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
On 23 April 2014 17:11, 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
s/happend/happen
in the cpu runqueue.
Why don't you want all above to be present in git log ? This is all useful information and should go into logs.
On hotplug, all the tasks/works/timers/hrtimers are migrated to other online CPUs..
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 cpureturn 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 $cpustate=$(get_online $cpu)if [ $state -ne 0 ]; thenecho "fail: cpu offline failed at " $ireturn 1fiset_online $cpustate=$(get_online $cpu)if [ $state -ne 1 ]; thenecho "fail: cpu online failed at " $ireturn 1fi- done
- end=$(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
linaro-dev mailing list linaro-dev@lists.linaro.org http://lists.linaro.org/mailman/listinfo/linaro-dev
On Thursday 24 April 2014 09:33 AM, Viresh Kumar wrote:
On 23 April 2014 17:11, 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
s/happend/happen
thanks, ack
in the cpu runqueue.
Why don't you want all above to be present in git log ? This is all useful information and should go into logs.
right thanks, ack
On hotplug, all the tasks/works/timers/hrtimers are migrated to other online CPUs..
hmm thanks, was thinking what could lead to the freeze then.
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 cpureturn 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 $cpustate=$(get_online $cpu)if [ $state -ne 0 ]; thenecho "fail: cpu offline failed at " $ireturn 1fiset_online $cpustate=$(get_online $cpu)if [ $state -ne 1 ]; thenecho "fail: cpu online failed at " $ireturn 1fi- done
- end=$(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
linaro-dev mailing list linaro-dev@lists.linaro.org http://lists.linaro.org/mailman/listinfo/linaro-dev
On 24 April 2014 10:01, Sanjay Singh Rawat sanjay.rawat@linaro.org wrote:
On hotplug, all the tasks/works/timers/hrtimers are migrated to other online CPUs..
hmm thanks, was thinking what could lead to the freeze then.
Hmm.. So you are actually debugging a problem here? I thought it is yet another 'good to have' test case..
What't the problem? The 500 hotplug thing you mentioned? Please enabled CONFIG_DEBUG and show the last few logs...
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 cpureturn 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 $cpustate=$(get_online $cpu)if [ $state -ne 0 ]; thenecho "fail: cpu offline failed at " $ireturn 1fiset_online $cpustate=$(get_online $cpu)if [ $state -ne 1 ]; thenecho "fail: cpu online failed at " $ireturn 1fi- done
- end=$(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
-
Amit Kucheria -
Sanjay Singh Rawat -
Viresh Kumar