Skip to contentSkip to conversation

USB MIDI driver

PATCH In review Clemens Ladisch 7 participants 41 messages Jul 1, 2002 – Jul 5, 2002

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 USB MIDI driver

Attached is a patch for a USB MIDI driver. Essentially, it does the same as my USB MIDI daemon.

The source is put into the alsa-kernel tree, not into alsa-driver like the usb-audio driver.

I don't know whether my changes to scripts/Modules.dep make any sense, please review them before applying.

Clemens

#1 lore raw

Hi,

At Mon, 01 Jul 2002 19:07:40 +0200,

Quoted 6 lines · Clemens Ladisch

Attached is a patch for a USB MIDI driver. Essentially, it does the same as my USB MIDI daemon.

The source is put into the alsa-kernel tree, not into alsa-driver like the usb-audio driver.

i prefer putting it onto alsa-driver at first unless it's confirmed as enough stable after tarball release. alsa-kernel is the tree to be merged into 2.5 kernel. also, the location of usb drivers is still a question, whether on sound tree or on drivers/usb tree.

anyway, i'll check the code into cvs after review (if there is no objection).

thanks,

Takashi

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

#2 lore raw

USB recording - repetitive peaks

Thanks to your patch Takashi we have two people who have been able to record through the quattro.

We have both found that there is a constant repetitive peak in the signal which reminds me of the problems I had with the cmipci for a brief period.

Thorsten Hass has recorded a perfect example for you which he did using aplay and arecord. It is a sine wave so is easy on the ears.

http://www.thorstenhaas.de/alsa/audio.file

He has also generated a jpeg plotted by an oscilloscope while recording which I can send you if it will help.

He has also asked if there is any chance you will have time to look at this in the two weeks. I believe he has to do some work with his device so needs to make a decision about which OS to use for the job.

Signature · 12 lines
Patrick Shirkey - Boost Hardware Ltd.
For the discerning hardware connoisseur
Http://www.boosthardware.com
Http://www.boosthardware.com/LAU/guide/
========================================



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
#31 lore raw
In reply to Takashi Iwai

At Tue, 02 Jul 2002 12:36:55 +0200,

Quoted 10 lines · I wrote:

At Mon, 01 Jul 2002 19:07:40 +0200,

Quoted 6 lines · Clemens Ladisch

Attached is a patch for a USB MIDI driver. Essentially, it does the same as my USB MIDI daemon.

The source is put into the alsa-kernel tree, not into alsa-driver like the usb-audio driver.

Quoted 7 lines

i prefer putting it onto alsa-driver at first unless it's confirmed as enough stable after tarball release. alsa-kernel is the tree to be merged into 2.5 kernel. also, the location of usb drivers is still a question, whether on sound tree or on drivers/usb tree.

anyway, i'll check the code into cvs after review (if there is no objection).

ok, now it's there. as written above, i put it to alsa-driver tree. let us move the usb stuff later to the official kernel place.

your codes have been slightly modified:

- usb compatibility hack was moved to wrapper.h, so that it can be shared with usbuadio.c

- newly added virmidi registrations.
that means, snd-usb-midi will have its own rawmidi devices automatically as /dev/snd/midiCxDx.
this will be good for applications which speak only with rawmidi.

the latter one was not tested.
can anyone update the cvs if it works?

thanks.

Takashi

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

#3 lore raw

Re: USB recording - repetitive peaks

In reply to Patrick Shirkey

Hi Patrick,

At Tue, 02 Jul 2002 23:28:08 +0900,

Quoted 4 lines · Patrick Shirkey

Thanks to your patch Takashi we have two people who have been able to record through the quattro.

Quoted 8 lines

We have both found that there is a constant repetitive peak in the signal which reminds me of the problems I had with the cmipci for a brief period.

Thorsten Hass has recorded a perfect example for you which he did using aplay and arecord. It is a sine wave so is easy on the ears.

http://www.thorstenhaas.de/alsa/audio.file

hmm, this sounds like the mismatching of packet size and copy destination. i heard your ogg file, and this was similar, too.

under which condition were they recorded?
with 24bit/96kHz/2ch?
or does it happen generally on every condition?

you can see the running status and condition on streamX proc file during playback/capture.

Quoted 2 lines

He has also generated a jpeg plotted by an oscilloscope while recording which I can send you if it will help.

can you figure out how long the period of puls is?
i guess this has relation between the status shown in the proc file above.

Takashi

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

#4 lore raw

Re: USB recording - repetitive peaks

Quoted 14 lines · Takashi Iwai

Hi Patrick,

At Tue, 02 Jul 2002 23:28:08 +0900,

Quoted 7 lines · Patrick Shirkey

Thanks to your patch Takashi we have two people who have been able to record through the quattro.

We have both found that there is a constant repetitive peak in the signal which reminds me of the problems I had with the cmipci for a brief period.

Thorsten Hass has recorded a perfect example for you which he did using aplay and arecord. It is a sine wave so is easy on the ears.

http://www.thorstenhaas.de/alsa/audio.file

Quoted 7 lines

hmm, this sounds like the mismatching of packet size and copy destination. i heard your ogg file, and this was similar, too.

under which condition were they recorded?
with 24bit/96kHz/2ch?
or does it happen generally on every condition?

They were arecord -f cd

Quoted 3 lines

you can see the running status and condition on streamX proc file during playback/capture.

----

# cat /proc/asound/card2/stream1
M Audio USB AudioSport Quattro (tm) : USB Audio #1

Playback:
Status: Stop
Altset 3

    Format: S16_LE
    Channels: 2
    Endpoint: 3 OUT (ADAPTIVE)
    Rates: 11025, 22050, 44100, 48000
  Altset 2
    Format: S24_3LE
    Channels: 2
    Endpoint: 3 OUT (ADAPTIVE)
    Rates: 11025, 22050, 44100, 48000
  Altset 1
    Format: S24_3LE
    Channels: 2
    Endpoint: 3 OUT (ADAPTIVE)
    Rates: 88200, 96000

Capture:
Status: Running
Packets = 20
URBs = 5
Packet Size = 192
Nominal freq = 44. 1682 Altset 3
Format: S16_LE
Channels: 2
Endpoint: 5 IN (SYNC)
Rates: 11025, 22050, 44100, 48000
Altset 2

    Format: S24_3LE
    Channels: 2
    Endpoint: 5 IN (SYNC)
    Rates: 11025, 22050, 44100, 48000
  Altset 1
    Format: S24_3LE
    Channels: 2
    Endpoint: 5 IN (SYNC)
    Rates: 88200, 96000

----

Quoted 6 lines
Quoted 3 lines

He has also generated a jpeg plotted by an oscilloscope while recording which I can send you if it will help.

can you figure out how long the period of puls is?

They are so regular you can set a use them as a metronome:) Would you like me to have a look at the timing in an editor or is there an more precise way to find out?

Quoted 7 lines

i guess this has relation between the status shown in the proc file above.

Takashi

Signature · 12 lines
Patrick Shirkey - Boost Hardware Ltd.
For the discerning hardware connoisseur
Http://www.boosthardware.com
Http://www.boosthardware.com/LAU/guide/
========================================



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

Re: USB recording - repetitive peaks

Hi Patrick,

i found out an obvious bug there. i forgot to correct the offset value of capture destination after a minor modification. it's fixed on cvs.
could you update the cvs and test whether it works?

Takashi

At Wed, 03 Jul 2002 00:35:03 +0900,

Quoted 92 lines · Patrick Shirkey
Quoted 18 lines · Takashi Iwai

Hi Patrick,

At Tue, 02 Jul 2002 23:28:08 +0900,

Quoted 11 lines · Patrick Shirkey

Thanks to your patch Takashi we have two people who have been able to record through the quattro.

We have both found that there is a constant repetitive peak in the signal which reminds me of the problems I had with the cmipci for a brief period.

Thorsten Hass has recorded a perfect example for you which he did using aplay and arecord. It is a sine wave so is easy on the ears.

http://www.thorstenhaas.de/alsa/audio.file

Quoted 7 lines

hmm, this sounds like the mismatching of packet size and copy destination. i heard your ogg file, and this was similar, too.

under which condition were they recorded?
with 24bit/96kHz/2ch?
or does it happen generally on every condition?

They were arecord -f cd

Quoted 3 lines

you can see the running status and condition on streamX proc file during playback/capture.

----

# cat /proc/asound/card2/stream1
M Audio USB AudioSport Quattro (tm) : USB Audio #1

Playback:

   Status: Stop
   Altset 3
     Format: S16_LE
     Channels: 2
     Endpoint: 3 OUT (ADAPTIVE)
     Rates: 11025, 22050, 44100, 48000
   Altset 2
     Format: S24_3LE
     Channels: 2
     Endpoint: 3 OUT (ADAPTIVE)
     Rates: 11025, 22050, 44100, 48000
   Altset 1
     Format: S24_3LE
     Channels: 2
     Endpoint: 3 OUT (ADAPTIVE)
     Rates: 88200, 96000

Capture:
Status: Running
Packets = 20
URBs = 5
Packet Size = 192
Nominal freq = 44. 1682 Altset 3
Format: S16_LE
Channels: 2
Endpoint: 5 IN (SYNC)
Rates: 11025, 22050, 44100, 48000
Altset 2

     Format: S24_3LE
     Channels: 2
     Endpoint: 5 IN (SYNC)
     Rates: 11025, 22050, 44100, 48000
   Altset 1
     Format: S24_3LE
     Channels: 2
     Endpoint: 5 IN (SYNC)
     Rates: 88200, 96000

----

Quoted 7 lines
Quoted 4 lines

He has also generated a jpeg plotted by an oscilloscope while recording which I can send you if it will help.

can you figure out how long the period of puls is?

They are so regular you can set a use them as a metronome:) Would you like me to have a look at the timing in an editor or is there an more precise way to find out?

Quoted 17 lines
Quoted 7 lines

i guess this has relation between the status shown in the proc file above.

Takashi

Signature · 5 lines
Patrick Shirkey - Boost Hardware Ltd.
For the discerning hardware connoisseur
Http://www.boosthardware.com
Http://www.boosthardware.com/LAU/guide/
========================================

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

#30 lore raw

Re: Re: USB recording - repetitive peaks

In reply to Patrick Shirkey
Quoted 3 lines · Patrick Shirkey

They are so regular you can set a use them as a metronome:) Would you like me to have a look at the timing in an editor or is there an more precise way to find out?

36.5720 -> 36.5820

0.3710 -> 0.7430 -> 1.1140

Looking at the wave in an editor at almost 1:1 the distance between peaks is the above. There are mini peaks every 10ms and there are large peaks every 37.10 ms.

Signature · 12 lines
Patrick Shirkey - Boost Hardware Ltd.
For the discerning hardware connoisseur
Http://www.boosthardware.com
Http://www.boosthardware.com/LAU/guide/
========================================



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

Re: Re: USB recording - repetitive peaks

In reply to Takashi Iwai
Quoted 7 lines · Takashi Iwai

Hi Patrick,

i found out an obvious bug there. i forgot to correct the offset value of capture destination after a minor modification. it's fixed on cvs.
could you update the cvs and test whether it works?

I get this in alsa-utils

----
/usr/local/src/music/alsa/alsa-utils# ./cvscompile --with-cards=intel8x0,cmipci,usb-audio,usb-midi --with-sequencer=yes;make;make install
aclocal: configure.in: 11: macro `AM_PATH_ALSA' not found in library
/usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL
/usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL
/usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL /usr//share/automake/am/lang-compile.am: AMDEP does not appear in AM_CONDITIONAL /usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL
/usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL
/usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL /usr//share/automake/am/lang-compile.am: AMDEP does not appear in AM_CONDITIONAL /usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL
/usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL
/usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL /usr//share/automake/am/lang-compile.am: AMDEP does not appear in AM_CONDITIONAL /usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL
/usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL
/usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL /usr//share/automake/am/lang-compile.am: AMDEP does not appear in AM_CONDITIONAL /usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL
/usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL
/usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL /usr//share/automake/am/lang-compile.am: AMDEP does not appear in AM_CONDITIONAL /usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL
/usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL
/usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL /usr//share/automake/am/lang-compile.am: AMDEP does not appear in AM_CONDITIONAL
configure.in:4: error: possibly undefined macro: AM_INIT_AUTOMAKE
configure.in:11: error: possibly undefined macro: AM_PATH_ALSA
configure.in:28: error: possibly undefined macro: AM_CONFIG_HEADER
CFLAGS=-O2 -Wall -pipe -g
./configure --with-cards=intel8x0,cmipci,usb-audio,usb-midi --with-sequencer=yes
./configure: line 922: syntax error near unexpected token `AM_INIT_AUTOMAKE(alsa-utils,'
./configure: line 922: `AM_INIT_AUTOMAKE(alsa-utils, 0.9.0rc2)'
make: *** No targets specified and no makefile found. Stop.
make: *** No rule to make target `install'. Stop.

----

Signature · 12 lines
Patrick Shirkey - Boost Hardware Ltd.
For the discerning hardware connoisseur
Http://www.boosthardware.com
Http://www.boosthardware.com/LAU/guide/
========================================



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

Re: Re: USB recording - repetitive peaks

At Wed, 03 Jul 2002 01:37:30 +0900,

Quoted 10 lines · Patrick Shirkey
Quoted 7 lines · Takashi Iwai

Hi Patrick,

i found out an obvious bug there. i forgot to correct the offset value of capture destination after a minor modification. it's fixed on cvs.
could you update the cvs and test whether it works?

Quoted 4 lines

I get this in alsa-utils

----
/usr/local/src/music/alsa/alsa-utils# ./cvscompile

                            ^^^^^^^^^^

shouldn't be alsa-driver?

Takashi

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

#8 lore raw

Re: Re: USB recording - repetitive peaks

Quoted 13 lines · Takashi Iwai

At Wed, 03 Jul 2002 01:37:30 +0900,

Quoted 10 lines · Patrick Shirkey
Quoted 7 lines · Takashi Iwai

Hi Patrick,

i found out an obvious bug there. i forgot to correct the offset value of capture destination after a minor modification. it's fixed on cvs.
could you update the cvs and test whether it works?

Quoted 8 lines
Quoted 4 lines

I get this in alsa-utils

----
/usr/local/src/music/alsa/alsa-utils# ./cvscompile

                            ^^^^^^^^^^

shouldn't be alsa-driver?

Yeah I had to zap the cvs because of the patch confusing things so I'm trying to install alsa-utils but I get that error. alsa-driver and alsa-lib installed correctly.

Another thing is that even though I can rmmod all the other sound modules for other cards I cannot rmmod the following. I have checked with ps -aux that there are no stray processes. Currently I have to reboot. I think this only started with the new patches for recording.

snd-usb-audio          23136   1  (autoclean)
snd-pcm                48320   1  [snd-usb-audio]
snd-timer               9280   0  [snd-pcm]
snd                    23400   3  [snd-usb-audio snd-pcm snd-timer]
soundcore               3556   0  [snd]
Signature · 12 lines
Patrick Shirkey - Boost Hardware Ltd.
For the discerning hardware connoisseur
Http://www.boosthardware.com
Http://www.boosthardware.com/LAU/guide/
========================================



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

Re: Re: USB recording - repetitive peaks

At Wed, 03 Jul 2002 02:08:13 +0900,

Quoted 20 lines · Patrick Shirkey
Quoted 18 lines · Takashi Iwai

At Wed, 03 Jul 2002 01:37:30 +0900,

Quoted 14 lines · Patrick Shirkey
Quoted 7 lines · Takashi Iwai

Hi Patrick,

i found out an obvious bug there. i forgot to correct the offset value of capture destination after a minor modification. it's fixed on cvs.
could you update the cvs and test whether it works?

I get this in alsa-utils

----
/usr/local/src/music/alsa/alsa-utils# ./cvscompile

Quoted 7 lines
Quoted 3 lines
                            ^^^^^^^^^^

shouldn't be alsa-driver?

Yeah I had to zap the cvs because of the patch confusing things so I'm trying to install alsa-utils but I get that error. alsa-driver and alsa-lib installed correctly.

for alsa-utils you don't need any configure arguments.

the error message suggests that you didn't make install on alsa-lib. please check whether /usr/share/aclocal/alsa.m4 exists.

Quoted 10 lines

Another thing is that even though I can rmmod all the other sound modules for other cards I cannot rmmod the following. I have checked with ps -aux that there are no stray processes. Currently I have to reboot. I think this only started with the new patches for recording.

snd-usb-audio          23136   1  (autoclean)
snd-pcm                48320   1  [snd-usb-audio]
snd-timer               9280   0  [snd-pcm]
snd                    23400   3  [snd-usb-audio snd-pcm snd-timer]
soundcore               3556   0  [snd]

this is weird.
there has been no fundamental change about this.
please try the following

	# fuser -k /dev/snd/* /dev/dsp* /dev/mixer*

Takashi

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

#10 lore raw

Re: Re: USB recording - repetitive peaks

In reply to Patrick Shirkey
Quoted 2 lines

the error message suggests that you didn't make install on alsa-lib. please check whether /usr/share/aclocal/alsa.m4 exists.

I did run make install and have just done so again to make sure but it's not in /usr/local/share/aclocal either. I will copy it over and try recording again (see below).

Also I get this for snd-usb-midi
----

# modprobe snd-usb-midi
/lib/modules/2.4.19-rc1/kernel/sound/usb/snd-usb-midi.o: unresolved symbol snd_virmidi_new_Rc7dd286c
/lib/modules/2.4.19-rc1/kernel/sound/usb/snd-usb-midi.o: insmod /lib/modules/2.4.19-rc1/kernel/sound/usb/snd-usb-midi.o failed
/lib/modules/2.4.19-rc1/kernel/sound/usb/snd-usb-midi.o: insmod snd-usb-midi failed
----

I just tried recording using the new code and completely hung the computer where only alt+sysreq+b worked. That was during recording.

The other person who is debugging with me has experienced simlilar as a kernel oops. He is currently using the patch to cvs but he will update soon I think.

This is what he said:
----
2) Kernel Oops:
The compilation of the source was rather a nice thing. I build some RPMs from yesterdays snapshot to be able to clean up things after I installed them. After installing, configuring modules.conf and removing my old asound.conf the module probed finely.
I attached a copy of the aoutput of "arecord -l"

Recording was a bit messy, due to killing the arecord process freezes my box.
Deterministically: whenever the process is interuppted (Ctrl-C) or killed (sig 9).
I attached a copy of the kernel-oops.

Oops: 0002

CPU: 0

EIP: 0010:[<d40c824a>]

EFLAGS: 00010002

eax: 000000b0 ebx: d16bfc00 ecx: 0000002c edx: d45f4950

esi: d16bfc00 edi: 00002cc8 ebp: 0000002c esp: c02e5e78

ds: 0018 es: 0018 ss: 0018

Process swapper (pid: 0, stackpage=c02e5000)

Stack: d45f4950 d45f49c0 d4cc19a0 d3efe5e0 00004000 00000000 00000004 d40c8607
       d45f4950 d54a7c00 d3efe5e0 d3efe5e0 00000000 00000000 d3131e60 

d51b1f58

       d3efe5e0 00000246 d3efe5e0 d51b2660 d3efe5e0 d50f6800 04000001 

d89d8000

Call Trace: [<d40c8607>] [<d51b1f58>] [<d51b2660>] [<d51b3260>] [<c010814e>]

[<c01082ae>] [<c010a118>] [<c01ad465>] [<c01ad2d4>] [<c0105407>]
[<c0105000>]

[<c0105027>]

Code: f3 a5 a8 02 74 02 66 a5 a8 01 74 01 a4 9c 59 fa 8b 5c 24 20

<0>Kernel panic: Aiee, killing interrupt handler!

In interrupt handler - not syncing

----

Signature · 12 lines
Patrick Shirkey - Boost Hardware Ltd.
For the discerning hardware connoisseur
Http://www.boosthardware.com
Http://www.boosthardware.com/LAU/guide/
========================================



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

Re: Re: USB recording - repetitive peaks

Quoted 6 lines · Patrick Shirkey

I did run make install and have just done so again to make sure but it's not in /usr/local/share/aclocal either. I will copy it over and try recording again (see below).

Quoted 4 lines

Recording was a bit messy, due to killing the arecord process freezes my >> box.
Deterministically: whenever the process is interuppted (Ctrl-C) or killed (sig 9).

This seems to be the same here now. The file was recorded and with no distortions to the sound but arecord hangs my rig when it stops. I'll check with JACK+ardour.

Signature · 12 lines
Patrick Shirkey - Boost Hardware Ltd.
For the discerning hardware connoisseur
Http://www.boosthardware.com
Http://www.boosthardware.com/LAU/guide/
========================================



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

Re: Re: USB recording - repetitive peaks

Quoted 6 lines · Patrick Shirkey
Quoted 3 lines · Patrick Shirkey

I did run make install and have just done so again to make sure but it's not in /usr/local/share/aclocal either. I will copy it over and try recording again (see below).

Quoted 1 line
Quoted 10 lines
Quoted 1 line
Quoted 1 line

Recording was a bit messy, due to killing the arecord process

freezes my >> box.

Quoted 1 line
Quoted 1 line

Deterministically: whenever the process is interuppted (Ctrl-C) or

killed

Quoted 1 line
Quoted 1 line

(sig 9).

This seems to be the same here now. The file was recorded and with no distortions to the sound but arecord hangs my rig when it stops. I'll check with JACK+ardour.

Ouch. JACK+ardour causes an instant hang. Using alt+sysreq I can sync unmount and reboot but nothing else. Also the sound quality is significantly degraded while running through JACK. At least the native drivers are sounding correct now though :)

Signature · 12 lines
Patrick Shirkey - Boost Hardware Ltd.
For the discerning hardware connoisseur
Http://www.boosthardware.com
Http://www.boosthardware.com/LAU/guide/
========================================



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
#32 lore raw
In reply to Takashi Iwai
Quoted 12 lines

- usb compatibility hack was moved to wrapper.h, so that it can be shared with usbuadio.c

- newly added virmidi registrations.
that means, snd-usb-midi will have its own rawmidi devices automatically as /dev/snd/midiCxDx.
this will be good for applications which speak only with rawmidi.

the latter one was not tested.
can anyone update the cvs if it works?

I can't compile current CVS usb drivers:

gcc -E -M -DALSA_BUILD -D__KERNEL__ -DMODULE=1 -I/home/plc/kernelsrc/ALSA/cvstree/alsa-driver/include -I/lib/modules/2.4.18/build/include -O2 -mpreferred-stack-boundary=2 -march=k6 -DLINUX -Wall -Wstrict-prototypes -fomit-frame-pointer -pipe usbaudio.c usbaudio.h usbmidi.c usbmixer.c wrapper.h > .depend wrapper.h:1:40: missing binary operator before '<' wrapper.h:15:40: missing binary operator before '<' make[1]: *** [fastdep] Error 1

Pedro.

-------------------------------------------------------
Sponsored by:
ThinkGeek at http://www.ThinkGeek.com/

#33 lore raw
Quoted 3 lines · Me

I can't compile current CVS usb drivers:

Sorry, It had been fixed 4 hours ago.
I have compiled it succesfully now. And also works! (only preliminary tests).

$ cat /proc/asound/seq/clients
Client info
  cur  clients : 5
  peak clients : 6
  max  clients : 192
Client   0 : "System" [Kernel]
  Port   0 : "Timer" (Rwe-)
  Port   1 : "Announce" (R-e-)
    Connecting To: 63:0

Client 63 : "OSS sequencer" [Kernel]
Port 0 : "Receiver" (-we-)
Connected From: 0:1
Client 64 : "External MIDI 0" [Kernel]
Port 0 : "MIDI 0-0" (RWeX)
Client 72 : "EDIROL UM-2" [Kernel]
Port 0 : "UM-2 Port 0" (RWeX)

    Connecting To: 80:0

Port 1 : "UM-2 Port 1" (RWeX)
Client 80 : "Midiman Midisport 2x2" [Kernel]
Port 0 : "Midisport 2x2 Port 0" (RWeX)
Connected From: 72:0
Port 1 : "Midisport 2x2 Port 1" (RWeX)

Thanks,
Pedro

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
No, I will not fix your computer.
http://thinkgeek.com/sf

#28 lore raw

Re: Re: USB recording - repetitive peaks

In reply to Patrick Shirkey
Quoted 10 lines

I get this in alsa-utils

----
/usr/local/src/music/alsa/alsa-utils# ./cvscompile --with-cards=intel8x0,cmipci,usb-audio,usb-midi --with-sequencer=yes;make;make install
aclocal: configure.in: 11: macro `AM_PATH_ALSA' not found in library /usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL /usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL /usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL

[MUNCH]

Same here from just downloaded CVS.
I noticed that alsa-lib's make install did not install the alsa.m4 file... Maybe it has something to do with this problem?
-- Fernando

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
No, I will not fix your computer.
http://thinkgeek.com/sf

#29 lore raw

Re: Re: USB recording - repetitive peaks

At Tue, 2 Jul 2002 22:18:30 -0700 (PDT),

Quoted 11 lines · Fernando Pablo Lopez-Lezcano
Quoted 11 lines

I get this in alsa-utils

----
/usr/local/src/music/alsa/alsa-utils# ./cvscompile --with-cards=intel8x0,cmipci,usb-audio,usb-midi --with-sequencer=yes;make;make install
aclocal: configure.in: 11: macro `AM_PATH_ALSA' not found in library /usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL /usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL /usr//share/automake/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL

Quoted 5 lines

[MUNCH]

Same here from just downloaded CVS.
I noticed that alsa-lib's make install did not install the alsa.m4 file... Maybe it has something to do with this problem?

oh, yeah. it happens on a new automake.
fixed now on cvs.

thanks,

Takashi

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
No, I will not fix your computer.
http://thinkgeek.com/sf

#14 lore raw

Re: Re: USB recording - repetitive peaks

In reply to Patrick Shirkey

At Wed, 03 Jul 2002 03:04:39 +0900,

Quoted 10 lines · Patrick Shirkey
Quoted 3 lines

the error message suggests that you didn't make install on alsa-lib. please check whether /usr/share/aclocal/alsa.m4 exists.

I did run make install and have just done so again to make sure but it's not in /usr/local/share/aclocal either. I will copy it over and try recording again (see below).

Also I get this for snd-usb-midi
----

Quoted 8 lines
# modprobe snd-usb-midi
/lib/modules/2.4.19-rc1/kernel/sound/usb/snd-usb-midi.o: unresolved 
symbol snd_virmidi_new_Rc7dd286c
/lib/modules/2.4.19-rc1/kernel/sound/usb/snd-usb-midi.o: insmod 
/lib/modules/2.4.19-rc1/kernel/sound/usb/snd-usb-midi.o failed
/lib/modules/2.4.19-rc1/kernel/sound/usb/snd-usb-midi.o: insmod 
snd-usb-midi failed
----

fixed.

Quoted 3 lines

I just tried recording using the new code and completely hung the computer where only alt+sysreq+b worked. That was during recording.

hmmm.. i'll check this.

Takashi

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
No, I will not fix your computer.
http://thinkgeek.com/sf

#34 lore raw
In reply to Takashi Iwai
Quoted 8 lines · Takashi Iwai

- newly added virmidi registrations.
that means, snd-usb-midi will have its own rawmidi devices automatically as /dev/snd/midiCxDx.
this will be good for applications which speak only with rawmidi.

the latter one was not tested.
can anyone update the cvs if it works?

Using the cable number as index for the rawmidi device will not work if there is more than one endpoint (I don't think such a device exists; my driver supports more than one endpoint only because the OSS and NetBSD drivers do).

I changed it to use the sequencer port number instead.

There are devices with more output than input ports (e.g. MU1000), and virmidi with SNDRV_VIRMIDI_SEQ_ATTACH doesn't support input anyway, so I moved the rmidi pointer from usbmidi_in_port to usbmidi_out_port. (The virmidi device applies to both input and output ports, so, theoretically, it should go into a usbmidi_endpoint structure. That one hasn't been needed until now.)

Index: usbmidi.c
===================================================================
RCS file: /cvsroot/alsa/alsa-driver/usb/usbmidi.c,v retrieving revision 1.1

usbmidi.c +23 -17
diff -u -r1.1 usbmidi.c
@@ -202,6 +202,7 @@
 	int nports;
 	struct usbmidi_out_port {
 		usbmidi_out_endpoint_t* ep;
+		snd_rawmidi_t *rmidi;
 		uint8_t cable;		/* cable number << 4 */
 		uint8_t sysex_len;
 		uint8_t sysex[2];
@@ -213,7 +214,6 @@
 	urb_t* urb;
 	struct usbmidi_in_port {
 		int seq_port;
-		snd_rawmidi_t *rmidi;
 		snd_midi_event_t* midi_event;
 	} ports[0x10];
 };
@@ -573,12 +573,9 @@
 		}
 		usb_free_urb(ep->urb);
 	}
-	for (i = 0; i < 0x10; ++i) {
+	for (i = 0; i < 0x10; ++i)
 		if (ep->ports[i].midi_event)
 			snd_midi_event_free(ep->ports[i].midi_event);
-		if (ep->ports[i].rmidi)
-			snd_device_free(ep->umidi->card, ep->ports[i].rmidi);
-	}
 	snd_magic_kfree(ep);
 }
 
@@ -695,6 +692,11 @@
  */
 static void snd_usbmidi_out_endpoint_delete(usbmidi_out_endpoint_t* ep)
 {
+	int i;
+
+	for (i = 0; i < ep->nports; ++i)
+		if (ep->ports[i].rmidi)
+			snd_device_free(ep->umidi->card, ep->ports[i].rmidi);
 	if (ep->tasklet.func)
 		tasklet_kill(&ep->tasklet);
 	if (ep->urb) {
@@ -789,13 +791,14 @@
 
 /*
  * After input and output endpoints have been initialized, create
- * the ALSA port for each input/output port pair.
+ * the ALSA port for each input/output port pair in the endpoint.
+ * *port_idx is the port number, which must be unique over all endpoints.
  */
 static int snd_usbmidi_create_endpoint_ports(usbmidi_t* umidi, int ep,
 					     int* port_idx)
 {
 	usbmidi_endpoint_info_t* ep_info = &umidi->device_info.endpoints[ep];
-	int c, out_port;
+	int c, out_port, err;
 	int cap, type, port;
 	int out, in;
 	snd_seq_port_callback_t port_callback;
@@ -813,7 +816,6 @@
 		if (out) {
 			port_callback.event_input = snd_usbmidi_port_input;
 			port_callback.private_data = &umidi->out_endpoints[ep]->ports[out_port];
-			++out_port;
 			cap |= SNDRV_SEQ_PORT_CAP_WRITE |
 				SNDRV_SEQ_PORT_CAP_SUBS_WRITE;	
 		}
@@ -828,7 +830,7 @@
 		type = SNDRV_SEQ_PORT_TYPE_MIDI_GENERIC;
 		/* TODO: read port name from jack descriptor */
 		sprintf(port_name, "%s Port %d",
-			umidi->device_info.product, (*port_idx)++);
+			umidi->device_info.product, *port_idx);
 		port = snd_seq_event_port_attach(umidi->seq_client,
 						 &port_callback,
 						 cap, type, port_name);
@@ -837,23 +839,27 @@
 		if (in)
 			umidi->in_endpoints[ep]->ports[c].seq_port = port;
 
-		if (c < SNDRV_MINOR_RAWMIDIS) {
+		if (out && *port_idx < SNDRV_MINOR_RAWMIDIS) {
 			snd_rawmidi_t *rmidi;
 			snd_virmidi_dev_t *rdev;
-			if (snd_virmidi_new(umidi->card, c, &rmidi) < 0)
-				continue;
-			umidi->in_endpoints[ep]->ports[c].rmidi = rmidi;
-			rdev = snd_magic_cast(snd_virmidi_dev_t, rmidi->private_data, continue);
+			err = snd_virmidi_new(umidi->card, *port_idx, &rmidi);
+			if (err < 0)
+				return err;
+			rdev = snd_magic_cast(snd_virmidi_dev_t, rmidi->private_data, return -ENXIO);
 			strcpy(rmidi->name, port_name);
 			rdev->seq_mode = SNDRV_VIRMIDI_SEQ_ATTACH;
 			rdev->client = umidi->seq_client;
 			rdev->port = port;
-			if (snd_device_register(umidi->card, rmidi) < 0) {
+			err = snd_device_register(umidi->card, rmidi);
+			if (err < 0) {
 				snd_device_free(umidi->card, rmidi);
-				continue;
+				return err;
 			}
-			umidi->in_endpoints[ep]->ports[c].rmidi = rmidi;
+			umidi->out_endpoints[ep]->ports[out_port].rmidi = rmidi;
 		}
+		if (out)
+			++out_port;
+		++*port_idx;
 	}
 	return 0;
 }

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
No, I will not fix your computer.
http://thinkgeek.com/sf

#35 lore raw

At Wed, 03 Jul 2002 14:52:13 +0200,

Quoted 16 lines · Clemens Ladisch
Quoted 8 lines · Takashi Iwai

- newly added virmidi registrations.
that means, snd-usb-midi will have its own rawmidi devices automatically as /dev/snd/midiCxDx.
this will be good for applications which speak only with rawmidi.

the latter one was not tested.
can anyone update the cvs if it works?

Using the cable number as index for the rawmidi device will not work if there is more than one endpoint (I don't think such a device exists; my driver supports more than one endpoint only because the OSS and NetBSD drivers do).

Quoted 4 lines

I changed it to use the sequencer port number instead.

There are devices with more output than input ports (e.g. MU1000), and virmidi with SNDRV_VIRMIDI_SEQ_ATTACH doesn't support input anyway,

hmm, this mode wasn't tested, too. usb-midi is the first case. doesn't rawmidi read work at all?

Quoted 4 lines

so I moved the rmidi pointer from usbmidi_in_port to usbmidi_out_port. (The virmidi device applies to both input and output ports, so, theoretically, it should go into a usbmidi_endpoint structure. That one hasn't been needed until now.)

yep.

anyway, the patch is applied. thanks.

Takashi

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
No, I will not fix your computer.
http://thinkgeek.com/sf

#36 lore raw
Quoted 5 lines · Takashi Iwai
Quoted 1 line · Clemens Ladisch

and virmidi with SNDRV_VIRMIDI_SEQ_ATTACH doesn't support input anyway,

hmm, this mode wasn't tested, too. usb-midi is the first case. doesn't rawmidi read work at all?

The driver calls snd_seq_kernel_client_dispatch() to get rid of an event. To receive a sequencer event, virmidi would need its own sequencer client (or port), like it does with SNDRV_VIRMIDI_SEQ_DISPATCH, but then we would have two sequencer clients.

Alternatively, usb-midi could call snd_virmidi_receive() directly, but that smells somewhat like a hack. (I've noticed that snd_virmidi_receive() isn't declared static, although it isn't exported either. :-)

Clemens

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
No, I will not fix your computer.
http://thinkgeek.com/sf

#13 lore raw

Re: Re: USB recording - repetitive peaks

In reply to Patrick Shirkey

At Wed, 03 Jul 2002 03:38:41 +0900,

Quoted 23 lines · Patrick Shirkey
Quoted 10 lines · Patrick Shirkey
Quoted 7 lines · Patrick Shirkey

I did run make install and have just done so again to make sure but it's not in /usr/local/share/aclocal either. I will copy it over and try recording again (see below).

Quoted 3 lines

Recording was a bit messy, due to killing the arecord process

freezes my >> box.

Quoted 10 lines
Quoted 1 line
Quoted 1 line

Deterministically: whenever the process is interuppted (Ctrl-C) or

killed

Quoted 1 line
Quoted 1 line

(sig 9).

This seems to be the same here now. The file was recorded and with no distortions
to the sound but arecord hangs my rig when it stops. I'll check with JACK+ardour.

Quoted 4 lines

Ouch. JACK+ardour causes an instant hang. Using alt+sysreq I can sync unmount and reboot but nothing else. Also the sound quality is significantly degraded while running through JACK. At least the native drivers are sounding correct now though :)

the hang-up at capture close was at least fixed on cvs.
it was due to the combination of ASYNC_UNLINK and complete callbacks.

not sure whether the bug with jack is fixed, too, but i think it was caused by the same reason. please update the cvs tree.

Takashi

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Caffeinated soap. No kidding.
http://thinkgeek.com/sf

#37 lore raw
In reply to Clemens Ladisch

At Wed, 03 Jul 2002 16:42:43 +0200,

Quoted 5 lines · Clemens Ladisch
Quoted 3 lines · Takashi Iwai
Quoted 1 line · Clemens Ladisch

and virmidi with SNDRV_VIRMIDI_SEQ_ATTACH doesn't support input anyway,

Quoted 7 lines
Quoted 2 lines

hmm, this mode wasn't tested, too. usb-midi is the first case. doesn't rawmidi read work at all?

The driver calls snd_seq_kernel_client_dispatch() to get rid of an event. To receive a sequencer event, virmidi would need its own sequencer client (or port), like it does with SNDRV_VIRMIDI_SEQ_DISPATCH, but then we would have two sequencer clients.

oh, yes, you're right.

Quoted 3 lines

Alternatively, usb-midi could call snd_virmidi_receive() directly, but that smells somewhat like a hack. (I've noticed that snd_virmidi_receive() isn't declared static, although it isn't exported either. :-)

i think it's the easiest and cleanest way.
actually this was declared as external intentionally for such a purpose, but just forgotten to be exported :)

i changed virmidi codes for easier use, and added
snd_virmidi_receive() to usbmidi.c.
could you check the updated driver?

ciao,

Takashi

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Caffeinated soap. No kidding.
http://thinkgeek.com/sf

#38 lore raw
Quoted 3 lines · Takashi Iwai

i changed virmidi codes for easier use, and added
snd_virmidi_receive() to usbmidi.c.
could you check the updated driver?

It is obvious that you have not been able to check it. :-)

My snd_usbmidi_port_input() function actually does the _output_ (I named it 'input' because it began its life as the function to be assigend to port_callback.event_input -- well, that's input _to_ the usb-midi port. A better name for it would be snd_usbmidi_event_for_output).

The call to snd_virmidi_receive() should go into snd_usbmidi_input_packet.

And now it turns out that moving the rmidi pointer from the input to the output endpoint structure was rather stupid by me because rmidi is needed for input only. That change needs to be reverted, too.

The virmidi creation in snd_usbmidi_create_endpoint_ports should look like the following: (warning: untested and made by hand because I don't have access to my Linux box right now. patch probably won't accept those line numbers! :-)

@@ at the end of snd_usbmidi_create_endpoint_ports

	if (in)
		umidi->in_endpoints[ep]->ports[c].seq_port = port;

- if (out && *port_idx < SNDRV_MINOR_RAWMIDIS) {

+	if (in && *port_idx < SNDRV_MINOR_RAWMIDIS) {
		snd_rawmidi_t *rmidi;
		snd_virmidi_dev_t *rdev;
		err = snd_virmidi_new(umidi->card, *port_idx, &rmidi);
@@ a few lines later
			snd_device_free(umidi->card, rmidi);
			return err;
		}
-		umidi->out_endpoints[ep]->ports[out_port].rmidi = rmidi;
+		umidi->in_endpoints[ep]->ports[c].rmidi = rmidi;
	}
	if (out)
		++out_port;

This will not create virmidi ports for output only ports, but it should actually work for the rest. :)

I think I'll be able to submit a patch for storing rmidi independently of the input/output structures tomorrow.

Clemens

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Caffeinated soap. No kidding.
http://thinkgeek.com/sf

#15 lore raw

Re: Re: USB recording - repetitive peaks

In reply to Takashi Iwai
Quoted 4 lines · Takashi Iwai

the hang-up at capture close was at least fixed on cvs.
it was due to the combination of ASYNC_UNLINK and complete callbacks.

not sure whether the bug with jack is fixed, too, but i think it was >caused by the same reason. please update the cvs tree.

I can now record successfully using arecord without hanging the system
but now I get an instant lock up while starting ardour. :( I have a suspicion that it happens when ardour registers with jackd and asks for the available input devices. (I'm not sure about the sound quality. I will see if I can generate a test tone to use instead of actual music).

The only way to get out of it is reboot using sysreq.

I will check some more. Please note this wasn't happening so quickly yesterday.

I'm not sure about the sound quality it is definitely better. I will see if I can generate a test tone to use instead of actual music.

Signature · 12 lines
Patrick Shirkey - Boost Hardware Ltd.
For the discerning hardware connoisseur
Http://www.boosthardware.com
Http://www.boosthardware.com/LAU/guide/
========================================



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Caffeinated soap. No kidding.
http://thinkgeek.com/sf
#16 lore raw

Re: Re: USB recording - repetitive peaks

At Fri, 05 Jul 2002 01:23:18 +0900,

Quoted 6 lines · Patrick Shirkey
Quoted 3 lines · Takashi Iwai

the hang-up at capture close was at least fixed on cvs.
it was due to the combination of ASYNC_UNLINK and complete callbacks.

Quoted 8 lines
Quoted 2 lines

not sure whether the bug with jack is fixed, too, but i think it was caused by the same reason. please update the cvs tree.

I can now record successfully using arecord without hanging the system but now I get an instant lock up while starting ardour. :( I have a suspicion that it happens when ardour registers with jackd and asks for the available input devices. (I'm not sure about the sound quality. I will see if I can generate a test tone to use instead of actual music).

then could you remove the mask USB_ASYNC_UNLINK from line 826 of usbaudio.c?
this might also leads to a lock-up when usb_submit_urb() fails, but this must happen rarely.

Takashi

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Caffeinated soap. No kidding.
http://thinkgeek.com/sf

#17 lore raw

Re: Re: USB recording - repetitive peaks

Quoted 10 lines · Takashi Iwai

At Fri, 05 Jul 2002 01:23:18 +0900,

Quoted 7 lines · Patrick Shirkey
Quoted 4 lines · Takashi Iwai

the hang-up at capture close was at least fixed on cvs.
it was due to the combination of ASYNC_UNLINK and complete callbacks.

not sure whether the bug with jack is fixed, too, but i think it was caused by the same reason. please update the cvs tree.

Quoted 8 lines
Quoted 2 lines

I can now record successfully using arecord without hanging the system
but now I get an instant lock up while starting ardour. :( I have a suspicion that it happens when ardour registers with jackd and asks for the available input devices. (I'm not sure about the sound quality. I will see if I can generate a test tone to use instead of actual music).

then could you remove the mask USB_ASYNC_UNLINK from line 826 of usbaudio.c? this might also leads to a lock-up when usb_submit_urb() fails, but this must happen rarely.

First I tried

		u->urb->transfer_flags = USB_ISO_ASAP; /*  |  USB_ASYNC_UNLINK; */

but that made almost no difference. I was able to load ardour for about 10 seconds before the machine hung.

Now I have commented out the line and I get this from jackd

----

$ jackd -v -d alsa -d quattro2
jackd 0.37.1
Copyright 2001-2002 Paul Davis and others.
jackd comes with ABSOLUTELY NO WARRANTY
This is free software, and you are welcome to redistribute it
under certain conditions; see the file COPYING for details

2616 waiting for signals
creating alsa driver ... quattro2|1024|2|48000|swmon
new client: alsa_pcm, id = 1 type 1 @ 0x80664b0 fd = 12
port alsa_pcm:in_1 buf shm key 0x30945346 at offset 4096 bi = 0x8065880 registered port alsa_pcm:in_1, offset = 4096
port alsa_pcm:in_2 buf shm key 0x30945346 at offset 8192 bi = 0x8065890 registered port alsa_pcm:in_2, offset = 8192
registered port alsa_pcm:out_1, offset = 0
registered port alsa_pcm:out_2, offset = 0
-- jack_rechain_graph():
client alsa_pcm: inprocess client, execution_order=0.
ALSA: poll time out polled for 44800.813378
driver wait function failed, exiting
telling signal thread that the engine is done
jack main caught signal 1

----

Signature · 12 lines
Patrick Shirkey - Boost Hardware Ltd.
For the discerning hardware connoisseur
Http://www.boosthardware.com
Http://www.boosthardware.com/LAU/guide/
========================================



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Caffeinated soap. No kidding.
http://thinkgeek.com/sf
#18 lore raw

Re: Re: USB recording - repetitive peaks

Quoted 5 lines

client alsa_pcm: inprocess client, execution_order=0.
ALSA: poll time out polled for 44800.813378 driver wait function failed, exiting
telling signal thread that the engine is done jack main caught signal 1

which means that the JACK alsa driver/client didn't return from poll on the ALSA streams within the given timeout: ie. one or both of them is not running.

--p

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Caffeinated soap. No kidding.
http://thinkgeek.com/sf

#22 lore raw

Re: Re: USB recording - repetitive peaks

In reply to Patrick Shirkey

Patrick, I'll checkout cvs tomorrow morning, compile and test around a bit with a rc-generator and the oscilloscope. I'll send the results to the list.

Am Donnerstag, 4. Juli 2002 18:23 schrieb Patrick Shirkey:

Quoted 13 lines
Quoted 5 lines · Takashi Iwai

the hang-up at capture close was at least fixed on cvs.
it was due to the combination of ASYNC_UNLINK and complete callbacks.

not sure whether the bug with jack is fixed, too, but i think it was caused by the same reason. please update the cvs tree.

I can now record successfully using arecord without hanging the system but now I get an instant lock up while starting ardour. :( I have a suspicion that it happens when ardour registers with jackd and asks for the available input devices. (I'm not sure about the sound quality. I will see if I can generate a test tone to use instead of actual music).

Quoted 7 lines

The only way to get out of it is reboot using sysreq.

I will check some more. Please note this wasn't happening so quickly yesterday.

I'm not sure about the sound quality it is definitely better. I will see if I can generate a test tone to use instead of actual music.

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Caffeinated soap. No kidding.
http://thinkgeek.com/sf

#23 lore raw

Re: Re: USB recording - repetitive peaks

Quoted 2 lines
Quoted 2 lines

I'm not sure about the sound quality it is definitely better. I will see if I can generate a test tone to use instead of actual music.

I checked out cvs this morning and compiled. I grabbed my function generator and fed my quattro usb with some signal; 400Hz, 0,7V peak-to-peak.

It's working finely now. (no more peaks)

I did the same thing some days before. ( Whoever is interested in the results and the output of the oscilloscope: http://www.thorstenhaas.de/alsa.php)

Am Donnerstag, 4. Juli 2002 20:59 schrieb Thorsten Haas:

Quoted 23 lines

Patrick, I'll checkout cvs tomorrow morning, compile and test around a bit with a rc-generator and the oscilloscope. I'll send the results to the list.

Am Donnerstag, 4. Juli 2002 18:23 schrieb Patrick Shirkey:

Quoted 13 lines
Quoted 5 lines · Takashi Iwai

the hang-up at capture close was at least fixed on cvs.
it was due to the combination of ASYNC_UNLINK and complete callbacks.

not sure whether the bug with jack is fixed, too, but i think it was caused by the same reason. please update the cvs tree.

I can now record successfully using arecord without hanging the system but now I get an instant lock up while starting ardour. :( I have a suspicion that it happens when ardour registers with jackd and asks for the available input devices. (I'm not sure about the sound quality. I will see if I can generate a test tone to use instead of actual music).

Quoted 4 lines

The only way to get out of it is reboot using sysreq.

I will check some more. Please note this wasn't happening so quickly yesterday.

Quoted 8 lines

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Caffeinated soap. No kidding.
http://thinkgeek.com/sf
_______________________________________________
Alsa-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/alsa-devel

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Bringing you mounds of caffeinated joy.
http://thinkgeek.com/sf

#39 lore raw
In reply to Clemens Ladisch
Quoted 2 lines · I wrote:

I think I'll be able to submit a patch for storing rmidi independently of the input/output structures tomorrow.

And here it is.

- moved rmidi to new structure usbmidi_endpoint
- moved call to snd_virmidi_receive to snd_usbmidi_input_packet
- renamed snd_usbmidi_port_input to snd_usbmidi_event_input
- removed dynamic allocation of usbmidi_out_port structures

Index: usbmidi.c
===================================================================
RCS file: /cvsroot/alsa/alsa-driver/usb/usbmidi.c,v retrieving revision 1.3

usbmidi.c +50 -53
diff -u -r1.3 usbmidi.c
@@ -155,6 +155,7 @@
 typedef struct usbmidi usbmidi_t;
 typedef struct usbmidi_device_info usbmidi_device_info_t;
 typedef struct usbmidi_endpoint_info usbmidi_endpoint_info_t;
+typedef struct usbmidi_endpoint usbmidi_endpoint_t;
 typedef struct usbmidi_out_endpoint usbmidi_out_endpoint_t;
 typedef struct usbmidi_out_port usbmidi_out_port_t;
 typedef struct usbmidi_in_endpoint usbmidi_in_endpoint_t;
@@ -183,9 +184,11 @@
 	int dev;
 	int seq_client;
 	usbmidi_device_info_t device_info;
-
-	usbmidi_out_endpoint_t* out_endpoints[MAX_ENDPOINTS];
-	usbmidi_in_endpoint_t* in_endpoints[MAX_ENDPOINTS];
+	struct usbmidi_endpoint {
+		usbmidi_out_endpoint_t* out;
+		usbmidi_in_endpoint_t* in;
+		snd_rawmidi_t* rmidi[0x10];
+	} endpoints[MAX_ENDPOINTS];
 };
 
 struct usbmidi_out_endpoint {
@@ -199,18 +202,17 @@
 	int data_size;
 	spinlock_t buffer_lock;
 
-	int nports;
 	struct usbmidi_out_port {
 		usbmidi_out_endpoint_t* ep;
-		snd_rawmidi_t *rmidi;
 		uint8_t cable;		/* cable number << 4 */
 		uint8_t sysex_len;
 		uint8_t sysex[2];
-	} ports[0];
+	} ports[0x10];
 };
 
 struct usbmidi_in_endpoint {
 	usbmidi_t* umidi;
+	usbmidi_endpoint_t* ep;
 	urb_t* urb;
 	struct usbmidi_in_port {
 		int seq_port;
@@ -258,7 +260,8 @@
 	static const uint8_t cin_length[] = {
 		0, 0, 2, 3, 3, 1, 2, 3, 3, 3, 3, 3, 2, 2, 3, 1
 	};
-	usbmidi_in_port_t* port = &ep->ports[packet[0] >> 4];
+	int cable = packet[0] >> 4;
+	usbmidi_in_port_t* port = &ep->ports[cable];
 	snd_seq_event_t ev;
 
 	if (!port->midi_event)
@@ -271,6 +274,8 @@
 		ev.dest.client = SNDRV_SEQ_ADDRESS_SUBSCRIBERS;
 		snd_seq_kernel_client_dispatch(ep->umidi->seq_client,
 					       &ev, 1, 0);
+		if (ep->ep->rmidi[cable])
+			snd_virmidi_receive(ep->ep->rmidi[cable], &ev);
 	}
 }
 
@@ -454,8 +459,8 @@
 /*
  * Converts an ALSA sequencer event into USB MIDI packets.
  */
-static int snd_usbmidi_port_input(snd_seq_event_t* ev, int direct,
-				  void* private_data, int atomic, int hop)
+static int snd_usbmidi_event_input(snd_seq_event_t* ev, int direct,
+				   void* private_data, int atomic, int hop)
 {
 	usbmidi_out_port_t* port = (usbmidi_out_port_t*)private_data;
 	int err;
@@ -554,7 +559,6 @@
 	default:
 		return 0;
 	}
-	snd_virmidi_receive(port->rmidi, ev);
 	tasklet_hi_schedule(&port->ep->tasklet);
 	return 0;
 }
@@ -617,7 +621,7 @@
  */
 static int snd_usbmidi_in_endpoint_create(usbmidi_t* umidi,
 					  usbmidi_endpoint_info_t* ep_info,
-					  usbmidi_in_endpoint_t** rep)
+					  usbmidi_endpoint_t* rep)
 {
 	usbmidi_in_endpoint_t* ep;
 	int do_int_transfer;
@@ -626,11 +630,12 @@
 	unsigned int pipe;
 	int length, i, err;
 
-	*rep = NULL;
+	rep->in = NULL;
 	ep = snd_magic_kcalloc(usbmidi_in_endpoint_t, 0, GFP_KERNEL);
 	if (!ep)
 		return -ENOMEM;
 	ep->umidi = umidi;
+	ep->ep = rep;
 	for (i = 0; i < 0x10; ++i)
 		ep->ports[i].seq_port = -1;
 
@@ -674,7 +679,7 @@
 			}
 		}
 
-	*rep = ep;
+	rep->in = ep;
 	return 0;
 }
 
@@ -693,11 +698,6 @@
  */
 static void snd_usbmidi_out_endpoint_delete(usbmidi_out_endpoint_t* ep)
 {
-	int i;
-
-	for (i = 0; i < ep->nports; ++i)
-		if (ep->ports[i].rmidi)
-			snd_device_free(ep->umidi->card, ep->ports[i].rmidi);
 	if (ep->tasklet.func)
 		tasklet_kill(&ep->tasklet);
 	if (ep->urb) {
@@ -716,17 +716,15 @@
  */
 static int snd_usbmidi_out_endpoint_create(usbmidi_t* umidi,
 					   usbmidi_endpoint_info_t* ep_info,
-			 		   usbmidi_out_endpoint_t** rep)
+			 		   usbmidi_endpoint_t* rep)
 {
 	usbmidi_out_endpoint_t* ep;
-	int nports, i;
+	int i;
 	unsigned int pipe;
 	void* buffer;
 
-	*rep = NULL;
-	nports = snd_usbmidi_count_bits(ep_info->out_cables);
-	ep = snd_magic_kcalloc(usbmidi_out_endpoint_t,
-			       nports * sizeof(usbmidi_out_port_t), GFP_KERNEL);
+	rep->out = NULL;
+	ep = snd_magic_kcalloc(usbmidi_out_endpoint_t, 0, GFP_KERNEL);
 	if (!ep)
 		return -ENOMEM;
 	ep->umidi = umidi;
@@ -749,16 +747,13 @@
 	spin_lock_init(&ep->buffer_lock);
 	tasklet_init(&ep->tasklet, snd_usbmidi_out_tasklet, (unsigned long)ep);
 
-	ep->nports = nports;
-	nports = 0;
 	for (i = 0; i < 0x10; ++i)
 		if (ep_info->out_cables & (1 << i)) {
-			ep->ports[nports].ep = ep;
-			ep->ports[nports].cable = i << 4;
-			++nports;
+			ep->ports[i].ep = ep;
+			ep->ports[i].cable = i << 4;
 		}
 
-	*rep = ep;
+	rep->out = ep;
 	return 0;
 }
 
@@ -768,7 +763,7 @@
 static int snd_usbmidi_seq_device_delete(snd_seq_device_t* seq_device)
 {
 	usbmidi_t* umidi;
-	int i;
+	int i, j;
 
 	umidi = (usbmidi_t*)SNDRV_SEQ_DEVICE_ARGPTR(seq_device);
 
@@ -776,16 +771,21 @@
 		snd_seq_delete_kernel_client(umidi->seq_client);
 		umidi->seq_client = -1;
 	}
-
 	for (i = 0; i < MAX_ENDPOINTS; ++i) {
-		if (umidi->out_endpoints[i]) {
-			snd_usbmidi_out_endpoint_delete(umidi->out_endpoints[i]);
-			umidi->out_endpoints[i] = NULL;
-		}
-		if (umidi->in_endpoints[i]) {
-			snd_usbmidi_in_endpoint_delete(umidi->in_endpoints[i]);
-			umidi->in_endpoints[i] = NULL;
-		}
+		usbmidi_endpoint_t* ep = &umidi->endpoints[i];
+		if (ep->out) {
+			snd_usbmidi_out_endpoint_delete(ep->out);
+			ep->out = NULL;
+		}
+		if (ep->in) {
+			snd_usbmidi_in_endpoint_delete(ep->in);
+			ep->in = NULL;
+		}
+		for (j = 0; j < 0x10; ++j)
+			if (ep->rmidi[j]) {
+				snd_device_free(umidi->card, ep->rmidi[j]);
+				ep->rmidi[j] = NULL;
+			}
 	}
 	return 0;
 }
@@ -799,13 +799,12 @@
 					     int* port_idx)
 {
 	usbmidi_endpoint_info_t* ep_info = &umidi->device_info.endpoints[ep];
-	int c, out_port, err;
+	int c, err;
 	int cap, type, port;
 	int out, in;
 	snd_seq_port_callback_t port_callback;
 	char port_name[40];
 
-	out_port = 0;
 	for (c = 0; c < 0x10; ++c) {
 		out = ep_info->out_cables & (1 << c);
 		in = ep_info->in_cables & (1 << c);
@@ -815,8 +814,8 @@
 		memset(&port_callback, 0, sizeof(port_callback));
 		port_callback.owner = THIS_MODULE;
 		if (out) {
-			port_callback.event_input = snd_usbmidi_port_input;
-			port_callback.private_data = &umidi->out_endpoints[ep]->ports[out_port];
+			port_callback.event_input = snd_usbmidi_event_input;
+			port_callback.private_data = &umidi->endpoints[ep].out->ports[c];
 			cap |= SNDRV_SEQ_PORT_CAP_WRITE |
 				SNDRV_SEQ_PORT_CAP_SUBS_WRITE;	
 		}
@@ -838,9 +837,9 @@
 		if (port < 0)
 			return port;
 		if (in)
-			umidi->in_endpoints[ep]->ports[c].seq_port = port;
+			umidi->endpoints[ep].in->ports[c].seq_port = port;
 
-		if (out && *port_idx < SNDRV_MINOR_RAWMIDIS) {
+		if (*port_idx < SNDRV_MINOR_RAWMIDIS) {
 			snd_rawmidi_t *rmidi;
 			snd_virmidi_dev_t *rdev;
 			err = snd_virmidi_new(umidi->card, *port_idx, &rmidi);
@@ -856,10 +855,8 @@
 				snd_device_free(umidi->card, rmidi);
 				return err;
 			}
-			umidi->out_endpoints[ep]->ports[out_port].rmidi = rmidi;
+			umidi->endpoints[ep].rmidi[c] = rmidi;
 		}
-		if (out)
-			++out_port;
 		++*port_idx;
 	}
 	return 0;
@@ -879,13 +876,13 @@
 			continue;
 		if (ep_info->out_cables) {
 			err = snd_usbmidi_out_endpoint_create(umidi, ep_info,
-							      &umidi->out_endpoints[i]);
+							      &umidi->endpoints[i]);
 			if (err < 0)
 				return err;
 		}
 		if (ep_info->in_cables) {
 			err = snd_usbmidi_in_endpoint_create(umidi, ep_info,
-							     &umidi->in_endpoints[i]);
+							     &umidi->endpoints[i]);
 			if (err < 0)
 				return err;
 		}
@@ -936,8 +933,8 @@
 	}
 
 	for (i = 0; i < MAX_ENDPOINTS; ++i)
-		if (umidi->in_endpoints[i])
-			snd_usbmidi_submit_urb(umidi->in_endpoints[i]->urb,
+		if (umidi->endpoints[i].in)
+			snd_usbmidi_submit_urb(umidi->endpoints[i].in->urb,
 					       GFP_KERNEL);
 	return 0;
 }

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Bringing you mounds of caffeinated joy.
http://thinkgeek.com/sf

#24 lore raw

Re: Re: USB recording - repetitive peaks

In reply to Thorsten Haas
Quoted 11 lines · Thorsten Haas
Quoted 3 lines
Quoted 2 lines

I'm not sure about the sound quality it is definitely better. I will see if I can generate a test tone to use instead of actual music.

I checked out cvs this morning and compiled. I grabbed my function generator and fed my quattro usb with some signal; 400Hz, 0,7V peak-to-peak.

It's working finely now. (no more peaks)

I did the same thing some days before. ( Whoever is interested in the results and the output of the oscilloscope: http://www.thorstenhaas.de/alsa.php)

That looks much better.

We still face the problem of accessing all 4 channels at once. AFAICT in the .asoundrc file the multi interface option does not have the ability to seperately access the channels in non sequential order. Therefore we are currently unable to access all the channels on the card because pcm0 and pcm1 point to the same physical channels input 1 and input 2.

We need to either reorganise the pcm devices for the card or make it possible to secify channels in non sequential order using the multi interface option in a .asoundrc.

Signature · 12 lines
Patrick Shirkey - Boost Hardware Ltd.
For the discerning hardware connoisseur
Http://www.boosthardware.com
Http://www.boosthardware.com/LAU/guide/
========================================



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Bringing you mounds of caffeinated joy.
http://thinkgeek.com/sf
#25 lore raw

Re: Re: USB recording - repetitive peaks

Am Freitag, 5. Juli 2002 10:34 schrieb Patrick Shirkey:

Quoted 22 lines
Quoted 7 lines · Thorsten Haas
Quoted 2 lines
Quoted 2 lines

I'm not sure about the sound quality it is definitely better. I will see if I can generate a test tone to use instead of actual music.

I checked out cvs this morning and compiled. I grabbed my function generator and fed my quattro usb with some signal; 400Hz, 0,7V peak-to-peak.

Quoted 5 lines

It's working finely now. (no more peaks)

I did the same thing some days before. ( Whoever is interested in the results and the output of the oscilloscope:
http://www.thorstenhaas.de/alsa.php)

That looks much better.

We still face the problem of accessing all 4 channels at once. AFAICT in the .asoundrc file the multi interface option does not have the ability to seperately access the channels in non sequential order. Therefore we are currently unable to access all the channels on the card because pcm0 and pcm1 point to the same physical channels input 1 and input 2.

Quoted 3 lines

We need to either reorganise the pcm devices for the card or make it possible to secify channels in non sequential order using the multi interface option in a .asoundrc.

I wish I was much more experienced in programming C++ and the structure of ALSA. Actually I am not. So I can hardly follow you what this will mean concerning effort and complexity.

So, what will it be?

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Bringing you mounds of caffeinated joy.
http://thinkgeek.com/sf

#26 lore raw

Re: Re: USB recording - repetitive peaks

I like to use two soundcards in my system.

I have an usb quattro and a trident compatible card. I read the manual at
http://www.boosthardware.com/LAU/guide/ and configured modules.conf:#

------------------------- snip -------------------------------------- ### NEW ALSA 0.9 from CVS 2002-07-05

## General
# ALSA portion
alias char-major-116 snd
# OSS/Free portion
alias char-major-14 soundcore

## Native Devices: ALSA portion

# quattro usb
  alias snd-card-0 snd-usb-audio
  options snd-usb-audio snd_index=0 snd_id="Q4"
# sis onboard (trident)
  alias snd-card-1 snd-card-trident
  options snd-card-trident snd_index=1 snd_id="SiS"

## Native Devices: OSS/Free portion
# OSS Free Portion - Card 1 - quattro usb
alias sound-slot-0 snd-card-0
alias sound-service-0-0 snd-mixer-oss
alias sound-service-0-1 snd-seq-oss
alias sound-service-0-3 snd-pcm-oss
alias sound-service-0-8 snd-seq-oss
alias sound-service-0-12 snd-pcm-oss
# OSS Free Portion - Card 2 - Trident
alias sound-slot-1 snd-card-1
alias sound-service-1-0 snd-mixer-oss
alias sound-service-1-3 snd-pcm-oss
alias sound-service-1-12 snd-pcm-oss
------------------------- snap --------------------------------------

After that i ran depmod -a and tried

linda:~ # modprobe snd-card-trident /lib/modules/2.4.10-4GB/misc/snd-rawmidi.o: unresolved symbol snd_control_unregister_ioctl /lib/modules/2.4.10-4GB/misc/snd-rawmidi.o: unresolved symbol snd_switch_remove
/lib/modules/2.4.10-4GB/misc/snd-rawmidi.o: unresolved symbol snd_switch_free /lib/modules/2.4.10-4GB/misc/snd-rawmidi.o: unresolved symbol snd_switch_add /lib/modules/2.4.10-4GB/misc/snd-rawmidi.o: unresolved symbol snd_switch_prepare
/lib/modules/2.4.10-4GB/misc/snd-rawmidi.o: unresolved symbol snd_info_create_entry
/lib/modules/2.4.10-4GB/misc/snd-rawmidi.o: unresolved symbol snd_control_register_ioctl /lib/modules/2.4.10-4GB/misc/snd-rawmidi.o: unresolved symbol snd_switch_new /lib/modules/2.4.10-4GB/misc/snd-rawmidi.o: unresolved symbol snd_switch_free_one
/lib/modules/2.4.10-4GB/misc/snd-rawmidi.o: insmod /lib/modules/2.4.10-4GB/misc/snd-rawmidi.o failed /lib/modules/2.4.10-4GB/misc/snd-rawmidi.o: insmod snd-card-trident failed

Can someone please give me ahint what I did wrong here?

Thorsten

PS: SORRY for the length.

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Bringing you mounds of caffeinated joy.
http://thinkgeek.com/sf

#40 lore raw
In reply to Clemens Ladisch

At Fri, 05 Jul 2002 09:55:27 +0200,

Quoted 8 lines · Clemens Ladisch
Quoted 2 lines · I wrote:

I think I'll be able to submit a patch for storing rmidi independently of the input/output structures tomorrow.

And here it is.

- moved rmidi to new structure usbmidi_endpoint

ok, that's less confusing than before :)

Quoted 3 lines

- moved call to snd_virmidi_receive to snd_usbmidi_input_packet
- renamed snd_usbmidi_port_input to snd_usbmidi_event_input
- removed dynamic allocation of usbmidi_out_port structures

thanks, now applied to cvs.

ciao,

Takashi

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Bringing you mounds of caffeinated joy.
http://thinkgeek.com/sf

#19 lore raw

Re: Re: USB recording - repetitive peaks

In reply to Patrick Shirkey

At Fri, 05 Jul 2002 02:15:39 +0900,

Quoted 26 lines · Patrick Shirkey
Quoted 18 lines · Takashi Iwai

At Fri, 05 Jul 2002 01:23:18 +0900,

Quoted 5 lines · Patrick Shirkey
Quoted 3 lines · Takashi Iwai

the hang-up at capture close was at least fixed on cvs.
it was due to the combination of ASYNC_UNLINK and complete callbacks.

Quoted 8 lines
Quoted 2 lines

not sure whether the bug with jack is fixed, too, but i think it was caused by the same reason. please update the cvs tree.

I can now record successfully using arecord without hanging the system but now I get an instant lock up while starting ardour. :( I have a suspicion that it happens when ardour registers with jackd and asks for the available input devices. (I'm not sure about the sound quality. I will see if I can generate a test tone to use instead of actual music).

Quoted 5 lines

then could you remove the mask USB_ASYNC_UNLINK from line 826 of usbaudio.c?
this might also leads to a lock-up when usb_submit_urb() fails, but this must happen rarely.

Quoted 6 lines

First I tried

		u->urb->transfer_flags = USB_ISO_ASAP; /*  |  USB_ASYNC_UNLINK; */

but that made almost no difference. I was able to load ardour for about 10 seconds before the machine hung.

i (hopefully) found the spot.
it happend when snd_pcm_stop() is called during complete callbacks, so the working urb won't be released properly.
now the fixed version is on cvs.

Takashi

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Bringing you mounds of caffeinated joy.
http://thinkgeek.com/sf

#27 lore raw

Re: Re: USB recording - repetitive peaks

In reply to Thorsten Haas

At Fri, 5 Jul 2002 11:06:06 +0200,

Quoted 15 lines · Thorsten Haas

I like to use two soundcards in my system.

I have an usb quattro and a trident compatible card. I read the manual at
http://www.boosthardware.com/LAU/guide/ and configured modules.conf:#

------------------------- snip -------------------------------------- ### NEW ALSA 0.9 from CVS 2002-07-05

## General
# ALSA portion
alias char-major-116 snd
# OSS/Free portion
alias char-major-14 soundcore

Quoted 6 lines

## Native Devices: ALSA portion

# quattro usb
  alias snd-card-0 snd-usb-audio
  options snd-usb-audio snd_index=0 snd_id="Q4"
# sis onboard (trident)
  alias snd-card-1 snd-card-trident
                     ^^^^^^^^^^^^^^^^  snd-trident

snd-card-xxx is an obsolete name for alsa 0.5.x.

Quoted 1 line

options snd-card-trident snd_index=1 snd_id="SiS"

            ^^^^^^^^^^^^^^^^ here too
Quoted 4 lines

After that i ran depmod -a and tried

linda:~ # modprobe snd-card-trident /lib/modules/2.4.10-4GB/misc/snd-rawmidi.o: unresolved symbol

you have both alsa 0.5.x and 0.9.0 modules on the same kernel tree. please remove old ones (snd-*.o) under /lib/modules/2.4.10-4GB/misc/.

ciao,

Takashi

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Bringing you mounds of caffeinated joy.
http://thinkgeek.com/sf

#20 lore raw

Re: Re: USB recording - repetitive peaks

In reply to Takashi Iwai
Quoted 6 lines · Takashi Iwai

i (hopefully) found the spot.
it happend when snd_pcm_stop() is called during complete callbacks, so the working urb won't be released properly.
now the fixed version is on cvs.

Apparently not the g-spot.

I just tried using jack+ardour and the system hung again. This time it may have been triggered when I tried to switch virtual desktops. (I'm running enlightenment).

Everything was working (apart from peaks in the sound quality similar to the ones we were hearing natively only more pronounced). When I tried to switch into another part of the desktop everything hung.

I will try again because it may have been a time issue and therefore a complete coincidence.

Signature · 12 lines
Patrick Shirkey - Boost Hardware Ltd.
For the discerning hardware connoisseur
Http://www.boosthardware.com
Http://www.boosthardware.com/LAU/guide/
========================================



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Bringing you mounds of caffeinated joy.
http://thinkgeek.com/sf
#21 lore raw

Re: Re: USB recording - repetitive peaks

Quoted 9 lines · Patrick Shirkey
Quoted 6 lines · Takashi Iwai

i (hopefully) found the spot.
it happend when snd_pcm_stop() is called during complete callbacks, so the working urb won't be released properly.
now the fixed version is on cvs.

Quoted 9 lines

Apparently not the g-spot.

I just tried using jack+ardour and the system hung again. This time it may have been triggered when I tried to switch virtual desktops. (I'm running enlightenment).

Everything was working (apart from peaks in the sound quality similar to the ones we were hearing natively only more pronounced). When I tried to switch into another part of the desktop everything hung.

I will try again because it may have been a time issue and therefore a complete coincidence.

It seems to be directly related to stress on xfree which causes an xrun in jackd.

This time when I started up I had a clear signal and could switch screens without dropouts but when I downloaded some messages I caused an xrun and then the sound got all choppy. Still no hang so I tried moving the ardour editor window to a different screen and that was it.

I have a usb mouse and am using the nvidia drivers for my video card which is a 64MB geforce2. Plus I was running jackd as normal user.

However I can get normal recording and xfree performance if I don't use jack. That is whether I am root user or not.

So it sounds like you have sorted the problem (unless you can think of anything else) and I need to do some serious fine tuning to get JACK working correctly.

We still need a way to access all four channels at the same time though.

Signature · 12 lines
Patrick Shirkey - Boost Hardware Ltd.
For the discerning hardware connoisseur
Http://www.boosthardware.com
Http://www.boosthardware.com/LAU/guide/
========================================



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Bringing you mounds of caffeinated joy.
http://thinkgeek.com/sf