ALC880 + Intel 915GM - ICH6 results in Kernel panic (snd-hda-intel)

[mp3cube]

I'm using the latest preview release of Ubuntu breezy with kernel 2.6.12-8-386
(default kernel)

Installation was succesfull, but during the first boot, got the following problem :
Stacktrace :

PCI: Setting latency timer of device 0000:00:1b.0 to 64
hda_codec: Unknown model for ALC880, trying auto-probe from BIOS...
Unable to handle kernel NULL pointer dereference at virtual address 00000000
printing eip:
f8f421d4
*pde = 00000000
Oops: 0002 [#1]
PREEMPT
last sysfs file:
Modules linked in: snd_hda_intel snd_hda_codec snd_pcm snd_timer snd soundcore
snd_page_alloc
CPU: 0
EIP: 0060:[<f8f421d4>] Not tainted VLI
EFLAGS: 00010293 (2.6.12-8-386)
eax: fffffffe ebx: f7bf6748 ecx: 00000000 edx: 00000000
esi: f7bf6600 edi: 00000000 ebp: 00000006 esp: f718fde4
ds: 007b es: 007b ss: 0068
Process modprobe (pid: 6993, threadinfo=f718e000 task=f74d7a90)
Stack: 00000000 00000000 00000000 00000000 f7bf6600 f7bf6748 f7689000 f8f42713
f7bf6600 f7689000 f8f4298d f8f460d8 f8f45556 00000001 f8f3e77e 00000f00
00000005 00000000 f7689000 f753b240 00000000 f8f3e984 00000f00 00000002
Call Trace:
[<f8f42713>]
[<f8f4298d>]
[<f8f3e77e>]
[<f8f3e984>]
[<f8f3e592>]
[<c0254b3e>]
[<f8851871>]
[<f8851100>]
[<f8851220>]
[<f88524a3>]
[<c02619ca>]
[<c0261a06>]
[<c0293273>]
[<c02933a9>]
[<c029292a>]
[<c02933c6>]
[<c0293370>]
[<c0292e1c>]
[<c0293795>]
[<c0261c68>]
[<c0135be9>]
[<c0102c7b>]
Code: 31 c0 53 83 ec 10 89 d3 89 e7 f3 ab 8b 12 31 ff 83 fa 00 7e 45 89 f6 0f b7 44
7b 04 8d 48 ec 66 83 f9 03 77 13 8b 56 3c 83 e8 16 <66> 89 04 7a 8b 13 c7 04 8c 01 00
00 00 47 39 fa 7f da 31 ff 83

Usefull link on same problem : http://www.thisishull.net/showthread.php?
s=499dbf5ed874743ab2736780b9fd5732&t=110980&page=3&pp=10

Feel free to ask any further questions.
Many thanks in advance

[Frank Niedermann (fbn)]

I have the same issue on my system with onboard sound Intel High
Definition Audio / Realtek ALC880. Installation was successful but
during the first boot system stops at Starting hotplug subsystem.

If I disable the onboard sound in the BIOS booting is fine (but
without sound).

I don't know how to copy the boot messages in case of the error
because no log entries are written at this time, here is what I've
written off:

Oops: 0002 [#1]
Modules linked in: snd_hda_intel snd_hda_codec snd_pcm_oss ...
 ... some kernel messages here ...
Process modprobe (pid: 4948, threadinfo=f7fac000 task=df9fe520)
Stack: ....
 ... more kernel messages here ...
 <3>hw_random: RNG not detected

Tested with the Ubuntu Hoary 5.10 installation CD.

[Gouchi (gouchii)]

Same here using LivecD Breezy Colony 4, Ubuntu 5.10 Preview LiveCD and upgrade
to Breezy.

Here is the log :

[4294909.799000]
[4294909.799000] oops : 0002[#1]
Modules linked in : snd_hda_intel snd_hda_codec snd_pcm_oss snd_mixer_oss
snd_pcm snd_timer snd soundcore snd_page_alloc shpchp pci_hotplug intel_agp
agpgart psmouse lp mod ipv6 dm_snapshot dm_mod loop cloop af_packet ipw2200
firmware_class ieee80211_crypt skge nsl_cp437 pcmcia sr_mod sbp2 isofs ide_cd
cdrom ide_disk ide_floppy ide_generic pdc202xx_new aec62xx alim15x3 amd74 atiixp
cmd64x cs5520 cs5530 cy82c693 generic hpt34x ns87415 opti621 pdc202xx old rz1000
sc1200 serverworks siimage sis5513 slc90e66 triflex trù290 via82cxxx usb_storage
scsi_mod parport_pc parport mousedev fbcon tileblit font bitblit vga16fb
vgastate vesafb cfbcopyarea cfbimgblt cfbfillrect softcursor usbserial usbkbd
thermal processor fan usbhid ohci1394 ieee1394 yenta_socket rsrc_nonstatic
pcmcia_core piix ide_core ehci_hcd uhci_hcd usbcore evdev unix

[4294909.799000] CPU : [w]0
[4294909.799000] EIP : [<e0e545b1>] Not tainted VLI
[4294909.799000] EFLAGS : 00010246 (2.6.12-8-386)
[4294909.799000] EIP is at alc880_auto_fill_dac_nids+0x39/0x92 [snd_hda_codec]
[4294909.799000] eax : 00000000 ebx : deb67748 ecx : 00000000 edx : fffffffe
[4294909.799000] esi : deb67600 edi : 00000000 dbp : dd501000 esp : de937e1c
[4294909.799000] ds : 007b es 007b ss : 0068
[4294909.799000] Process modprobe (pid : 21982, threadinfo=de936000 task=dde22060)
[4294909.799000] stack : 00000000 00000000 00000000 00000000 deb67600 deb67748
00000006 e0e549f0 [4294909.799000] : deb67600 deb67748 dev67600 dd501000
e0e54b49 dd501000 dd050100
[4294909.799000] : d3586480 00000000 e0e517a1 dd501000 00001007 d358b40
e0e5cb04 d358b480
[4294909.799000] : Call Trace :
[4294909.799000] [<e0e549f0>] alc_880_parse_auto_config+0x27/0xda [snd_hda_codec]
[4294909.799000] [<e0e54b49>] patch_alc880+0x82/0x1e7 [snd_hda_codec]
[4294909.799000] [<e0e517a1>] snd_hda_codec_new+0x13f/0x18f [snd_hda_codec]
[4294909.799000] [<e0e7a698>] azx_codec_create+0x73/0xa1 [snd_hda_intel]
[4294909.799000] [<e0e7a0f2>] azx_send_cmd+0x0/0x60 [snd_hda_intel]
[4294909.799000] [<e0e7a1d5>] azx_get_response+0x0/0x5b [snd_hda_intel]
[4294909.799000] [<e0e7b0f7>] azx_probe+0xdc/0x150 [snd_hda_intel]
[4294909.799000] [<c01a2cb5>] pci_device_probe_static+0x2e/0x41
[4294909.799000] [<c01a2ce7>] __pci_device_probe+0x1f/0x32
[4294909.799000] [<c01a2d16>] pci_device_probe+0x1c/0x31
[4294909.799000] [<c01ebffa>] driver_probe_device+0x36/0x54
[4294909.799000] [<c01ec0d2>] driver_attach+0x39/0x6a
[4294909.799000] [<c01ec458>] bus_add_driver+0x5e/0x80
[4294909.799000] [<c01ec7b9>] driver_register+0x23/0x25
[4294909.799000] [<c01a2ed9>] pci_register_driver+0x5f/0x72
[4294909.799000] [<e0e0700a>] alsa_card_azx_init+0xa/0xc [snd_hda_intel]
[4294909.799000] [<c012967a>] sys_init_module+0xb5/0x172
[4294909.799000] [<c0102e9f>] sysenter_past_esp+0x54/0x75
[4294909.799000] Code : 74 24 20 8b 5c 24 24 89 e7 f3 ab 31 ff 3b 3b 7d
28 0f b7 54 7b 04 8d 42 ec 66 83 f8 03 77 17 0f b7 c2 8b 4e 3c 83 e8 14 8d 50 fe
<66> 89 14 79 c7 04 84 01 00 00 00 47 eb d4 31 ff 8b 13 39 d7 7d...

[Frank Niedermann (fbn)]

Can somebody give a statement about this bug, please?

Will it be fixed in Breezy final?

Is there a way to bypass the issue until it's fixed?

[Frank Niedermann (fbn)]

another user has this problem:
http://ubuntuforums.org/showthread.php?p=348721#post348721

[Daniel T Chen (crimsun)]

This issue has been resolved in alsa-driver 1.0.10rc1.

Because the kernel is essentially frozen, we need to test the following ASAP. If
alsa-driver 1.0.10rc1 works for you, then I'll generate a patch against our kernel.

*** NOTE: You will end up clobbering the Ubuntu /lib/modules/$(uname -r) if you
choose to test! Be sure you keep a copy of an older working kernel! ***

Here's what you need to do:

(1) Install build-essential, gcc-3.4-base, and linux-headers-$(uname -r)
(2) Download ftp://ftp.alsa-project.org/pub/driver/alsa-driver-1.0.10rc1.tar.bz2
(3) Modify alsa-kernel/pci/ac97/ac97_bus.c according to
https://bugtrack.alsa-project.org/alsa-bug/view.php?id=1356
(4) Configure alsa-driver with: CPP=cpp-3.4 CC=gcc-3.4 ./configure
--with-oss=yes --with-sequencer=yes --with-isapnp=no --with-pcmcia=kernel
--with-kernel=/lib/modules/$(uname -r)/build --with-cards=hda-intel
(5) Compile alsa-driver with: make && sudo make install

You'll need to reboot, and after the system has come up, make sure the ALSA
version is 1.0.10rc1:
$ cat /proc/asound/version

I'd like to avoid pulling in additional changes from cvs for
alsa-kernel/pci/hda, but we may have no choice.

[Frank Niedermann (fbn)]

Daniel,

gcc-3.4 package is also needed - just in case somebody else also want's to try it :)

I've done the steps you suggested, system is booting up but there are some
errors while starting alsa:
 Loading alsa ...
 amixer: hw:0 Invalid argument
 (message repeats about * 20)
and in /var/log/dmesg:
 [4294700.731000] hda_codec: Unknown model for ALC880, trying auto-probe from
BIOS...

If I try to run alsamixer or alsamixergui:
alsamixer: function snd_mixer_load failed: Invalid argument

Without alsamixer I don't know how to unmute the soundcard, because of that I
had no way to test if sound is working ...

fbn@paul:~$ cat /proc/asound/version
Advanced Linux Sound Architecture Driver Version 1.0.10rc1.
Compiled on Sep 13 2005 for kernel 2.6.12-8-686.

The card worked fine with Hoary, what changed that it's having these issues with
new alsa version?

[Daniel T Chen (crimsun)]

*** Bug 21564 has been marked as a duplicate of this bug. ***

[Daniel T Chen (crimsun)]

(In reply to comment #6)
> gcc-3.4 package is also needed - just in case somebody else also want's to try
it :)

Thanks, been a while since I brushed up on the deps.

> I've done the steps you suggested, system is booting up but there are some
> errors while starting alsa:
> Loading alsa ...
> amixer: hw:0 Invalid argument
> (message repeats about * 20)
> and in /var/log/dmesg:
> [4294700.731000] hda_codec: Unknown model for ALC880, trying auto-probe from
> BIOS...

Please attach full output of dmesg.

> If I try to run alsamixer or alsamixergui:
> alsamixer: function snd_mixer_load failed: Invalid argument
>
> Without alsamixer I don't know how to unmute the soundcard, because of that I
> had no way to test if sound is working ...

amixer can be used, e.g., amixer sset Master on (which would unmute Master).

> The card worked fine with Hoary, what changed that it's having these issues with
> new alsa version?

Short answer: "too many changes." Long answer: various initialization and codec
changes.

You may need a checkout of the alsa-driver and alsa-kernel cvs modules. Let's
defer that until we rule out the BIOS issue...

$ cvs -d:pserver:<email address hidden>:/cvsroot/alsa login
$ cvs -z3 -d:pserver:<email address hidden>:/cvsroot/alsa co alsa-kernel
alsa-driver

[Frank Niedermann (fbn)]

Created an attachment (id=3788)
/var/log/dmesg with alsa 1.0.10rc1.

[Frank Niedermann (fbn)]

amixer sset Master on
amixer: Mixer default load error: Invalid argument

I've some difficulties compiling alsa-driver from cvs:
./configure: line 9231: ALSA_TOPLEVEL_SELECT: command not found
./configure: line 9241: ALSA_TOPLEVEL_OUTPUT: command not found
configure: creating ./config.status
config.status: creating version
config.status: creating Makefile.conf
config.status: creating snddevices
config.status: creating utils/alsa-driver.spec
config.status: creating utils/buildrpm
config.status: error: cannot find input file: toplevel.config.in
root@paul:/root/alsa-driver# find . | grep -i toplevel
root@paul:/root/alsa-driver#

About the alsa patch (alsa bug 8112):
> -#include <linux/module.h>
> +#include <sound/driver.h>
This was not fixed as I did a cvs checkout, is it right to replace the first
line with the second line?

Could you send me a binary version of the module that you've compiled on your or
another system? I'm running 2.6.12-8-686

dmesg attached

[Daniel T Chen (crimsun)]

(In reply to comment #10)
> I've some difficulties compiling alsa-driver from cvs:

All right, I've done a cvs checkout of alsa-kernel and alsa-driver as of ~25
minutes ago and rolled it into a deb compiled on current Breezy against
linux-headers-2.6.12-8-686 that can be found here:
http://sh.nu/~crimsun/alsa-modules-2.6.12-8-686_1.0.9b-3ubuntu1~2_i386.deb
(md5sum eab278b69d4038b6c2274e08140004c6)

In order to cleanly install the above deb, you'll need to "unclobber" your
/lib/modules/$(uname -r). To do that, you need to download linux-image-$(uname
-r) and linux-restricted-modules-$(uname -r), then remove the entire
/lib/modules/$(uname -r) subdirectory. Finally:

$ sudo apt-get --reinstall install linux-image-$(uname -r)
linux-restricted-modules-$(uname -r)
$ sudo dpkg -i /path/to/alsa-modules-2.6.12-8-686_1.0.9b-3ubuntu1~2_i386.deb

Please reboot afterward, and attach the output from dmesg.

[Frank Niedermann (fbn)]

(In reply to comment #11)

> In order to cleanly install the above deb, you'll need to "unclobber" your
> /lib/modules/$(uname -r). To do that, you need to download linux-image-$(uname
> -r) and linux-restricted-modules-$(uname -r),

What do you mean by downloading?

> then remove the entire
> /lib/modules/$(uname -r) subdirectory. Finally:

Thats not possible:
rm -rf /lib/modules/2.6.12-8-686
rm: cannot remove directory `/lib/modules/2.6.12-8-686/volatile': Device or
resource busy

I restored my backup of that directory (cp /lib/modules/2.6.12-8-686-ORIG
/lib/modules/2.6.12-8-686), no error occured.

> $ sudo apt-get --reinstall install linux-image-$(uname -r)
> linux-restricted-modules-$(uname -r)
> $ sudo dpkg -i /path/to/alsa-modules-2.6.12-8-686_1.0.9b-3ubuntu1~2_i386.deb

dpkg -i alsa-modules-2.6.12-8-686_1.0.9b-3ubuntu1~2_i386.deb
(Reading database ... 91650 files and directories currently installed.)
Unpacking alsa-modules-2.6.12-8-686 (from
alsa-modules-2.6.12-8-686_1.0.9b-3ubuntu1~2_i386.deb) ...
dpkg: error processing alsa-modules-2.6.12-8-686_1.0.9b-3ubuntu1~2_i386.deb
(--install):
 trying to overwrite `/lib/modules/2.6.12-8-686/modules.pcimap', which is also
in package linux-image-2.6.12-8-686
Errors were encountered while processing:
 alsa-modules-2.6.12-8-686_1.0.9b-3ubuntu1~2_i386.deb

ls -l /lib/modules/2.6.12-8-686/modules.pcimap
ls: /lib/modules/2.6.12-8-686/modules.pcimap: No such file or directory

[Daniel T Chen (crimsun)]

(In reply to comment #12)
> What do you mean by downloading?

Since you've already restored the original /lib/modules/2.6.12-8-686, this
instruction is moot.

> dpkg -i alsa-modules-2.6.12-8-686_1.0.9b-3ubuntu1~2_i386.deb
> (Reading database ... 91650 files and directories currently installed.)
> Unpacking alsa-modules-2.6.12-8-686 (from
> alsa-modules-2.6.12-8-686_1.0.9b-3ubuntu1~2_i386.deb) ...
> dpkg: error processing alsa-modules-2.6.12-8-686_1.0.9b-3ubuntu1~2_i386.deb
> (--install):
> trying to overwrite `/lib/modules/2.6.12-8-686/modules.pcimap', which is also
> in package linux-image-2.6.12-8-686
> Errors were encountered while processing:
> alsa-modules-2.6.12-8-686_1.0.9b-3ubuntu1~2_i386.deb

$ dpkg -i --force-overwrite alsa-modules-2.6.12-8-686_1.0.9b-3ubuntu1~2_i386.deb

[Frank Niedermann (fbn)]

Created an attachment (id=3808)
dmesg with alsa-modules-2.6.12-8-686_1.0.9b-3ubuntu1~2_i386.deb

This one was a bit tricky as the system fist did not get up anymore, I hope I
have the right version of the modules now :)

dpkg -i --force-overwrite alsa-modules-2.6.12-8-686_1.0.9b-3ubuntu1 ~2_i386.deb

(Lese Datenbank ... 91650 Dateien und Verzeichnisse sind derzeit installiert.)
Entpacke alsa-modules-2.6.12-8-686 (aus
alsa-modules-2.6.12-8-686_1.0.9b-3ubuntu 1~2_i386.deb) ...
dpkg - Warnung: Problem wird übergangen, weil --force angegeben ist:
 versuche »/lib/modules/2.6.12-8-686/modules.pcimap« zu überschreiben,
welches a uch in Paket linux-image-2.6.12-8-686 ist
dpkg - Warnung: Problem wird übergangen, weil --force angegeben ist:
 versuche »/lib/modules/2.6.12-8-686/modules.dep« zu überschreiben, welches
auch in Paket linux-image-2.6.12-8-686 ist
dpkg - Warnung: Problem wird übergangen, weil --force angegeben ist:
 versuche »/lib/modules/2.6.12-8-686/modules.ieee1394map« zu überschreiben,
welc hes auch in Paket linux-image-2.6.12-8-686 ist
dpkg - Warnung: Problem wird übergangen, weil --force angegeben ist:
 versuche »/lib/modules/2.6.12-8-686/modules.usbmap« zu überschreiben,
welches a uch in Paket linux-image-2.6.12-8-686 ist
dpkg - Warnung: Problem wird übergangen, weil --force angegeben ist:
 versuche »/lib/modules/2.6.12-8-686/modules.ccwmap« zu überschreiben,
welches a uch in Paket linux-image-2.6.12-8-686 ist
dpkg - Warnung: Problem wird übergangen, weil --force angegeben ist:
 versuche »/lib/modules/2.6.12-8-686/modules.isapnpmap« zu überschreiben,
welche s auch in Paket linux-image-2.6.12-8-686 ist
dpkg - Warnung: Problem wird übergangen, weil --force angegeben ist:
 versuche »/lib/modules/2.6.12-8-686/modules.inputmap« zu überschreiben,
welches auch in Paket linux-image-2.6.12-8-686 ist
dpkg - Warnung: Problem wird übergangen, weil --force angegeben ist:
 versuche »/lib/modules/2.6.12-8-686/modules.seriomap« zu überschreiben,
welches auch in Paket linux-image-2.6.12-8-686 ist
dpkg - Warnung: Problem wird übergangen, weil --force angegeben ist:
 versuche »/lib/modules/2.6.12-8-686/modules.alias« zu überschreiben,
welches au ch in Paket linux-image-2.6.12-8-686 ist
dpkg - Warnung: Problem wird übergangen, weil --force angegeben ist:
 versuche »/lib/modules/2.6.12-8-686/modules.symbols« zu überschreiben,
welches auch in Paket linux-image-2.6.12-8-686 ist
Richte alsa-modules-2.6.12-8-686 ein (1.0.9b-3ubuntu1~2) ...
Terminating processes: 7229 7320 4607 7209 (failed: processes still using sound
devices: 4625(gnome-settings-) 4652(esd)).
/etc/init.d/alsa: Warning: Processes using sound devices: 4625(gnome-settings-)
4652(esd) 4665(esd).
Unloading ALSA sound driver modules: snd-hda-intel snd-hda-codec snd-pcm-oss
snd -mixer-oss snd-pcm snd-timer snd-page-alloc (failed: modules still loaded:
snd-h da-intel snd-hda-codec snd-mixer-oss snd-pcm snd-timer snd-page-alloc).
Loading ALSA sound driver modules: snd-hda-intel snd-hda-codec snd-pcm-oss
snd-m ixer-oss snd-pcm snd-timer snd-page-allocFATAL: Error inserting
snd_pcm_oss (/li b/modules/2.6.12-8-686/updates/alsa/acore/oss/snd-pcm-oss.ko):
Unknown symbol in module, or unknown parameter ...

[Daniel T Chen (crimsun)]

(In reply to comment #14)
> amixer sset Master on
> amixer: Mixer default load error: Invalid argument

Ok, then you will need to manually try every model parameter until you find one
that works. Please unload snd-hda-intel after boot and try each of the following
models (see alsa-kernel/pci/hda/patch_realtek.c#alc880_cfg_tbl[] for a full
listing). Please try the ones that most closely match your hardware
configuration first.

$ sudo modprobe snd-hda-intel model=foo

where foo is one of:

3stack (/* Back 3 jack, front 2 jack */)
3stack-digout (/* Back 3 jack plus 1 SPDIF out jack, front 2 jack */)
5stack (/* Back 5 jack, front 2 jack */)
5stack-digout (/* Back 5 jack plus 1 SPDIF out jack, front 2 jack */)
w810
z71v
6stack
6stack-digout
asus
uniwill
F1734

[Frank Niedermann (fbn)]

I had to add snd_hda_intel to /etc/hotplug/blacklist because I was unable to
rmmod it. After rebooting the system no snd* module was loaded. I tried every
model on your list and "full" and "test". Every time I hear a "bump" on my
speakers, this is what gets logged into /var/log/syslog and dmesg:

Sep 15 16:33:50 paul kernel: [4295095.493000] ACPI: PCI Interrupt
0000:00:1b.0[A] -> GSI 16 (level, low) -> IRQ 16
Sep 15 16:33:50 paul kernel: [4295095.493000] PCI: Setting latency timer of
device 0000:00:1b.0 to 64
Sep 15 16:33:50 paul kernel: [4295095.511000] hda_codec: Unknown model for
ALC880, trying auto-probe from BIOS...
Sep 15 16:33:50 paul kernel: [4295095.826000] hda_codec: num_steps = 0 for NID=0x8
Sep 15 16:33:50 paul kernel: [4295095.854000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:33:50 paul kernel: [4295095.857000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:33:50 paul kernel: [4295095.862000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:33:50 paul kernel: [4295095.866000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:33:50 paul kernel: [4295095.871000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:33:50 paul kernel: [4295095.875000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:33:50 paul kernel: [4295095.879000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:33:50 paul kernel: [4295095.884000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:33:50 paul kernel: [4295095.888000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:33:50 paul kernel: [4295095.893000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:33:50 paul kernel: [4295095.898000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:33:50 paul kernel: [4295095.903000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:33:50 paul kernel: [4295095.906000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:33:50 paul kernel: [4295095.911000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:33:50 paul kernel: [4295095.915000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:33:50 paul kernel: [4295095.919000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:33:50 paul kernel: [4295095.924000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:33:50 paul kernel: [4295095.928000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:33:50 paul kernel: [4295095.933000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:33:50 paul kernel: [4295095.937000] hda_codec: num_steps = 0 for NID=0xb
Sep 15 16:34:03 paul kernel: [4295108.362000] hda_codec: num_steps = 0 for NID=0xb

alsamixer still has the error message
alsamixer: function snd_mixer_load failed: Invalid argument

If I check the alsa version (/proc/asound/version) I see this:
Advanced Linux Sound Architecture Driver Version 1.0.10rc1.
Compiled on Sep 14 2005 for kernel 2.6.12-8-686.

Is version 1.0.10rc1 the right one because you're package is named
alsa-modules-2.6.12-8-686_1.0.9b-3ubuntu1~2_i386.deb and this is what I've done:
dpkg -i alsa-modules-2.6.12-8-686_1.0.9b-3ubuntu1~2_i386.deb
(Lese Datenbank ... 91864 Dateien und Verzeichnisse sind derzeit installiert.)
Vorbereiten zum Ersetzen von alsa-modules-2.6.12-8-686 1.0.9b-3ubuntu1~2 (durch
alsa-modules-2.6.12-8-686_1.0.9b-3ubuntu1~2_i386.deb) ...
Entpacke Ersatz für alsa-modules-2.6.12-8-686 ...

[Daniel T Chen (crimsun)]

Please attach the output of ''lspci -vv''.

[Frank Niedermann (fbn)]

Created an attachment (id=3838)
lspci -vv

[Daniel T Chen (crimsun)]

And now, please ''lspci -nvv''. Thanks.

[Frank Niedermann (fbn)]

Created an attachment (id=3848)
lspci-nvv

[Frank Niedermann (fbn)]

Daniel,

what do you want to figure out about the lspci output?

[Daniel T Chen (crimsun)]

(In reply to comment #21)
> Daniel,
>
> what do you want to figure out about the lspci output?

I needed the subvendor:subdevice identifiers and have confirmed that your
specific combination is not supported currently in ALSA. I'll push the report
upstream.

[Frank Niedermann (fbn)]

Today I've installed Realtek's ALC880 driver [1], it's named
realtek-linux-audiopack-3.4-0.tar.bz2 and it's based on
alsa-driver-1.0.9b_10.tar.bz2 - this is what I've done:

cd alsa-driver-1.0.9b_10
./configure
make
make install
./snddevices

modprobe snd_hda_intel

cat /proc/asound/version
Advanced Linux Sound Architecture Driver Version 1.0.9b.
Compiled on Sep 17 2005 for kernel 2.6.12-8-686.

amixer sset Master on
amixer: Unable to find simple control 'Master',0

amixer
Simple mixer control 'Headphone',0
  Capabilities: pswitch
  Playback channels: Front Left - Front Right
  Front Left: Playback [on]
  Front Right: Playback [on]
Simple mixer control 'Front',0
  Capabilities: pvolume pswitch
  Playback channels: Front Left - Front Right
  Limits: Playback 0 - 64
  Front Left: Playback 37 [58%] [on]
  Front Right: Playback 37 [58%] [on]
  ... more lines ...

Now I can start playing audio files, but the quality is very bad (much crackle).

[1]
http://www.realtek.com.tw/downloads/dlhd-2.aspx?lineid=2004052&famid=2004052&series=2004061&Software=True

[Frank Niedermann (fbn)]

Any updates here? There is not much going on on
https://bugtrack.alsa-project.org/alsa-bug/view.php?id=1429 ...

[Frank Niedermann (fbn)]

I've downloaded the new (2005/09/29) Realtek driver package
realtek-linux-audiopack-3.4-3.tar.bz2 from here:
http://www.realtek.com.tw/downloads/dlhd-2.aspx?lineid=2004052&famid=2004052&series=2004061&Software=True

Compiled and installed alsa-driver-1.0.9b_13.tar.bz2 as the Readme.txt described:

Step 3. Complied source code
 a. cd alsa-driver-1.0.9b_1.
 b. ./configure
 c. make
 d. make install
 e. ./snddevices

After that I did a reboot and modprobe snd_hda_intel. Gnome recognizes the audio
card twice (two popups), I'm able to play an audio file with mp3blaster as root.
Sound is really bad (crackles all the time).

[Daniel T Chen (crimsun)]

Created an attachment (id=4274)
dpatch (take 1) for ALC880 OOPS on boot

apt-get source linux-image-$(uname -r)

bunzip2 this patch and place it in the
linux-source-2.6.12-2.6.12/debian/patches/ subdirectory

Append the name of the patch (strip .dpatch) to 00list-9.19

Compile the linux-image deb from the top-level extracted directory with:
debuild binary

[japj (japj)]

I've also got this problem on a Medion MD 95500 Notebook (I tried the 5.10
Preview Release LiveCD of 2005/09/09 and it hung during boot with this kernel
panic).

Any idea if this will be fixed for the final release?

[Daniel T Chen (crimsun)]

(In reply to comment #27)
> I've also got this problem on a Medion MD 95500 Notebook (I tried the 5.10
> Preview Release LiveCD of 2005/09/09 and it hung during boot with this kernel
> panic).
>
> Any idea if this will be fixed for the final release?

Frank and I are testing a workaround. I've added a comment above (see #26 if you
wish to help test).

[japj (japj)]

If I read correctly the solution at #26 requires an already installed Ubuntu
environment to test with.

Can this 'test' also be performed with the LiveCD? (I have not installed Ubuntu
on this laptop yet, I only tried to boot the LiveCD).

[Frank Niedermann (fbn)]

(In reply to comment #29)
> Can this 'test' also be performed with the LiveCD?

I don't think so, the described steps will create a new kernel image which has
to be installed into the system. I don't see how this could be done with the
Live-CD. If you have some free disk space you could do a Breezy installation, 4G
should be enough.

[Daniel T Chen (crimsun)]

(In reply to comment #29)
> If I read correctly the solution at #26 requires an already installed Ubuntu
> environment to test with.
>
> Can this 'test' also be performed with the LiveCD? (I have not installed Ubuntu
> on this laptop yet, I only tried to boot the LiveCD).

You'd have to patch alsa-source from universe, then compile it, install it,
unload the ALSA drivers, then reload snd-hda-intel. So yes, it's certainly
feasible. In fact, this is pretty much what Frank already did in #25.

[Daniel T Chen (crimsun)]

Created an attachment (id=4329)
dpatch (take 2) for ALC880 OOPS on boot

Please test this patch. The same instructions apply.

You may find it more efficient to remove the unneeded configs/directories from
debian/config.

[Frank Niedermann (fbn)]

Daniel,

(In reply to comment #32)
> Created an attachment (id=4329) [edit]
> dpatch (take 2) for ALC880 OOPS on boot
> Please test this patch. The same instructions apply.

this time it works and I can compile the kernel but after some minutes
(compiling) I get the following:

...
dpkg-gencontrol -DArchitecture=i386 -isp \
                        -plinux-image-2.6.12-9-686-smp -Pdebian/tmp-image/
chmod -R og=rX debian/tmp-image
chown -R root:root debian/tmp-image
dpkg --build debian/tmp-image ..
dpkg-deb: building package `linux-image-2.6.12-9-686-smp' in
`../linux-image-2.6.12-9-686-smp_2.6.12-9.19_i386.deb '.
rm -f -r debian/tmp-image
echo done > stamp-image
make[2]: Leaving directory
`/home/fbn/linux-source-2.6.12-2.6.12/debian/build/install-686-smp'
make[1]: Leaving directory
`/home/fbn/linux-source-2.6.12-2.6.12/debian/build/install-686-smp'
dh_testdir
dh_testroot
# unpack the kernels into a temporary directory
mkdir -p debian/d-i-i386
# XXX: this stuff finds the kernels that need upacking according to
# kernel-versions and unpack them into the temp dir.
imagelist=$(cat kernel-versions | grep ^i386 | awk '{print $4}') && \
for i in $imagelist; do \
  dpkg -x $(ls debian/build/linux-image-$i\_*i386.deb) debian/d-i-i386; \
done
ls: debian/build/linux-image-2.6.12-9-386_*i386.deb: No such file or directory
dpkg-deb: --extract needs a target directory.
Perhaps you should be using dpkg --install ?
make: *** [binary-udebs] Error 2
debuild: fatal error at line 836:
couldn't exec debian/rules:
root@paul:~/linux-source-2.6.12-2.6.12#

Is this because the script tries to build i386 debs? (I've removed everything
except 686-smp)

[Daniel T Chen (crimsun)]

I was able to create an amd64-generic linux-image deb successfully, but I don't
think it will be much use to you.

Instead of erasing all the non-686* configs, erase all but the 386 generic one.
That should build successfully.

Btw, this dpatch works with 2.6.12-9.21.

[Daniel T Chen (crimsun)]

In case any of you guys have difficulty building a linux-image deb, I've place a
386 one based on 2.6.12-9.21 at
http://www.sh.nu/~crimsun/linux-image-2.6.12-9-386_2.6.12-9.21_i386.deb

The sha1sum is: 78b8702ea30d1d3d6e7dc04742812f397e823d2e

Please install and test.

[Raymond Sneekes (raymond-sneekes)]

(In reply to comment #35)
> In case any of you guys have difficulty building a linux-image deb, I've place a
> 386 one based on 2.6.12-9.21 at
> http://www.sh.nu/~crimsun/linux-image-2.6.12-9-386_2.6.12-9.21_i386.deb
>
> The sha1sum is: 78b8702ea30d1d3d6e7dc04742812f397e823d2e
>
> Please install and test.
This kernel works for me.

[Frank Niedermann (fbn)]

(In reply to comment #35)
> In case any of you guys have difficulty building a linux-image deb, I've place a
> 386 one based on 2.6.12-9.21 at
> http://www.sh.nu/~crimsun/linux-image-2.6.12-9-386_2.6.12-9.21_i386.deb
>
> The sha1sum is: 78b8702ea30d1d3d6e7dc04742812f397e823d2e
>
> Please install and test.

Does not work for me, I was booting the new kernel, loaded snd_hda_intel
manually (is in hotplug/blacklist) and got this in syslog:

Oct 5 22:38:59 localhost kernel: [4294910.669000] hda_codec: num_steps = 0 for
NID=0xb
Oct 5 22:39:08 localhost kernel: [4294919.877000] hda_codec: num_steps = 0 for
NID=0xb
Oct 5 22:39:43 localhost kernel: [4294954.920000] hda_codec: num_steps = 0 for
NID=0xb
Oct 5 22:39:48 localhost kernel: [4294959.766000] hda_codec: num_steps = 0 for
NID=0xb

Gnome recognizes soundcard but alsamixer not:
alsamixer: function snd_mixer_load failed: Invalid argument

Means no sound at all for me.

Is there a way to confirm running the new kernel? I was wondering why I still
have the nvidia module loaded on boot ... or have you packed this module into
the deb file?

[Daniel T Chen (crimsun)]

(In reply to comment #36)
> > Please install and test.
> This kernel works for me.

Do you mean the kernel boots and you can hear undistorted sound? Please clarify.

[Daniel T Chen (crimsun)]

(In reply to comment #37)
> Does not work for me, I was booting the new kernel, loaded snd_hda_intel
> manually (is in hotplug/blacklist) and got this in syslog:
>
> Oct 5 22:38:59 localhost kernel: [4294910.669000] hda_codec: num_steps = 0 for
> NID=0xb
[...]
> Gnome recognizes soundcard but alsamixer not:
> alsamixer: function snd_mixer_load failed: Invalid argument
>
> Means no sound at all for me.
>
> Is there a way to confirm running the new kernel? I was wondering why I still
> have the nvidia module loaded on boot ... or have you packed this module into
> the deb file?

$ cat /proc/version

The kernel that I compiled does not include any nvidia.ko. Are you sure you're
running the 2.6.12-2.21 image that I built?

[Matt Zimmerman (mdz)]

(In reply to comment #39)
> (In reply to comment #37)
> > Does not work for me, I was booting the new kernel, loaded snd_hda_intel
> > manually (is in hotplug/blacklist) and got this in syslog:
> >
> > Oct 5 22:38:59 localhost kernel: [4294910.669000] hda_codec: num_steps = 0 for
> > NID=0xb
> [...]
> > Gnome recognizes soundcard but alsamixer not:
> > alsamixer: function snd_mixer_load failed: Invalid argument
> >
> > Means no sound at all for me.
> >
> > Is there a way to confirm running the new kernel? I was wondering why I still
> > have the nvidia module loaded on boot ... or have you packed this module into
> > the deb file?
>
> $ cat /proc/version
>
> The kernel that I compiled does not include any nvidia.ko. Are you sure you're
> running the 2.6.12-2.21 image that I built?

There's no reason that installing this .deb should cause nvidia.ko (provided by
linux-restricted-modules) to go away...

[Raymond Sneekes (raymond-sneekes)]

(In reply to comment #38)
> (In reply to comment #36)
> > > Please install and test.
> > This kernel works for me.
>
> Do you mean the kernel boots and you can hear undistorted sound? Please clarify.

Yes.
Removed the snd-hda-intel from my hotplug blacklist and it booted without any
errors. Also sound is excellent.

[Frank Niedermann (fbn)]

(In reply to comment #39)
> (In reply to comment #37)
> > Does not work for me, I was booting the new kernel, loaded snd_hda_intel
> Are you sure you're running the 2.6.12-2.21 image that I built?

No I'm not - I'll try your deb with the brand new Breezy RC1 which I'm
downloading atm :)

[Frank Niedermann (fbn)]

Daniel,

your kernel does not work for me. Brand new Breezy RC1 installation (disabled
sound card in BIOS), installed your kernel (dpkg -i), reboot with sound card
enabled.

Hotplug loads and ALSA gets initialized. In Gnome I can unmute sound but I don't
hear any sound (tried the Gnome system events and two ogg files).

Tried to run amixer and alsamixer:
amixer: Mixalsamixer: function snd_mixer_load failed: Invalid argument
alsamixer: function snd_mixer_load failed: Invalid argument

cat /proc/version
Linux version 2.6.12-9-386 (root@garnish) (gcc version 3.4.5 20050809
(prerelease) (Ubuntu 3.4.4-6ubuntu8)) #1 Tue Oct 4 23:08:57 EDT 2005

Is this your kernel?

This is what I see in the logfiles (syslog, messages, dmesg):

[4294739.381000] hda_codec: num_steps = 0 for NID=0xb
[4294750.445000] hda_codec: num_steps = 0 for NID=0xb
[4294759.294000] hda_codec: num_steps = 0 for NID=0xb
[4295007.335000] hda_codec: num_steps = 0 for NID=0xb
[4295010.738000] hda_codec: num_steps = 0 for NID=0xb
[4295015.819000] hda_codec: num_steps = 0 for NID=0xb
[4295252.879000] hda_codec: num_steps = 0 for NID=0xb
[4295254.981000] hda_codec: num_steps = 0 for NID=0xb
[4295314.999000] hda_codec: num_steps = 0 for NID=0xb
[4295316.792000] hda_codec: num_steps = 0 for NID=0xb

It's a bit weired that I can unmute the card and even totem is not telling me
about not finding any audio device while playing ogg files - but I just don't
hear anything.

I really hope it's not a silly mistake I've done but I've checked everything in
my mind ...

[Daniel T Chen (crimsun)]

(In reply to comment #43)
> your kernel does not work for me. Brand new Breezy RC1 installation (disabled
> sound card in BIOS), installed your kernel (dpkg -i), reboot with sound card
> enabled.

Ok, this is what I expected. Please remember that the original objective was to
prevent a regression from Hoary - that is, the kernel should not OOPS on boot.
Given that support for the high definition audio chipsets was not included in
Hoary, and given that this patch is unavoidably intrusive, it has been decided
that this patch will not be included in Breezy's kernel.

For the record, please confirm that installing a fresh RC _without_ installing
my kernel results in a kernel panic.

[...]
> It's a bit weired that I can unmute the card and even totem is not telling me
> about not finding any audio device while playing ogg files - but I just don't
> hear anything.

Unfortunately there is currently no way around this except to use the upstream
Realtek patches against 1.0.9b. An attempt to integrate them into Breezy's
kernel results in an even larger, even more intrusive patch than the second
attempt above, so if the one that simply prevents an OOPS on boot (that results
in no sound for you but _does_ result in sound for #36) was rejected, there is
essentially no chance that a larger Realtek patch (that results in sound for
you, albeit horrid quality) will be considered.

[japj (japj)]

(In reply to comment #44)
> Ok, this is what I expected. Please remember that the original objective was to
> prevent a regression from Hoary - that is, the kernel should not OOPS on boot.

Just to confirm my understanding, is it still the objective that the kernel should
not OOPS on boot (for both the Breezy InstallCD & LiveCD)?

(If I have to choose between being able to boot or having sound, I choose boot :)

Or is this specific problem being moved to after the Breezy release is done? (and
thus having an OOPS for this specific situation in Breezy)?

[Daniel T Chen (crimsun)]

(In reply to comment #45)
> Just to confirm my understanding, is it still the objective that the kernel should
> not OOPS on boot (for both the Breezy InstallCD & LiveCD)?

That is the objective given the condition that the OOPS is not a regression from
Hoary. Because the drivers are not enabled in Hoary's configuration, this is not
a regression.

> (If I have to choose between being able to boot or having sound, I choose boot :)

As would I, but as you can see, the dpatch (take 2) is quite intrusive and has
been rejected - even if it prevents an OOPS.

> Or is this specific problem being moved to after the Breezy release is done? (and
> thus having an OOPS for this specific situation in Breezy)?

Yes, this specific problem will be resolved following Breezy's release. The OOPS
is specific to this situation.

[Daniel T Chen (crimsun)]

I've posted
http://www.sh.nu/~crimsun/linux-image-2.6.12-9-386_2.6.12-9.22_i386.deb which
contains the security fix from .22 as well as the dpatch (take 2) above.

sha1sum below:
a0d20a4bce5a63db976b8f1a4624e33e130ae0ce
linux-image-2.6.12-9-386_2.6.12-9.22_i386.deb

[Frank Niedermann (fbn)]

Daniel,

your kernel image seems to work for me too - at least sometimes. I did a fresh
RC1 installation, installed your deb, rebooted, soundcard available but no sound
and no alsamixer possible. After that I installed the default 686-smp kernel,
rebootet and wohoo I had soundcard available, sound and even alsamixer was working.

I first thought that the 686-smp kernel had a new snd-hda-intel module that
works but that's not the case because now it doesn't work anymore.

I don't remember doing anything else special on my installation but I remember I
did not install the Realtek drivers ;)

I can confirm that Breezy RC1 will not boot with default 386 and 686-smp kernels
if ALC880 is enabled.

[Matthias-volltext (matthias-volltext)]

(In reply to comment #47)
> I've posted
> http://www.sh.nu/~crimsun/linux-image-2.6.12-9-386_2.6.12-9.22_i386.deb which
> contains the security fix from .22 as well as the dpatch (take 2) above.
>
> sha1sum below:
> a0d20a4bce5a63db976b8f1a4624e33e130ae0ce
> linux-image-2.6.12-9-386_2.6.12-9.22_i386.deb

/i am submitter of duplicate bug 23064/

sound on my asus w6a works now with this kernel

i had to use
------>8---------------------------------
option snd-hda-intel model=asus
------>8---------------------------------
in /etc/modutils.d/alsa-base as found in
http://jeriko.l-tech.org/2005/09/04/intel-hda-unter-debian-sarge-installieren/
(german)

without that i was missing some controls in alsamixer (no
front/surround/center/lfe, only headphones/pcm/line/mic)

other kernels didnt not recognize the card for me

cheers m

[Florent CHANTRET (florent-chantret)]

I have tried the pre-packaged patched kernel and it works perfectly ! So nice.

For info, my computer is a Medion MD95257.

Will you add the patch for the Breezy updates ? Else, I will have to permanently
patch your updated kernel in order to keep both my Breezy secured and with sound

[Thomas Hood (jdthood)]

(In reply to comment #49)
> i had to use
> ------>8---------------------------------
> option snd-hda-intel model=asus
> ------>8---------------------------------
> in /etc/modutils.d/alsa-base

I suggest putting this line in a separate file such as /etc/modutils.d/snd-hda-intel
so that you can continue using an unmodified /etc/modutils.d/alsa-base and thus
continue to benefit from automatic updates of the latter file.

[Thomas Hood (jdthood)]

*** Bug 21688 has been marked as a duplicate of this bug. ***

[japj (japj)]

I could not find any information on this, but is it possible to disable
alsa/sound from loading on the live CD?
Without such a boot option, I will not be able to boot the LiveCD at all because
of the OOPS.

[Matt Zimmerman (mdz)]

(In reply to comment #53)
> I could not find any information on this, but is it possible to disable
> alsa/sound from loading on the live CD?
> Without such a boot option, I will not be able to boot the LiveCD at all because
> of the OOPS.

There is no convenient way to do this, no. You'd have to disable it from the
command line (alt+f2) during the initialization process

[Florent CHANTRET (florent-chantret)]

As I said before, the dpatch works perfectly for me but I have noticed a problem
related to it. Now, there is no freeze and sound works perfectly but when I
shutdown the computer, there is a message at the end of the shutdown process like :

[xxxxx.yyyyy] Power down.

Where x and y are some dynamic numbers (I think this is not always the same each
time I try to shutdown the computer).

But the computer stayed powered-on. If I add again the snd-hda-intel to the
hotplug blacklist file, the computer shut off correctly so I think this is
related to this dpatch.

[japj (japj)]

Hello,

Has there been any progress on solving this problem?
If not yet, is there any information on whether this bug is going to be solved
in Drake?
(I can help test a live cd if needed)

[Matt Zimmerman (mdz)]

*** Bug 24392 has been marked as a duplicate of this bug. ***

[japj (japj)]

I can confirm this bug to still exist in the Dapper Flight-1 Live i386 CD.

[Florent CHANTRET (florent-chantret)]

It's really important this bug is fixed in Dapper Drake. It should even have
been corrected for Breezy Badger as it's a gravous bug, especially for a lot of
Centrino laptops owner and as the dpatch seems to work perfectly. As I said
before, the only stuff with the dpatch is the computer isn't powered off
correctly at the end of the shutdown sequence.

[Daniel T Chen (crimsun)]

Please test upstream cvs (post-1.0.10) alsa-driver.

https://bugtrack.alsa-project.org/alsa-bug/view.php?id=1429

[Ben Collins (ben-collins)]

This is fixed in dapper.

[Ben Collins (ben-collins)]

This can not easily be fixed in breezy.

[Frank Niedermann (fbn)]

Did somebody already confirm that it is fixed in Dapper?

[Daniel T Chen (crimsun)]

It's fixed in Dapper.

[Andrew Ash (ash211)]

With no activity on this bug in almost a year and Breezy's life coming to a close, I'm going to assume that it's safe to close this bug. If it's still a problem for anyone, feel free to reopen it.