Post by thread:MPASM question

Finn Amundsen (.....famKILLspam@spam@cap-computas.no) wrote: >Is there a way to avoid the warnings that occur when data in memory >banks 1-3 is accessed? >Processor type is 16C57. > >Example: > >Definition: > ORG 70H ; BANK3 >RAMXX RES 1 > >Code: > BSF FSR,BA1 ; BANK3 > BSF FSR,BA2 >Warning: Argument out of range: RAMDE (112). Least significant bits used. > CLRF RAMXX Finn: Sure... There are a number of ways. First, in case you don't understand what exactly is causing the warning, here's a quick explanation: The 16C57's 12-bit instructions are only wide enough to hold an opcode and a 5-bit register address. They work fine for the 16C54, which only has 32 registers, but they can't handle the 7-bit addresses of the 16C57's registers. That's why the high two bits of those register-addresses are held in the FSR bits you call BA1 and BA2. When you equate the symbol RAMDE to address 70 (hex), then try to do a CLRF RAMDE, the assembler sees that the address won't fit in 5 bits, so it ignores the high two bits, assembles the instruction using the low 5 bits, and generates the warning message. It's only a warning; your code will work just fine. The easiest (and, from some points of view, the worst) way to keep the warnings from appearing is to use the assembler directive "LIST W=1", which will keep warning messages from appearing in your .LST file. Another way is to equate your symbols, not at their actual addresses, but at the 5-bit equivalent of those addresses. This is difficult to do if you use ORG and RES to reserve space for registers; it's easy if you use EQU. For example: RAMG0 EQU 0x08 ;FIRST GLOBALLY-ACCESSIBLE REG. RAMG1 EQU RAMG0+1 ;NEXT ONE. RAMG2 EQU RAMG1+1 ;NEXT... .... RAM00 EQU 0x10 ;FIRST REGISTER ON PAGE 0. RAM01 EQU RAM00+1 ;ETC... .... RAM10 EQU 0x30 & 0x1F ;FIRST REGISTER ON PAGE 1. RAM11 EQU RAM10+1 ;ETC... .... RAM20 EQU 0x50 & 0x1F ;FIRST REGISTER ON PAGE 2. RAM21 EQU RAM20+1 ;ETC... .... RAM30 EQU 0x70 & 0x1F ;FIRST REGISTER ON PAGE 3. RAM31 EQU RAM30+1 ;ETC... This will assemble beautifully, and will generate no warnings when you use the registers, but the PIC-Master emulator (and probably MPSIM, too) will be a little confused. This confusion manifests itself as an inability to specify watch-window variables by name; since the emulator thinks that RAM30 is at address 0x10, it will display 0x10's contents for RAM30, instead of 0x70's. This is a pretty minor thing, easily worked around. You can also equate your symbols to their real addresses, then do the "& 0x1F" masking in the body of your source-code. For example: ORG 0x70 RAMDE RES 1 ;FIRST REGISTER ON PAGE 3. .... CLRF RAMDE & 0x1F This will generate no warnings, and PIC-Master/MPSIM will have a slightly-easier time with it. With the current version of MPASM, it's easy to write macros that redefine the PIC opcodes, so the assembler can even be made to automatically insert the "& 0x1F" for you when necessary. Unfortunately, when MPASM version 2.0 is released, it will not allow redefinition of opcodes, so you'll probably want to do the bit-masking in the EQUs or just turn off the warning-message display. Good luck... -Andy -- Andrew Warren - fastfwdKILLspamix.netcom.com Fast Forward Engineering, Vista, California

Hello all, I know this question will bring a hailstorm of advice and comments about which assembler(s) are best. I am going to ask it anyway. I have been using the Parallax assembler (PASM) to generate 16C57 code. In several instances my code has grown large enough to utilize most of the program memory as well as the RAM in the PIC. I have found that it gets very confusing trying to manage all the bank switching and all required for the 16C57. Does the Microchip assembler (MPASM) offer any features which simplify this dilemma? ie, can it be set up to automate the bank switching from within the assembler? Thanks for your comments, Martin Kirk Arizona State University .....mlkKILLspam.....asu.edu (602) 582-5718

Martin Kirk (EraseMEmlkspam_OUTTakeThisOuTASU.EDU) wrote: > I have been using the Parallax assembler (PASM) to generate 16C57 > code. .... I have found that it gets very confusing trying to > manage all the bank switching and all required for the 16C57. > > Does the Microchip assembler (MPASM) offer any features which > simplify this dilemma? ie, can it be set up to automate the bank > switching from within the assembler? Martin: MPASM doesn't automate the process, but its macro capability (which Parallax's PASM lacks) does make the process a WHOLE lot easier. -Andy Andrew Warren - fastfwdspam_OUTix.netcom.com Fast Forward Engineering, Vista, California

Philip Lalone wrote: > > Is it possible to write a MPASM macro that will print a string > using: PRINT "This is the string"? The only way i've been able to get > anything like this working is writing seperate code for each string I > want to print, any theory or code would be appriciated. I need to do this > for serial data and a LCD, which I've written code for, so all the macro > would need to do is call the serial/lcd routine with each character or a > pointer to the string. Philip, I'm not sure if this is exactly what you want... But several months ago I posted some code that allowed strings to be looked up in a table. Matthew Rowe copied it into his "Virtual Notebook" (see the write_string function): http://hobbes.king.ac.uk:80/matt/pic/tables.html You will still need to supply the LCD infra-structure software. That should be no problem since you see a new version every month or so... Matthew has some other PIC stuff at http://hobbes.king.ac.uk:80/matt/pic/ (Hey Matthew, are you still out there?) Scott

On Mon, 23 Sep 1996, Philip Lalone wrote: > Is it possible to write a MPASM macro that will print a string > using: PRINT "This is the string"? The only way i've been able to get > anything like this working is writing seperate code for each string I > want to print, any theory or code would be appriciated. I need to do this > for serial data and a LCD, which I've written code for, so all the macro > would need to do is call the serial/lcd routine with each character or a > pointer to the string. I'm not certain about the syntax, but I believe there's a built-in that will take a string (?) and assemble a series of retlw instructions, which suggests that you could get the macro to generate something like this: goto str1_end str1: retlw 'T' retlw 'h' retlw 'i' retlw 's' . . . retlw 0 str1_end: movlw high(str1) movwf StrPtrHi movlw low(str1) call RoutineToOutputStringToWherever Where the output routine would handle stepping the pointer and fetching bytes until the terminating NUL was reached. Alternately, I believe you could make the first byte returned be a count. In either case, the string's (far) address is in StrPtrHi:W on entry to the output routine. Hmmm... I may have a use for this in a project that's been simmering on the back burner for a while. :-)

Martin J. Maney wrote: > > I'm not certain about the syntax, but I believe there's a built-in that > will take a string (?) and assemble a series of retlw instructions.... dt "your string", 0 will be assembled by MPASM as though if it were RETLW 'y' RETLW 'o' RETLW 'u' RETLW 'r' . . . RETLW 0 Scott PS. Did the previous message I posted on this thread get out there? I got confirmation from the PICLIST server, yet I also got an error from what I believe to be a Compuserve server saying that the "mail-box" was full. I requested this month's archive and did see my post. However, Martin's question/response suggests that my post never made it (or maybe you didn't/ couldn't check out the web link, Martin?). At any rate, I can a) repost the original message, or b) Post a message that has the code (so you don't have to surf).

try this: 0008 301B 00031 MOVLW 00011111b ;value I always pad out the binary to 8 chars. Also try '00011111'b with ' ' chars either side as early versions of mpasm seemed to need this. Also - I do not use the tris reg. Instead use the fsr and indf reg, eg movlw trisa ;get the address of the tris reg into w movwf fsr ;store this address at the indirect setup reg movlw 00001111b ;set the io bits and store them at movwf indf ;the indirect register which is pointing to trisa This method stops having to use page 0,1 bit setups and stops compiler messages Happy Xmas ----- Stephen H Alsop ----- email: @spam@s.ssystemsKILLspameasynet.co.uk www : http://easyweb.easynet.co.uk/~s.ssystems S&S Systems Ltd, Bretton Court, Manor Road, Wales Village, Sheffield S31 8PD, England. Tel: 01909 773399 * Fax: 01909 773645 ---------- : From: Tony Matthews <KILLspamtonyKILLspamMAGICNET.NET> : To: Multiple recipients of list PICLIST <RemoveMEPICLISTTakeThisOuTMITVMA.MIT.EDU> : Subject: mpasm question : Date: 22 December 1996 00:21 : : Hello : Would someone be willing to explain what I am doing wrong here This is : mpasm .lst file clipping for a 16c84? : : Warning[202]: Argument out of range. Least significant bits used. : 0008 301B 00031 MOVLW 11111b ;value : used to initialize data direction : Message[302]: Register in operand not in bank 0. Ensure that bank bits : are correct. : 0009 0085 00032 MOVWF TRISA ;set : ra <0:5> to input mode

>Warning[202]: Argument out of range. Least significant bits used. >0008 301B 00031 MOVLW 11111b It doesn't understand you are giving it a binary number. You should say: MOVLW b'00011111' >Message[302]: Register in operand not in bank 0. Ensure that bank bits >are correct. >0009 0085 00032 MOVWF TRISA PICs can't address registers above 7F directly, and TRISA is 85, and the assembler is telling you it is addressing 05 or 85, it doesn't know which. Which one depends on the setting of bit 5 of the status register (this bit is usually referred to as "RP0"). In other words, of the eight bits you need for a register's address, you can only specify the lower 7. The eighth bit is ignored, and RP0 is used instead. To make sure that you are really talking to TRISA, you would need to do this: movlw b'00011111' ;the direction pattern you wanted bsf STATUS,5 ;set RP0 to the 80's movwf TRISA ;move w into TRISA bcf STATUS,5 ;switch back to the lower bank of registers You'll still get the "Message[302]: Register in operand not in bank 0. Ensure that bank bits" warning, but ignore it or tell the assembler to suppress it. Most people will tell you that changing to register bank 1 is not needed, since there is a perfectly good way to do just what you wanted. I'd do it, too! movlw b'00011111' ;your bit pattern tris 5 ;5 is the address of port A This will give you an assembler warning, but ignore it. The tris command is still valid and you should use it whenever you want to. Cheers, Bob

yes it is easy: the macros can even take parameters. for e.g. the following macro simulates an 8051 instruction (compare and jump if not equal) the macro begins with a "macro" declaration and ends with a "endm" statement. if the "reg" value is not equal to the constant the program branches to "label". cjne macro reg,constant,label movlw constant subwf reg,w btfss status,zero goto label endm {Quote hidden}> > Is it possible in MPASM to create a #define-macro which translates to > - multiple lines of 'code', among which > - a start-of-macro-definition (the "label MACRO parameter,..." type of > macro) line > > I want to create > #procedure(procname) > and > #endproc > macros which (among other things) store the intervening lines as a macro > definition. > > regards, > Wouter. >

Thank you for the info , I will see how things pan out Regards- David At 03:22 PM 13/08/97 -0700, you wrote: {Quote hidden}>yes it is easy: >the macros can even take parameters. > >for e.g. the following macro simulates >an 8051 instruction (compare and jump if not equal) >the macro begins with a "macro" declaration and ends with >a "endm" statement. if the "reg" value is not equal >to the constant the program branches to "label". > > >cjne macro reg,constant,label > movlw constant > subwf reg,w > btfss status,zero > goto label > > endm > > > >> >> Is it possible in MPASM to create a #define-macro which translates to >> - multiple lines of 'code', among which >> - a start-of-macro-definition (the "label MACRO parameter,..." type of >> macro) line >> >> I want to create >> #procedure(procname) >> and >> #endproc >> macros which (among other things) store the intervening lines as a macro >> definition. >> >> regards, >> Wouter. >> > >

Ooijen,Wouter van <Ooijen,Wouter van <spamBeGonePICLISTspamBeGoneMITVMA.MIT.EDU>> wrote: > I want a #define-style macro to translate to mutiple lines, one of > which would be the beginning (or end) of a label-MACRO-args style > macro. > > something like: > > #define procedure(name) \ > name macro > > #define endproc(name) \ > endm Wouter: I'm not sure that I understand what you want... But if you just want to type procedure (name) and have MPASM translate it to name macro then you can do it with the following #define: #define procedure(name) ((name) macro) The corresponding #define for "endproc" is simply: #define endproc(name) (endm) Depending upon the way you indent your "procedure" and "endproc" invocations, you may need to add the following to the top of your source file: LIST FIXED=0 I hope this helps... But, again, I really don't think I understood your question. -Andy === Andrew Warren - TakeThisOuTfastfwdEraseMEspam_OUTix.netcom.com === Fast Forward Engineering - Vista, California === http://www.geocities.com/SiliconValley/2499 === For PICLIST help (including "unsubscribe" instructions), === put the single word "help" in the body of a message and === send it to: RemoveMElistservTakeThisOuTmitvma.mit.edu

I tried andrew's suggestion: #define open(x) ((x) macro) but that does not work (in mpasm V01.50 from MPLAB). "open(a)" seems to be translated to "(a) macro" or "((a) macro)", which is not accepted by mpasm as the start of a macro definition. BTW, I can start a macro open (see below), but I can not end a macro with close! #define open(name)name macro #define close endm ... open(my_macro) ; macro body close Anyone understands why? To rephrase the second part of my question: can a #define macro translate to multiple lines? regards, Wouter. {Quote hidden}> ---------- > From: Andrew Warren[SMTP:fastfwdEraseME.....IX.NETCOM.COM] > Reply To: pic microcontroller discussion list > Sent: Thursday 14 August 1997 10:17 > To: EraseMEPICLISTMITVMA.MIT.EDU > Subject: Re: mpasm question > > Ooijen,Wouter van <Ooijen,Wouter van <RemoveMEPICLISTEraseMEEraseMEMITVMA.MIT.EDU>> wrote: > > > I want a #define-style macro to translate to mutiple lines, one of > > which would be the beginning (or end) of a label-MACRO-args style > > macro. > > > > something like: > > > > #define procedure(name) \ > > name macro > > > > #define endproc(name) \ > > endm > > Wouter: > > I'm not sure that I understand what you want... But if you just want > to type > > procedure (name) > > and have MPASM translate it to > > name macro > > then you can do it with the following #define: > > #define procedure(name) ((name) macro) > > The corresponding #define for "endproc" is simply: > > #define endproc(name) (endm) > > Depending upon the way you indent your "procedure" and "endproc" > invocations, you may need to add the following to the top of your > source file: > > LIST FIXED=0 > > I hope this helps... But, again, I really don't think I understood > your question. > > -Andy > > === Andrew Warren - RemoveMEfastfwdspam_OUTKILLspamix.netcom.com > === Fast Forward Engineering - Vista, California > === http://www.geocities.com/SiliconValley/2499 > > === For PICLIST help (including "unsubscribe" instructions), > === put the single word "help" in the body of a message and > === send it to: RemoveMElistservTakeThisOuTspammitvma.mit.edu >

Ooijen,Wouter van <Ooijen,Wouter van <EraseMEPICLISTspamspamBeGoneMITVMA.MIT.EDU>> wrote: > I tried andrew's suggestion: > #define open(x) ((x) macro) > but that does not work (in mpasm V01.50 from MPLAB). > "open(a)" seems to be translated to "(a) macro" or "((a) macro)", > which is not accepted by mpasm as the start of a macro definition. Whoops... Sorry about that, Wouter; I wasn't thinking. > BTW, I can start a macro open (see below), but I can not end a macro > with close! > #define open(name)name macro > #define close endm > ... > open(my_macro) > ; macro body > close > Anyone understands why? Yeah... Now that I've actually tried to do what I said you could do, I see that it's not possible; MPASM doesn't like anything to be #defined to "endm". Sorry for wasting your time. > To rephrase the second part of my question: can a #define macro > translate to multiple lines? No. -Andy === Andrew Warren - RemoveMEfastfwdKILLspamix.netcom.com === Fast Forward Engineering - Vista, California === http://www.geocities.com/SiliconValley/2499 === For PICLIST help (including "unsubscribe" instructions), === put the single word "help" in the body of a message and === send it to: listservSTOPspamspam_OUTmitvma.mit.edu

part 0 1258 bytes Friends, I came across some code that has this in it:         movlw           (high(TEXTSTART)) Can somebody plese explain to me what this means. I would appreciate it very much. Thank you in advance!! Best regards, Steve Steven Kosmerchock Father/Student/Engineering Technician http://www.geocities.com/researchtriangle/lab/6584 "Great spirits have always encountered violent oppposition from mediocre minds."--A.Einstein </BODY> </HTML> </x-html>

part 0 1253 bytes content-type:text/html; "Kosmerchock, Steve" <spamBeGoneSteve.KosmerchockSTOPspamEraseMERFSWORLD.COM> on 02/28/2000 01:45:11 PM Please respond to pic microcontroller discussion list <KILLspamPICLISTspamBeGoneMITVMA.MIT.EDU> To: EraseMEPICLISTEraseMEMITVMA.MIT.EDU cc: (bcc: Andrew Kunz/TDI_NOTES) Subject: MPASM Question Steve, It means that the value of TEXTSTART doesn't fit in 8 bits, and that you want to get the portion of the value in the high byte (bits 15-8) into the W register. This is a common way of breaking a 16-bit value into two 8-bit chunks, so it can be manipulated by an 8-bit processor. Andy Friends, I came across some code that has this in it: movlw (high(TEXTSTART)) Can somebody plese explain to me what this means. I would appreciate it very much. Thank you in advance!! Best regards, Steve Steven Kosmerchock Father/Student/Engineering Technician http://www.geocities.com/researchtriangle/lab/6584 "Great spirits have always encountered violent oppposition from mediocre minds."--A.Einstein Content-type: text/html; name="att1.htm" Content-Disposition: attachment; filename="att1.htm" Content-Description: Internet HTML Attachment converted: birth:att1.htm (TEXT/MSIE) (00006858)

part 0 2664 bytes <TITLE>MPASM Question</TITLE> <META content='"MSHTML 4.72.3110.7"' name=GENERATOR> </HEAD> <BODY> <DIV>PIC processors process data in 8-bit bytes.  The physical memory address represented by a label in your code is more than 8 bits, since PIC's have more than 256 ROM locations.  So, in order to break code addresses into 8-bit chunks that can be processed by a PIC, the assembler provides high and low operators which instruct the assembler to use the high or low 8 bits of the memory address.  These operators become essential when dealing with memory pages and when coding case statements.</DIV> <BLOCKQUOTE style="BORDER-LEFT: #0000ff solid 2px; MARGIN-LEFT: 5px; PADDING-LEFT: 5px"> <DIV align=left class=OutlookMessageHeader DIR = LTR>{Original Message removed}