Package: pulseaudio
Version: 0.9.5-5
Severity: minor

I noticed today that pulseaudio daemon was spending 5% of cpu time
(this is a 200MHz pentium) constantly even when it was supposed to be
completely idle.

strace shows that

poll([{fd=3, events=POLLIN}, {fd=12, events=POLLOUT}, {fd=11, events=POLLIN}, 
{fd=14, events=POLLIN, revents=POLLIN}, {fd=13, events=POLLIN}, {fd=19, 
events=POLLIN}, {fd=18, events=POLLIN}, {fd=17, events=POLLIN}, {fd=16, 
events=POLLIN}, {fd=15, events=POLLIN}, {fd=10, events=POLLIN}, {fd=7, 
events=POLLIN}, {fd=5, events=POLLIN}], 13, 634) = 1
gettimeofday({1186470396, 895595}, NULL) = 0
ioctl(14, 0x800c4151, 0xbfaad140) = 0
gettimeofday({1186470396, 896480}, NULL) = 0

is repeated again and again. The problematic fd 14 seems to be
connected to /dev/snd/pcmC0D0c (I don't know what it is, I'm not too
familiar with ALSA yet):

$ ls -l /proc/1975/fd
total 19
lr-x------ 1 root root 64 Aug  7 10:08 0 -> /dev/null
lrwx------ 1 root root 64 Aug  7 10:08 1 -> /dev/console
lrwx------ 1 root root 64 Aug  7 10:08 10 -> socket:[5087]
lrwx------ 1 root root 64 Aug  7 10:08 11 -> /dev/snd/controlC0
lrwx------ 1 root root 64 Aug  7 10:08 12 -> /dev/snd/pcmC0D0p
lrwx------ 1 root root 64 Aug  7 10:08 13 -> /dev/snd/controlC0
lrwx------ 1 root root 64 Aug  7 10:08 14 -> /dev/snd/pcmC0D0c
lrwx------ 1 root root 64 Aug  7 10:08 15 -> socket:[5109]
lrwx------ 1 root root 64 Aug  7 10:08 16 -> socket:[5134]
lrwx------ 1 root root 64 Aug  7 10:08 17 -> socket:[5135]
lrwx------ 1 root root 64 Aug  7 10:08 18 -> socket:[5142]
lrwx------ 1 root root 64 Aug  7 10:08 19 -> socket:[5143]
lrwx------ 1 root root 64 Aug  7 10:08 2 -> /dev/console
lr-x------ 1 root root 64 Aug  7 10:08 3 -> pipe:[5076]
l-wx------ 1 root root 64 Aug  7 10:08 4 -> pipe:[5076]
lr-x------ 1 root root 64 Aug  7 10:08 5 -> pipe:[5078]
l-wx------ 1 root root 64 Aug  7 10:08 6 -> pipe:[5078]
lr-x------ 1 root root 64 Aug  7 10:08 7 -> pipe:[5079]
l-wx------ 1 root root 64 Aug  7 10:08 8 -> pipe:[5079]

This condition seem to be reproducible in the sense that restarting
pulseaudio won't make it use less cpu time. I've made minimal changes
to /etc/pulse but if you like to look at them I've uploaded the whole
/etc/pulse to

http://iki.fi/lindi/dandelion.pulse.tar.gz


best regards,
Timo Lindfors


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Reply via email to