Re: [avrdude-dev] Re: [bug #22248] Read efuse error

[Shaun Jackman]

On Feb 7, 2008 7:49 AM, Joerg Wunsch <address@hidden> wrote:
> As Joerg Wunsch wrote:
>
> > When browsing avrdude.conf, I noticed that we are currently not very
> > consistent about the unused bits in efuse.  Some devices always read
> > (and display) 8 bits including the unused ones, and only mask bits
> > off when writing.  Other devices (like the ATmega88/168 you are
> > perhaps using) apply the mask as applicable for the respective
> > device even when reading.  We should probably make that consistent.
>
> I'd like to hear other users'/developers' opinions about *which*
> direction to make this consistent into.  Should we:
>
> [ ]  Use 'x' consistently for all efuse bits that are not defined for
>      a particular part, or
> [ ]  always use 'i' for all bit positions of the entire efuse byte?

I prefer the 'i' for all bit positions, only because for the AT90PWM3
parts it was wrong for some time (and might still be) and took me way
too long to figure out that I couldn't write the efuse because the
mask in avrdude.conf was wrong.

Cheers,
Shaun