Re: [avr-gcc-list] inline debugging stk200/stk500 and gdb 5.x

[Klaus Rudolph]

Theodore Roth wrote:
> 
> Hi,
> 
> Avr-mon needs to be updated to use the gdb remote serial protocol. The
> protocol is implemented in simulavr in the gdb.c file. I've looked at
> bringing avr-mon up-to-date with the current gdb and it looks straight
> forward but slightly involved since I didn't do a great job of
> separating gdb.c from simulavr internals.
> 
> If enough people are loud enough with requests, I will tackle it. ;-) I
> will also help anyone who wishes to work on it.
> 
Enough people... no problem, i´m going on to get 125 new EMail acounts.
And write some mails, ok? :-)

OK, if i wish to start the job, whats to do.
I have some big understanding problems in the way how gdb works
with the target implementation. There is one routine, the
exception_handler
Here all the stuff is dispatched for the different commands, or? 
Sorry, talking about the stub-file. So whats to do is "only"
bring up memory, registers and breakpoints to and from the target.
How is memory adresed? Is this like in linker script with offsets 
for I/O, RAM, EEPROM, and progmem, so that i can decide in the
handler from the delivered address from gdb which memory i have to
read/write (in/out/ EEPROM access, lpm and so on)?

Which format must i take to send/receive the registers. The inline
assembly looks a bit strange to me.


Have i start with all commands working? Is there a way to implement
first a few functions and serial io, later make stepping available and
register access? 

Do you think the usage of timer0 for interrupting after each instruction
could
work? Seen such things in avrmon, but also inline assembler, which is
not easy
readable :-(

Sorry, have not the howle day to work on it, but i hope i can spend some
ours this
week. :-)

Thanks
        Klaus
avr-gcc-list at http://avr1.org