RE: [avr-libc-dev] fuse API on the xmega16a4

[Weddington, Eric]

> -----Original Message-----
> From: 
> address@hidden 
> [mailto:address@hidden
> org] On Behalf Of Galen Seitz
> Sent: Saturday, March 13, 2010 11:56 AM
> To: address@hidden
> Subject: Re: [avr-libc-dev] fuse API on the xmega16a4

> > My guess is that the API is intentionally different. 
> However, it's a bug in that avr-libc is not correctly 
> accounting for the differences. That's something we need to fix.
> 
> Would something like this be appropriate?
> 
> --- /usr/avr/include/avr/fuse.h 2010-03-10 16:32:49.000000000 -0800
> +++ fuse.h      2010-03-13 10:54:26.000000000 -0800
> @@ -261,8 +261,12 @@
>   #endif
> 
>   #ifndef FUSES
> +#ifdef __AVR_XMEGA__
> +#define FUSES NVM_FUSES_t __fuse FUSEMEM
> +#else
>   #define FUSES __fuse_t __fuse FUSEMEM
>   #endif
> +#endif
> 
>   #endif /* !__ASSEMBLER__ */
> 

Yes, absolutely! That's exactly what I was thinking of.

I have an additional problem though, that I noticed last night: I'm in the 
middle of adding support for a new xmega device, the ATxmega128A1U, which is a 
variant on the regular xmega128A1, and I have the header file generated and 
committed. I noticed that this header file does not contain an NVM_FUSES_t 
structure. It's the only xmega header file that doesn't have this structure. So 
I think that there might be something wrong with the XML device file (which is 
the source for the generated header file), that it might have this structure 
missing. I have to go track that down on my side...

Please let me know if you have any further suggestions or patches.

Thanks,
Eric Weddington