[alsa-devel] ASoC: TWL4030 driver support multi channel playback
Hi Peter,
I am using linux 2.6.28 kernel,
there is a requirement to support multi (4) channel play back .
I guess TWL4030 ASoC driver now has full fledged support for both the DAC1(TX1) and DAC2(TX2) paths in option 1 of codec operation.
I am looking forward to back port the ASoC stuff drivers from 2.6.30 rc7 to 2.6.28.
Is there any better way of meeting the requirement with out back porting the stuff
Suggestions please.
Sorry for the english,,
Hello,
On Monday 01 June 2009 17:20:01 ext Naveen Krishna Ch wrote:
Hi Peter,
I am using linux 2.6.28 kernel,
there is a requirement to support multi (4) channel play back .
I guess TWL4030 ASoC driver now has full fledged support for both the DAC1(TX1) and DAC2(TX2) paths in option 1 of codec operation.
Yes it does. It has support for TX1L/R, TX2L/R, RX1L/R and RX2L/R (four channel playback and recording)
I am looking forward to back port the ASoC stuff drivers from 2.6.30 rc7 to 2.6.28.
Is there any better way of meeting the requirement with out back porting the stuff
Well, I think this is not going to be an easy ride. It is not that hard, but takes some time. I could think of two approaches to get the twl4030 codec work backported (plus one): 1. Backport the the whole ASoC changes to 2.6.28 2. Backport the twl4030 codec driver with minimal amount of ASoC changes 3. Convince whoever you need to convince that you need to move to 2.6.30 ;)
1. Since ASoC has been dramatically changed since 2.6.28 this is not going to be easy, but you will have the current ASoC features and it will be easier to follow upstream. 2. Pick the changes to the twl4030 codec one-by-one apply it, compile the result, if it breaks, take the needed patches to fix the errors. Fair amunt of work, but you will end up a working twl4030 codec, but you need to backport the new patches from upstream every time.. 3. This only exist in the ideal world.
In overall, I think your best bet is 1) You can take the 2) if you really know what you are doing...
Suggestions please.
Sorry for the english,,
Cheers (: Naveen Krishna Ch :)
Hi Peter,
Thanks for your quick reply, I have an issue, probably the last one.
2009/6/2 Peter Ujfalusi peter.ujfalusi@nokia.com
Hello,
On Monday 01 June 2009 17:20:01 ext Naveen Krishna Ch wrote:
Hi Peter,
I am using linux 2.6.28 kernel,
there is a requirement to support multi (4) channel play back .
I guess TWL4030 ASoC driver now has full fledged support for both the DAC1(TX1) and DAC2(TX2) paths in option 1 of codec operation.
Yes it does. It has support for TX1L/R, TX2L/R, RX1L/R and RX2L/R (four channel playback and recording)
I am looking forward to back port the ASoC stuff drivers from 2.6.30 rc7 to 2.6.28.
Is there any better way of meeting the requirement with out back porting the stuff
Well, I think this is not going to be an easy ride. It is not that hard, but takes some time. I could think of two approaches to get the twl4030 codec work backported (plus one):
- Backport the the whole ASoC changes to 2.6.28
- Backport the twl4030 codec driver with minimal amount of ASoC changes
I tried this last night, i was able to boot the board... Except the following thing
Advanced Linux Sound Architecture Driver Version 1.0.18rc3. No device for DAI twl4030 No device for DAI twl4030 Voice No device for DAI omap-mcbsp-dai-0 No device for DAI omap-mcbsp-dai-1 No device for DAI omap-mcbsp-dai-2 No device for DAI omap-mcbsp-dai-3 No device for DAI omap-mcbsp-dai-4 omap3apple SoC init ALSA device list: No soundcards found.
I tried to fix this, it seems like a dev structure is missing during the codec registration. Can u help me find where i went wrong..
- Convince whoever you need to convince that you need to move to 2.6.30 ;)
There is some powermanagment issue which restricting us to move forward to 2.6.30
- Since ASoC has been dramatically changed since 2.6.28 this is not going
to be easy, but you will have the current ASoC features and it will be easier to follow upstream. 2. Pick the changes to the twl4030 codec one-by-one apply it, compile the result, if it breaks, take the needed patches to fix the errors. Fair amunt of work, but you will end up a working twl4030 codec, but you need to backport the new patches from upstream every time.. 3. This only exist in the ideal world.
In overall, I think your best bet is 1) You can take the 2) if you really know what you are doing...
Suggestions please.
Sorry for the english,,
Cheers (: Naveen Krishna Ch :)
-- Péter
On Tuesday 02 June 2009 08:33:59 ext Naveen Krishna Ch wrote:
- Backport the twl4030 codec driver with minimal amount of ASoC changes
I tried this last night, i was able to boot the board... Except the following thing
Advanced Linux Sound Architecture Driver Version 1.0.18rc3. No device for DAI twl4030 No device for DAI twl4030 Voice No device for DAI omap-mcbsp-dai-0 No device for DAI omap-mcbsp-dai-1 No device for DAI omap-mcbsp-dai-2 No device for DAI omap-mcbsp-dai-3 No device for DAI omap-mcbsp-dai-4 omap3apple SoC init ALSA device list: No soundcards found.
I tried to fix this, it seems like a dev structure is missing during the codec registration. Can u help me find where i went wrong..
To resolve this you need to backport the changes in sound/soc/omap-* also and probably take a look at arch/arm/plat-omap/mcbsp.c as well.
On Tue, Jun 02, 2009 at 08:16:41AM +0300, Peter Ujfalusi wrote:
Well, I think this is not going to be an easy ride. It is not that hard, but takes some time. I could think of two approaches to get the twl4030 codec work backported (plus one):
Backport the the whole ASoC changes to 2.6.28
Backport the twl4030 codec driver with minimal amount of ASoC changes
Convince whoever you need to convince that you need to move to 2.6.30 ;)
Since ASoC has been dramatically changed since 2.6.28 this is not going to
be easy, but you will have the current ASoC features and it will be easier to follow upstream.
I'd recommend just backporting ASoC in its entirety - it is relatively well insulated against changes in the rest of the kernel other that any arch code the machine drivers rely on so this process normally goes fairly smoothly. It's much easier than trying a partial backport.
participants (3)
-
Mark Brown
-
Naveen Krishna Ch
-
Peter Ujfalusi