On Sun, Jan 18, 2009 at 3:25 PM, Tobias Powalowski wrote:

Hi guys, new kernel adresses the following things: - bump to latest version - fixed bugs: http://bugs.archlinux.org/task/12821 http://bugs.archlinux.org/task/12816 http://bugs.archlinux.org/task/12820

please signoff for both arches

greetings tpowa --

Systems booted and everything seems to be running fine. Signoff both arches. Eric

2009/1/19 David Rosenstrauch :

FYI - I'm seeing a problem with the 2.6.28-3 kernel. For some reason there's no sound coming out of either the speakers or the headphones on my laptop (Dell Precision M4400) after the upgrade.

Sound module is snd-hda-intel. I checked volume is up, mute is not on, yada yada, but no sound coming out. Problem goes away as soon as I downgrade the kernel.

Anyone else seeing this and/or have any idea what the problem might be?

Maybe there's a regression in your soundcard autodetection? Do you pass any "model=" parameter to snd-hda-intel in /etc/modprobe.conf? Anyway, there have been a couple of commits in that driver, if you want you can track them searching for "alsa: hda" in the changelog (http://kernel.org/pub/linux/kernel/v2.6/ChangeLog-2.6.28.1). Corrado

2009/1/19 David Rosenstrauch :

No, I don't pass any model parameter, partly because I didn't know about this parm until recently, and partly because I'm not sure what the correct value should be. ("head -n 1 /proc/asound/card0/codec#0" says "IDT 92HD71B7X", but I'm not sure where to go to look up the model number from that.)

hda-intel, like the infamous ac97, is a widely-used codec, and the kernel module supports a whole lot of different devices. I don't have a lot of time to check right now, but a quick grep on the sources shows your card's support is included in sound/pci/hda/patch_sigmatel.c. From a quick search, the list of possible options can be found here: http://free-electrons.com/kerneldoc/latest/sound/alsa/ALSA-Configuration.txt and your best bet could be dell-m44 (similar to 4400, but marked as Inspiron) or dell-m43 (near to 44 but marked as Precision). HTH, Corrado

bardo wrote:

2009/1/19 David Rosenstrauch :

...

No, I don't pass any model parameter, partly because I didn't know about this parm until recently, and partly because I'm not sure what the correct value should be. ("head -n 1 /proc/asound/card0/codec#0" says "IDT 92HD71B7X", but I'm not sure where to go to look up the model number from that.)

hda-intel, like the infamous ac97, is a widely-used codec, and the kernel module supports a whole lot of different devices. I don't have a lot of time to check right now, but a quick grep on the sources shows your card's support is included in sound/pci/hda/patch_sigmatel.c. From a quick search, the list of possible options can be found here: http://free-electrons.com/kerneldoc/latest/sound/alsa/ALSA-Configuration.txt and your best bet could be dell-m44 (similar to 4400, but marked as Inspiron) or dell-m43 (near to 44 but marked as Precision).

HTH, Corrado

Hmmm .... IDT 92HD71B7X actually sounds like it matches more closely to one of these: STAC92HD71B* ref Reference board dell-m4-1 Dell desktops dell-m4-2 Dell desktops dell-m4-3 Dell desktops I'll check them out - thanks. DR

David Rosenstrauch wrote:

bardo wrote:

...

hda-intel, like the infamous ac97, is a widely-used codec, and the kernel module supports a whole lot of different devices. I don't have a lot of time to check right now, but a quick grep on the sources shows your card's support is included in sound/pci/hda/patch_sigmatel.c. From a quick search, the list of possible options can be found here: http://free-electrons.com/kerneldoc/latest/sound/alsa/ALSA-Configuration.txt

and your best bet could be dell-m44 (similar to 4400, but marked as Inspiron) or dell-m43 (near to 44 but marked as Precision).

HTH, Corrado

Hmmm .... IDT 92HD71B7X actually sounds like it matches more closely to one of these:

STAC92HD71B* ref Reference board dell-m4-1 Dell desktops dell-m4-2 Dell desktops dell-m4-3 Dell desktops

I'll check them out - thanks.

DR

Hmmm ... I tried testing this again, loading the module with each of the dell-m4-* model parameters, in turn, but no joy. *Some*thing's broken here. But I'm not enough of a kernel hacker to be able to figure out what or how to fix it. There's both a kernel bug and an ALSA bug about this; I guess I'll follow up there. DR

bardo wrote:

2009/1/19 David Rosenstrauch :

...

FYI - I'm seeing a problem with the 2.6.28-3 kernel. For some reason there's no sound coming out of either the speakers or the headphones on my laptop (Dell Precision M4400) after the upgrade.

Sound module is snd-hda-intel. I checked volume is up, mute is not on, yada yada, but no sound coming out. Problem goes away as soon as I downgrade the kernel.

Anyone else seeing this and/or have any idea what the problem might be?

Fun thing: I already upgraded the kernel when I answered your mail, but I hadn't rebooted. Today I did, and same thing happened :) I'm using snd-hda-intel too, on a 13" macbook, latest gen. My mixer is much thinner than with 2.6.28 (5 entries vs ~10), and there's no way to get some sound. Downgrading fixed. Did you find out anything?

C.

Not really. In theory this should be a matter of doing like bardo said: find out which chipset your machine uses, and then find the correct "model" parameter to supply for that chipset (via this page: http://free-electrons.com/kerneldoc/latest/sound/alsa/ALSA-Configuration.txt). (Well, technically, in theory you shouldn't have to specify a model at all, and the module should just autodetect it.) But that doesn't appear to be working, as I tried all 4 models listed for my chipset. So I have to figure then that this is a kernel or ALSA bug, and so monitor the upstream bug(s) to see when it'll be fixed. These 2 bug entries appear to be the relevant ones: http://bugzilla.kernel.org/show_bug.cgi?id=12336 https://bugtrack.alsa-project.org/alsa-bug/view.php?id=4306 So might be a while before this one gets fixed. HTH, DR

bardo wrote:

2009/1/20 David Rosenstrauch :

...

Not really. In theory this should be a matter of doing like bardo said: find out which chipset your machine uses, and then find the correct "model" parameter to supply for that chipset (via this page: http://free-electrons.com/kerneldoc/latest/sound/alsa/ALSA-Configuration.txt). (Well, technically, in theory you shouldn't have to specify a model at all, and the module should just autodetect it.) But that doesn't appear to be working, as I tried all 4 models listed for my chipset.

Did you reboot before every try? Best way is either blacklisting the module and loading manually or changing the model parameter in modprobe.conf before rebooting. This is because the BIOS needs to be reset too. It is also mentioned in the kernel bug you linked.

Egads! I have to reboot too?!?!? Sigh. Alright, I'll try that too when I get a moment.

...

So I have to figure then that this is a kernel or ALSA bug, and so monitor the upstream bug(s) to see when it'll be fixed. These 2 bug entries appear to be the relevant ones:

http://bugzilla.kernel.org/show_bug.cgi?id=12336 https://bugtrack.alsa-project.org/alsa-bug/view.php?id=4306

If we both have the same problem, I'm afraid these bug reports aren't related: I have sound working nicely with 2.6.28, problems started to appear with 2.6.28.1. Did you go straight from 2.6.27.10 to 2.6.28.1? If so, try testing 2.6.28.

C.

I haven't tried 2.6.28.1. The module was working in 2.6.27.10, and stopped working in the 2.6.28(-3) kernel package in the core repository. Thanks, DR

Dale Blount wrote:

...

I haven't tried 2.6.28.1. The module was working in 2.6.27.10, and stopped working in the 2.6.28(-3) kernel package in the core repository.

Try blacklisting the PC speaker module and rebooting. This "fixed" it for me.

Dale

I'm doing this in my rc.conf: MODULES=(snd-pcm-oss snd-hda-intel !snd-pcsp !pcspkr tun fuse loop) Is that not sufficient? DR

Dale Blount wrote:

On Tue, 2009-01-20 at 14:53 -0500, David Rosenstrauch wrote:

...

Dale Blount wrote:

...

...

I haven't tried 2.6.28.1. The module was working in 2.6.27.10, and stopped working in the 2.6.28(-3) kernel package in the core repository.

Try blacklisting the PC speaker module and rebooting. This "fixed" it for me.

Dale I'm doing this in my rc.conf:

MODULES=(snd-pcm-oss snd-hda-intel !snd-pcsp !pcspkr tun fuse loop)

Is that not sufficient?

Did you reboot afterward? I could not unload mine as it was "in use".

Yep. I'll have to try bardo's suggestion of rebooting after trying each of the models too. Thanks, DR

2009/1/20 Attila :

I suggest to do this

MODULES=(!pcspkr tun fuse loop)

but from my view the most problem with this "very usefull" -)) pcspkr is that he wants to grab the first slot.

You can avoid this with a file and mine looks so:

# cat /etc/modprobe.d/sound options snd-hda-intel model=6stack-dig power_save=60 enable=1 index=0 options snd-pcsp enable=1 index=2 alias snd-card-0 snd-hda-intel alias sound-slot-0 snd-hda-intel

For me this works as the best and perhaps for you too.

pcspkr and snd-pcsp are two different modules. The former provides you with the usual "beep" sounds, the latter tries to use the speaker as a sound card when your sound card is not supported or you don't have one at all. As far as I know, only snd-pcsp, sharing the alsa backend, can conflict with other modules. So there should be no need to blacklist pcspkr if it is wanted. It's ok to blacklist snd-pcsp or use the index=2 option, though. Corrado

On Dienstag, 20. Januar 2009 23:27 bardo wrote:

pcspkr and snd-pcsp are two different modules. The former provides you with the usual "beep" sounds, the latter tries to use the speaker as a sound card when your sound card is not supported or you don't have one at all. As far as I know, only snd-pcsp, sharing the alsa backend, can conflict with other modules. So there should be no need to blacklist pcspkr if it is wanted. It's ok to blacklist snd-pcsp or use the index=2 option, though.

Yes you be right and your explanation is better than mine. I blacklist pcspkr only because i don't need it and i want to to suppress the messages in the logfile. See you, Attila

2009/1/21 David Rosenstrauch :

I tried everything that everyone suggested in this thread, but no dice. Definitely a bug in 2.6.28. :-(

I managed to get a partial success (most of my mixer got back, but sound still doesn't come out, and I think it's because of a missing switch) by using the latest alsa-1.0.19. This, be warned, means recompiling the kernel with some PKGBUILD modifications, nothing really difficult, but takes some time. Instructions are here: http://alsa.opensrc.org/index.php/AlsaBuild2.6#Overwrite_alsa_in_the_kernel_... Corrado

bardo schrieb:

2009/1/21 David Rosenstrauch :

...

I tried everything that everyone suggested in this thread, but no dice. Definitely a bug in 2.6.28. :-(

I managed to get a partial success (most of my mixer got back, but sound still doesn't come out, and I think it's because of a missing switch) by using the latest alsa-1.0.19. This, be warned, means recompiling the kernel with some PKGBUILD modifications

It doesn't: You can download the standalone alsa-driver package, compile it and install the modules to /lib/modules/2.6.28-ARCH/updates (I think you can just replace /kernel/ with /updates/ in a Makefile, you'll have to figure out the details). The module in /updates/ will override the ones delivered with our kernel package, but you won't have file conflicts.