The solution suggested works possibly with the application code written but we have the same error using the call to the printf library functions. Typically the linker puts these functions at the end, so in the upper 128K. The trampoline fails anyway. Any chance to fix definitively this issue?
Rino Tallarico _________________________ via Pola 26 - 10135 Torino tel. +39 347 2261196 mailto: palmerinotallar...@vodafone.it -----Messaggio originale----- Da: avr-gcc-list-bounces+palmerinotallarico=vodafone...@nongnu.org [mailto:avr-gcc-list-bounces+palmerinotallarico=vodafone...@nongnu.org] Per conto di Joerg Wunsch Inviato: mercoledì 20 gennaio 2010 20.10 A: avr-gcc-list@nongnu.org Oggetto: Re: [avr-gcc-list] Function pointers on > 128 KiB devices, how to? Dale Whitfield <d...@4drealtime.co.za> wrote: > This will create an explicit trampoline. The rest of your code would > remain the same. I've used this technique in a slightly different > way previously. Caution: Compiled and inspected but not > tested/simulated! Thanks for the ideas. That's still not quite ideal for me (if possible, I'd like to hide all compiler dependencies in a single header file, as this is a multi-compiler project), but it might at least be a way around. We've also got function pointers as callbacks (so Stu's idea of putting everything into a single segment won't work anyway), which so far I don't have a real idea for. Well, maybe I should declare them as uint32_t (hidden in a typedef), and then write a FAR_CALL() macro wrapper that emits some inline asm EICALL. -- cheers, J"org .-.-. --... ...-- -.. . DL8DTL http://www.sax.de/~joerg/ NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-) _______________________________________________ AVR-GCC-list mailing list AVR-GCC-list@nongnu.org http://lists.nongnu.org/mailman/listinfo/avr-gcc-list _______________________________________________ AVR-GCC-list mailing list AVR-GCC-list@nongnu.org http://lists.nongnu.org/mailman/listinfo/avr-gcc-list
