[En-Nut-Discussion] eboot - bootloader under 4k

Bernard Fouché bernard.fouche at kuantic.com
Thu May 3 11:24:59 CEST 2007


FYI a while back I did change appload. Now I have a version that:

- fits under 4k (4068 bytes with gcc 3.4.5 and avr-libc 1.4.5)
- has a .bss segment and a .data segment as usual apps (however .data is 
unused at the moment)
- flash only modified pages from a boot to another, sparing flash write 
cycles when rebooting often. If you reboot and get the same file, no 
flashing is done.
- read jumpers on port F to select if network info from EEPROM is to be 
read before booting and to select if network bootloading is to be done. 
That allows you to bootload once and then keep your current application 
if it has to be used in a place you can't bootload.
- be able to read and flash a usual binary file or use files that have 
CRC's prefixed at the beginning of the file thanks to an external C 
tool. ('addcrc'). This is to be sure that you flash a correct file (not 
relying on UDP checksums)
- be aware of gateway issues (you can TFTP from a remote site if your 
firewall correctly manages TFTP)

I sent this to Harald a long time ago but I guess he did not get enough 
time to have a look. I just had a look recently, adding the jumper thing 
and sparing a few more bytes. If someone is interested, I can post it 
there and if these modifications are considered usefull, one may include 
them in NutOs.

 Bernard

Thiago A. Corrêa wrote:
> Hi guys,
>
>    Ever since last year
> (http://www.egnite.de/pipermail/en-nut-discussion/2005-May/004446.html)
> eboot wasn't compiling under somewhat recent gcc versions. Today I
> gave it a try to optimize a bit of the space based on the code from
> appload. The only problem is that my SP Duo died. I don't know when I
> will have one again since its currently sold out at Embedded
> creations.
>
>   So, anyway, if someone could check it out it would be great. I have
> attached a patch based on current CVS then I can commit it if it's
> reported to work.
>
>    There are still other places where the same optimizations can be
> applied further reducing the size.
>
> Regards,
>     Thiago A. Correa
> _______________________________________________
> http://lists.egnite.de/mailman/listinfo/en-nut-discussion
>
>   




More information about the En-Nut-Discussion mailing list