sound again [was: Re: xf4 bug report]

sound again [was: Re: xf4 bug report]

[Iain Sandoe]

On Tue, Jul 25, 2000, Steffen Haeuser wrote:
> Am 25-Jul-00 schrieb Iain Sandoe:
[...]
>>> Don't bother, that's a known problem as well. Blame Apple for not releasing
>>> specs for the latest sound chips...
>
>>which machine, which kernel (and, if possible, which sound chip - can be
>>found from the MacOS side if not the Linux) ?
>
> A "blue/white" G3 Mac first Revision. Kernel is: 2.2.17pre10-ben2 (with the
> older
> kernel Sound worked - at least I cannot remember that an error message during
> OS-booting happened, I never tried out the sound then.
> But the new Kernel is needed for the X Server V4)
>

OK, which was the last working version (prob. easiest is if you mail me
drivers/sound/dmasound.c).  I can diff that to what we have under 17preXX
and see if there are any clues...

I did a back-port of the 2.4.0-test4 dmasound stuff to 2.2.17pre10-ben2
(will check on pre13-ben1 later):
http://www.drfruitcake.com/linux/linuxppc.html

For some people this helps - but it seems that there are problems (some
minor, some major) across the board.   I would much prefer to work on the
2.4.0-testX code (more useful in the grand scheme of things) and so would
encourage you to test the back-port :-)

Having said that, I would expect your machine to work - I think it's an
AWACS sound chip of some variety?  If it has a DACA you'll have to wait..
:-(

Iain.

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

Re: sound again [was: Re: xf4 bug report]

[Steffen Haeuser]

Am 25-Jul-00 schrieb Iain Sandoe:

Hi!

>OK, which was the last working version (prob. easiest is if you mail me
>drivers/sound/dmasound.c).  I can diff that to what we have under 17preXX
>and see if there are any clues...

Hmmm, I never touched the source code... the latest with working sound was
2.2.15
... well, actually I never tried the sound there... but I *think* this error
message
during bootup concerning sound came first with the new kernel (something about
dma
it said... if that helps I can look up the exact error message...).

>I did a back-port of the 2.4.0-test4 dmasound stuff to 2.2.17pre10-ben2
>(will check on pre13-ben1 later):
>http://www.drfruitcake.com/linux/linuxppc.html

I guess the dmasnd-v2a.patch is the file to get ? Hmmm, is there also a Binary
Version available somewhere ? I do not have the latest Kernel Source on my
machine
and would like to avoid having to download it if possible...

BTW: Note about this page: at the start there is written
"PowerMac/Amiga/Atari/etc."
There is no such thing as a PowerAtari... there is a PowerAmiga, yes... but no
PowerAtari... Atari never got rid of the 68k CPU Architecture...
(and never, never name Amiga and Atari in one sentence :) )

>Having said that, I would expect your machine to work - I think it's an
>AWACS sound chip of some variety?  If it has a DACA you'll have to wait..
>:-(

No clue what soundchip... it is a out-of-the-box "Blue/White" G3 Mac 1st
Revision...
no changes...

Steffen


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

Re: sound again [was: Re: xf4 bug report]

[Geert Uytterhoeven]

On Tue, 25 Jul 2000, Steffen Haeuser wrote:
> BTW: Note about this page: at the start there is written
> "PowerMac/Amiga/Atari/etc."
> There is no such thing as a PowerAtari... there is a PowerAmiga, yes... but no
> PowerAtari... Atari never got rid of the 68k CPU Architecture...

Please note that it's _not_ written "Power{Mac/Amiga/Atari}".

> (and never, never name Amiga and Atari in one sentence :) )

Why not? Linux/m68k first ran on the Amiga, then on the Atari, and only much
later on the Mac. You can even compile a kernel that will run on both Amiga
and Atari! :-)

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds


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

Re: sound again [was: Re: xf4 bug report]

[Steffen Haeuser]

Am 25-Jul-00 schrieb Geert Uytterhoeven:

Hi!

>Why not? Linux/m68k first ran on the Amiga, then on the Atari, and only much

Some people might feel upset :)

>later on the Mac. You can even compile a kernel that will run on both Amiga
>and Atari! :-)

Would be only a 68k Kernel then though and not use the PowerPC Board in case
of the Amiga...

Steffen


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

Re: sound again [was: Re: xf4 bug report]

[Michael Schmitz]

> >Why not? Linux/m68k first ran on the Amiga, then on the Atari, and only much
>
> Some people might feel upset :)

Right. The very people that first brought Linux to the m68k Mac might feel
upset. One was a well known ix86 kernel hacker, the other a m68k hacker
from the Atari/Amiga camp.
While I personally never used an Amiga, and did work on mostly Atari code,
I've considered myself m68k kernel hacker first, Atari second (or third,
or....). Similar for most others on linux-m68k. With a bit of hindsight,
I'd say I'd not mention the Mac in the same sentence as all of the others.
But Amiga/Atari rivalry has never existed in the m68k team for as long as
I've followed it.

> >later on the Mac. You can even compile a kernel that will run on both Amiga
> >and Atari! :-)

And Mac, and VME, perhaps even hp300. The odd ones are Q40 and Sun3.

> Would be only a 68k Kernel then though and not use the PowerPC Board in case
> of the Amiga...

Right, but can you compile a kernel that will run on PowerMac/CHRP/PReP
and APUS? Which one's serial will work?

	Michael


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

Re: sound again [was: Re: xf4 bug report]

[Steffen Haeuser]

Am 26-Jul-00 schrieb Michael Schmitz:

Hi!

>Right. The very people that first brought Linux to the m68k Mac might feel
>upset. One was a well known ix86 kernel hacker, the other a m68k hacker
>from the Atari/Amiga camp.

There is no Atari/Amiga camp... actually Amiga and Atari people cannot stand
each other... :) Mac... yes... but Atari... no... :)

>> Would be only a 68k Kernel then though and not use the PowerPC Board in
case
>> of the Amiga...

>Right, but can you compile a kernel that will run on PowerMac/CHRP/PReP
>and APUS? Which one's serial will work?

The Linux for Amiga PowerPC Boards uses the APUS stuff. Only difference is
a special bootloader.

Steffen Haeuser


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

Re: sound again [was: Re: xf4 bug report]

[Michael Schmitz]

> >Right. The very people that first brought Linux to the m68k Mac might feel
> >upset. One was a well known ix86 kernel hacker, the other a m68k hacker
> >from the Atari/Amiga camp.
>
> There is no Atari/Amiga camp... actually Amiga and Atari people cannot stand
> each other... :) Mac... yes... but Atari... no... :)

BS. Geert, Roman Z., Jes Sorensen are Amiga users. /me, Roman H., Andreas
Schwab are Atari users. We get along fine. I'm talking Linux/m68k here, in
case it wasn't obvious.

> >Right, but can you compile a kernel that will run on PowerMac/CHRP/PReP
> >and APUS? Which one's serial will work?
>
> The Linux for Amiga PowerPC Boards uses the APUS stuff. Only difference is
> a special bootloader.

I'm aware of the connection between APUS and LinuxPPC for Amiga. My
question was: can you compile a kernel that will load and run on all of
the major PPC machines, namely CHRP/PREP, PowerMac, VME boards (sorry I
forgot you guys), and Amiga with PPC boards? Not configure and build a
single kernel for each of these, but configure and build one kernel for
all of them?

All I've got configured so far is a (PowerMac/PReP/CHRP) Machine Type
combination. Not what I asked for, as you will no doubt notice.

	Michael


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

Re: sound again [was: Re: xf4 bug report]

[Michel Dänzer]

Michael Schmitz wrote:
>
> > >Right. The very people that first brought Linux to the m68k Mac might
> > >feel upset. One was a well known ix86 kernel hacker, the other a m68k
> > >hacker from the Atari/Amiga camp.
> >
> > There is no Atari/Amiga camp... actually Amiga and Atari people cannot
> > stand each other... :) Mac... yes... but Atari... no... :)
>
> BS. Geert, Roman Z., Jes Sorensen are Amiga users. /me, Roman H., Andreas
> Schwab are Atari users. We get along fine. I'm talking Linux/m68k here, in
> case it wasn't obvious.

Right.

I personally don't like Ataris too much either (though there were times I
almost bought one before I was enlightened by the Amiga ;), but those wars
should really be over now - have they done any good for the Amiga?

In any case, they're highly inappropriate as far as Linux is concerned IMHO.


> > >Right, but can you compile a kernel that will run on PowerMac/CHRP/PReP
> > >and APUS? Which one's serial will work?
> >
> > The Linux for Amiga PowerPC Boards uses the APUS stuff. Only difference is
> > a special bootloader.
>
> I'm aware of the connection between APUS and LinuxPPC for Amiga. My
> question was: can you compile a kernel that will load and run on all of
> the major PPC machines, namely CHRP/PREP, PowerMac, VME boards (sorry I
> forgot you guys), and Amiga with PPC boards? Not configure and build a
> single kernel for each of these, but configure and build one kernel for
> all of them?
>
> All I've got configured so far is a (PowerMac/PReP/CHRP) Machine Type
> combination. Not what I asked for, as you will no doubt notice.

When Jesper Skov was the strong leader of APUS development, he worked on
making that possible, but he hasn't succeeded. Not that I think it's really
needed...


Michel


--
I'm so hungry, I could almost eat health food.
______________________________________________________________________________
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86, Team *AMIGA*, AUGS

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

Re: sound again [was: Re: xf4 bug report]

[Michael Schmitz]

> > All I've got configured so far is a (PowerMac/PReP/CHRP) Machine Type
> > combination. Not what I asked for, as you will no doubt notice.
>
> When Jesper Skov was the strong leader of APUS development, he worked on
> making that possible, but he hasn't succeeded. Not that I think it's really
> needed...

:-) Jesper's another one from the m68k port. I don't think it's really
needed for PPC either, all I wanted to point out that it's been possible
on m68k, because it was designed that way in the first place. And to
support what Geert said in response to the silly Amiga/Atari topic - for a
long time, the kernels that were released by the m68k maintainer had
support for both, mainly because he was too lazy to compile twice. I've
benefited far too much from the joint efforts from Amiga and Atari hackers
to be told there's no way these would ever work together.

Sorry for veering so far off topic.

	Michael


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

Re: sound again [was: Re: xf4 bug report]

[Steffen Haeuser]

Am 26-Jul-00 schrieb Michael Schmitz:

Hi!

>I'm aware of the connection between APUS and LinuxPPC for Amiga. My
>question was: can you compile a kernel that will load and run on all of
>the major PPC machines, namely CHRP/PREP, PowerMac, VME boards (sorry I
>forgot you guys), and Amiga with PPC boards? Not configure and build a
>single kernel for each of these, but configure and build one kernel for
>all of them?

>All I've got configured so far is a (PowerMac/PReP/CHRP) Machine Type
>combination. Not what I asked for, as you will no doubt notice.

Hmmm... okay, here you are right. Still I'd prefer (if I'd like to install
Linux on an
Amiga, I "only" have Linux on my Mac runninh here though) LinuxPPC on the
Amiga
to Linux68k anytimes :)

Steffen


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

Re: sound again [was: Re: xf4 bug report]

[Geert Uytterhoeven]

On Wed, 26 Jul 2000, Steffen Haeuser wrote:
> Am 26-Jul-00 schrieb Michael Schmitz:
> >Right. The very people that first brought Linux to the m68k Mac might feel
> >upset. One was a well known ix86 kernel hacker, the other a m68k hacker
> >from the Atari/Amiga camp.
>
> There is no Atari/Amiga camp... actually Amiga and Atari people cannot stand
> each other... :) Mac... yes... but Atari... no... :)

That was how it was before Commodore and Atari died. We've been united since
Linux/m68k exists :-)

> >> Would be only a 68k Kernel then though and not use the PowerPC Board in
> case
> >> of the Amiga...
>
> >Right, but can you compile a kernel that will run on PowerMac/CHRP/PReP
> >and APUS? Which one's serial will work?
>
> The Linux for Amiga PowerPC Boards uses the APUS stuff. Only difference is
> a special bootloader.

You cannot compile a kernel that works on all of the above due to memory
mappings. Either compile for PowerMac/CHRP/PReP, or for APUS. Anyway, the
Amiga serial driver is a sane driver that won't be activated on anything but
an Amiga.

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds


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

Re: sound again [was: Re: xf4 bug report]

[Steffen Haeuser]

Am 27-Jul-00 schrieb Geert Uytterhoeven:

Hi!

>That was how it was before Commodore and Atari died. We've been united since
>Linux/m68k exists :-)

Hmmm... most Amiga users I know don't care for 68k anymore :) But I suggest
we drop that matter now... :) Probably not really interesting for most people
on the list :)))

Steffen


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

Re: sound again [was: Re: xf4 bug report]

[Iain Sandoe]

Tue, Jul 25, 2000,  Steffen Haeuser wrote:
> Am 25-Jul-00 schrieb Iain Sandoe:
[...]
> during bootup concerning sound came first with the new kernel (something about
> dma
> it said... if that helps I can look up the exact error message...).

can you do a 'dmesg' and find any lines which have "dmasound" in them?
Or... are you using sound as modules?
I assume you are using Ben's binary (pre10-ben2)?

>>I did a back-port of the 2.4.0-test4 dmasound stuff to 2.2.17pre10-ben2
>>(will check on pre13-ben1 later):
>>http://www.drfruitcake.com/linux/linuxppc.html
>
> I guess the dmasnd-v2a.patch is the file to get ? Hmmm, is there also a Binary
> Version available somewhere ? I do not have the latest Kernel Source on my
> machine
> and would like to avoid having to download it if possible...

OK - I'll do it later (just this once :-) I don't want to as a general rule
- because I have to pay for download bandwidth -- if there's a general move
for people to want to download binaries - I'll have to beg some space on an
ftp server...

I can't just send you the binaries for the modules - because they wouldn't
load against an un-patched 2.2.17pre10.

> BTW: Note about this page: at the start there is written
> "PowerMac/Amiga/Atari/etc."
> There is no such thing as a PowerAtari... there is a PowerAmiga, yes... but no
> PowerAtari... Atari never got rid of the 68k CPU Architecture...
> (and never, never name Amiga and Atari in one sentence :) )

OK, OK, maybe I wasn't clear ==> the dmasound stuff has bits for several
machines including some of the 68k ones...

>>Having said that, I would expect your machine to work - I think it's an
>>AWACS sound chip of some variety?  If it has a DACA you'll have to wait..
>>:-(
>
> No clue what soundchip... it is a out-of-the-box "Blue/White" G3 Mac 1st
> Revision...

You can find this using "Display Name Registry" on the apple side [it's
available on the Apple web site IIRC] look under davbus:sound.

- or by looking in the /proc/device-tree on the linux side (I hope -
although, of course, this could be the root of the problem :).

Iain.

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

Re: sound again [was: Re: xf4 bug report]

[Steffen Haeuser]

Am 25-Jul-00 schrieb Iain Sandoe:

Hi!

>can you do a 'dmesg' and find any lines which have "dmasound" in them?
>Or... are you using sound as modules?
>I assume you are using Ben's binary (pre10-ben2)?

I just checked the message. I was wrong about the DMA stuff... sorry...
it reads

modprobe: can't locate module sound
rc.sysinit: Loading Sound module failed

Sounds as if the module is missing... but I installed the modules
exactly how I found them on the site... and the kernel and system.map
also...

>OK - I'll do it later (just this once :-) I don't want to as a general rule
>- because I have to pay for download bandwidth -- if there's a general move

Well, same here...

>for people to want to download binaries - I'll have to beg some space on an
>ftp server...

>I can't just send you the binaries for the modules - because they wouldn't
>load against an un-patched 2.2.17pre10.

Ah...

>OK, OK, maybe I wasn't clear ==> the dmasound stuff has bits for several
>machines including some of the 68k ones...

Ah :)

>You can find this using "Display Name Registry" on the apple side [it's
>available on the Apple web site IIRC] look under davbus:sound.

Not sure if I know what you mean...

>- or by looking in the /proc/device-tree on the linux side (I hope -
>although, of course, this could be the root of the problem :).

Well, I found the /proc/device-tree... what to look for there ? There are
a lot of files and directories...

Steffen


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

Re: sound again [was: Re: xf4 bug report]

[Franz Sirl]

At 01:30 26.07.00, Steffen Haeuser wrote:

>Am 25-Jul-00 schrieb Iain Sandoe:
>
>Hi!
>
> >can you do a 'dmesg' and find any lines which have "dmasound" in them?
> >Or... are you using sound as modules?
> >I assume you are using Ben's binary (pre10-ben2)?
>
>I just checked the message. I was wrong about the DMA stuff... sorry...
>it reads
>
>modprobe: can't locate module sound
>rc.sysinit: Loading Sound module failed
>
>Sounds as if the module is missing... but I installed the modules
>exactly how I found them on the site... and the kernel and system.map
>also...

It may be missing, but most likely it is compiled directly into the kernel.
Ben does that usually.

Franz.


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

Re: sound again [was: Re: xf4 bug report]

[Iain Sandoe]

Hi Steffen,

On Tue, Jul 25, 2000, Steffen Haeuser wrote:
> Am 25-Jul-00 schrieb Iain Sandoe:
[...]
>>I did a back-port of the 2.4.0-test4 dmasound stuff to 2.2.17pre10-ben2
>>(will check on pre13-ben1 later):
>>http://www.drfruitcake.com/linux/linuxppc.html
>
> I guess the dmasnd-v2a.patch is the file to get ? Hmmm, is there also a Binary
> Version available somewhere ? I do not have the latest Kernel Source on my
> machine
> and would like to avoid having to download it if possible...

OK I've done a build - hope it's got enough in it to boot your machine :-)
it's linked to the page given above along with the modules & system map.

=====

To find the sound chip info (on my machine) under linux:

cat /proc/device-tree/pci/mac-io/davbus/sound/compatible

of course, different machines have different IO structures so YMMV...

a bit of poking around under proc should reveal something sensible (it's not
that big a tree really)...

Tschuss,

Iain

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

Re: sound again [was: Re: xf4 bug report]

[Steffen Haeuser]

Am 25-Jul-00 schrieb Iain Sandoe:

>Hi Steffen,

Hi!

>OK I've done a build - hope it's got enough in it to boot your machine :-)
>it's linked to the page given above along with the modules & system map.

Thanks, I am downloading it right now :)

>=====

>To find the sound chip info (on my machine) under linux:

>cat /proc/device-tree/pci/mac-io/davbus/sound/compatible

The filename is

/proc/device-tree/pci/pci-bridge/mac-io/davbus/sound/compatible

and the file contains

burgundyawacs

if this says anything to you :)

Steffen


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

Re: sound again [was: Re: xf4 bug report]

[Steffen Haeuser]

Am 25-Jul-00 schrieb Iain Sandoe:

>Hi Steffen,


Hi!

Okay, installed the new one...

Well, it behaves just the same like the old kernel... same error-message...

Steffen


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

Re: sound again [was: Re: xf4 bug report]

[Iain Sandoe]

Hi Steffen,
On Wed, Jul 26, 2000, Steffen Haeuser wrote:
> Am 25-Jul-00 schrieb Iain Sandoe:
[...]

> and the file contains
>
> burgundyawacs
>
> if this says anything to you :)

AH HA!!! - you are the first "burgundy" reply I've had... I will be *very*
interested to hear how it goes...

Iain.

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

Re: sound again [was: Re: xf4 bug report]

[Steffen Haeuser]

Am 25-Jul-00 schrieb Iain Sandoe:

>Hi Steffen,

Hi!

>On Wed, Jul 26, 2000, Steffen Haeuser wrote:
>> Am 25-Jul-00 schrieb Iain Sandoe:
>[...]

>> and the file contains
>>
>> burgundyawacs
>>
>> if this says anything to you :)

>AH HA!!! - you are the first "burgundy" reply I've had... I will be *very*
>interested to hear how it goes...

What does this mean ? This burgundy stuff ?

Steffen


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

Re: sound again [was: Re: xf4 bug report]

[Iain Sandoe]

 Wed, Jul 26, 2000,  Steffen Haeuser wrote:
> Am 25-Jul-00 schrieb Iain Sandoe:

> Okay, installed the new one...
>
> Well, it behaves just the same like the old kernel... same error-message...

Glad to hear it boots OK (I wasn't sure that the AppleSCSI driver wouldn't
panic on boot - it used to on my Lombard)...

But... does the sound work when you load the modules by hand? (see web-page
for details).  I suspect you have a bogus entry in your conf.modules...

Iain.

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

Re: sound again [was: Re: xf4 bug report]

[Steffen Haeuser]

Am 25-Jul-00 schrieb Iain Sandoe:

Hi!

> Wed, Jul 26, 2000,  Steffen Haeuser wrote:
>> Am 25-Jul-00 schrieb Iain Sandoe:

>> Okay, installed the new one...
>>
>> Well, it behaves just the same like the old kernel... same error-message...

>Glad to hear it boots OK (I wasn't sure that the AppleSCSI driver wouldn't
>panic on boot - it used to on my Lombard)...

Appearently the problem was a wrong entry in /etc/conf.modules... it had

alias sound dmasound

there while correct would be

alias sound dmasound_awacs

After I fixed this the sound modules can be correctly loaded during bootup,
but there is still no sound. Maybe I should go back to the old kernel and try
if it works with that, or don't you think this will make a difference ?

Steffen


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

Re: sound again [was: Re: xf4 bug report]

[Iain Sandoe]

HI Steffen,
[...]
>
>>AH HA!!! - you are the first "burgundy" reply I've had... I will be *very*
>>interested to hear how it goes...
>
> What does this mean ? This burgundy stuff ?

It's a more "advanced" version of the AWACS chip - should actually be easier
to deal with since you can actually read the registers in it (instead of
having to remember what you wrote - I always *did* hate write only
registers)...
 :-)

Iain.

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

Re: sound again [was: Re: xf4 bug report]

[Steffen Haeuser]

Am 25-Jul-00 schrieb Iain Sandoe:

Hi!

Just wanted to tell that after putting earphones to the Mac, the sound comes.
Appearently it is only the internal speaker which does not work.

The Gnome-Event sounds sound "sort-of" correct (but a big "hard and noisy"),
but this might be correct, as MPG123 sounds perfectly :)

When I started Sin I just fell from my chair of this BIG NOISE, but this is
probably our fault (the guy who coded the sound was a Linux-x86 guy, so it
might be an Endian Problem, I have to check this... We'll surely manage
to fix this :) ).

Anyways, thanks for your help, guys :)

Steffen


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

Re: sound again [was: Re: xf4 bug report]

[Michel Dänzer]

Steffen Haeuser wrote:

> Just wanted to tell that after putting earphones to the Mac, the sound
> comes. Appearently it is only the internal speaker which does not work.

Just for the record:

My Pismo shows the same behaviour, the 'compatible' file in /proc/device-tree
says 'screamerawacs'.


Michel


--
0 and 1. Now what could be so hard about that?
______________________________________________________________________________
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86, Team *AMIGA*, AUGS

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

Re: sound again [was: Re: xf4 bug report]

[Benjamin Herrenschmidt]

>> Just wanted to tell that after putting earphones to the Mac, the sound
>> comes. Appearently it is only the internal speaker which does not work.
>
>Just for the record:
>
>My Pismo shows the same behaviour, the 'compatible' file in /proc/device-tree
>says 'screamerawacs'.

I'm not completely sure, but some of these machines have the sound output
routed to an external I2C "SRS"/ampilfier module that drives the
speakers. We have code to set it up via Cuda on older machines, but not
for newer Core99 machines.

An interesting note also: When using this external module, Darwin driver
"fix" the AWACS volume to a constant value and uses the amplifier volume
control via I2C.

Ben.


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

Re: sound again [was: Re: xf4 bug report]

[Topi Kanerva]

On Wed, 26 Jul 2000, Steffen Haeuser wrote:

>Just wanted to tell that after putting earphones to the Mac, the sound comes.
>Appearently it is only the internal speaker which does not work.
>

so this behaviour is also happening on Burgundy. I've got an AWACS rev 3
here in my Pismo (device-tree name for the chip is 'screamerawacs')
and everything else works for me _except_ the damned built-in speakers.
as long as I don't have any external speakers to connect to, this is _very_
annoying, I have to boot to MacOS to listen to CD's etc etc... first I
thought the problem was in the new version of the sound chip, but since
Burgundies seem to have the same problem, I'm not so sure about it anymore.
However, poking around the mixer code doesn't shed any more light into the
problem.

Now, what do the Burgundy and AWACS rev.3 have in common? Most certainly
not the registers.


-topi


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

Re: sound again [was: Re: xf4 bug report]

[Benjamin Herrenschmidt]

>so this behaviour is also happening on Burgundy. I've got an AWACS rev 3
>here in my Pismo (device-tree name for the chip is 'screamerawacs')
>and everything else works for me _except_ the damned built-in speakers.
>as long as I don't have any external speakers to connect to, this is _very_
>annoying, I have to boot to MacOS to listen to CD's etc etc... first I
>thought the problem was in the new version of the sound chip, but since
>Burgundies seem to have the same problem, I'm not so sure about it anymore.
>However, poking around the mixer code doesn't shed any more light into the
>problem.
>
>Now, what do the Burgundy and AWACS rev.3 have in common? Most certainly
>not the registers.

New "screamer" revisions of AWACS have more registers than defined in the
current Linux driver. I'm not sure this is the cause of our problem, but
it could be. I thinking about a problem setting up an external amplifier.

Note also that a working sound driver wont give you CD playback "out of
the box". On all new machines, the CD audio output is _not_ wired to the
sound chip. You have to use a player software that actually reads the
digital audio datas from the CD via ATAPI and sends them back to the
sound driver. There's an XMMS plugin for that, I managed to use it but it
still contains a lot of bugs.

Ben.


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

Re: sound again [was: Re: xf4 bug report]

[Topi Kanerva]

On Wed, 26 Jul 2000, Benjamin Herrenschmidt wrote:

>New "screamer" revisions of AWACS have more registers than defined in the
>current Linux driver. I'm not sure this is the cause of our problem, but
>it could be. I thinking about a problem setting up an external amplifier.
>

Now, do you propose any method of determining exactly *where* the problem
lies? How can we manage to get hold of the specs of the new audio chips?
I mean, it works on MacOS, someone ought to know how to program the device!

>Note also that a working sound driver wont give you CD playback "out of
>the box". On all new machines, the CD audio output is _not_ wired to the
>sound chip. You have to use a player software that actually reads the
>digital audio datas from the CD via ATAPI and sends them back to the

I know. I use "cdparanoia 1- - | play -"
However it seems to be quite solid, i.e. I don't ever hear any 'blips'
resulting from jitter, so maybe the DVD-ROM drive is just reading the
audio data with full fidelity.

-topi


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

Re: sound again [was: Re: xf4 bug report]

[Iain Sandoe]

Hmmm...
[...]
>>Glad to hear it boots OK (I wasn't sure that the AppleSCSI driver wouldn't
>>panic on boot - it used to on my Lombard)...
>
> Appearently the problem was a wrong entry in /etc/conf.modules... it had
>
> alias sound dmasound
>
> there while correct would be
>
> alias sound dmasound_awacs
>
> After I fixed this the sound modules can be correctly loaded during bootup,
> but there is still no sound. Maybe I should go back to the old kernel and try
> if it works with that, or don't you think this will make a difference ?

No, I'm afraid it won't make any difference.

The dmasound_awacs entry only arrives with the back-port.  As Franz said,
the error you were getting before was probably due to the conf.modules entry
- when Ben had built the dmasound in.

So the long & short of it is that you don't get sound out with the "current"
set of code.  == there *should* be no difference in the implementation
between 2.2.17pre10 and the back-port (I checked that fairly carefully)...
and this seems to bear that out.

So.. the next thing to do is to find out why you get no sound.

1. what does dmesg now show?
2. have you tried headphones/versus built-in speaker?  (others have reported
a problem here - which is why the finger of blame is currently pointing at
the mixer abstraction)...

3. what does cat /dev/sndstat say?

4. what are you using to try and get sound out (e.g. some of the kde
implementations had broken sound anyway) - I never got *any* beeps &
whistles until I upgraded.

I can send or post a prog. that sets the snd output to 16bit LE - so that
you can just cat <some_file.wav> >/dev/dsp  if that's any use.

xmms usually works for me (although I've been using the cvs version)...

Iain.

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

Re: sound again [was: Re: xf4 bug report]

[Steffen Haeuser]

Am 25-Jul-00 schrieb Iain Sandoe:

Hi!

>The dmasound_awacs entry only arrives with the back-port.  As Franz said,
>the error you were getting before was probably due to the conf.modules entry
>- when Ben had built the dmasound in.

Ah.

>1. what does dmesg now show?

DMA sound driver core [Ed 3]

4 Buffers of 32 K for Output
Same for Input

AWACS rev 100

registers:

1400 782d 7810 0001 00a1 0021 0000 0000
0000 0000 0000 0000 0000 0000 0000 0000
0000 0000 4000 0000 26bb 0010 0000 0002
0001 0000 0000 0000 0000 0000 0000 0000

>2. have you tried headphones/versus built-in speaker?  (others have reported
>a problem here - which is why the finger of blame is currently pointing at
>the mixer abstraction)...

Up to now only built-in speakers. Will try with headphones :)

>3. what does cat /dev/sndstat say?

PowerMac (AWACS rev 100) DMA sound driver:
  sound.format = 0x20 (signed 16 bit big)
  sound.speed = 44100Hz (phys. 44100Hz)
  sound.stereo = 0x1 (stereo)
  write_sq.block_size = 32768 write_sq.max_count = 4 write_sq.max_active = 4
  write_sq.count = 0 write_sq.rear_size = 4096
  write_sq.active = 0 write_sq.syncing = 0
  read_sq.block_size = 0 read_sq.max_count = 0 read_sq.max_active = 0
  read_sq.count = 0 read_sq.rear_size = 0
  read_sq.active = 0 read_sq.syncing = 0

>4. what are you using to try and get sound out (e.g. some of the kde
>implementations had broken sound anyway) - I never got *any* beeps &
>whistles until I upgraded.

I tried in the Gnome Sound Settings. And I tried Sin naturally :) (Sound
works on Linux-x86, and does not contain any x86-specific code, so normally
it should work on LinuxPPC straight away - if sound is correctly configured,
that is :) )

>I can send or post a prog. that sets the snd output to 16bit LE - so that
>you can just cat <some_file.wav> >/dev/dsp  if that's any use.

>xmms usually works for me (although I've been using the cvs version)...

Well, I have some sound programs here... at least the Sin-stuff should work...
quite straight WAV-replaying...

Steffen


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

Re: sound again [was: Re: xf4 bug report]

[Iain Sandoe]

Hi Steffen,

I think we should take this off-list... (it's getting to be a bit
long-winded)...

Ben, Franz (if you are 'listening' or have an interest) let me know and I'll
cc you otherwise we'll report back when there's some solution.

>>1. what does dmesg now show?
>
> DMA sound driver core [Ed 3]

Good that's definitely the back-port code...

> 4 Buffers of 32 K for Output
> Same for Input

You've abbreviated this (I hope)...

> AWACS rev 100
>
> registers:
>
> 1400 782d 7810 0001 00a1 0021 0000 0000
> 0000 0000 0000 0000 0000 0000 0000 0000
> 0000 0000 4000 0000 26bb 0010 0000 0002
> 0001 0000 0000 0000 0000 0000 0000 0000

Good, it's recognised your Burgundy chip...

>>2. have you tried headphones/versus built-in speaker?  (others have reported
>>a problem here - which is why the finger of blame is currently pointing at
>>the mixer abstraction)...
>
> Up to now only built-in speakers. Will try with headphones :)
>
[snip]

The rest looks OK too...

Now, I'll have a quick look at the amp-enable code (again) - that did have a
problem at one point.

I'm out (on paid work :) tomorrow... so not much action 'till late...

Iain.

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

Re: sound again [was: Re: xf4 bug report]

[Iain Sandoe]

Hey guys, I've used all of them at one time or another...
Mac(s), Amiga, Atari(s) & VME...
Never meant to start a war :-))))

now next message --- back to sorting out the problem...
Iain...

>>Right. The very people that first brought Linux to the m68k Mac might feel
>>upset. One was a well known ix86 kernel hacker, the other a m68k hacker
>>from the Atari/Amiga camp.
>
> There is no Atari/Amiga camp... actually Amiga and Atari people cannot stand
> each other... :) Mac... yes... but Atari... no... :)
>
>>> Would be only a 68k Kernel then though and not use the PowerPC Board in
> case
>>> of the Amiga...
>
>>Right, but can you compile a kernel that will run on PowerMac/CHRP/PReP
>>and APUS? Which one's serial will work?
>
> The Linux for Amiga PowerPC Boards uses the APUS stuff. Only difference is
> a special bootloader.


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

Re: sound again [was: Re: xf4 bug report]

[Iain Sandoe]

Hello all,

I'd say that a summary of today's (most interesting) thread is...

That:

(a) There's no real difference in behaviour between the back-port and the
2.2.17preXX monolithic dmasound module.

(b) The finger of blame is still pointing where I suspected (i.e. the mixer
abstraction).

OK -

1/  Is there an "official" maintainer?

2/  If not, is there anyone else (other than me) who wants to volunteer to
get to the bottom of it?

3/ Is there any reason to work on the 2.2.17preXX monolithic version or can
we continue with the 2.4.0-testxx with conditionals? (my preference as it
will transport directly to 2.4.0).

FYI: there is another party working on the I2C bus to bring support to the
iBook etc. - I've already made contact but he wants to get things going
before merging...

I *am* willing to have a go at doing a further split & debug job (certainly
on the pmac stuff).  Is there any news as to whether the code is OK of the
other platforms?

Thanks everyone for your input & an exciting read after a long and tiring
day out of the office :-)

Iain.

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