Re: [avr-libc-dev] Change EEPROM SFR definition.

[Björn Haase]

E. Weddington wrote on Dienstag, 19. Juli 2005 17:03 :
> Joerg Wunsch wrote:
> >I wonder whether we should merge the EE* changes before 1.2.4, too,
> >mainly since these changes also fixed a bug report.  I don't see any
> >risk in them.
>
> If it fixes the problem of having eeprom routines for a number of
> devices that didn't previously have them, then I'm all for it. This
> issue is getting to be an FRB (Frequently Reported Bug).
>
> Thanks again Anatoly!
>
> Eric
>
Don't know whether Anatoly's patch as well adds support for the devices with 
different eeprom register locations. In case that it does not: Have a look at 
the patch I have posted about one week ago.

I admit that the workaround I have used in the build system that treats the 
eeprom assembler functions in the misc subdirectory differently as the other 
modules is not ideal. However, seeing that there are presently only two sets 
of different address register locations, I think that this could be a 
possible solution until a transition to true multi-libs takes place.

Yours,

Björn

BTW. I have found one issue that needs to be fixed in the patch I have posted: 
In case of calling eeprom block transfer functions with a compile-time 
constant byte count of zero, the function transfers 256 bytes instead of 
nothing. In this case one should add a check for 0 and skip all eeprom 
accesses accordingly.