open /dev/snd/seq failed

open /dev/snd/seq failed

[John Rigg]

I just compiled alsa-lib, alsa-tools and alsa-utils 1.0.11
on a new Debian Etch amd64 installation and keep getting the following error
message from apps that try to connect to the sequencer:

ALSA lib seq_hw.c:457:(snd_seq_hw_open) open /dev/snd/seq failed: No such file
or directory
can't open sequencer

This is with Debian 2.6.15-1 kernel and with kernel.org 2.6.16.20 (compiled
using Debian .config but with CONFIG_PREEMPT=y).

Contents of /dev/snd:

controlC0 controlC1 controlC2 midiC0D0 midiC1D0 midiC2D0
pcmC0D0c pcmC0D0p pcmC1D0c pcmC1D0p pcmC2D0c pcmC2D0p timer

lsmod:

snd_ice1712            67360  0 
snd_ice17xx_ak4xxx      6720  1 snd_ice1712
snd_ak4xxx_adda         9216  2 snd_ice1712,snd_ice17xx_ak4xxx
snd_cs8427             12672  1 snd_ice1712
snd_ac97_codec        103868  1 snd_ice1712
snd_pcm                96008  2 snd_ice1712,snd_ac97_codec
snd_timer              27336  1 snd_pcm
snd_page_alloc         13904  1 snd_pcm
snd_ac97_bus            4928  1 snd_ac97_codec
snd_i2c                 8192  2 snd_ice1712,snd_cs8427
snd_mpu401_uart        10432  1 snd_ice1712
snd_rawmidi            29664  1 snd_mpu401_uart
snd_seq_device         11792  1 snd_rawmidi
joydev                 13120  0 
evdev                  13952  0 
shpchp                 47008  0 
snd                    62304  10 snd_ice1712,snd_ak4xxx_adda,snd_cs8427,snd_ac97_codec,snd_pcm,snd_timer,snd_i2c,snd_mpu401_uart,snd_rawmidi,snd_seq_device

There are three ice1712 cards (Delta 1010s) installed.
The kernel config contains the following:

# Linux kernel version: 2.6.15-1-amd64-k8-smp
...
CONFIG_SND=m
CONFIG_SND_AC97_CODEC=m
CONFIG_SND_AC97_BUS=m
CONFIG_SND_TIMER=m
CONFIG_SND_PCM=m
CONFIG_SND_HWDEP=m
CONFIG_SND_RAWMIDI=m
CONFIG_SND_SEQUENCER=m
CONFIG_SND_SEQ_DUMMY=m
CONFIG_SND_OSSEMUL=y
CONFIG_SND_MIXER_OSS=m
CONFIG_SND_PCM_OSS=m
CONFIG_SND_SEQUENCER_OSS=y
CONFIG_SND_RTCTIMER=m
CONFIG_SND_SEQ_RTCTIMER_DEFAULT=y
# CONFIG_SND_VERBOSE_PRINTK is not set
# CONFIG_SND_DEBUG is not set
CONFIG_SND_GENERIC_DRIVER=y
#
# Generic devices
#
CONFIG_SND_MPU401_UART=m
CONFIG_SND_OPL3_LIB=m
CONFIG_SND_VX_LIB=m
CONFIG_SND_DUMMY=m
CONFIG_SND_VIRMIDI=m
CONFIG_SND_MTPAV=m
CONFIG_SND_SERIAL_U16550=m
CONFIG_SND_MPU401=m
...

Everything (jackd, Ardour) works fine except the sequencer.
Although this is AMD64 SMP the RTCs sync properly (two single core
Opterons, not dual core). Any ideas?


John

Re: open /dev/snd/seq failed

[John Rigg]

On Sun, Jun 11, 2006 at 02:21:57PM +0100, John Rigg wrote:
> Although this is AMD64 SMP the RTCs sync properly

Sorry, typo. The TSCs sync properly.

John

Re: open /dev/snd/seq failed

[Lee Revell]

On Sun, 2006-06-11 at 14:21 +0100, John Rigg wrote:
> snd_ice1712            67360  0 
> snd_ice17xx_ak4xxx      6720  1 snd_ice1712
> snd_ak4xxx_adda         9216  2 snd_ice1712,snd_ice17xx_ak4xxx
> snd_cs8427             12672  1 snd_ice1712
> snd_ac97_codec        103868  1 snd_ice1712
> snd_pcm                96008  2 snd_ice1712,snd_ac97_codec
> snd_timer              27336  1 snd_pcm
> snd_page_alloc         13904  1 snd_pcm
> snd_ac97_bus            4928  1 snd_ac97_codec
> snd_i2c                 8192  2 snd_ice1712,snd_cs8427
> snd_mpu401_uart        10432  1 snd_ice1712
> snd_rawmidi            29664  1 snd_mpu401_uart
> snd_seq_device         11792  1 snd_rawmidi
> joydev                 13120  0 
> evdev                  13952  0 
> shpchp                 47008  0 
> snd                    62304  10 snd_ice1712,snd_ak4xxx_adda,snd_cs8427,snd_ac97_codec,snd_pcm,snd_timer,snd_i2c,snd_mpu401_uart,snd_rawmidi,snd_seq_device

Of course it won't work, snd-seq-midi is not loaded.

Lee

Re: open /dev/snd/seq failed

[John Rigg]

On Sun, Jun 11, 2006 at 01:11:00PM -0400, Lee Revell wrote:
> Of course it won't work, snd-seq-midi is not loaded.

Thanks. Stupidity caused by lack of sleep.

John

Re: open /dev/snd/seq failed

[Lee Revell]

On Sun, 2006-06-11 at 19:30 +0100, John Rigg wrote:
> On Sun, Jun 11, 2006 at 01:11:00PM -0400, Lee Revell wrote:
> > Of course it won't work, snd-seq-midi is not loaded.
> 
> Thanks. Stupidity caused by lack of sleep.

I can't see why udev does not load this module on demand when the
sequencer is first used.

Lee

Re: open /dev/snd/seq failed

[John Rigg]

On Sun, Jun 11, 2006 at 03:51:01PM -0400, Lee Revell wrote:
> I can't see why udev does not load this module on demand when the
> sequencer is first used.

Debian testing and unstable kernels are now using initramfs instead
of initrd. According to initramfs docs there are possible issues
with hotplug. I fixed the problem temporarily by adding snd_seq_midi
to /etc/modules, but I'll try changing back to initrd.

John

Re: open /dev/snd/seq failed

[Lee Revell]

On Mon, 2006-06-12 at 09:09 +0100, John Rigg wrote:
> On Sun, Jun 11, 2006 at 03:51:01PM -0400, Lee Revell wrote:
> > I can't see why udev does not load this module on demand when the
> > sequencer is first used.
> 
> Debian testing and unstable kernels are now using initramfs instead
> of initrd. According to initramfs docs there are possible issues
> with hotplug. I fixed the problem temporarily by adding snd_seq_midi
> to /etc/modules, but I'll try changing back to initrd.

What do initramfs and initrd have to do with it?  I do not use either
one, and snd-seq-midi does not autoload for me.

Lee

Re: open /dev/snd/seq failed

[John Rigg]

On Mon, Jun 12, 2006 at 10:51:02AM -0400, Lee Revell wrote:
> On Mon, 2006-06-12 at 09:09 +0100, John Rigg wrote:
> > On Sun, Jun 11, 2006 at 03:51:01PM -0400, Lee Revell wrote:
> > > I can't see why udev does not load this module on demand when the
> > > sequencer is first used.
> > 
> > Debian testing and unstable kernels are now using initramfs instead
> > of initrd. According to initramfs docs there are possible issues
> > with hotplug. I fixed the problem temporarily by adding snd_seq_midi
> > to /etc/modules, but I'll try changing back to initrd.
> 
> What do initramfs and initrd have to do with it?

Nothing, apparently. Just tried plain initrd and it's the same.

John

Re: open /dev/snd/seq failed

[Takashi Iwai]

At Sun, 11 Jun 2006 15:51:01 -0400,
Lee Revell wrote:
> 
> On Sun, 2006-06-11 at 19:30 +0100, John Rigg wrote:
> > On Sun, Jun 11, 2006 at 01:11:00PM -0400, Lee Revell wrote:
> > > Of course it won't work, snd-seq-midi is not loaded.
> > 
> > Thanks. Stupidity caused by lack of sleep.
> 
> I can't see why udev does not load this module on demand when the
> sequencer is first used.

It's a chicken-and-egg problem.  Since the device files doesn't exist,
it can't be triggered.


Takashi

Re: open /dev/snd/seq failed

[Matthias Koenig]

Lee Revell <rlrevell@joe-job.com> writes:
> On Sun, 2006-06-11 at 19:30 +0100, John Rigg wrote:
>> On Sun, Jun 11, 2006 at 01:11:00PM -0400, Lee Revell wrote:
>> > Of course it won't work, snd-seq-midi is not loaded.
>> 
>> Thanks. Stupidity caused by lack of sleep.
>
> I can't see why udev does not load this module on demand when the
> sequencer is first used.

I don't think this can be done in this way you mention, because udev 
actually creates the device nodes. 
But of course a missing sequencer module has bothered me too.

One possibility would be to add some modprobe configuration
to load snd-seq-midi when your MIDI device driver gets loaded.
E.g. If you want to load snd-seq-midi automatically with loading 
of snd-foo, add this to some file in /etc/modprobe.d/:

install snd-foo /sbin/modprobe --ignore-install snd-foo && \
{ /sbin/modprobe\ --quiet snd-seq-midi ; }

If snd-foo gets loaded, it also loads snd-seq-midi. But somehow
I don't like this solution.

Another solution could trigger the loading of snd-seq-midi by
udev at Midi device detection time -- the time /dev/snd/midiC*
get created, we could start a script, which loads snd-seq-midi.
The udev rule could be something like:
KERNEL=="midiC*", RUN+="/path/to/alsa-seq-load"

Matthias