Re: Sound on G4 tower, almost

Re: Sound on G4 tower, almost

[Dave Ray]

Well, my /dev/mixer devices seem okay:

> cd /dev; ls -l mix*
lrwxrwxrwx    1 root     root            6 Feb 26
2001 mixer -> mixer0
crw-rw-rw-    1 root     17        14,   0 Oct  4
2000 mixer0
crw-rw-rw-    1 root     17        14,  16 Oct  4
2000 mixer1
crw-rw-rw-    1 root     17        14,  32 Oct  4
2000 mixer2
crw-rw-rw-    1 root     17        14,  48 Oct  4
2000 mixer3

The syslog messages that I posted previously do show
some errors when loading dmasound_pmac. I'm wondering
if those errors are related to the symptoms. Any help
appreciated.

-Dave



Derrik Pates <dpates@dsdk12.net> wrote:

> > Oct 19 20:54:38 garuda modprobe: modprobe: Can't
> > locate module sound-slot-1
> > Oct 19 20:54:38 garuda modprobe: modprobe: Can't
> > locate module sound-service-1-0
>
> seem to indicate something is trying to open
/dev/mixer1 instead of /dev/mixer0 (which /dev/mixer
should either be a symlink pointing to, or a dev node
with the same major/minor numbers). What does your
/dev/mixer look like? It shuld be major 14, minor 0.
Check /usr/src/linux/Documentation/devices.txt for all
the major/minor numbers for devices (including sound
devices).
>
> In other words, the Texas chip is probably
functioning fine - you have to tell it to do something
now, though...


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Sound on G4 tower, almost

[Derrik Pates]

On Mon, 22 Oct 2001, Dave Ray wrote:

> The syslog messages that I posted previously do show
> some errors when loading dmasound_pmac. I'm wondering
> if those errors are related to the symptoms. Any help
> appreciated.

I looked at the output you posted, and I didn't see any errors in those
messages. I'm using the same driver (albeit on a FireWire iBook, which has
a DACA chip, but the messages are similar), and so I know there is nothing
in any of those messages to indicate any non-functioning (in fact, it
claimed to have found the Texas/Tumbler I2C mixer device). What program
are you using to try to change mixer settings? Have you tried playing any
audio, and if so, what happened? Did you receive an error, or did it just
play silently, or what?

Derrik Pates      |   Sysadmin, Douglas School   |    #linuxOS on EFnet
dpates@dsdk12.net |     District (dsdk12.net)    |    #linuxOS on OPN


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Sound on G4 tower, almost

[Dave Ray]

--- Derrik Pates <dpates@dsdk12.net> wrote:
> On Mon, 22 Oct 2001, Dave Ray wrote:
>
> > The syslog messages that I posted previously do
> show
> > some errors when loading dmasound_pmac. I'm
> wondering
> > if those errors are related to the symptoms. Any
> help
> > appreciated.
>
> I looked at the output you posted, and I didn't see
> any errors in those
> messages. I'm using the same driver (albeit on a
> FireWire iBook, which has
> a DACA chip, but the messages are similar), and so I
> know there is nothing
> in any of those messages to indicate any
> non-functioning (in fact, it
> claimed to have found the Texas/Tumbler I2C mixer
> device). What program
> are you using to try to change mixer settings? Have
> you tried playing any
> audio, and if so, what happened? Did you receive an
> error, or did it just
> play silently, or what?

All sound apps just play silently without complaint,
but errors appear in the syslog.

Just to give some more background, I'm running the
latest audiofile libs and latest esound libs,
enlightenment 0.16.5, kernel 2.4.13-BenH. Under
previous hardware, audio (awacs) always worked, and I
generally ran esound.

However for this debugging I'm leaving esound off and
I should still be able to get terminal beeps and play
mp3's from gqmpeg. All audio programs happily run,
without complaint, but no sound comes out. When I try
to adjust volume I get the syslog errors. I'm stumped
and open to any suggestions.

The following lines in the syslog looked suspicious:

Oct 19 20:53:48 garuda kernel: tumbler:no device
.....
.....
Oct 19 20:53:48 garuda kernel: i2c-keywest: Ack Status
on addr expected but got: 0x09 on addr: 0x34
.....
(see original post for full syslog)

I suspect that the driver is failing to open the real
hardware and registering a dummy device.

-Dave


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Sound on G4 tower, almost

[Derrik Pates]

On Tue, 23 Oct 2001, Dave Ray wrote:

> However for this debugging I'm leaving esound off and
> I should still be able to get terminal beeps and play
> mp3's from gqmpeg. All audio programs happily run,
> without complaint, but no sound comes out. When I try
> to adjust volume I get the syslog errors. I'm stumped
> and open to any suggestions.

What errors are those? Nothing you've posted so far looks like an error to
me.

> Oct 19 20:53:48 garuda kernel: tumbler:no device

That's from not having a devnode for "amp-mute", "headphone-mute" or
"headphone-detect". I don't think that's a problem.

> Oct 19 20:53:48 garuda kernel: i2c-keywest: Ack Status
> on addr expected but got: 0x09 on addr: 0x34

That's just because the tas3001 driver code (the mixer for the
Tumbler/Texas) is searching all available I2C buses, and the first one
doesn't have what it's looking for.

> I suspect that the driver is failing to open the real
> hardware and registering a dummy device.

I suspect your mixer app is opening the wrong mixer devnode.

Derrik Pates      |   Sysadmin, Douglas School   |    #linuxOS on EFnet
dpates@dsdk12.net |     District (dsdk12.net)    |    #linuxOS on OPN


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Sound on G4 tower, success

[Dave Ray]

--- Derrik Pates <dpates@dsdk12.net> wrote:
>
> What errors are those? Nothing you've posted so far
> looks like an error to
> me.
>
> > Oct 19 20:53:48 garuda kernel: tumbler:no device
>
> That's from not having a devnode for "amp-mute",
> "headphone-mute" or
> "headphone-detect". I don't think that's a problem.
>
> > Oct 19 20:53:48 garuda kernel: i2c-keywest: Ack
> Status
> > on addr expected but got: 0x09 on addr: 0x34
>
> That's just because the tas3001 driver code (the
> mixer for the
> Tumbler/Texas) is searching all available I2C buses,
> and the first one
> doesn't have what it's looking for.

Thanks for your insight to these error messages. After
seeing your confidence that syslog looked okay, I
hypothesized that the output jacks may be live even
though the built-in speaker is dead silent. I plugged
the output audio jacks to an amplifier, and voila,
sound is is working there just fine.

Under MacOS (and Open Firmware) the internal speaker
is on by default, and turns off only when the audio
jack is physically connected. I assumed it was a
hardware switch and that if sound is working, I should
expect to hear it on the internal speaker. Not the
case apparently with dmasound_pmac. The speaker is
completely dead, and the output jack is always on,
regardless of whether a plug is connected. Is this a
bug? (Don't get me wrong, I'm delighted that it
works!)

This should go in some FAQ. Who's maintaining an FAQ
for the Mac G4 towers?

Thanks for all your help looking over my data,

Dave


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Sound on G4 tower, success

[Derrik Pates]

On Tue, 23 Oct 2001, Dave Ray wrote:

> Thanks for your insight to these error messages. After
> seeing your confidence that syslog looked okay, I
> hypothesized that the output jacks may be live even
> though the built-in speaker is dead silent. I plugged
> the output audio jacks to an amplifier, and voila,
> sound is is working there just fine.

Yeah. Apparently there's some weird magic crap that you have to do to make
the internal speaker actually work. Far as I know, no one's decoded how to
actually make it work in Linux, and no one seems overly interested in
figuring out how to anyway (since the internal speaker isn't anything
startling sound-quality-wise anyhow).

> regardless of whether a plug is connected. Is this a
> bug? (Don't get me wrong, I'm delighted that it
> works!)

No, as I noted above, it's not a bug, it's just something we don't yet
know how to control, and no one's felt it pressing to figure it out (and
the lack of documentation on it would make it something of a task to
figure out).

Derrik Pates      |   Sysadmin, Douglas School   |    #linuxOS on EFnet
dpates@dsdk12.net |     District (dsdk12.net)    |    #linuxOS on OPN


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Sound on G4 tower, success

[Benjamin Herrenschmidt]

>
>Yeah. Apparently there's some weird magic crap that you have to do to make
>the internal speaker actually work. Far as I know, no one's decoded how to
>actually make it work in Linux, and no one seems overly interested in
>figuring out how to anyway (since the internal speaker isn't anything
>startling sound-quality-wise anyhow).
>
>> regardless of whether a plug is connected. Is this a
>> bug? (Don't get me wrong, I'm delighted that it
>> works!)
>
>No, as I noted above, it's not a bug, it's just something we don't yet
>know how to control, and no one's felt it pressing to figure it out (and
>the lack of documentation on it would make it something of a task to
>figure out).

Well, I think I know how to control it. It's just that I don't have
the HW to test on ;)

I'll try to hack something soon.

Ben.


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Sound on G4 tower, success

[Lars Brinkhoff]

Benjamin Herrenschmidt <benh@kernel.crashing.org> writes:
> Well, I think I know how to control it. It's just that I don't have
> the HW to test on ;)

I'd be happy to test new sound hacks on a 2001 G4 Tower.

--
Lars Brinkhoff          http://lars.nocrew.org/     Linux, GCC, PDP-10
Brinkhoff Consulting    http://www.brinkhoff.se/    programming

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Sound on G4 tower, success

[Benjamin Herrenschmidt]

>Benjamin Herrenschmidt <benh@kernel.crashing.org> writes:
>> Well, I think I know how to control it. It's just that I don't have
>> the HW to test on ;)
>
>I'd be happy to test new sound hacks on a 2001 G4 Tower.

Can you tell me if the version of dmasound in my current rsync
helps getting sound our of speaker ? It should mute the speaker
when the jack is plugged and unmute it when unplugged. Can you
verify this behaviour ?

Ben


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Sound on G4 tower, success

[James Baughman]

> Can you tell me if the version of dmasound in my current rsync
> helps getting sound our of speaker ? It should mute the speaker
> when the jack is plugged and unmute it when unplugged. Can you
> verify this behaviour ?

The speaker is working on my dual 800/G4 from the kernel in your tree of
10pm Thurs. night.  Thanks!  I've not tried to plug anything into the
jack yet though.

Also, it seems as though the mixer isn't working quite right.  The
actual  volume (using aumix or xmixer) doesn't seem to correspond to the
level indicated on the mixer GUI.  If I set the volume to its lowest
setting, the actual volume could be cranked.  As I increment the volume
slider, the volume seems to randomly get louder and softer all the way
up to the slider's maximum.

--
Jim Baughman
Applied Research Laboratory
University of Texas, Austin


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Sound on G4 tower, success

[Lars Brinkhoff]

> Can you tell me if the version of dmasound in my current rsync helps
> getting sound our of speaker ? It should mute the speaker when the
> jack is plugged and unmute it when unplugged. Can you verify this
> behaviour ?

These are my results:

* kernel compilation without powerbock power management fails.
  (I resynced serveral times.)

* dmasound_pmac as module can't be loaded because of missing
  symbols (pmu_register_sleep_notifier, pmu_unregister_sleep_notifier).

* dmasound_pmac can be successfully compiled into the kernel, but
  there is no sound from neither the internal speaker nor from the
  loudspeakers.  I haven't tried the headphone output.

--
Lars Brinkhoff          http://lars.nocrew.org/     Linux, GCC, PDP-10
Brinkhoff Consulting    http://www.brinkhoff.se/    programming

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Sound on G4 tower, success

[Benjamin Herrenschmidt]

>
>* kernel compilation without powerbock power management fails.
>  (I resynced serveral times.)
>
>* dmasound_pmac as module can't be loaded because of missing
>  symbols (pmu_register_sleep_notifier, pmu_unregister_sleep_notifier).

Looks like your build is screwed in some way.

Either you have dangling root files or versions files. Try make mrproper
as root to clean the tree (after backing up your .config).

>* dmasound_pmac can be successfully compiled into the kernel, but
>  there is no sound from neither the internal speaker nor from the
>  loudspeakers.  I haven't tried the headphone output.

Did you try playing with a mixer like gmix ? The best way to try if
the sound driver works is the console beep. There are plenty of
reasons why a userland app would fail emitting sound, but the
console beep is entirely kernel driven.

Ben.


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Sound on G4 tower, success

[Lars Brinkhoff]

Benjamin Herrenschmidt <benh@kernel.crashing.org> writes:
> Can you tell me if the version of dmasound in my current rsync
> helps getting sound our of speaker ? It should mute the speaker
> when the jack is plugged and unmute it when unplugged. Can you
> verify this behaviour ?

A freshly rsynced tree works perfectly (apart from the config.in
thing).  Yes, the speaker mutes and unmutes as intended.

--
Lars Brinkhoff          http://lars.nocrew.org/     Linux, GCC, PDP-10
Brinkhoff Consulting    http://www.brinkhoff.se/    programming

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/