2008-04-18 06:20:33     Adeos/Xenomai on BF527

Document created by Aaronwu Employee on Aug 6, 2013
Version 1Show Document
  • View in full screen mode

2008-04-18 06:20:33     Adeos/Xenomai on BF527

Frans Klaver (NETHERLANDS)

Message: 54473    Are there any views on when the ipipe patch will be suitable for the bf52x series?

QuoteReplyEditDelete

 

 

2008-04-18 12:01:34     Re: Adeos/Xenomai on BF527

Robin Getz (UNITED STATES)

Message: 54484    Frans:

 

I take it by this, you tried it, and it didn't work?

 

-Robin

QuoteReplyEditDelete

 

 

2008-04-21 03:45:58     Re: Adeos/Xenomai on BF527

Frans Klaver (NETHERLANDS)

Message: 54591    It doesn't work because it isn't yet made for it. The entire patch depends on things that are only defined for BF533,537 and 561. During editing I stumbled upon #define PRIO_GPIODEMUX(irq) ... (include/asm-blackfin/ipipe.h), which is set to different values for every processor. I'm in the dark as to what it should be set to in the 527.

 

Edit: The code snippet from ipipe.h in question is:

 

#if defined(CONFIG_BF533)

#define IRQ_SYSTMR      IRQ_TMR0

#define IRQ_PRIOTMR     CONFIG_TIMER0

#define PRIO_GPIODEMUX(irq) CONFIG_PFA

#elif defined(CONFIG_BF537)

#define IRQ_SYSTMR      IRQ_TMR0

#define IRQ_PRIOTMR     CONFIG_IRQ_TMR0

#define PRIO_GPIODEMUX(irq) CONFIG_IRQ_PROG_INTA

#elif defined(CONFIG_BF561)

#define IRQ_SYSTMR      IRQ_TIMER0

#define IRQ_PRIOTMR     CONFIG_IRQ_TIMER0

#define PRIO_GPIODEMUX(irq) ((irq) == IRQ_PROG0_INTA ? CONFIG_IRQ_PROG0_INTA : \

                 (irq) == IRQ_PROG1_INTA ? CONFIG_IRQ_PROG1_INTA : \

                 CONFIG_IRQ_PROG2_INTA)

#define bfin_write_TIMER_DISABLE(val)   bfin_write_TMRS8_DISABLE(val)

#define bfin_write_TIMER_ENABLE(val)    bfin_write_TMRS8_ENABLE(val)

#define bfin_write_TIMER_STATUS(val)    bfin_write_TMRS8_STATUS(val)

#endif

 

Adding to my confusion here is that I do not understand the use of CONFIG_PFA and CONFIG_IRQ_PROG_INTA, as neither is defined anywhere I can find, as opposed to PFA and IRQ_PROG_INTA.

QuoteReplyEditDelete

 

 

2008-04-21 05:29:58     Re: Adeos/Xenomai on BF527

Yi Li (CHINA)

Message: 54603    Adeos/Xenomai is not ported to BF52x, BF54x yet. Those macros are for BF561, I will check BF561 here.

QuoteReplyEditDelete

 

 

2008-04-21 05:43:15     Re: Adeos/Xenomai on BF527

Yi Li (CHINA)

Message: 54607    Unfortunately, Adeos on 2008R1 cannot be built on BF561, BF52x, and BF548 - I may need to port it.

QuoteReplyEditDelete

 

 

2008-04-21 05:44:04     Re: Adeos/Xenomai on BF527

Frans Klaver (NETHERLANDS)

Message: 54608    I've been mocking around with the code a bit. I got the kernel to compile (with a whole bunch of warnings) and it very effectively failed to boot. Will require a bit more than 2 hours of work.

QuoteReplyEditDelete

 

 

2008-04-24 09:29:29     Re: Adeos/Xenomai on BF527

Frans Klaver (NETHERLANDS)

Message: 54863    I've made some more changes to the patched code. By now linux is trying to boot, but fails after the memory initialization. There is no kernel panic, though. It seems as if something is just blocking the boot progress.

Attachments

    Outcomes