Post by thread:[PIC]: is it ok to have stack underflow ?

In my 16F628 program I call a subroutine that waits for an event to happen. If the event does not happen within a certain length of time can I use a goto instruction to restart the program from a known point, without using a return instruction to pop the return address back off the stack ? I know this will cause a stack underflow(?) condition, but since the stack is a circular buffer surely it won't matter because subsequent calls will have the required return to go with them - or will it ? Start: call initialize StartEnd: call setup Mainloop: call Subroutine call sub2 goto Mainloop Subroutine: clrf counter SubLoop: btfss Input1 retlw 0x00 btfss Input2 retlw 0x01 decfsz counter,f goto SubLoop goto StartEnd ;timeout so restart again Thanks, James -- http://www.piclist.com hint: To leave the PICList spam_OUTpiclist-unsubscribe-requestTakeThisOuTmitvma.mit.edu

> -----Original Message----- > From: James Hillman [SMTP:.....jamesKILLspam@spam@IND-INTERFACE.CO.UK] > Sent: Monday, February 25, 2002 9:16 AM > To: PICLISTKILLspamMITVMA.MIT.EDU > Subject: [PIC]: is it ok to have stack underflow ? > > In my 16F628 program I call a subroutine that waits for an event to > happen. > If the event does not happen within a certain length of time can I use a > goto instruction to restart the program from a known point, without using > a > return instruction to pop the return address back off the stack ? > I know this will cause a stack underflow(?) condition, but since the stack > is a circular buffer surely it won't matter because subsequent calls will > have the required return to go with them - or will it ? > In this case it's fine. As you say, the stack is circular so it dosen't matter where the PIC starts pushing within the buffer. Stack underflow is more RETURNS than CALLS, so this is technicaly overflow, as you are executing a CALL without the matching RETURN. Mike -- http://www.piclist.com hint: To leave the PICList .....piclist-unsubscribe-requestKILLspam.....mitvma.mit.edu

{Quote hidden}> ----- Original Message ----- > From: "James Hillman" <EraseMEjamesspam_OUTTakeThisOuTIND-INTERFACE.CO.UK> > To: <PICLISTspam_OUTMITVMA.MIT.EDU> > Sent: Monday, February 25, 2002 3:15 AM > Subject: [PIC]: is it ok to have stack underflow ? > > > > In my 16F628 program I call a subroutine that waits for an event to > happen. > > If the event does not happen within a certain length of time can I use a > > goto instruction to restart the program from a known point, without > using > a > > return instruction to pop the return address back off the stack ? > > I know this will cause a stack underflow(?) condition, but since the > stack > > is a circular buffer surely it won't matter because subsequent calls > will > > have the required return to go with them - or will it ? > > > > Start: > > call initialize > > StartEnd: > > call setup > > Mainloop: > > call Subroutine > > call sub2 > > goto Mainloop > > > > Subroutine: > > clrf counter > > SubLoop: > > btfss Input1 > > retlw 0x00 > > btfss Input2 > > retlw 0x01 > > decfsz counter,f > > goto SubLoop > > goto StartEnd ;timeout so restart again > > {Original Message removed}

On Mon, Feb 25, 2002 at 09:09:21AM -0600, Lawrence Lile wrote: > This sounds like skating on thin ice to me. I would avoid it. Maybe what > you really want to do is trigger a watchdog timeout, and restart the PIC for > real? Lawrence, James' logic is perfectly correct. You can throw away the stack at any time as long as you have no intention of accessing the previous data from the stack. Just as long as after the jump the code doesn't do a return/retlw before doing a call everything is fine. Note that StartEnd immediately does a subroutine call. So the return stack is instantly reset. This will work with absolutely no problem. BAJ > > --Lawrence > > {Original Message removed}

> In my 16F628 program I call a subroutine that waits for an event to happen. > If the event does not happen within a certain length of time can I use a > goto instruction to restart the program from a known point, without using a > return instruction to pop the return address back off the stack ? Yes. I've done this. In one case I used the RB0 interrupt as an externally forced GOTO 4. > I know this will cause a stack underflow(?) Overflow, actually, but that is only from the debugger's point of view. > condition, but since the stack > is a circular buffer surely it won't matter because subsequent calls will > have the required return to go with them Right. ******************************************************************** Olin Lathrop, embedded systems consultant in Littleton Massachusetts (978) 742-9014, @spam@olinKILLspamembedinc.com, http://www.embedinc.com -- http://www.piclist.com hint: To leave the PICList KILLspampiclist-unsubscribe-requestKILLspammitvma.mit.edu

> This sounds like skating on thin ice to me. I would avoid it. Maybe what > you really want to do is trigger a watchdog timeout, and restart the PIC for > real? As long as you understand what is happening and the code is properly documented, I don't see why this is "skating on thin ice". ******************************************************************** Olin Lathrop, embedded systems consultant in Littleton Massachusetts (978) 742-9014, RemoveMEolinTakeThisOuTembedinc.com, http://www.embedinc.com -- http://www.piclist.com hint: To leave the PICList spamBeGonepiclist-unsubscribe-requestspamBeGonemitvma.mit.edu

> In my 16F628 program I call a subroutine that waits for an event to happen. > If the event does not happen within a certain length of time can I use a > goto instruction to restart the program from a known point, without using a > return instruction to pop the return address back off the stack ? > I know this will cause a stack underflow(?) condition, but since the stack > is a circular buffer surely it won't matter because subsequent calls will > have the required return to go with them - or will it ? Stack overflow is never a problem, it is the stack underflow (that in most cases unavoidably follows a stack overflow) that is the problem. But what you describe is one of the specialised uses of stack overflow that does not cause a later underflow (because you skip a return), so this use ie perfectly OK. Wouter van Ooijen -- Van Ooijen Technische Informatica: http://www.voti.nl Jal compiler, Wisp programmer, WLoader bootloader, PICs kopen -- http://www.piclist.com hint: To leave the PICList TakeThisOuTpiclist-unsubscribe-requestEraseMEspam_OUTmitvma.mit.edu

> The position of the stack > pointer within the stack is completely unknown at power up It is easier to regard the pic stack as not having a stack pointer but as a chain of registers. Maybe it is even implemented that way. Consider a shelve of 8 books. You can add one at the right side, and the leftmost book will fall of the shelve and is lost forever. You can take off the rightmost and a magic copying machine duplicates the leftmost book. There are ALWAYS exactly 8 books on the shelve. The pic starts with 8 'random' books. A wise man does not return to one of these books. Wouter van Ooijen -- Van Ooijen Technische Informatica: http://www.voti.nl Jal compiler, Wisp programmer, WLoader bootloader, PICs kopen -- http://www.piclist.com hint: To leave the PICList RemoveMEpiclist-unsubscribe-requestTakeThisOuTmitvma.mit.edu

James Hillman <PICLISTEraseME.....mitvma.mit.edu> wrote: > In my 16F628 program I call a subroutine that waits for an event to > happen. If the event does not happen within a certain length of time > can I use a goto instruction to restart the program from a known > point, without using a return instruction to pop the return address > back off the stack ? I know this will cause a stack underflow(?) > condition, but since the stack is a circular buffer surely it won't > matter because subsequent calls will have the required return to go > with them - or will it ? James: As others have said, it'll work fine. If you try debugging your code on one of Microchip's emulators, however, you may have problems. At the very least, you'll have to disable the "stack overflow/underflow" warning, and -- depending on which emulator you use -- you may have to change the code to avoid the overflow. Speaking of which... Why not change your code to avoid the overflow, anyway? You only need to add 5 lines and remove one, and you'll end up with a subroutine that behaves as most people would expect it to: Mainloop: call Subroutine BTFSC TIMEOUT ;*** ADD GOTO StartEnd ;*** ADD call sub2 goto Mainloop Subroutine: BCF TIMEOUT ;*** ADD clrf counter SubLoop: btfss Input1 retlw 0x00 btfss Input2 retlw 0x01 decfsz counter,f goto SubLoop BSF TIMEOUT ;*** ADD RETLW 0x02 ;*** ADD ;*** REMOVE goto StartEnd ;timeout so restart again -Andy === Andrew Warren -- EraseMEaiwcypress.com === Principal Design Engineer === Cypress Semiconductor Corporation === === Opinions expressed above do not === necessarily represent those of === Cypress Semiconductor Corporation -- http://www.piclist.com hint: To leave the PICList RemoveMEpiclist-unsubscribe-requestEraseMEEraseMEmitvma.mit.edu