record files are in fact identical with respect to the information passed to some utility loader for the M6800 microprocessor Of course since the sizes of each record are not the same across the two versions checksum values will also be different For the users who are not convinced the final part of this chapter explains the fields of a typical record with sufficient clarity to enable the doubters to verify the claim for themselves 9 6 A Few Words About the S Record Format For the interested reader a complete des
62. 22 CHAPTER 3 ADDRESSING MODES FOR EXECUTABLE INSTRUCTIONS indexed addressing For example a user might write SUBA FIELD X or ANDB 5 X to specify indexed addressing 3 4 Extended Addressing Extended addressing provides a 16 bit address field so an instruction that uses extended addressing can conveniently access an operand anywhere in the microprocessor s entire memory space The operand field for an instruction with extended addressing contains an expression that specifies the memory address of the desired operand For example INC COUNT uses extended addressing to increment the contents of a memory location named COUNT and STX 1234 uses extended addressing to store the 16 bit value of the XR into memory locations 1234 and 1235 with the most significant byte of the XR value being stored into location 1234 and with the least significant byte of the XR value being stored into location 1235 The assembler always recognizes a leading to indicate a hexadecimal number 3 5 Direct Addressing A programmer can use direct addressing with some instructions to access operands that reside in page zero of memory where page zero of memory is defined to be the 256 bytes from location 0000 through location 00FF Since the top 8 bits of a direct address are always zero by definition direct addressing requires only an 8 bit field to specify the address of an operand An instruction that uses direct addressing is theref
63. 5 38 5 39 BTEX T vet ao ate sarc ees a et ake Seagate cee aed pp aaa douse 55 CHKSUM va akate a ea Eea pa i au PA EN a iiA 56 ASA A e gees Dar RE 57 CSOT wach be So ed ead eco aceesl ee oa Gash Anke wuld 58 SOAs 4 ih e No A ec at ca il Hg at 58 BIRGI mro ra e ane es 58 PND anae Ste ely gay Smt ia be io da oe re 59 A ene Ses Maia eng As oie olin Sh te tao ter a etal 59 VO Svar SR saan oth Ses E AS Led een ah Basen aot AS pe oe a 59 EQUREG 6809 id Ys bape Anker yest amp Being Wiebe Ade lind iia Be 59 ETE E ee 59 PORT peri Os A th A Fhe alee 60 OB Nh ae ete tat eu ia fede oe ee ae Cite at ee iy oS oe 60 OBS ns Seat eel at dl eed Sarin Oe set ge nal ae tle OH oe 61 O eat a ea eee nee a ea es A E A 61 O hoa ane dee eh aaa E elie are 61 POB a e ts setts eh Bical aoe ica eee a A 61 COIR 04 sis le trap ty eee See he ate ea eee ai Seg a 62 OZ oy a aia de ghana danse rd aR 62 COTO 5 3 60 A Geechee etn ee 63 GOW So sed aS a ay eels nee oS ate hod bud 64 UN a a tel etl tht te anal e o atte 64 INPORT A aa 65 DIVE ROIS a ns da pe ra 66 INCLUDE 2s a het Ad e wl Siete hee ee 66 END os ware Pe Hage ee os Hes BE Sa a eee Bees ee he 66 MACRO 5 dices to Aw tate Bele aid ties he 66 MENDE san as ts Se chee a a hss tae aa ai Attn cat hee ae 66 MENET Mots seins argh Bake Bt a Anes Medics he aah o oe wees 67 MON A a tet ek e et leu MON eo 67 NAMES A Ra 67 ili ASSEMBLER USER S GUIDE 5 40 5 41 5 42 5 43 5 44 5 45 5 46 OPT eiea I
64. DB statements The assembler must evaluate these expression values completely and can t defer the final evaluation to the linker RV A relocatable or imported value is not allowed for an 8 bit field Relocatable Value error An expression that is used for an 8 bit field contains a symbol that is defined in another module or the expression has a relocatable program section value that is not ASCT or BSCT The assembler doesn t allow such values for 8 bit fields because relocation could cause such a value to exceed 8 bits at link time SE A synchronization error between passes has occurred for this label Synchronization Error The value that the assembler computed during pass 2 for the label in the label field doesn t match the value if any that was recorded in the symbol table for the label during pass 1 Synchronization errors are usually caused by forward references in expressions that must be evaluated during pass 1 For example a forward reference in an EQU statement or an ORG statement or an RMB statement would create problems The root cause of a synchronization error is frequently at some statement before the statement at which the synchronization error occurs For example a forward reference in an ORG or an RMB statement could cause synchronization errors for subsequently defined labels Syn chronization errors can also be caused by errors that are due to forward references in the definitions of labels that are used for controll
65. DERR option is EOERR 5 40 8 OPT FORMAT NOFORMAT The FORMAT option directs the assembler to automatically format the user s assembly language listing into neatly aligned columns for the label opcode operand and comment fields The NOFORMAT option on the other hand leaves the formatting to the user through the use of space and tab characters The no format mode is assumed by default at the beginning of each pass of the assembly but the user can change between no format and format modes as often as may be desired Format mode is of little use in view of the fact that users can conveniently use horizontal tabs but the format mode is retained as 72 CHAPTER 5 ASSEMBLER DIRECTIVES an optional feature of the assembler to provide compatibility with earlier versions of the assembler that existed before the advent of modern disk based text editors The default for this option is NOFORMAT 5 40 9 OPT IGNORE NOIGNORE By default the assembler ignores or suppresses from the assembly listing any source line that starts with a period Often these dot lines are the commands for a text formatter that the user will employ to improve the quality and readability of the listing However when a program is still under development the user may not want to squander the computer resources to preprocess the source with the formatter when the listing is only a preliminary version The IGNORE option permits a reasonably clear listing since it causes th
66. E SBTTLE 1 1 Level 2 SBTTLE 1 2 Level 2 SBTTLE SBTTLE 1 2 1 Level 3 SBTTLE 1 2 2 Level 3 SBTTLE 1 2 3 Level 3 SBTTLE SBTTLE 1 3 Level 2 SBTTLE 1 4 Level 2 SBTTLE SBTTLE 2 Level 1 SBTTLE SBTTLE 2 1 Level 2 SBTTLE 2 2 Level 2 SBTTLE SBTTLE 3 Level 1 The leading period in each subtitle string starts the string but is not printed in the listing so the lines in the resulting table of contents are properly indented to show the logical organization of the program You can also use the leading period feature to adjust the horizontal position of the title that is printed at the top of each page of the listing A circumflex 7 as the first character of a subtitle string tells the assembler that you want to start a new page in the table of contents with the current subtitle so you can now control pagination in the table of contents The leading circumflex character itself will not be printed as part of the subtitle string but the remaining part of the subtitle string will 81 ASSEMBLER USER S GUIDE be at the top of a new page in the table of contents If for some reason you want a literal circumflex as the first character of a subtitle string you can simply start the subtitle string with a leading period which will be ignored and then type the circumflex after the period In this case the circumflex will be a literal part of the subtitle string 5 52 SET The SET statement is similar to the EQ
67. Growing The set of run time options is being extended to more completely compliment those selectable via the assembly OPT statement At this time a more comprehensive interactive help facility is also being considered 110 Chapter 7 Conditional Assembly The M6800 assembler provides the user with conditional assembly capabilities so that a single source file can produce multiple versions of a program Conditional assembly allows the user to include or omit blocks of source lines automatically and dynamically based on the logical verity of an arbitrary expression that can reference a 16 bit conditional assembly pa rameter The assembler supports two separate protocols for conditional assembly Though the most recent version is preferred this chapter documents both Conditional assembly greatly eases program maintenance because it allows people to maintain only one source file of a program even if the program exists in several different object versions Programmers who must keep several different versions of a program up to date frequently forget to edit the latest changes into every source version of the program but conditional assembly eliminates this possibility of errors and frustration by making it unnecessary to keep more than one source element for any given program Beginning programmers seldom use conditional assembly except as an exercise but systems programmers use it extensively Software houses commonly employ conditional
68. IMPORT statement tells the assembler to import the specified label s from other program s to make them available for use in the program that is being assembled The IMPORT statement is valid only in a relocatable assembly and the label s that it specifies must be defined as exported labels in other module s that will eventually be linked with the program that is being assembled The format for the IMPORT statement is as follows IMPORT label_1 label_2 label_n The IMPORT statement requires at least one label as an operand and it accepts any number of labels beyond the first one The IMPORT statement s label field must be empty The B size qualifier can follow a label in the import list for the IMPORT statement The following example illustrates the syntax IMPORT LABEL1 B LABEL2 LABEL3 B LABEL4 LABEL5 LABEL6 B The B size qualifier tells the assembler that the imported label is a byte sized label that is not relocatable beyond the range of an 8 bit address so the assembler can choose to use direct addressing for an access to the label even though the label s actual value is unknown at assembly time For more information regarding the IMPORT statement please refer to the chapter that describes relocatable assemblies and also refer to the description of the related EXPORT statement 65 ASSEMBLER USER S GUIDE 5 32 IMPRTB The IMPRTB statement is just like the IMPORT statement but all of the labels in the imp
69. LOG2 VALUE is invalid but NOT LOG2 VALUE is correct and also more readable As you would expect the unary operators take precedence over the binary operators For example the expression 1 LOG2 VALUE is equivalent to 1 LOG2 VALUE and to 1 LOG2 VALUE The last example is probably the best for clarity The unary positive sign i e emphasizes that the following term is positive and the unary negative sign i e negates the following term Note that the positive sign is not necessary because all terms without a positive or negative sign are positive For example 3 4 12 and 3 4 12 The BITS operator returns a count of the one bits in its term For example BITS EXPRESSION returns a count of the number of one bits in EXPRESSION A per son might use the BITS operator to compute the number of registers in a register mask The logical NOT operator converts any true value into a false value and it converts a false value into a true value The assembler considers any nonzero value to be true and it considers zero to be false The NOT operator produces the value zero for a false output and it produces the value one for a true output The logical not operator is intended for use 40 CHAPTER 4 EXPRESSIONS in logical expressions such as MPU EQ 68XX and NOT D2KIT The NOT operator does not produce a one s complement but the assembler provides a COM operator for that function The COM o
70. LZ statement allows a user to perform conditional assembly by skipping a conditional section of a source program or including the conditional section of the source program as part of the program during the assembly process Also like the GOIFZ statement the GOULZ statement is no longer recommended for use and the assembler still recognizes the GOULZ statement only for purposes of maintaining compatibility with old programs Anything that can be done with the GOULZ statement can be done more clearly with the GOIF statement so programmers should avoid using the GOULZ statement in any new programs The format for the GOULZ statement is as follows GOULZ conditional_expression target If the conditional expression ANDed with the assembler s CAMASK parameter is zero the assembler simply continues the assembly with the next line and the GOULZ statement essentially has no effect If on the other hand the conditional expression ANDed with the CAMASK parameter is not zero the assembler goes forward to the target conditional label thus skipping the assembly of all intervening lines The CAMASK parameter is a 16 bit value that a programmer can dynamically specify as a run time parameter when the programmer invokes the assembler The CAMASK parameter is discussed in the section that explains the assembler s run time options The assembler must evaluate the conditional expression in a GOULZ statement during the first pass of the assembly so forward
71. M6800 Family Assembler User s Manual David C Pheanis Western Microsystems Third Edition January 2003 I have used reasonable care to make the assembler and this manual as accurate and as usable as possible but I have probably overlooked a few mistakes In the interest of quality therefore I am offering a reward of 1 00 in U S funds to the first finder of each error whether it is technical typographical grammatical or otherwise Any student who uses the assembler and this manual in my class may elect to receive 20 points of class credit in lieu of a cash reward for reporting an error I hope that this offer will encourage people to report my mistakes so I can correct them I shall appreciate receiving positive suggestions for improving this manual or the as sembler in any way Many of the features that are already implemented in the assembler evolved from discussions with users and future suggestions for improvements will be more than welcome Disclaimer The reward that I mentioned above applies only to the first four chapters The rest of the manual is in preliminary form and I added it rather hastily in an effort to make the manual as complete as possible as soon as possible DCP Copyright 1985 2003 by David C Pheanis All rights reserved No part of this publication may be reproduced stored in any retrieval system photocopied recorded or otherwise duplicated or transmitted in any form or by any means ele
72. N eA OR ps ER 67 5 40 1 OPT ABS REL quita dre dow late eh Late de was 70 5 40 2 OPT CALIST CANOLIST otci0h 48 a aa 70 5 40 3 OPT CMTCOL ouaou a aaa 71 5 40 4 OPT DB8 DB10 DB16 ear e ll 71 5 40 5 OPT DEBUG oaaae aaa a a aa 71 5 40 6 OPT DIRECT NODIRECT lt a 72 5 40 7 OPT EOERR NOEOERR ooo ouaaa aa 72 5 40 8 OPT FORMAT NOFORMAT ouaaa aaa aaa 72 5 40 9 OPT IGNORE NOIGNORE noaua aaa a 73 5 40 10 OPT INCLIST INCNOLIST oo ouou aaa 73 5 40 11 OPT LBLLIMIT A 73 5 40 12 OPT LINENUMBER ote oahu cect oe a aa 73 5 40 13 OPT LIST NOLIST ooo a 74 5 40 14OPT LLRANGE ata a 74 5 40 15 OPT LONG SHORT Messi 75 5 40 16 OPT LPAGE 4 sive apes a 75 5 40 17 OPT MPERR NOMPERR ooa aa aa a 75 5 40 18 OPT MPU ota a a 76 5 40 190PT MSHOW NOMSHOW 0 00000 76 5 40 20 OP T PAGENUM usura wie ehh ee oA 76 FAQ OOPS PARITY x e ll 76 5 40 22 OPT RV8 NORV8 baaa odds 77 SAD DORE A A Se a ea 77 5 40 240PT TMERR NOTMERR co coo 0 77 ORG ienaa ee ere Be Bates Be hab Tees AE AE SEB aes 78 PAC Eiee O E aa e dese ange Bode 78 II EEE 78 A cove Oa eHCA Saas fen Me ee ee 79 PCR fini RN Tk ute he 79 O A O See Se NS 79 lv CONTENTS DAC ROBU tere sti g th sae ge ee ae eh ar Geet gh Eee POE ME EY hake De Ae Gog 79 DAS RDB re A athe bos eek es ALR A Boh eh ee ee da ek 79 DAD RIM Bint Bote tah atk e e gM O td hike atk be a sek We ote 80 DO SB LTD i bein ae a ata A A ee Re a ee ee 80 BIOL SB BELE cute eaten p de te dela w la
73. S commands such as PRINT and TYPE The reader is also expected to have some previous experience with assembly language and with the use of a normal two pass assembler The M6800 family assembler is a two pass assembler so it has many of the characteristics that are inherent in any two pass assembler The M6800 family assembler features a run time option that allows it to assemble pro grams for any of the following microprocessors M6800 M6801 M6802 M6803 M6808 and H6301 The first five microprocessors in the family were all developed by Motorola and they are manufactured both by Motorola and by other manufacturers who serve as second sources The H6301 is a CMOS microprocessor that was developed by Hitachi as an extension of the M6801 architecture All six of these microprocessors share the basic architectural features of the M6800 microprocessor but the later processors in the family include new instructions and or other features that aren t available with the M6800 which was the first processor in the family Please refer to the appropriate microprocessor man ual for detailed information regarding the architecture and instruction set of the particular microprocessor that you want to use The M6800 family assembler also features a run time option that allows it to assemble programs for a 6502 microprocessor or for a GTE 65150 microprocessor The 6502 is similar to the M6800 microprocessor in many respects but it is considerably different in
74. SSEMBLER DIRECTIVES 5 44 PSCT The PSCT statement tells the assembler to start using the relocatable location counter that applies to the program section of memory so the PSCT statement is allowed only in a relocatable assembly not in an absolute assembly Like the common section and the data section the program section can be relocated to any memory address at link time but the program section normally contains executable code and constants while the common section and the data section normally contain variables For more information regarding the PSCT statement please refer to Chapter 9 5 45 RCB The repeat constant byte RCB directive is a useful extension of the FCB statement The count expression in the operand field defines the number of times that the assembler will perform the FCB function on the remaining expression values The result is as if the corresponding FCB statement had appeared count times The syntax of the statement is shown below and further information on the acceptable values of the expressions is available under the FCB entry in this chapter label RCB count valuel comment field 5 46 RCBS The RCBS directive is like RCB but the RCBS statement specifies that the assembler should interpret and validate the expression value as a signed number 5 47 RCBU The RCBU directive is like RCB but the RCBU statement specifies that the assembler should interpret and validate the expression value as an unsigned num
75. SSEMBLER USER S GUIDE Since the programmer may determine that nothing of value was lost in the expanded line only the comment field was truncated the assembler allows the user to suppress reporting of the expansion overflow error The user should not routinely suppress the error it is best to suppress the error to get a clean listing only after ensuring that nothing important is lost The EOerr option tells the assembler to report any expansion overflow errors and the NEOerr option suppresses reporting expansion overflow There are also assembler directive OPT switches EOERR and NOEOERR which override the run time options The default is EOerr 6 3 7 ERROR When the Error option is selected error messages are sent to the terminal as well as to the listing The NError option on the other hand suppresses errors from being sent to the terminal If the terminal is the output listing device this option is effectively ignored since errors are automatically reported to the assembly listing see the REPORT run time option for suppressing errors in the assembly listing Error is the default for this option 6 3 8 FFEED The FFeed option directs the assembler to issue form feeds rather than the slower carriage return lt CR gt and line feed lt LF gt combination The NFFeed option directs the assembler to replace each form feed with an appropriate number of lt CR gt lt LF gt combina tions The selection of the FFeed op
76. Self Relative Addressing A branch class instruction in the M6800 family uses self relative addressing to spec ify its destination address Instead of containing an actual 16 bit destination address a branch class instruction contains a signed 8 bit self relative offset value that can be used to determine the desired destination address The microprocessor computes a self relative destination address by sign extending the signed 8 bit offset to 16 bits and then adding the sign extended offset value to the 16 bit value of the PR The PR which is also known as the 24 CHAPTER 3 ADDRESSING MODES FOR EXECUTABLE INSTRUCTIONS PC or the program counter is a register that always points to the next sequential byte in the instruction stream When the PR is being used to compute the self relative destination of a branch class instruction it points to the byte that immediately follows the second byte of the branch class instruction in memory Obviously the signed 8 bit offset that appears in the second byte of a branch class instruction must equal the difference between the desired branch destination and the PR value The value of the offset is therefore DEST PR where DEST represents the desired destination address and PR represents the relevant value of the PR Since the relevant PR value is the memory address of the byte that immediately follows the second byte of the branch class instruction in memory the value of the offset can be written as
77. The MACRO instruction introduces a macro definition Everything from the MACRO pseudo op to the next MEND pseudo op is interpreted as part of the macro definition Since the MACRO instruction is not associated with the location counter as it does not generate object code the MACRO instruction must not be labeled Refer to Chapter 8 for information on the use of macros 5 36 MEND The MEND instruction ends a macro definition When the assembler encounters a MACRO pseudo op it processes everything until the next MEND instruction as lines of the macro 66 CHAPTER 5 ASSEMBLER DIRECTIVES definition The MEND statement must not be labeled and it is only valid in a macro definition If a MEND instruction is labeled the macro processor will not terminate the definition If a MEND occurs outside of a macro definition the assembler will report a non fatal error Refer to Chapter 8 for information on the use of macros 5 37 MEXIT The MEXIT macro exit instruction tells the macro expansion processor to exit a macro definition When a macro call occurs program lines are generated by expansion of the macro definition model lines Normally program lines are retrieved from the source file Expansion ends when a MEXIT or a MEND instruction is encountered Unlike the MEND in struction MEXIT instructions do not end a macro definition only the expansion of that definition MEXIT is sometimes used in conjunction with conditional assembly to preven
78. U statement but the SET statement defines a re definable label If you define a label with the SET statement you can re define that label as often as desired with other SET statements An attempt to re define a SET defined label as a normal i e not re definable label is not allowed A SET defined label is available for use from the time you define it until you re define it to another value A forward reference to a SET defined label is therefore invalid The assembler provides a separate concordance entry for each definition of a SET defined label with multiple definitions and the entries in each definition s reference list indicate references to that definition of the re definable label The assembler warns you if you try to export a re definable label 5 53 SETREG 6809 The SETREG statement is analogous to the SET statement but the SETREG statement takes a register list instead of an expression as its operand 5 54 SK1 The SK1 SKip 1 byte instruction tells the assembler to generate an instruction that will skip the next one byte instruction during execution of the assembled user s program By employing this instruction as opposed to a branch instruction the programmer will save one byte of memory For example the following instruction sequences execute the same way but the code on the right uses one less byte of memory BRA 1F SK1 SHIFT LSRA SHIFT LSRA 1H STAA 0 X STAA 0 X The generated opcode value 85 for a SK1 in
79. a B cannot appear in a decimal number 77BT The BTEXT or ETEXT statement is redundant BTEXT ETEXT error A BTEXT or ETEXT statement appears redundantly so a BTEXT or ETEXT statement is probably missing somewhere 148 CHAPTER 10 ASSEMBLY ERRORS CI Object values are not permitted in CSECT memory CSCT Initialization error Any directive or instruction that generates an object value for the blank common area is illegal Blank common is shared by all modules as a read write area of memory so no module is allowed to contain object values that are assigned to blank common If two or more modules contained object values for blank common the object values from the second module would overwrite the object values from the first module To prevent this possibility the assembler prohibits any module from generating object values for blank common 77C0 The first character of this source line is invalid COmment error The first character in the source line is not valid so the line is treated as a comment line Valid first characters are a tab blank a letter for a symbol a digit for a local label an asterisk for a comment line or a dollar sign for a conditional assembly label A carriage return may also be the first character in a line in which case the line is treated as a blank comment line DF Divide fault or division by zero occurs in an expression Divide Fault error Divide fault occurs when a user attempt
80. abs as necessary to advance from the operand field to the comment field with correct field alignment 2 6 Comment Field The fourth and last field of a typical statement is the comment field The comment field can contain anything and the assembler essentially ignores the comment field except for including it in the program s symbolic listing The user should note however that string substitution can occur in the comment field during macro expansion a feature that is more fully explained in the chapter on macros Although the assembler doesn t require or examine the comment field complete com ments are critically important to the quality of any program Most users soon become aware of the value of good documentation so they make extensive use of the comment field 13 ASSEMBLER USER S GUIDE Programmers normally group a few instructions together as a logical step and then they write a paragraph of commentary to explain the purpose and function of the logical step rather than explaining individual instructions A single paragraph of comments therefore continues across the comment fields of several consecutive lines The assembler considers a blank or empty line to be a comment line so a programmer can conveniently use a blank line to separate one logical group of instructions from the next logical group of instructions The following example is only a brief fragment of a program but it illustrates the method for grouping instru
81. al Numbers A prefix character followed by series of one or more hextets 0 9 A F or a f is a hexadecimal number commonly known as a hex number Hex numbers like numbers in 28 CHAPTER 4 EXPRESSIONS the other bases can contain as many as 16 significant bits so a hex value can contain as many as four significant hextets The smallest acceptable hexadecimal number is 0 and the largest acceptable hex number is FFFF 4 1 5 Symbolic Labels Like any other term a symbolic label has a 16 bit value If the assembler is performing an absolute assembly every label that appears as a term in an expression must be defined somewhere in the program For a relocatable assembly the assembler also accepts a label as a term in an expression if 1 the label is declared as a label that is defined in another module and 2 the context in which the expression is used permits a value from another module Some assembler directives have operand fields or subfields that must be evaluated ac curately during pass 1 so the expressions that appear in these particular fields or subfields must not contain any forward references For example ORG EQU BSS and RMB statements prohibit forward references If a label that appears as a term in an expression hasn t already been defined to have a particular value when the assembler encounters the expression that contains the label the label is being used as a forward reference Therefore any label that is ini
82. an equal sign and of course a value form part of the required syntax The other class of run time options is a group of Boolean flags A user can select a Boolean option by simply typing the option name or at least the capitalized significant portion of the option name However a user can turn off or deselect a Boolean option by prefixing the option name with the letter N or n to signify the negated action In any case if an option is invalid or assigned an improper value the assembler outputs an appropriate error message and then solicits new input with the OPTIONS prompt again All options that were correctly entered before the error was noted will remain in effect as specified though the rest of the input line will have been ignored Effective use of the run time options may be made more clear by the following sug gestions If a program is always run with the same options and they are not the normal default options it would be best if the user included these options in his or her source code On the other hand some options allow the user to format the output listing and to obtain different listings of the same source program by specifying the options prior to the program s assembly and these are best specified at run time The available run time options are tabled on the next few pages in alphabetical order for the user s convenience Subsequent to the table appear more complete descriptions of each of
83. and field Nearly all programmers use three tabs in this case though to maintain proper field alignment for purposes of readability 2 5 Operand Field The third field of a typical statement is the operand field which naturally con tains the operand or operands for the instruction Most instructions e g STX SAM and ADDA JOE require at least a simple operand field while other instructions e g RCB RDB and the 6301 s AIM MASK MEMORY require a minimum of two subfields in the operand field Many instructions e g LDAA 3 X and STAA O X option ally allow two subfields in the operand field and a few instructions e g the 6301 s OIM MASK OFFSET X allow three subfields in the operand field Some instructions e g FCB FDB CONC and OPT permit more than three subfields in the operand field Fi nally some instructions such as NEGA SKIP1 and SKIP2 don t require any operand field at all If an operand field contains two or more subfields a comma must be used to terminate every subfield except the last one The assembler allows but does not require white space after the comma that terminates one subfield and before the beginning of the next subfield This rule makes the assembler s source syntax compatible with a standard typing rule that requires a blank space after a comma The operand field can be terminated by white space or by a carriage return Most users type one or two t
84. arameter Remarks EJECT Start a new page in the symbolic listing END Terminate the assembly The assembler ig nores anything that appears after the END statement ENTRY Define the entry point of a program Note that the host environment that executes the program must support this feature in order for this directive to be useful EQU expression Define the label to equal the value of the expression ETEXT End text Terminate text mode The ETEXT statement ends a block of comments EXPORT label1 Export the specified label s which must be defined somewhere in this module to make them available for use as imported labels in other modules that will be linked with this module The EXPORT statement can be used only in a relocatable assembly See IMPORT for related information FCB valuel Form constant byte Generate one data byte for each expression and name the first data byte label Each object byte contains the 8 bit signed or unsigned value of the corre sponding expression FCBS valuel Similar to the FCB directive except that as sembler ensures that the given value is in the 8 bit signed range i e 128 127 FCBU valuel Similar to the FCB directive except that as sembler ensures that the given value is in the 8 bit unsigned range i e 0 255 FCC string Form constant characters This statement is the same
85. arameters See Chap ter 5 for an explanation of the CMTCOL option If no alphabetic character appears within CMTCOL columns following the macro name then the macro uses positional parameters 118 CHAPTER 8 MACRO ASSEMBLY 8 2 1 Positional Parameters Positional parameters do not have formal names but are identified by an index of the form n where n denotes the position of the parameter in the macro call Parameter index values must be in the range 1 through 255 1 denotes the parameter that is associated with the first parameter 2 denotes the second parameter and so forth The user simply scripts a n index wherever a parameter string is to be substituted When the assembler stores the macro definition it determines how many parameters the macro uses i e the largest index value Then when the assembler encounters a macro call the macro processor expects to find that number of strings The assembler reports an error if there are too few or too many parameters The missing parameter and too many parameters messages may be suppressed using the source line OPT switches NOMPERR and NOTMERR or the run time options NMP and NTM Refer to Chapters 5 and 6 for information on the use of options and refer to Chapter 10 for information on error messages The macro call model line in the definition of a positional parameter macro must not contain anything within CMTCOL columns following the macro name The following simple example illus
86. as a 16 bit value and it uses that value as a 16 36 CHAPTER 4 EXPRESSIONS Table 4 4 Shift and Rotate Operators Operator Function ASL same as LSL ASR rr S 2 0 a LSL same as ASL LSR 0 115 01 ROL 15 0 ROR 15 0 bit unsigned shift count that specifies the number of bit positions by which the assembler should arithmetically shift the current 16 bit expression value In an arithmetic shift to the right the assembler inserts copies of the original sign bit at the left end of the 16 bit expression value and it discards bits as they pass out of bit zero to the right As a result the sign bit is duplicated or extended to the right according to the shift count and any shift count greater than 14 yields a 16 bit expression that is either 1 or zero depending on the original sign The LSL operator logically shifts the 16 bit expression value to the left This operator behaves exactly like the ASL operator Refer to the documentation about the ASL operator for more information about the behavior of this operator The LSR operator logically shifts the 16 bit expression value to the right The as sembler interprets the term following the LSR operator as an unsigned 16 bit shift count In a logical shift to the right
87. assembler recognizes either of these characters as a request for immediate addressing and the expression that immediately follows the equality sign or the number sign specifies the constant value of the instruction s operand For example LDAA 5 loads the constant value 5 into the AR and SUBB 47 subtracts the constant value 47 from the BR Similarly LDX 3 loads the value 3 into the XR and CPX 0 compares the XR against the value zero Some instructions such as LDX and LDS require 16 bit operands and other instructions such as ADDA and CMPB use 8 bit operands When a programmer specifies immediate addressing with an instruction that uses a 16 bit operand the assembler accepts any 16 bit value for the immediate operand A programmer can therefore specify any expression value as long as that expression value can be represented in a 16 bit number system A programmer can consider an immediate value to be either signed or unsigned so the assembler accepts both signed and unsigned immediate operand values For example LDX 65535 and LDX 1 are two ways of writing the same instruction because the unsigned value 65 535 is the same 16 bit number as the signed value 1 When a programmer specifies immediate addressing with an instruction that uses an 8 bit operand the assembler checks the expression value and reports an error for any im mediate operand value that can t be correctly represented in an 8 b
88. assembly terminates When LAB1 ASM does exist the file is assembled and the two companion files LAB1 LST and LAB1 0BJ are created in the current directory Generally assembly errors found in the source are reported both to the terminal and within the listing file though these actions can be controlled with run time options that are described later in this chapter By assuming the default file extension ASM for source files and by automatically appending LST and OBJ to the listing and object files the assembler promotes the recom mended organization files related to M6800 assembly on the VAX VMS system Not only will the user be able to save a few keystrokes in specifying files but he or she will also accrue the advantage of simplifying communication with other users who adhere to the standard FILES SRC LST OBJ sAM SAM Joe SAM JoE JOE lt cr gt However as this example shows the user is not obligated to adhere to any logical scheme of naming the source listing and object files Although the assembler will not be confused can the user say the same Incidentally VMS file identification is not case sensitive which means that any combination of uppercase and lowercase alphabetic characters can be used in a file name The assembler of course matches this attribute and will recognize the file sAM SAM in the current directory even though VMS displays the name as SAM SAM when a DIR command is issued FILES SRC LST OBJ LAB1 NONE
89. at ap pears at the top of each page of the symbolic listing A user can change the title string as often as desired VERIFY boolean_expression Verifies that the given expression is true dur ing assembly This directive helps ensure that assumptions in the code remain true XDEF label1 The XDEF statement is equivalent to the EXPORT statement but the EXPORT state ment is preferred XREF label1 The XREF statement is equivalent to the IMPORT statement but the IMPORT state ment is preferred 5 1 ASCII The ASCIT statement which requires a quoted string as an argument tells the assembler to generate an ASCII data byte for each character of the string The string must be enclosed between two identical delimiters each of which is a single nonwhite character that doesn t appear as a character in the string itself The following statements illustrate some correct methods for specifying quoted strings ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII Who is John Galt Duke Leto is dead Frodo exclaimed Watch out Gandalf A Q 7 amp O _ 3 5 lt gt XYZA Go to the Tardis the Doctor ordered mow Live long and prosper am 1991 Same as 99 RADAR Same as ADA 50 CHAPTER 5 ASSEMBLER DIRECTIVES The following statements show incorrect uses of the ASCII directive ASCII Out out brief candle ASCII The de
90. ately obvious to a person who reads the program Secondly if any intervening code is inserted or deleted after the BRA 5 and before the target address of the branch the program will no longer operate correctly If the offending instruction is changed to a BRA 1F instruction however the program will be easier to read and will continue to perform correctly when intervening code is inserted or deleted between the branch and the destination of the branch 4 1 9 RAD 40 Hash String A question mark immediately preceding a string of usually three label characters specifies a RAD 40 hash string as a term of an expression Any character that is valid as a character in a label is also valid as a character in a RAD 40 hash string The RAD 40 hash code is a compact encoding scheme that allows a user to encode three label characters into only two bytes instead of requiring three bytes as ASCII would so a programmer can profitably use the RAD 40 code to compress storage requirements for strings of label characters Each label character has a RAD 40 code value in the range from 1 through 39 and the RAD 40 hash code for a string of three characters is the radix 40 number whose three base 40 digits correspond to the RAD 40 character codes for the three characters of the string For example the RAD 40 character codes for the characters P I and G are 28 21 and 19 respectively so the RAD 40 hash string PIG specifies a term with
91. ations control switch to the MPU position and turn on the power to the microcomputer kit Be sure the baud rate is set appropriately 300 baud if you re using a phone dial up 9600 baud for a direct port Then push the reset button on the microcomputer to generate a reset RSI interrupt and type LD to initiate a MUDBUG load operation with a zero displacement 3 Move the communications control switch to the load position and then type TYPE LAB5 0BJ at the terminal After the VAX has received the carriage re turn at the end of the line it will download the object program directly into the microcomputer When the VAX has finished downloading the program move the communications control switch on the microcomputer back to the CPU position before proceeding with the VAX run 6 6 Advanced Topics Most users can ignore the material in this section especially during a first reading This section is included to provide the interested user an insight into the full capabilities of the M6800 assembler 106 CHAPTER 6 RUNNING THE ASSEMBLER 6 6 1 Parsing the Files and Options Specification Input String Perhaps the singular case of the noun string in the title above will clue the reader that the Section 6 2 description is not complete with regard to the user s input specifications In fact a user is able to specify both the files and all the run time options in a single input string in certain conditions that are outl
92. ber 5 48 RDB The repeat double byte RDB directive is an extension of the FDB statement The count expression in the operand field defines the number of times that the assembler will per form the FDB function on the remaining expression values The result is much as if the corresponding FDB statement had appeared count times in the source file The syntax of the statement is as shown below and further information on the acceptable values of the expressions is available under the FDB entry in this chapter label RDB count valuel comment field 79 ASSEMBLER USER S GUIDE 5 49 RMB The RMB Reserve Memory Byte instruction is equivalent to the BSS instruction and the RMB instruction is included only for compatibility with Motorola s definition of the assembly language Please refer to the description of the BSS statement for more information 5 50 SBTTL The SBTTL assembler directive is very similar to the SBTTLE directive described in detail below except that SBTTL does not solicit the EJECT function automatically SBTTL spelled without the trailing E is mnemonic for no page E jection 5 51 SBTTLE The SBTTLE statement tells the assembler to use the string from the operand field as the current subtitle The assembler always prints the current subtitle if any on the second line at the top of each page of the symbolic listing The first line at the top of each page is used of course for the title line that
93. bility reasons uses the pseudo opcodes GOIFZ go if zero and GOULZ go unless zero to skip over or to include conditionally assembled blocks of code Regardless of the mode of conditional assembly the GOTO pseudo instruction is employed to skip from one block of conditional assembly code to another While the GOIF and GOUNLS statements depend on the value of an expression and the GOIFZ and GOULZ statements depend on the value of an expression ANDed with the conditional assembly parameter the GOTO statement is evaluated unconditionally Conditional assembly also uses system labels A system label is not entered into the symbol table and it is uniquely identified with a dollar sign followed by a hexadecimal value Since system labels are not put into the symbol table the same system label can appear several times in one program System labels are used as targets for the GOIF GOUNLS or GOIFZ and GOULZ and GOTO statements Conditional assembly statements are normally listed in the pass 2 symbolic listing but they may be suppressed from the listing either with a run time option confer Chapter 6 or with an OPT pseudo instruction as discussed in Chapter 5 of this user s guide Perhaps the functions of conditional assembly can best be understood through the use of examples GOIF XX 10 means If the expression XX is true i e evaluates to a nonzero value then skip over the intervening source lines until reaching the target label
94. ble 5 1 which extends over the next several pages contains a complete list of the assembler s directives and the remaining part of this chapter contains brief descriptions of the various assembler directives The assembler directives are presented in alphabetical order so a reader who wants to refer to the documentation for a particular assembler directive can find the appropriate section quickly and easily Anything that is shown in the Parameter column of Table 5 1 is a descriptive name for a parameter that a programmer can supply in an actual program The notation in the table indicates that the assembler allows multiple occurrences of the designated field or subfield Table 5 1 Summary of Assembler Directives Directive Parameter Remarks ASCII string Generate object byte s that contain the ASCII value s of the character s of the string and define label as the address of the first data byte of the string Continued 43 ASSEMBLER USER S GUIDE Table 5 1 Continued Directive Parameter Remarks ASCIIC string Generate a data byte that contains an un signed count of the character s in the string and define label as the address of that data byte Then generate data byte s that con tain the ASCII value s of the character s of the string ASCIIZ string Same as ASCII but generate a zero null byte after the last byte of the string ASCT For reloca
95. ble 6 1 Continued Option Remarks SEQuence NSEQuence Include the line number or sequence field in the listing output SLine NSLine Include the line number of definition for labels included in the symbol table SRC NSRC Include the source code field in the listing output Symbol NSymbol Generate a symbol table of all labels and their values and include it in the listing output TMerr NTMerr Report an error if a macro call has too many parameter sub stitution strings Toc NToc Include a table of contents generated from SBTTL or SBTTLE statements in the listing output Value NValue Include the value or object code field in the listing output Width n Set the maximum number of columns in a line of the listing output to n 6 3 1 ABSOLUTE The assembler is capable of generating either absolute or relocatable object code pro vided that the programmer has taken care to observe the requirements of a particular mode with respect to label definition and usage The absolute mode of assembly is selected with the ABSolute run time option and the companion mode is selected with the RELocatable option More information on relocatable assembly can be found in Chapter 9 and in the discussions on the assembler directives ASCT BSCT CSCT DSCT EXPORT IMPORT and PSCT found in Chapter 5 The mode of assembly can also be controlled from within the source file using the REL and ABS switches with the OPT assembler directive The
96. ble instructions in that they don t have operand fields In some cases an instruction in this group such as WAT simply doesn t use an operand and the lack of an operand obviously explains the lack of an operand field for the instruction In other cases an instruction in the inherent group uses one or more operands but the instruction s operand s are inherently specified in the definition of the instruction itself For example the ABA instruction adds the BR to the AR so its operands are inherently defined to be the BR and the AR Similarly the CLRA instruction clears the AR so its operand is inherently defined to be the AR The instruc tions that use inherently defined operands don t have operand fields because their operands are implicitly specified by their opcodes 3 2 Immediate Addressing Many instructions allow the use of a mode that is commonly known as immediate ad dressing With immediate addressing the actual value of the instruction s operand appears immediately after the instruction s opcode in memory Instead of being a variable the operand is a constant value that is known at assembly time or perhaps at link time and a 21 ASSEMBLER USER S GUIDE programmer types the operand s value rather than its address in the instruction s operand field A user who wants to specify immediate addressing must use either an equality sign or a number sign as the first character of the operand field The
97. bler allows various styles of parentheses to force precedence as desired Parentheses can be nested to virtually any level When parentheses are not used to dictate precedence the assembler simply evaluates expressions from left to right with all operators having equal precedence so the user doesn t need to memorize any arbitrary precedence relationships among the many operators The assembler allows statement continuation from one source line to another so a pro grammer can conveniently continue a single statement over any desired number of lines The continuation feature gives the user added control over the source format and consequently over the listing format For example line continuation permits a user to break up long CHAPTER 1 INTRODUCTION statements that would otherwise ruin the uniform appearance of the listing by extending into the columns that are normally used for the comment field The assembler has an extensive set of assembler directives Besides recognizing assembler directives that generate object values reserve memory space and or define labels the assembler also features directives for generating and formatting a table of contents adding references to the concordance and controlling several assembler options Many of the assembler s options can be controlled either by option statements in the user s source file or by run time options that can be dynamically specified from the terminal or job stream when the
98. bler uses only the 16 least significant bits of the specified value The distance between two lines is defined to be the absolute value of the difference between their line numbers The distance from a line to itself is therefore zero The default local label range limit is 104 lines 74 CHAPTER 5 ASSEMBLER DIRECTIVES 5 40 15 OPT LONG SHORT The SHORT option tells the assembler to print only the first line of any statement e g FCC FCB or FDB that generates multiple output lines of object values and the LONG option tells the assembler to generate all object lines of a multiple line statement This option allows the user to have only the significant line of a multiple line generating instruction output to the symbolic listing if one was selected and shortens the listing by not listing lines that are not important The SHORT LONG options may be specified as many times as desired in the program The default for the SHORT LONG option is LONG 5 40 16 OPT LPAGE The LPAGE n option tells the assembler to produce n lines of output per logical page of the symbolic output listing These n lines include the lines that the assembler automatically produces for the heading of each page The assembler reports an invalid option 10 error unless LPMIN lt n lt PHYPG where LPMIN is somewhat arbitrarily defined to be 20 and PHYPG is the number of lines that exist from perforation to perforation on a physical page The LPAGE parameter is in
99. bly is output at the end of pass 2 When a complete pass 2 symbolic listing is output all the error lines in the listing are chained together via their line numbers into a linked list of sorts so the user can easily find all of the errors in a program even if the program is quite long The user can also use a text editor to search for the XX string that is used to identify each error message In any case the user should be able to locate all errors quite easily A typical error line consists of two fields The first as alluded to above contains char acters of the form XX where XX represents a unique two or three letter error mnemonic corresponding to each of the possible assembly errors Besides the convenience of the XX form in conjunction with a search command in a text editor the appearance of a unique 147 ASSEMBLER USER S GUIDE code within the first few characters of a line ensures that these important messages will not go unheralded even in a severely truncated listing The second field of the error line contains a complete English sentence which offers a concise explanation of the assembly error The message is short enough to occupy only one line of the listing and in fact is kept under 80 characters to avoid the sometimes distracting line wrap around on standard CRT display devices All possible assembly error messages are reproduced below in alphabetic order for con venience of the user Each message is accompanie
100. cal label 1H the branch following TEST will branch into the macro which is probably not what the programmer intended The following example illustrates another problem with using local labels with a macro Macro LOOPNE was previously defined in the section describing the MEXIT assembler directive 1H 1H LOOPNE 1B Obviously the programmer wants a conditional branch to the top of the loop at the first 1H local label We tend to think of macros as occupying a single line when we write the call but the expanded macro resides on more than one line Inside LOOPNE is a branch to 1 Guess where that is Not where the programmer expected the 1B substituted for 1 points to the macro call itself This code can hang the program into an infinite loop that never gets out of the expanded macro Macros can be used for several purposes Obviously anytime a sequence of instructions must be repeated frequently a macro can save time However before defining a macro the 127 ASSEMBLER USER S GUIDE programmer should consider the functionality of that sequence of instructions Does the proposed macro perform a function Or does the group of instructions just happen to fall together often Would a macro clarify the flow of a program by raising the level of the code Or would it obscure the meaning of the program by hiding part of the processing One common use for macros is to implement opcodes that would be nice to have For example the M6800
101. call The programmer may have tried to input more parameters than the macro supports Since the programmer may deliberately use a comma to delimit the final string as a shortcut to avoid using lt gt syntax to specify a null string the assembler allows the user to suppress reporting of the too many parameters error The TMerr option tells the assembler to report an error whenever too many parameters are detected and the NTMerr option suppresses reporting these errors There are also assembler directive TMERR and NOTMERR switches in the OPT set which override the run time options 104 CHAPTER 6 RUNNING THE ASSEMBLER The default for the TMerr NTMerr option is TMerr 6 3 28 TOC The TOC option tells the assembler to output a table of contents for the program that is being assembled The table of contents is derived from the SBTTL E statements that are included in the program and the assembler prints the table of contents at the beginning of the listing Each subtitle is listed in the table of contents along with the line number on which it occurs The TOC option is especially valuable for large programs that contain many different parts The NTOC option tells the assembler to suppress the table of contents If no SBTTL E statements are included in the program the TOC option is ineffective The default value for the TOC NTOC option is TOC 6 3 29 VALUE The Value option tells the assembler to output the generated object
102. cally provides top and bottom margins of six lines each there are normally 66 12 54 lines printed on each page Two of these lines are always used for a title line and a blank line and one more line is used for a subtitle if the user includes subtitles Therefore the assembler normally prints 51 or 52 lines of the user s program on each page 6 3 17 MPerr By default the assembler reports an error if the macro call does not specify enough parameter substitution strings For a macro that uses positional parameters an error is detected if the macro call contains fewer parameters than required For a macro that uses formal parameters an error is detected if any parameter that is not associated with a default 101 ASSEMBLER USER S GUIDE string is not assigned in the macro call The macro expansion processor substitutes a null string for any unspecified parameters Since the programmer may intend that the unspecified parameter s be replaced by the null string the assembler allows the user to suppress reporting of the missing parameters error The user should not routinely suppress the error message it is best to suppress the error message to get a clean listing only after ensuring that nothing important is omitted The MPerr option tells the assembler to report any missing parameter errors and the NMPerr option suppresses reporting missing parameters There are also assembler directive OPT switches MPERR and NOMPERR wh
103. ce file INTRPT ASM begins here TITLE INTRPT ASM BTEXT This is the companion module to MAIN ASM It contains a series of possible NMI interrupt handlers that the user will select interactively from within the main routine ETEXT OPT REL EXPORT NMI1i NMI2 These two labels which are the starting addresses of the interrupt handlers will be passed to MAIN ASM via the EXPORT to IMPORT connection BSCT SHORT RMB 1 Since interrupt handlers need to be very 138 CHAPTER 9 RELOCATABLE ASSEMBLY AND FAST GLOBAL TRANS1 TRANS2 NMT1 NMI2 RMB RMB CSCT RMB DSCT FCC FCC PSCT LDAA STAB RTI LDAA STAB RTI A gr FAST GLOBAL SHORT GLOBAL efficient it may well be that the direct addressing mode will be profitably used Perhaps one of the routines is very memory efficient and the other is execution time efficient Even though MAIN ASM calls this variable PARAM the accesses to GLOBAL in this module get mapped to the same physical memory address This method of global communication is similar to FORTRAN common equivalencies The data section here contains the beginnings of two translation tables that the two interrupt handlers might use to quickly convert data they obtain from the stack NMI1 is the fast one Note that register and stack usage is not explicitly shown here as for any real interrupt handler Before returning NMI1 will store some value in the
104. ces as white space characters This esoteric feature is completely transparent to many users but it is often helpful for programmers who use text formatters to format their source files If a programmer uses grave accents as nonexpandable spaces for a text formatter then the 19 ASSEMBLER USER S GUIDE programmer doesn t need to waste time formatting the source file for every intermediate assembly Instead the programmer can skip the formatting step altogether and submit the raw source file directly to the assembler and the assembler interprets the grave accents in the raw source file as white space characters Some programmers use grave accents as nonexpandable blanks between the subfields of multi part operand fields and the assembler correctly recognizes these grave accents as white space characters 20 Chapter 3 Addressing Modes for Executable Instructions The operand field of an executable instruction often specifies not just the location or value of the operand but also the addressing mode that the instruction should use to access the operand The addressing modes that are available in the M6800 family are known as inherent addressing immediate addressing indexed addressing extended addressing direct addressing and self relative addressing 3 1 Inherent Addressing Several instructions e g ABA CLRA PSHB DAA and SWI are said to use inherent ad dressing The instructions in this group are unique among executa
105. cidentally edit invisible characters into their source files Except where characters appear in quoted strings or as quoted ASCII literals in expres sions the assembler ignores the case of alphabetic characters A user can therefore type labels opcodes and expressions in any desired mixture of uppercase and lowercase and the assembler considers each lowercase character to be equivalent to its uppercase counter part Although the assembler is insensitive to case programmers conventionally use pure uppercase for everything but comments For comments of course people normally use an appropriate mixture of uppercase and lowercase just as this paragraph uses a mixture of uppercase and lowercase to enhance readability 2 2 Comments If the first column of a source statement contains an asterisk the assembler considers the entire line to be a comment line An asterisk in column one is therefore known as a comment indicator and it provides one method for typing a comment Better methods 6 CHAPTER 2 SOURCE SYNTAX of typing comments are available though and the assembler recognizes the asterisk com ment indicator chiefly for purposes of backward compatibility to earlier assemblers The BTEXT ETEXT feature which is described later is best for blocks of full width commentary such as the introduction that normally appears at the beginning of a routine or with the definition of a data structure and comments that relate to short groups of
106. command file or batch job 109 ASSEMBLER USER S GUIDE 6 7 Final Notes The M6800 assembler should not be viewed as a fixed and finished software product This assembler has history and it has a future it is an ongoing process While the claim is made that the assembler is fast and robust it is the intent of the author that future growth is in the direction of better meeting the needs of the user For this reason it is hoped that the user of this assembler will take the time to explore its features and what is more important will offer suggestions for its improvement Of particular interest are documented examples of assembler behavior that seem to operate contrary to that outlined in this user s guide Any report of a combination of files and options that cause the assembler to exit in any but a controlled fashion will be promptly acted upon and any changes that are necessary will be conscientiously completed The Assembler is Robust As an exercise the M6800 assembler itself which is written in Macro 32 assembly lan guage for the VAX was run through the assembler as a source file In the course of ninety seconds of execution the assembler found 21 748 errors with itself For that matter the binary executable image of the M6800 assembler was also assembled and 53 errors were reported after five seconds of execution The Assembler is Fast When it wants to be Felix qui potuit rerum cognescere causas The Assembler is
107. constant characters opcode mnemonic is simply another name for the ASCII opcode mnemonic The assembler recognizes the FCC mnemonic to provide com patibility with other assemblers but most users prefer the more descriptive ASCII opcode mnemonic The FCC statement which requires a quoted string as an argument tells the assembler to generate an ASCII data byte for each character of the string Please refer to the description of the ASCII statement for more details 5 25 FDB The FDB form double byte assembler directive is similar to the FCB form constant byte assembler directive but an FDB statement generates a 16 bit constant that occupies two consecutive object bytes while an FCB statement generates an 8 bit constant that occupies a single object byte The FDB statement is useful for creating a 16 bit constant value that can be accessed by a program at execution time The FDB statement tells the assembler to evaluate the operand expression and to generate two object data bytes that contain the 16 bit signed or unsigned value of the expression The most significant half of the 16 bit expression value for an FDB statement occupies the first object byte and the least significant half of the 16 bit expression value occupies the second object byte Note that the assembler allows both signed and unsigned values for FDB expressions The format for an FDB statement is as follows 61 ASSEMBLER USER S GUIDE label FDB value_1 value_2
108. cription of the S Record format object code is provided On a casual or first reading this material can be omitted since usually it is transparent to the user Notice first of all that all object values are output in ASCII coded hexadecimal so object modules can easily be handled by timesharing systems Object modules can also be listed and manually interpreted rather easily The format of an object record is best explained by an example The following ASCII string illustrates a typical absolute object data record with 16 data bytes 511301008E0132FE0137C603960AA1022705095A59 144 CHAPTER 9 RELOCATABLE ASSEMBLY The first two characters are S1 to indicate that we have an absolute object data record These two characters are the same for every absolute object data record The next two characters 13 are interpreted as a hex number to tell how many more pairs of characters there are in the object data record This particular record contains 13 i e 19 character pairs beyond the first four characters so the total record contains 4 19 2 42 characters The next four characters 0100 are interpreted as a 16 bit hex number and this number is used as the memory origin for the data values that follow in the record In this example the data values would be loaded into consecutive memory locations of a microcomputer starting with location 0100 The next two characters 8E are interpreted as an 8 bit unsigned hex number and t
109. ct the larger of two signed values MINUNS Select the smaller of two unsigned values MAXUNS Select the larger of two unsigned values 34 CHAPTER 4 EXPRESSIONS Table 4 3 Unary Operators for Assembler Expressions Operator Remarks Do not negate the following term Negate the following term BITS Count the number of one bits in the following term COM Unary one s complement of the following term 0000 gt FFFF LOG2 Unary base 2 logarithm of the following term NOT Unary logical not of the following term True gt False 65 535 The assembler discards any remainder from an unsigned division and it trun cates the unsigned quotient toward zero For example FFFF DIVUNS 2 7FFF and 60000 DIVUNS 50000 1 The assembler s MOD operator performs the modulo function for signed values so it should be used with values in the range from 32 768 through 32 767 The assembler defines the result of lt expression gt MOD lt term gt to be the signed remainder that results from dividing the signed value of the expression on the left by the signed value of the term on the right For example 5 MOD 3 2 5 MOD 3 2 5 MOD 3 2 and 5 MOD 3 2 The assembler s MODUNS operator performs the modulo function for unsigned val ues so it should be used with values in the range from zero through 65 535 The re sult of
110. ctions Each type of section is then restricted to contain code or data of specific attributes so that a program like the LK6800 linkage editor can successfully incorporate the relocatable objects of several modules into one coherent pack age At a general level of understanding relocatable address and expression references are defined to within an additive constant i e the yet to be determined value of the base address of the particular section and the assembler must leave markers or flags within the object code records to inform the linker where adjustment must be made A more precise specification of the communication between the LK6800 linker and the M6800 assembler can be found in the documentation for the LK6800 linkage editor since for the most part these specifications remain transparent to the user For now it is enough to describe the five specific types of program sections and their attributes that the M6800 assembler supports in relocation mode Each of the sections have a corresponding pseudo op that causes the assembler to update the location counter variable that it internally maintains to match the current state of the location counter uniquely defined within each section ASCT or absolute section is a non relocatable section There may be a number of absolute sections in a user s program These sections are used to allocate or initialize memory locations that are assigned by the programmer rather than by the LK6800 linkage
111. ctions into logical steps and explaining the logical steps NEGA Set the next DELTA_T value in the list STAA DELTA_T X equal to its original value minus the value that we 11 use for the new entry s DELTA_T value This step is necessary to account for the new DELTA_T value that will be inserted into the priority list STX NEWENT Save a pointer to the position for the LDX LISTEND new list entry and initialize the XR to point to the location following the end of the list We want to move the tail of the list down to make room for the new entry 2H DEX This loop moves the tail of the list down LDAA 0 X by one entry to make room for the new STAA ENTSIZE X entry We move one byte at a time instead CPX NEWENT of moving an entire entry at a time because BNE 2B we don t have a register that s big enough to hold an entire entry Good programmers organize their comments as well structured paragraphs of complete and proper English sentences with no arcane terms or cryptic abbreviations The only intended audience for a comment is a human being so programmers should always write comments with the same quality and style that are appropriate for any report document or article that humans will read 2 7 Statement Continuation The assembler allows statement continuation from one source line to the next so a programmer can conveniently continue a single statement over any desired number of lines Since the assembler s continuation feat
112. ctronic mechanical or otherwise without the prior written permission of the author Contents 1 Introduction 1 2 Source Syntax 5 2 1 Source Characters 2 trata ote ne ee ee ee a 6 2 2 Comments i arae a Ee A op Bee A a ae 6 2 3 Label Field a cu Bee oe ew A ee Be A a 8 2 3 1 Symbole Labels raesade Ga o Bote Oa Ba we HY aE 8 232 Local Labels 24 proh ea E A ee a EAS 9 2 3 3 Conditional Labels i e e do We ida adha 12 2A Opcode Field eo ras 4 a ts a a ee e 12 2 5 Operand Field e ee 13 2 6 Comment Field e e ee 13 2 7 Statement Continuation 20 0 000 ee ee 14 2 8 Assembler Support for Text Formatters a 17 2 8 1 Dot Commandds 00000 ee ee 18 2 8 2 Grave Accents ee 19 3 Addressing Modes for Executable Instructions 21 3 1 Inherent Addressing gt gt oet a oe rea aoua pa a Daa e pa d ee 21 3 2 Immediate Addressing e 21 3 3 Indexed Addressing oo o a 22 3 4 Extended Addressing ooo ee 23 3 5 Direct ATEN A e ey ee A 23 3 6 Self Relative Addressing 2 aa ee 24 ASSEMBLER USER S GUIDE 4 Expressions 27 AD EMS rt A Se ee E ee ek EE Oe BS 27 dk Binary Numbels ic tia not etal dota hae EE Bete at 27 41 2 Octal Numbers 2 222 ap Se ee be be oe as 28 4 1 3 Decimal Numbers s eos s a nipapa paia 000202 eee eee 28 4 1 4 Hexadecimal Numbers 0 0 020000 ee eee 28 A165 Symbolic Labels cier cria
113. currently selected A programmer can specify the parity option and change it as often as desired by using the OPT PARITY expression assembler directive which is fully explained elsewhere in this chapter If a programmer never uses an OPT instruction to specify the desired parity the assembler uses zero parity by default If zero parity is being used then for example the ASCII code for the character A is 41 If mark parity is being used on the other hand the ASCII code for the character A is C1 If the label field of an ASCIT statement contains a symbolic label or a local label the assembler defines the label to equal the memory address of the first byte of the string The assembler performs extensive error checking and reports an error if it finds anything wrong with the quoted string that appears as the argument of an ASCII statement For example the assembler reports an error if a nonwhite character immediately follows the closing string delimiter because the programmer probably tried to use the string delimiter as a character of the string in this case The assembler also reports an error if it fails to find the closing string delimiter and it similarly reports an error if the string is empty Programmers often use ASCII statements to generate character strings that can be used as output strings at execution time and programmers also use ASCII statements to generate strings that can be matched against user inputs at e
114. d by an explanation of the error probable causes of the error and when appropriate a cross reference to other chapters of this user s guide where further relevant examples of correct usage appear AD The operand address exceeds the range allowed for this instruction ADdressing error A branch class instruction is coded with a destination address that is outside the range of addresses that can be reached by the instruction AM The given addressing mode doesn t exist with the given instruction Addressing Mode error The user has correctly specified the addressing mode and the user has also correctly specified an instruction However the given addressing mode is not available with the given instruction 7AR It s too late to specify absolute or relocatable mode Absolute Relocatable error An OPT ABS or OPT REL statement occurs after the first non introductory statement and we can t change modes after we ve already started building the symbol table or generating object values 7BM A macro name must start with an alphabetic character Bad Macro name The first non white space character on the line following the MACRO pseudo op is not alphabetic 7BR An invalid numeral for a specified base appears in an expression Base Range error A hextet that appears in a number is too big for the number s base For example an 8 or a 9 cannot appear in an octal number and similarly an A or
115. d by the fact that it starts with a period in column 1 and the assembler considers dot commands to be special kinds of comment lines By default the assembler suppresses dot commands from the assembly listing but it can optionally print them as comment lines when desired Since the assembler normally suppresses dot commands from the assembly listing the listing appears essentially as it would appear if the dot commands didn t exist in the source file A programmer can therefore include dot commands in a source program to prepare it for final formatting but the programmer can still assemble the program directly without wasting time to format the source program for intermediate assemblies 2 8 2 Grave Accents When a text formatter processes and reformats text a single blank space in the original source text might change into multiple spaces or even a line break in the formatted output For example a space between words can change into a line break when a text formatter moves words from one line to another within a paragraph to produce lines with relatively uniform lengths This process is known as filling the text A single blank space can also change into multiple spaces if a text formatter inserts extra spaces between words to justify or align the right margin of a paragraph While most blank spaces are simply word separators that can be freely adjusted as described above a few blank spaces are different and should not be modified or expanded
116. d keys may be intimidating facile use of an editor will prove to be a quite valuable return on one s initial investment in learning to use it Files once created can be renamed deleted copied and printed The pertinent VMS commands are RENAME DELETE COPY and PRINT respectively and are very easy to use at times one may find the delete command disastrously easy Again it should be stressed that HELP on any of these topics is just four characters away A word on the relationship between the version number part of the full file identifier and the PURGE command may save the user from generating arbitrary file names to maintain uniqueness for a sequence of versions in an evolving source program A new file when created is given a version number of 1 If one edits this file and exits with changes using the same name VMS automatically saves the current edition of the file with a version number incremented by one Similarly if the assembler is told to generate a listing or object file with a file name that already exists in the current directory then the assembler creates these files with a version number one higher than the highest in the directory Most VMS commands automatically use the highest version numbered file by default when the rest of the file name is common to more than one file but the PURGE command deletes all but the highest version numbered file in such a set The user should be aware however that most accounts are set up by defaul
117. development edit compile link run edit compile link run In the specific case that concerns the assembly language programmer the high level language constructs are the opcode mnemonics and pseudo ops that comprise the assembly language itself It is the task of the assembler to take as input these constructs and based upon some well defined syntactic and semantic rules refine the high level meaning into a sequence of numeric data known as object code that will be eventually executed in a specific microprocessor It is in the realm of the word eventually in the last sentence that the concept of relocation is important Consider first the simple situation of a dedicated microcomputer without any special memory management techniques One hits the reset button and hardware functions clear the registers and perhaps load the program counter with the address of some simple boot strap ROM based routine that first loads then transfers control to a rudimentary operating system that resides in some basal area of the address space Say just to continue a hypo thetical situation that this operating system can accept through a serial port a sequence of bytes of numeric data that it can store into memory at arbitrary locations as directed by specifically encoded information bytes at the beginning of each of the transmitted se quences Finally when the transfer is complete a user should be able to pass control to some starting address that
118. discussion is made of possible errors and the messages that occur during the interactive front end processing aspect of the assembler s execution Information about file handling errors at the system level or incorrect run time option specifications is discussed in Chapter 6 itself As such no discussion is made of possible errors and the messages that occur during the interactive front end processing aspect of the assembler s execution Information about file handling errors at the system level or incorrect run time option specifications is discussed in Chapter 6 on running the M6800 assembler The M6800 assembler produces an error line for each error it finds up to a limit of five errors per line If the NReport option is specified errors are not reported to the symbolic listing If the NError option is also selected errors are not reported to the terminal either More information on these two run time options and their interaction can be found in Chapter 6 The user should be careful in the selection of the two error suppression options since fatal errors may be hidden from the user The default error reporting is to send the error messages both to the assembly listing if one has been specified and to the terminal If the terminal has been selected as the listing output device the errors are reported only once to avoid confusion and time consuming duplication A message that shows the total number of errors that were recorded during the assem
119. e CONC statement eliminates the deficiency of the invisible references by generating an artificial concordance entry for each variable name that appears in the CONC statement s operand field 5 12 CSCT The CSCT statement tells the assembler to start using the relocatable location counter that applies to the common section of memory so the CSCT statement is allowed only in a relocatable assembly not in an absolute assembly The common section is a relocatable area of uninitialized memory that is shared among modules usually for common variables or for scratch memory Each module s common section overlays the same area of memory so the common section is analogous to FORTRAN s blank common For additional infor mation concerning the CSCT statement please refer to the chapter that describes relocatable assemblies 5 13 DSCT The DSCT statement tells the assembler to start using the relocatable location counter that applies to the data section of memory so the DSCT statement is allowed only in a relocatable assembly not in an absolute assembly The data section can be relocated to any memory address at link time so accesses to items in the data section require extended addressing as opposed to direct addressing The data section normally resides in read write memory and it is normally used for variables that can t be put into the base section For more information regarding the DSCT statement please refer to the chapter that describes
120. e assembler it directly generates an executable object module that is in loader format In this case no linking step is necessary The assembler provides a powerful set of directives for conditional assembly The condi tional assembly can be conditional on items that are specified in the source file and it can also be conditional on a value that the user can specify as a run time option when invoking the assembler The ability to make assemblies conditional on a value that is specified as a run time option allows a user to generate multiple object versions of a program without editing the program s source file The assembler includes a flexible macro facility that allows a user to customize or extend the assembler s instruction set The assembler s macro facility provides the following fea tures positional or formal parameters default strings for formal parameters nested macro definitions nested macro calls string concatenation automatic label field generation de limited or undelimited strings continuation of undelimited strings user controlled listing or suppression of expanded lines and controllable syntax error reporting that can be partially suppressed by the user ASSEMBLER USER S GUIDE Some assemblers and compilers aren t very helpful when they report errors For example some assemblers issue meaningless error numbers such as error 735 or cryptic general purpose error messages such as the ever popular
121. e assembler to omit these temporarily extraneous source lines from the listing The NOIGNORE option allows the user to display the dot lines in the assembly listing Run time variants of this option are described in Chapter 6 5 40 10 OPT INCLIST INCNOLIST OPT INCLIST tells the assembler to list INCLUDE statements in the symbolic list ing and OPT INCNOLIST tells the assembler to suppress INCLUDE statements from the symbolic listing The assembler now lists INCLUDE statements by default unless user includes an OPT INCNOLIST statement You can nest the OPT INCLIST and OPT INCNOLIST statements much as you can nest BEGIN END statements in Algol and a nested INCLIST INCNOLIST pair preserves the current listing option 5 40 11 OPT LBLLIMIT The LBLLIMIT n option tells the assembler to consider only the first n characters of a label to be significant and to ignore all characters after the nth one Labels can be as long as the user wants up to and including a limit of 80 characters but only the first n characters of a label are significant The default value for the LBLLIMIT option is 80 but the user can reduce its value to conserve space in the symbol table and to make the assembly faster Note that care must be taken not to change the label limit between the occurrence of a label in the label field and its use in an operand field since the labels will not be the same if there is even a single character difference in the sig
122. e comment lines Command files are executed by prefixing the file name with an sign o ON ERROR THEN EXIT If something goes wrong terminate this command file a useful command for new users of command files Also the assembler returns an exit code via a system exit call so that the command file is able to detect errors in assembly o ASM P1 TT none Nlist o The P1 in quotes represents the first parameter passed to the command file as DEBUG COM FILENAME Note that the options are taken from the input stream of 108 CHAPTER 6 RUNNING THE ASSEMBLER this command file Data records have no leading sign This command file is set up to run the assembler for a file in the debugging stages when only errors are to be reported to the terminal H YN NO YN SP EXIT This next example is a command file named FINAL COM and is used to tailor the as sembler for a source file once it has been debugged and finalized Notice that all of the arguments are listed on the ASM command line which LIB GET_FOREIGN will pass to the assembler This is FINAL COM and is executed as QOFINAL filename ASM P1 NEO NMP NTM EXIT Of course these two samples just hint at the possibilities of command file operation used in conjunction with the M6800 assembler Furthermore once command files are un derstood they can be used with the VMS submit command as jobs in the batch
123. e macro in the opcode field The macro processor copies the lines of the macro model one at a time as though they were part of the source file The assembler replaces any occurrence of the parameters with the actual substitution strings specified in the macro call The process of copying the macro model lines and replacing the parameters is called macro expansion and the lines produced by macro expansion are called macro expanded lines The macro processor provides the following features Positional or formal parameters Default strings for formal parameters Nested macro calls String concatenation Automatic label field generation e Delimited or undelimited strings 117 ASSEMBLER USER S GUIDE e Continuation of undelimited strings e User controlled listing or suppression of expanded lines e Syntax error reporting some of which can be suppressed The following shows the structure of a macro definition MACRO lt macro name gt lt formal parameter list gt lt model statements gt MEND 8 1 Macro Names A macro name consists of an alphanumeric string that must begin with an alphabetic character Unlike normal symbol names only the first six characters of a macro name are significant This restriction is consistent with the fact that macros are used like user defined opcode mnemonics which also contain only six significant characters Also unlike normal symbol names macro names can be redefined as often a
124. e run time options The default for the MSHOW NOMSHOW option is MSHOW 5 40 20 OPT PAGENUM The PAGENUM n option tells the assembler to set the page number of the next page to n and to restart the page numbering sequence from this point This option statement is best used immediately before an EJECT statement This option is also available as a run time option see Chapter 6 but the run time option applies to the initial page number of the symbolic listing and it cannot modify page numbers in the listing after the first page If both assembly and run time options are selected the page number will be reset at the occurrence of the PAGENUM option in the source stream as if the run time option had not been selected The default value for the page number is 1 5 40 21 OPT PARITY The user may specify what parity an ASCII character in the program is to have with the OPT PARITY n option Currently only mark and zero parity are allowed If n equals 1 we have mark parity and if n equals 0 of course we have zero parity Any other value for n results in an invalid option 10 error The option selected controls the values of the 76 CHAPTER 5 ASSEMBLER DIRECTIVES characters in FCC strings as well as quoted character values in expressions For example the character A will have the value C1 with mark parity and an A will have the value 41 with zero parity The user can switch between zero parity and mark parity values as
125. e value unchanged A programmer can think of the rotate count for a ROR operator as being either signed or unsigned For example VALUE ROR 10 and VALUE ROL 10 both yield the same result For a typical example of a good use of a shift operator return to the situation in which a person wants to load the address not the contents of location VECTOR into the AR and the BR with the AR receiving the most significant byte of the address of location VECTOR and with the BR receiving the least significant byte of the address of location VECTOR A programmer could write LDAA VECTOR LSR 8 as an appropriate instruction to get the most significant byte of the address of location VECTOR into the AR The LSR operation shifts the most significant byte of the address of location VECTOR into the least significant byte of the expression value thus making the expression value correct for the LDAA instruction with immediate addressing A typical example of a good use of a rotate operator occurs when a programmer wants to test a particular bit of the AR The programmer could write an instruction such as BITA 1 ROL 5 to test bit 5 of the AR or the programmer could write an instruction such as BITA 1 ROL OVER_RUN assuming that the label OVER_RUN is defined as a symbolic name for the bit number of a status bit that is set when a data overrun error occurs for a serial I O port 4 2 4 Relational Operators The assembler supports a complete set
126. ecimal number that matches the value of a target conditional label that appears in the label field of a subsequent target line The assembler uses hexadecimal numbers for conditional labels much the same as FORTRAN uses decimal numbers for statement labels The unconditional GOTO statement often appears in conjunction with the conditional GOIF statement to perform an if then else conditional function for a user A programmer frequently wants to assemble one of two alternative sections of a program In this case the programmer can use a GOIF statement to skip or include the first alternative section and 63 ASSEMBLER USER S GUIDE the programmer can use a GOTO statement at the end of the first alternative section of the program to skip the second alternative section when the GOIF statement selects the first alternative section The GOTO statement is also useful when a programmer wants to omit a section of a program without deleting that section from the program s source file The programmer simply uses a GOTO statement to skip the section in question at assembly time but the skipped section can be recovered easily since it is still present in the program s source file For more information regarding the GOTO statement please refer to the chapter that describes conditional assembly 5 29 GOULZ The GOULZ go unless zero statement is the logical complement of the GOIFZ go if zero statement Like the GOIFZ statement the GOU
127. ed its maximum value FF LOP An invalid logical operator appears in an expression Logical OPerator error A logical operator that is not recognized appears in an expres sion For example the user might have typed ANC instead of typing AND as intended The assembler assumes a AND operator for the invalid logical operator and continues to evaluate the expression LOT The termination character for a logical operator must be a period Logical Operator Termination error The termination character for a logical operator 152 CHAPTER 10 ASSEMBLY ERRORS is invalid For example a user might have typed AND instead of typing AND as desired The assembler ignores the invalid termination character and assumes a period in its place LR A local label reference exceeds the local label range Local label Range error The instruction refers to a local label that is not defined within LLRNGE lines The user can set LLRNGE to any desired value via an OPT statement MC Assembler syntax requires a comma that is missing Missing Comma error The assembler s syntax requires a comma and the comma is missing 77MD A label that is defined here is also defined elsewhere Multiply Defined symbol error The label that appears in the label field is a multiply defined symbol That is the label that appears in the label field of this statement also appears in the label field of at least one other statement The user can induce m
128. ediately followed by a nonblank character indicates a probable error in which the user accidentally used the delimiter character as a character in the quoted string In the case of an omitted string delimiter the assembler processes the string to the end of the source line In the case of a nonblank character immediately following the closing delimiter the assembler processes the string to the closing delimiter 77RE Relocation directives are not allowed in absolute mode RElocation directives error A relocation assembler directive such as ASCT BSCT CSCT DSCT PSCT IMPORT or EXPORT appears in a program that is being assembled in the absolute mode The assembler reports an error and essentially treats the invalid relocation directive as if it were a comment line RL An expression s relocation factor is invalid ReLocation factor error The relocation factor for an expression must be in the range from 1 through 1 and the relocation factor must be zero for all except one section We report an error if these restrictions are violated 156 CHAPTER 10 ASSEMBLY ERRORS RLV A relocatable expression value is used where it is not allowed A relocatable expression value appears in a context where it is not allowed For example a relocatable expression value is not allowed with BSS RMB ORG GOIF GOUNLS GOIFZ GOULZ or GOTO statements and a relocatable expression value is not allowed as the first expression with the RCB and R
129. editor ASCT can be used to define the locations of PIA s or ACIA s for example Since the ORG pseudo op is only defined as an absolute construct its appearance in a source line has the same effect as switching to an absolute section with an ASCT statement BSCT or base section is a relocatable section There is only one base section The linkage editor assigns portions of the base section to each module that requires space in BSCT The base section is generally used for variables that are to be accessed using the direct addressing mode BSCT is restricted to memory locations 00 FF inclusive If the location counter for this section ever exceeds this one page limit the assembler generates a LO error CSCT or blank common section is a relocatable section There is only one blank common section CSCT is similar to blank common used in FORTRAN The blank common section 131 ASSEMBLER USER S GUIDE cannot be initialized hence the only legitimate statements in this section are those that request the allocation of variable storage such as RMB or BSS Violations of this convention are flagged by a CI error Regardless of the names that various modules assign to their allocations of the CSCT the first byte and hence all subsequent bytes of each CSCT address the same byte of the physical address space DSCT or data section is a relocatable section There is only one data section The linkage editor assigns portions of this sec
130. ek Ae eo a eS 101 0317 MR CrP Megs foe AA a AR ee IA 101 6 318 MRU a cin ohn ate ata e E Se as A ee Ral 102 6 3 19 MSHOW us ana Bie ek eS ada ae ee ae 102 0 3 201 PAGE weht A ee AE A BGP eh a 103 6 3 21 RELOCATABLE jo cn coda ci a ae ee ee E g 103 6 3 22 REPORT aii B44 e Re bok alin Bok otek ee a 103 6 3 23 SEQUENCE 2 2040 244 Bb eee ba ee Pa es 103 6 53 24 SEINE 3 oA a al eed ed ate AS A Wk ah ee a 104 63 20 SRE 2 rt a Le oe oe Bie oe eee oe LA a A ae on ae 104 63 2067 SYMBOL cate 2h toot A he Ee et a ia h 104 6 327 TMERRKos oa th A ae eet ee ee a ee ak Besos ke aie 104 63 28 TOO seg ise feats Seeks pee a Me E hae LS GE A Spe a 105 6 33 29 VALUE A AA Aa Di a h 105 6 3 30 WIDTH i eaoaai aa e a a eae a ela eb Dae ke ae 105 64 Afte The Assembly s purr A be Be E doe RO OO BAe Be ed 105 6 5 Downloading Object Files from VAX VMS 04 106 6 6 Advanced Topics e 106 6 6 1 Parsing the Files and Options Specification Input String 107 6 6 2 Alternative Modes of Assembler Execution 108 620 Final Notes e is la dr A gage Bt gh ek ot Gal EB eile ok Se aa 110 7 Conditional Assembly 111 vi CONTENTS 8 Macro Assembly 117 81 Macro Names unida ia ii fotos Shiga Boo Baie high Y 118 8 2 Macro Parameters osos miss Se ee a a od wees 118 8 2 1 Positional Parameters e 119 8 2 2 Formal Param tels socca 2 0000 cae da 120 8 2 3 Parameter Substitution Strings
131. el KEY or the label POINTER but an expression can be significantly more complex This chapter explains the rules that govern expressions The processors in the M6800 family have 16 bit address spaces so the assembler uses 16 bit arithmetic to evaluate expressions An expression can be simply a single term or it can contain multiple terms connected by binary operators An expression can also include a unary operator that applies to the first term of the expression The assembler evaluates expressions from left to right with no operator precedence rules but a programmer can use various styles of parentheses to dictate any desired order of evaluation An expression that appears within parentheses is called a subexpression and the assembler treats a parenthesized subexpression as a single term of the expression that contains it The same rules that apply to expressions also apply to parenthesized subexpressions 4 1 Terms An expression always contains one or more terms so a term is a basic component of any expression Table 4 1 summarizes the kinds of terms that the assembler accepts and the following paragraphs explain the various kinds of terms in detail 4 1 1 Binary Numbers A prefix character followed by a series of one or more bits Os and 1s is a binary number A binary number can contain as many as 16 significant bits so a user can specify full 16 bit values in binary if the need arises The smallest acceptable binary number is 0
132. eld a specified 8 bit checksum target value The CHKSUM statement is therefore useful to a programmer who wants to write a program that verifies the integrity of a ROM at execution time The format for the CHKSUM statement is as follows label CHKSUM target_value The expression in the CHKSUM statement s operand field specifies the desired target value for the checksum Most programmers use 1 as the target value but the assembler allows any 8 bit value Although the assembler accepts any 8 bit target value the value zero is specifically not recommended because a common kind of hardware failure can cause every location in an entire ROM to read as a zero byte thus passing a checksum verification if the target checksum is zero A programmer typically uses a CHKSUM statement to generate a checksum byte as the last byte of a ROM or as the last byte of a group of ROMs The programmer who wants to use the checksum byte to validate the ROM writes a self test routine that computes the 8 bit sum of all of the bytes of the ROM including the checksum byte The self test routine simply ignores both overflows and carries that occur during the dynamic computation of the checksum value because only the final 8 bit result is meaningful The self test routine compares the computed checksum value to the known target value and the routine reports an error in the ROM if the computed checksum value doesn t match the target value The following example code segm
133. ement except that the SBTTLE statement also starts a new page in the sym bolic listing SET expression The SET directive is similar to the EQU state ment but the SET statement defines a re definable label SK1 The SK1 statement generates a one byte ex ecutable instruction that skips the next one byte instruction at execution time SK2 The SK2 statement generates a one byte ex ecutable instruction that skips the next two bytes at execution time SKIP1 The SKIP1 statement is a more mnemonic name for the older SK1 instruction The SKIP1 statement generates a one byte exe cutable instruction that skips the next one byte instruction at execution time SKIP2 The SKIP2 statement is a more mnemonic name for the older SK2 instruction The SKIP2 statement generates a one byte exe cutable instruction that skips the next two bytes at execution time SPACE line count Skip over line count blank lines in the sym bolic listing but don t skip any lines at the top of a new page Continued 49 ASSEMBLER USER S GUIDE Table 5 1 Continued Directive Parameter Remarks SPC line_count The SPC statement is an older name for the SPACE statement SPACE is now preferred TEMPLT offset Starts a template section The given expres sion in offset is the value of the location counter for this section TITLE string Establish the string as a title line th
134. ended addressing If an is not included in the instruction the assembler chooses the addressing mode automatically much as the Motorola assembler does e The assembler accepts a string of label characters following a as a term in an expression and its value is the 16 bit RAD 40 code of the character string The RAD 40 code value is the same code that the assembler uses internally e The assembler treats empty lines as comment lines in the source input An empty line occurs whenever a carriage return is found with no effective characters preceding it Otherwise the assembler ignores empty lines e Motorola s definition of the character string with an FCC instruction is ambiguous With their definition for example 25 2 could mean 2 followed by 24 blanks or it could mean 5 with the character 2 as a string delimiter To eliminate this ambiguity we do not allow the count comma text definition of a string A string in an FCC instruction is therefore introduced by any nonblank character that the assembler accepts for input and it is terminated by the next occurrence of the same character The following examples illustrate ways to code FCC strings THIS IS A SIMPLE STRING ISN T IT THOUGH XLOOK AT ALL THESE amp amp CHARACTERS X e The assembler requires one or more blank spaces or horizontal tab characters between any two fields of an assembly language statement The assembler has pre set tabs consistent wit
135. ent a debugging program e g MUDBUG can set the program counter to the program s entry point when you download the program The following example illustrates the ENTRY statement START ENTRY Program execution begins here The START label is optional If a label is present it gets the value of the location counter An ENTRY statement does not generate any object code itself It merely indicates where execution of the program should begin An ENTRY statement is not appropriate in a module such as a subroutine module that does not run as an independent program 5 17 EQU The EQU statement equates the label in the label field to the value of the expression that appears in the operand field Since the assembler records values for labels during pass 1 no forward references are allowed in the operand field of an EQU statement 5 18 EQUREG 6809 The EQUREG statement is analogous to the EQU statement but the EQUREG statement takes a register list instead of an expression as its operand 5 19 ETEXT The ETEXT end text mode statement terminates a text block that was started by a BTEXT begin text mode statement Like the BTEXT statement the ETEXT statement is not 59 ASSEMBLER USER S GUIDE printed in the symbolic listing so a text mode comment block appears in the listing without the distracting clutter of the BTEXT and ETEXT statements that bracket it If the assembler encounters an ETEXT statement without first processin
136. ent is a self test routine that checksums a ROM The ROM starts at location ROM and contains LENGTH bytes and the target value for the checksum was specified as TARGET CLRA LDX ROM 1H ADDA 0 X INX CPX ROM LENGTH BNE 1B CMPA TARGET BNE ERROR In some cases the checksum byte can t be the last byte of a ROM because a system s hardware might require special values such as interrupt vectors at the end of a ROM In 56 CHAPTER 5 ASSEMBLER DIRECTIVES such a case a clever programmer can place the checksum byte elsewhere and still include the interrupt vectors in the checksum The programmer can simply use an ORG statement to put the interrupt vectors at their special locations and then use another ORG statement followed by a CHKSUM statement to generate the checksum byte at its location Thus the checksum byte can cover an entire ROM without being at the last location in the ROM The assembler also allows the CHKSUM statement in a relocatable assembly and the linker properly generates the specified checksum value to cover the object values generated in all program sections together If you want the checksum to cover contributions from all modules you must specify the module containing the CHKSUM statement as the last module in the link 5 11 CONC The CONC statement generates a concordance entry to indicate a reference for each label that appears in the CONC statement s operand field The purpose of the CONC statement is t
137. er symbols For example successive expansions of macro CLRBLK will define labels LOOP_A LOOP_B etc The use of the special character along with some other naming conventions can prevent conflicts with LOOPA LOOPB or other labels the user has defined directly Information on automatic label field generation is presented later in this chapter One advantage of formal parameter names is that they provide the ability to specify default substitution strings For example if the earlier definition were modified as shown below the user could call the macro without having to specify the parameters Assume that the most common use for macro CLRBLK is to clear an 80 byte buffer named BUFF We might change the first part of the definition to specify MACRO CLRBLK FROM BUFF TO BUFF 79 Now any macro call that does not expressly specify substitution strings for FROM and TO will be equivalent to listing the default strings in the call Default values can be quite convenient when a macro habitually performs the same specific task 8 2 3 Parameter Substitution Strings A string must satisfy one of the following definitions 121 ASSEMBLER USER S GUIDE e Any group of characters not containing any blanks tabs carriage returns or commas and not beginning with lt or a hyphen followed by blanks or tabs This last situation will be interpreted as line continuation so the hyphen and any characters that follow the hyphen will be ignored
138. ere are also run time MP and NMP options but the OPT source code options override the run time options The default for the MPERR NOMPERR option is MPERR 75 ASSEMBLER USER S GUIDE 5 40 18 OPT MPU The assembler is now capable of assembling source lines as if they belonged to any of a number of M6800 closely related microprocessors While the default MPU is the 6800 the user is able to specify that subsequent lines of source code are to be assembled for the following MPU s 6801 6802 6803 6808 6809 68HC11 6301 and 65150 This option is also expressible at run time see Chapter 6 5 40 19 OPT MSHOW NOMSHOW The assembler shows macro expansion by default but the user can change between the list mode and the no list mode for macro expanded statements as often as desired by selecting the MSHOW or NOMSHOW option respectively When MSHOW is in effect every line generated by the macro expansion processor except for any MEND MEXIT or other instruction that is never listed will be listed when MSHOW is not in effect only the macro call statement will be shown unless there are errors Any macro expanded line that generates an error will be shown regardless of the state of the MSHOW option Of course the user can suppress all error messages with a run time option There is also a run time MS NMS option but any OPT MSHOW NOMSHOW source code option overrides the MS NMS run time option Refer to Chapter 6 for information on th
139. espect to the positional parameters does make a semantic difference For the rest of this discussion any of these three symbols will be represented by TT When TT is specified for the source file the assembler understands that the user is going to enter from the keyboard the entire source file one line at a time with no editing once the carriage return at the end of each line is entered In most situations this would not be the preferred method of generating a source file since an editor is a far superior method for entering and modifying source line text However for the user who is both a careful organized typist and who is perhaps only testing one or two lines of questionable assembler syntax the method will reduce the overhead involved with invoking an editor Because the assembler requires two passes at the source file to complete the assembly there is a side effect of this terminal as source mode of operation Each line as it is entered is therefore copied to a source file named TEMP ASM which resides on disk in the current directory The assembler displays a message that it is creating this file before any lines are accepted to ensure that potential file name collisions are brought to the attention of the user If for some reason the assembler is unable to create a file in this directory the assembly is aborted and an error message is reported to the terminal The preferred method of ending this mode of source input is for the user to
140. f the presence or absence of subtitles Most users employ the SBTTLE statement at the beginning of each new subroutine or at the beginning of each logical section of a program If the user has asked the assembler to generate a table of contents the assembler uses all of the SBTTLE statements in the program to generate the table of contents Please refer to the section that describes run time options for more information 80 CHAPTER 5 ASSEMBLER DIRECTIVES If the first character of a title or subtitle string is a period the assembler now throws that period away and does not print it as part of the title or subtitle In the past you could not include any white space at the beginning of a title or subtitle string because the assembler assumed as it still does that the string started with the first nonblank character after the white space that terminated the opcode field Now however you can effectively include leading white space in a title or subtitle string by using a period to start the string Any white space that immediately follows the period will effectively become leading white space when the title or subtitle is printed You can use the new leading period feature to generate a table of contents with in dentation that graphically illustrates the logical structure of your program The following example illustrates how you might want to use this feature to obtain indentation in your table of contents SBTTLE 1 Level 1 SBTTL
141. fect the SK2 instruction may modify the N bit the Z bit and the V bit of the condition codes 5 56 SKIP1 The SKIP1 assembler directive is just a more mnemonic name for the SK1 directive described above 5 57 SKIP2 The SKIP2 assembler directive is just a more mnemonic name for the SK2 directive described above 83 ASSEMBLER USER S GUIDE 5 58 SPACE The SPACE pseudo instruction is used to insert blank lines into a pass 2 symbolic output listing for improved readability The SPACE instruction must not be labeled and the value from its operand field tells the assembler how many lines to skip in the listing For example SPACE 3 tells the assembler to leave three blank lines and SPACE N tells the assembler to leave N blank lines where N was previously equated to some number Regardless of the ultimate operand value at least one blank line will be inserted into the listing The SPACE command is counted as a line but it is not printed in the listing Most programmers use the SPACE instruction to separate their assembly language instructions into logical groups so their listing will be attractive and easy to read Users needn t worry about crossing a page boundary and generating any unnecessary blank lines at the top of the next page with the SPACE command because the assembler automatically stops skipping lines when it reaches the top of a page A SPACE command that just happens to occur at the top of a page therefo
142. for the remaining expression values The result is similar to what would occur if the corresponding FCB statement appeared count times RCBS count valuel Similar to the RCB directive except that as sembler ensures that the given value is in the 8 bit signed range i e 128 127 RCBU count valuel Similar to the RCB directive except that as sembler ensures that the given value is in the 8 bit unsigned range i e 0 255 RDB count valuel Repeat double byte The RDB statement is similar to the FDB statement The count ex pression defines a count that tells how many times the assembler should perform the FDB function for the remaining expression values The result is similar to what would occur if the corresponding FDB statement appeared count times Continued 48 CHAPTER 5 ASSEMBLER DIRECTIVES Table 5 1 Continued Directive Parameter Remarks RMB block_size Reserve memory bytes Reserve i e skip block_size bytes and name the first one la bel The RMB statement is equivalent to the BSS statement SBTTL string Establish string as the current subtitle and also insert string as an entry in the table of contents The subtitle string appears on the subtitle line of each page of the symbolic listing and a user can change the subtitle as often as necessary SBTTLE string The SBTTLE statement is identical to the SBTTL stat
143. formation at link time The 6800 microprocessor hardware vectors the NMI through this location It is an absolute location in the address space The main routine monitors this variable which is set by an interrupt handlers This is global information at run time The data section is used for variables and constants Here the input prompt is created with the FCC directive The main routine begins here and notice that no ORG statements have been used to lift the code above the MUDBUG utility in low memory This feature will be taken care of by the linker The LDAA instruction is representative of initialization In each of these blocks only one or two statements are given so that the final object file will not be too unwieldy 137 ASSEMBLER USER S GUIDE LDX PROMPT This represents the first step involved in displaying the input prompt at the terminal Actually this code could well be in a library of i o routines in another module CMPA 1 This comparison shows the last step in BNE 1F a routine which obtained input from the user and left it converted to numeric data in the AR 1H LDX NMI2 Use the relocatable imported value of STX NMIVCT the second NMI interrupt handler as the one selected by the interactive input Note that an absolute address is being referenced here WAIT CMPB PARAM Wait in this sense loop until the BNE WAIT interrupt handler has stored the quit Signal in PARAM END The sour
144. g Concatenation It is sometimes desirable to concatenate substitution strings The concatenation opera tor is an invisible delimiter A single vertical bar in a macro model line is never copied to the expanded line Instead the character is simply skipped For example consider the following simple macro MACRO CONCAT LEFT RIGHT ASCII LEFT RIGHT MEND The following call and expansion illustrate the effect of the concatenation operator CONCAT LEFT DOWN RIGHT TOWN ASCII DOWNTOWN The concatenation operator delimits the parameter names LEFT and RIGHT so they can be identified by the assembler but the character is not copied to the expanded line To include a vertical bar in the expanded line put two consecutive vertical bars into the source line In general the expanded line will contain one fewer vertical bars than appear in a sequence of vertical bars i e three vertical bars get expanded as two 11 and one vertical bar gets expanded as none 123 ASSEMBLER USER S GUIDE 8 5 Automatic Label Field Generation The concatenation operator also has one special usage if a vertical bar is immediately followed by a pound sign the assembler automatically generates a label field that depends on the number of times the macro has been called The first time a macro is called each occurrence of the string is replaced by the string A On the second call the string B is
145. g a matching BTEXT statement to start a text block the assembler reports an error Please refer to the description of the BTEXT statement for related information 5 20 EXPORT The EXPORT statement tells the assembler to export the specified label s to make them available for use as imported labels in other program s that will eventually be linked with the program that is being assembled The EXPORT statement is valid only in a relocatable assembly and the label s that it specifies must be defined somewhere in the program that is being assembled The format for the EXPORT statement is as follows EXPORT label_i label_2 label_n The EXPORT statement requires at least one label as an operand and it accepts any number of labels beyond the first one The EXPORT statement s label field must be empty For more information regarding the EXPORT statement please refer to the chapter that describes relocatable assemblies and also refer to the description of the related IMPORT statement 5 21 FCB The FCB form constant byte statement tells the assembler to evaluate the operand expression and to generate an object data byte that contains the 8 bit signed or unsigned value of the expression The FCB statement is therefore useful for creating a constant value that can be accessed by a program at execution time The assembler allows both signed and unsigned values for FCB expressions so the assembler accepts an expression if its value fits
146. g mode In these situations the OPT switch NODIRECT can be selected By default the assembler generates object code with direct addressing whenever expressions in the operand field of instructions that permit this mode can be evaluated in pass 1 5 40 7 OPT EOERR NOEOERR By default the assembler reports an error if the expanded macro line overflows the source line buffer i e exceeds 126 characters This error occurs when the length of the substitution string exceeds the length of the parameter name and the macro model line did not have enough unused spaces at the end to absorb the difference For example if 1 is replaced by ALPHA BRAVO CHARLIE DELTA the resulting macro expanded line is going to contain more characters than the model line does Since the programmer may determine that nothing of value was lost in the expanded line only the comment field was truncated the assembler allows the user to suppress reporting of the expansion overflow error The user should not routinely suppress the error it is best to suppress the error to get a clean listing only after ensuring that nothing important is lost The user can change between the report expansion overflow mode and the suppress expansion overflow error mode as often as desired by selecting the EOERR or NOEOERR option respectively There are also run time EO and NEO options and the OPT source code options override the run time options The default for the EOERR NOE
147. gram The VERIFY statement provides a convenient method for improving robustness and maintainability The syntax for the VERIFY statement is as follows VERIFY lt expression gt If the VERIFY statement s expression value is true i e nonzero the assembler simply lists the VERIFY statement in the program listing If the VERIFY statement s expression value is false i e zero however the assembler reports a VF error to alert the user to the fact that the stated assumption is false A user might employ the VERIFY statement as shown in the following examples 85 ASSEMBLER USER S GUIDE VERIFY LIST_SIZE LS MAX_SIZE VERIFY SPARES GE O VERIFY TABLE_SIZE GT 0 VERIFY TBL1SZ EQ TBL2SZ SKIP2 VERIFY 2F eq 2 1H ADDA 10 2H INCB 5 63 XDEF The XDEF statement is equivalent to the EXPORT statement but the EXPORT statement is preferred 5 64 XREF The XREF statement is equivalent to the IMPORT statement but the IMPORT statement is preferred 86 Chapter 6 Running the Assembler 6 1 Getting Started How many times has a new user of a software package been confronted with a thick user s guide and the dubiously helpful assurance from some more experienced user that It s all in the manual somewhere Well for just those people including but not limited to harried students attempting to obtain a working program mere hours before it is due the initial portion of this chapter on running the M6800 assembler
148. h either signed or unsigned numbers In each case the assembler retains the 16 least significant bits of the result and discards any bits that carry beyond the 16 least significant bits The final 16 bit result of an addition subtraction or multiplication is the same for either a signed operation or an unsigned operation The assembler recognizes two division operators because the results of signed divisions and unsigned divisions are often different in ways that can t be reconciled by mere trunca tion Similarly the assembler accepts a signed and unsigned modulo operator The division operator specifies a signed integer division so this operator should be used only with values in the range from 32 768 through 32 767 The assembler simply discards any remainder from a signed integer division and it truncates the signed quotient toward zero For example 8 3 2 8 3 2 8 3 2 and 8 3 2 The DIVUNS division operator specifies an unsigned integer division so the DIVUNS division operator should be used with unsigned values in the range from zero through 33 ASSEMBLER USER S GUIDE Table 4 2 Binary Operators for Assembler Expressions Operator Remarks Add signed or unsigned Subtract signed or unsigned Multiply signed or unsigned 16 bit product Divide signed integer division remainder discarded DIVUNS Divide unsigned integer division remainder
149. h instruction is short only 80 back or 7F forward Sometimes we can get away with coding BNE LOOP but sometimes the top of the loop is out of range The following macro handles that problem MACRO Automatically generate a short or long LOOPNE conditional branch to the top of a loop GOIF 2 1 If the destination is out of range of a GT 80 1 branch then go to generate a long jump 124 CHAPTER 8 MACRO ASSEMBLY BNE t Generate the short branch not equal and MEXIT quit macro expansion 1 BEQ LNE Since the destination is out of range JMP M branch around a long jump if the contrary LNE condition exists We would like to doa branch not equal but the top of the loop is too far away Instead we branch equal around a long jump MEND Notice that there are two possible ways macro LOOPNE can be expanded If the destina tion address is within byte range the macro call LOOPNE TOP will generate a short BNE instruction then the assembler will process the MEXIT instruction and stop macro expan sion However if the destination is too far away the assembler will be in conditional skip mode when the MEXIT instruction is expanded In skip mode the assembler is looking for the target label 1 in this case that ends the skip Thus the assembler never decodes the MEXIT instruction and continues to expand the rest of the lines 8 8 Macro Call As the preceding examples illustrate a macro is called simply by p
150. h the PDP initial tab stops i e 9 17 73 The tab stops may be changed by the user with the OPT TABS directive see Chapter 5 e This assembler does not permit forward references to future symbols in some pseudo opcode instructions For example SAM EQU JOE is valid only if the label JOE has been defined before the EQU statement is encountered in the source stream Refer to Chapter 4 for a discussion of the pseudo opcodes and which may not contain forward references e The front end processor allows users to select file names and run time options inter actively with the assembler and the user doesn t need to wait until the end of a run to find out that the options or file names were not acceptable to the assembler e The BLO branch lower than and BHS branch higher or same opcodes have been added The BLO instruction is equivalent to the BCS instruction and the BHS instruc tion is equivalent to the BCC instruction These two opcodes are provided to make life easier for the user who wants to make comparisons between unsigned values 162 CHAPTER 11 SUMMARY OF NONSTANDARD FEATURES e The other primary differences between this assembler and Motorola s definition of the assembly language concern pseudo opcodes OPT options and expressions These items are all thoroughly documented in the previous chapters 163
151. he BCC instruction which branches if the carry borrow bit in the condition codes is cleared Although the BHS instruction is the same as the BCC instruction the BHS opcode mnemonic conveniently provides a logical extension of the instruction set When a microprocessor in the 6800 family subtracts one unsigned number the subtrahend or bottom number from another unsigned number the minuend or top number the carry borrow bit in the condition codes is cleared if and only if the unsigned value of the minuend or top number is higher than or the same as the unsigned value of the subtrahend or bottom number The BHS instruction which branches if the carry borrow bit is cleared is therefore useful for comparing unsigned values and the function that the BHS instruction performs for unsigned numbers is analogous to the function that the BGE instruction performs for signed values While the BCC mnemonic is usually appropriate after a shift a rotation or an addition the BHS mnemonic is normally appropriate after a subtraction or a comparison The avail ability of the BHS mnemonic allows a programmer to use the mnemonic BHS or BCC that most clearly describes the function that the programmer intends to perform 5 6 BLO The BLO branch if lower opcode mnemonic is simply another name for the BCS instruc tion which branches if the carry borrow bit in the condition codes is set Although the BLO instruction is the same as the BCS instruction the BLO
152. his value is a data value to be loaded 8E will be loaded into location 0100 01 will be loaded into location 0101 32 will be loaded into location 0102 and 5A will be loaded into location 010F There are 16 data bytes in a full object data record so object files of full object data records can easily be broken down into segments for individual ROMs PROMs or EPROMs The last two characters of the example object data record 59 are interpreted as an ASCII coded hexadecimal number and this number is the one s complement 8 bit checksum for the object data record 59 is therefore the 8 bit one s complement of the 8 bit two s complement arithmetic sum of 13 01 00 8E 01 32 54 Notice that the checksum includes the record length characters and the block origin characters as well as the object data values Motorola may have used a one s complement checksum instead of a two s complement checksum just to make their object format different from Intel s 8080 object format or they may have used a one s complement checksum because a two s complement checksum allows a record of all zeros to pass In any case the standard Motorola object format requires a one s complement checksum 145 ASSEMBLER USER S GUIDE 146 Chapter 10 Assembly Errors This chapter is confined to assembly errors those abnormalities of source line syntax and semantics found in the assembly process itself As such no
153. his error message by specifying the NOMPERR option 7MQ The statement is missing a literal character argument Missing Quoted character An apostrophe or a quotation mark has been found in an expression to indicate that the next character is to be taken literally That is the ASCII code of the next character is to be used as the value of a term However the apostrophe or quotation mark is the last character of the source line and the next character is missing The assembler uses a blank space as a default character in place of the missing quoted character MR Macro Recursion is not allowed at any level Macro Recursion error During macro expansion recursion was detected Either the macro called itself or a subordinate macro called one of its predecessors 154 CHAPTER 10 ASSEMBLY ERRORS MR6 A required right parenthesis is missing or misplaced The assembler s syntax requires a right parenthesis to close a specification for indirect addressing and the right parenthesis does not appear where it belongs MRP A left parenthesis appears with no matching right parenthesis Missing right parenthesis A left parenthesis appears in an expression with no balancing right parenthesis to close the parenthesized subexpression The assembler automatically supplies the missing right parenthesis at the end of the expression MS A label in the operand field is a multiply defined symbol Multiply defined Symbol error A label tha
154. ho work with computers typically use automatic text formatters such as runoff compose prose nroff or troff to format nearly everything they write A text formatter is an extremely useful program that formats text according to standard rules and or formatting commands to improve both the appearance of the text and the consistency of its format In terms of format and appearance a text formatter essentially converts a rough draft into a final copy so a text formatter helps a person produce neat formatted text quickly and with a minimum amount of manual effort Besides reducing manual effort a text formatter also provides flexibility and generates output that is superior to the output that a person without a text formatter could expect to produce in a reasonable amount of time Additionally a text formatter encourages people to edit clarifications and other improvements into their text because the text formatter painlessly reformats the text and produces a new final copy that incorporates the improvements The writer doesn t need 17 ASSEMBLER USER S GUIDE to worry about things like margins or pagination because the text formatter takes care of those details automatically A source program in assembly language is essentially a text file that has been carefully prepared in a particular format Some parts of a source program contain assembly language instructions that have a distinctive columnized format and other parts of a source prog
155. ich override the run time option The default for the MPerr NMPerr option is MPerr 6 3 18 MPU The assembler now has a run time option for specifying the MPU for which you want to assemble the program The format is MPU xxxx where xxxx is 6800 6801 6802 6803 6808 6809 6301 or 65150 The default is 6800 The 6800 6801 6802 6803 6808 and 6809 are all Motorola products The 6301 is produced by Hitachi and it is an extension of Motorola s 6801 microprocessor The option to assemble source lines for another but closely related microprocessor is also expressible within the source file using the MPU variant of the assembler directive OPT The default microprocessor is the 6800 6 3 19 MSHOW The MShow option tells the assembler to show the program lines generated by macro expansion The NMShow option suppresses listing of the expanded lines When MShow is in effect every line generated by the macro expansion processor except for any MEND MEXIT or other instruction that is never listed will be listed when MShow is not in effect only the macro call statement will be shown unless there are errors Any macro expanded line that generates an error will be shown regardless of the state of the MShow option Of course the user can suppress all error messages with the NReport and NError run time options This option may also be specified in the source program by entering the MSHOW op tion or deselected by the NOMSHOW option and an
156. ider trying the following incantation as with all ritu als they must be reproduced exactly to obtain the desired results The mysterious symbol lt cr gt stands for striking the carriage return key and lt cntrl z gt means the simultaneous depressing of both the lt control gt and lt z gt keys copy sys input first asm lt cr gt org 200 lt cr gt alabel ldaa 100 a simple instruction lt cr gt end lt cr gt lt cntrl z gt The file first asm now exists in the current directory and it can be tried out in the role of a source file in the assembly process outlined above As a harbinger to the novice this example program includes an assembly error 6 2 The Informed User Since this user s guide describes a M6800 cross assembler written for the VAX VMS operating system it seems apparent that any knowledge one has about the host environment would aid in potent use of this software VMS documentation is notoriously complete and as a minimum it behooves the user to obtain at least a comfortable relationship with the following topics If what is discussed below lacks clarity the VMS on line help facility is suggested as a more complete explanation The concept of a file is of basic importance In files one organizes information into pre defined formats that pieces of software can manipulate In particular the file that contains the assembly statements is known as the source file The M6800 assembler ingests a source file
157. ier points in the program remain in effect after a subordinate CANOLIST CALIST pair have been processed The CANOLIST option suppresses GOIF GOUNLS GOIFZ GOULZ and GOTO statements statements with conditional assembly labels and skipped statements i e statements that won t be assembled because of a conditional assembly action from the output listing See Chapter 7 for a discussion of conditional assembly statements and see Chapter 6 for a description of the corresponding run time list conditional assembly option The default for the CANOLIST CALIST option is CALIST i e zero 5 40 3 OPT CMTCOL The CMTCOL n option allows the user to specify the number of blank columns that may occur between the end of the label field and the occurrence of a nonblank character before the line is considered to be a comment statement If there is no label in the label field the column counting starts with column one An invalid option error 10 is reported unless 1 lt n lt 80 CMTCOL is set to a default value of 10 at the beginning of each pass but it may be changed as often as desired Refer to the example in Chapter 2 under the discussion of the label field This feature alleviates the user from having to start comment lines with an asterisk provided that there are at least n blank columns preceding the first nonblank character of the comments The default value for the CMTCOL parameter is 10 5 40 4 OPT DB8 DB10 DB16 The user may
158. in normal programs Any label that either starts or ends with a special character is reserved by convention for use with system software that is properly part of the operating system A programmer who heedlessly starts or ends a label with a special character in a normal application program may therefore encounter serious but richly deserved difficulties Subsequent characters after the first character of a symbolic label can be alphabetic characters A Z or a z numeric characters 0 9 or special characters or _ Al though the special characters can be used freely as embedded characters within labels a label in a normal application program should never end with a special character because any label that ends with a special character is reserved by convention for use with system software A symbolic label can be as short as one character or as long as 80 characters By default the assembler interprets and records a maximum of 80 significant characters per label but a programmer can use the OPT LBLLIM expression statement which is described in another section to tell the assembler to limit the number of significant characters per 8 CHAPTER 2 SOURCE SYNTAX symbolic label to any desired length in the range one through 80 The assembler effectively ignores any characters beyond the number of significant characters in a label The assembler allows a programmer to define and use any label that syntactically qual ifies a
159. ined below Mere microseconds after the user enters the command ASM to execute the M6800 as sembler a call is made to a VMS system procedure LIB GET_FOREIGN to recover any string that may appear on the same line as the ASM verb itself This string if it is not empty is passed to the parser and is treated as if it were input in response to the files specification prompt In order for this system call to function the ASM command must have previously been defined as a special string symbol e g in the LOGIN COM file at the user or group level similar to that shown below which is pertinent to the situation at the VAX VMS installation at Arizona State University The leading dollar sign is the critical feature of this string definition ASM SYS USERDISK UD GROUP OWNER M6800ASM EXE 19 El In the discussion that follows white space is defined as any character that precedes in the standard ASCII ordering i e ASCII value 21 is definition automatically includes tabs 09 and spaces 20 The parser recognizes either the comma or the virgule as a separator between tokens Now from the beginning If LIB GET_FOREIGN returns with a null string the files specification prompt is issued and the user s input string is turned over to the parser If the string is still empty or if it contains only white space before the first separator is found no source file has been entered and the files prompt is continually re is
160. ing conditional assembly operations Addi tionally synchronization errors arise if symbol table overflow occurs because label values can t be put into the symbol table after symbol table overflow has occurred ST This label is undefined due to symbol table overflow Symbol Table overflow error The amount of memory that the assembler has available for the symbol table is insufficient for assembling the program and the label in the label field of the current statement cannot be put into the symbol table This label is therefore not defined and undefined symbol errors occur wherever this label is used The user can use an OPT statement to reduce LBLLIM to a smaller value or the user can use fewer la bels or shorter labels or more local labels or the user can deselect the concordance i e cross reference option Another alternative is to break the program into sections that are assembled separately SZ An expression value is outside the range allowed for its field SiZe error An expression value is too large to fit into the field in which it is used The expression value is truncated on the left so it will fit into the desired field 157 ASSEMBLER USER S GUIDE TM A macro call contains too many parameters Too Many parameters error When retrieving parameter substitution strings for each of the defined parameters the macro call statement appeared to offer at least one more string This probably occurred because the user
161. instruction set is missing many instructions that the M6809 provides One could define a library of macros that would simulate the 6809 instruction set or a subset thereof on the 6800 Another way of using macros is to streamline subroutine calls For example if a sub routine has the following calling sequence JSR SUBR FCB ARG1 FDB ARG2 FCB ARG3 the macro MACRO SUB JSR SUBR FCB M FDB 2 FCB 3 MEND allows calls of the form SUB ARG1 ARG2 ARG3 In addition to providing a high level call statement such an approach simplifies later changes to the calling sequence For example say we want to change the order of the arguments or we want to push the arguments onto the stack before jumping to the sub routine We can change the macro definition and all of the subroutine calls will be revised automatically 128 Chapter 9 Relocatable Assembly 9 1 Preliminary Discussion Before outlining the specifics of relocatable assembly as it has been implemented within the M6800 assembler it may prove useful to briefly circumscribe the scenario that oper ationally defines relocation The digression concerns the process that begins with some high level language constructs and ends with that unique sequence of binary information that in the von Neumann sense implements the algorithm on a specific machine The user probably knows this process couched in terms of the components to a commonly observed vicious cycle of program
162. instructions are best typed as automatic side comments which are described in the paragraphs that follow If seventeen or more columns of white space occur at the beginning of a source statement the assembler automatically considers the entire line to be a comment Similarly if seventeen or more columns of white space occur immediately after the label field the assembler processes the label field and automatically considers the rest of the line to be a comment Notice that the assembler conveniently counts columns of white space not white space characters A tab character is therefore equivalent to an appropriate number of columns of white space depending on the position of the tab character in the source line and the assembler s interpretation of a tab agrees with the user s intuitive feeling regarding columns of white space The number seventeen that appears in the paragraph above is not rigidly fixed In stead it is merely the standard default value for the parameter that is known as the as sembler s CMTCOL parameter If the default value of the CMTCOL parameter is unsuitable for some particular application a programmer can easily set the CMTCOL parameter to a more appropriate value by using the OPT CMTCOL expression assembler directive which is explained in detail later The assembler allows a user to change the CMTCOL value as often as desired in a single source program but most programmers simply use the standard defau
163. ion is invalid and produces an RL error since two relocation counts are nonzero Trying to use a relocatable value to allocate space produces an RLV error since this information must be known during pass one of the assembly While using a relocatable expression in an EQU directive is permitted an imported label acts as a forward reference and will cause the assembler to generate an IMP error Even though 1 d1 is only di this line has an UO error since only addition and subtraction are supported by the linker 135 ASSEMBLER USER S GUIDE threea LDAA d1 im1 im2 This expression is well formed with respect to relocation count but an 8 bit field is implied and relocatable values are therefore not allowed An RV error is reported threeb LDX d1 im1 im2 The same expression is fine when it is slated for a 16 bit field threec LDX d1 im1 im2 An invalid relocation count error RL occurs here since both the di and the which associates with PSCT have nonzero relocation counts DSCT four EQU d2 If the program section is changed to DSCT then the location counter symbol also changes meaning Here d2 gives a zero relocation count five FDB al ci c2 im1 Absolute imported and relocatable symbols can appear together as long as the above rules are observed 9 5 Example Usage of Relocatable Assembly In this section a simple and contrived example of relocatable assembly is presented which demonstrates the
164. ion limit as often as desired in a single program by using the OPT LLRANGE expression directive which is described in detail in another section The distance between two source lines is defined to be the absolute value of the difference between their line numbers 29 ASSEMBLER USER S GUIDE A local label in an expression has the form jB j backward or jF j forward where j is a single digit in the range from 0 through 9 A jB local label in an expression refers to the closest previous definition of the jH j here local label and a jF local label in an expression refers to the closest following definition of the jH local label A jB or jF term in an expression never refers to a local label in the label field of the same statement that contains the jB or jF term so a jB or jF term always refers to a label that is defined in a previous or subsequent statement For example the instruction 7H BRA 7B refers to some previously defined 7H label not to the 7H label in the same instruction Similarly the instruction 4H BRA 4F refers to some 4H label that will presumably be defined by a subsequent statement in the program A local label of the form jF is always a forward reference so a jF term is illegal in an expression that is used in a context where forward references aren t allowed For example the expression value for an ORG EQU or RMB statement must be evaluated during pass 1 so these assembler directives don t all
165. it the assembler reports an error in the format of the local label White space terminates the label field and a user customarily types a single tab charac ter to advance from the label field to the opcode field A programmer who wants to define a local label on a line by itself can use a carriage return to terminate the label field If a local label appears on a line by itself or if CMTCOL columns of white space appear between the local label and the next nonwhite character the assembler defines the local label to have the value of the location counter A local label that appears on a line by itself therefore receives the same value that it would have if an executable instruction appeared in the same statement with the local label 11 ASSEMBLER USER S GUIDE Local labels may initially be a little disquieting to a person who has never seen them before but they soon become completely natural to anyone who uses them in a few pro grams Most programmers who have used local labels wonder how anyone could ever get along without them 2 3 3 Conditional Labels If a dollar sign appears in the first column of a statement the assembler assumes that the label field contains a conditional label A conditional label which is also known as a system label is used to label a target line in a conditional assembly When the label field of a line contains a conditional label the assembler ignores the rest of the line Conditional labels and c
166. it field A programmer can consider an 8 bit value to be either signed or unsigned so the assembler accepts both signed and unsigned immediate operand values A signed 8 bit value lies in the range from 128 through 127 and an unsigned 8 bit value lies in the range from zero through 255 3 3 Indexed Addressing Indexed addressing which is available for many instructions provides flexibility by al lowing a user s program to compute an operand address dynamically at execution time With indexed addressing the microprocessor computes the memory address of the instruc tion s operand by adding the current 16 bit value of the XR i e X register or index register to the unsigned 8 bit offset or base address that is specified in the instruction Since the offset is an 8 bit unsigned value the operand can reside in memory at any lo cation in the range from XR 0000 through XR 00FF where XR denotes the value of the XR and where a leading indicates a hexadecimal number The XR of course can contain any 16 bit value at execution time so an instruction that uses indexed addressing can potentially access any memory location in the microprocessor s entire memory space A programmer who wants to use indexed addressing must include two subfields in the operand field The first subfield contains an expression that specifies the desired 8 bit unsigned offset or base address and the second subfield contains an X or XR to indicate
167. itialized at the beginning of each pass to a value that provides top and bottom margins of one inch but it may be changed as often as the user desires The default logical page size is set according to the physical page size that the user specifies with the LPP run time option Since the default value for the LPP parameter is 66 the standard default value for the LPAGE parameter is 66 12 54 5 40 17 OPT MPERR NOMPERR By default the assembler reports an error if the macro call does not specify enough parameter substitution strings For a macro that uses positional parameters an error is detected if the macro call contains fewer parameters than required For a macro that uses formal parameters an error is detected if any parameter that is not associated with a default string is not assigned in the macro call The macro expansion processor substitutes a null string for any unspecified parameters Since the programmer may intend that the unspecified parameter s be replaced by the null string the assembler allows the user to suppress reporting of the missing parameters error The user should not routinely suppress the error message it is best to suppress the error message to get a clean listing only after ensuring that nothing important is omitted The user can change between the report missing parameters mode and the suppress missing parameters error mode as often as desired by selecting the MPERR or NOMPERR option respectively Th
168. itional parameter by using the two comma separators FILES SRC LST OBJ LAB1 FINAL LIS lt cr gt After the assembly given this specification a listing file named FINAL LIS and an object file named LAB1 0BJ will be created in the current directory Note that it was not necessary to skip over to the third positional parameter by placing a comma after FINAL LIS because the user specified the default action with respect to the object file FILES SRC LST OBJ LAB1 LST LAB1 OBJ lt cr gt This specification results in the message You must enter a source file name and indeed that does seem to be a requirement for an assembly The user will be prompted again for the file specification input string until a source file name is entered The default action works only for the specification of listing or object files from a source file base not in the opposite direction FILES SRC LST OBJ NONE NONE NONE lt cr gt This example however is not an error provided that the file NONE ASM exists in the current directory 91 ASSEMBLER USER S GUIDE FILES SRC LST OBJ TT KB terminal lt cr gt One final file specification feature is demonstrated in this example All three symbols as well as their uppercase and lowercase permutations are considered by the assembler to represent special nondisk files While the three symbols are truly synonyms the user would probably only ever type the shortest where they appear with r
169. l must be defined elsewhere in the current program If the symbols contained in the EXPORT instruction are not defined in the program an IS invalid symbol error will be reported The IMPORT or equivalently XREF instruction is used to declare externally referenced symbols symbols that may be referenced in the current source program but that are defined via an EXPORT instruction in some other program The general form for the IMPORT instruction is IMPORT SYMBOL1 SYMBOL2 SYMBOL3 132 CHAPTER 9 RELOCATABLE ASSEMBLY The IMPORT statement must not be labeled Each symbol in the operand field must be declared by an EXPORT pseudo instruction in the program in which it was defined The user can specify several externally referenced symbols by typing a delimiter comma following each symbol and she or he can also type any number of blank spaces between the delimiter comma and the first character of next symbol As a limitation of the LK6800 linker only a maximum of 251 symbols can be imported and the assembler will generate an IML error if this limit is exceeded As with the five section swapping pseudo instructions EXPORT and IMPORT or XDEF and XREF are only valid when the assembler is in the relocatable assembly mode otherwise an RE error is reported 9 4 Implications of Relocatable Assembly To the user accustomed to the absolute assembly mode the profusion of program sections and symbol swapping might appear o
170. limiters must match ASCII gt The apostrophe is the user s mistake Although the assembler accepts any nonwhite character as a string delimiter most programmers use quotation marks to delimit strings A programmer who wishes to include a quotation mark as a character in a string ordinarily uses an apostrophe as a string delimiter and a user who wishes to include both quotation marks and apostrophes in a string usually chooses some other special character such as amp 3 or as a string delimiter A programmer should avoid the mistake of trying to use a hyphen as a string delimiter for a string that starts with white space because a hyphen followed by white space at the beginning of the operand field tells the assembler to ignore the rest of the line and to continue the statement with the first nonwhite character of the next line In this case the assembler does exactly what the user requests but that might not be what the user really wants Note that a hyphen followed by white space within a quoted string is perfectly valid as part of the string because the assembler takes all characters within a string literally as characters of the string The assembler generates one object data byte for each character of the quoted string and each object data byte contains the 8 bit ASCII code of the corresponding string character with the parity bit either set or reset according to the parity option that is
171. ls the assembler to generate an ASCII data byte for each character of the string In addition the ASCIIZ statement tells the assembler to generate a zero i e null byte immediately following the last byte of the string A programmer can conveniently use the zero termination byte as a flag that marks the end of the string in memory The following example illustrates an ASCIIZ statement along with a set of other state ments that are functionally equivalent to the given ASCIIZ GAUL ASCIIZ Veni vidi vici GAUL ASCII Veni vidi vici FCB 0 52 CHAPTER 5 ASSEMBLER DIRECTIVES As with the ASCIIC statement the assembler allows an empty string such as or with an ASCIIZ statement An empty string with an ASCIIZ statement causes the assembler to generate just the zero termination byte 5 4 ASCT The ASCT statement tells the assembler to start using the location counter that applies to the absolute section of memory Contrary to what one might first think the ASCT state ment is allowed only in a relocatable assembly not in an absolute assembly A programmer can use the absolute section in a relocatable assembly to define I O ports timers and other such items that reside at particular locations in the memory map For more informa tion concerning the ASCT statement please refer to the chapter that describes relocatable assemblies 5 5 BHS The BHS branch if higher or same opcode mnemonic is simply another name for t
172. lt CMTCOL value and never change it at all The CMTCOL feature allows a programmer to have an automatic comment field without the bothersome requirement to type an asterisk comment indicator in column one of each comment line This feature is especially convenient for side comments that are being used to describe a short group of instructions because the comments frequently extend one or more lines beyond the last instruction line of the code segment For example the following two program segments are functionally identical but the second segment takes advantage of the automatic comment feature while the first uses asterisk comment indicators STAA 0 X Store the maximum value from the AR DEX into the bottom entry of the list Then BRA LOOP decrement the list pointer and branch back to the top of the outer sort loop to see if the list is now completely sorted STAA 0 X Store the maximum value from the AR DEX into the bottom entry of the list Then BRA LOOP decrement the list pointer and branch back to the top of the outer sort loop to see if the list is now completely sorted ASSEMBLER USER S GUIDE The benefit of being able to omit the asterisk comment indicators from column one doesn t appear to be significant in the short example above However the ability to omit comment indicators from an entire program provides a considerable advantage for the person who must type the source file Also the program listing is cleaner a
173. lues from 1 through 39 correspond to the label characters amp 0 9 A Z and _ respectively The RAD 40 code values for the lowercase alphabetic characters a z are 13 through 38 the same as the RAD 40 code values for the uppercase alpha betic characters A RAD 40 hash string is therefore insensitive to case and we lose case information when we hash a string into its RAD 40 representation Note that the collating sequence for RAD 40 strings is the same as the collating sequence for ASCII strings The RAD 40 code can be extremely useful for an application in which a user wishes to condense the storage requirements for strings of characters However the user must realize that the RAD 40 code applies only to label characters not to all ASCII characters and the user must further realize that the RAD 40 code does not preserve any case information The RAD 40 code definitely has limitations but it can be highly beneficial for a programmer who understands how to use it 4 1 10 Conditional Assembly Mask The assembler accepts a special predefined token CAMASK as a term of an expression The CAMASK term represents the 16 bit value of a conditional assembly mask that can be specified as a run time parameter when a user invokes the assembler so a programmer has access to a parameter that can be dynamically specified as any desired value for a particular assembly The chapter that explains how to run the assembler includes information about the
174. lways need this option and the option is not one of the default values it is best to include the option in the 67 ASSEMBLER USER S GUIDE program using the OPT pseudo instruction If on the other hand this option is a one time selection it is best to use the run time option if it is available For example if a pro gram s output listing is to always start at page number 500 by including the assembly time statement OPT PAGENUM 500 in the source program the user doesn t have to remember to specify the starting page number each time the program is assembled If on the other hand the program is to be assembled with a starting page number of 500 for just one special run it is better to specify the initial page number with the Page n run time option so the user needn t insert a nonstandard instruction into the source code which he or she just might forget to remove after the assembly Some of the options are either set or reset i e selected or not some have levels associated with them and still others require specific value assignments A sample input for the option selections for outputting in base 10 an initial page number of 5 and a label length limit of 8 may be specified as OPT DB10 PAGENUM 5 LBLLIM 8 The options can be specified in any order and options can be specified in any combina tion of uppercase and lowercase characters However it is good programming practice to use uppercase for the inst
175. lways starts in column 1 as required by the assembler By convention the opcode field starts at the first tab stop which is in column 9 Similarly the operand field starts at the second tab stop which is in column 17 Also by convention the comment field starts at the fourth not the third tab stop and the fourth tab stop is in column 33 2 1 Source Characters The assembler accepts a maximum of 126 characters in a single source line and the assembler simply truncates any source lines that are too long Most users automatically limit their source lines to fewer than 80 characters because lines that contain 80 or more characters are less convenient to edit In fact most users limit their source lines to a maximum of 75 characters per line because that line length is standard for several applications Notice that a tab character is a single character in a source line but a tab appears to occupy several columns of a line when the line is displayed or printed A line that appears to exceed 126 columns might therefore be acceptable to the assembler without truncation if some of the characters in the line are tabs The assembler essentially ignores i e does not process delete codes and control char acters other than tabs but those invisible characters do count toward the limit of 126 characters per line just as tabs do The assembler s policy of ignoring invisible characters prevents subtle problems that might otherwise occur for users who ac
176. major features discussed above Specifically two modules MAIN ASM and INTRPT ASM were separately relocatably assembled and then linked with the LK6800 editor to produce an S Record format object file The source files an example relocatable assembly listing and a transcript of the LK6800 linking session are given below Next the two modules were combined into one program that was rewritten for the absolute assembly code generation mode The source file for this combined program is also given Finally both of the resultant object files are listed to confirm the comparability of the two processes The source file MAIN ASM begins here TITLE MAIN ASM BTEXT This is a contrived but hopefully informative example of relocatable assembly techniques The program consists of two modules One contained in this file is a main routine which obtains interactive input from the 136 CHAPTER 9 RELOCATABLE ASSEMBLY user regarding special interrupt handler routines The second file INTRPT ASM contains a series of these interrupt handlers NMIVCT PARAM PROMPT MAIN ETEXT OPT IMPORT ASCT EQU CSCT RMB DSCT FCC PSCT LDAA REL NMI1 NMI2 FFFC OPTIONS Since relocation will always be needed with this source code the OPT REL selection is preferred over the front end RELocatable mode Bring in the labels that mark the starting points of the interrupt handlers This is global in
177. mbol table listing if a symbol table output listing was requested see the Symbol option Hence to obtain a symbol table without the line numbers of definition the user needs to select both the Symbol and the NSLine run time options The default for the SLine NSLine option is SLine 6 3 25 SRC The SRC option tells the assembler to output the user s source code to the assembly listing and the NSRC option tells the assembler to suppress the user s source code from the listing This option may be used to dramatically reduce the output listing width and it is a valuable aid in determining how well someone understands the actual assembly code This feature is definitely not recommended for any output listings you may wish to read again in the future The default for the SRC NSRC option is SRC 6 3 26 SYMBOL The Symbol option tells the assembler to create and output a symbol table listing and the NSymbol option tells the assembler not to output a symbol table listing The symbol table listing provides an alphabetic listing of all nonlocal labels used in the source program and it also includes each label s value The default for the Symbol NSymbol option is NSymbol 6 3 27 TMERR By default the assembler reports an error if the macro call contains too many parameter substitution strings This error occurs when the macro call processor detects a comma following the last parameter substitution string of a positional parameter macro
178. meter is set dynamically by the user is important because it permits the user to assemble multiple versions of the same program from the same source file without editing the file For the M6800 assembler the conditional assembly parameter is set via the CAmask n run time option confer Chapter 6 If no value is specified by the user for the conditional assembly parameter the assembler uses a default value of zero Depending upon the mode of conditional assembly this parameter can interact with the expression in a conditional assembly statement in two ways The first is directly through the special assembler recognized token CAMASK which always takes on the value of the run time parameter CAmask The CAMASK token may appear anywhere in an arbitrary expression on the conditional assembly statement source line This method is the newer and clearer and hence to be preferred mode of conditional assembly supported by the M6800 assembler The pseudo opcodes GOIF go if true and GOUNLS go unless true are used to skip over or to include conditionally assembled blocks of code The second mode uses the conditional assembly parameter in a more rigid manner by always evaluating the logical AND of the CAmask run time value and the 16 bit value of the expression from the conditional assembly statement in the source line This mode of conditional assembly which will eventually be phased out of operation though is still being supported for compati
179. mismatch errors for expressions like 5 4 2 3 4 7 but a user who typed the same expression with just one style of parentheses would never be aware of the error The assembler allows a user to nest parentheses to virtually any level so programmers don t need to worry about writing expressions that are too complex for the assembler to evaluate In fact while an average programmer typically has difficulty understanding more than six levels of nested subexpressions the assembler easily handles 10 000 levels 42 Chapter 5 Assembler Directives The assembler has an extensive set of assembler directives Assembler directives which are sometimes known as pseudo instructions don t usually correspond to executable machine language instructions on a one to one basis as normal assembly language instruc tions do Instead an assembler directive might tell the assembler to reserve memory space for a variable to generate a particular object value for use as a constant or to define a label to have a specified value An assembler directive can also tell the assembler to take some desired action or to select some option For example the assembler includes directives for generating and formatting a table of contents adding special references to the concordance and controlling the format of the program listing Additionally the assembler recognizes directives that allow a programmer to control several of the assembler s options Ta
180. mments for debugging purposes To select the indentation the user specifies INdent n where n is an unsigned decimal number The default value for this option is zero i e no indentation 6 3 12 LCA The LCA option tells the assembler to send any lines associated with conditional assembly statements to the assembly listing The NLCA option suppresses GOIF GOUNLS GOTO GOULZ and GOIFZ statements and statements with conditional assembly labels as well as condi tionally skipped statements from the assembly listing This option may also be specified in the source program by entering the CALIST or deselected by the CANOLIST pseudo opcode instructions The LCA or NLCA options are one time selectable but the CALIST and CANOLIST options may be employed in the source program to change between the list mode and the no list mode for conditional assembly statements as often as desired The default for the LCA NLCA option is LCA 6 3 13 LINE NUMBER The LINe number n option tells the assembler that the beginning line number of the assembly listing is n where n is an unsigned decimal integer Note that the largest line number that may be output is 65 535 after which the line number sequence will start over with zero This option is convenient when writing modular program segments that will later be brought together with proper sequence numbering Occasionally blocks of line numbers are allocated before the start of a large programming project to ens
181. mplates for data structures and stack frames For example the following 54 CHAPTER 5 ASSEMBLER DIRECTIVES instruction sequence defines a template for the stack frame that appears on the stack im mediately after an interrupt in the 6800 microprocessor STACK EQU 0 TEMPLT STACK CC BSS 1 BR BSS 1 AR BSS 1 XR BSS 2 PC BSS 2 The statements in the example might initially appear to reserve memory locations start ing at location 0000 but they are actually being used to define meaningful labels as relative offsets from the first byte of the stack frame which can be anywhere in memory at execution time For example the label AR equals 0002 and the value of the microprocessor s AR resides in byte number two of the stack frame where the first byte of the stack frame is byte number zero A programmer could use a TSX instruction after an interrupt to point the XR to the first byte of the stack frame and then the programmer could use an instruction such as STAB AR X to modify the AR value in the stack frame 5 9 BTEXT The BTEXT begin text mode statement marks the beginning of a text block and it tells the assembler to treat all subsequent lines with certain exceptions as comment lines until the next occurrence of an ETEXT end text mode statement The assembler recognizes EJECT and PAGE SPACE and SPC SBTTL and SBTTLE statements even in text mode because those statements are often just as useful in comment blocks
182. n is too deep The assembler ignores the excessive levels of parentheses The maximum allowable nesting level can be increased to virtually any level by changing the label MAXNEST 155 ASSEMBLER USER S GUIDE NS A number exceeds 16 bits when it is converted to binary Number Size error A number exceeds 16 bits when it is converted to binary This error is repeated for each extra digit or hextet or octet or bit of the number after the number once exceeds 16 bits 70M Out of memory unable to process macro definition completely Out of Memory error While processing a macro definition a request for more symbol table memory was denied 0N Out of memory null strings used for specified parameter strings Out of memory Null strings used error A request for additional symbol table memory was denied so we are unable to store the default parameter substitution strings 70P The string in the opcode field is not a valid opcode mnemonic OPcode error The string in the opcode field is not recognized as a valid opcode mnemonic QT A quoted string is improperly terminated Quote Termination error Something is wrong with the termination of the quoted string in an FCC statement Either the closing string delimiter is missing or else the closing string delimiter is present but is immediately followed by a nonblank character An omitted closing string delimiter is obviously an error A closing string delimiter that is imm
183. name are significant regardless of the setting of the LBLLIM option Since the parameter names are associated with a particular macro definition the same parameter name can be used in more than one macro definition with no conflict or ambiguity The user can also use the same name for a parameter and a normal symbol but that may cause an unexpected result The macro processor will substitute for any parameter it finds in the macro model For example if the macro definition defines X as a parameter name and a model line references the variable X the assembler will replace the variable X with the specified string As the following example illustrates formal parameters are used in much the same way as positional parameters except that they are explicitly named Note that the first parameter name must begin within CMTCOL columns after the macro name The example shows the use of a generated label field which is explained later in this chapter MACRO Clear block shows use of formal CLRBLK FROM TO parameters STX SAVE_X Save the current value of the XR and LDX FROM load the address of the first location in the block LOOP_ CLR 0 X Clear each memory location and advance INX the index register to address the next CPX T0 1 location Loop until the last location BNE LOOP_ has been cleared LDX SAVE_X Restore the XR MEND The following illustrates a macro call and the resulting expansion CLRBLK TAB1 Clear table 1 TAB1 TABSIZ 1
184. nd in Chapter 5 The mode of assembly can also be controlled from within the source file using the REL and ABS switches with the OPT assembler directive The default assembly mode is absolute 6 3 22 REPORT The Report option tells the assembler to report errors normally in the assembly list ing and the NReport option tells the assembler to suppress the reporting of errors in the assembly listing The NReport option can be used as a desperation measure to suppress diagnostic warnings for a final clean assembly listing but it should be used with extreme caution since it can hide errors from the user This option has no effect on the reporting of errors to the terminal as selected by the Error NError option The default for the Report NReport option is Report 6 3 23 SEQUENCE The SEQuence option tells the assembler to include sequence numbers i e line numbers in the assembly listing and the NSEQuence option tells the assembler to omit sequence numbers from the listing The selection of the NSEQuence option may be used to reduce the width of the listing when necessary The default action is SEQuence 103 ASSEMBLER USER S GUIDE 6 3 24 SLINE The SLine option tells the assembler to create and output a symbol table listing that includes an alphabetic listing of all nonlocal labels each label s value and its line number of definition The NSLine option on the other hand suppresses outputting of the line numbers in the sy
185. nd less cluttered without the unnecessary distraction of the asterisks in column one of each comment line Finally with an old style assembler that requires comment indicators a comment indicator that is accidentally omitted can cause an assembly error and require the programmer to waste time by editing and assembling again 2 3 Label Field The label field starts in column one i e the first column and white space in column one indicates that the label field has been omitted A user who wants to omit the label field normally types a tab in column one to skip directly to the opcode field If the label field of a statement isn t blank it can contain a symbolic label a local label or a conditional label A symbolic label commonly known simply as a label is the ordinary kind of label that almost every programming language allows A local label is a special kind of label that can be defined over and over for various local uses and a conditional label sometimes known as a system label is a label that can be used as a target in conditional assembly The paragraphs that follow discuss the three varieties of labels 2 3 1 Symbolic Labels The first character of a symbolic label in a normal application program must be an alphabetic character A Z or a z The assembler also recognizes a colon an ampersand 2 or an underscore _ as the first character of a symbolic label but these special characters should not be used to start labels
186. ng is processed as option tokens and the default action is taken for both the listing and object files Empty white space or separator only strings that are passed to the options processor simple result in all of the default run time options to remain in effect In all cases once the options processor is invoked as a result of the first input string the options prompt will only be issued again due to an error or at the user request for help The next topic on alternative modes of assembler execution may help explain the reasoning behind the parser as it has been implemented in the M6800 assembler 6 6 2 Alternative Modes of Assembler Execution The Section 6 2 description of the assembler execution focused upon the interactive mode In this section two other modes of assembler operation within a command file and as a submitted job to the batch queue are discussed by example A command file is a sequence of commands that normally would be entered at the prompt in a VMS interactive session The advantage of command files is that often repeated multi stage tasks can be executed by simply typing one command file name In some sense a command file can be viewed as a macro to the command line interpreter of the VMS interactive shell Command files are also used to tailor the environment in preparation for the execution of a particular program Consider these two examples This command file is named DEBUG COM o Exclamation points ar
187. nificant lengths An invalid option error is reported unless n gt 0 The default label limit is 80 5 40 12 OPT LINENUMBER The LINENUMBER n option tells the assembler to set the line number of this line equal to n and to restart the line numbering sequence from this point The assembler accepts any 73 ASSEMBLER USER S GUIDE 16 bit value for the line number and it interprets the 16 bit value as an unsigned number The default initial line number is one but the user can start a new line number sequence as many times as desired in the program The initial line number may also be selected with the LIN LINe number run time option see Chapter 6 The default value for the LINENUM option is 1 5 40 13 OPT LIST NOLIST The LIST option tells the assembler to produce a symbolic listing of the program pro vided of course that the user has specified an output device or file for the listing and the NOLIST option suppresses any symbolic assembly listing The LIST and NOLIST options may be used in various portions of the program to generate and suppress lines from the symbolic output listing e g suppress previously debugged routines from being output The LIST NOLIST option uses listing levels to control the listing in a way that allows the programmer to have local listing control without affecting the overall listing When the listing level is negative the assembler suppresses lines of code from the symbolic listing The listing
188. o generate concordance entries for labels that are accessed invisibly by a nearby statement The format for the CONC statement is as follows CONC lt expression 1 gt lt expression 2 gt lt expression n gt The CONC statement requires at least one expression and it accepts any number of expressions beyond the first one In actual practice each expression is normally just a single label An example might be beneficial Suppose that a program maintains several variables COUNT ID LINK MACRO SEGMENT SIZE VALUE and WEIGHT Suppose furthermore that by design these variables reside together in a sequential array and that an initialization routine in the program is supposed to clear all of these variables The initialization routine could clear each variable individually by name or it could simply use a loop to clear the entire array of variables The following example code segment illustrates the use of a loop to clear the array of variables LDX ARRAY 1H CLR 0 X CONC COUNT ID LINK MACRO SEGMENT SIZE VALUE WEIGHT INX CPX ARRAY LENGTH BNE 1B The CLR instruction in the example clears each variable in turn but it doesn t use any of the variable names The CLR instruction therefore makes invisible references to 57 ASSEMBLER USER S GUIDE the variables and these invisible references don t show up in the concordance listing thus severely handicapping anyone who later tries to maintain or modify the program Th
189. od programmer however always includes an ORG statement at the beginning of a program for improved clarity and maintainability 5 42 PAGE The PAGE statement is equivalent to the EJECT statement and the PAGE statement is included only for compatibility with Motorola s definition of the assembly language Please refer to the description of the EJECT statement for more information 5 43 PRINT The PRINT statement is useful for displaying user controlled messages in the listing file In addition these messages will appear at the user s terminal in the course of the assembly The syntax of the PRINT statement is PRINT expression comment When the assembler encounters this statement in the assembly the assembler evaluates the expression and displays the expression value and the comment at the terminal during pass two The assembler prints the value of the expression in both hexadecimal and decimal If the assembler is generating a listing the output from the PRINT statement will also appear in the listing One possible use for this statement is in conjunction with conditional assembly For example if two mutually exclusive sections of code were both erroneously assembled in a large program a thoughtful positioning of a PRINT statement could flag the user that an error in conditional assembly logic had occurred An error code of several values via the expression field could further pinpoint the exact nature of the error 78 CHAPTER 5 A
190. ogrammers who don t use local labels usually end up using relatively meaningless symbolic labels like LOOP LOOP1 and LOOP2 or NEXT NEXT1 10 CHAPTER 2 SOURCE SYNTAX and NEXT2 Occasionally a desperate programmer without access to local labels even resorts to symbolic labels like SAM JOE NANCY and MARY As a general rule an instruction label that doesn t have global significance to the pro gram as a whole should probably be a local label Variable names of course should ordi narily be meaningful symbolic labels A good programmer typically uses a normal symbolic label to define the entry point of a routine and then the programmer uses local labels as local branch points within the routine The same local labels can be used again as local branch points within the next routine with no conflict As a matter of good style most programmers use local labels in ascending order starting with 1H rather than OH in each routine to enhance readability consistency and maintainability but the assembler doesn t enforce any such requirement The assembler has a special feature to guard against subtle logic errors that can occur when a user accidentally omits the definition of a local label Suppose that a programmer refers to the 3F local label and then forgets to define it With ordinary symbolic labels an omitted label definition would cause the assembler to report an error because of an undefined symbol With local labels however
191. onditional assembly are discussed in detail in another section 2 4 Opcode Field The second field of a source statement in assembly language for a microprocessor in the M6800 family is the opcode field The opcode field can contain one of the instruction mnemonics that is listed in the programmer s reference manual for the microprocessor or it can contain the mnemonic for an assembler directive The assembler directives which are sometimes known as pseudo instructions are all documented in another section of this manual Unlike some M6800 assemblers the M6800 family assembler requires the user to type the entire opcode mnemonic as a single string with no embedded spaces For example while some assemblers accept LDA A and LDA B as valid opcodes the M6800 family assembler requires LDAA and LDAB with no embedded spaces The idea of allowing a space as some assemblers do between the basic instruction mnemonic and the character that specifies the accumulator sounds appealing at first but it unfortunately introduces certain ambiguities For example consider the CLR instruction and the CLRA instruction The CLR instruction clears the memory location that is specified by the operand field and the CLRA instruction which has no operand field clears the microprocessor s AR i e its A register or A accumulator If the assembler allowed a programmer to use embedded spaces in an opcode mnemonic then a CLR A statemen
192. ore one byte shorter than the corresponding instruction with extended addressing and an instruction that uses direct addressing consequently executes more quickly than the same instruction with extended addressing Clearly direct addressing offers some important advantages over extended addressing but direct addressing can be used only for operands in page zero whereas extended addressing can be used for operands throughout the microprocessor s memory space The operand field for an instruction with direct addressing contains an expression that specifies the page zero memory address of the operand so the source syntax for direct addressing is essentially no different from the source syntax for extended addressing For example an instruction such as LDAB FLAG could use either direct or extended addressing since the source syntax in this case is the same for both addressing modes The assembler must choose between direct and extended addressing based on the value of the expression and other considerations as explained below The assembler automatically selects direct addressing instead of extended addressing if 1 direct addressing is available for the particular instruction that is specified in the opcode field and 2 the expression in the operand field specifies an operand address in page zero and 3 the expression in the operand field doesn t contain any forward references A forward reference is a symbolic label or a jF local label
193. ort list for an IMPRTB statement are byte sized imports with or without the B size qualifier 5 33 INCLUDE The INCLUDE statement requires a standard VMS file specification in its operand field and it tells the assembler to include the specified source file as part of the program that the assembler is assembling The specified source file effectively appears in the program imme diately after the INCLUDE statement The assembler treats source lines from an INCLUDE file exactly the same as lines from the main source file The INCLUDE statement itself does not appear in the listing so users may find it a good practice to make the first line of an included file a comment that identifies the included file The assembler permits nesting of INCLUDE statements to any depth but it does not support recursion at any level No included file can include itself nor can any nested include file contain an INCLUDE statement that refers to a file higher up in the nested hierarchy Invalid file names or nonexistent files file read system errors or the unavailability of virtual memory to process file control blocks are types of error conditions that the include processor reports See Chapter 10 for more details on these errors 5 34 END An END statement in an included file will terminate the assembly and the assembler will ignore any subsequent source lines in the main source file and in the possibly nested structure of include files 5 35 MACRO
194. orward to the target conditional label thus skipping the assembly of all interven ing lines The GOULZ statement provides conditional assembly but it is no longer recommended GOUNLS expression target Go unless If the expression is false i e zero go forward to the target conditional label thus skipping the assembly of all inter vening lines The GOUNLS statement is used with conditional assembly Continued 46 CHAPTER 5 ASSEMBLER DIRECTIVES Table 5 1 Continued Directive Parameter Remarks IMPORT labell Import the specified label s from other module s to make them available for use in this module The imported labels must be defined as exported labels by the other modules that will be linked with this mod ule The IMPORT statement can be used only in a relocatable assembly See EXPORT for re lated information IMPRTB labell Similar to the IMPORT directive except that all labels are 8 bit values INCLUDE file specification Include the specified source file as part of the program that is being assembled The specified source file effectively appears in the program immediately after the INCLUDE statement MACRO Define a macro with the specified name and the specified parameters Note that the macro name appears in the opcode field of the line following the MACRO statement MEND Macro end Terminate the definition of a macro
195. ositional param eters or the last parameter was delimited by a comma and we treated a comment as a parameter name 7US A label that appears in the operand field is undefined Undefined Symbol error A label that appears in the operand field is undefined A default value of zero is used 158 CHAPTER 10 ASSEMBLY ERRORS WS A nonblank character appears where white space is expected White Space error A field that must be terminated by white space is terminated by a nonblank character For example a user might erroneously say LDAA 0 X where the comma following the immediate value is invalid 159 ASSEMBLER USER S GUIDE 160 Chapter 11 Summary of Nonstandard Features The primary differences between this M6800 assembler and the standard Motorola defi nition of the M6800 assembler are listed below Some of the points are merely informative but should be of interest to most users Please note that any reference to the Motorola assembler refers to the M68SAM 6800 assembler that was produced by Motorola Incorpo rated Motorola has now produced other and better versions of their 6800 assembler and deficiencies that exist in the M68SAM Motorola assembler may not exist in newer Motorola assemblers e Immediate addressing can be specified by an as well as by the Motorola defined character e A quoted character in an expression can be specified for example as X Motorola s definition or as
196. ow forward references 4 1 7 Quoted ASCII Characters A single printable ASCII character that is enclosed by quotation marks or apostrophes e g A A or merely preceded by a quotation mark or an apostrophe e g A or A is always valid as a term of an expression The 16 bit value that the assembler uses for the quoted character is the right justified zero filled 8 bit ASCII code of the character with the parity bit either set or reset according to the parity option that is currently selected A programmer can easily specify the parity option and change it as often as desired by using the OPT PARITY expression assembler directive which is fully explained in the chapter that discusses assembler directives If a programmer never uses an OPT instruction to specify the desired parity the assembler uses zero parity by default If zero parity is being used then the 16 bit value of the term A for example is 0041 If mark parity is being used on the other hand then the 16 bit value of the term A is 00C1 A user normally specifies a quoted ASCII character in an expression either by enclosing the character with quotation marks e g A or by enclosing it with apostrophes e g A The assembler accepts the funny looking A and A forms primarily for the sake of maintaining compatibility with Motorola s original definition of the M6800 assembly language and most programmers no longer use those forms 4 1 8 Present
197. ows a casual or novice user to skip some chapters entirely and still be able to use the assembler proficiently at the desired level of understanding For example a beginning user can easily start using the assembler without reading about conditional assembly or macros Chapter 2 Source Syntax A source statement in assembly language for a microprocessor in the M6800 family can contain as many as four fields These fields are known from left to right as the label field the opcode field the operand field and the comment field The following line is a typical source statement that illustrates all four fields as they would be typed into a source file LABEL LDAA CHAR X Get the next character into the AR Any combination of horizontal tab s and or blank space s is known as white space and the assembler normally recognizes white space as a field delimiter The assembler also allows but does not require white space between the subfields of a multi part field as illustrated by the space that appears between the comma and the X in the operand field of the example statement above The assembler does not require white space after the last field of a line because the carriage return at the end of the line terminates the last field of the line The definition that was given above for white space isn t complete Besides recognizing horizontal tabs and blank spaces as white space the assembler also accepts a grave accent as a white space cha
198. perator produces the 16 bit one s complement of a term Note the difference between the NOT operator and the COM operator The NOT operator performs a logical operation whereas the COM operator performs a bitwise inversion The LOG2 operator produces the integer portion of the base 2 logarithm of a term If you express a value in binary the integer portion of the base 2 logarithm of the value is the bit number of the value s most significant bit The LOG2 operator is useful for example if you want to substitute a shift for a division or if you want to verify that an expression is a power of 2 LDAA L0G2 COUNT 1H LSRB DECA BGT 1B VERIFY 1 LSL LOG2 COUNT EQ COUNT 4 3 Parentheses The assembler computes the value of an expression from left to right with equal prece dence for all operators so a user doesn t need to memorize any arbitrary precedence rela tionships For example a user doesn t need to know the precedence relationships among the AND LSR GT MAX and MOD operators because all operators have equal precedence Although the assembler evaluates expressions from left to right with no operator prece dence rules a programmer can use parentheses to dictate any desired order of evaluation For example the expression 5 3 4 evaluates to 5 7 and then to 35 because the paren theses force the assembler to evaluate the parenthesized subexpression as a single term of the expression that con
199. pression YY ANDed with the conditional assembly pa rameter is not zero skip source lines until a source statement that has the system label 200 is found If the result of the YY expression value ANDed with the conditional assembly pa rameter was zero the GOULZ statement would have been ignored As a general rule a target for a conditional assembly statement cannot appear before the conditional assembly statement in the source program because the assembler skips only in a forward direction Furthermore since the assembler must know in pass 1 which statements it is to assemble the expression values in the GOIFZ and GOULZ statements must be defined in pass 1 Therefore the expressions in the operand fields of these instructions may not contain any forward references The following example illustrates the use of conditional assembly with the preferred mode in a program segment Of course conditional assembly could also be used to se lectively define tables of constants in a data segment corresponding to the specific device driver code being assembled in the example TAPE EQU 1 ROL 3 Bit 3 of the conditional assembly parameter corresponds to the symbol TAPE DISK EQU 1 ROL 4 Bit 4 of the conditional assembly parameter corresponds to the symbol DISK This block of code is common for the three versions of the program that may be produced in this program segment GOIF DISK AND If bit 4 of the conditional assembly CAMASK EQ 0 parame
200. properly into 8 bits either as a signed value or as an unsigned value The format for an FCB statement is as follows label FCB value_1 value_2 value_n The FCB statement requires at least one operand expression and it allows any number of operand expressions after the first one An FCB statement with multiple operand expressions is effectively a shorthand notation for a series of individual FCB statements so a single FCB statement with multiple operand expressions is equivalent to a series of individual FCB statements with the same set of operand expressions 60 CHAPTER 5 ASSEMBLER DIRECTIVES The assembler advances the location counter after it processes each expression of an FCB statement so the location counter symbol in an FCB statement always refers to the memory address of the data byte that is being generated by the expression that contains the location counter symbol If a label appears in the label field of an FCB statement the assembler defines that label s value to be the memory address of the first data byte that is generated by the FCB statement 5 22 FCBS The FCBS directive is like FCB but the FCBS statement specifies that the assembler should interpret and validate the expression value as a signed number 5 23 FCBU The FCBU directive is like FCB but the FCBU statement specifies that the assembler should interpret and validate the expression value as an unsigned number 5 24 FCC The FCC form
201. queue For example if an assembly job is quite involved and expensive the user can usually reduce costs by submitting the job to run during off peak hours A typical submit command for SAVINGS COM would look as follows SUBMIT NOPRINT AFTER 13 JAN 1986 02 35 SAVINGS COM The NOPRINT option keeps the log file which captures all output from batch job from being printed and deleted The AFTER date time parameter specifies when the job is to be executed Parameters can also be passed to batch files and the interested user is referred to the extensive literature The wary reader may wonder what would happen if the special assembler file name TT or its equivalent is specified as the source file within a command file or a batch job submission VMS clearly answers this question by defining two logical labels SYS INPUT and SYS O0UTPUT The assembler always gets information from SYS INPUT and this includes the source lines when TT is specified as the source file The assembler always outputs information to SYS QUTPUT In interactive mode both of these system symbols refer to the user s terminal as was discussed in Section 6 2 For a command file SYS INPUT is the command file itself see DEBUG COM but SYS OUTPUT remains the terminal In a batch job SYS INPUT is again the submitted command file but SYS OUTPUT becomes the log file that is associated with each batch process Hence one can only advise the user against using TT as the source in a
202. racter so white space actually consists of any combination of horizontal tab s and or blank space s and or grave accent s The assembler s recognition of a grave accent as a white space character provides a benefit that relates to compatibility with text formatters such as runoff and compose The advantages of using a grave accent as a white space character in certain situations will be discussed later Most programmers use one or more tab characters to separate fields in a source line because standard tab stops conveniently align the source file and ultimately the symbolic listing if any into neat columns Tabs are generally better than blank spaces as field delimiters because field alignment is more difficult for anyone who uses blank spaces as delimiters The assembler does have a FORMAT option that can provide automatic field alignment of the listing even if the source file is badly misaligned but the FORMAT option is ASSEMBLER USER S GUIDE intended for the special case in which a user needs to resurrect an antiquated source file that was originally produced in a primitive environment that didn t support tabs The FORMAT option will be described in detail later In the meantime the reader is advised that tabs are universally preferred as field delimiters Although the assembler is relatively flexible regarding columnization nearly all users adhere rigorously to standard columns as a matter of common practice The label field a
203. ram contain blocks of comments with the same kind of format that people use for ordinary text such as the text in a report or a business letter A good text formatter can easily handle the columnized instruction format as well as the normal text format so programmers might well consider the possibility of using text formatters to format their source programs Although a text formatter can easily provide the columnized format that is customary for assembly language instructions most programmers format their instructions manually while they type them because they can easily tab from one field to the next Most programmers also format their comments to one degree or another while they type them because many kinds of formatting are quick and easy with a modern screen oriented editor such as the E editor that is available on a VAX VMS system Some kinds of formatting are done most easily by text formatters though and a number of programmers use text formatters to polish the final appearance of their source programs For example a programmer might use a small portion of a text formatter s power simply to justify i e align the right margin in all of a program s comment blocks and a programmer might also use a text formatter to center headings or to format tables 2 8 1 Dot Commands A person who wants to use a text formatter to improve the format of a source program usually needs to include a few formatting commands in the program s source file
204. re does nothing If a SPACE command occurs immediately after an EJECT command however the SPACE instruction generates the requested number of blank lines at the top of the page This last feature allows a user to skip lines intentionally at the top of a page 5 59 SPC The SPC instruction is an abbreviated form of the SPACE pseudo instruction which is described above 5 60 TEMPLT The TEMPLT directive allows the user to build a template for a data structure The following example illustrates a simple use of the TEMPLT statement to build a template for a stack frame STKFRM EQU 0 The TEMPLT statement s argument TEMPLT STKFRM specifies the initial value for the SAVE_AR RMB 1 location counter in the template SAVE_BR RMB 1 block The initial value for the RET_ADR RMB 2 location counter is often zero A template defines the structure or organization of a data structure A template does not reserve memory for the data structure that it defines The assembler reports an error if you try to generate any object values in a template block The next ORG or TEMPLT or SECT statement terminates a template block Note that TEMPLT blocks are available in both absolute and relocatable assemblies 84 CHAPTER 5 ASSEMBLER DIRECTIVES A TEMPLT block is equivalent to the OFFSET block that is available with the 68000 family assembler As a convenience for users with 68000 experience the 6800 family assembler accepts OFFSET as a synonym for TEMPLT
205. re to break the expression before each operator The second statement in the example above illustrates the format that most users seem to prefer for continued expressions but the assembler has enough flexibility in this area to allow for a wide variation of individual preferences Notice that the assembler can tell the difference between a minus sign and a hyphen continuation character by the context in which the character appears White space or a carriage return immediately follows a hyphen continuation character but the character immediately following a minus sign is the next character of the expression or a hyphen continuation character In any case the character immediately following a minus sign is a nonwhite character Notice also that a user must not insert white space immediately before a hyphen contin uation character in an expression because white space preceding the hyphen would normally terminate the expression In general white space terminates an expression except in the case of a white space character that appears in an expression as a quoted ASCII character White space can of course appear before a hyphen continuation character that is used to break a statement between two fields such as the label field and the opcode field or between two subfields of the operand field The following examples illustrate some more useful applications of line continuation LABEL1 FDB EXPR1 This example shows a convenient method EXPR2 for
206. references are not allowed in a GOULZ conditional expression For more information regarding the GOULZ statement please refer to the chapter that describes conditional assembly 5 30 GOUNLS The GOUNLS go unless statement which allows a user to perform conditional assembly is the logical complement of the GOIF statement The format for the GOUNLS statement is as follows 64 CHAPTER 5 ASSEMBLER DIRECTIVES GOUNLS conditional_expression target If the conditional expression is true i e not equal to zero the assembler simply con tinues the assembly with the next line and the GOUNLS statement essentially has no effect If however the conditional expression is false i e zero the assembler goes forward to the target conditional label thus skipping the assembly of all intervening lines The assem bler must evaluate the GOUNLS conditional expression during the first pass of the assembly so forward references are not allowed in a GOUNLS conditional expression The GOUNLS statement isn t really necessary because a programmer can always use a GOIF statement to do anything that a GOUNLS statement can do The assembler supports the GOUNLS statement however because the reverse logic of the GOUNLS statement can be more clear than the GOIF statement s positive logic in some situations For more information regarding the GOUNLS statement please refer to the chapter that describes conditional assembly 5 31 IMPORT The
207. represents the first instruction in the downloaded program By now the user may have recognized a similarity between this example and the con ventional means of using the M6800 assembler in its default absolute object code generation 129 ASSEMBLER USER S GUIDE mode in conjunction with the monitor utility MUDBUG The key point to realize about absolute object code generation is that all address references must be completely defined at the end of the assembly process since the downloaded object code contains quite definite information for the monitor with respect to the microcomputer s address space Consider next the problems facing a group of individual programmers each developing a separate module of assembly code for a large system Quite clearly if each programmer hard wired the data structures and code into specific locations of the memory map then when the pieces were all put together collisions would almost surely occur A similar situation exists when developing libraries of routines that a user would like to be able to link in with some main program To accommodate this degree of freedom in the placement of code in the address space requires another step in the overall assembly process The assembler in relocatable mode defines object code addresses up to additive constant of their ultimate destinations It then becomes the job of a linker to take several modules of code and resolve ambiguities by assigning a specific value to
208. ructions and to use lowercase only in the comment field Except for the TABS option several options may be specified in a single OPT statement by simply separating one option field from the next by a comma or a comma and white space Op tion names that are under six characters must appear exactly as listed in Table 5 2 Option names given in Table 5 2 that are longer than six characters these are generally spelled out for the purposes of readability can be specified in any fashion as long as the first six characters of the name exactly match that given in the table If the assembler finds an invalid option it reports an invalid option error 10 and continues to process the option list The remainder of this section describes the various OPT assembly options and although it is not specifically mentioned in the following paragraphs each of the options selected must be in the operand field and an OPT pseudo opcode must be specified in the opcode field of the same statement An alphabetized table of OPT options is also provided as a quick and convenient reference Table 5 2 Summary of Assembler OPT Statements Option Remarks ABS REL Select between absolute or relocatable object code generation CALIST CANOLIST Increment or decrement respectively the conditional assembly list level Continued 68 CHAPTER 5 ASSEMBLER DIRECTIVES Table 5 2 Continued Option Remarks CMTCOL n DB8 DB10 DB16
209. s SO pera Wh a ts ae th ide 80 AS oe ogi ee aoa A aod BA ee PG a ee A Bee A pak aod 82 553 SETREG 6809 a SI A ee 82 Bib4 SK ua a A a a a 82 AN A 83 A tes Me ah 83 DOS RIP ena ve RAS A A AA a le a 83 DOSS PAGE A E SH A A A A ARE 84 BOO SPE e id ow alee Be Ae See ie a Bog 84 5 60 TEMPLE w a e egu Bed ot A ekg A et ac ee AA dale Bt 84 OLITE a hata Ea ei Mg A Te eo Po 85 B62 VERIFY aora th pis eek it by he wag a OA lama hE eke E 85 5 63 XDEEF 6 5 4 3 wah pase ed See ke ee a ae le eal bok we ed 86 DOA XREF y zri Sheed ats edt HB Ao Soe eae a a A pe aes 86 Running the Assembler 87 6 1 Getting Started 87 6 2 The Informed User ss d ri aa ee Wie ee ee ee a le 8 ae 88 6 3 Run Time Options 2 2 0 20 a 93 6 31 ABSOLUTE qu baci aan tle Pd ete hee eed de eo Pes 96 0 32 CAMASK se sad ii E ae ta bh a Ae Ban er ia 96 6 3 3 COMMENT usada Sete ih a a a A we a Bs 97 6 34 CROSS 20 Sth A AN a a ee eet tls 97 63 55 DEBUG ri ssai ioa a wba eb aa a BP ae 97 6 3 6 HOERRe o aurot d us el ent ean ie eee ee lata da 97 6 34 ERROR os wack goes a nha Swe aA ES AG ade als 98 6 3 9 PREBE Ds gh 2 A Ph Pn ee aA ik 98 ASSEMBLER USER S GUIDE 6 3 9 HELD Te dhe ake AD Se Be nee eee Gee Ang Berd be Ag 98 6 3 10 IGNORE inca Pk Pe a Qed Ai 99 63ALLANDENT o tn 99 CIAN LOA ui a A de eS Oe ta 100 6 3 13 LINE NUMBER e e 100 A AAA 100 623 15 HOC wait ah a A A o hd 101 63162 LRP cack ee ee ee Bie AGUS e
210. s a label even if the label happens to match a macro name an opcode mnemonic or some other predefined mnemonic Therefore a user doesn t need to remember any arbitrary list of reserved words that can t be used as labels White space terminates the label field and a user customarily types a single tab charac ter to advance from the label field to the opcode field A programmer who wants to define a label on a line by itself can use a carriage return to terminate the label field If a label appears on a line by itself or if CMTCOL columns of white space appear between the label and the next nonwhite character the assembler defines the label to have the value of the location counter A label that appears on a line by itself therefore receives the same value that it would have if an executable instruction appeared in the same statement with the label 2 3 2 Local Labels While a normal symbolic label can be defined only once in a program a local label can be defined over and over any number of times Local labels can therefore be used for local applications by several different sections of a single program and various sections that define and use the same local label don t conflict with each other A programmer defines a local label by using a jH j here entry in the label field where j is any digit from 0 through 9 and a programmer can refer to a local label by writing jB j backward or jF j forward as a term of an expression For e
211. s desired for any given assembly so the assembler doesn t waste any time or resources performing assembly options that aren t needed The option specifications are mostly one to three character mnemonics and they may be entered in any mixture of uppercase and lowercase The various option specifications may be typed in any order but multiple option specifications must be separated from each other by commas Spaces and tabs are permitted anywhere within the option request fields but the option request line may not exceed 132 characters and blanks and tabs are counted as input characters The user is therefore warned against an extravagant use of blanks and tabs when numerous options are going to be selected The option mnemonics themselves have mandatory and optional entries For example to set the beginning line number of an assembly listing the user may specify LIN n LINe n LINe n n LINe nu n LINe num n LINe numb n LINe numbe n or LINe number n All of these entries are acceptable to the assembler for this option as long as the mandatory characters LIN are included The mandatory characters are shown in uppercase and the optional characters are shown in lowercase Basically two classes of run time options exist One class of options require a number to be input with them e g line number page number lines per page conditional assembly 93 ASSEMBLER USER S GUIDE mask etc For these options
212. s relocatability is therefore severely restricted and the base section is guaranteed to reside entirely in the memory area that extends from location 0000 through location 00FF For more information regarding the BSCT statement please refer to the chapter that describes relocatable assemblies 5 8 BSS The BSS block starting at symbol directive reserves and names but does not back ground a specified amount of memory for a variable a table or some other storage area The format for a BSS statement is as follows label BSS block_size The expression in the operand field specifies the number of bytes to be reserved Since the assembler must evaluate this expression during the first pass of the assembly the ex pression must not contain any forward references The assembler effectively reserves the specified amount of memory by adding the value of the expression to the location counter thus skipping the designated number of bytes in memory If a label is present in the label field of a BSS statement the assembler defines the label to have the original value of the location counter so the label represents the memory address of the first byte of the reserved area The assembler does not generate any object values for the byte s that it skips and it doesn t mark the bytes as being permanently reserved in any way A programmer can therefore use BSS statements not only to reserve read write memory for variables but also to construct te
213. s the user desires and a macro call will result in expansion of the most recent definition If a macro name is the same as one of the native opcode mnemonics the macro has precedence over the opcode Note that macro names are stored in a different list from normal symbols so the program can define and use a macro of the same name as a label without any conflict or ambiguity The macro name belongs in the opcode mnemonic field In the macro definition the macro name should be in the opcode field of the first line following the MACRO pseudo op which is described in Chapter5 The line with the macro name must not be labeled Besides the fact that a label would be meaningless the assembler would interpret the label field as the macro name and the macro name as a formal parameter name 8 2 Macro Parameters The macro assembler recognizes two types of parameters Formal parameters are named and positional parameters are recognized only by their position in a macro call A macro can have no parameters positional parameters or formal parameters but formal and positional parameters are not recognized for the same macro Regardless of the type of parameter used the assembler will replace each occurrence of a parameter name in the macro model with the substitution string specified in the macro call If the macro call model statement in the definition contains a parameter name within CMTCOL columns after the macro name then the macro uses formal p
214. s to divide 8000 by 1 thus generating a result that doesn t fit correctly into 16 bits We also report this same error when a user divides by zero in an expression In either case the assembler ignores the division and retains the original expression value DI Direct addressing is not available with the given instruction DIrect addressing error Direct addressing is requested by the use of an to start the operand field and direct addressing is not available with the given instruction DP The operand value exceeds 8 bits Direct Page register error The format of the operand expression for a DPR pseudo instruction is incorrect The proper format is 00XX where XX is the low byte of the operand expression and denotes the value that the user wants for the direct page register The high byte of the operand expression must always be zero 7E0 An expanded macro line contains too many characters macro Expansion Overflow error The source line buffer SRCLIN has overflowed during macro expansion because the expanded line is too long Perhaps only comments remain on the current line but we report an error to draw the user s attention to a possible problem 149 ASSEMBLER USER S GUIDE 7ER A right parenthesis appears with no matching left parenthesis Extra Right parenthesis A character that functions as a right parenthesis appears in an expression with no matching character that functions as a left parenthesis
215. single expression can be continued over as many lines as desired so the length of an expression is unlimited The assembler does not allow statement continuation in the middle of a token where a token is a label an opcode a single term of an expression an operator in an expression a quoted string a delimited macro argument or a keyword such as an option mnemonic The assembler also disallows statement continuation in the comment field or anywhere after the end of the last field that precedes the comment field Continuation of the comment field is unnecessary because the assembler s CMTCOL feature which was explained earlier effectively makes the continuation of comments totally automatic Incidentally the CMTCOL feature doesn t apply to the white space if any at the begin ning of a line that is a continuation of the previous line The assembler s CMTCOL feature applies only at the beginning of a statement not at the beginning of a line that is merely a continuation of a statement A programmer can therefore indent as far as desired at the be ginning of a continuation line and the conventional practice is to indent each continuation line to the standard starting column for the field that is currently being typed If white space rather than a carriage return immediately follows the hyphen continu ation character then the assembler considers the rest of the line following the white space to be a comment The significant part of
216. specifies a 16 bit logical exclusive OR operation The assembler uses mnemonic tokens such as AND OR and XOR instead of using special characters for operators beyond the simple arithmetic operators because the mnemonic tokens are easier for people to understand and remember For example XOR is easy to understand and remember but people might have trouble trying to remember a particular combination of special characters such as to specify an exclusive OR operation As a typical example of a good use of a logical operator consider a situation in which a person wants to load the address not the contents of location VECTOR into the AR and the BR with the AR receiving the most significant byte of the address of location VECTOR and with the BR receiving the least significant byte of the address of location VECTOR A programmer could write LDAB VECTOR AND 00FF as an appropriate instruction to get the least significant byte of the address of location VECTOR into the BR The AND operation eliminates the most significant byte of the address of location VECTOR and it therefore leaves the least significant byte of the memory address as the value of the expression 4 2 3 Shift and Rotate Operators The assembler accepts operators for both arithmetic and logical shifts and it also rec ognizes operators that specify rotations either to the left or to the right In each case the assembler uses the 16 bit value of the term immediatel
217. specify either the DB8 DB10 or the DB16 option and the generated pass 2 values will then be output in base 8 octal base 10 decimal or base 16 hexadecimal character format respectively If none of the options are specified the assembler will automatically output the values in hexadecimal The display base may be changed as often as desired in the program Some programmers who have trouble figuring out displacement values in hexadecimal may find the decimal base selection as an aid in debugging but it is not recommended since most programmers are expected to become experts in binary octal and hexadecimal arithmetic The value of the location counter is always printed in hexadecimal regardless of the display base but object values are printed according to the display base selected The default for the display base option is base 16 i e hexadecimal 5 40 5 OPT DEBUG By using this option the user can include a symbol table in the generated object file Each entry in the symbol table consists of an ASCII string label and its corresponding 71 ASSEMBLER USER S GUIDE numeric value Thus a debugging tool e g MUDBUG can help the programmer by providing labels in place of magic numbers This debugging option is available in both absolute and relocatable modes 5 40 6 OPT DIRECT NODIRECT Under certain conditions the user may need to restrict the assembler from generating object code that uses the direct addressin
218. statement only for purposes of backward compatibility The format of the GOIFZ statement is as follows GOIFZ conditional_expression target If the conditional expression ANDed with the assembler s CAMASK parameter is nonzero the assembler simply continues the assembly with the next line and the GOIFZ statement essentially has no effect If on the other hand the conditional expression ANDed with the CAMASK parameter is zero the assembler goes forward to the target conditional label thus skipping the assembly of all intervening lines The CAMASK parameter is a 16 bit value that a programmer can dynamically specify as a run time parameter when the programmer invokes the assembler The CAMASK parameter is discussed in the section that explains the assembler s run time options The assembler must evaluate the conditional expression in a GOIFZ statement during the first pass of the assembly so forward references are not allowed in a GOIFZ conditional expression For more information regarding the GOIFZ statement please refer to the chapter that describes conditional assembly 5 28 GOTO Like the GOIF statement the GOTO statement is used to skip a section of source code during the assembly process Unlike the GOIF statement however the GOTO statement unconditionally skips a section of source code The format of the GOTO statement is as follows GOTO target The target value in the GOTO statement s operand field is a 16 bit hexad
219. struction is the same value that is gen erated for the first byte of a two byte BITA instruction with immediate addressing When the SK1 instruction is executed the microprocessor takes the following byte as the second 82 CHAPTER 5 ASSEMBLER DIRECTIVES byte of a BITA instruction with immediate addressing The net effect is to skip one byte As a side effect the SK1 instruction may modify the N bit the Z bit and the V bit of the condition codes 5 55 SK2 The SK2 SKip 2 bytes instruction tells the assembler to generate an instruction that causes the next two memory bytes to be skipped at execution time The SK2 instruction is similar to the SK1 instruction and it can be used to save the programmer one byte of memory For example the following instruction sequences execute the same but the code on the right uses one less byte of memory BRA 1F SK2 LOOP ADDB 4 LOOP ADDB 4 1H STAB 2 X STAB 2 X The opcode value 8C that is generated for a SK2 instruction is the same value that is generated for the first byte of a three byte CPX instruction with immediate addressing When the SK2 instruction is executed therefore the microprocessor takes the following two bytes as the second and third bytes of a CPX instruction with immediate addressing The net effect is to skip two bytes The SK2 instruction may therefore be used to skip two single byte instructions or one double byte instruction during program execution As a side ef
220. sued with an explanatory note until a nonempty input is obtained The first token that appears in the input string is taken to be the source file If no separator follows this token the default action as described in Section 6 2 is taken with respect to the listing and object files and then the options prompt is issued and processed as usual If all that follows the first token is only a separator then it is assumed that the default options are acceptable and the assembly begins immediately Otherwise the next action depends upon which separator is found and the next two paragraphs exhaustively trace the possible outcomes of each path of this bifurcation If the separator is a comma then the next token with leading white space stripped pertains to the listing file though recall that as a positional parameter comma white space comma specifies default action for the listing file If nothing remains in the input string the default action for the object file is taken and the options prompt is issued If only a separator remains then the options prompt is foregone and assembly begins immediately 107 ASSEMBLER USER S GUIDE If a token follows this separator it is associated with the object file It is at this point where the level two description ends however if more characters are yet to be examined in the input string they will be passed as tokens to the options processor If the separator is a virgule the remaining stri
221. t expansion of lines that will be skipped anyway The MEXIT instruction must not be labeled and it is only valid in a macro definition If a MEXIT instruction is labeled or if a MEXIT occurs outside of a macro definition the assembler will report a non fatal error Refer to Chapter 8 for information on the use of macros 5 38 MON The MON instruction is equivalent to the END statement and the MON statement is included only for compatibility with the Motorola definition of the assembly language Please refer to the description of the END statement for more information 5 39 NAME The NAME statement is equivalent to the TITLE statement and the NAME statement is included only for compatibility with the Motorola definition of the assembly language Please refer to the description of the TITLE statement for more information 5 40 OPT The OPT pseudo instruction is used to set or reset various assembly options during pass 1 and or pass 2 of the assembly Some of the options are also selectable during run time processing at the user interface level as described in Chapter 6 An OPT command in a source program however will override any run time option selected by the user if the options are in conflict All of the options are initially set to their most commonly used values by the assembler and so most users don t need to worry about setting options during assembly of their programs When an option is needed in a program and the program will a
222. t appears in the operand field is a multiply defined symbol That is a label that is used in the current line is defined in at least two different places so the value that is used for the label may not be the value that is desired 7MT An expression is missing a term or is otherwise malformed Missing Term error An expression ends with an operator that is not followed by a term or else an expression field is completely empty Perhaps a blank space was inadvertently typed between an operator and the next term of an expression or perhaps two consecutive commas were inadvertently typed Another common cause of this error is the presence of a comma after the last expression field of an instruction that can have multiple expression fields The comma indicates the presence of another expression field but the last expression field is missing The assembler uses zero as a default value for any term that is missing MX MEND and MEXIT are invalid except in a macro Illegal Macro termination error A macro end MEND or macro exit MEXIT pseudo op appears outside a macro definition i e in the normal program flow 7MY A macro definition is different in pass 2 synchronization error Macro sYnchronization error On the second pass we were unable to match the macro definition with a macro node with the same information 7NL The nesting level for parentheses is too deep Nesting Level error The nesting level for parentheses in an expressio
223. t could be either 1 a CLR instruction that clears a location that has the label A or 2 a CLRA instruction The CLR instruction is just one example of an entire class of ambiguities Obviously the assembler could resolve each ambiguity by imposing some arbitrary rules and restrictions but special rules and restrictions are undesirable because they are unnatural and hard for users to remember Therefore the M6800 family assembler eliminates the ambiguities entirely by simply requiring the user to type the opcode mnemonic as a contiguous string with no embedded spaces This single requirement is both natural and easy to remember Like the label field the opcode field can be terminated by white space A programmer ordinarily uses a single tab to advance from the opcode field to the operand field if the 12 CHAPTER 2 SOURCE SYNTAX instruction has an operand field Some instructions such as INX and DEX however don t have operand fields For an instruction without an operand field the opcode field can be terminated by a carriage return if the user doesn t want a comment field More commonly a programmer uses three tabs to advance from the opcode field of an instruction with no operand field directly to the comment field The assembler knows which instructions don t have operand fields of course so even a single column of white space is sufficient to advance from the opcode field to the comment field for an instruction with no oper
224. t to automatically purge any files of a set in excess of the top three version numbers Files themselves are organized into structures known as directories books organized on a shelf or folders in a file cabinet drawer may be useful analogies Most beginning users of VMS wakeup in their home directories when they login to the computer If that is the directory from which the M6800 assembler is invoked then any listing or object files will also appear in that directory and that is where the assembler searches for the file that was named as the input source An educated user can easily override this default action by careful inclusion of the directory part of the complete file identifier A user can protect both files and directories with respect to such functions as reading writing and executing The default protection that the system provides each user should be sufficient for most purposes However the user with sensitive information can more closely regulate his or her files using the SET PROTECTION command as always on line help is available for this command Be advised that the assembler requires read permission on any source file that it is to process Generally the M6800 assembler is invoked within an interactive session in VMS other modes of operation are discussed in Section 6 6 In this situation the user issues the ASM command as a response to the prompt from the VMS interactive command line interpreter NOTE the ASM
225. table assembly only start using the location counter for the absolute section BHS destination Branch if higher or same This instruction which is just another name for the BCC in struction is the unsigned version of BGE BLO destination Branch if lower This instruction which is just another name for the BCS instruction is the unsigned version of BLT BSCT For relocatable assembly only start using the location counter for the base section BSS block_size Block starting at symbol Skip i e reserve block_size bytes and name the first one label BTEXT Begin text Treat the source lines between this statement and the next ETEXT statement as comment lines except for EJECT PAGE SPACE SPC SBTTL and SBTTLE statements CHKSUM target_value Generate a checksum data byte that can be added to the sum of all object bytes since the previous CHKSUM statement if any to yield the specified target_value CONC labell Create a concordance entry to indicate a ref erence for each specified label The CONC statement is used to compensate for invisi ble references to labels CSCT For relocatable assembly only start using the location counter for the common section DSCT For relocatable assembly only start using the location counter for the data section Continued 44 CHAPTER 5 ASSEMBLER DIRECTIVES Table 5 1 Continued Directive P
226. tains the parenthesized subexpression The assembler allows a programmer to use four different styles of parentheses In addi tion to accepting regular parentheses the assembler also accepts square brackets and braces and and angular brackets lt and gt Since the assembler accepts multiple styles of parentheses we use the generic term parenthesis in this section to mean any character that the assembler accepts as a logical parenthesis not just a or character The assembler s multiple styles of parentheses are valuable because they allow a programmer to enhance the readability of a complex expression that contains nested parentheses For example consider the following expressions 8 5 2 7 4 9 3 4 2 2 3x 2 1 7 6 442 5 Al ASSEMBLER USER S GUIDE 8x 54 2 7 4 9 3 lt 4 2 gt 4 2 3 2 1 7 6x 4 2 5 The two example expressions are functionally identical but the second expression con tains multiple styles of parentheses that improve its readability While the second expression is still difficult to read correctly it isn t nearly as bad as the first expression The assembler checks each closing parenthesis to verify that it matches the style of the corresponding opening parenthesis so the multiple styles of parentheses provide some addi tional error checking that wouldn t be possible with only a single style of parentheses For example the assembler reports
227. ted correctly during pass 1 and the assembler obviously can t evaluate the expression completely when it contains an imported label 77ING The 6502 doesn t work right when the pointer address ends with FF The user has specified a 6502 JMP instruction with the 1234 indirect addressing mode and the address of the indirect pointer is of the form xxFF The 6502 believe it or not doesn t work right with such an instruction so we warn the user The 6502 correctly gets the LSB of the indirect address from location xxFF but it incorrectly gets the MSB of the indirect address from location xx00 instead of getting it from location xxFF 1 The processors in the 65150 group don t have this error IN The specified file could not be included as source input INclude error An error has occurred while attempting to access a file as source for an INCLUDE statement Probably the file name in the operand field is in error though perhaps a system error occurred on opening or connecting to this file it would also have been reported in this case Another possibility is that the file was already opened i e an improper nesting of INCLUDE statements or that no virtual memory was available to accommodate the Include control blocks in the front end 10 An invalid option or option range has been specified Invalid Option error An invalid option mnemonic is specified or an option value is outside the allowed range for its field 151
228. ter is zero skip statements until 770 finding the system label 770 113 ASSEMBLER USER S GUIDE This block of code is assembled only if bit 4 of the conditional assembly parameter is set Otherwise it is skipped This code therefore corresponds to the DISK version of the program GOTO 1000 Skip the other versions 770 GOIF TAPE AND If bit 3 of the conditional assembly CAMASK parameter is set skip all statements until 900 finding the system label 900 This block of code is assembled only if bits 3 and 4 of the conditional assembly parameter are zero This code therefore corresponds to the nonDISK and nonTAPE version of the program GOTO 1000 Skip the TAPE portion of the program 900 This block of code is assembled if bit 4 of the conditional assembly parameter is not set but bit 3 is set This portion therefore corresponds to the nonDISK TAPE portion of the program 1000 This is a dummy target line for all three routines to enter and the rest of the program consists of more common code END Since the conditional assembly parameter contains 16 bits and the GOUNLS and GOIF pseudo ops involve arbitrary operators not necessarily AND as in the example in expres sions several different conditional variables can appear in the same assembly and condi tional assembly can become quite a bit more complex than this simple example indicates For the sake of completeness the critical lines of the e
229. that hasn t already been defined to have a particular value by a previous statement in the source program 23 ASSEMBLER USER S GUIDE The assembler must establish the length or size of each instruction during pass 1 of the assembly so the assembler must choose between direct and extended addressing during pass 1 If an operand expression contains a forward reference the assembler obviously can t evaluate the expression completely during the first pass of the assembly Therefore the assembler can t reliably determine during pass 1 whether or not an operand with a forward reference resides in page zero and the only safe course for an operand expression that contains a forward reference is to choose extended addressing If the operand with the forward reference later turns out to be in page zero the extended addressing is slightly wasteful but at least works correctly Consider the alternative If the assembler had chosen direct addressing during pass 1 and the operand with the forward reference later turned out to reside somewhere outside page zero the instruction couldn t address the desired operand and therefore wouldn t even work At first glance forward references might seem to be a potential source of significant irritation since they can possibly cause the assembler to select extended addressing when direct addressing would work more efficiently This undesirable situation seldom occurs in actual practice however because
230. the base address of the abstract address sections that all relocatable addresses are related to by assembler defined offsets One such linker is the LK6800 and the exact nuances of relocation as implemented on the M6800 assembler are defined by the requirements and features of this specific software product These features will be more completely described in this chapter The key to remember however about relocatable assembly is that addresses will be exactly defined at linkage time though they must be defined to within an arbitrary additive constant at assembly time Some users may wonder where the concept of position independent code falls within this discussion since it often accompanies the relocation concept in conversation If the simple microprocessor described above is expanded into a multi user environment perhaps with virtual memory then a case could be made for requiring code that could be executed from any position in the address space In particular for code to be sharable it must be position independent The key here is that all address references can only be made with respect to some value usually contained in the program counter or perhaps the stack pointer that is finally defined at execution time By keeping all address references ambiguous to this degree allows the loader that piece of software that deposits a program in memory and transfers control to it such as the VMS RUN command complete freedom in assigning the code
231. the source lines that begin with a period in column one in the listing output Recall that these dot commands are often associated with text formatters that the user may employ to justify and align blocks of text in the listing As the word ignore suggests when this option is in effect the dot lines do not appear in the listing They may of course be included by specifying the negated variant NIGNore of this option The analogous action within the assembler directive OPT regime is also possible with the IGNORE and NOIGNO switches The default action of IGNore is to suppress listing of period initiated lines 6 3 11 INDENT The INdent option allows the user to set the left print margin for his or her output listing The listing may not be indented to such an extent that no meaningful output will result Therefore the maximum indentation allowed must provide for 29 print columns i e INDENT 29 lt WIDTH If the selected output width is 80 columns for example the maximum indentation is 51 since 80 29 51 The INdent option in conjunction with the Width option allows the user to set the left and right output margins It should be noted however that any line exceeding the right margin will be truncated This feature is useful when the user wants assembled values and source code without any comments For 99 ASSEMBLER USER S GUIDE example a user might want to display the listing file at a CRT terminal with no co
232. the statement continues with the first nonwhite character of the next line The first nonwhite character of the next line can even be another hyphen followed by white space in which case the statement continues again to the next line Most programmers use statement continuation with long operand fields to prevent the operand field from extending into the columns that are normally used for the comment field Consider the following examples 15 ASSEMBLER USER S GUIDE LABEL LDX TABLE ENTRY_SIZE ENTRY_NUMBER F IELD_OFFSET ADJUSTMENT LABEL LDX TABLE The comment field is independent of line ENTRY_SIZE continuation so programmers who use line ENTRY_NUMBER continuation to continue long statements FIELD_OFFSET can conveniently type their comments just ADJUSTMENT as they would ordinarily type them LABEL LDX TABLE This statement like the ones above sets ENTRY_SIZE the XR to point to an individual byte of a ENTRY_NUMBER particular field of a specified entry of a FIELD_OFFSET sequential table Each entry in the table ADJUSTMENT contains ENTRY_SIZE bytes The three statements above are all functionally equivalent to each other The first state ment doesn t use the continuation feature and its operand field extends into the columns that are normally used for comments The second statement uses the continuation feature to break the expression after each operator and the third statement uses the continuation featu
233. the style of the first argument in the macro call determines the style for the rest of the string substitutions The assembler does not support a mixed mode of argument specification with respect to formal parameters A macro call for a positional parameter macro is of the form lt macro_name gt lt string gt lt string gt where each string is assigned to the positional parameter at the associated index The first string is assigned to positional parameter 1 the second string is assigned to positional parameter 2 and so forth The assembler will assign the null string to any parameters for which the macro call did not specify a substitution string In that case the assembler will report a missing parameter error see Chapter 10 unless the user has suppressed MP error messages see Chapters 5 and 6 Sometimes it is appropriate to assign the null string for a parameter The syntax lt gt specifies the null string The user may also use a comma to delimit the end of a null string For example the following methods assign three null strings for the given positional parameter macro TEST lt gt lt gt lt gt TEST gt The following calls produce the same result as the two calls listed above but the first one generates a too many parameters error message and the second generates a missing parameter error message Both of these error messages may be suppressed see Chapters 5 and 6 TEST NE Extra comma delimits
234. the value 28 x 40 21 x 40 19 x 40 or 28 x 1600 21 x 40 19 x 1 or 45 659 or B25B If a RAD 40 hash string contains fewer than three label characters the assembler effec tively left justifies the RAD 40 character codes in the radix 40 number to obtain the hash value of the shorter RAD 40 string For example the two character RAD 40 hash string PI specifies a term with the value 28 x 40 21 40t 0 x 40 or 28 x 1600 21 x 40 or 45 640 or B248 If a RAD 40 hash string contains more than three label characters the assembler effectively ignores all characters beyond the first three Why does the assembler use radix 40 as the basis for an encoding scheme The reason is simple The largest possible three place base 40 value is 39 40 39 x 40t 39 x 40 or 39 x 1600 39 x 40 39x 1 or 63 999 or F9FF so the value for any three character RAD 40 hash string fits into a 16 bit field as an unsigned value On the other hand the 31 ASSEMBLER USER S GUIDE largest possible three place base 41 value is 68 920 or 10D38 which exceeds the capacity of a 16 bit unsigned value Radix 40 is therefore the largest radix we can use if we want to code a three place value into 16 bits Obviously any encoding scheme needs one value usually zero to represent a null character so we are left with the ability to encode a three character string from any set of 39 characters into a 16 bit unsigned value The RAD 40 code va
235. these options Those options that may also be selected in the source code via the OPT assembler directive are noted Table 6 1 Summary of Run Time Options Option Remarks ABSolute Generate absolute object code The assembler generates ab solute object code by default though relocatable assembly can be selected via RELocatable CAmask hhhh Set the conditional assembly parameter to the 16 bit hex adecimal value hhhh COMment NCOMment Include all comment lines in the listing output This option does not apply to any comments that appear in the comment field of a source statement Cross NCross Generate a cross reference or concordance and include it in the listing output at the end of pass 2 of the assembly Elerr NEOerr Report expansion overflow errors that when a macro ex pansion exceeds the maximum line length in the model definition Error NError Report assembly errors to both the terminal and the listing file but only once if they coincide Continued 94 CHAPTER 6 RUNNING THE ASSEMBLER Table 6 1 Continued Option Remarks FFeed NFFeed Help NHelp IGNore NIGNore INdent n LCA NLCA LINenumber n List NList LOC NLOC LPP n MPerr NMPerr MPU xxxx MShow NMShow Page n RELocatable Report NReport Generate page breaks using the form feed character This op tion should be used unless the output device does not support form feeds Inform the
236. tially defined on a subsequent line in the program is a forward reference A label is similarly a forward reference in the case of a relocatable assembly if the label is declared as a label that is defined in another module Recall that the assembler allows a programmer to define and use any label that syntac tically qualifies as a label even if the label happens to match a macro name an opcode mnemonic or some other predefined mnemonic For example the instruction LDAA X X has the label X in the first subfield of its operand field and it has the predefined mnemonic X in the second subfield of its operand field The assembler knows what belongs in each subfield so it properly uses the label X as an offset value with indexed addressing just as the programmer intended In a related example the assembler correctly interprets the instruction DEC DEC to be a DEC instruction that decrements the value at location DEC 4 1 6 Local Labels The assembler accepts a local label as a term of an expression if the statement that defines the local label isn t too far away from the expression that uses the local label The assembler considers a local label s definition to be too far away from a reference to the local label when the definition is more than 104 lines about two full pages away from the reference The number 104 is merely the assembler s standard default limit and a programmer can conveniently redefine the local label separat
237. tion 6 3 4 CROSS The Cross option tells the assembler to generate a cross reference listing and output it to the listing device file at the end of pass 2 of the assembly The NCross option suppresses the outputting of the cross reference listing which also suppresses the undefined and un referenced symbols listings The cross reference listing is a valuable tool for debugging a program and it is almost mandatory for determining all the places in a program that refer to use or modify a memory location The default is the Cross option 6 3 5 DEBUG By using this run time option the user can include a symbol table in the generated object file Each entry in the symbol table consists of an ASCII string label and its corresponding numeric value Thus a debugging tool e g MUDBUG can help the programmer by providing labels in place of magic numbers This debugging option is available in both absolute and relocatable modes 6 3 6 EOERR By default the assembler reports an error if the expanded macro line overflows the source line buffer i e exceeds 126 characters This error occurs when the length of the substitution string exceeds the length of the parameter name and the macro model line did not have enough unused spaces at the end to absorb the difference For example if 1 is replaced by ALPHA BRAVO CHARLIE DELTA the resulting macro expanded line is going to contain more characters than the model line does 97 A
238. tion may speed up output processing considerably especially for long listings The NFFeed option is necessary when the listing file is to be listed at a device that does not support form feeds The default is the FFeed option 6 3 9 HELP The Help option can be selected by the user who while in the context of the options processor would like to be briefly refreshed on the default states and correct syntax of the run time options After the help message is delivered the options prompt is again issued and another input string is processed If an error was detected in the input string to the options processor regardless of the state of the Help flag this same action will occur The default state of this option is to offer help only when requested or when a run time option error has occurred The help message as it appears at the user s terminal follows These are the options available for the front end interface 98 CHAPTER 6 RUNNING THE ASSEMBLER With Parameters Default values CAmask 0 MPU 6800 INdent 0 Page 1 LINenumber 1 Width 132 LPP 66 As Boolean State Flags an N prefix toggles flag ABSolute LCA SEQuence COMment List SLine Cross LOC SRC Elerr MPerr Symbol Error MShow TMerr FFeed RELocatable TOC IGNore Report Value Entries marked with are active by default Help yields this message 6 3 10 IGNORE The IGNore run time option controls the appearance of
239. tion to each program that requires space in DSCT DSCT is generally used to contain variables which are in RAM and are to be accessed using the extended addressing mode PSCT or program section is a relocatable section There is only one program section PSCT is similar to DSCT However it is generally used to contain program instructions The use of DSCT and PSCT facilitates creation of programs that reside in ROM but access variables in RAM The use of these section swapping pseudo ops is only supported in relocatable mode and the assembler displays RE errors when it encounters them in absolute mode In addition to the five section types defined above relocatable assembly permits the use of two label definition constructs that enable location and offset information to be passed between separately assembled modules An EXPORT or equivalently XDEF instruction is used to declare an externally defined symbol a symbol that may be referenced by any separately assembled relocatable program The general form for the EXPORT instruction is EXPORT SYMBOL1 SYMBOL2 SYMBOL3 The EXPORT statement must not be labeled and the operand field must contain at least one symbol The user may specify several externally defined symbols in one EXPORT instruc tion by typing a comma immediately following each symbol the assembler also accepts blank spaces and or tabs between the delimiter comma and the first character of next symbol Each symbo
240. tional expression is true i e nonzero the assembler goes forward to the target conditional label thus skipping the assembly of all intervening lines The target value in the GOIF statement is a 16 bit hexadecimal number that matches the value of a target conditional label that appears in the label field of a subsequent target line The assembler requires the use of hexadecimal numbers for conditional labels much the same as FORTRAN requires the use of decimal numbers for statement labels Obviously the assembler must evaluate the GOIF conditional expression during the first pass of the assembly so forward references are not allowed in a GOIF conditional expres sion For more information regarding the GOIF statement please refer to the chapter that describes conditional assembly 5 27 GOIFZ Like the GOIF statement the GOIFZ go if zero statement allows a user to perform conditional assembly by skipping a conditional section of source code or including the con ditional section of source code as part of the program during the assembly process The 62 CHAPTER 5 ASSEMBLER DIRECTIVES GOIFZ statement was the primary vehicle for performing conditional assembly before the GOIF statement was added to the assembler s repertoire but the GOIFZ statement is no longer recommended for use Anything that can be done with the GOIFZ statement can be done more clearly with the GOIF statement and the assembler still recognizes the old GOIFZ
241. trates the use of positional parameters MACRO Example usage of positional parameters TEST ADDA M Add 1 to the A register and store the STAA 2 result in 2 MEND When the assembler reads the definition above it stores the information needed to handle macro calls for the macro named TEST The assembler saves the model lines and the number of parameters 2 The following illustrates a macro call and the resulting expansion String syntax is explained later in this chapter TEST COUNT lt 0 X gt ADDA COUNT Add COUNT to the A register and store the STAA 0 X result in 0 X When the assembler encounters the macro name TEST in the opcode field it gathers the parameter substitution strings from the call and substitutes those strings into each line of the model as the assembler expands each line Notice that wherever the assembler found 1 in the model it substituted the first string COUNT and 2 was replaced by the second string 0 X It does not matter what field the positional parameter index appears in In this example parameter indices appear in the operand field and in the comment field 119 ASSEMBLER USER S GUIDE 8 2 2 Formal Parameters The use of formal parameter names can make the macro definition easier to read and provides the capability of specifying default parameter substitution strings A parameter name is an alphanumeric string that must begin with an alphabetic character All char acters of the
242. uate a label to the CAMASK term and then use the label instead of using the CAMASK term For example a user might say CAMASK EQU CAMASK and then use the CAMASK label instead of using the CAMASK token 4 1 11 Parenthesized Subexpression A complete expression enclosed by parentheses is known as a parenthesized subexpres sion and a parenthesized subexpression is valid as a term in an expression A programmer can use parentheses to dictate the order of evaluation in an expression and a programmer can also use parentheses to add clarity to an expression Another section in this chapter explains the use of parentheses in detail 4 2 Operators The assembler has an extensive set of operators that can be used in expressions For example besides recognizing the traditional arithmetic operators the assembler also recog nizes logical operators shift and rotate operators signed and unsigned relational operators and a few miscellaneous operators Table 4 2 and Table 4 3 summarize the binary and unary operators respectively The following paragraphs explain each operator in detail 4 2 1 Arithmetic Operators The assembler recognizes binary arithmetic operators for addition subtraction multi plication and division and the assembler has provisions that allow a user to perform these operations with both signed and unsigned values The addition subtraction and multiplication operators can all be used equally well wit
243. ue Invalid macro parameter Index error A backslash is followed by something other than a decimal number in the range from 1 through 255 Positional parameters are referenced as n where n is the position of the parameter in the macro call This error can occur if an invalid digit string or no digit at all follows a backslash in the macro model IL The label field for this statement must be empty Invalid Label error A label or a local label appears in the label field of the statement and the statement is a statement for which a label is not allowed For example a label is 150 CHAPTER 10 ASSEMBLY ERRORS not allowed with a SPACE statement or an EJECT statement If the invalid label is a local label of the form jH then any local label of the form jF with the same j value in the expression field of the statement has been incorrectly evaluated because of the effect of the invalid label IM Immediate addressing is not permitted with this instruction IMmediate addressing error Immediate addressing is specified for an instruction that does not have immediate addressing IML Too many imported labels Only 251 are allowed Importing too Many Labels error The program contains too many externally referenced symbols No more than 251 symbols may be specified as external references 7IMP An imported label is used where it is not allowed IMPorted label error An imported label appears in an expression that must be evalua
244. ultiply defined symbol errors in a seemingly correct program by using an OPT statement to set LBLLIM to a value that is too small only the first LBLLIM characters of each label are used to distinguish labels from each other and subsequent characters are effectively ignored Therefore any labels that are identical in the first LBLLIM characters are considered to be the same symbol 77ME No END statement has been found to mark the end of the source file Missing End error The end of the source file has been encountered and no END state ment has been found Perhaps the user employed a BTEXT statement but forgot to include a matching ETEXT statement or perhaps the user employed conditional assembly but did not include the desired target label 7MI Immediate addressing is required for the first operand field Missing Immediate error Immediate addressing is required for the first operand field of an HD6301 immediate to memory instruction and the user has omitted the immediate addressing symbol The assembler automatically uses the correct addressing mode but reports the error to alert the user to the incorrect syntax 7ML Missing parameter string delimiter in a macro call Missing deLimiter error A parameter substitution string that begins with a delimiter is not ended with a delimiter before the end of the line 153 ASSEMBLER USER S GUIDE MM The left and right parentheses aren t the same style MisMatch error The st
245. ure permits a user to continue a statement over an 14 CHAPTER 2 SOURCE SYNTAX unlimited number of lines programmers don t need to worry about any arbitrary restrictions on the total length of a statement The continuation feature gives the user added control over the source format and con sequently over the listing format For example statement continuation allows a user to break up long statements that would otherwise ruin the uniform appearance of the listing by extending into the columns that are normally used for the comment field A hyphen followed by either white space or a carriage return is known as a continu ation indicator Wherever white space is allowed in the noncomment part of a statement a programmer can use a hyphen followed by either white space or a carriage return to indicate that the noncomment part of the statement continues with the first nonwhite character of the next line A user can therefore break a statement and continue it to the next line follow ing the label field following the opcode field or between any two subfields of the operand field A programmer can also specify statement continuation in an expression by typing a hyphen followed by either white space or a carriage return at any logical break in the expression A logical break occurs in an expression immediately before and after every term immediately before and after every operator and immediately before and after every parenthesis A
246. ure that confusion about the location of certain segments of code won t result during code merging This option is also selectable in the source program through the use of the LINENUM option with the OPT statement which will restart the line numbering sequence from the point of its occurrence in the program The default beginning line number is one 6 3 14 LIST The List option tells the assembler to send an assembly listing to the file or device if any that was specified for the listing output The NList option on the other hand tells the assembler to suppress its output of the assembly listing The NList option suppresses only the source lines that do not contain errors All lines with assembly language errors will still be sent to the listing file device if one was specified The NList option is therefore very useful for obtaining a listing of all lines that contain errors and the NList option is often specified when the terminal is selected as the output listing device 100 CHAPTER 6 RUNNING THE ASSEMBLER The list and no list modes may also be selected in the source program by specifying the LIST or NOLIST options with the OPT statement The List and NList run time options are one time selectable but the LIST and NOLIST options with the OPT statement may be se lected in the program to list or not list various segments of the program at the programmer s discretion see Chapter 5 for a discussion of the LIST NOLIST option
247. user of all the run time options with a brief help message displayed at the terminal but remain in the option gathering mode Suppress source lines that start with a period in column one from the listing output Indent the left margin of the output listing by n character spaces INdent and Width can be used to select the left and right margins of the listing Lines associated with conditional assembly are to be included in the listing output Begin the assembly listing with the value n as the first line number Include all source lines in the listing output NList is used to suppress all lines except those that contain assembly errors from the listing output Include the location counter field in the assembly listing output Set the number of lines in the physical page of the output device that receives the listing to n Report an error if a macro call does not specify enough pa rameter substitution strings Assemble the source files for the M6800 related microproces sor denoted by xxxx For example n 6800 represents the M6800 microprocessor default value Show the source lines generated by a macro expansion in the listing output Number the first page of the assembly listing as n Generate relocatable object code Report assembly errors to the listing output This option does not affect errors sent to the terminal as selected by the Error run time option Continued 95 ASSEMBLER USER S GUIDE Ta
248. user runs the assembler The assembler recognizes an INCLUDE statement that includes a specified source file at a specified point in the assembly The INCLUDE feature allows nested includes to any desired level subject only to the maximum number of files that the user s VMS account is allowed to have open at any one time The VMS system manager can validate an individual user s account to permit the required level of include processing and the assembler automatically provides the maximum level of nesting that is currently allowed for the account The INCLUDE feature is especially valuable for large programs because it allows a user to structure a program as a tree of logical modules The INCLUDE feature is also particularly useful to a chief programmer or a project manager who is managing a large project that includes many modules developed by several programmers One of the assembler s options allows it to run either as an absolute assembler or as a relocatable assembler The assembler runs as an absolute assembler by default because most applications for the M6800 family require programs that reside in ROM at specified memory locations However the assembler can also run as a relocatable assembler and produce relocatable object modules that are suitable as inputs to the M6800 family linker The user must then invoke the linker to link the desired relocatable object module s into an executable object module When the assembler runs as an absolut
249. utting the macro name into the opcode mnemonic field of an instruction If the macro uses parameters then their substitution strings may be specified in the call Formal parameter macro calls are of the form lt macro_name gt lt param_name gt lt string gt where each formal parameter is named and a substitution string is directly assigned For example CLRBLK FROM TO TABI TAB1 TABSIZ 1 Since formal parameters are named they may be assigned in any order Furthermore since formal parameters may be assigned a default substitution string in the macro defini tion parameters that are not listed in a macro call will be assigned their associated default strings If a parameter that has no specified default string is not assigned in a macro call the assembler will assign the null string to that parameter Such an omission is flagged as a 125 ASSEMBLER USER S GUIDE missing parameter error see Chapter 10 unless the user has suppressed MP error messages see Chapters 5 and 6 A call to a macro that was modeled with formal parameters can also be invoked with the arguments arranged as positional parameters In other words if the user is cautious to match the order and number of parameters that were defined in the macro model with the macro call argument strings then the sometimes bothersome requirement of typing the formal parameter name and its accompanying equals sign can be omitted The user should be aware that
250. verwhelming However in practice good programming techniques dictate just the type of constant variable and program statement segmentations that the LK6800 linkage editor facilitates After one has accepted the structure imposed by the five types of sections a possible next step is to chain the context of modules together by passing symbols with IMPORT and EXPORT statements To aid in keeping the origin and use of these symbols separate from the standard use symbols symbol table references are specially marked with an I or an E to stand for IMPORT and EXPORT respectively The structure of the main source listing is also modified by the assembler in relocatable mode to account for the different possible program counters and label section attributes See the sample listing in the next section for specific details The final threshold to attain before mastering relocatable assembly is probably the use in expressions of a mixture of external symbols and so called relocatable labels defined in the various types of program sections that support relocation Indeed in the relocatable mode the expression evaluating routines in the assembler need to perform a more cautious check to ensure for example that 8 bit fields do not contain relocatable expressions which at link time could possibly exceed the field allocated to them this situation is flagged as an RV error More precisely the following rules must be met in order for the expression to be valid
251. well shuffled deck of confusion 92 CHAPTER 6 RUNNING THE ASSEMBLER Of course the user is not required to replace all fields of the file specification with TT or its equivalent as in the above example Any desired combination of the special symbols TT KB TERMINAL NONE and standard VMS file names may be employed to meet the needs of the user FILES SRC LST OBJ TT NONE lt cr gt The user in this example desires to input the source file from the terminal one line at a time knowing that the source will be saved in a disk file TEMP ASM in the current directory The user also desires a listing file which will be named TEMP LST and no object file will be generated 6 3 Run Time Options After a user has successfully specified the files the assembler outputs the prompting message OPTIONS to solicit run time options from the user The user should respond by typing any desired option specifications followed by a carriage return If the user wants the default settings for all run time options he or she can respond with just a carriage return The default values of the options are the most useful values so users can often respond to the options prompt with just a carriage return A snapshot of the default state of the run time option environment appears in the discussion of the run time Help option subsequent to Table 6 1 By specifying run time options a user can conveniently perform the precise assembly task that i
252. xample suppose that a programmer defines the label 3H in the label field of a statement The programmer can refer to that label as 3B in the operand field of an instruction that appears in the source program somewhere after the statement that defines the 3H label and the programmer can refer to the label as 3F in the operand field of an instruction that appears somewhere before the statement that defines the 3H label The use of a jB term in an expression refers to the closest previous definition of the jH local label and the use of a jF term in an expression refers to the closest following definition of the jH local label A jB or jF term in an expression never refers to a local label in the label field of the same statement that contains the jB or jF term so a backward or forward reference always refers to a previous or following statement never to the current statement A jH entry is valid in the label field but a jH entry is not valid in an expression Similarly a jB or jF entry can be used in an expression but a jB or jF entry is not valid in the label field as a label that is to be defined The following example shows two sequences of instructions that are functionally identi cal but the sequence on the left uses symbolic labels while the sequence on the right uses local labels ASSEMBLER USER S GUIDE LBLA EQU 29 1H EQU 29 ORG LBLA 35 ORG 1B 35 LBLB FDB LBLC LBLD 2H FDB 3F 4F LBLC FCB LBLD LBLB 3H FCB 4F 2B LBLD RMB
253. xample are repeated using the older and less preferred mode of conditional assembly In this mode the interaction of the conditional assembly parameter with the expression in the source line statement is implicit and hence more difficult to document and maintain 114 CHAPTER 7 CONDITIONAL ASSEMBLY GOIFZ DISK 770 If bit 4 of the conditional assembly parameter is zero skip statements until the system label 770 is found 770 GOULZ TAPE 900 If bit 3 of the conditional assembly parameter is set skip all statements until the system label 900 is found 115 ASSEMBLER USER S GUIDE 116 Chapter 8 Macro Assembly The M6800 assembler provides the user with macro assembly capabilities so that a single source code line can generate multiple program lines Macro assembly allows the user to define a commonly used sequence of instructions only once then insert that sequence of instructions by naming the macro The macro processor is basically a text manipulating tool that works in conjunction with the rest of the assembler A macro definition which is introduced by the MACRO pseudo op consists of a model call statement and a series of model statements followed by the MEND pseudo op The macro definition can optionally contain parameters When the macro is later called the parameters in the macro definition will be replaced by the actual strings specified by the user in the macro call A macro is called by simply naming th
254. xecution time ol ASSEMBLER USER S GUIDE 5 2 ASCIIC The ASCIIC statement which is an extension of the ASCII statement produces a counted ASCII string The assembler s first object output for an ASCIIC statement is a data byte that contains an 8 bit unsigned count of the characters if any in the string Then the assembler generates an ASCII data byte for each character of the string just as it would do for an ASCII statement The following example illustrates an ASCIIC statement along with a set of other state ments that are essentially equivalent to the given ASCIIC statement RUBICON ASCIIC Alea jacta est RUBICON FCB 2F 1F 1H ASCII Alea jacta est 2H EQU Notice from the example that the value of a label that appears in the label field of an ASCIIC statement is the memory address of the count byte not the address of the first character of the string Although the assembler reports an error for an empty string with an ASCII statement the assembler allows an empty string such as or gt with an ASCIIC statement An empty string with an ASCIIC statement has legitimate uses in some applications and it causes the assembler to generate a count byte with the value zero to indicate the character count for the empty string 5 3 ASCIIZ The ASCIIZ statement which is another extension of the ASCII statement produces a string that is terminated by a zero byte Like the ASCII statement the ASCIIZ state ment tel
255. y multiple object lines in the assem bly listing Assembler directives such as FCC or FDB can produce multiple object lines Set the length of the logical page to n Minus the two standard header lines LPAGE is the number of lines of user output that will appear per physical page Report or do not report any macro missing parameter errors to the assembly listing Continued 69 ASSEMBLER USER S GUIDE Table 5 2 Continued Option Remarks MPU xxxx Assemble the subsequent lines for the specified microprocessor xxxx Valid values for xxxx are 6800 6801 6802 6803 6805 6808 6809 6811 6301 6502 and 65150 MSHOW NOMSHOW Include or suppress respectively source lines that result from macro expansion PAGENUMBER n Begin the pagination in the assembly listing with the number n PARITY n Set the parity bit of any ASCII characters generated in the assembly to n When n equals 1 mark parity is selected When n equals 0 space or zero parity is selected RV8 NORV8 In relocatable mode give a warning if an 8 bit expression may exceed the 8 bit range during link time if NORV8 is selected Otherwise do not give such warnings TABS n1 n2 nk Set the tab stop in the columns denoted by n1 through nk A TABS OPT statement must appear in its own OPT statement TMERR NOTMERR Report or do not report respectively the macro too many parameters errors to the assembly listing 5 40 1 OPT ABS REL
256. y following the shift or rotate opera tor as an unsigned bit count that indicates the number of bit positions by which the current 16 bit expression value should be shifted or rotated For example 5217 ROL 4 means to rotate the value 5217 to the left by 4 bit positions thus obtaining the value 2175 A bit count of zero is valid but it always specifies a null shift or rotate and effectively does nothing Table 4 4 graphically illustrates the functions of the shift and rotate operators and the following paragraphs explain the shift and rotate operators in detail The discussions of the shift and rotate operators refer to the leftmost bit of a 16 bit value as bit 15 and they refer to the rightmost bit as bit zero The ASL operator specifies an arithmetic shift to the left The assembler evaluates the term following the ASL operator as a 16 bit value and it uses that value as a 16 bit unsigned shift count The shift count specifies the number of bit positions by which the assembler should arithmetically shift the current 16 bit expression value In an arithmetic shift to the left the assembler inserts zero bits at the right end of the 16 bit expression value and it discards bits as they pass to the left out of bit 15 the sign bit Obviously any shift count greater than 15 results in a final expression value of zero The ASR operator specifies an arithmetic shift to the right The assembler evaluates the term following the ASR operator
257. y such source code options override any MShow NMShow run time options Refer to Chapter 5 for information on the assembler di rective OPT The default for the MShow NMShow option is MShow 102 CHAPTER 6 RUNNING THE ASSEMBLER 6 3 20 PAGE The Page n option specifies the initial page number for the symbolic output listing if any This option is convenient when a person is writing a program that is to be included as part of a report The program listing can be generated to start at the correct page number of the overall report For example if the program listing is to begin on page 75 of a report the user could type Page 75 as a run time option to tell the assembler to start the page numbering sequence with page 75 For related information please refer to the description of the OPT PAGENUM n state ment in Chapter 5 The default value for the page number option is 1 6 3 21 RELOCATABLE The assembler is capable of generating either absolute or relocatable object code pro vided that the programmer has taken care to observe the requirements of a particular mode with respect to label definition and usage The relocatable mode of assembly is selected with the RELocatable run time option and the companion mode is selected with the ABSolute option More information on relocatable assembly can be found in Chapter 9 and in the discussions on the assembler directives ASCT BSCT CSCT DSCT EXPORT IMPORT and PSCT fou
258. yle of the right parenthesis doesn t match the style of the left parenthesis For example a parenthesized subexpression might be enclosed by and or lt and instead of being enclosed by parentheses of the same style The user has probably made a typing error MN A radix symbol has been used but was not followed by any numerals Missing Number error A prefix character such as a or has been found in an expression to indicate that a number in a particular base should follow but there is no number The assembler uses zero as a default value for the missing number 7M0 The operand field is missing and it is required Missing Operand error The operand field for an FCC ASCII ASCIIZ or ASCIIC state ment is missing entirely MOP An operator is missing in an expression multiplication assumed Missing OPerator error An operator that should appear between two terms in an expression is missing The assembler assumes a multiplication operator by default and continues evaluating the expression A parenthesized subexpression is equivalent to a term so an expression like 2 3 4 evaluates to 2 x 7 or 14 7MP The macro call is missing a parameter string Missing Parameter error The macro call statement contains fewer positional parameter strings than required or the call does not specify the string to be substituted for a non default formal parameter In either case we substitute the null string The user can turn off t

M6800 Family Assembler User's Manual

Contents

Download Pdf Manuals

Related Search

Related Contents