Re: [ECOS] Create interrupt problem

[Gary Thomas <gary at mlbassoc dot com>]

On Tue, 2004-10-05 at 08:24, Vikas K. Prasad wrote:
> Hi,
> 
> I was unable to attach an ISR properly.  I am  using a custom
> board (with excalibur on it) on which the basic ecos features
> are working and now I am testing the interrupts.
> 
> I used the following command:
> cyg_interrupt_create((cyg_vector_t)TIMER1_VECTOR, /* Vector           */
>                      (cyg_priority_t)TIMER1_PRI,  /* Priority         */
>                      TIMER1_DAT,                  /* Interrupt data   */
>                      (cyg_ISR_t*)timer1_isr,      /* ISR handler      */
>                      (cyg_DSR_t*)timer1_dsr,      /* DSR handler      */
>                      &timer1_handle,              /* Interrupt handle */
>                      &timer1_obj);
> 
> With this after execution of the command I see that some
> junk values are stored in timer1_obj.
> 
> (gdb) p/x * (cyg_interrupt*)timer1_obj
> $16 = {vector = 0xe3a00a40, priority = 0xe3a01880, isr = 0xe0410000,
>   dsr = 0xe04ff000, data = 0xe1a00000, dsr_count = 0xe59f022c,
>   next_dsr = 0xe3a01007}

This command makes no sense.  You told eCos to put the interrupt
object in 'timer1_obj', but you are asking GDB to use the contents
of 'timer1_obj' as a pointer.

You really should write:
  (gdb) p/x timer1_obj

> 
> Because of this, hal_default_isr() is called and my ISR is never
> hit. Can you please give me some clues as to where the problem
> lies ? I am suspecting that the ISR routines are not not attached
> properly. Anything that I need to specifically do in the platform
> port for this?

After you created the interrupt, did you attach it?  Did you
unmask it?  

Try looking at the myriad of uses of the cyg_interrupt_XXX()
functions in the eCos code base and read the documentation.
Just creating an interrupt object does not allow interrupts
to happen.

-- 
Gary Thomas <gary@mlbassoc.com>
MLB Associates


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss