[PATCH 0/2] dma: edma: Allow to disable eDMA IRQ during cyclic transfer

[PATCH 0/2] dma: edma: Allow to disable eDMA IRQ during cyclic transfer

[Peter Ujfalusi]

Hi,

After this series clients can ask to not receive notifications after each period.
In this case we can disable the completion interrupt since the position reporting
does not rely on it for cyclic mode.
Patchset for ASoC part has been sent which allows users space to take adventage
of SNDRV_PCM_INFO_NO_PERIOD_WAKEUP:
[1] http://mailman.alsa-project.org/pipermail/alsa-devel/2014-July/078993.html

When both series applied on top of linux-next:

# cache the audio file in memory
cat some_music.mp3 > /dev/null
# start PA
pulseaudio -v
# in another terminal, play the music via PA
mplayer -ao pulse some_music.mp3
# yet another terminal to monitor the interrupts
watch cat /proc/interrupts
# note the non increasing number of edma interrupts ;)

Regards,
Peter
---
Peter Ujfalusi (2):
  dma: edma: Update caps->residue_granularity to match with reality
  dma: edma: Support to suppress the period interrupts in cyclic mode

 drivers/dma/edma.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

-- 
2.0.0

[PATCH 0/2] dma: edma: Allow to disable eDMA IRQ during cyclic transfer

[Peter Ujfalusi]

Hi,

After this series clients can ask to not receive notifications after each period.
In this case we can disable the completion interrupt since the position reporting
does not rely on it for cyclic mode.
Patchset for ASoC part has been sent which allows users space to take adventage
of SNDRV_PCM_INFO_NO_PERIOD_WAKEUP:
[1] http://mailman.alsa-project.org/pipermail/alsa-devel/2014-July/078993.html

When both series applied on top of linux-next:

# cache the audio file in memory
cat some_music.mp3 > /dev/null
# start PA
pulseaudio -v
# in another terminal, play the music via PA
mplayer -ao pulse some_music.mp3
# yet another terminal to monitor the interrupts
watch cat /proc/interrupts
# note the non increasing number of edma interrupts ;)

Regards,
Peter
---
Peter Ujfalusi (2):
  dma: edma: Update caps->residue_granularity to match with reality
  dma: edma: Support to suppress the period interrupts in cyclic mode

 drivers/dma/edma.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

-- 
2.0.0

[PATCH 2/2] dma: edma: Support to suppress the period interrupts in cyclic mode

[Peter Ujfalusi]

If the client (audio) does not request interrupts for every period we can
disable them.
With updated audio driver stack we can play audio w/o the need to process
any edma interrupts.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
---
 drivers/dma/edma.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/dma/edma.c b/drivers/dma/edma.c
index d5017abcf7cb..69f1f5250120 100644
--- a/drivers/dma/edma.c
+++ b/drivers/dma/edma.c
@@ -723,10 +723,10 @@ static struct dma_async_tx_descriptor *edma_prep_dma_cyclic(
 		edesc->absync = ret;
 
 		/*
-		 * Enable interrupts for every period because callback
-		 * has to be called for every period.
+		 * Enable period interrupt only if it is requested
 		 */
-		edesc->pset[i].param.opt |= TCINTEN;
+		if (tx_flags & DMA_PREP_INTERRUPT)
+			edesc->pset[i].param.opt |= TCINTEN;
 	}
 
 	/* Place the cyclic channel to highest priority queue */
-- 
2.0.0

[PATCH 2/2] dma: edma: Support to suppress the period interrupts in cyclic mode

[Peter Ujfalusi]

If the client (audio) does not request interrupts for every period we can
disable them.
With updated audio driver stack we can play audio w/o the need to process
any edma interrupts.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
---
 drivers/dma/edma.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/dma/edma.c b/drivers/dma/edma.c
index d5017abcf7cb..69f1f5250120 100644
--- a/drivers/dma/edma.c
+++ b/drivers/dma/edma.c
@@ -723,10 +723,10 @@ static struct dma_async_tx_descriptor *edma_prep_dma_cyclic(
 		edesc->absync = ret;
 
 		/*
-		 * Enable interrupts for every period because callback
-		 * has to be called for every period.
+		 * Enable period interrupt only if it is requested
 		 */
-		edesc->pset[i].param.opt |= TCINTEN;
+		if (tx_flags & DMA_PREP_INTERRUPT)
+			edesc->pset[i].param.opt |= TCINTEN;
 	}
 
 	/* Place the cyclic channel to highest priority queue */
-- 
2.0.0

[PATCH 1/2] dma: edma: Update caps->residue_granularity to match with reality

[Peter Ujfalusi]

The edma can report accurate DMA position so update the residue_granularity
to DMA_RESIDUE_GRANULARITY_BURST.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi-l0cyMroinI0@public.gmane.org>
---
 drivers/dma/edma.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/dma/edma.c b/drivers/dma/edma.c
index fe55f78ea137..d5017abcf7cb 100644
--- a/drivers/dma/edma.c
+++ b/drivers/dma/edma.c
@@ -1001,7 +1001,7 @@ static int edma_dma_device_slave_caps(struct dma_chan *dchan,
 	caps->directions = BIT(DMA_DEV_TO_MEM) | BIT(DMA_MEM_TO_DEV);
 	caps->cmd_pause = true;
 	caps->cmd_terminate = true;
-	caps->residue_granularity = DMA_RESIDUE_GRANULARITY_DESCRIPTOR;
+	caps->residue_granularity = DMA_RESIDUE_GRANULARITY_BURST;
 
 	return 0;
 }
-- 
2.0.0

[PATCH 1/2] dma: edma: Update caps->residue_granularity to match with reality

[Peter Ujfalusi]

The edma can report accurate DMA position so update the residue_granularity
to DMA_RESIDUE_GRANULARITY_BURST.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
---
 drivers/dma/edma.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/dma/edma.c b/drivers/dma/edma.c
index fe55f78ea137..d5017abcf7cb 100644
--- a/drivers/dma/edma.c
+++ b/drivers/dma/edma.c
@@ -1001,7 +1001,7 @@ static int edma_dma_device_slave_caps(struct dma_chan *dchan,
 	caps->directions = BIT(DMA_DEV_TO_MEM) | BIT(DMA_MEM_TO_DEV);
 	caps->cmd_pause = true;
 	caps->cmd_terminate = true;
-	caps->residue_granularity = DMA_RESIDUE_GRANULARITY_DESCRIPTOR;
+	caps->residue_granularity = DMA_RESIDUE_GRANULARITY_BURST;
 
 	return 0;
 }
-- 
2.0.0

Re: [PATCH 0/2] dma: edma: Allow to disable eDMA IRQ during cyclic transfer

[Peter Ujfalusi]

On 07/16/2014 03:29 PM, Peter Ujfalusi wrote:
> Hi,
> 
> After this series clients can ask to not receive notifications after each period.
> In this case we can disable the completion interrupt since the position reporting
> does not rely on it for cyclic mode.
> Patchset for ASoC part has been sent which allows users space to take adventage
> of SNDRV_PCM_INFO_NO_PERIOD_WAKEUP:
> [1] http://mailman.alsa-project.org/pipermail/alsa-devel/2014-July/078993.html

ping.

> When both series applied on top of linux-next:
> 
> # cache the audio file in memory
> cat some_music.mp3 > /dev/null
> # start PA
> pulseaudio -v
> # in another terminal, play the music via PA
> mplayer -ao pulse some_music.mp3
> # yet another terminal to monitor the interrupts
> watch cat /proc/interrupts
> # note the non increasing number of edma interrupts ;)
> 
> Regards,
> Peter
> ---
> Peter Ujfalusi (2):
>   dma: edma: Update caps->residue_granularity to match with reality
>   dma: edma: Support to suppress the period interrupts in cyclic mode
> 
>  drivers/dma/edma.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 


-- 
Péter

Re: [PATCH 0/2] dma: edma: Allow to disable eDMA IRQ during cyclic transfer

[Peter Ujfalusi]

On 07/16/2014 03:29 PM, Peter Ujfalusi wrote:
> Hi,
> 
> After this series clients can ask to not receive notifications after each period.
> In this case we can disable the completion interrupt since the position reporting
> does not rely on it for cyclic mode.
> Patchset for ASoC part has been sent which allows users space to take adventage
> of SNDRV_PCM_INFO_NO_PERIOD_WAKEUP:
> [1] http://mailman.alsa-project.org/pipermail/alsa-devel/2014-July/078993.html

ping.

> When both series applied on top of linux-next:
> 
> # cache the audio file in memory
> cat some_music.mp3 > /dev/null
> # start PA
> pulseaudio -v
> # in another terminal, play the music via PA
> mplayer -ao pulse some_music.mp3
> # yet another terminal to monitor the interrupts
> watch cat /proc/interrupts
> # note the non increasing number of edma interrupts ;)
> 
> Regards,
> Peter
> ---
> Peter Ujfalusi (2):
>   dma: edma: Update caps->residue_granularity to match with reality
>   dma: edma: Support to suppress the period interrupts in cyclic mode
> 
>  drivers/dma/edma.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 


-- 
Péter

Re: [PATCH 0/2] dma: edma: Allow to disable eDMA IRQ during cyclic transfer

[Vinod Koul]

On Wed, Jul 16, 2014 at 03:29:19PM +0300, Peter Ujfalusi wrote:
> Hi,
> 
> After this series clients can ask to not receive notifications after each period.
> In this case we can disable the completion interrupt since the position reporting
> does not rely on it for cyclic mode.
> Patchset for ASoC part has been sent which allows users space to take adventage
> of SNDRV_PCM_INFO_NO_PERIOD_WAKEUP:
> [1] http://mailman.alsa-project.org/pipermail/alsa-devel/2014-July/078993.html

Applied, thanks

Please use right subsystem name for patches, i have fixed that while
applying

-- 
~Vinod