Midiman Oxygen8 on Fedora 9

[Peter Howard]

My old laptop died and when setting up my new (to me) one I decided to
go straight to Fedora 9.  Which has been good up to the point I decided
to connect my Oxygen8 keyboard.  I discovered to my joy that
midisport-firmware is now packaged for Fedora so installed it.
Interestingly the current version (1.2) doesn't mention the need to
provide the actual firmware as the 0.X version did.  Anyway I plug in
the keyboard and it appears to initialise.  But it doesn't appear
anywhere.   My normal way of connecting it is via the qjackctl "connect"
window.  But it isn't there.  All I see in /var/log/messages is:

Jul  6 11:10:46 wagner kernel: hub 1-0:1.0: unable to enumerate USB device on port 3
Jul  6 11:10:47 wagner kernel: usb 3-1: new full speed USB device using uhci_hcd and address 4
Jul  6 11:10:47 wagner kernel: usb 3-1: configuration #1 chosen from 1 choice
Jul  6 11:10:47 wagner kernel: usb 3-1: New USB device found, idVendor=0763, idProduct=1014
Jul  6 11:10:47 wagner kernel: usb 3-1: New USB device strings: Mfr=0, Product=0, SerialNumber=0


sometimes accompanied by:


Jul  6 11:11:18 wagner pulseaudio[2592]: module-alsa-sink.c: Error opening PCM device front:0: Device or resource busy
Jul  6 11:17:12 wagner pulseaudio[2592]: module-alsa-sink.c: Error opening PCM device front:0: Device or resource busy


Are the pulseaudio errors relevant?
Do you still need the *.SYS firmware blobs with midisport-firmware 1.2?
I tried manually installing "ye olde" midisport-firmware 0.2 with the
firmware blobs but that didn't work either :-(

Hoping someone out there knows the answer (I also have a Midisport 4x4
and I'm really hoping to have both working soon).


--
Peter Howard <pjh@...>

_______________________________________________
PlanetCCRMA mailing list
PlanetCCRMA@...
http://ccrma-mail.stanford.edu/mailman/listinfo/planetccrma

[Ebrahim Mayat]

On Sun, 2008-07-06 at 12:14 +1000, Peter Howard wrote:

> My old laptop died and when setting up my new (to me) one I decided to
> go straight to Fedora 9.  Which has been good up to the point I decided
> to connect my Oxygen8 keyboard.  I discovered to my joy that
> midisport-firmware is now packaged for Fedora so installed it.
> Interestingly the current version (1.2) doesn't mention the need to
> provide the actual firmware as the 0.X version did.  Anyway I plug in
> the keyboard and it appears to initialise.  But it doesn't appear
> anywhere.   My normal way of connecting it is via the qjackctl "connect"
> window.  But it isn't there.  All I see in /var/log/messages is:
>
> Jul  6 11:10:46 wagner kernel: hub 1-0:1.0: unable to enumerate USB device on port 3
> Jul  6 11:10:47 wagner kernel: usb 3-1: new full speed USB device using uhci_hcd and address 4
> Jul  6 11:10:47 wagner kernel: usb 3-1: configuration #1 chosen from 1 choice
> Jul  6 11:10:47 wagner kernel: usb 3-1: New USB device found, idVendor=0763, idProduct=1014
> Jul  6 11:10:47 wagner kernel: usb 3-1: New USB device strings: Mfr=0, Product=0, SerialNumber=0
>
>
> sometimes accompanied by:
>
>
> Jul  6 11:11:18 wagner pulseaudio[2592]: module-alsa-sink.c: Error opening PCM device front:0: Device or resource busy
> Jul  6 11:17:12 wagner pulseaudio[2592]: module-alsa-sink.c: Error opening PCM device front:0: Device or resource busy
>
>
> Are the pulseaudio errors relevant?
> Do you still need the *.SYS firmware blobs with midisport-firmware 1.2?
> I tried manually installing "ye olde" midisport-firmware 0.2 with the
> firmware blobs but that didn't work either :-(
>
> Hoping someone out there knows the answer (I also have a Midisport 4x4
> and I'm really hoping to have both working soon).
>

Yes, I also have a similar problem with my MIDI interface which I posted to this list previously :

> OTOH, my MIDI is not fully functional. Virtual onboard MIDI works if
> I
> route MIDI events between applications e.g. vkeybd --> softsynth or
> aplaymidi ---> kmidimon. There was also no problem in rendering a MIDI
> file
>
> $ aplaymidi -p 129 MIDIFile.mid (129 = qsynth)
>
> BUT...I am unable to trigger MIDI events from my external MIDI
> controller.
>
> I suspect that this is because the driver for my FastLane MIDI
> interface
> is now broken on F9. (FastLane works on OS X so it's not the hardware
> device.) While this problem has been documented before, this MIDI
> interface has been working flawlessly on FC6, FC7 and FC8. So,
> something
> has indeed changed.
>
> Long story short...the hardware MIDI device is recognized in F9 when
> checked using "amidi -l", "dmesg" and "cat /proc/asound/devices".
> Also see <http://hg.alsa-project.org/alsa-kernel/log?rev=motu>
>
> The "Messages" window of qsynth, however, returns the following
> error:
> fluidsynth: warning: Could not connect to any physical jack ports;
> fluidsynth is unconnected
>
>
> All the relevent modules are loaded:
>
> # lsmod | grep snd
> snd_aoa_codec_tas      11616  2
> snd_aoa_fabric_layout    11048  3
> snd_aoa                15360  2
> snd_aoa_codec_tas,snd_aoa_fabric_layout
> snd_aoa_i2sbus         19396  1
> snd_usb_audio          83808  0
> snd_usb_lib            16928  1 snd_usb_audio
> snd_rawmidi            23168  1 snd_usb_lib
> snd_hwdep               8324  1 snd_usb_audio
> snd_powermac           43704  0
> snd_seq_dummy           2820  0
> snd_seq_oss            36692  0
> snd_seq_midi_event      6144  1 snd_seq_oss
> snd_seq                60224  5
> snd_seq_dummy,snd_seq_oss,snd_seq_midi_event
> snd_seq_device          7084  4
> snd_rawmidi,snd_seq_dummy,snd_seq_oss,snd_seq
> snd_pcm_oss            49568  0
> snd_mixer_oss          16576  2 snd_pcm_oss
> snd_pcm                77796  4
> snd_aoa_i2sbus,snd_usb_audio,snd_powermac,snd_pcm_oss
> snd_timer              21764  2 snd_seq,snd_pcm
> snd_page_alloc          8584  1 snd_pcm
> snd                    56084  21
> snd_aoa_codec_tas,snd_aoa_fabric_layout,snd_aoa,snd_aoa_i2sbus,snd_usb_audio,snd_usb_lib,snd_rawmidi,snd_hwdep,snd_powermac,snd_seq_dummy,snd_seq_oss,snd_seq,snd_seq_device,snd_pcm_oss,snd_mixer_oss,snd_pcm,snd_timer
> soundcore               5700  2 snd
> snd_aoa_soundbus        4964  2 snd_aoa_fabric_layout,snd_aoa_i2sbus
>
> The modprobe.conf looks like this:
>
> alias eth1 gem
> remove snd-powermac { /usr/sbin/alsactl store 0 >/dev/null 2>&1
> || : ; }; /sbin/modprobe -r --ignore-remove snd-powermac
> options snd cards_limit=8
> alias snd-card-0 snd-powermac
> options snd-powermac index=0
> alias snd-card-7 snd-usb-audio
> options snd-usb-audio index=7
>
> So I edited it to
>
> alias eth1 b43
> remove snd-powermac { /usr/sbin/alsactl store 0 >/dev/null 2>&1
> || : ; }; /sbin/modprobe -r --ignore-remove snd-powermac
> options snd cards_limit=8
> alias snd-card-0 snd-powermac
> options snd-powermac index=0
> alias snd-card-7 snd-usb-audio
> options snd-usb-audio index=1
>
> Still external MIDI controller events are not registered.
>
> Finally "amidi --dump" does give the following error message:
> # amidi --dump
> ALSA lib rawmidi_hw.c:233:(snd_rawmidi_hw_open) open /dev/snd/midiC0D0
> failed: No such file or directory
> cannot open port "default": No such file or directory

As a short-term (maybe long-term) fix I re-installed the old
2.6.22.10-1.rt9.1.fc8rt kernel (with which external MIDI triggering has
worked flawlessly and still works).

E

_______________________________________________
PlanetCCRMA mailing list
PlanetCCRMA@...
http://ccrma-mail.stanford.edu/mailman/listinfo/planetccrma

[Peter Howard]

On Sun, 2008-07-06 at 07:44 -0400, Ebrahim Mayat wrote:

> On Sun, 2008-07-06 at 12:14 +1000, Peter Howard wrote:
> > My old laptop died and when setting up my new (to me) one I decided to
> > go straight to Fedora 9.  Which has been good up to the point I decided
> > to connect my Oxygen8 keyboard.  I discovered to my joy that
> > midisport-firmware is now packaged for Fedora so installed it.
> > Interestingly the current version (1.2) doesn't mention the need to
> > provide the actual firmware as the 0.X version did.  Anyway I plug in
> > the keyboard and it appears to initialise.  But it doesn't appear
> > anywhere.   My normal way of connecting it is via the qjackctl "connect"
> > window.  But it isn't there.  

[snip]

> Yes, I also have a similar problem with my MIDI interface which I posted to this list previously :
>

Thanks.  This wasn't my problem (read below), but your details got me
out of my "It's stopped working!!!! I DON'T want to have to think about
this!!!!" mindset :-)

> > OTOH, my MIDI is not fully functional. Virtual onboard MIDI works if
> > I
> > route MIDI events between applications e.g. vkeybd --> softsynth or
> > aplaymidi ---> kmidimon. There was also no problem in rendering a MIDI
> > file
> >
> > $ aplaymidi -p 129 MIDIFile.mid (129 = qsynth)
> >
> > BUT...I am unable to trigger MIDI events from my external MIDI
> > controller.
> >
> > I suspect that this is because the driver for my FastLane MIDI
> > interface
> > is now broken on F9. (FastLane works on OS X so it's not the hardware
> > device.) While this problem has been documented before, this MIDI
> > interface has been working flawlessly on FC6, FC7 and FC8. So,
> > something
> > has indeed changed.
> >
> > Long story short...the hardware MIDI device is recognized in F9 when
> > checked using "amidi -l", "dmesg" and "cat /proc/asound/devices".
> > Also see <http://hg.alsa-project.org/alsa-kernel/log?rev=motu>
> >
> > The "Messages" window of qsynth, however, returns the following
> > error:
> > fluidsynth: warning: Could not connect to any physical jack ports;
> > fluidsynth is unconnected
> >
> >
> > All the relevent modules are loaded:
> >
> > # lsmod | grep snd
> >

[snip]

> > Finally "amidi --dump" does give the following error message:
> > # amidi --dump
> > ALSA lib rawmidi_hw.c:233:(snd_rawmidi_hw_open) open /dev/snd/midiC0D0
> > failed: No such file or directory
> > cannot open port "default": No such file or directory
>

The fact modules _were_ missing in my list, and the fact that "amidi -l"
crashed, let alone "amidi --dump" confirmed to me that the keyboard
_wasn't_ being initialised.  So then I did what I should have done at
the start - cross check the Vendor/ProdID values given out
of /proc/bus/usb/devices against what was
in /etc/udev/rules.d/81-midisport.rules (as provided by the
midisport-firmware RPM).  And, lo and behold, the Oxygen8 values weren't
in there.

'DOH!

pulling the values out of ye olde midisport_firmware 0.2 (I've confirmed
they're also in the 1.2 you pull directly from sourceforge) and manually
adding them to the rules file solved my problem.  

So from that I can answer my own questions (plus some more)
      * The old hotplug-based midisport-firmare releases (0.x) don't
        work at all with Fedora 9
      * It appears that you no longer need to separately supply the
        firmware blobs - they're there in the 1.X series of
        midisport-firmware.
      * The rules file included in the midisport-firmware RPM for F9
        lacks most of the devices listed in the upstream package.



> As a short-term (maybe long-term) fix I re-installed the old
> 2.6.22.10-1.rt9.1.fc8rt kernel (with which external MIDI triggering has
> worked flawlessly and still works).

Once it's configured properly, the keyboard is working fine with the
2.6.24.7-1.rt3.2.fc9.ccrmart kernel.  Don't know if it's relevant to you
or not, but if you were failing on earlier kernels it's worth a try
(note: as of right now all I've done is play the keyboard - no fiddling
with wheels or controllers - just notes).

Anyway thanks again for getting me thinking in the right direction.

PJH


_______________________________________________
PlanetCCRMA mailing list
PlanetCCRMA@...
http://ccrma-mail.stanford.edu/mailman/listinfo/planetccrma

[Fernando Lopez-Lezcano]

On Thu, 2008-07-10 at 10:10 +1000, Peter Howard wrote:

> The fact modules _were_ missing in my list, and the fact that "amidi -l"
> crashed, let alone "amidi --dump" confirmed to me that the keyboard
> _wasn't_ being initialised.  So then I did what I should have done at
> the start - cross check the Vendor/ProdID values given out
> of /proc/bus/usb/devices against what was
> in /etc/udev/rules.d/81-midisport.rules (as provided by the
> midisport-firmware RPM).  And, lo and behold, the Oxygen8 values weren't
> in there.
>
> 'DOH!
>
> pulling the values out of ye olde midisport_firmware 0.2 (I've confirmed
> they're also in the 1.2 you pull directly from sourceforge) and manually
> adding them to the rules file solved my problem.  
>
> So from that I can answer my own questions (plus some more)
>       * The old hotplug-based midisport-firmare releases (0.x) don't
>         work at all with Fedora 9
>       * It appears that you no longer need to separately supply the
>         firmware blobs - they're there in the 1.X series of
>         midisport-firmware.
>       * The rules file included in the midisport-firmware RPM for F9
>         lacks most of the devices listed in the upstream package.

Good debugging!!

Would it be possible for you to file a bug in the redhat bugzilla
system? This is definitely something that should be corrected in the
Fedora package...

> > As a short-term (maybe long-term) fix I re-installed the old
> > 2.6.22.10-1.rt9.1.fc8rt kernel (with which external MIDI triggering has
> > worked flawlessly and still works).
>
> Once it's configured properly, the keyboard is working fine with the
> 2.6.24.7-1.rt3.2.fc9.ccrmart kernel.

Good to know. I could not see how the kernel could be relevant, but you
never know...

> Don't know if it's relevant to you
> or not, but if you were failing on earlier kernels it's worth a try
> (note: as of right now all I've done is play the keyboard - no fiddling
> with wheels or controllers - just notes).
>
> Anyway thanks again for getting me thinking in the right direction.

Thanks for the report (and happy that you got it working...)!
-- Fernando


_______________________________________________
PlanetCCRMA mailing list
PlanetCCRMA@...
http://ccrma-mail.stanford.edu/mailman/listinfo/planetccrma

[Peter Howard]

On Tue, 2008-07-08 at 10:51 +0200, Fernando Lopez-Lezcano wrote:

> Good debugging!!
>
> Would it be possible for you to file a bug in the redhat bugzilla
> system? This is definitely something that should be corrected in the
> Fedora package...

Raised.  Bug #454610.

PJH

_______________________________________________
PlanetCCRMA mailing list
PlanetCCRMA@...
http://ccrma-mail.stanford.edu/mailman/listinfo/planetccrma

[Fernando Lopez-Lezcano]

On Thu, 2008-07-10 at 07:03 +1000, Peter Howard wrote:
> On Tue, 2008-07-08 at 10:51 +0200, Fernando Lopez-Lezcano wrote:
>
> > Good debugging!!
> >
> > Would it be possible for you to file a bug in the redhat bugzilla
> > system? This is definitely something that should be corrected in the
> > Fedora package...
>
> Raised.  Bug #454610.

Thanks++!
That really helps...
-- Fernando


_______________________________________________
PlanetCCRMA mailing list
PlanetCCRMA@...
http://ccrma-mail.stanford.edu/mailman/listinfo/planetccrma