Multiple PowerPC BSPs Fail to BUIld

[Chris Johns]

On 10/3/2024 8:40 am, Joel Sherrill wrote:
> Hi
> 
> In looking at the build sweep log, I see what looks like motorola_powerpc and
> beatnik all have a common build failure that results in 53 configurations
> failing per
> 
> https://lists.rtems.org/pipermail/build/2024-March/051698.html
> 
> This is one but there are a lot.
> 
> debug powerpc/mvme2307 build:
>       configure: /home/tester/rtems-cron-6/rtems/waf configure\
>       --prefix=/home/tester/rtems-cron-6/tools/6/bsps --top=/home/tester\
>       /rtems-cron-6/rtems --rtems-config=config-powerpc-mvme2307-debug.ini
>      error: home/tester/bsps/powerpc/include/bsp/irq_supp.h:87:65: error:
>             'rtems_irq_connect_data' {aka 'struct
>             __rtems_irq_connect_data__'} has no member named
>             'next_handler'
> 
> Anyone care to fix that?

There has been some PowerPC interrupt changes recently.

I do not have access to a mvme2700 to test these changes and awhen I can is not
known at this point in time. My concern is the interconnected relationship these
BSPs have makes them fragile. The last release that had working interrupts on
the mvme2700 was 4.10 and 5 did not work until I updated the release branch. We
had broken interrupts for years. I would like to avoid this if possible.

Chris