Post by thread:[PIC]: Help! :) Error with FSR0 and MPLAB-C18?

It seems that I am encountering an error in MPLAB-C18. I have a procedure that has a parameter passed to it. That parameter is used as an index into an array of structures. If that procedure is interrupted, when the interrupt handler returns, the wrong index is used to finish the assignments. For example, if this is the code that is running (ID is the parameter, and is equal to 1 before the interrupt): DEBUG_ROLL=TIMERS[ID].ROLLOVERS_LEFT; ** Interrupt occurs here ** DEBUG_ELAPSED=TIMERS[ID].ELAPSED; DEBUG_TIME=TIMERS[ID].TIME; When the return from interrupt occurs, ID is now 0. So the code ends up indexing the wrong element of the array... Anybody else seen anything like this? What should I post to help out? Thanks, Jeff Meyer Rockwell Collins, Inc. -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email spam_OUTlistservTakeThisOuTmitvma.mit.edu with SET PICList DIGEST in the body

> DEBUG_ROLL=TIMERS[ID].ROLLOVERS_LEFT; > > ** Interrupt occurs here ** > > DEBUG_ELAPSED=TIMERS[ID].ELAPSED; > DEBUG_TIME=TIMERS[ID].TIME; > > When the return from interrupt occurs, ID is now 0. So the code ends up > indexing the wrong element of the array... Sounds like the interrupt is trashing some state. Is the interrupt routine saving/restoring FSR? Take a look at the assembly code generated for the lines above and see what resources it uses. ***************************************************************** Olin Lathrop, embedded systems consultant in Devens Massachusetts (978) 772-3129, .....olinKILLspam@spam@embedinc.com, http://www.embedinc.com -- http://www.piclist.com hint: To leave the PICList piclist-unsubscribe-requestKILLspammitvma.mit.edu

Here is the assembly code generated for the first part of the interrupt handler: 0014fa cfe7 MOVFF 0xfe7,0xfe6 void Interrupt_Handler(void) { 0014fc ffe6 0014fe cfe9 MOVFF 0xfe9,0xfe6 001500 ffe6 001502 cfea MOVFF 0xfea,0xfe6 001504 ffe6 001506 cfd9 MOVFF 0xfd9,0xfe6 001508 ffe6 00150a cfe1 MOVFF 0xfe1,0xfd9 00150c ffd9 00150e 0e01 MOVLW 0x1 001510 26e1 ADDWF 0xe1,0x1,0x0 unsigned char TEMP; 001512 a4f2 BTFSS 0xf2,0x2,0x0 if (INTCONbits.TMR0IF) { The registers for the 18C452 are: FE7=INDF1 FE6=POSTINC1 FE9=FSR0L FEA=FSR0H FD9=FSR2L FE1=FSR1L Upon further testing and analysis, it seems that I may have misstated the original problem. I don't know if it is just FSR0 getting corrupted, or if it is more than that. I originally thought that the index term (ID) was being corrupted, but it looks like the entire data stack is behaving funny. I should also note that the procedure that is being interrupted is meant to be reentrant. This is the reason I found this "bug" in the first place. One of the interrupts that can interrupt the procedure is the serial data interrupt. If that happens, the same procedure is called by the serial data interrupt handler (with a different value for ID). Thanks for the help , Jeff. -- http://www.piclist.com hint: To leave the PICList .....piclist-unsubscribe-requestKILLspam.....mitvma.mit.edu

Sorry, I forgot to put the last bit of code that restores the data: 00154c 0e02 MOVLW 0x2 00154e 5ee1 SUBWF 0xe1,0x1,0x0 001550 cfe5 MOVFF 0xfe5,0xfd9 001552 ffd9 001554 cfe5 MOVFF 0xfe5,0xfea 001556 ffea 001558 cfe5 MOVFF 0xfe5,0xfe9 00155a ffe9 00155c 0011 RETFIE 0x1 FE5=POSTDEC1 Thanks, Jeff. -- http://www.piclist.com hint: To leave the PICList EraseMEpiclist-unsubscribe-requestspam_OUTTakeThisOuTmitvma.mit.edu

> I should also note that the procedure that is being interrupted is meant to > be reentrant. This smells like a stack overflow problem. ***************************************************************** Olin Lathrop, embedded systems consultant in Devens Massachusetts (978) 772-3129, olinspam_OUTembedinc.com, http://www.embedinc.com -- http://www.piclist.com hint: To leave the PICList @spam@piclist-unsubscribe-requestKILLspammitvma.mit.edu

Re: > >When the return from interrupt occurs, ID is now 0. So the code ends up >indexing the wrong element of the array... > Make sure you are saving the temporary data segment in the declaration for your ISR -- I think it's called ".tmpdata". Also, make sure you have the latest and greatest version of MPLAB-C18 -- several updates occurred in the past few months, and I, too was bitten by a int-save bug until I got the update AND I additionally saved that segment. -- ______________________________________ Andrew E. Kalman, Ph.D. KILLspamaekKILLspampumpkininc.com -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

>Make sure you are saving the temporary data segment in the >declaration for your ISR -- I think it's called ".tmpdata". Andrew, Thanks for the reply. I heard something like this from the guys at Microchip as well. I did try to save the ".tmpdata" area, but I could not get the syntax right (I kept getting invalid symbol errors and what not.) Do you remember the correct way to do this? I have tried numerous things, including: #pragma interrupt Interrupt_Handler save=.tmpdata #pragma interrupt Interrupt_Handler save=tmpdata but I (of course) get the "invalid symbol" or "symbol name expected" errors. I also have MPLAB-C18 v 1.00.13. Thanks, Jeff. -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email RemoveMElistservTakeThisOuTmitvma.mit.edu with SET PICList DIGEST in the body

Hi Jeff. Re: >Thanks for the reply. I heard something like this from the guys at >Microchip as well. I did try to save the ".tmpdata" area, but I could not >get the syntax right (I kept getting invalid symbol errors and what not.) >Do you remember the correct way to do this? I have tried numerous things, >including: Below is a clip out of some working code. Sorry about the delay in my original posting -- the PICLIST server rejected it for about a week before it showed up. ***** begin code section ******* #elif OSCOMPILER == OSMPLAB_C18 && OSTARGET == OSPIC18 void IntVector ( void ); /* locator for high-priority ISR */ #pragma code highVector = 0x08 void athighVector ( void ) { _asm GOTO IntVector _endasm } #pragma code /* end of ISR locator */ #pragma interrupt IntVector save=section(".tmpdata"),PROD void IntVector( void ) { if ( TMR2IE && TMR2IF ) { TMR2IF = 0; OSTimer(); } PORTB ^= 0x20; } #endif /* #if OSCOMPILER == OSHT_PICC etc. */ -- ______________________________________ Andrew E. Kalman, Ph.D. spamBeGoneaekspamBeGonepumpkininc.com -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email TakeThisOuTlistservEraseMEspam_OUTmitvma.mit.edu with SET PICList DIGEST in the body

> Below is a clip out of some working code. > > >#elif OSCOMPILER == OSMPLAB_C18 && OSTARGET == OSPIC18 >#pragma interrupt IntVector save=section(".tmpdata"),PROD > Thanks! This worked like a charm! BTW, where did you find the "section" statement? Jeff. -- http://www.piclist.com hint: To leave the PICList RemoveMEpiclist-unsubscribe-requestTakeThisOuTmitvma.mit.edu