Post by thread:[PIC]: asm question

Hi all, I have just started learning asm and playing around with a 16f84. I've already read a couple of tutorials from the net and now have a question. How do I create a function that uses a bit orientated command and pass its arguments via a register. I have included a small piece of code to try and illustrate. #include "p16f84.inc" #DEFINE PAGE0 bcf STATUS,5 #DEFINE PAGE1 bsf STATUS,5 RBpin equ 0x0C ; Port B pin number org 4 goto start start clrf PORTB ; clear portb PAGE1 ; select bank 1 clrf PORTB ; portb all output's PAGE0 ; select bank 0 loop movlw 0 ; bit 0 portb movwf RBpin ; move bit select into register call foobar ; test function goto loop foobar bsf PORTB,RBpin return END When I compile this I get the error "Argument out of range. Least significant bits used". I guess this is because bsf expects a single bit arg but I pass it a byte. I tried loading W with my bit number and then read W from within my function which worked fine, but if I need to pass 2 args I can't use W as one arg overwrites the other. Any suggestions ? Thanks Russell.

See remars between code On Mon, Jun 26, 2000 at 11:17:16AM +0100, Russell Farnhill wrote: {Quote hidden}> Hi all, > > I have just started learning asm and playing around with a 16f84. I've > already read > a couple of tutorials from the net and now have a question. How do I create > a function > that uses a bit orientated command and pass its arguments via a register. I > have > included a small piece of code to try and illustrate. > > #include "p16f84.inc" > #DEFINE PAGE0 bcf STATUS,5 > #DEFINE PAGE1 bsf STATUS,5 > > RBpin equ 0x0C ; Port B pin number > > > org 4 > goto start > > start > clrf PORTB ; clear portb > PAGE1 ; select bank 1 > clrf PORTB ; portb all output's > PAGE0 ; select bank 0 > > loop movlw 0 ; bit 0 portb > movwf RBpin ; move bit select into register > call foobar ; test function > goto loop > > foobar > > bsf PORTB,RBpin What you are doing here is effectively a bsf PORTB,0x0C So it indeed can't work. {Quote hidden}> return > > END > > When I compile this I get the error "Argument out of range. Least > significant bits used". > I guess this is because bsf expects a single bit arg but I pass it a byte. I > tried loading > W with my bit number and then read W from within my function which worked > fine, but if I > need to pass 2 args I can't use W as one arg overwrites the other. > > Any suggestions ? What I would do is something like : RBpin EQU 0x0C ;bit to set is in RBpin, ; eg. b'00001000' foobar COMF RBpin,w ; set b'11110111' in w , complement of RBpin ANDWF PORTB,w ; make sure bit to set is 0, store in w XORWF RBpin,w ; make that bit '1' now, store in w MOVWF PORTB ; set ouput port return I hope I got this correct. There may be a faster (smaller) solution, but can't think of any. Hans

The bit manipulation insturctions (BTFxx, BSF and BCF) all include the bit number as an immediate value within the instruction. You cannot pass a bit number to these instructions in a register. However, you can get the same effect by using a bit _mask_ rather than a bit _number_ and using the ANDWF or IORWF instructions. You can convert a bit number to a bit mask using a short table lookup function. Bob Ammerman RAm Systems (high performance, high function, low-level software) {Original Message removed}

Perhaps a good way to understand this is to look at the actual bit pattern of a BSF instruction: This example is for the 16C84: The instruction is 14 bits long, as follows: 0 1 0 0 B B B F F F F F F F where "0 1 0 0" is the opcode that identifies this as a BSF instruction, "B B B" is a 3-bit binary number in the range of 0 to 7 that identifies the bit to be set, and "F F F F F F F" is a 7-bit binary number in the range of 0 to 127 (ie 0x7F) that defines the file register to be affected. You will notice that there is no way to indicate in this format that the bit number should be obtained from the W register, or a file register, or any other such location. Here is some code that could be helpful (all based on the 16C84). (Note that this code assumes that 'read-modify-write' instructions to ports will work OK in your application. See notes in the Microchip datasheets about BSF and BCF on ports if you don't understand what I mean by this). CBLOCK xxx bitno ; register containing bit to set/reset ENDC ; To turn on bit 'bitno' on PORTB (leaves rest of PORTB alone) movf bitno,W call BitToMask iorwf PORTB,F ; To turn off bit 'bitno' on PORTB (leaves rest of PORTB alone) movf bitno,W call BitToMask xorlw 0xFF ; Complement all the bits andwf PORTB,F ; To see if bit 'bitno' is set on PORTB movf bitno,W call BitToMask andwf PORTB,W btfss STATUS,Z goto bit_was_set goto bit_was_clear ; A subroutine to convert from a bit number to the corresponding bit mask: ; ; On entry: WREG is a bit number in the range 0..7 ; On exit: WREG is the mask for the corresponding bit BitToMask: addwf PCL,F retlw 0x01 retlw 0x02 retlw 0x04 retlw 0x08 retlw 0x10 retlw 0x20 retlw 0x40 retlw 0x80 ----- Original Message ----- From: Russell Farnhill <spam_OUTr.farnhillTakeThisOuTSTOR-WAVE.CO.UK> To: <.....PICLISTKILLspam@spam@MITVMA.MIT.EDU> Sent: Monday, June 26, 2000 11:01 AM Subject: Re: [PIC]: asm question > So basically what I was doing was passing the address of > the register rather than its contents ? > > I could write this so much easier in C but I'am > determined to get to grips with asm. > > Ok thanks guys, I'll keep at it. > > Russ. > > > {Original Message removed}