[gentoo-amd64] problem with multi-channel alsa on hda-intel

[gentoo-amd64] problem with multi-channel alsa on hda-intel

[Raffaele BELARDI]

Until last world update I was happily using the following .asoundrc [1]
to duplicate stereo sound from front to rear channels (the former is
connected to a 2.1 system, the latter to a hi-fi equipment):

pcm.ch40dup {
 type route
 slave.pcm surround40
 slave.channels 4
 ttable.0.0 1
 ttable.1.1 1
 ttable.0.2 1
 ttable.1.3 1
 }

and then "mplayer -channels 4 -ao alsa:device=ch40dup" gave me the
duplicated stereo sound on 4 channels.

After last update alsa-lib went from 1.0.19 to 1.0.20 and kernel 2.6.29
to 2.6.30 and mplayer sound stopped working from any channel with the
following error:

Trying preferred audio driver 'alsa', options 'device=ch40dup'
alsa-init: requested format: 48000 Hz, 2 channels, 9
alsa-init: using ALSA 1.0.20
alsa-init: setup for 1/2 channel(s)
alsa-init: using device ch40dup
alsa-init: pcm opened in blocking mode
[AO_ALSA] Unable to get initial parameters: Invalid argument
Failed to initialize audio driver 'alsa:device=ch40dup'
Could not open/initialize audio device -> no sound.

If I remove the "-ao alsa:device..." bit from the mplayer command line I
get stereo sound from the front speakers only.

This is with HDA-intel driver on Nvidia nForce430/AD1986 chipset, icewm
desktop. I just read from the kernel doc [2] that there is a "model
name" parameter that defaults to '6stack' while my model is '3stack' so
tonight I will try that.

Any other hints before I try upstream?

thanks,

raf

[1]
http://alsa.opensrc.org/index.php/Playing_stereo_on_surround_sound_setup_%28Howto%29

[2] /usr/src/linux/Documentation/sound/alsa/HD-Audio-Models.txt

[gentoo-amd64] Re: problem with multi-channel alsa on hda-intel

[Duncan]

Raffaele BELARDI posted on Tue, 25 Aug 2009 08:39:14 +0200 as excerpted:

> This is with HDA-intel driver on Nvidia nForce430/AD1986 chipset, icewm
> desktop. I just read from the kernel doc [2] that there is a "model
> name" parameter that defaults to '6stack' while my model is '3stack' so
> tonight I will try that.
> 
> Any other hints before I try upstream?

Just that the alsa hda drivers are undergoing quite a lot of changes 
upstream ATM, so you might expect possible behavior changes for 2.6.31 as 
you seem to have gotten for 2.6.30.  I don't have that audio device here, 
mine's much older, but I do run direct git kernels, and have developed 
the habit of running a git whatchanged ORIG_HEAD.. after every git pull, 
just to follow the changes as they happen.  As such, I've seen a LOT of 
those alsa:hda commits going by, so I know it's quite active ATM.  
However, it does appear most of those are simply adding tweaks and new 
model numbers to the list of what it supports, not major changes, so it's 
not /that/ bad.  But there's still enough change that you might see bugs 
if you're running that hardware and nobody that's testing is running your 
particular model of it.  But if they're reported, they should be fixed in 
short order, and even if you don't report them, the hardware's apparently 
common enough on mew machines that it'll probably be fixed in a stable 
update or two. (2.6.30 to 2.6.30.1 or .2, for instance, noting of course 
that those are the upstream versions, Gentoo uses -rX notation in place 
of the additional version digit, for its gentoo-sources packages.)

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman

Re: [gentoo-amd64] Re: problem with multi-channel alsa on hda-intel

[Raffaele BELARDI]

On Tue, 2009-08-25 at 16:24 +0000, Duncan wrote:
> Raffaele BELARDI posted on Tue, 25 Aug 2009 08:39:14 +0200 as excerpted:
> 
> > This is with HDA-intel driver on Nvidia nForce430/AD1986 chipset, icewm
> > desktop. I just read from the kernel doc [2] that there is a "model
> > name" parameter that defaults to '6stack' while my model is '3stack' so
> > tonight I will try that.
> > 
> > Any other hints before I try upstream?
> 
> Just that the alsa hda drivers are undergoing quite a lot of changes 
> upstream ATM, so you might expect possible behavior changes for 2.6.31 as 
> you seem to have gotten for 2.6.30.  I don't have that audio device here, 

I solved the issue by switching from '2ch' to '4ch' in alsamixer. Thanks
anyway.

> update or two. (2.6.30 to 2.6.30.1 or .2, for instance, noting of course 
> that those are the upstream versions, Gentoo uses -rX notation in place 
> of the additional version digit, for its gentoo-sources packages.)

I thought -rX was an internal Gentoo release number for ebuild. The
kernel is an exception, I guess.

raf

[gentoo-amd64] Re: problem with multi-channel alsa on hda-intel

[Duncan]

Raffaele BELARDI posted on Wed, 26 Aug 2009 11:37:03 +0200 as excerpted:

>> (2.6.30 to 2.6.30.1 or .2, for instance, noting of course
>> that those are the upstream versions, Gentoo uses -rX notation in place
>> of the additional version digit, for its gentoo-sources packages.)
> 
> I thought -rX was an internal Gentoo release number for ebuild. The
> kernel is an exception, I guess.

-rX is an internal Gentoo release number.  But with the kernel, you'll 
note that they don't normally track the upstream 4th level release 
numbers, so that ends up getting wrapped into the -rX number as well.

Do note however that they don't necessarily exactly correspond.  Thus, 
2.6.30-r1 may or may not correspond to upstream's 2.6.30.1.  Generally, 
there will be bumps about the same time, but if Gentoo finds and patches 
other issues between upstream bumps, it might bump the -rX twice for a 
single upstream 2.6.x.y bump, and the reverse occurs as well.  
Occasionally there'll be an upstream bump, and then another one maybe two 
hours later the same day, if they found a brown-paper-bag issue (in LKML 
lingo, a really embarrassing mistake, such that one wishes to cover their 
head with a bag if they appear in public immediately thereafter), or if 
there's an exceedingly urgent but also exceedingly simple security fix.  
These would probably be wrapped into the same Gentoo -rX bump.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman

Re: [gentoo-amd64] Re: problem with multi-channel alsa on hda-intel

[Raffaele BELARDI]

On Wed, 2009-08-26 at 14:37 +0000, Duncan wrote:
> -rX is an internal Gentoo release number.  But with the kernel, you'll 
> note that they don't normally track the upstream 4th level release 
> numbers, so that ends up getting wrapped into the -rX number as well.
> 
> Do note however that they don't necessarily exactly correspond.  Thus, 
> 2.6.30-r1 may or may not correspond to upstream's 2.6.30.1.  Generally, 
> there will be bumps about the same time, but if Gentoo finds and patches 
> other issues between upstream bumps, it might bump the -rX twice for a 
> single upstream 2.6.x.y bump, and the reverse occurs as well.  

it make sense, thanks.

raf