[PlanetCCRMA]Re: /dev/midi and pd

[Kevin Ernste]

Hi -

Relating to the last part of Jon's initial post on this thread (the
posibility of alsa-seq support in PD), it seems that the alsa-seq
patch he mentioned sems to have made it upstream to 0.38-2.

I downloaded the 0.38-2 tarball from Miller's page the other day,
built it ("./configure --enable-alsa --enable-jack --enable-setuid",
make) and ran it without installing and PD shows up as an ALSA-aware
midi device (Qjackctl sees it and allows routing to/from other MIDI
apps).  Needless to say, this is very promising.

Are there any plans for 0.38 to get into Planet or is it still
considered unstable?

Kevin


On Thu, 10 Feb 2005 19:20:13 -0800 (PST), Fernando Lopez-Lezcano
<nando@ccrma.stanford.edu> wrote:
> On Thu, 2005-02-10 at 18:14, Jon B wrote:
> > > > Yes, I just found that.  So I tried it, but it doesn't work, because
> > > > my midiC_D_ files have changed names.
> > > >
> > > > > There is a way to tell udev to recreate devices automatically on boot
> > > > > (it's on the fc3 release notes). Copy the devices from /dev/ to
> > > > > /etc/udev/devices/. Again, would be better to alter the creation rules
> > > > > or really find the bug :-)
> >
> > It seems like I should be figuring out how to change the creation rules...
> >
> > > In reality those files should not be links, they should be normal device
> > > files (made with mknod). This is what I get on an old rh9 machine:
> > >
> > > > ls -l /dev/midi*
> > > crw-------    1 nando    root      35,   0 Jan 30  2003 /dev/midi0
> > > crw-------    1 nando    root      14,   2 Jan 30  2003 /dev/midi00
> > > crw-------    1 nando    root      14,  18 Jan 30  2003 /dev/midi01
> > > crw-------    1 nando    root      14,  34 Jan 30  2003 /dev/midi02
> > > crw-------    1 nando    root      14,  50 Jan 30  2003 /dev/midi03
> > > crw-------    1 nando    root      35,   1 Jan 30  2003 /dev/midi1
> > > crw-------    1 nando    root      35,   2 Jan 30  2003 /dev/midi2
> > > crw-------    1 nando    root      35,   3 Jan 30  2003 /dev/midi3
> >
> > I see this:
> >
> > 1. Find the major device number by typing 'cat /proc/devices'. Look
> > for the device called "ir", and note its number (it will probably be
> > 254 or 253).
> > 2. As root, create a character device called '/dev/ir' with that major
> > number, using the 'mknod' program.
> 
> The "major" number is the first number in the entries above, that is,
> for /dev/midi0 the major of the device is 35, the minor number is 0.
> 
> > But are these placed in /dev/ or /etc/udev/devices/ ?  And will they
> > always be the same numbers?
> 
> Yes, these should always be the same numbers.
> >From what I've read you can create them in /dev to use them, if you want
> them recreated on boot put them in /etc/udev/devices/
> 
> >  Can't this be done automatically somehow?
> 
> Yes, I still have to read up, that would involve adding rules so that
> udev automatically creates the devices on demand. Most probably I could
> add this to the alsa-driver package in a future release (until this is
> fixed in fcx).
> 
> > These are mine:
> >
> > crw-------  1 omegatron root  14, 34 Feb  9 23:50 /dev/snd/midi2
> > crw-------  1 omegatron root 116, 72 Feb  9 23:50 /dev/snd/midiC2D0
> > crw-------  1 omegatron root 116, 73 Feb  9 23:50 /dev/snd/midiC2D1
> > crw-------  1 omegatron root 116, 74 Feb  9 23:50 /dev/snd/midiC2D2
> > crw-------  1 omegatron root 116, 75 Feb  9 23:50 /dev/snd/midiC2D3
> >
> > but pd is looking for /dev/midi*.  Is there any way to make pd look in
> > the other directory?  That might be easier...
> 
> Well, they are different things. "/dev/midi*" are the OSS midi devices,
> "/dev/snd/midi*" are the ALSA midi devices. When running ALSA (as we
> are) the OSS part is handled through emulation modules. Anyway, both
> devices are not the same thing, although maybe things will work if you
> point one to the other.
> 
> Besides this bug (missing devices in the standard udev configuration in
> fc3 - I have to check in redhat's bugzilla to see if this has been
> reported before), the _real_ solution would be for pd to use the alsa
> sequencer interface to midi devices, instead of the oss raw interface.
> 
> > But then I still have to worry about plugging in a USB device after
> > boot or before boot and having the order different...  Argh.
> 
> -- Fernando
> 
> 
> _______________________________________________
> PlanetCCRMA mailing list
> PlanetCCRMA@ccrma.stanford.edu
> http://ccrma-mail.stanford.edu/mailman/listinfo/planetccrma
>