30 May
2010
30 May
'10
4:12 p.m.
I recently did a fresh UbuntuStudio 10.4 amd64 install. I found that I
could only
start qjackctl once without rebooting. When I attempted to close jackd with
the
Quit button on qjackctl, the next attempt to run it results in the following
message:
JACK compiled with System V SHM support.
loading driver ..
Enhanced3DNow! detected
SSE2 detected
apparent rate = 48000
creating alsa driver ... hw:0|hw:0|256|2|48000|0|0|nomon|swmeter|-|32bit
control device hw:0
23:08:49.544 JACK was stopped successfully.
23:08:49.544 Post-shutdown script...
23:08:49.545 killall jackd
23:08:49.549 JACK has crashed.
jackd: no process found
23:08:49.993 Post-shutdown script terminated with exit status=256.
23:08:50.025 Could not connect to JACK server as client. - Overall operation
failed. -
Unable to connect to server. Please check the messages window for more info.
Neither "sudo kill -9 <PIDs of jackd, qjackctl and qjackctl.bin>" nor
"sudo
pkill -9 qjackctl"
nor closing the terminal in which qjackctl was executed eliminates this
problem.
Since this was a clean install with very little added beyond ambdec,
tetraproc and mhwaveedit,
I don't have any explanation for this problem except bad karma between the
current UbuntuStudio
and qjackctl. On one occasion when starting the reboot, I got an error
message asking if I
wanted to wait for process Unknown to complete.
The good side to this is that UbuntuStudio 10.4 boots up quickly.
Does anyone have any ideas?
Thanks,
John
30 May
30 May
5:01 p.m.
On 05/30/2010 05:12 PM, John Ouzts wrote:
I recently did a fresh UbuntuStudio 10.4 amd64
install. I found that
I could only
start qjackctl once without rebooting. When I attempted to close jackd
with the
Quit button on qjackctl, the next attempt to run it results in the
following message:
JACK compiled with System V SHM support.
loading driver ..
Enhanced3DNow! detected
SSE2 detected
apparent rate = 48000
creating alsa driver ... hw:0|hw:0|256|2|48000|0|0|nomon|swmeter|-|32bit
control device hw:0
23:08:49.544 JACK was stopped successfully.
23:08:49.544 Post-shutdown script...
23:08:49.545 killall jackd
23:08:49.549 JACK has crashed.
jackd: no process found
23:08:49.993 Post-shutdown script terminated with exit status=256.
23:08:50.025 Could not connect to JACK server as client. - Overall
operation failed. -
Unable to connect to server. Please check the messages window for more
info.
Neither "sudo kill -9 <PIDs of jackd, qjackctl and qjackctl.bin>" nor
"sudo pkill -9 qjackctl"
nor closing the terminal in which qjackctl was executed eliminates
this problem.
Since this was a clean install with very little added beyond ambdec,
tetraproc and mhwaveedit,
I don't have any explanation for this problem except bad karma between
the current UbuntuStudio
and qjackctl. On one occasion when starting the reboot, I got an error
message asking if I
wanted to wait for process Unknown to complete.
The good side to this is that UbuntuStudio 10.4 boots up quickly.
Does anyone have any ideas?
Thanks,
John
What output do you get when doing: lsof | grep snd
7:40 p.m.
On Sun, 30 May 2010, John Ouzts wrote:
I recently did a fresh UbuntuStudio 10.4 amd64
install. I found that I
could only
start qjackctl once without rebooting. When I attempted to close jackd with
the
Quit button on qjackctl, the next attempt to run it results in the following
message:
...
Since this was a clean install with very little added
beyond ambdec,
tetraproc and mhwaveedit,
I don't have any explanation for this problem except bad karma between the
current UbuntuStudio
and qjackctl. On one occasion when starting the reboot, I got an error
message asking if I
wanted to wait for process Unknown to complete.
Are you getting bus errors from other programs and weird window and
keyboard behavior in GNOME as well? I had something like this happening
just yesterday, reading this gave some insight:
https://bugs.launchpad.net/ubuntu/+source/jack-audio-connection-kit/+bug/49…
Exiting PA (pacmd exit) and deleting /dev/shm/pulse* before starting
qjackctl seems to do the trick for me. I upgraded my Ubuntustudio from
9.10, so YMMV.
--
Raine M. Ekman ...... raine(a)iki.fi
the Radio Project ... http://www.mikseri.net/artists/?id=36109&displ_lang=en
11:31 p.m.
On Sun, May 30, 2010 at 1:40 PM, Raine M. Ekman <raine(a)iki.fi> wrote:
Are you getting bus errors from other programs and weird window and
keyboard behavior in GNOME as well? I had something like this happening just
yesterday, reading this gave some insight:
https://bugs.launchpad.net/ubuntu/+source/jack-audio-connection-kit/+bug/49…
Exiting PA (pacmd exit) and deleting /dev/shm/pulse* before starting
qjackctl seems to do the trick for me. I upgraded my Ubuntustudio from 9.10,
so YMMV.
--
Raine M. Ekman ...... raine(a)iki.fi
the Radio Project ...
http://www.mikseri.net/artists/?id=36109&displ_lang=en
Raine,
Thank you. "du -m /dev/shm" shows 320 megabytes of memory locked out by
pulseaudio, even after its processes have been killed. So my mileage did
not vary at all.
Are you putting the "/usr/bin/pacmd exit; rm /dev/shm pulse*" script into
the qjackctl>Setup>Options box? The default script has long been
"artsshell -terminate" which I always thought was to kill the ancient KDE
arts daemon. If so, that option could use an update. I suppose the
replacement
script would need to be setuid, since at least one of the pulseaudio files
could not be removed as user. That in turn, I think, means this needs to be
a
compiled program rather than a shell script for security reasons. Does that
make sense?
Again many thanks,
John
31 May
31 May
9:49 a.m.
On Sun, 30 May 2010, John Ouzts wrote:
Are you putting the "/usr/bin/pacmd exit; rm
/dev/shm pulse*" script into
the qjackctl>Setup>Options box?
I actually only did this cleanup from the command line so far, but that is
the place where I'd put the commands. Any more commands and I'd migrate
it to ~/bin/jack-prestart.sh or something.
I suppose the replacement script would need to be
setuid, since at least
one of the pulseaudio files could not be removed as user. That in turn,
I think, means this needs to be a compiled program rather than a shell
script for security reasons. Does that make sense?
I wouldn't worry about that one file (left from playing some startup
sound?) unless it actually causes problems. I'm of course assuming a
single user setup, where there wouldn't be leftovers from other users'
sessions.
--
Raine M. Ekman tel: 0400 838 395
raine(a)iki.fi www: http://www.iki.fi/%7Eraine/
2:33 p.m.
Thanks Raine. My current understanding is that artsshell -q terminate does
in fact stop the pulseaudio processes but it does not release enough memory
(by removing the contents of /dev/shm) to allow a second startup of
qjackctl. The exact symptoms would vary with the amount of memory installed
in a given machine.
John
On Mon, May 31, 2010 at 3:49 AM, Raine M. Ekman <raine(a)iki.fi> wrote:
On Sun, 30 May 2010, John Ouzts wrote:
Are you putting the "/usr/bin/pacmd exit; rm
/dev/shm pulse*" script into
the qjackctl>Setup>Options box?
I actually only did this cleanup from the command line so far, but that is
the place where I'd put the commands. Any more commands and I'd migrate it
to ~/bin/jack-prestart.sh or something.
I suppose the replacement script would need to be setuid, since at least
one of the pulseaudio files could not be removed
as user. That in turn,
I think, means this needs to be a compiled program rather than a shell
script for security reasons. Does that make sense?
I wouldn't worry about that one file (left from playing some startup
sound?) unless it actually causes problems. I'm of course assuming a single
user setup, where there wouldn't be leftovers from other users' sessions.
--
Raine M. Ekman tel: 0400 838 395
raine(a)iki.fi www: http://www.iki.fi/%7Eraine/
8:39 p.m.
A little more digging shows that qjackctl in Ubuntu 10.4 first tries to
reroute audio through JACK and then suspending PA when starting up, so
that takes care of PA stepping on JACKs toes. All that actually is needed
to get everything working is this line in /etc/pulse/client.conf or
~/.pulse/client.conf:
enable-shm = no
--
Raine M. Ekman tel: 0400 838 395
raine(a)iki.fi www: http://www.iki.fi/%7Eraine/
1 Jun
1 Jun
5:27 a.m.
On Mon, May 31, 2010 at 2:39 PM, Raine M. Ekman <raine(a)iki.fi> wrote:
A little more digging shows that qjackctl in Ubuntu 10.4 first tries to
reroute audio through JACK and then suspending PA when starting up, so that
takes care of PA stepping on JACKs toes. All that actually is needed to get
everything working is this line in /etc/pulse/client.conf or
~/.pulse/client.conf:
enable-shm = no
Raine,
Your suggestion works like a charm. "du -m /dev/shm" shows 1 MB being used,
while "ls /dev/shm" shows two (not six) pulse files. Starting qjackctl
increases "du -m /dev/shm" to 64MB, which remain when qjackctl is killed. On
the other hand, I can restart qjackctl at least 5 times without problem
and without increasing the 64MB of memory. May I ask where you were digging
to come up with this?
Fons and Alexander,
This looks like a fairly clean solution without taking any drastic action
that might cause upgrade/dependency issues down the road.
Many thanks at all,
John
11:05 a.m.
On Mon, 31 May 2010, John Ouzts wrote:
Your suggestion works like a charm. "du -m
/dev/shm" shows 1 MB being used,
while "ls /dev/shm" shows two (not six) pulse files. Starting qjackctl
increases "du -m /dev/shm" to 64MB, which remain when qjackctl is killed. On
the other hand, I can restart qjackctl at least 5 times without problem
and without increasing the 64MB of memory. May I ask where you were digging
to come up with this?
/usr/binqjackctl is a shell script wrapper (at least on Ubuntu), easy to
follow the startup logic there. And for the PA config, "man -k pulse" or a
Google search on "pulseaudio shm" gives good leads to follow.
--
Raine M. Ekman tel: 0400 838 395
raine(a)iki.fi www: http://www.iki.fi/%7Eraine/
31 May
31 May
10:47 p.m.
On Sun, May 30, 2010 at 05:31:46PM -0400, John Ouzts wrote:
Thank you. "du -m /dev/shm" shows 320
megabytes of memory locked out by
pulseaudio, even after its processes have been killed. So my mileage did
not vary at all.
If an audio subsystem is using 320MB of memory it's broken. Just dump it.
Ciao,
--
FA
O tu, che porte, correndo si ?
E guerra e morte !
1 Jun
1 Jun
12:19 a.m.
On 05/31/2010 10:42 PM, fons(a)kokkinizita.net wrote:
On Sun, May 30, 2010 at 05:31:46PM -0400, John Ouzts
wrote:
I think you followed Lennart's presentation @LAC2010 [1] explaining the
motivation for this: Pulseaudio keeps buffers sometimes 10 secs long or
even more. The reason for that is to have the CPU wake up as little as
possible in order to save power.
That PA does not release that memory is either a bug or because of the
way it got killed.
FWIW jackd also locks ~80MB in /dev/shm here.
Cheers!
robin
[1]
http://lac.linuxaudio.org/2010/recordings/day1_1400_Pro_Audio_is_Easy_Consu…
Thank you. "du -m /dev/shm" shows 320
megabytes of memory locked out by
pulseaudio, even after its processes have been killed. So my mileage did
not vary at all.
If an audio subsystem is using 320MB of memory it's broken. Just dump it.
9:53 a.m.
On Tue, Jun 01, 2010 at 12:18:57AM +0200, Robin Gareus wrote:
On 05/31/2010 10:42 PM, fons(a)kokkinizita.net wrote:
This is a bogus argument.
Keeping software buffers does not help to reduce the number of
times the CPU has to wake up - this depends *only* on how much
the audio *hardware* is able to buffer. For this reason chips
designed for portable use have something like 2 seconds of
internal buffering.
320 MB corresponds to more than ** 13 minutes ** of stereo signal
in floating point format at 48 kHz.
If an audio subsystem is using 320MB of memory
it's broken. Just dump it.
I think you followed Lennart's presentation @LAC2010 [1] explaining the
motivation for this: Pulseaudio keeps buffers sometimes 10 secs long or
even more. The reason for that is to have the CPU wake up as little as
possible in order to save power. FWIW jackd also locks ~80MB in /dev/shm here.
The computer I'm writing this on has just 512 MB. ATM it's
playing back 32 channels in Ardour, and processing them
with jconvolver. I still have 300 MB free.
Ciao,
--
FA
O tu, che porte, correndo si ?
E guerra e morte !
10:20 a.m.
fons(a)kokkinizita.net wrote:
On Tue, Jun 01, 2010 at 12:18:57AM +0200, Robin Gareus
wrote:
The first recording studio I helped assemble was a reasonably funded student
radio studios, it was fully analogue, 8-track, good mics .. at the time digital
was beginning to be available, but way beyond our means. It was possible to buy
a digital delay of maybe 10 seconds - I think they were mostly used as a buffer
on the main feed of talkback radio, or as part of very expensive effects units -
but it would have cost as much as the rest of the studio put together, which had
some reasonable gear, and substantially more than a modest sized home in the
cheapish inner suburb we were in.
Simon
On 05/31/2010 10:42 PM, fons(a)kokkinizita.net
wrote:
This is a bogus argument.
Keeping software buffers does not help to reduce the number of
times the CPU has to wake up - this depends *only* on how much
the audio *hardware* is able to buffer. For this reason chips
designed for portable use have something like 2 seconds of
internal buffering.
320 MB corresponds to more than ** 13 minutes ** of stereo signal
in floating point format at 48 kHz. If an audio subsystem is using 320MB of memory
it's broken. Just dump it.
I think you followed Lennart's presentation @LAC2010 [1] explaining the
motivation for this: Pulseaudio keeps buffers sometimes 10 secs long or
even more. The reason for that is to have the CPU wake up as little as
possible in order to save power.
5089
days inactive
5091
days old
linux-audio-user@lists.linuxaudio.org
12 comments
6 participants
participants (6)
-
alexander -
fons＠kokkinizita.net -
John Ouzts -
Raine M. Ekman -
Robin Gareus -
Simon Wise