avr-libc-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [avr-libc-dev] Adding new device to LibC


From: Colin O'Flynn
Subject: Re: [avr-libc-dev] Adding new device to LibC
Date: Tue, 14 Jan 2003 18:47:40 -0800 (PST)

Hi,

Well the AVR core has a massive 2 Kbyte program memory
so its all well within an RJMP...

As it happened the number I used (6) ended up being
right :p Yes thats right it has two interrupts.. and
neither of them are on any other AVR's that I know of!

So what should I do now with this file, post it too
this list? I have to say that the core is like a
'classic' AVR (AT90S2313 for example), its
peripherials are quite different... it doesn't even
have a PORTD or something, it calls it IO_DATOUT (eqiv
of PORTD) and IO_DATIN (equiv of PIND). 

I got the .o file to be created by adding
crtat86rf401.o to the line that says "AVR_CRT_OTHER ="
in the crt1\Makefile file, and adding the same thing
to the same line the the Makefile in the main
directory...

Actually my testing of it consisted of just changing
the io.h in my current avr-gcc, and copying over the
new .h and .o files... although everything compiled OK
I only used the new .o and .h files.. not sure if I
should have...

BTW this might be a dumb question, but for the
"Copyright" at the top of the device header file, do I
just leave that guys name or do I put mine? I'm not
sure what the normal thing to do is in this project..

 Thanks again!

   -Colin

PS: I don't have the silicon yet to do a full test of
the device header file...

--- "Theodore A. Roth" <address@hidden> wrote:
> 
> 
> On Tue, 14 Jan 2003, Colin O'Flynn wrote:
> 
> :) Hi again,
> :)
> :) Just a quick update:
> :)
> :) I've managed to build the .o file for the
> at86rf401,
> :) and compiled a program for with with AVR-GCC...
> so I
> :) think it should work.
> :)
> :) I'm still not sure what to put in #define
> :) _VECTORS_SIZE though, I just put a random number
> to
> :) get it to compile...
> 
> There's two cases:
> 
>   (num_vectors + 1) * 4
> 
> or
> 
>   (num_vectors + 1) * 2
> 
> Use 4 if the vector table uses jmp and 2 if it uses
> rjmp. You can figure
> this out by disassembling your resulting object
> code.
> 
> Hope that helps.
> 
> Send us a patch when you get it working. ;-)
> 
> Ted Roth
> 





reply via email to

[Prev in Thread] Current Thread [Next in Thread]