12 Nov
2005
12 Nov
'05
8:40 p.m.
Hi all,
I'd be glad to hear any comments on the following annoying problem. If
we could locate and eliminate the reason for this bug, this would be
great.
I have various USB audio/midi devices:
* Midisport 8x8
* Terratec Aureon USB
* m-audio Trigger Finger
Not always but often when I disconnect them from my machine, I loose my
keyboard so I have to restart the machine. During the shutdown, the
machine hangs when shutting down hotplug.
After the reboot, some hardware is still messed, e.g. no DSL connection
can be established. I have to reboot the machine again, then it works.
I tried to monitor the messages-log and the xorg-log but I got no
valuable information. I see the following misbehaviours:
* No keyboard at all
* Konsole (the KDE console emulation) cannot be closed. It hangs
* The machine hangs at shutdown when trying to stop hotplug. I see a
kernel Ooops, and it reports a segfaut.
see http://www.christeck.de/files/DSCN1061.JPG
for details.
A thing that all used devices share is that they do not register
properly on the USB bus. Instead of registering as usb audio devices,
these register as HID (human input device) or similar on the bus.
The question is:
* Anyone else having this behaviour?
* Does anyone have a clue which software (hotplug, X.org, KDE) causes
the keyboard disconnect?
* Any hints which other software components can cause this?
I'd be very glad if we could find the reason for this behaviour and
report it to the responsible project so the devices can be
blacklisteted there.
Thanks & best regards
ce
12 Nov
12 Nov
8:59 p.m.
On Sun, 2005-11-13 at 01:40 +0100, Christoph Eckert wrote:
Hi all,
I'd be glad to hear any comments on the following annoying problem. If
we could locate and eliminate the reason for this bug, this would be
great.
I have various USB audio/midi devices:
* Midisport 8x8
* Terratec Aureon USB
* m-audio Trigger Finger
Not always but often when I disconnect them from my machine, I loose my
keyboard so I have to restart the machine. During the shutdown, the
machine hangs when shutting down hotplug.
This bug was thought to have been fixed a long, long time ago. How old
is your kernel and ALSA version?
Lee
13 Nov
13 Nov
6:18 a.m.
Hi,
This bug was thought to have been fixed a long, long
time ago. How
old is your kernel and ALSA version?
I'm still running a 2.6.10 vanilla with the Rt-LSM patch applied, ALSA
is at Driver Version 1.0.6.
Is this too old? But if it already is fixed I'd be glad.
Best regards
ce
1:17 p.m.
Lee Revell wrote:
On Sun, 2005-11-13 at 01:40 +0100, Christoph Eckert
wrote:
i have this same bug on fedora core 3 planetccrma kernel. any help
would be great...
Hi all,
I'd be glad to hear any comments on the following annoying problem. If
we could locate and eliminate the reason for this bug, this would be
great.
I have various USB audio/midi devices:
* Midisport 8x8
* Terratec Aureon USB
* m-audio Trigger Finger
Not always but often when I disconnect them from my machine, I loose my
keyboard so I have to restart the machine. During the shutdown, the
machine hangs when shutting down hotplug.
This bug was thought to have been fixed a long, long time ago. How old
is your kernel and ALSA version?
Lee
1:53 p.m.
Christoph Eckert wrote:
[nunativs@localhost ~]$ uname -a
Linux localhost.localdomain 2.6.10-2.1.ll.rhfc3.ccrma #1 Thu Dec 30
03:08:13 EST 2004 i686 athlon i386 GNU/Linux
[nunativs@localhost ~]$ cat /proc/asound/version
Advanced Linux Sound Architecture Driver Version 1.0.9rc1.
Compiled on Mar 22 2005 for kernel 2.6.10-2.1.ll.rhfc3.ccrma.
[nunativs@localhost ~]$
i have this
same bug on fedora core 3 planetccrma kernel. any help
would be great...
could you please post your kernel and ALSA version? Simply type
uname -a
cat /proc/asound/version
in a shell window.
Best regards
ce
2:13 p.m.
OK,
Linux localhost.localdomain 2.6.10-2.1.ll.rhfc3.ccrma
#1 Thu Dec 30
03:08:13 EST 2004 i686 athlon i386 GNU/Linux
[nunativs@localhost ~]$ cat /proc/asound/version
Advanced Linux Sound Architecture Driver Version 1.0.9rc1.
as Lee wrote it can be the bug is meanwhile fixed (2.6.10 is rather
dated). Maybe there's a newer Kernel version you could install?
Best regards
ce
4 p.m.
On Sunday 13 November 2005 17:45, Christoph Eckert wrote:
Hi Christoph. I'm on FC3 at the moment, only have a usb-midi-keyboard on the
USB. I tried switching the keyboard off an on. No problems with the computer
keyboard. Tried unplugging USB, but no keyboard problems. Plug it back in,
and everything works ok. This machine is on a LAN, going through a Smoothwall
Express2 firewall to a serial modem.
Kernel: 2.6.11-0.3rdt.rhfc3.ccrma
Alsa version: 1.0.9rc1
All the best. Nigel.
i have this
same bug on fedora core 3 planetccrma kernel. any help
would be great...
could you please post your kernel and ALSA version? Simply type
uname -a
cat /proc/asound/version
in a shell window.
Best regards
ce
5:55 p.m.
Hi,
Kernel: 2.6.11-0.3rdt.rhfc3.ccrma
Alsa version: 1.0.9rc1
thanks for the info.
At least it appears that the bug is solved. I hesitate to update the
kernel, because the last time it caused *lots* of troubles on my
machine.
Still dreaming about a stock kernel being perfect for audio use :) .
Best regards
ce
14 Nov
14 Nov
5:30 a.m.
On Sun, 2005-11-13 at 22:55 +0100, Christoph Eckert wrote:
Hi,
2.6.14 works great here.
BTW upgrading your ALSA modules is not enough to fix this bug - since
the bug was in the USB subsystem the kernel must be upgraded.
Lee
Kernel: 2.6.11-0.3rdt.rhfc3.ccrma
Alsa version: 1.0.9rc1
thanks for the info.
At least it appears that the bug is solved. I hesitate to update the
kernel, because the last time it caused *lots* of troubles on my
machine.
Still dreaming about a stock kernel being perfect for audio use :) .
2:02 p.m.
2.6.14 works great here.
BTW upgrading your ALSA modules is not enough to fix this bug - since
the bug was in the USB subsystem the kernel must be upgraded.
Thanks a lot for this info.
Seems I have to reserve some time on the weekend for trying a new kernel
again :) .
Best regards
ce
15 Nov
15 Nov
10:05 a.m.
On Sun, 13 Nov 2005 at 18:17 -0500, Lee Revell wrote:
On Sun, 2005-11-13 at 22:55 +0100, Christoph Eckert
wrote:
Not sure if I'm seeing the same problem, but it sounds like it might be.
My midi controller (keystation) works fine for day after day of minimal
use, and then one day I'll go to use it and it doesn't work. qjackctl
sees it and can establish connections, dmesg reports nothing unusual. It
just doesn't seem to be generating any events. I unplug its usb and it
disappears from qjackctl. I plug it back in and it reappears, but still
doesn't work. So I unplug the power and plug it back in, same thing. So
I unload snd-usb-audio and unplug the usb and plug it back in and the
same thing. Finally I give up and reboot.
Yesterday after rebooting, it still didn't work, so I unplugged all my
USB devices (keyboard, mouse, scanner, keystation) and plugged the
keystation back in, and then it worked.
Very frustrating and since dmesg doesn't say anything interesting and
afaict alsa things it's fine, I don't know where to look for more
information.
2.6.13-rt13, and the problem has been around for ages.
--
Hans Fugal ; http://hans.fugal.net
There's nothing remarkable about it. All one has to do is hit the
right keys at the right time and the instrument plays itself.
-- Johann Sebastian Bach
Hi,
2.6.14 works great here.
BTW upgrading your ALSA modules is not enough to fix this bug - since
the bug was in the USB subsystem the kernel must be upgraded. Kernel: 2.6.11-0.3rdt.rhfc3.ccrma
Alsa version: 1.0.9rc1
thanks for the info.
At least it appears that the bug is solved. I hesitate to update the
kernel, because the last time it caused *lots* of troubles on my
machine.
Still dreaming about a stock kernel being perfect for audio use :) .
4 p.m.
Hi,
Not sure if I'm seeing the same problem, but it
sounds like it might
be. My midi controller (keystation) works fine for day after day of
minimal use, and then one day I'll go to use it and it doesn't work.
qjackctl sees it and can establish connections, dmesg reports nothing
unusual. It just doesn't seem to be generating any events. I unplug
its usb and it disappears from qjackctl. I plug it back in and it
reappears, but still doesn't work. So I unplug the power and plug it
back in, same thing. So I unload snd-usb-audio and unplug the usb and
plug it back in and the same thing. Finally I give up and reboot.
did you try amidimon in a shell window (amidimon --ports all) or
kmidimon to check if there realy is no MIDI coming in?
Yesterday after rebooting, it still didn't work,
so I unplugged all
my USB devices (keyboard, mouse, scanner, keystation) and plugged the
keystation back in, and then it worked.
Just to clearify: I loose my keyboard which is a notebook keyboard
connected internally via psaux.
In your case, maybe it's a problem of coldplug vs. hotplug. I'd try to
unplug the devices during boot time and plug it in after you logged in.
Furthermore, ensure that in your kernel config, drivers, usb you have
*disabled* usb-midi and usb-audio. Both are OSS and conflict with the
ALSA snd_usb_audio which handles four USB devices perfectly on my
machine (except for the keyboard loss, of course :) .
Best regards
ce
6:04 p.m.
On Tue, 15 Nov 2005 at 21:00 +0100, Christoph Eckert wrote:
Hi,
No, but I tried aseqview. I'll try one of those next time.
Not sure if I'm seeing the same problem, but
it sounds like it might
be. My midi controller (keystation) works fine for day after day of
minimal use, and then one day I'll go to use it and it doesn't work.
qjackctl sees it and can establish connections, dmesg reports nothing
unusual. It just doesn't seem to be generating any events. I unplug
its usb and it disappears from qjackctl. I plug it back in and it
reappears, but still doesn't work. So I unplug the power and plug it
back in, same thing. So I unload snd-usb-audio and unplug the usb and
plug it back in and the same thing. Finally I give up and reboot.
did you try amidimon in a shell window (amidimon --ports all) or
kmidimon to check if there realy is no MIDI coming in? In your case, maybe it's a problem of coldplug vs.
hotplug. I'd try to
unplug the devices during boot time and plug it in after you logged in.
I think it is probably *plug related. I used to think that something in
the usb drivers was getting in a wierd state, but this last time being
able to fix it without a reboot is making me doubt that. Or it could
have been a coincidental unrelated problem.
Furthermore, ensure that in your kernel config,
drivers, usb you have
*disabled* usb-midi and usb-audio. Both are OSS and conflict with the
ALSA snd_usb_audio which handles four USB devices perfectly on my
machine (except for the keyboard loss, of course :) .
I'll double-check. Plugging and unplugging with snd-usb-audio works
great most of the time.
--
Hans Fugal ; http://hans.fugal.net
There's nothing remarkable about it. All one has to do is hit the
right keys at the right time and the instrument plays itself.
-- Johann Sebastian Bach
6:27 p.m.
I'll double-check. Plugging and unplugging with
snd-usb-audio works
great most of the time.
If you do not need hotplug for other devices, you could completely
remove them from your boot process and load snd_usb_audio statically
during boot time in one of the boot scripts. That's how I do it,
because it also allows to pass parameters to the module.
Best regards
ce
8:57 p.m.
On Tue, 15 Nov 2005 at 23:27 +0100, Christoph Eckert wrote:
Well, you can also blacklist it so hotplug ignores it. I actually do
want hotplug to do its thing though because it loads the firmware, and
it works 90% of the time. In fact, when it stops working it's usually
not even been unplugged - I suspect usb confusion when I plug/unplug
other things on the same bus.
--
Hans Fugal ; http://hans.fugal.net
There's nothing remarkable about it. All one has to do is hit the
right keys at the right time and the instrument plays itself.
-- Johann Sebastian Bach
I'll double-check. Plugging and unplugging
with snd-usb-audio works
great most of the time.
If you do not need hotplug for other devices, you could completely
remove them from your boot process and load snd_usb_audio statically
during boot time in one of the boot scripts. That's how I do it,
because it also allows to pass parameters to the module. 
10:02 p.m.
New subject: [linux-audio-user] getting around hotplug (was Re: Keyboard loss @ usb audio disconnect)
On Tue, Nov 15, 2005 at 11:27:41PM +0100, Christoph Eckert wrote:
If you do not need hotplug for other devices, you
could completely
remove them from your boot process and load snd_usb_audio statically
during boot time in one of the boot scripts. That's how I do it,
because it also allows to pass parameters to the module.
Do you know if there's a deterministic way other than hotplug to get
the dynamic device numbers that a USB device has ben assigned? If
there were, I'd abandon hotplug completely. It's horribly
overcomplicated, but I can't figure out a way to reproduce it's
ability to know what device numbers my stuff has been assigned.
Thanks!
--
Ross Vandegrift
ross(a)lug.udel.edu
"The good Christian should beware of mathematicians, and all those who
make empty prophecies. The danger already exists that the mathematicians
have made a covenant with the devil to darken the spirit and to confine
man in the bonds of Hell."
--St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
10:38 p.m.
New subject: [linux-audio-user] getting around hotplug (was Re: Keyboard loss @ usb audio disconnect)
Ross Vandegrift wrote:
On Tue, Nov 15, 2005 at 11:27:41PM +0100, Christoph
Eckert wrote:
Hi Ross,
You can use udev rules to setup devices so they always get assigned the
to same device numbers. Not sure whether this would work with audio
devices but it works fine for USB drive bays and cameras etc.
Ross. :)
If you do not need hotplug for other devices, you
could completely
remove them from your boot process and load snd_usb_audio statically
during boot time in one of the boot scripts. That's how I do it,
because it also allows to pass parameters to the module.
Do you know if there's a deterministic way other than hotplug to get
the dynamic device numbers that a USB device has ben assigned? If
there were, I'd abandon hotplug completely. It's horribly
overcomplicated, but I can't figure out a way to reproduce it's
ability to know what device numbers my stuff has been assigned.
Thanks!
16 Nov
16 Nov
1:31 p.m.
New subject: [linux-audio-user] getting around hotplug (was Re: Keyboard loss @ usb audio disconnect)
Do you know if there's a deterministic way other
than hotplug to get
the dynamic device numbers that a USB device has ben assigned? If
there were, I'd abandon hotplug completely. It's horribly
overcomplicated, but I can't figure out a way to reproduce it's
ability to know what device numbers my stuff has been assigned.
I guess you mean the order in which they appear in ALSA, is it?
http://alsa.opensrc.org/MultipleUSBAudioDevices
AFAIK udev will not help on this issue.
Best regards
ce
2:01 p.m.
New subject: [linux-audio-user] getting around hotplug (was Re: Keyboard loss @ usb audio disconnect)
On Tue, 2005-11-15 at 21:02 -0500, Ross Vandegrift wrote:
On Tue, Nov 15, 2005 at 11:27:41PM +0100, Christoph
Eckert wrote:
This sounds like a huge waste of time - rather than abandoning something
as immensely useful as hotplug, just get the bug in the USB driver
fixed! AFAICT all the people having this problem are using really old
kernels anyway. It's idiotic to abandon hotplug just because you don't
want to upgrade your kernel to fix a known bug.
Lee
If you do not need hotplug for other devices, you
could completely
remove them from your boot process and load snd_usb_audio statically
during boot time in one of the boot scripts. That's how I do it,
because it also allows to pass parameters to the module.
Do you know if there's a deterministic way other than hotplug to get
the dynamic device numbers that a USB device has ben assigned? If
there were, I'd abandon hotplug completely. It's horribly
overcomplicated, but I can't figure out a way to reproduce it's
ability to know what device numbers my stuff has been assigned.
17 Nov
17 Nov
12:50 p.m.
New subject: [linux-audio-user] getting around hotplug (was Re: Keyboard loss @ usb audio disconnect)
On Wed, Nov 16, 2005 at 12:59:14PM -0500, Lee Revell wrote:
This sounds like a huge waste of time - rather than
abandoning something
as immensely useful as hotplug, just get the bug in the USB driver
fixed! AFAICT all the people having this problem are using really old
kernels anyway. It's idiotic to abandon hotplug just because you don't
want to upgrade your kernel to fix a known bug.
Whoa, hang on. I don't have the problem identified by the original
poster. I'm just a thread hijacker ::-).
My particular bias against hotplug is two-fold:
1) It is far more difficult to understand than /etc/rc.d
2) The only USB device I have is a MIDI interface, which I
just want to leave plugged in.
#1 - sometimes, hotplug doesn't "coldplug" my device properly. I
don't really know why, and re-plugging the thing in usually works.
I don't understand why they should be different processes.
It's a minor headache that I'd rather just relieve with a shell
script. Problem is, my Midisport needs firmware loaded. I need to
make use of hotplug's firmware facilities, since I cannot predict
where in /proc/bus/usb/ the device node will appear.
#2 - I don't need hotplugging! This is a desktop machine that never
changes. I have all of one device. hotplug seems a little like
overkill in my degenerate (but probably very common) case.
Trust me - if I was losing mouse and keyboard, I'd be all over that
kernel upgrade!
--
Ross Vandegrift
ross(a)lug.udel.edu
"The good Christian should beware of mathematicians, and all those who
make empty prophecies. The danger already exists that the mathematicians
have made a covenant with the devil to darken the spirit and to confine
man in the bonds of Hell."
--St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
16 Nov
16 Nov
6:58 p.m.
Still dreaming
about a stock kernel being perfect for audio use :)
.
2.6.14 works great here.
BTW upgrading your ALSA modules is not enough to fix this bug - since
the bug was in the USB subsystem the kernel must be upgraded.
6948
days inactive
6953
days old
linux-audio-user@lists.linuxaudio.org
21 comments
7 participants
participants (7)
-
Christoph Eckert -
Hans Fugal -
Ken -
Lee Revell -
Nigel Henry -
ross＠jose.lug.udel.edu -
Rossco