Wayne Thornton created an issue: https://gitlab.rtems.org/rtems/rtos/rtems/-/issues/5582
Assignee: Wayne Thornton ## Summary In response to @chris comment (https://gitlab.rtems.org/rtems/rtos/rtems/-/merge_requests/1193#note_150641) on my DHRL Merge Request (https://gitlab.rtems.org/rtems/rtos/rtems/-/merge_requests/1193) I will be looking at adding a barrier API to the self-contained API set in `cpukit/include/rtems/thread.h` Every time the DHRL fast-path execution loop calls `rtems_barrier_wait()`, the OS has to disable interrupts, look up that ID in the Object Manager's internal tables, validate it, and then execute the barrier logic. By embedding the barrier directly into the `rtems_dhrl_control` structure, we can interact directly with the memory address. There is zero ID lookup latency, making it beneficial for performance-critical subsystems. By implementing this, there will be a secondary benefit: we no longer need the `RTEMS_DHRL_BARRIERS_PER_INSTANCE` configuration macro in `dhrl.h`, as self-contained objects don't consume global Workspace resources. ## Steps to reproduce <!-- Pre-set options - milestone --> -- View it on GitLab: https://gitlab.rtems.org/rtems/rtos/rtems/-/issues/5582 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
