Sound in VMWare

Sound in VMWare

[emr]

Hi, I got 9front up for testing :-)

I cant get audio working, though:

"rc: play cannot open /dev/audio: mounted directory forbids creation"

... meaning there is no /dev/audio device.

The 9front wiki lists AC97 on the features list (
http://code.google.com/p/plan9front/wiki/features ), but it seems like
the device and vendor ID of the card created by VMWare is not known to
plan9/9front.

This is the card created by VMWare Workstation 7.1:

test@ubuntu:~$ lspci -s 02:02.0 -v -nn
02:02.0 Multimedia audio controller [0401]: Ensoniq ES1371
[AudioPCI-97] [1274:1371] (rev 02)
	Subsystem: Ensoniq Creative Sound Blaster AudioPCI64V, AudioPCI128
[1274:1371]
	Flags: bus master, medium devsel, latency 64, IRQ 16
	I/O ports at 2080 [size=64]
	Kernel driver in use: ENS1371
	Kernel modules: snd-ens1371

With the es1371 / ac97 chip, I get the speaker symbol from VMWare turn
to green, but kernel output shows no audio chip detected.

Please add this device ID to the relevant source files (unless they
are already there, but I am doing it wrong).

I tried changing the virtual audio device in VMWare by editing
the .vmx file:

  sound.virtualDev = "hdaudio"

instead of the default

  sound.virtualDev = "es1371"

but still I receive same error message in plan9/9front. This is the
Intel HDA device created by VMWare:

test@ubuntu:~$ lspci -s 02:02.0 -v -nn
02:02.0 Audio device [0403]: Intel Corporation 631xESB/632xESB High
Definition Audio Controller [8086:269a] (rev 02)
	Flags: bus master, fast devsel, latency 64, IRQ 16
	Memory at d8900000 (32-bit, non-prefetchable) [size=16K]
	Kernel driver in use: HDA Intel
	Kernel modules: snd-hda-intel

With HDA device, the VMWare speaker symbol does not show or turn to
green. Interestingly, the plan9 kernel shows "#A0: hda codec" etc.
Still same error message, though when trying to run "play
somesong.mp3".

Under virtual Linux, both device types work and I can play sound fine.

Any way to get these two cards working?

Greets, emr

Re: Sound in VMWare

[emr]

For added details, this is the exact kernel output when HDA device is
enabled:

#A0: hda mem e0000000 irq 5
#A0: waitup timeout for reg=4a, mask=8000, set=8000
#A0: codec #0, vendor 10ec0885, rev 00100103
#A0: no output pins found!

I am not sure what this last line (error message?) means..

Would still be great to get es1397 AC97 card working too as it is the
default card for VMWare.

Greets, emr

Re: Sound in VMWare

[cinap_lenrek]

yeah, and theres no driver for the es pci card. you can get audio working
in vmware using sb16 driver. i just tested this with the latest 9front
iso. our sb16 driver is different from the bell-labs one (i program the
dma controllers in cyclic mode to avoid audio dropouts in bochs and qemu)
and it works here in vmware player 4.0.1 on windows xp.

http://ninetimes.cat-v.org/tips/2008/08/06/1_Audio_support_in_VmWare/

if you still want to write a driver for the es:

ftp://ftp.alsa-project.org/pub/manuals/ensoniq/
http://fxr.watson.org/fxr/source/dev/pci/eap.c?v=OPENBSD

--
cinap

Re: Sound in VMWare

[Ethan Grammatikidis]

On Wed, Jun 13, 2012, at 06:38 AM, emr wrote:
> Hi, I got 9front up for testing :-)
> 
> I cant get audio working, though:
> 
> "rc: play cannot open /dev/audio: mounted directory forbids creation"
> 
> ... meaning there is no /dev/audio device.

Did you bind '#A' /dev? cpurc does not bind #A by default. (termrc does.)

Re: Sound in VMWare

[Ethan Grammatikidis]

On Sat, Jul 7, 2012, at 12:46 PM, Ethan Grammatikidis wrote:

> Did you bind '#A' /dev?

Woops, need -b or -a option to bind.

Re: Sound in VMWare

[cinap_lenrek]

the '#A' device is bound in termrc right at the top
see:

# bind all likely devices
for(i in S f k t m i v L P u U A '$' Σ κ)
	bind -a '#'^$i /dev >/dev/null >[2=1]

just check if you can:

ls '#A'

you need to enable sb16 emulation to get audio working in
vmware:

http://ninetimes.cat-v.org/tips/2008/08/06/1_Audio_support_in_VmWare/

--
cinap

Re: Sound in VMWare

[Matthew Veety]

On Jul 7, 2012, at 9:30 AM, cinap_lenrek@gmx.de wrote:

> the '#A' device is bound in termrc right at the top
> see:
> 
> # bind all likely devices
> for(i in S f k t m i v L P u U A '$' Σ κ)
> 	bind -a '#'^$i /dev >/dev/null >[2=1]
> 
> just check if you can:
> 
> ls '#A'
> 
> you need to enable sb16 emulation to get audio working in
> vmware:
> 
> http://ninetimes.cat-v.org/tips/2008/08/06/1_Audio_support_in_VmWare/
> 
> --
> cinap

Unrelated: Who was the jackass that wanted to use the sigma there?

--
Veety

Re: Sound in x61 (was VMWare)

[Ethan Grammatikidis]

On Sat, 7 Jul 2012 15:30:49 +0200
cinap_lenrek@gmx.de wrote:

> the '#A' device is bound in termrc right at the top

as I said, it is not bound in cpurc. if he's running a cpu server he
needs to bind it.

i'm having a little trouble myself with my x61. everything seems in
order (and i've bound #A), but play(1)ing a (known-good) mp3 fails to
produce any sound. play exits after exactly the expected time, verified
with date ; play 01* ; date. volume seems to be up, but i'm surprised
there's only master and no others:

% cat /dev/volume
master 100 100
speed 44100
delay 882

it's the first time i've tried to play anything with it.

-- 
This is obviously some strange usage of the 
word "simple" that I was previously unaware of.

Re: Sound in VMWare

[Ethan Grammatikidis]

On Sat, 7 Jul 2012 10:59:26 -0400
Matthew Veety <mveety@gmail.com> wrote:

> 
> On Jul 7, 2012, at 9:30 AM, cinap_lenrek@gmx.de wrote:
> 
> > the '#A' device is bound in termrc right at the top
> > see:
> > 
> > # bind all likely devices
> > for(i in S f k t m i v L P u U A '$' Σ κ)
> > 	bind -a '#'^$i /dev >/dev/null >[2=1]
> > 
> > just check if you can:
> > 
> > ls '#A'
> > 
> > you need to enable sb16 emulation to get audio working in
> > vmware:
> > 
> > http://ninetimes.cat-v.org/tips/2008/08/06/1_Audio_support_in_VmWare/
> > 
> > --
> > cinap
> 
> Unrelated: Who was the jackass that wanted to use the sigma there?

just type alt, *, S; or alt, *, k for the kappa on the end of the list.

-- 
This is obviously some strange usage of the 
word "just" that I was previously unaware of.

Re: Sound in VMWare

[Julius Schmidt]

Shut up Wesley

On Sat, 7 Jul 2012, Matthew Veety wrote:

>
> On Jul 7, 2012, at 9:30 AM, cinap_lenrek@gmx.de wrote:
>
>> the '#A' device is bound in termrc right at the top
>> see:
>>
>> # bind all likely devices
>> for(i in S f k t m i v L P u U A '$' Σ κ)
>> 	bind -a '#'^$i /dev >/dev/null >[2=1]
>>
>> just check if you can:
>>
>> ls '#A'
>>
>> you need to enable sb16 emulation to get audio working in
>> vmware:
>>
>> http://ninetimes.cat-v.org/tips/2008/08/06/1_Audio_support_in_VmWare/
>>
>> --
>> cinap
>
> Unrelated: Who was the jackass that wanted to use the sigma there?
>
> --
> Veety

Re: Sound in x61 (was VMWare)

[sl]

> i'm having a little trouble myself with my x61. everything seems in
> order (and i've bound #A), but play(1)ing a (known-good) mp3 fails to
> produce any sound. play exits after exactly the expected time, verified
> with date ; play 01* ; date.

On my x61t 7767-01u:

% pci -v # irrelevant portions ommitted
0.27.0:	aud  04.03.00 8086/284b  11 0:fe020004 16384 1:00000000 16
	Intel Corporation 888 Microsoft UAA bus for HD audio

% cat '#'A/audiostat
bufsize   1024 buffered      0
codec  0 pin  17
pin  17 out jack ext right hpout green
pin  18 out fix int N/A speaker ? eapd
pin  19 out nothing int rear speaker ? eapd
pin  20 in jack ext right micin red
pin  21 in fix int N/A micin ?
pin  22 out nothing int special cd ?
pin  23 in nothing int top micin ?
pin  24 in nothing int top micin ?
pin  26 in fix int rear other black
pin  27 out nothing sep rear spdifout black
pin  28 out jack sep rear micin red

% cat '#'A/volume
master 100 100
speed 44100
delay 882

Note that the lowest pin number is the line out. Assuming your
hardware is the same as mine, to use the speaker you would use
to do:

% echo pin 18 >/dev/audioctl

I just put this in /cfg/$sysname/termrc.

-sl

Re: Sound in VMWare

[sl]

> Unrelated: Who was the jackass that wanted to use the sigma there?

Undoubtedly, someone affiliated with Bell Labs. UTF-8 is used
throughout the system. What's your alternative for 1-character
special device names after the 26-character English alphabet
has been exhausted?

-sl