Skip to contentSkip to conversation

unresolved bugs in the Debian BTS

DISCUSSION Jordi Mallach 2 participants 4 messages Mar 6, 2003 – Mar 10, 2003

lore
Share & tools
lore
#0 lore raw

Hello!

As promised, here are the bugs we have on our bug tracking system. I hope some will have been fixed in some recent release and we haven't been able to confirm. Others have been reported against rc7 (what is in Debian, currently).

All the logs for a given bugs are in http://bugs.debian.org/bugnumber.

#######################
Bug 183316: OSS-emulation MIDI timer problems

SEQ_START_TIMER(), if used more than once for a given /dev/sequencer session, causes odd things to happen. I'm not quite sure what's going on here, but investigaton using lxmusserv and rosegarden shows that timer events seem to be occurring in bursts.

With Doom, let some piece of music be repeated then try to do something (start a game, quit) and you'll find that it doesn't happen immediately; instead, it happens some time later; mwanwhile, the same music continues to be played as the game engine is blocked, waiting for musserv to read in the new music data.

With rosegarden, the effect can be seen by starting, stopping and restarting the music: watch the timer apparently make several large jumps (each corresponding to many SNDCTL_SEQ_SYNC + SEQ_WAIT_TIME pairs) instead of lots of small jumps (each of which corresponds to one SNDCTL_SEQ_SYNC + SEQ_WAIT_TIME pair).

########################
Bug 183315: Sequencer device opening fails with ENOMEM

After some time (several open/close cycles of /dev/sequencer?), the next attempt to use the device will cause any open attempt to return ENOMEM. The usage counts of snd-emu10k1-synth, snd-emux-synth and snd-emu10k1 become non-zero and are incremented for each open failure, rendering them (AFAIK) unremovable without a reboot.

To trigger this, I've found it sufficient to run some lxmusserv-using Doom port several times [1]; though I don't think that it matters which program is used, I don't currently know whether it matters if AWE emulation/support is used.

The problem appears to only affect OSS sequencer usage; pmidi and programs which use other OSS devices still work.

########################
Bug 182840: hw_pointer warning with 845PE chipset

this system has an Asustek P4PE motherbooard with an Intel 845PE chipset. Here is the lspci output for the integrated audio controller:

00:1f.5 Multimedia audio controller: Intel Corp. 82801DB AC'97 Audio (rev 02)

I finally managed to get sound out of this motherboard with alsaplayer (not with the ALSA xmms output plugin yet).

So all is well, except that I saw this warning message in the dmesg output:

ALSA ../alsa-kernel/core/pcm_lib.c:187: Unexpected hw_pointer value (stream = 0, delta: -1168, max jitter = 8192): wrong interrupt acknowledge?

PS: the other soundcard (ens1371) is managed by OSS/Free.

#########################
Bug 181890: System crash after using sound device

The submitter says the kernel and the modules were compiled with the same compiler version. Sound card is a Crystal Sound 4237B or 4236B, he's not sure. He gets a Null pointer dereference kernel error.

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=181890

#########################
Bug 180653: Bad audio quality on VT8233 with rc7

Quoting:
-----------------
The quality of the audio output has decreased very strongly from rc6 to rc7. With rc6, everything was working fine, but with rc7 there is always a bit of noise when sound is played (I had the same problem with older versions of alsa, but rc6 was working fine).

My sound card is an on-board VT8233 chip, and I'm running kernel 2.4.20. I didn't try rc7-1.
------------------

Could this be related to the bad detection code in rc7 for VIA8233?

#########################

Bug 177522: aplay segfaults on multiple files

Submitted against RC3, but I think it's still there:

-----------------
sorry fellas, when one mixes .au and .wav, or does more than one of the same file, e.g.

$ y=/usr/share/emacs/site-lisp/emacspeak/sounds/default-8k/yank-object.au
$ aplay $y $y $y $y
one wont hear all what is requested, and will even seg fault!

aplay: pcm.c:5692: snd_pcm_unlink_ptr: Assertion `0' failed.
Aborted...
aplay: pcm_plug.c:68: snd_pcm_plug_close: Assertion `plug->slave == plug->req_slave' failed. Aborted...
aplay: pcm_plug.c:68: snd_pcm_plug_close: Assertion `plug->slave == plug->req_slave' failed.

that line repeated 4000 times, then
Segmentation fault
-----------------

Is this a bug in alsa-utils, or in an old rc3 driver?

########################

The following bugs are a bit older, so I'll just give urls for the bug logs. We would appreciate some feedback, if you know they have been fixed in recent release candidates

########################
Bug 174690: alsamixer silently fails with ALI5451 South Bridge card

(against beta12)

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=174690

########################
Bug 172967: "AC'97 0:0 does not respond"

(against rc6)

He said:
When attempting to load the es1968 driver, built against kernel 2.4.18, 2.4.19 or 2.4.20, I get the following message from the ALSA code:

    kernel: ALSA ../../alsa-kernel/pci/ac97/ac97_codec.c:1529: AC'97 0:0 does not respond - RESET [REC_GAIN = 0x0]

The driver fails to load.

Same for rc5, works with beta12.

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=172967

#######################
Bug 126797: Rear channels no longer work on SB Live ! 512

(against beta10)

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=126797

Probably fixed?

#######################
Bug 123800: maestro3 breaks irda

ancient bug, submitter doesn't have the hardware anymore. The bug has a pointer to a patch, but it's a 404 now, so I couldn't check.

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=123800

########################
Bug 108572: PCI IRQ sharing with USB will crash ALSA (against beta 4)

Null pointer dereference when loading the cs45xx module. Another user confirmed this for rc6, but now with kernel panic.

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=108572

Would be nice to have a look at this one.

#########################

That's it, I think. Any feedback very welcome.

Jordi

Signature · 3 lines
Jordi Mallach Pérez  --  Debian developer     http://www.debian.org/
[email protected]     [email protected]     http://www.sindominio.net/
GnuPG public key information available at http://oskuro.net/~jordi/
#1 lore raw
Quoted 1 line · Jordi Mallach

As promised, here are the bugs we have on our bug tracking system.

Damn, forgot a nice one:

Bug #100041: /dev/music MIDI adds unwanted staccatos on first "long" note after seq_reset()

Have a look at http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=100041

It includes a testcase, which I guess it wouldn't help me much, given my 0-knowledge about music :)
The bug was reported against beta4, but it'd be nice to see if the bug still applies.

Jordi

Signature · 3 lines
Jordi Mallach Pérez  --  Debian developer     http://www.debian.org/
[email protected]     [email protected]     http://www.sindominio.net/
GnuPG public key information available at http://oskuro.net/~jordi/
#2 lore raw
In reply to Jordi Mallach

Hi Jordi,

thanks for the summary.

At Thu, 6 Mar 2003 19:17:13 +0100,

Quoted 10 lines · Jordi Mallach

########################
Bug 183315: Sequencer device opening fails with ENOMEM

After some time (several open/close cycles of /dev/sequencer?), the next attempt to use the device will cause any open attempt to return ENOMEM. The usage counts of snd-emu10k1-synth, snd-emux-synth and snd-emu10k1 become non-zero and are incremented for each open failure, rendering them (AFAIK) unremovable without a reboot.

Quoted 7 lines

To trigger this, I've found it sufficient to run some lxmusserv-using Doom port several times [1]; though I don't think that it matters which program is used, I don't currently know whether it matters if AWE emulation/support is used.

The problem appears to only affect OSS sequencer usage; pmidi and programs which use other OSS devices still work.

fixed on the latest rc.

Quoted 12 lines

########################
Bug 182840: hw_pointer warning with 845PE chipset

this system has an Asustek P4PE motherbooard with an Intel 845PE chipset. Here is the lspci output for the integrated audio controller:

00:1f.5 Multimedia audio controller: Intel Corp. 82801DB AC'97 Audio (rev 02)

I finally managed to get sound out of this motherboard with alsaplayer (not with the ALSA xmms output plugin yet).

Quoted 6 lines

So all is well, except that I saw this warning message in the dmesg output:

ALSA ../alsa-kernel/core/pcm_lib.c:187: Unexpected hw_pointer value (stream = 0, delta: -1168, max jitter = 8192): wrong interrupt acknowledge?

PS: the other soundcard (ens1371) is managed by OSS/Free.

likely a hardware problem. not a fatal bug.
the above appears only when the driver is compiled with the debug option.

Quoted 10 lines

#########################
Bug 180653: Bad audio quality on VT8233 with rc7

Quoting:
-----------------
The quality of the audio output has decreased very strongly from rc6 to rc7. With rc6, everything was working fine, but with rc7 there is always a bit of noise when sound is played (I had the same problem with older versions of alsa, but rc6 was working fine).

Quoted 5 lines

My sound card is an on-board VT8233 chip, and I'm running kernel 2.4.20. I didn't try rc7-1.
------------------

Could this be related to the bad detection code in rc7 for VIA8233?

rc7 had a bug on VIA8233A. fixed on the later versions.

Quoted 9 lines

#########################

Bug 177522: aplay segfaults on multiple files

Submitted against RC3, but I think it's still there:

-----------------
sorry fellas, when one mixes .au and .wav, or does more than one of the same file, e.g.

Quoted 6 lines
$ y=/usr/share/emacs/site-lisp/emacspeak/sounds/default-8k/yank-object.au
$ aplay $y $y $y $y
one wont hear all what is requested, and will even seg fault!

aplay: pcm.c:5692: snd_pcm_unlink_ptr: Assertion `0' failed.

this happend only on a certain chip.
should have been fixed already on rc7.

Quoted 6 lines

########################
Bug 174690: alsamixer silently fails with ALI5451 South Bridge card

(against beta12)

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=174690

already fixed on rc6 or rc7.

Quoted 9 lines

########################
Bug 172967: "AC'97 0:0 does not respond"

(against rc6)

He said:
When attempting to load the es1968 driver, built against kernel 2.4.18, 2.4.19 or 2.4.20, I get the following message from the ALSA code:

Quoted 7 lines
    kernel: ALSA ../../alsa-kernel/pci/ac97/ac97_codec.c:1529: AC'97 0:0 does not respond - RESET [REC_GAIN = 0x0]

The driver fails to load.

Same for rc5, works with beta12.

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=172967

this might be due to the detection of modem codec. fixed on the recent version (hopefully).

Quoted 8 lines

#######################
Bug 126797: Rear channels no longer work on SB Live ! 512

(against beta10)

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=126797

Probably fixed?

should work - needs another tests with the latest version.

Quoted 6 lines

#######################
Bug 123800: maestro3 breaks irda

ancient bug, submitter doesn't have the hardware anymore. The bug has a pointer to a patch, but it's a 404 now, so I couldn't check.

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=123800

already fixed for Dell Inspiron 4000 and 8000.
for other notebooks, we need the PCI subsystem vendor and device id numbers to add the quirks in the driver code.

Quoted 10 lines

########################
Bug 108572: PCI IRQ sharing with USB will crash ALSA (against beta 4)

Null pointer dereference when loading the cs45xx module. Another user confirmed this for rc6, but now with kernel panic.

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=108572

Would be nice to have a look at this one.

hmm, it's difficult to guess.
we need a trace from kernel oops.

ciao,

Takashi

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

#3 lore raw
Quoted 1 line · Takashi Iwai

thanks for the summary.

Thanks for the feedback. Very helpful :)

Quoted 7 lines
Quoted 7 lines

########################
Bug 108572: PCI IRQ sharing with USB will crash ALSA (against beta 4)

Null pointer dereference when loading the cs45xx module. Another user confirmed this for rc6, but now with kernel panic.

Quoted 6 lines
Quoted 3 lines

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=108572

Would be nice to have a look at this one.

hmm, it's difficult to guess.
we need a trace from kernel oops.

Will ping the submitter, and close it if we get no answer.

Thanks for the rest of hints!

Jordi

Signature · 3 lines
Jordi Mallach Pérez  --  Debian developer     http://www.debian.org/
[email protected]     [email protected]     http://www.sindominio.net/
GnuPG public key information available at http://oskuro.net/~jordi/