Skip to contentSkip to conversation

problems with external MIDI port

DISCUSSION Ivan Middleton 2 participants 4 messages Sep 18, 2002 – Sep 20, 2002

lore
Share & tools
lore
#0 lore raw

Hi all,

I'm having a problem with the ALSA sequencer. First I was using 0.9.0rc2, and then I switched to cvs, hoping that things would be fixed there. But alas, the problems occur in both. I was going to try the 0.5 release, but it didn't want to compile for me.

The problem is as follows: I am running a program (called Linux Pipe Organ) which reads from the external MIDI port (64:0) via an ALSA sequencer connection, and which also starts a TiMidity process and sends events to it via ALSA sequencer connections. (Basically, the program is just relaying MIDI events, after some modifications.) This works fine for a few minutes, and then suddenly... nothing. The ALSA sequencer stops relaying events from the external MIDI port. Now, my digital piano is connected to the MIDI port, and it sends something like 30 MIDI events per second (I don't know exactly what these are) even when no keys are being pressed. The problem I just described occurs even if I never press any keys on the piano; I can tell by watching /proc/asound/seq/clients, looking at the following lines

Client 132 : "Linux Organ" [User]
Port 0 : "Linux Organ" (rWe-)
Connected From: 64:0
Input pool :

    Pool size          : 200
    Cells in use       : 0
    Peak cells in use  : 30
    Alloc success      : 27513
    Alloc failures     : 0

and noticing that after a while the "Alloc success" number stops changing (when everything is working it goes up by about 30 every second).

Now, I wanted to test whether this problem is the fault of Linux Pipe Organ (hereafter LPO) or the ALSA sequencer. So I recompiled LPO to have it listen to T. Iwai's Virtual Keyboard instead of the external MIDI port. I then ran those programs, while simultaneously having 'aseqview' listen to 64:0 (the external MIDI port). The same thing happened again--after a little while, 'aseqview' stopped registering anything from the MIDI port, even though LPO, Virtual Keyboard, and TiMidity hadn't been communicating at all, only sitting there with established connections. After this, LPO was still listening to the Virtual Keyboard just fine, meaning that the ALSA sequencer wasn't completely trashed or anything. When I had 'aseqview' listen to 64:0 in the absence of LPO, it seemed to work fine indefinitely.

So it seems that the ALSA sequencer has problems with the external MIDI port in the presence of too many sequencer clients (8-10 in my situation). I really have no idea what's going on here. I should mention that my kernel is 2.5.33 and my sound card is sbawe. Another funny thing I found was that in order to get the external MIDI port working again, I had to unload and reload the ALSA modules *twice* (I think this was the case every time, anyway).

I'm hoping that the developers can reproduce this or at least have ideas for things I should try in order to probe this problem some more.

Thanks for your time,
Ivan Middleton

__________________________________________________
Do you Yahoo!?
Yahoo! News - Today's headlines
http://news.yahoo.com

-------------------------------------------------------
This SF.NET email is sponsored by: AMD - Your access to the experts on Hammer Technology! Open Source & Linux Developers, register now for the AMD Developer Symposium. Code: EX8664
http://www.developwithamd.com/developerlab

#1 lore raw

Hi,

At Wed, 18 Sep 2002 14:10:19 -0700 (PDT),

Quoted 31 lines · Ivan Middleton

Hi all,

I'm having a problem with the ALSA sequencer. First I was using 0.9.0rc2, and then I switched to cvs, hoping that things would be fixed there. But alas, the problems occur in both. I was going to try the 0.5 release, but it didn't want to compile for me.

The problem is as follows: I am running a program (called Linux Pipe Organ) which reads from the external MIDI port (64:0) via an ALSA sequencer connection, and which also starts a TiMidity process and sends events to it via ALSA sequencer connections. (Basically, the program is just relaying MIDI events, after some modifications.) This works fine for a few minutes, and then suddenly... nothing. The ALSA sequencer stops relaying events from the external MIDI port. Now, my digital piano is connected to the MIDI port, and it sends something like 30 MIDI events per second (I don't know exactly what these are) even when no keys are being pressed. The problem I just described occurs even if I never press any keys on the piano; I can tell by watching /proc/asound/seq/clients, looking at the following lines

Client 132 : "Linux Organ" [User]
Port 0 : "Linux Organ" (rWe-)
Connected From: 64:0
Input pool :

    Pool size          : 200
    Cells in use       : 0
    Peak cells in use  : 30
    Alloc success      : 27513
    Alloc failures     : 0
Quoted 3 lines

and noticing that after a while the "Alloc success" number stops changing (when everything is working it goes up by about 30 every second).

this could be due to the timer routine.
could you check /proc/asound/seq/queue and /proc/asound/seq/timer whether the timer and queue are still running correctly in such a situation?

thanks,

Takashi

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

#2 lore raw
Quoted 42 lines · --- Takashi Iwai

Hi,

At Wed, 18 Sep 2002 14:10:19 -0700 (PDT),

Quoted 5 lines · Ivan Middleton

Hi all,

I'm having a problem with the ALSA sequencer. First I was using 0.9.0rc2, and then I switched to cvs, hoping that things would be

fixed

Quoted 1 line

there. But alas, the problems occur in both. I was going to try

the

Quoted 3 lines

0.5 release, but it didn't want to compile for me.

The problem is as follows: I am running a program (called Linux

Pipe

Quoted 4 lines

Organ) which reads from the external MIDI port (64:0) via an ALSA sequencer connection, and which also starts a TiMidity process and sends events to it via ALSA sequencer connections. (Basically, the program is just relaying MIDI events, after some modifications.)

This

Quoted 1 line

works fine for a few minutes, and then suddenly... nothing. The

ALSA

Quoted 1 line

sequencer stops relaying events from the external MIDI port. Now,

my

Quoted 2 lines

digital piano is connected to the MIDI port, and it sends something like 30 MIDI events per second (I don't know exactly what these

are)

Quoted 14 lines

even when no keys are being pressed. The problem I just described occurs even if I never press any keys on the piano; I can tell by watching /proc/asound/seq/clients, looking at the following lines

Client 132 : "Linux Organ" [User]
Port 0 : "Linux Organ" (rWe-)
Connected From: 64:0
Input pool :

    Pool size          : 200
    Cells in use       : 0
    Peak cells in use  : 30
    Alloc success      : 27513
    Alloc failures     : 0
Quoted 8 lines
Quoted 3 lines

and noticing that after a while the "Alloc success" number stops changing (when everything is working it goes up by about 30 every second).

this could be due to the timer routine.
could you check /proc/asound/seq/queue and /proc/asound/seq/timer whether the timer and queue are still running correctly in such a situation?

am I doing something wrong? I don't see anything in either /proc/asound/seq/queue or /proc/asound/seq/timer, no matter what.

Ivan

__________________________________________________
Do you Yahoo!?
New DSL Internet Access from SBC & Yahoo!
http://sbc.yahoo.com

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

#3 lore raw

At Thu, 19 Sep 2002 20:14:00 -0700 (PDT),

Quoted 7 lines · Ivan Middleton
Quoted 4 lines

could you check /proc/asound/seq/queue and /proc/asound/seq/timer whether the timer and queue are still running correctly in such a situation?

am I doing something wrong? I don't see anything in either /proc/asound/seq/queue or /proc/asound/seq/timer, no matter what.

sorry, /proc/asound/seq/queues

                             ^

both files are empty when no timer is running, that is, no sequencer client is playing.

ciao,

Takashi

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf