Active Topics
-
The N900 is now 15 years old! (9)
to Nokia N900 by Arno_11 - 8 hrs ago -
Sailfish OS for the Motorola Moto G7 Power (XT1955-5) - (ocean) (4)
to SailfishOS by Maemish - 5 days, 20 hrs ago - more...
|
2016-02-18
, 18:25
|
|
Posts: 5 |
Thanked: 1 time |
Joined on Feb 2016
|
#472
|
I did it and I'm still on the same version. 
|
|
2016-02-18
, 18:26
|
|
Posts: 5 |
Thanked: 1 time |
Joined on Feb 2016
|
#473
|
Originally Posted by jellyroll
OK, I'll do it. Thanks
You need to start with
and set the latest version.Code:devel-su sed -i -e "s|^adaptation=.*$|adaptation=http://repo.merproject.org/obs/nemo:/testing:/hw:/lge:/hammerhead/sailfish_latest_armv7hl/|" /usr/share/ssu/repos.ini
It's also better to change your language settings back to English.Code:ssu release 2.0.1.7
|
|
2016-02-19
, 23:53
|
|
Posts: 426 |
Thanked: 1,812 times |
Joined on Dec 2013
|
#474
|
I created an alternative workaround for the sensor services running amok in the background and eating cpu thus battery life.
I was not so happy that I need to compile a python library for the proposed fix to test, so I took a look at the python script and wrote my own version of it in a more simpler way in bash script.
You can download it from here: https://www.dropbox.com/s/86wvsayyg5...ensors.sh?dl=0
You don't need python just execute the script and let it running.
You can also use the service for systemd and edit it to use this bash script instead of the python one. I also did that here: https://www.dropbox.com/s/xs7rkzcpdh...l.service?dl=0
(see the .service file check_sensors.sh needs to be put into /home/nemo if you don't want to change the location)
In fact the bash script is so small I can attach the 29 lines directly here. What it does should be self explanatory:
If anyone with an bugzilla account could forward this to the bugtracker it would be great. (I don't have an account there and registration is disabled)
I was not so happy that I need to compile a python library for the proposed fix to test, so I took a look at the python script and wrote my own version of it in a more simpler way in bash script.
You can download it from here: https://www.dropbox.com/s/86wvsayyg5...ensors.sh?dl=0
You don't need python just execute the script and let it running.
You can also use the service for systemd and edit it to use this bash script instead of the python one. I also did that here: https://www.dropbox.com/s/xs7rkzcpdh...l.service?dl=0
(see the .service file check_sensors.sh needs to be put into /home/nemo if you don't want to change the location)
In fact the bash script is so small I can attach the 29 lines directly here. What it does should be self explanatory:
Code:
#!/bin/bash
# This checks every 1/2 minute and kills if
# sensors.qcom process is eating more than 8% CPU for 1 minute
TERM=linux
export TERM
kill_count=0
while true; do
sensors_pid=$(ps aux | grep sensors.qcom | grep -v grep | awk '{print $2}')
cpu_usage=$(top -b -p $sensors_pid -n1 | grep $sensors_pid | awk '{print $9}')
echo $cpu_usage
echo "CPU Usage of sensors.qcom: ${cpu_usage/.*}"
if [[ ${cpu_usage/.*} -ge 8 ]]; then
if [[ $kill_count -ge 1 ]]; then
echo "CPU Usage of sensors.qcom too high restarting..."
systemctl restart sensorfwd
kill_count=0
else
echo "CPU Usage of sensors.qcom too high! Setting kill_count + 1"
kill_count=$((kill_count+1))
fi
else
echo "Nothing to do"
fi
sleep 30
done
 |
|
2016-02-20
, 07:29
|
|
Posts: 411 |
Thanked: 1,105 times |
Joined on Jan 2010
@ Europe
|
#475
|
Originally Posted by llelectronics
Done, there's a small typo in the service, killall command.
I created an alternative workaround for the sensor services running amok in the background and eating cpu thus battery life.
I was not so happy that I need to compile a python library for the proposed fix to test, so I took a look at the python script and wrote my own version of it in a more simpler way in bash script.
...
If anyone with an bugzilla account could forward this to the bugtracker it would be great. (I don't have an account there and registration is disabled)
| The Following 2 Users Say Thank You to BluesLee For This Useful Post: | ||
|
|
2016-02-20
, 09:09
|
|
Posts: 426 |
Thanked: 1,812 times |
Joined on Dec 2013
|
#476
|
Thanks fixed that typo 
| The Following User Says Thank You to llelectronics For This Useful Post: | ||
|
|
2016-02-20
, 17:54
|
|
Posts: 69 |
Thanked: 85 times |
Joined on Mar 2013
|
#477
|
i'm also looking to this sensors.qcom error except i rather would like to find the reason for its behavior.
i don't see errors when enabling debug for sensorfwd
(it are exactly the same things that appear in journalctl -af on jolla phone and on nexus5)
temporary disabled sensors.qcom in init.hammerhead.rc for now
i've seen that in cm 13 and 12.1 the build.prop fiole has changed.
it's nor longer ro.qc.sensor.xxx but it's now ro.qti.sensors.xxx
i would like to find a way to disable each specific sensors to find which one exactly is the reason for the high cpu usage
someone else a suggestion to achieve this?
regards
i don't see errors when enabling debug for sensorfwd
(it are exactly the same things that appear in journalctl -af on jolla phone and on nexus5)
temporary disabled sensors.qcom in init.hammerhead.rc for now
i've seen that in cm 13 and 12.1 the build.prop fiole has changed.
it's nor longer ro.qc.sensor.xxx but it's now ro.qti.sensors.xxx
i would like to find a way to disable each specific sensors to find which one exactly is the reason for the high cpu usage
someone else a suggestion to achieve this?
regards
|
|
2016-02-20
, 20:48
|
|
Posts: 69 |
Thanked: 85 times |
Joined on Mar 2013
|
#478
|
I think i found a solution.
I've replaced sensors.msm8974.so in /vendor/lib/hw with the one from the 6.0.1 android rom and now sensors.qcom en sensorfwd do stay below 1%
Can somenone else try this too?
(unzip attached file before copying into the proper directory)
I've replaced sensors.msm8974.so in /vendor/lib/hw with the one from the 6.0.1 android rom and now sensors.qcom en sensorfwd do stay below 1%
Can somenone else try this too?
(unzip attached file before copying into the proper directory)
|
|
2016-02-20
, 21:58
|
|
Posts: 426 |
Thanked: 1,812 times |
Joined on Dec 2013
|
#479
|
Hmm... replacing it on my system does not work.
After a restart sensors won't work at all.
Did you modified anything else to make it work ?
journalctl shows me this when trying to start sensorfwd service:
After a restart sensors won't work at all.
Did you modified anything else to make it work ?
journalctl shows me this when trying to start sensorfwd service:
Feb 20 22:54:48 Jolla sensorfwd[2268]: Node ' "proximityadaptor" ' state changed to invalid
Feb 20 22:54:48 Jolla sensorfwd[2268]: SensorManagerError: "adaptor 'proximityadaptor' can not be started"
Feb 20 22:54:48 Jolla sensorfwd[2268]: Node ' "proximitysensor" ' state changed to invalid
Feb 20 22:54:48 Jolla sensorfwd[2268]: "proximitysensor instantiation failed"
Feb 20 22:54:48 Jolla sensorfwd[2268]: SensorManagerError: "sensor has not been instantiated"
Feb 20 22:54:48 Jolla sensorfwd[2268]: SensorManagerError: "requested sensor id 'alssensor' not registered"
Feb 20 22:54:48 Jolla sensorfwd[2268]: int HybrisManager::handleForType(int) No sensor of type: 1
Feb 20 22:54:48 Jolla sensorfwd[2268]: Node ' "accelerometeradaptor" ' state changed to invalid
Feb 20 22:54:48 Jolla sensorfwd[2268]: SensorManagerError: "adaptor 'accelerometeradaptor' can not be started"
Feb 20 22:54:48 Jolla sensorfwd[2268]: ASSERT: "accelerometerAdaptor_" in file accelerometerchain.cpp, line 48
Feb 20 22:54:48 Jolla systemd[1]: sensorfwd.service: main process exited, code=killed, status=6/ABRT
|
|
2016-02-20
, 22:12
|
|
Posts: 69 |
Thanked: 85 times |
Joined on Mar 2013
|
#480
|
rebooted afterwards...
have put debug mode back to warning
changed the name of the original file to sensors.msm8974.so.bak
to be sure that the copy infiletug root worked well, maybe filepermissions?
have put debug mode back to warning
changed the name of the original file to sensors.msm8974.so.bak
to be sure that the copy infiletug root worked well, maybe filepermissions?
Last edited by jellyroll; 2016-02-18 at 17:42.