From: Con Kolivas Rework the priority matrix used by the staircase-deadline cpu scheduler. Ensuring every nice level uses priority slot 139 means that niced tasks will not cause expiration prematurely for less niced tasks that have been sleeping. This also increases the frequency less niced tasks preempt niced tasks, and simplifies greatly the code that generates the priority matrix. Update the documentation accordingly and explain why the priority matrix exists in the first place. Signed-off-by: Con Kolivas Signed-off-by: Andrew Morton --- Documentation/sched-design.txt | 13 +++++++------ 1 files changed, 7 insertions(+), 6 deletions(-) diff -puN Documentation/sched-design.txt~sched-implement-staircase-deadline-scheduler-rework-priomatrix-doc Documentation/sched-design.txt --- a/Documentation/sched-design.txt~sched-implement-staircase-deadline-scheduler-rework-priomatrix-doc +++ a/Documentation/sched-design.txt @@ -268,13 +268,14 @@ there are 40 priority slots where a task and the allocation of slots is dependant on nice level. In the following table, a zero represents a slot where the task may run. +PRIORITY:0..................20.................39 nice -20 0000000000000000000000000000000000000000 -nice -10 1001000100100010001001000100010010001000 -nice 0 0101010101010101010101010101010101010101 -nice 5 1101011010110101101011010110101101011011 -nice 10 0110111011011101110110111011101101110111 -nice 15 0111110111111011111101111101111110111111 -nice 19 1111111111111111111011111111111111111111 +nice -10 1000100010001000100010001000100010010000 +nice 0 1010101010101010101010101010101010101010 +nice 5 1011010110110101101101011011010110110110 +nice 10 1110111011101110111011101110111011101110 +nice 15 1111111011111110111111101111111011111110 +nice 19 1111111111111111111111111111111111111110 As can be seen, a nice -20 task runs in every priority slot whereas a nice 19 task only runs one slot per major rotation. This dithered table allows for the _