diff options
| author | John Ogness <john.ogness@linutronix.de> | 2019-03-28 15:29:25 +0106 |
|---|---|---|
| committer | John Ogness <john.ogness@linutronix.de> | 2019-03-28 15:29:25 +0106 |
| commit | b8f92e275f24599253d4a66c3d3d2ebd70c1eb93 (patch) | |
| tree | 152d2160845c6d9be7eb51fccdf6c23ce47816a2 | |
| parent | 115114f896fbb1a89467ea07476d2b6d1b18f1b8 (diff) | |
schulung_tools: scheduling_tests: disable rt throttling
On a machine with few CPUs, these tests could trigger RT
throttling.
Signed-off-by: John Ogness <john.ogness@linutronix.de>
| -rw-r--r-- | schulung_tools/scheduling_tests/README | 9 |
1 files changed, 7 insertions, 2 deletions
diff --git a/schulung_tools/scheduling_tests/README b/schulung_tools/scheduling_tests/README index cf6c601..7fdbbf6 100644 --- a/schulung_tools/scheduling_tests/README +++ b/schulung_tools/scheduling_tests/README @@ -1,15 +1,20 @@ +#!/bin/sh + +# disable rt throttling +echo -1 | sudo tee /proc/sys/kernel/sched_rt_runtime_us > /dev/null + # restrict current shell to CPU0 for tests # (the busy.sh script restricts tests to CPU1) taskset -p 1 $$ > /dev/null echo 'default vs. max unnice (unnice wins)' sudo ./busy.sh task1_other_default & \ - sudo nice -n -20 ./busy.sh task2_other_nice-20 + sudo nice -n -20 ./busy.sh task2_other_unnice wait echo 'max unnice vs. FIFO (FIFO wins)' sudo chrt -f 1 ./busy.sh task1_fifo_prio1 & \ - sudo nice -n -20 ./busy.sh task2_other_nice-20 + sudo nice -n -20 ./busy.sh task2_other_unnice wait echo 'FIFO vs. FIFO (first FIFO wins)' |