I put my bootloaders at the bottom of the memory and have GOTOs at the
start of the moveable code which jump to the start of the main code, and the
interrupt service vectors.
 I have one byte at the top of the address space which is set to 0x0FF when
programmed, the bootloaders checks this, if it's set to 0x0FF it will
execute the main code, if not, it starts the bootloader.  When reprogramming
the reflash utility clears this byte first so if there are any problems it
will reboot into the bootloader.
 One thing you do have to make sure of is that you put GOTOs to all the
interrupts you are likely to use, as you can't add any more later.

 Another aproach would be to put the bootloader at the top of the code and
just jump straight to it from the reset vector, write your bootloader so
that this address cannot be modified, as well as making sure your bootloader
can't be modified.  Then you have complete freedom with interrupt routines,
I would think is a better approach in terms of flexibility.

Martin


-----Original Message-----
From: pic microcontroller discussion list
[mailto:PICLIST@MITVMA.MIT.EDU]On Behalf Of Harold Hallikainen
Sent: 06 December 2002 18:30
To: PICLIST@MITVMA.MIT.EDU
Subject: [PIC]: boot loader


     I am STARTING to look at boot loaders for an 18f6720 project that
currently is all written in C. From what I've read, everyone seems to be
putting boot loaders at the very "top" end of the memory map (high ROM
addresses). It SEEMS that since the boot code always runs at startup, we
could put it at low memory addresses and put the changing code above it,
protecting the low memory addresses. I see a problem when the interrupt
service routines move, but perhaps the ISR vectors could point to a fixed
location in the changeable code that either has the ISR or a goto to the
actual ISR, which is free to move. Putting the loader at low memory
addresses and never overwriting these would keep a corrupted download from
ever messing up the reset vector. The boot code would always execute on
reset, then goto a fixed ROM location where the application code would
ALWAYS start. There would similarly be a fixed ROM location where the ISR
would ALWAYS be located (or, in either case, a goto to the actual code would
be loc
ated at this fixed location).
     It then SEEMS I should be able to use a #pragma sectiontype to tell the
linker where to put the boot loader. I'm guessing this would best be written
in assembly so I would not have to go through C initialization before
running the boot loader. Instead, I'd goto the normal C entry point (where
stacks are set up, initialized variables are initialized, etc.) after my
bootloader timeout.
     So... where (if anywhere) am I going wrong? Is this a reasonable
approach? Or, should I put the bootloader up in high memory?

THANKS!

Harold




FCC Rules Online at http://www.hallikainen.com/FccRules/




________________________________________________________________
Sign Up for Juno Platinum Internet Access Today
Only $9.95 per month!
Visit www.juno.com

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listserv@mitvma.mit.edu with SET PICList DIGEST in the body

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listserv@mitvma.mit.edu with SET PICList DIGEST in the body