Post by thread:Bit setting via a variable

James Tilley <spam_OUTPICLISTTakeThisOuTMITVMA.MIT.EDU> wrote: > I am trying to set a specific bit on portb, and I wanted to do this > by using a variable and the bsf function. Problem is, the bsf > function evaluates the variable as the literal file register > address. Is there a way to do this aside from writing to the whole > of portb? Like some compiler directive to evaluate the contents of > the file register instead of its address? No; the BSF and BCF instructions take two literal operands. -Andy === Andrew Warren - .....fastfwdKILLspam@spam@ix.netcom.com === Fast Forward Engineering - Vista, California === http://www.geocities.com/SiliconValley/2499

On Thu, 19 Feb 1998 23:08:15 -0800 Andrew Warren <fastfwdKILLspamIX.NETCOM.COM> writes: >James Tilley <.....PICLISTKILLspam.....MITVMA.MIT.EDU> wrote: > >> I am trying to set a specific bit on portb, and I wanted to do this >> by using a variable and the bsf function. Problem is, the bsf >> function evaluates the variable as the literal file register >> address. Is there a way to do this aside from writing to the whole >> of portb? Like some compiler directive to evaluate the contents of >> the file register instead of its address? > > No; the BSF and BCF instructions take two literal operands. > Although you should be able to point FSR at the port, then do a BSF or BCF on INDF, right? Harold _____________________________________________________________________ You don't need to buy Internet access to use free Internet e-mail. Get completely free e-mail from Juno at http://www.juno.com Or call Juno at (800) 654-JUNO [654-5866]

On Thu, 19 Feb 1998, James Tilley wrote: > I am trying to set a specific bit on portb, and I wanted to do this by > using a variable and the bsf function. Problem is, the bsf function > evaluates the variable as the literal file register address. Is there a > way to do this aside from writing to the whole of portb? bcf and bsf do a read/modify/write cycle, so they also write to the whole of portb. Only the pins set as outputs by trisb will actually change. Bitmasks work almost like bit pointers. The bitmask 00000001b is the "bit pointer" to bit 0. 00001000b points to bit 3. Say your "bit pointer" is in the file register "BitP". "bsf BitP" and "bcf BitP" can be simulated with these little functions. bsfbitp ; subroutine to set the bit in portb pointed to by BitP movf BitP,w iorwf portb,f return bcfbitp ; subroutine to clear the bit in portb pointed to by BitP comf BitP,w andwf portb,f return Use rotates to increment and decrement these "bit pointers". If the carry bit gets set, you've reached the end of the byte. I'm sure someone on the list has a clever way of converting a number in the range 0 to 7 into a bitmask with one bit set and vice versa. -- KILLspampaulhKILLspamhamjudo.com http://www.hamjudo.com The April 97 WebSight magazine describes me as "(presumably) normal".

I often write to PORTB indirectly like this: BSF MIRROR, 3 MOVF MIRROR, W MOVWF PORTB This avoids reading PORTB and possible errors associated with that route. To use this method you must insure that the variable MIRROR always reflects the proper states for PORTB and you never write directly to PORTB without also writing to MIRROR. {Quote hidden}> I'm not sure if I'm missing something completely obvious our what, > but here is my problem: I am trying to set a specific bit on portb, > and I wanted to do this by using a variable and the bsf function. > Problem is, the bsf function evaluates the variable as the literal > file register address. Is there a way to do this aside from writing > to the whole of portb? Like some compiler directive to evaluate the > contents of the file register instead of its address? > > Thanks for any help... > > James Best Regards, Lawrence Lile

On Thu, 19 Feb 1998 23:50:37 -0800 Andrew Warren <RemoveMEfastfwdTakeThisOuTIX.NETCOM.COM> writes: >Harold M Hallikainen <spamBeGonePICLISTspamBeGoneMITVMA.MIT.EDU> wrote: > >> > the BSF and BCF instructions take two literal operands. >> >> Although you should be able to point FSR at the port, then do a BSF >> or BCF on INDF, right? > > Right, Harold... But as I inmterpreted the question, he was > asking about using a variable for the SECOND operand (i.e., the > bit number). > Oh, I see! I'd probably do that with a jump table. I had something like that where I had to enable demultiplexors based on a channel number (which I shifted right three places since the demux was 1 in 8 out). The jump table looked at w, then jumped to the appropriate "goto enablemux1" or whatever. Harold _____________________________________________________________________ You don't need to buy Internet access to use free Internet e-mail. Get completely free e-mail from Juno at http://www.juno.com Or call Juno at (800) 654-JUNO [654-5866]

At 09:31 AM 2/20/98 -0400, you wrote: <snip> {Quote hidden}>Bitmasks work almost like bit pointers. The bitmask 00000001b is the "bit >pointer" to bit 0. 00001000b points to bit 3. Say your "bit pointer" is >in the file register "BitP". "bsf BitP" and "bcf BitP" can be simulated >with these little functions. > >bsfbitp ; subroutine to set the bit in portb pointed to by BitP > movf BitP,w > iorwf portb,f > return > >bcfbitp ; subroutine to clear the bit in portb pointed to by BitP > comf BitP,w > andwf portb,f > return > >Use rotates to increment and decrement these "bit pointers". If the carry >bit gets set, you've reached the end of the byte. > >I'm sure someone on the list has a clever way of converting a number in >the range 0 to 7 into a bitmask with one bit set and vice versa. > >-- >TakeThisOuTpaulhEraseMEspam_OUThamjudo.com http://www.hamjudo.com >The April 97 WebSight magazine describes me as "(presumably) normal". > I have never really participated in the discussions about neat programming algorithims but I have always been fascinated by them. Since the challenge has been issued for a "clever" way of creating a bitmask, I decided to give it a try. The methods given here are probably not the most efficient in size or speed, but I am looking to improve my skills at programming tight algorithims so I thought I would throw them out here for critique. Here is the fastest routene I could come up with to convert a register INDEX to a bitmask in BitP This routene ignores the upper five bits of INDEX, and sets the bit in BitP which corresponds to the value of the lower three bits of INDEX. It preserves W and needs no other registers besides INDEX and BitP. It is 13 words long and takes 12 cycles for INDEX[bits 0,1,2] < 4 and 13 cycles otherwise. bcf STATUS,C clrf BitP,F bsf BitP,0 btfss INDEX,2 goto NOT4 clrf BitP,F bsf BitP,4 NOT4 btfsc INDEX,1 rlf BitP,F btfsc INDEX,1 rlf BitP,F btfsc INDEX,0 rlf BitP,F Here's the smallest routene: It looks at the whole INDEX register, sets the bit in BitP which corresponds to the value in INDEX. It trashes W, also needs no additional registers, and is only 8 words long. It takes 8 cycles for INDEX < 2 and 8 + 4*(INDEX-1) for INDEX > 1 bcf STATUS,C movlw 1 movwf BitP addwf INDEX,F ALOOP rlf BitP,F decfsz INDEX,F goto ALOOP rrf BitP,F DISCLAIMER: Anyone can use these routenes for anything, however, I have not tested them. Sean +--------------------------------+ | Sean Breheny | | Amateur Radio Callsign: KA3YXM | | Electrical Engineering Student | +--------------------------------+ Fight injustice, please look at http://homepages.enterprise.net/toolan/joanandrews/ Personal page: http://www.people.cornell.edu/pages/shb7 RemoveMEshb7TakeThisOuTcornell.edu Phone(USA): (607) 253-0315

Paul Haas <paulhEraseME.....hamjudo.com> wrote: > I'm sure someone on the list has a clever way of converting a number > in the range 0 to 7 into a bitmask with one bit set and vice versa. Paul: Here's the most obscure way I could come up with to convert from a number in the range [0-7] to a bitmask. 9 words of program space and between 3 and 10 cycles to execute. Note that this routine is subject to the usual "direct modification of the Program Counter" limitations (e.g., on a 16C5x, the whole thing must be in the first half of a page; and on the other parts, PCLATH must be set apropriately before the "ADDWF" is executed). ADDWF PCL ;CALCULATE 2^W, WHERE W IS IN THE ADDLW 0 ;RANGE [0-7]. ADDLW -1 ; ADDLW -3 ; ADDLW -7 ; ADDLW -15 ; ADDLW -31 ; ADDLW -63 ; ADDLW 121 ; -Andy === Andrew Warren - EraseMEfastfwdix.netcom.com === Fast Forward Engineering - Vista, California === http://www.geocities.com/SiliconValley/2499

On Fri, 20 Feb 1998 15:46:34 -0500 Sean Breheny <RemoveMEshb7EraseMEEraseMECORNELL.EDU> writes: [...] > bcf STATUS,C > movlw 1 > movwf BitP > addwf INDEX,F >ALOOP rlf BitP,F > decfsz INDEX,F > goto ALOOP > rrf BitP,F This is a good start, but can be simplified quite a bit: setc clrf BitP incf INDEX,f ALOOP rlf BitP,f decfsz INDEX,f goto ALOOP Here we start with BitP=0 and C=1, and rotate left INDEX+1 times. The worst-case execution time (when INDEX=7) is about 34 cycles. This routine doesn't affect W, but it does affect the C and Z flags. INDEX must be in the range of 0 to 7. There is a trick method that works for 4 bits. It replaces a 2-bit number in a variable with a 4-bit mask: incf BitP,w ;W = 0001 0010 0011 0100 btfsc BitP,1 ;If 0 or 1, result is almost correct now. iorwf BitP,f ;BitP = [0000] [0001] 0011 0111 incf BitP,f ;BitP = 0001 0010 0100 1000 It could be the core of an 8-bit routine thus: ; Convert 3-bit number (0-7) in INDEX to a 8-bit mask (00000001 ... ; 10000000) in BitP. movfw INDEX andlw b'00000011' ;Start with half of the mask. movwf BitP incf BitP,w ;The 4-bit converter btfsc BitP,1 iorwf BitP,f incf BitP,f btfsc INDEX,2 ;Is it high 4 bits? swapf BitP,f This is more code space than the repeated shifting methods, but fast execution of only 9 cycles. Also, the high 5 bits of INDEX are ignored. The routine doesn't offer any trememdous advantage over using a table other than not having to pay attention to the placement of the routine, set up PCLATH, etc. I think someone did present a better 8-bit routine the last time this topic came up. _____________________________________________________________________ You don't need to buy Internet access to use free Internet e-mail. Get completely free e-mail from Juno at http://www.juno.com Or call Juno at (800) 654-JUNO [654-5866]