Re: [PATCH PMQA] cpuhotplug: add subtest to stress hotplug feature

On Wed, Apr 23, 2014 at 5:11 PM, Sanjay Singh Rawat <sanjay.rawat@linaro.org

wrote:

Signed-off-by: Sanjay Singh Rawat sanjay.rawat@linaro.org

• This will help in catching issue like

https://lkml.org/lkml/2014/3/25/122

• Tested on my x86 dev m/c

Output: running hotplug stress on cpu : cpu1 pass: cpu hotplug count = 200 duration = 16.98 seconds running hotplug stress on cpu : cpu2 pass: cpu hotplug count = 200 duration = 16.36 seconds running hotplug stress on cpu : cpu3 pass: cpu hotplug count = 200 duration = 15.05 seconds

• Issue: For a hotplug count of 500 the system just hangs. Need to manually restart. Is it safe to simply plugout the cpu? What will happend to the tasks in the cpu runqueue.

---

cpuhotplug/cpuhotplug_09.sh | 70 ++++++++++++++++++++++++++++++++++++++++++ cpuhotplug/cpuhotplug_09.txt | 1 + 2 files changed, 71 insertions(+) create mode 100755 cpuhotplug/cpuhotplug_09.sh create mode 100644 cpuhotplug/cpuhotplug_09.txt

diff --git a/cpuhotplug/cpuhotplug_09.sh b/cpuhotplug/cpuhotplug_09.sh new file mode 100755 index 0000000..be0c91c --- /dev/null +++ b/cpuhotplug/cpuhotplug_09.sh @@ -0,0 +1,70 @@ +#!/bin/bash +# +# PM-QA validation test suite for the power management on Linux +# +# Copyright (C) 2014, Linaro Limited. +# +# This program is free software; you can redistribute it and/or +# modify it under the terms of the GNU General Public License +# as published by the Free Software Foundation; either version 2 +# of the License, or (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program; if not, write to the Free Software +# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. +# +# Contributors: +# Sanjay Singh Rawat sanjay.rawat@linaro.org +# - initial API and implementation +#

+# URL : https://wiki.linaro.org/WorkingGroups/PowerManagement/Resources/TestSuite/Pm...

+source ../include/functions.sh +STRESS_COUNT=200

+cpuhotplug_stress() {

• local cpu=$1
• local dirpath=$CPU_PATH/$1
• local state=
• local start=
• local end=
• shift 1
• if [ "$cpu" == "cpu0" ]; then

•   # skipping main cpu
  

•   return 0
  

• fi
• echo "running hotplug stress on cpu :" $cpu
• start=$(echo $(awk '{print $1}' /proc/uptime))
• for i in $(seq 1 $STRESS_COUNT);
• do

•    set_offline $cpu
  

•    state=$(get_online $cpu)
  

•    if [ $state -ne 0 ]; then
  

•        echo "fail: cpu offline failed at " $i
  

•       return 1
  

•    fi
  

•    set_online $cpu
  

•    state=$(get_online $cpu)
  

•    if [ $state -ne 1 ]; then
  

•        echo "fail: cpu online failed at " $i
  

•       return 1
  

•    fi
  

• done
• end=$(echo $(awk '{print $1}' /proc/uptime))
• echo "pass: cpu hotplug in/out count =" $i "duration =" $(echo

"$end-$start" | bc) "seconds"

• return 0

+}

+for_each_cpu cpuhotplug_stress +test_status_show diff --git a/cpuhotplug/cpuhotplug_09.txt b/cpuhotplug/cpuhotplug_09.txt new file mode 100644 index 0000000..5b0f8ac --- /dev/null +++ b/cpuhotplug/cpuhotplug_09.txt @@ -0,0 +1 @@

+test to stress cpu hotplug feature

1.7.10.4

This is a good start, but too regular. I would like more randomness to the hotplug process in terms of order of cpus being hotplugged in and out. I tried to come up with something long ago[1] but didn't quite finish it.

The script should start with a list of cpus, then maintain a list of online and offline cpus (can be probed repeatedly at runtime too) and then at random select either the online or offline operation and then a random cpu from that list to online or offline.

Regards, Amit

[1] https://bugs.launchpad.net/linaro-big-little-system/+bug/1097213/comments/2