Re: [avr-libc-dev] boot.h (was: Re: avr-libc-dev archives)

[E. Weddington]

On 23 Dec 2002 at 0:45, Volkmar Dierkes wrote:

> On Tue, 17 Dec 2002 15:57:16 -0700, E. Weddington wrote:
> 
> > What I'm proposing is a *low-level* API that right now a user has to
> > write themselves, more of a boot loader *support* API:
> >
> > boot_page_fill(address, data)       // Fill the temporary page write
> > buffer boot_page_erase(address)     // Erase page
> > boot_page_write(address)            // Write page boot_mem_read(address)    
> >         //
> > Read memory byte boot_rww_enable()                  // Enable RWW section
> > boot_lock_bits_set(__lock_bits) // Set lock bits
> > boot_spm_interrupt_enable() // Enable SPM interrupt
> > boot_spm_interrupt_disable()        // Disable SPM interrupt
> > is_boot_spm_interrupt()             // Check if SPM interrupt enabled
> > boot_rww_busy()                             // Check if RWW section is busy
> > boot_spm_busy()                             // Check if SPM is busy 
> > boot_spm_busy_wait()            //
> > Wait while SPM is busy
> >
> >
> > This doesn't provide a whole bootloader; that is up to the
> > programmer to write. But it does provide the *tools* for a
> > programmer to write one. This API pretty much corresponds to the
> > Boot Loader Support section in all the AVR data sheets. Right now
> > avrlibc does not provide any such functions.
> 
> I agree with your proposal. I think I got a copy of your avr-boot.h in
> an indirect way ;-) I found a small bug because your are testing it on
> a mega128 and at the moment I try to get it working with a mega8.
> 
> Look in __boot_mem_read_normal__, change
>             : "=r" ((unsigned char)__result),  \
> into
>             : "=r" ((unsigned char)__result)   \
> 
> I deleted the comma at the end.

Thanks for catching that! Now corrected. That's what I get for cut 
and paste. ;-)

> But I have also a question. There are some differences in the 
> mega's: Especially the pagesize is different. Is there a possibility
> to include such things in the iomxxx.h (or whereever)?
> 

I didn't know the pagesize differed between processors. IMO, the io 
headers would be the place to do that.

Joerg, Ted, do you have an opinion on this?

Eric