Thanks Luca,
I'll download the svn image.

For the question related to the process part that I posted in the prevoius message,
do you have any suggestion?

Thanks
J

Hi jaws,
I'm glad that external memory works!

Regarding the kernel issue, I don't understand if the micro resets or you just see the 'Hello world' print from main process. If you see the '*** Bertos debug start ***' mark each time, then the micro is resetting itself.

If the micro doesn't reset and you only see 'Hello world', try to enable kernel signals as explained in the developer's FAQ.

I'm in a hurry, more on this later.

Cheers,
Luca

Hi Jaws and Luca,

I'm facing same behavior described by Jaws!!! Whenever timer_delay() executes, the board is resetting itself!

I'm using an Arduino Mega2560 with the sample code created with wizard (considering Arduino1280 wizard creation). It just blinks a led and prints monitor_report. But resets during timer_delay().

I created (via wizard) the same sample for Arduino UNO (considering Arduino 2009 creation), and the sample worked perfectly!

Please, do you have any idea about this problem?

Thanks in advance,

Yep! It seems to be related to EIND. It should be pushed onto stack with funcaddr for both ATmega2560/2561.

From "ATmega640/1280/1281/2560/2561 Preliminary", item "7.6 Stack Pointer":
"The Stack Pointer is decremented by one when data is pushed onto the Stack with the PUSH instruction, and it is decremented by two for ATmega640/1280/1281 and three for
ATmega2560/2561 when the return address is pushed onto the Stack with subroutine call or
interrupt. The Stack Pointer is incremented by one when data is popped from the Stack with the POP instruction, and it is incremented by two for ATmega640/1280 /1281 and three for ATmega2560/2561 when data is popped from the Stack with return from subroutine RET or return from interrupt RETI."

So, in bertos\cpu\frame.h, edit proper CPU_PUSH_CALL_FRAME to:

#if CPU_AVR_ATMEGA2560 || CPU_AVR_ATMEGA2561
#define CPU_PUSH_CALL_FRAME(sp, func) \
do { \
uint16_t funcaddr = (uint16_t)(func); \
CPU_PUSH_WORD((sp), funcaddr); \
CPU_PUSH_WORD((sp), funcaddr>>8); \
CPU_PUSH_WORD((sp), EIND); \
} while (0)
#else
#define CPU_PUSH_CALL_FRAME(sp, func) \
do { \
uint16_t funcaddr = (uint16_t)(func); \
CPU_PUSH_WORD((sp), funcaddr); \
CPU_PUSH_WORD((sp), funcaddr>>8); \
} while (0)
#endif

Don't forget:
#if CPU_AVR_ATMEGA2560 || CPU_AVR_ATMEGA2561
#include <avr/io.h>
#endif

Does anyone knows any other side effect of this 3 bytes onto stack?

Thanks,

Hi led,
I think you will need to change the context switching routine to take into account for 3 bytes stack pointer. The file is located in cpu/avr/hw/switch_ctx_avr.S

Apart from that, I don't see any other problem in the patch.

Hi Lucca,

Please, could you point me where to change that file. I reviewed it, but I didn't find where the proc address is saved/restored. I am pretty new to avr processors and its risc (or, 'risky'? ) assembly language.

Thanks in advance,

Actually, it seems that stack pointers are always 2 bytes in all Mega AVRs, but you have two extra registers to access the flash memory (RAMPZ and EIND), so I guess they both should be saved.

You could do something like this in bertos/cpu/avr/hw/switch_ctx_avr.S:


and below, when restoring the context (line 94):


Of course you will need to add a
CPU_PUSH_WORD((sp), RAMPZ); \
to account for the extra register.

I don't have a Mega2560 to test the code, I hope you can make it work correctly!
Cheers,
Luca