Xinhong Hu commented: 
https://gitlab.rtems.org/rtems/rtos/rtems/-/issues/3791#note_142834


Hi @gedare,

Thank you for your guidance and for pointing me to the User's Forum for GSoC 
discussions. I have moved the discussion there as requested. For full details, 
including the corrected test code and output, I have posted a comprehensive 
report in the User's Forum: [GSoC 2026 : POSIX Message Queue thread release 
order is not priority based 
(#3791)](https://users.rtems.org/t/gsoc-2026-posix-message-queue-thread-release-order-is-not-priority-based-3791/495)

Following your suggestion, I traced the `mq_send` path and revisited my test 
logic. I realized my initial misunderstanding of the POSIX-to-RTEMS priority 
mapping led to a flawed test setup. After correcting the test to ensure the 
lower RTEMS priority thread (mapped from POSIX 10) blocks first and the higher 
RTEMS priority thread (mapped from POSIX 20) blocks second, the message queue 
**does** wake the correct higher-priority thread. The only necessary change to 
guarantee POSIX compliance is setting the discipline to 
`CORE_MESSAGE_QUEUE_DISCIPLINES_PRIORITY` in `_POSIX_Message_queue_Create()` 
within `mqueueopen.c`, which I confirmed in my latest tests on RTEMS 7.

Thank you again for your help. Any further feedback would be greatly 
appreciated.

-- 
View it on GitLab: 
https://gitlab.rtems.org/rtems/rtos/rtems/-/issues/3791#note_142834
You're receiving this email because of your account on gitlab.rtems.org.


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

Reply via email to