Re: [PATCH] ASoC: tegra: use dmaengine based dma driver

[Laxman Dewangan]

On Friday 29 June 2012 10:32 PM, Stephen Warren wrote:
> On 06/29/2012 05:34 AM, Laxman Dewangan wrote:
> > Use the dmaengine based Tegra APB DMA driver for
> > data transfer between SPI fifo and memory in
> > place of legacy Tegra APB DMA.
> >
> > Because generic soc-dmaengine-pcm uses the DMAs API
> > based on dmaengine, using the exported APIs provided
> > by this generic driver.
> >
> > The new driver is selected if legacy driver is not
> > selected and new dma driver is enabled through config
> > file.
> This works just fine with the existing non-dmaengine DMA driver enabled.
>
> However, I can't get it to work with dmaengine:
>
> > # aplay ~/abba-dq-48000-stereo.wav
> > [  151.613476] tegra20-i2s tegra20-i2s.0: dmaengine pcm open failed with err -6
> > [  151.620557] tegra20-i2s tegra20-i2s.0: can't open platform tegra20-i2s.0: -6
> > aplay: main:654: audio open error: No such device or address

With the error, it seems that dmachannel is not getting allocated.

> I do have the following in my local tree:
> 68a67b8 ARM: tegra: add device tree AUXDATA for APBDMA
> 0db7a96 ARM: tegra: dma: rename driver name for clock to "tegra-apbdma"

Some dma changes which I sent and already on linux-next are also require 
with the above change.
Alsong with that I have local change like to rename the dma driver to 
compatible with dts files (remove -new from driver).

I think I should send that patches so that you can test it by:
- Taking already applied dma driver change in linux-next.
- Apply my new patches which I am going to send.

And do local change in the tegra_defconfig to disable SYSTEM_DMA and 
enable dmaengine based dma driver.


> I also fixed the compatible values in drivers/dma/tegra20-apb-dma.c so
> the driver would get instantiated, which it does;
> /sys/devices/tegra-apbdma/dma has a bunch of dmaengine channels in it.

Possibly some patches in dma driver  which is already applied in next is 
not available in your tree. CYCLIC_DMA patch is important.

> (Note: This is on Ventana, although I doubt that makes much difference)

I tested on cardhu only but it should not matter.

> Is there something else I need to do to test this?

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/