[PATCH 1/3] i386/smp: Define unused CPU_Interrupt_frame to fix compiler error

[Amaan Cheval]

Context: https://lists.rtems.org/pipermail/devel/2018-March/020409.html

I just realized that it may make more sense to use uint8_t and a
CPU_INTERRUPT_FRAME_SIZE of 1 instead, if we _do_ want to leave it unused
but defined. Another option I considered was simply setting a preprocessor
flag to disable the use of CPU_Interrupt_frame, despite HAS_SMP being set
(for eg. DISABLE_INTERRUPT_FRAME, so only i386 can opt out of that?)

Let me know if either of those seem any better.

On Sat, Mar 10, 2018 at 7:42 PM Amaan Cheval <amaan.cheval at gmail.com> wrote:

> CPU_INTERRUPT_FRAME_SIZE needs to also be set to allow the
RTEMS_STATIC_ASSERT
> in percpuasm.c to be fulfilled.

> Updates #3331
> ---
>   cpukit/score/cpu/i386/include/rtems/score/cpu.h | 10 +++++++++-
>   1 file changed, 9 insertions(+), 1 deletion(-)

> diff --git a/cpukit/score/cpu/i386/include/rtems/score/cpu.h
b/cpukit/score/cpu/i386/include/rtems/score/cpu.h
> index f78149c24b..b8ac218acc 100644
> --- a/cpukit/score/cpu/i386/include/rtems/score/cpu.h
> +++ b/cpukit/score/cpu/i386/include/rtems/score/cpu.h
> @@ -271,12 +271,20 @@ typedef void (*cpuExcHandlerType)
(CPU_Exception_frame*);
>   extern cpuExcHandlerType _currentExcHandler;
>   extern void rtems_exception_init_mngt(void);

> +#ifdef RTEMS_SMP
> +  #define CPU_INTERRUPT_FRAME_SIZE 4
> +#endif
> +
>   /*
>    * This port does not pass any frame info to the
>    * interrupt handler.
> + * For SMP, RTEMS assumes CPU_Interrupt_frame exists,
> + * so we make it a fixed size which is never used.
>    */

> -typedef void CPU_Interrupt_frame;
> +typedef struct {
> +  uint32_t unused_field;
> +} CPU_Interrupt_frame;

>   typedef enum {
>     I386_EXCEPTION_DIVIDE_BY_ZERO      = 0,
> --
> 2.13.0