[PATCH 0/3] ASoC: multi-comp: OMAP platform fixes

[PATCH 0/3] ASoC: multi-comp: OMAP platform fixes

[Peter Ujfalusi]

Hello,

The following series introduces some cleanups for OMAP platforms in
multi-component branch.

- Enable audio for OMAP1 class
- Registers only the available amount of omap-mcbsp-dai devices for OMAP2,3,
  and 4

Please do verify, if the chages are correct, especially the number of McBSP
ports on different devices.

Liam: in case you take this, you can squash this to the corresponding commit.

Peter

---
Peter Ujfalusi (3):
  ASoC: multi-comp: OMAP McBSP: use macro for platform_device
  ASoC: multi-comp: OMAP McBSP: Enable audio on OMAP1 platform
  ASoC: multi-comp: OMAP2,3,4 McBSP: register correct number of ports

 arch/arm/mach-omap1/devices.c           |   24 +++++++++++++++
 arch/arm/mach-omap2/devices.c           |   50 ++++++++++++++++--------------
 arch/arm/plat-omap/include/plat/mcbsp.h |    7 ++++
 3 files changed, 58 insertions(+), 23 deletions(-)

--
1.7.2

[PATCH 1/3] ASoC: multi-comp: OMAP McBSP: use macro for platform_device

[Peter Ujfalusi]

Use macro (and define it in the plat/mcbsp.h) to create
the omap-mcbsp-dai devices.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
---
 arch/arm/mach-omap2/devices.c           |   29 +++++------------------------
 arch/arm/plat-omap/include/plat/mcbsp.h |    7 +++++++
 2 files changed, 12 insertions(+), 24 deletions(-)

diff --git a/arch/arm/mach-omap2/devices.c b/arch/arm/mach-omap2/devices.c
index 088b050..758d39b 100644
--- a/arch/arm/mach-omap2/devices.c
+++ b/arch/arm/mach-omap2/devices.c
@@ -297,30 +297,11 @@ static struct platform_device omap_pcm = {
 	.id	= -1,
 };
 
-static struct platform_device omap_mcbsp1 = {
-	.name	= "omap-mcbsp-dai",
-	.id	= OMAP_MCBSP1,
-};
-
-static struct platform_device omap_mcbsp2 = {
-	.name	= "omap-mcbsp-dai",
-	.id	= OMAP_MCBSP2,
-};
-
-static struct platform_device omap_mcbsp3 = {
-	.name	= "omap-mcbsp-dai",
-	.id	= OMAP_MCBSP3,
-};
-
-static struct platform_device omap_mcbsp4 = {
-	.name	= "omap-mcbsp-dai",
-	.id	= OMAP_MCBSP4,
-};
-
-static struct platform_device omap_mcbsp5 = {
-	.name	= "omap-mcbsp-dai",
-	.id	= OMAP_MCBSP5,
-};
+OMAP_MCBSP_PLATFORM_DEVICE(1);
+OMAP_MCBSP_PLATFORM_DEVICE(2);
+OMAP_MCBSP_PLATFORM_DEVICE(3);
+OMAP_MCBSP_PLATFORM_DEVICE(4);
+OMAP_MCBSP_PLATFORM_DEVICE(5);
 
 static void omap_init_audio(void)
 {
diff --git a/arch/arm/plat-omap/include/plat/mcbsp.h b/arch/arm/plat-omap/include/plat/mcbsp.h
index b4ff6a1..5b20103 100644
--- a/arch/arm/plat-omap/include/plat/mcbsp.h
+++ b/arch/arm/plat-omap/include/plat/mcbsp.h
@@ -30,6 +30,13 @@
 #include <mach/hardware.h>
 #include <plat/clock.h>
 
+/* macro for building platform_device for McBSP ports */
+#define OMAP_MCBSP_PLATFORM_DEVICE(port_nr)		\
+static struct platform_device omap_mcbsp##port_nr = {	\
+	.name	= "omap-mcbsp-dai",			\
+	.id	= OMAP_MCBSP##port_nr,			\
+}
+
 #define OMAP7XX_MCBSP1_BASE	0xfffb1000
 #define OMAP7XX_MCBSP2_BASE	0xfffb1800
 
-- 
1.7.2

[PATCH 2/3] ASoC: multi-comp: OMAP McBSP: Enable audio on OMAP1 platform

[Peter Ujfalusi]

OMAP1 class has 3 McBSP ports, if I'm not mistaken.
Enable the audio with multi-comp on OMAP1 class.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
---
 arch/arm/mach-omap1/devices.c |   24 ++++++++++++++++++++++++
 1 files changed, 24 insertions(+), 0 deletions(-)

diff --git a/arch/arm/mach-omap1/devices.c b/arch/arm/mach-omap1/devices.c
index 379100c..523bbc4 100644
--- a/arch/arm/mach-omap1/devices.c
+++ b/arch/arm/mach-omap1/devices.c
@@ -267,6 +267,29 @@ static inline void omap_init_sti(void)
 static inline void omap_init_sti(void) {}
 #endif
 
+#if defined(CONFIG_SND_SOC) || defined(CONFIG_SND_SOC_MODULE)
+
+static struct platform_device omap_pcm = {
+	.name	= "omap-pcm-audio",
+	.id	= -1,
+};
+
+OMAP_MCBSP_PLATFORM_DEVICE(1);
+OMAP_MCBSP_PLATFORM_DEVICE(2);
+OMAP_MCBSP_PLATFORM_DEVICE(3);
+
+static void omap_init_audio(void)
+{
+	platform_device_register(&omap_mcbsp1);
+	platform_device_register(&omap_mcbsp2);
+	platform_device_register(&omap_mcbsp3);
+	platform_device_register(&omap_pcm);
+}
+
+#else
+static inline void omap_init_audio(void) {}
+#endif
+
 /*-------------------------------------------------------------------------*/
 
 /*
@@ -299,6 +322,7 @@ static int __init omap1_init_devices(void)
 	omap_init_rtc();
 	omap_init_spi100k();
 	omap_init_sti();
+	omap_init_audio();
 
 	return 0;
 }
-- 
1.7.2

[PATCH 3/3] ASoC: multi-comp: OMAP2, 3, 4 McBSP: register correct number of ports

[Peter Ujfalusi]

Different OMAPs has different number of McBSP ports...
OMAP2420 has 2 McBSP ports
OMAP2430 has 5 McBSP ports
OMAP3 has 5 McBSP ports
OMAP4 has 4 McBSP ports

Since I don't have access to OMAP4 TRM, the number of ports on OMAP4
is guessed based on the original OMAP4 related McBSP code...

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
---
 arch/arm/mach-omap2/devices.c |   23 +++++++++++++++++++++++
 1 files changed, 23 insertions(+), 0 deletions(-)

diff --git a/arch/arm/mach-omap2/devices.c b/arch/arm/mach-omap2/devices.c
index 758d39b..d78cf6a 100644
--- a/arch/arm/mach-omap2/devices.c
+++ b/arch/arm/mach-omap2/devices.c
@@ -297,19 +297,42 @@ static struct platform_device omap_pcm = {
 	.id	= -1,
 };
 
+/*
+ * OMAP2420 has 2 McBSP ports
+ * OMAP2430 has 5 McBSP ports
+ * OMAP3 has 5 McBSP ports
+ * OMAP4 has 4 McBSP ports
+ */
+#if defined(CONFIG_ARCH_OMAP4) ||		\
+    defined(CONFIG_ARCH_OMAP3) ||	\
+    defined(CONFIG_ARCH_OMAP2430)
+#define OMAP_HAS_4_MCBSP_PORTS
+#endif
+#if defined(CONFIG_ARCH_OMAP3) || defined(CONFIG_ARCH_OMAP2430)
+#define OMAP_HAS_5_MCBSP_PORTS
+#endif
+
 OMAP_MCBSP_PLATFORM_DEVICE(1);
 OMAP_MCBSP_PLATFORM_DEVICE(2);
+#ifdef OMAP_HAS_4_MCBSP_PORTS
 OMAP_MCBSP_PLATFORM_DEVICE(3);
 OMAP_MCBSP_PLATFORM_DEVICE(4);
+#endif
+#ifdef OMAP_HAS_5_MCBSP_PORTS
 OMAP_MCBSP_PLATFORM_DEVICE(5);
+#endif
 
 static void omap_init_audio(void)
 {
 	platform_device_register(&omap_mcbsp1);
 	platform_device_register(&omap_mcbsp2);
+#ifdef OMAP_HAS_4_MCBSP_PORTS
 	platform_device_register(&omap_mcbsp3);
 	platform_device_register(&omap_mcbsp4);
+#endif
+#ifdef OMAP_HAS_5_MCBSP_PORTS
 	platform_device_register(&omap_mcbsp5);
+#endif
 	platform_device_register(&omap_pcm);
 }
 
-- 
1.7.2

Re: [PATCH 2/3] ASoC: multi-comp: OMAP McBSP: Enable audio on OMAP1 platform

[Peter Ujfalusi]

On Friday 06 August 2010 10:39:14 Ujfalusi Peter (Nokia-MS/Tampere) wrote:
> OMAP1 class has 3 McBSP ports, if I'm not mistaken.
> Enable the audio with multi-comp on OMAP1 class.
> 
> Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
> ---
>  arch/arm/mach-omap1/devices.c |   24 ++++++++++++++++++++++++
>  1 files changed, 24 insertions(+), 0 deletions(-)
> 
> diff --git a/arch/arm/mach-omap1/devices.c b/arch/arm/mach-omap1/devices.c
> index 379100c..523bbc4 100644
> --- a/arch/arm/mach-omap1/devices.c
> +++ b/arch/arm/mach-omap1/devices.c
> @@ -267,6 +267,29 @@ static inline void omap_init_sti(void)
>  static inline void omap_init_sti(void) {}
>  #endif

Hmm,

 +#include <plat/mcbsp.h>

Might help at compile time ;)

Liam, would you be so kind to fix this is you take the series.
I could as well resend...

Thanks,
Péter

> 
> +#if defined(CONFIG_SND_SOC) || defined(CONFIG_SND_SOC_MODULE)
> +
> +static struct platform_device omap_pcm = {
> +	.name	= "omap-pcm-audio",
> +	.id	= -1,
> +};
> +
> +OMAP_MCBSP_PLATFORM_DEVICE(1);
> +OMAP_MCBSP_PLATFORM_DEVICE(2);
> +OMAP_MCBSP_PLATFORM_DEVICE(3);
> +
> +static void omap_init_audio(void)
> +{
> +	platform_device_register(&omap_mcbsp1);
> +	platform_device_register(&omap_mcbsp2);
> +	platform_device_register(&omap_mcbsp3);
> +	platform_device_register(&omap_pcm);
> +}
> +
> +#else
> +static inline void omap_init_audio(void) {}
> +#endif
> +
>  /*------------------------------------------------------------------------
> -*/
> 
>  /*
> @@ -299,6 +322,7 @@ static int __init omap1_init_devices(void)
>  	omap_init_rtc();
>  	omap_init_spi100k();
>  	omap_init_sti();
> +	omap_init_audio();
> 
>  	return 0;
>  }

Re: [PATCH 0/3] ASoC: multi-comp: OMAP platform fixes

[Jarkko Nikula]

Hi, notes for Arun and Janusz below.

On Fri,  6 Aug 2010 10:39:12 +0300
Peter Ujfalusi <peter.ujfalusi@nokia.com> wrote:

> Hello,
> 
> The following series introduces some cleanups for OMAP platforms in
> multi-component branch.
> 
> - Enable audio for OMAP1 class

There's a on-going development going on to update ASoC framework. This
multi-compont work requires major changes and our problem is that we
don't have any OMAP1 boards to test these changes.

We'd glad if you could test these changes with Peter's patches on OMAP1
board.

The ASoC multi-component development happens in
origin/topic/multi-component branch of this tree:

git://git.kernel.org/pub/scm/linux/kernel/git/lrg/asoc-2.6.git


-- 
Jarkko

Re: [PATCH 2/3] ASoC: multi-comp: OMAP McBSP: Enable audio on OMAP1 platform

[Peter Ujfalusi]

On Friday 06 August 2010 10:50:09 Ujfalusi Peter (Nokia-MS/Tampere) wrote:
> Liam, would you be so kind to fix this is you take the series.
> I could as well resend...

I have found one cosmetic problem in the mach-omap2/devices.c changes.
I'll resend the series.

BTW: it seams Arun's email address has been changed (arunks@mistralsolutions.com 
is not alive)...

-- 
Péter

Re: [PATCH 3/3] ASoC: multi-comp: OMAP2, 3, 4 McBSP: register correct number of ports

[Jarkko Nikula]

On Fri,  6 Aug 2010 10:39:15 +0300
Peter Ujfalusi <peter.ujfalusi@nokia.com> wrote:

>  static void omap_init_audio(void)
>  {
>  	platform_device_register(&omap_mcbsp1);
>  	platform_device_register(&omap_mcbsp2);
> +#ifdef OMAP_HAS_4_MCBSP_PORTS
>  	platform_device_register(&omap_mcbsp3);
>  	platform_device_register(&omap_mcbsp4);
> +#endif
> +#ifdef OMAP_HAS_5_MCBSP_PORTS
>  	platform_device_register(&omap_mcbsp5);
> +#endif
>  	platform_device_register(&omap_pcm);
>  }
>  
This must be done at runtime since we can have multi omap builds now.
So use cpu_is_omapxxx instead of conditional compilation. Simplifies
patch too :-)


-- 
Jarkko

Re: [PATCH 2/3] ASoC: multi-comp: OMAP McBSP: Enable audio on OMAP1 platform

[Jarkko Nikula]

On Fri,  6 Aug 2010 10:39:14 +0300
Peter Ujfalusi <peter.ujfalusi@nokia.com> wrote:

> OMAP1 class has 3 McBSP ports, if I'm not mistaken.
> Enable the audio with multi-comp on OMAP1 class.
> 
...
> +static void omap_init_audio(void)
> +{
> +	platform_device_register(&omap_mcbsp1);
> +	platform_device_register(&omap_mcbsp2);
> +	platform_device_register(&omap_mcbsp3);
> +	platform_device_register(&omap_pcm);
> +}
> +
Not for sure but according to code only omap15xx and 16xx have three
ports, 7xx two and I'm unsure about 8xx. At least cpu_is_omap7xx seems
to be valid for both 730 and 850.


-- 
Jarkko

Re: [PATCH 3/3] ASoC: multi-comp: OMAP2, 3, 4 McBSP: register correct number of ports

[Peter Ujfalusi]

On Friday 06 August 2010 11:08:08 ext Jarkko Nikula wrote:
> This must be done at runtime since we can have multi omap builds now.
> So use cpu_is_omapxxx instead of conditional compilation. Simplifies
> patch too :-)

Yes, you are right, I'll do that.
I'll do the same with OMAP1 (2 vs 3 ports)..

Thanks,
Péter

Re: [PATCH 0/3] ASoC: multi-comp: OMAP platform fixes

[Mark Brown]

On Fri, Aug 06, 2010 at 10:39:12AM +0300, Peter Ujfalusi wrote:

> The following series introduces some cleanups for OMAP platforms in
> multi-component branch.
> 
> - Enable audio for OMAP1 class
> - Registers only the available amount of omap-mcbsp-dai devices for OMAP2,3,
>   and 4
> 
> Please do verify, if the chages are correct, especially the number of McBSP
> ports on different devices.

Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

from a code point of view only (ie, I've no idea if the definitions are
correct for any particular OMAP).

Re: [PATCH 0/3] ASoC: multi-comp: OMAP platform fixes

[Arun KS]

On Fri, Aug 6, 2010 at 4:56 PM, Jarkko Nikula <jhnikula@gmail.com> wrote:
> Hi, notes for Arun and Janusz below.
>
> On Fri,  6 Aug 2010 10:39:12 +0300
> Peter Ujfalusi <peter.ujfalusi@nokia.com> wrote:
>
>> Hello,
>>
>> The following series introduces some cleanups for OMAP platforms in
>> multi-component branch.
>>
>> - Enable audio for OMAP1 class
>
> There's a on-going development going on to update ASoC framework. This
> multi-compont work requires major changes and our problem is that we
> don't have any OMAP1 boards to test these changes.
>
> We'd glad if you could test these changes with Peter's patches on OMAP1
> board.
Jarkko,

I m sorry to inform you that I m not having access to OSK anymore. It
was with my previous employer.

Arun
>
> The ASoC multi-component development happens in
> origin/topic/multi-component branch of this tree:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/lrg/asoc-2.6.git
>
>
> --
> Jarkko
> _______________________________________________
> Alsa-devel mailing list
> Alsa-devel@alsa-project.org
> http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
>

Re: [PATCH 0/3] ASoC: multi-comp: OMAP platform fixes

[Janusz Krzysztofik]

Friday 06 August 2010 09:56:28 Jarkko Nikula wrote:
> Hi, notes for Arun and Janusz below.
>
> On Fri,  6 Aug 2010 10:39:12 +0300
>
> Peter Ujfalusi <peter.ujfalusi@nokia.com> wrote:
> > Hello,
> >
> > The following series introduces some cleanups for OMAP platforms in
> > multi-component branch.
> >
> > - Enable audio for OMAP1 class
>
> There's a on-going development going on to update ASoC framework. This
> multi-compont work requires major changes and our problem is that we
> don't have any OMAP1 boards to test these changes.
>
> We'd glad if you could test these changes with Peter's patches on OMAP1
> board.
>
> The ASoC multi-component development happens in
> origin/topic/multi-component branch of this tree:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/lrg/asoc-2.6.git

Hi,
If your request for testing Peter's patches is still valid, what exact tree 
and branch would you like me to test the patches against? I've tried to 
locate a tree with the patches already applied by Liam, but failed.

Meanwhile, I've verified that the patches, or their OMAP1 related bits, are 
harmless when applied against (almost) current l-o master. However, I'm not 
sure if testing the new platform devices not actually used by any 
corresponding platform drivers is what you expected.

Thanks,
Janusz

Re: [PATCH 0/3] ASoC: multi-comp: OMAP platform fixes

[Jarkko Nikula]

Hi

On Mon, 9 Aug 2010 19:09:57 +0200
Janusz Krzysztofik <jkrzyszt@tis.icnet.pl> wrote:

> If your request for testing Peter's patches is still valid, what exact tree 
> and branch would you like me to test the patches against? I've tried to 
> locate a tree with the patches already applied by Liam, but failed.
> 
Sure, we are glad if you can give a test to ASoC multi-component as
there are major changes coming and we don't have any omap1 boards to
try.

Peter's patches are already integrated into TI OMAP update of
multi-component so for you it's just enough to clone the development
tree and then going to topic/multi-component branch of it.

git://git.kernel.org/pub/scm/linux/kernel/git/lrg/asoc-2.6.git
and
git checkout -b multi-component_20100809 origin/topic/multi-component

There are no config changes so you should be able to test with your
regular .config. It should work as before if everything goes well.
There could be probing problems but usually fix is around the struct
snd_soc_dai_link in a machine driver (e.g. ams_delta_dai_link).


-- 
Jarkko

Re: [PATCH 0/3] ASoC: multi-comp: OMAP platform fixes

[Peter Ujfalusi]

Hello Janusz,

On Monday 09 August 2010 20:09:57 ext Janusz Krzysztofik wrote:
> 
> Hi,
> If your request for testing Peter's patches is still valid, what exact tree
> and branch would you like me to test the patches against? I've tried to
> locate a tree with the patches already applied by Liam, but failed.
> 
> Meanwhile, I've verified that the patches, or their OMAP1 related bits, are
> harmless when applied against (almost) current l-o master. However, I'm not
> sure if testing the new platform devices not actually used by any
> corresponding platform drivers is what you expected.

Since I do not have access to OMAP1 based boards, I can not test the patch 
which adds support for ASoC/multi-comp on OMAP1, I would be glad, if you could 
verify, if it is working on your board.

Liam has squashed the patches, so they are not visible.
In theory you can use your existing config, and the audio shall be working as 
it used to (before multi-comp).

It would be nice to know, if OMAP1 still works...

Thanks for taking time for this!
Péter