[linux-audio-user] Midi Woes

Clemens Ladisch cladisch at fastmail.net
Mon Nov 13 12:09:07 EST 2006


Giuseppe Zompatori wrote:
> 2006/11/13, Clemens Ladisch <cladisch at fastmail.net>:
> > The intel8x0 modules has nothing to do with MIDI.
> 
> It had to do with MIDI until a few kernel revisions ago.
> The intel8x0 had an "mpu_port" parameter which would activate the midi
> port (yes from within the codec module). In fact the mpu401 module
> wasn't even being loaded when the module was working this way, if I
> remember correctly.

It loaded the snd-mpu401-lib module, just as the snd-mpu401 module does
the PnP stuff and then loads the -lib module to do the actual work.

> That has been changed recently, mpu401 management has been moved from
> the intel8x0 module into the mpu401* ones and the mpu_port has been deprecated.
> You can hunt ALSA changelogs for the intel8x0 module if you don't believe me ;)

I do believe you - it was I who made the change.  :-)

The reason this was removed from intel8x0 was that it didn't actually
work, i.e., it would work only if the port address had the same value
that was already set by the BIOS.

> cmd: 0xff failed at 0x300 (status = 0xff, data = 0xff)

········································································
This means there is no MPU-401 device at this address (either the port
address is wrong, or the device is disabled).

> My MPU port is set at IRQ 0x300 DMA 5,

Is that how your BIOS labels these numbers?  :)

> modprobe -v snd_mpu401 pnp=0 port=0x300 irq=5

Oops, the module actually supports eight devices, and PnP is enabled
by default for all of them.  Use:

  modprobe snd-mpu401 pnp=0,0,0,0,0,0,0,0 port=0x300 irq=5

(Unless you have configured the kernel to use dynamic device numbers for
ALSA devices, then it supports 32 devices ...)

> [17179872.004000] pnp: Device 00:09 activated.
> [17179872.004000] pnp: Device 00:09 disabled.
> 
> I just noticed this now, it seems like it's trying to initialize the
> MPU port but silently fails?

There should be some message in case of an error.

> I am starting to think that this particular motherboard audio codec is
> either buggy, or not well supported by ALSA

The MIDI port is implemented in a different chip.

> as I also have issues with the HW mixer... things are swapped around,
> and there is no "ac97_quirk" module parameter for remapping things
> that works.

Which mixer controls are routed to which hardware outputs?
What mainboard model is this?

> > This were problems with the RTC used for MIDI playback timing
> > (not related to any particular driver).
> > Please try the latest ALSA drivers.
> 
> Latest as in latest released or CVS-checkout?

The latest release should work.  The latest Hg checkout has a fix not in
the release, but which shouldn't occur often.


Regards,
Clemens



More information about the Linux-audio-user mailing list