[PATCH] ASoC: atmel-pcm-dma: Remove SND_DMAENGINE_PCM_FLAG_NO_RESIDUE flag

[PATCH] ASoC: atmel-pcm-dma: Remove SND_DMAENGINE_PCM_FLAG_NO_RESIDUE flag

[Lars-Peter Clausen]

Since commit d48de6f1a ("DMA: AT91: Get residual bytes in dma buffer") the
at91-dma driver has support for residue reporting. So there is no need to
specify the SND_DMAENGINE_PCM_FLAG_NO_RESIDUE flag. This allows for a finer
grained resolution of the PCM pointer as well as avoids the race condition that
can occur with the period counting that is used when the dmaengine driver does
not support residue reporting.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>

---
I don't have the hardware, so only compile tested. It should work though if the
residue calculation in the dma driver is implemented correctly.
---
 sound/soc/atmel/atmel-pcm-dma.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/sound/soc/atmel/atmel-pcm-dma.c b/sound/soc/atmel/atmel-pcm-dma.c
index 06082e5..ba1366a 100644
--- a/sound/soc/atmel/atmel-pcm-dma.c
+++ b/sound/soc/atmel/atmel-pcm-dma.c
@@ -129,8 +129,7 @@ static const struct snd_dmaengine_pcm_config atmel_dmaengine_pcm_config = {
 
 int atmel_pcm_dma_platform_register(struct device *dev)
 {
-	return snd_dmaengine_pcm_register(dev, &atmel_dmaengine_pcm_config,
-			SND_DMAENGINE_PCM_FLAG_NO_RESIDUE);
+	return snd_dmaengine_pcm_register(dev, &atmel_dmaengine_pcm_config, 0);
 }
 EXPORT_SYMBOL(atmel_pcm_dma_platform_register);
 
-- 
1.8.0

Re: [PATCH] ASoC: atmel-pcm-dma: Remove SND_DMAENGINE_PCM_FLAG_NO_RESIDUE flag

[Bo Shen]

Hi Lars,

On 01/06/2014 10:30 PM, Lars-Peter Clausen wrote:
> Since commit d48de6f1a ("DMA: AT91: Get residual bytes in dma buffer") the
> at91-dma driver has support for residue reporting. So there is no need to
> specify the SND_DMAENGINE_PCM_FLAG_NO_RESIDUE flag. This allows for a finer
> grained resolution of the PCM pointer as well as avoids the race condition that
> can occur with the period counting that is used when the dmaengine driver does
> not support residue reporting.
>
> Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
>
> ---
> I don't have the hardware, so only compile tested. It should work though if the
> residue calculation in the dma driver is implemented correctly.
> ---
>   sound/soc/atmel/atmel-pcm-dma.c | 3 +--
>   1 file changed, 1 insertion(+), 2 deletions(-)

Thanks for your patch.

For the dma get residue code, we are improving it. So, I think we'd 
better waiting the improvement done and after let this patch go into 
mainline.

What do you think?

> diff --git a/sound/soc/atmel/atmel-pcm-dma.c b/sound/soc/atmel/atmel-pcm-dma.c
> index 06082e5..ba1366a 100644
> --- a/sound/soc/atmel/atmel-pcm-dma.c
> +++ b/sound/soc/atmel/atmel-pcm-dma.c
> @@ -129,8 +129,7 @@ static const struct snd_dmaengine_pcm_config atmel_dmaengine_pcm_config = {
>
>   int atmel_pcm_dma_platform_register(struct device *dev)
>   {
> -	return snd_dmaengine_pcm_register(dev, &atmel_dmaengine_pcm_config,
> -			SND_DMAENGINE_PCM_FLAG_NO_RESIDUE);
> +	return snd_dmaengine_pcm_register(dev, &atmel_dmaengine_pcm_config, 0);
>   }
>   EXPORT_SYMBOL(atmel_pcm_dma_platform_register);
>
>

Best Regards,
Bo Shen

Re: [PATCH] ASoC: atmel-pcm-dma: Remove SND_DMAENGINE_PCM_FLAG_NO_RESIDUE flag

[Lars-Peter Clausen]

On 01/07/2014 10:05 AM, Bo Shen wrote:
> Hi Lars,
> 
> On 01/06/2014 10:30 PM, Lars-Peter Clausen wrote:
>> Since commit d48de6f1a ("DMA: AT91: Get residual bytes in dma buffer") the
>> at91-dma driver has support for residue reporting. So there is no need to
>> specify the SND_DMAENGINE_PCM_FLAG_NO_RESIDUE flag. This allows for a finer
>> grained resolution of the PCM pointer as well as avoids the race condition
>> that
>> can occur with the period counting that is used when the dmaengine driver
>> does
>> not support residue reporting.
>>
>> Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
>>
>> ---
>> I don't have the hardware, so only compile tested. It should work though
>> if the
>> residue calculation in the dma driver is implemented correctly.
>> ---
>>   sound/soc/atmel/atmel-pcm-dma.c | 3 +--
>>   1 file changed, 1 insertion(+), 2 deletions(-)
> 
> Thanks for your patch.
> 
> For the dma get residue code, we are improving it. So, I think we'd better
> waiting the improvement done and after let this patch go into mainline.
> 
> What do you think?

If the residue reporting in the dmaengine driver doesn't work properly yet
we have to wait.

- Lars