AC97 audio broken on SAM9263EK?

AC97 audio broken on SAM9263EK?

[Bill Gatliff]

Guys:


It looks like AC97 support for the SAM9263EK isn't in the current
mainline kernel.  Or am I just missing something?

More specifically, it looks like the "atmel_ac97c" driver found in
sound/atmel/ac97c.c is what I need, based on my look at
at91_add_device_ac97().  But I see this in sound/atmel/Kconfig:

    config SND_ATMEL_AC97C
            tristate "Atmel AC97 Controller (AC97C) driver"
            select SND_PCM
            select SND_AC97_CODEC
            depends on DW_DMAC && AVR32

And I see this in sound/atmel/Makefile:

    snd-atmel-abdac-objs            := abdac.o
    snd-atmel-ac97c-objs            := ac97c.o

    obj-$(CONFIG_SND_ATMEL_ABDAC)   += snd-atmel-abdac.o
    obj-$(CONFIG_SND_ATMEL_AC97C)   += snd-atmel-ac97c.o

In other words, the driver isn't built unless I'm on an AVR32 platform
and I have a DesignWare DMAC.  Neither of those describe the AT91SAM9263EK.

Has anyone tested AC97 audio on the SAM9263EK lately?  Has it ever worked?


Thanks!

-- 
Bill Gatliff
Embedded systems training and consulting
http://billgatliff.com
bgat at billgatliff.com

AC97 audio broken on SAM9263EK?

[Sedji Gaouaou]

Hi Bill,

I have been working on the sam9263 AC97 driver. I have sent a patch to 
the alsa mailing list but it has not being integrated yet (as far as I 
know...)(http://mailman.alsa-project.org/pipermail/alsa-devel/2010-February/025119.html).

I have tested it and it worked, but I have recently made some changes so 
the atmel ac97.c driver could support full duplex mode.

So I will submit a new patch by the end of the day/week, and I will copy 
it to you if you are interested.

Regards,
Sedji Gaaouaou
Le 2/24/2010 8:06 PM, Bill Gatliff a ?crit :
> Guys:
>
>
> It looks like AC97 support for the SAM9263EK isn't in the current
> mainline kernel.  Or am I just missing something?
>
> More specifically, it looks like the "atmel_ac97c" driver found in
> sound/atmel/ac97c.c is what I need, based on my look at
> at91_add_device_ac97().  But I see this in sound/atmel/Kconfig:
>
>      config SND_ATMEL_AC97C
>              tristate "Atmel AC97 Controller (AC97C) driver"
>              select SND_PCM
>              select SND_AC97_CODEC
>              depends on DW_DMAC&&  AVR32
>
> And I see this in sound/atmel/Makefile:
>
>      snd-atmel-abdac-objs            := abdac.o
>      snd-atmel-ac97c-objs            := ac97c.o
>
>      obj-$(CONFIG_SND_ATMEL_ABDAC)   += snd-atmel-abdac.o
>      obj-$(CONFIG_SND_ATMEL_AC97C)   += snd-atmel-ac97c.o
>
> In other words, the driver isn't built unless I'm on an AVR32 platform
> and I have a DesignWare DMAC.  Neither of those describe the AT91SAM9263EK.
>
> Has anyone tested AC97 audio on the SAM9263EK lately?  Has it ever worked?
>
>
> Thanks!
>

AC97 audio broken on SAM9263EK?

[Bill Gatliff]

Sedji Gaouaou wrote:
> Hi Bill,
>
> I have been working on the sam9263 AC97 driver. I have sent a patch to
> the alsa mailing list but it has not being integrated yet (as far as I
> know...)(http://mailman.alsa-project.org/pipermail/alsa-devel/2010-February/025119.html).
>
>
> I have tested it and it worked, but I have recently made some changes
> so the atmel ac97.c driver could support full duplex mode.
>
> So I will submit a new patch by the end of the day/week, and I will
> copy it to you if you are interested.

Please do!  I'll also take a look at your submitted driver, and see how
it works for me.


Thanks!


b.g.

-- 
Bill Gatliff
Embedded systems training and consulting
http://billgatliff.com
bgat at billgatliff.com

AC97 audio broken on SAM9263EK?

[Sedji Gaouaou]

Hi Bill,

Here is the patch for the ac97 support on AT91:

http://mailman.alsa-project.org/pipermail/alsa-devel/2010-February/025629.html

Here is the patch for the full duplex support:

http://mailman.alsa-project.org/pipermail/alsa-devel/2010-February/025628.html

Best Regards,
Sedji Gaouaou

Le 2/25/2010 1:46 PM, Bill Gatliff a ?crit :
> Sedji Gaouaou wrote:
>> Hi Bill,
>>
>> I have been working on the sam9263 AC97 driver. I have sent a patch to
>> the alsa mailing list but it has not being integrated yet (as far as I
>> know...)(http://mailman.alsa-project.org/pipermail/alsa-devel/2010-February/025119.html).
>>
>>
>> I have tested it and it worked, but I have recently made some changes
>> so the atmel ac97.c driver could support full duplex mode.
>>
>> So I will submit a new patch by the end of the day/week, and I will
>> copy it to you if you are interested.
>
> Please do!  I'll also take a look at your submitted driver, and see how
> it works for me.
>
>
> Thanks!
>
>
> b.g.
>

AC97 audio broken on SAM9263EK?

[Bill Gatliff]

Sedji Gaouaou wrote:
> Hi Bill,
>
> Here is the patch for the ac97 support on AT91:
>
> http://mailman.alsa-project.org/pipermail/alsa-devel/2010-February/025629.html
>
>
> Here is the patch for the full duplex support:
>
> http://mailman.alsa-project.org/pipermail/alsa-devel/2010-February/025628.html
>

Hmm.  I applied both patches to 2.6.33-rc5.  There were two local
variables "err" and "ptcr" that weren't defined, which I corrected.  And
then there was this:

 CC      sound/atmel/ac97c.o
sound/atmel/ac97c.c: In function ?atmel_ac97c_capture_trigger?:
sound/atmel/ac97c.c:545: warning: unused variable ?casr?
sound/atmel/ac97c.c: In function ?atmel_ac97c_pcm_new?:
sound/atmel/ac97c.c:743: error: ?at91_ac97_pcm_defs? undeclared (first
use in this function)
sound/atmel/ac97c.c:743: error: (Each undeclared identifier is reported
only once
sound/atmel/ac97c.c:743: error: for each function it appears in.)
sound/atmel/ac97c.c:743: warning: type defaults to ?int? in declaration
of ?type name?
sound/atmel/ac97c.c:743: warning: type defaults to ?int? in declaration
of ?type name?
sound/atmel/ac97c.c:743: error: negative width in bit-field ?<anonymous>?
make[2]: *** [sound/atmel/ac97c.o] Error 1


b.g.

-- 
Bill Gatliff
Embedded systems training and consulting
http://billgatliff.com
bgat at billgatliff.com

AC97 audio broken on SAM9263EK?

[Sedji Gaouaou]

Hi,

Sorry my mistake, could you please try this one:
http://mailman.alsa-project.org/pipermail/alsa-devel/2010-March/025709.html

Regards,
Sedji

Le 2/28/2010 4:23 AM, Bill Gatliff a ?crit :
> Sedji Gaouaou wrote:
>> Hi Bill,
>>
>> Here is the patch for the ac97 support on AT91:
>>
>> http://mailman.alsa-project.org/pipermail/alsa-devel/2010-February/025629.html
>>
>>
>> Here is the patch for the full duplex support:
>>
>> http://mailman.alsa-project.org/pipermail/alsa-devel/2010-February/025628.html
>>
>
> Hmm.  I applied both patches to 2.6.33-rc5.  There were two local
> variables "err" and "ptcr" that weren't defined, which I corrected.  And
> then there was this:
>
>   CC      sound/atmel/ac97c.o
> sound/atmel/ac97c.c: In function ?atmel_ac97c_capture_trigger?:
> sound/atmel/ac97c.c:545: warning: unused variable ?casr?
> sound/atmel/ac97c.c: In function ?atmel_ac97c_pcm_new?:
> sound/atmel/ac97c.c:743: error: ?at91_ac97_pcm_defs? undeclared (first
> use in this function)
> sound/atmel/ac97c.c:743: error: (Each undeclared identifier is reported
> only once
> sound/atmel/ac97c.c:743: error: for each function it appears in.)
> sound/atmel/ac97c.c:743: warning: type defaults to ?int? in declaration
> of ?type name?
> sound/atmel/ac97c.c:743: warning: type defaults to ?int? in declaration
> of ?type name?
> sound/atmel/ac97c.c:743: error: negative width in bit-field ?<anonymous>?
> make[2]: *** [sound/atmel/ac97c.o] Error 1
>
>
> b.g.
>

AC97 audio broken on SAM9263EK?

[Yegor Yefremov]

> http://mailman.alsa-project.org/pipermail/alsa-devel/2010-March/025709.html

I tried your patch, but mpg123 says:

Playing MPEG stream 1 of 1: 05_Tabor_Romance.mp3 ...
Title:                                   Artist:
Comment:                                 Album:
Year:                                    Genre:  Other
MPEG 1.0 layer III, 128 kbit/s, 44100 Hz stereo
[audio.c:532] error: Error in writing audio (Input/output error?)!
[mpg123.c:557] error: Deep trouble! Cannot flush to my output anymore!

I'm using kernel 2.6.33 on the at91sam9g45-ek. For sound I use patches 0039 and 0044 from
ftp://ftp.linux4sam.org/devel/linux/2630_lnx4sam_patchset-tmp_20090828.tgz. With these I get proper sound from my head phones.

Regards,
Yegor

AC97 audio broken on SAM9263EK?

[Sedji Gaouaou]

Hi Yegor,

Le 3/9/2010 3:11 PM, Yegor Yefremov a ?crit :
>> http://mailman.alsa-project.org/pipermail/alsa-devel/2010-March/025709.html
>
> I tried your patch, but mpg123 says:
>
> Playing MPEG stream 1 of 1: 05_Tabor_Romance.mp3 ...
> Title:                                   Artist:
> Comment:                                 Album:
> Year:                                    Genre:  Other
> MPEG 1.0 layer III, 128 kbit/s, 44100 Hz stereo
> [audio.c:532] error: Error in writing audio (Input/output error?)!
> [mpg123.c:557] error: Deep trouble! Cannot flush to my output anymore!
>
> I'm using kernel 2.6.33 on the at91sam9g45-ek. For sound I use patches 0039 and 0044 from
> ftp://ftp.linux4sam.org/devel/linux/2630_lnx4sam_patchset-tmp_20090828.tgz. With these I get proper sound from my head phones.
>
> Regards,
> Yegor
>
I never tried with mpg123. But I just diff the 2 sets of patches, and 
the only difference I can see is that the one on linux4sam doesn't 
handle full duplex, whereas the one I posted on the alsa mailing does...

Yoy can find enclosed both files, may I ask you to test with both?
If it fails then I guess there is something wrong with the way the full 
duplex is done(even though I have tested it on my board...)

Regards,
Sedji


-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: ac97c_linux4sam.c
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20100309/f3483e6b/attachment-0002.c>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: ac97c_alsa.c
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20100309/f3483e6b/attachment-0003.c>

AC97 audio broken on SAM9263EK?

[Yegor Yefremov]

Hi Sedji,

Sedji Gaouaou wrote:
> Le 3/9/2010 3:11 PM, Yegor Yefremov a ?crit :
>>> http://mailman.alsa-project.org/pipermail/alsa-devel/2010-March/025709.html
>>>
>>
>> I tried your patch, but mpg123 says:
>>
>> Playing MPEG stream 1 of 1: 05_Tabor_Romance.mp3 ...
>> Title:                                   Artist:
>> Comment:                                 Album:
>> Year:                                    Genre:  Other
>> MPEG 1.0 layer III, 128 kbit/s, 44100 Hz stereo
>> [audio.c:532] error: Error in writing audio (Input/output error?)!
>> [mpg123.c:557] error: Deep trouble! Cannot flush to my output anymore!
>>
>> I'm using kernel 2.6.33 on the at91sam9g45-ek. For sound I use patches
>> 0039 and 0044 from
>> ftp://ftp.linux4sam.org/devel/linux/2630_lnx4sam_patchset-tmp_20090828.tgz.
>> With these I get proper sound from my head phones.
>>
>> Regards,
>> Yegor
>>
> I never tried with mpg123. But I just diff the 2 sets of patches, and
> the only difference I can see is that the one on linux4sam doesn't
> handle full duplex, whereas the one I posted on the alsa mailing does...
> 
> Yoy can find enclosed both files, may I ask you to test with both?
> If it fails then I guess there is something wrong with the way the full
> duplex is done(even though I have tested it on my board...)

I tried both files and the one really working is ac97c_alsa.c. As far as I understand it is combination of both patches for at91 support and full duplex?

I've tried the following applications under Debian 5.0:

mpg123
vlc
mplayer
amarok

Regards,
Yegor

AC97 audio broken on SAM9263EK?

[Sedji Gaouaou]

Hi,
>
> I tried both files and the one really working is ac97c_alsa.c. As far as I understand it is combination of both patches for at91 support and full duplex?
>
> I've tried the following applications under Debian 5.0:
>
> mpg123
> vlc
> mplayer
> amarok
>
> Regards,
> Yegor
>

ac97c_alsa.c is the result of the 2 patches I have send to the alsa 
mailing list 
(http://mailman.alsa-project.org/pipermail/alsa-devel/2010-March/025709.html 
and 
http://mailman.alsa-project.org/pipermail/alsa-devel/2010-February/025628.html). 
It indeed adds support for AT91(1/2) and full duplex(2/2).But is is 
strange that if you only apply only the first patch you encounter errors...

On my side I have only tested these patches with mplayer and aplay.
Plus to test the full duplex I have done like this:
"arecord -d 10 -f cd -t wav test.wav & mplayer my_file.mp3"

Regards,
Sedji

AC97 audio broken on SAM9263EK?

[Yegor Yefremov]

>> I tried both files and the one really working is ac97c_alsa.c. As far
>> as I understand it is combination of both patches for at91 support and
>> full duplex?
>>
>> I've tried the following applications under Debian 5.0:
>>
>> mpg123
>> vlc
>> mplayer
>> amarok
>>
>> Regards,
>> Yegor
>>
> 
> ac97c_alsa.c is the result of the 2 patches I have send to the alsa
> mailing list
> (http://mailman.alsa-project.org/pipermail/alsa-devel/2010-March/025709.html
> and
> http://mailman.alsa-project.org/pipermail/alsa-devel/2010-February/025628.html).
> It indeed adds support for AT91(1/2) and full duplex(2/2).But is is
> strange that if you only apply only the first patch you encounter errors...
> 
> On my side I have only tested these patches with mplayer and aplay.
> Plus to test the full duplex I have done like this:
> "arecord -d 10 -f cd -t wav test.wav & mplayer my_file.mp3"

When do you think the driver could hit the mainline kernel? Till 2.6.35 open window?

Best regards,
Yegor

AC97 audio broken on SAM9263EK?

[Sedji Gaouaou]

Hi,
>
> When do you think the driver could hit the mainline kernel? Till 2.6.35 open window?
>
> Best regards,
> Yegor
>

It has been acked, but I don't know when it will be in the mainline kernel.

Regards,
Sedji

AC97 audio broken on SAM9263EK?

[Yegor Yefremov]

Hi,

>> When do you think the driver could hit the mainline kernel? Till
>> 2.6.35 open window?
> 
> It has been acked, but I don't know when it will be in the mainline kernel.

have just tested 2.6.35-rc1 with mainline ac97 driver and found this issue. During playback I see following output on the console:

...
atmel_ac97c atmel_ac97c.0: channel A event
atmel_ac97c atmel_ac97c.0: channel A event
atmel_ac97c atmel_ac97c.0: channel A event
...

If I see it right these, lines come from sound/atmel/ac97c.c->atmel_ac97c_interrupt():

dev_info(&chip->pdev->dev, "channel A event%s%s%s%s%s%s\n",
                                casr & AC97C_CSR_OVRUN   ? " OVRUN"   : "",
                                casr & AC97C_CSR_RXRDY   ? " RXRDY"   : "",
                                casr & AC97C_CSR_UNRUN   ? " UNRUN"   : "",
                                casr & AC97C_CSR_TXEMPTY ? " TXEMPTY" : "",
                                casr & AC97C_CSR_TXRDY   ? " TXRDY"   : "",
                                !casr                    ? " NONE"    : "");

 

It is necessary to output this message in normal mode or is it a debug message?

Regards,
Yegor

AC97 audio broken on SAM9263EK?

[Sedji Gaouaou]

Hi Yegor,

It is an error indeed, it should be a debug message.

Regards,
Sedji

Le 6/1/2010 3:39 PM, Yegor Yefremov a ?crit :
> Hi,
>
>>> When do you think the driver could hit the mainline kernel? Till
>>> 2.6.35 open window?
>>
>> It has been acked, but I don't know when it will be in the mainline kernel.
>
> have just tested 2.6.35-rc1 with mainline ac97 driver and found this issue. During playback I see following output on the console:
>
> ...
> atmel_ac97c atmel_ac97c.0: channel A event
> atmel_ac97c atmel_ac97c.0: channel A event
> atmel_ac97c atmel_ac97c.0: channel A event
> ...
>
> If I see it right these, lines come from sound/atmel/ac97c.c->atmel_ac97c_interrupt():
>
> dev_info(&chip->pdev->dev, "channel A event%s%s%s%s%s%s\n",
>                                  casr&  AC97C_CSR_OVRUN   ? " OVRUN"   : "",
>                                  casr&  AC97C_CSR_RXRDY   ? " RXRDY"   : "",
>                                  casr&  AC97C_CSR_UNRUN   ? " UNRUN"   : "",
>                                  casr&  AC97C_CSR_TXEMPTY ? " TXEMPTY" : "",
>                                  casr&  AC97C_CSR_TXRDY   ? " TXRDY"   : "",
>                                  !casr                    ? " NONE"    : "");
>
>
>
> It is necessary to output this message in normal mode or is it a debug message?
>
> Regards,
> Yegor
>

AC97 audio broken on SAM9263EK?

[Yegor Yefremov]

Hi Sedji,

> It is an error indeed, it should be a debug message.

here is a patch:

atmel: set "channel A event" output to debug

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>

Index: b/sound/atmel/ac97c.c
===================================================================
--- a/sound/atmel/ac97c.c	2010-06-01 15:28:23.000000000 +0200
+++ b/sound/atmel/ac97c.c	2010-06-03 15:20:43.000000000 +0200
@@ -657,7 +657,7 @@
 	if (sr & AC97C_SR_CAEVT) {
 		struct snd_pcm_runtime *runtime;
 		int offset, next_period, block_size;
-		dev_info(&chip->pdev->dev, "channel A event%s%s%s%s%s%s\n",
+		dev_dbg(&chip->pdev->dev, "channel A event%s%s%s%s%s%s\n",
 				casr & AC97C_CSR_OVRUN   ? " OVRUN"   : "",
 				casr & AC97C_CSR_RXRDY   ? " RXRDY"   : "",
 				casr & AC97C_CSR_UNRUN   ? " UNRUN"   : "",