Re: [avrdude-dev] avrdude >= 5.4, butterfly protocol and byte mode devices

[www.mikrocontroller-projekte.de]

Joerg Wunsch schrieb:
> As Jens Wilhelm Wulf wrote:
>
>   
> > in reply to myself, as I had only written this on my homepage:
> >
> > avrdude refuses to use block mode because of 
> >
> >     if (pgm->paged_write != NULL && m->page_size != 0) {
> >                                  ^^^^^^^^^^^^^^^^^^^^
> > this condition, which is necessary in avr.c, avr_write() and avr_read().
> >     
>
> Well, if you can prove that any existing AVR910 could handle block
> mode even for target AVRs that cannot handle page mode, we can remove
> this.  So far, the term is "paged_write", rather than "block_write",
> which indicates that it will program an entire page.  Obviously, this
> does not make sense if the device does not implement page writes at
> all.  
I just checked this out with AVRprog.exe and this is what AVRprog.exe 
does: the received blocksize is set to the pagesize and AVRprog.exe 
sends the Bytes in  #blocksize packages, ragardless what device is 
selected. This is also true for avr109, as this is the way where the 
bootloader tells AVRprog.exe what pagesize the target has.
So I'm sure we can remove this "double check" if it has no implications 
to other programmers then avr910/avr109/butterfly.

If one avr910 programmer whould have a different implementation, the 
blockmode can be overwritten with the -x no_blockmode. But this is very 
unlikely as I expect all avr910 programmers are working with AVRprog.exe.

Kind Regards,
Klaus