Post by thread:[PIC]: Addresses for 18f452

Could someone explain to me why the following code produces a warning, and the address in the goto at address 0 is 1FF rather than 3FE?? I feel like it has to do with banks, but don't understand what I have to do to make it work. (Disclaimer: I'm new at this, but I'm working hard and reading the FM into the wee hours of the morning. And I know about the interrupt address and all that -- this only purpose of the code is to test the 20-bit address for the goto). LOC OBJECT CODE LINE SOURCE TEXT VALUE 00001 LIST P=18F452, F=INHX32 00002 #include <p18f452.inc> 0000 00004 org 0x00 Warning [202] : Argument out of range. Least significant bits used. 0000 EFFF F001 00005 goto MyStart 03FE 00006 org 0x3fe 03FE 00007 MyStart: 03FE 6AE8 00008 clrf WREG 0400 6E82 00009 movwf PORTC 0402 6E94 00010 movwf TRISC -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

I think it is working correctly. I can't comment on the warning that the assembler is generating since I can't reproduce it in MPLAB 6.1. I looked at the opcode generation in a few test cases and it appears that the address in the opcode is the address in 16-bit words, not in bytes. Note that 0x1FF * 2 = 0x3FE. For example, if you write GOTO 0x004, MPASM generates EF02 F000, which goes to address 2. The data sheet doesn't seem to mention this, but it can be deduced from the fact that the PIC18 series has a 2 MB address space for program memory, but the goto instruction only uses 20 bits, enough for 1 MB. The program memory is still normally addressed in bytes so that the TBLRD and TBLWT instructions can read and write individual bytes of program memory, instead of being limited to word boundary operations. Anyway, everything is working as you want it to. Daniel Imfeld On Saturday, March 08, 2003 5:50 PM, John Nall <spam_OUTjnall01TakeThisOuTALLTEL.NET> wrote: {Quote hidden}> Could someone explain to me why the following code produces a warning, > and the address in the goto at address 0 is 1FF rather than 3FE?? I > feel like it has to do with banks, but don't understand what I have to > do to make it work. (Disclaimer: I'm new at this, but I'm working > hard and reading the FM into the wee hours of the morning. And I > know about the interrupt address and all that -- this only purpose of > the code is to test the 20-bit address for the goto). > > LOC OBJECT CODE LINE SOURCE TEXT > VALUE > > 00001 LIST P=18F452, F=INHX32 > 00002 #include <p18f452.inc> > > 0000 00004 org 0x00 > Warning [202] : Argument out of range. Least significant bits used. > 0000 EFFF F001 00005 goto MyStart > 03FE 00006 org 0x3fe > 03FE 00007 MyStart: > 03FE 6AE8 00008 clrf WREG > 0400 6E82 00009 movwf PORTC > 0402 6E94 00010 movwf TRISC -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

At 07:22 PM 3/8/2003 -0800, Daniel Imfeld wrote: > >I think it is working correctly. I can't comment on the warning that >the assembler is generating since I can't reproduce it in MPLAB 6.1. The warning is generated by gpasm. When I run the same program with mpasm (the MPLAB 6.1 version) I did not get the warning there. But since I did get the same 1FF address I assumed that perhaps the two assemblers had different default list options set. So far as your explanation, I am going to spend some time studying it this morning. At a cursory glance it looked reasonable, but I want to understand exactly what you are telling me. Guess I will then run it through the simulator and see how it does. The proof of the pudding..... :-) Thanks much, John -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

> Could someone explain to me why the following code produces a warning, > and the address in the goto at address 0 is 1FF rather than 3FE?? Remember that instructions on the 18 family always start at an even address. The low bit of the target address therefore isn't stored in the GOTO instruction. As for the assembler warning, I don't understand it. It makes no sense. It appears to me you are doing everything right (other than using absolute mode of course). It is not possible for a GOTO to be out of range on the 18 family, and I've never gotten this message. Perhaps some obscure warning bug in absolute mode? > I > feel like it has to do with banks, but don't understand what I have to > do to make it work. Banks refers to the segmentation of the data memory space, not the program memory space. ***************************************************************** Embed Inc, embedded system specialists in Littleton Massachusetts (978) 742-9014, http://www.embedinc.com -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

> The warning is generated by gpasm. Arrgh! You might have mentioned you were using non-standard tools. I wasted a bit of time trying to figure out where that warning came from because it didn't look to me like there should be a warning at all. Use the MPLAB tools, or at least point out that you're not so most of us can just hit DELETE. ***************************************************************** Embed Inc, embedded system specialists in Littleton Massachusetts (978) 742-9014, http://www.embedinc.com -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

At 09:01 AM 3/9/2003 -0500, Olin wrote: > > The warning is generated by gpasm. > >Arrgh! > >You might have mentioned you were using non-standard tools. I wasted a bit >of time trying to figure out where that warning came from because it didn't >look to me like there should be a warning at all. Use the MPLAB tools, or >at least point out that you're not so most of us can just hit DELETE. Good point, and I will do that next time. However, my question (which is why the address was 1FF rather than 3FE) would still have been there with the MPLAB tools. Your explanation (in the earlier reply) was exactly what I needed, and I do apologise for spoiling your Sunday morning with gpasm. :-) I kind of feel duty bound to experiment with the Gnu tools even though it is clear that right now the MPLAB platform is the best way to go. One of these days they may start charging for it.... Thanks again, Olin. You've helped me several times and I appreciate it a lot. Underneath that rough exterior beats a heart of pure gold! :-) John -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

At 07:22 PM 3/8/2003 -0800, Daniel Imfeld wrote: > >The data >sheet doesn't seem to mention this, but it can be deduced from the fact >that the PIC18 series has a 2 MB address space for program memory, but >the goto instruction only uses 20 bits, enough for 1 MB. Yeah, that is one of those things that after I understand it, it then is so obvious that I wonder why in the world I didn't see it to begin with? Oh well. Back to the struggle with the data sheet. My hat is REALLY off to the people who worked with PIC's back when there was no expertise around to answer such questions. :-) John -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

On Sun, 9 Mar 2003, John Nall wrote: > At 07:22 PM 3/8/2003 -0800, Daniel Imfeld wrote: > > > >I think it is working correctly. I can't comment on the warning that > >the assembler is generating since I can't reproduce it in MPLAB 6.1. > > > The warning is generated by gpasm. When I run the same program with mpasm > (the MPLAB 6.1 version) I did not get the warning there. But since I did > get the same 1FF address I assumed that perhaps the two assemblers had > different default list options set. John Don't be afraid of Olin's bitching about gpasm. I use it nearly every day for the 18f452 without any problems. If there are any spurious warnings then those warnings are in MPASM versions less than 6.0 too. If you find a bug, then either send it to me, Craig Franklin, the GNUPIC mailing list, or even this list. Scott -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics