Gedare Bloom commented: 
https://gitlab.rtems.org/rtems/rtos/rtems/-/work_items/5629#note_153008


Thanks for bringing this up, it's definitely a problem. I took a quick look and 
agree, for example a question to ask is what happens in SMP builds when we use 
an RTEMS_COMPILER_MEMORY_BARRIER if __GNUC__ is not defined. It gets silently 
elided as an empty do-while statement. This propagates into _Atomic_Fence which 
becomes a nop without __GNUC__ or stdlib atomics.

I would tend to agree that the default / unsupported case should be an error, 
certainly for cases where it is a correctness problem. It could also be a 
warning in cases where the lack of support is not causing a correctness 
problem, for example in RTEMS_FALL_THROUGH the default case does nothing, but 
I'm not sure it should be considered mandatory?

-- 
View it on GitLab: 
https://gitlab.rtems.org/rtems/rtos/rtems/-/work_items/5629#note_153008
You're receiving this email because of your account on gitlab.rtems.org. 
Unsubscribe from this thread: 
https://gitlab.rtems.org/-/sent_notifications/4-awozlebcdnes262ufue6jybug-1d/unsubscribe
 | Manage all notifications: https://gitlab.rtems.org/-/profile/notifications | 
Help: https://gitlab.rtems.org/help


_______________________________________________
bugs mailing list
[email protected]
http://lists.rtems.org/mailman/listinfo/bugs

Reply via email to