[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Priority Aging and RT

This page is part of the web mail archives of SRFI 18 from before July 7th, 2015. The new archives for SRFI 18 contain all messages, not just those from before July 7th, 2015.



One of the discussion items is the issue of priority aging. While I do
not object to it, I suggest that it not be included as a specified
part of SRFI-18 on the basis that priority aging conflicts with hard
RT and hence would preclude a RT-SRFI. Or at the very least, priority
aging should be a thread specific (rather than global) option that can
be ignored in RT systems.

Note: QNX, a real-time operating system, allows tasks to have their
priorities aged by 1 if they exceed their quantum. However, tasks may
miss their deadlines unless the priority levels immediately below the
tasks employing priority aging are empty (effectively reducing the
number of levels by 2 in the worst case) and the amount of execution
time of the tasks with aging is half of the maximum that they can have
without missing their deadlines (effectively reducing utilization by 2
in the worst case). For all but the most trivial systems, the design
of the system would be complex. Coupled with the priority level and
utilization penalties, aging makes no sense for RT systems.

Mark

-- 
Mark K. Gardner
RADIANT Team
Network Engineering, CIC-5
Los Alamos National Laboratory
P.O. Box 1663, M.S. D451
Los Alamos, NM 87545
Email: mkg@xxxxxxxx
Phone: 1-505-665-4953
--