maemo.org - Talk

maemo.org - Talk (https://talk.maemo.org/index.php)
-   Nokia N900 (https://talk.maemo.org/forumdisplay.php?f=44)
-   -   Pulse Audio battery drain (occationally) (https://talk.maemo.org/showthread.php?t=47687)

thorbo 2010-03-18 15:44

Pulse Audio battery drain (occationally)
 
From time to time I notice that Pulse Audio "sticks" and runs the battery down in a short afternoon. Top/HTop shows that it is pulse audio, and if I put my ear to the speakers I can hear a slight hiss of static. I cannot kill the thread, it just restarts. When this is occurs it usually proceeded by a little double tweet or stutter in the system sound.

Has anyone else noticed this? Upon reboot it is gone, and batter consumption is per normal (which seems to be really good). Any further feedback would be appreciated.

-Thor

slender 2010-03-18 15:46

Re: Pulse Audio battery drain (occationally)
 
Pulseaudio cpu usage after phone call bug (hopefully) fixed in next FW update
https://bugs.maemo.org/show_bug.cgi?id=6868

.edit
Look from comments for workaround.

thorbo 2010-03-18 15:50

Re: Pulse Audio battery drain (occationally)
 
Thanks for the bug link, although I have had this pulseaudio issue both after phone calls, and sometimes where there is no phone call involved at all. That, frankly is the one that troubles me more. None the less, this is good information, and I guess I missed it in my search. Thanks again. -Thor

spanner 2010-03-18 17:34

Re: Pulse Audio battery drain (occationally)
 
Quote:

Originally Posted by thorbo (Post 572064)
Thanks for the bug link, although I have had this pulseaudio issue both after phone calls, and sometimes where there is no phone call involved at all.

Are you saying that this happens even when you haven't taken phone calls since the last reboot? If that's true then please file a new bug.

stas123 2010-04-02 21:24

Re: Pulse Audio battery drain (occationally)
 
Thanks for the bug report and workaround.

adhrie 2010-05-28 14:37

Re: Pulse Audio battery drain (occationally)
 
just experienced battery drain caused by pulseaudio
i'm on PR 1.2, installed titans kernel v37 but still on default settings

don't know what triggers it...
the last thing i did before i noticed a battery drain (and conky reported pulseaudio kept on using more than 5% cpu) was playing airport and brain party (both exits normally & without interruptions by calls or anything)

resolved with killall tonegend

unfortunately i haven't been able to reproduce this problem...

i thought this bug was supposed to be resolved in PR 1.2

or maybe this is a new bug??

http://i49.tinypic.com/24wcb42.png

bugelrex 2010-05-28 15:38

Re: Pulse Audio battery drain (occationally)
 
Quote:

Originally Posted by adhrie (Post 685798)
just experienced battery drain caused by pulseaudio
i'm on PR 1.2, installed titans kernel v37 but still on default settings

don't know what triggers it...
the last thing i did before i noticed a battery drain (and conky reported pulseaudio kept on using more than 5% cpu) was playing airport and brain party (both exits normally & without interruptions by calls or anything)

resolved with killall tonegend

unfortunately i haven't been able to reproduce this problem...

i thought this bug was supposed to be resolved in PR 1.2

or maybe this is a new bug??

http://i49.tinypic.com/24wcb42.png


I've been trying to keep an eye on pulseaudio also after 1.2. I usually run into it once every 2 weeks.

If it still occurs then its a big "F U" to Nokia software developers as I have to check "top" after every freakin phone call otherwise I'll have a dead battery in a few hours.

gianko 2013-10-27 12:10

Re: Pulse Audio battery drain (occationally)
 
i have exactly the same issue.i never had this issue beafore, but now i'm experiencing it often. i have latest cssu stable from a long time. i started to have this issue just few weeks ago. especially after using open media player and cutetube to watch videos. but pulseaudio stuck and cpu usage/battery drain also after phone calls or misscalls. is this bug solved? https://bugs.maemo.org/show_bug.cgi?id=6868 how can i fix the problem?

gianko 2014-04-30 01:00

Re: Pulse Audio battery drain (occationally)
 
Bump! No one have this issue of pulseaudio stuck and draining battery after a call on the N900? the bug has been marked as fixed, but on latest cssu stable is still present.

in this thread it seems it happens the same on the N9

the only solution is to kill pulseaudio everytime?

gianko 2014-05-01 14:48

Re: Pulse Audio battery drain (occationally)
 
i'd like to make a desktop launcher or a beecon widget to kill pulseaudio with one tap.

tried a launcher.desktop icon like:

[Desktop Entry]
Version=1.0
Encoding=UTF-8
Name=pulseaudiokiller
Icon=icon
Exec=/usr/bin/killall pulseaudio
Type=Application

but not working

how can i execute the killall pulseaudio? because is to be done as root

Maybe i can use Queen Beecon Widget with a command like: osso-x-term -e "root && killall pulseaudio"

but i'm not expert in shell, any suggestions?

shawnjefferson 2014-05-02 03:51

Re: Pulse Audio battery drain (occationally)
 
sudo killall pulseaudio

??

peterleinchen 2014-05-02 04:32

Re: Pulse Audio battery drain (occationally)
 
Yes.

Insert
Exec=sudo /usr/bin/killall pulseaudio
into the desktop file.

But to get that working you will need a sudoers entry. Add a file into /etc/sudsers.d (example in this directory) and run update-sudoers.
Or, if you do not know how to do, just install sudsers. But this is security mess as it will allow user to execute every binary to be executed with root rights!

gianko 2014-05-03 18:06

Re: Pulse Audio battery drain (occationally)
 
thanks.

also i found another way with queen beecon widget:

written killall pulseaudio in a file script.sh

then given to the queen beecon widget the command:

echo sh "path to script.sh" | root

btw is not clear still why the pulseaudio stays active and use cpu after phone calls...

gianko 2014-05-10 20:52

Re: Pulse Audio battery drain (occationally)
 
still investigating the issue with pulseaudio using cpu after closing a call

"powertop" , before killing pulseaudio, shows:

37% interrupt : DMA
34% interrupt : i2c_omap
21% interrupt : gp timer
2% i interrupt : mmc1

and "top" shows cpu usage around 24% for /usr/bin/pulseaudio -- system --high-priorit (user=pulse)

killing tonegend is not killing pulseaudio like said in similar bugs on bugs.maemo.org

should i open a new bug?
it needs other info like strace -p pidof pulseaudio?

gianko 2014-06-22 21:18

Re: Pulse Audio battery drain (occationally)
 
i tried to run from term
ls -thl /proc/`pidof pulseaudio`/fd

while pulseaudio stuck after a call

and i got this

Code:

lrwx------    1 pulse    pulse        64 May 10 22:06 0 -> /dev/null
lrwx------    1 pulse    pulse        64 May 10 22:06 1 -> /dev/null
lrwx------    1 pulse    pulse        64 May 10 22:06 10 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 11 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 12 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 13 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 14 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 15 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 16 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 17 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 18 -> /dev/snd/controlC0
lrwx------    1 pulse    pulse        64 May 10 22:06 19 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 2 -> /dev/null
lrwx------    1 pulse    pulse        64 May 10 22:06 20 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 21 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 22 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 23 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 24 -> socket:[24042]
lrwx------    1 pulse    pulse        64 May 10 22:06 25 -> socket:[35672]
lrwx------    1 pulse    pulse        64 May 10 22:06 26 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 27 -> socket:[24003]
lrwx------    1 pulse    pulse        64 May 10 22:06 28 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 29 -> anon_inode:[eventfd]
lr-x------    1 pulse    pulse        64 May 10 22:06 3 -> pipe:[23963]
lrwx------    1 pulse    pulse        64 May 10 22:06 30 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 31 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 32 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 33 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 34 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 35 -> /dev/cmt_speech
lrwx------    1 pulse    pulse        64 May 10 22:06 36 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 37 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 38 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 39 -> anon_inode:[eventfd]
l-wx------    1 pulse    pulse        64 May 10 22:06 4 -> pipe:[23963]
lrwx------    1 pulse    pulse        64 May 10 22:06 40 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 41 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 42 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 43 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 44 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 45 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 46 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 47 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 48 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 49 -> anon_inode:[eventfd]
lr-x------    1 pulse    pulse        64 May 10 22:06 5 -> pipe:[23964]
lrwx------    1 pulse    pulse        64 May 10 22:06 50 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 51 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 52 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 53 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 54 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 55 -> anon_inode:[eventfd]
lrwx------    1 pulse    pulse        64 May 10 22:06 56 -> socket:[24015]
lrwx------    1 pulse    pulse        64 May 10 22:06 57 -> socket:[24021]
lrwx------    1 pulse    pulse        64 May 10 22:06 58 -> socket:[24023]
lrwx------    1 pulse    pulse        64 May 10 22:06 59 -> socket:[24025]
l-wx------    1 pulse    pulse        64 May 10 22:06 6 -> pipe:[23964]
lrwx------    1 pulse    pulse        64 May 10 22:06 60 -> socket:[24028]
lrwx------    1 pulse    pulse        64 May 10 22:06 61 -> socket:[24029]
lrwx------    1 pulse    pulse        64 May 10 22:06 62 -> socket:[25535]
lrwx------    1 pulse    pulse        64 May 10 22:06 63 -> socket:[24033]
lrwx------    1 pulse    pulse        64 May 10 22:06 64 -> socket:[24036]
lrwx------    1 pulse    pulse        64 May 10 22:06 65 -> socket:[26414]
lrwx------    1 pulse    pulse        64 May 10 22:06 68 -> /dev/snd/pcmC0D0p
lrwx------    1 pulse    pulse        64 May 10 22:06 7 -> socket:[24019]
lrwx------    1 pulse    pulse        64 May 10 22:06 8 -> socket:[23985]
lrwx------    1 pulse    pulse        64 May 10 22:06 9 -> socket:[23988]

does it give any clues?

gianko 2014-07-10 14:25

Re: Pulse Audio battery drain (occationally)
 
[problem solved?]

maybe not but i have strong feelings that the issue is linked to sociality running in background.

if i activate "keep running in background" in sociality, i can notice pulseaudio stuck after a call, but if i deactivate that option, no issue after several calls

when sociality is running in background, you can see /usr/bin/sociality running via top in xterm or conky

having sociality running in background also generate issue with audio volume: for example if you have audio volume at medium level, after receiving a notification from sociality running in background (with the iphone-like sound), the audio volume automatically increase to high level (even if you are using earphones, that is bad)

somebody can confirm?


All times are GMT. The time now is 18:09.

vBulletin® Version 3.8.8