[PATCH 0/4] ALSA: hda: Bug fixes

[PATCH 0/4] ALSA: hda: Bug fixes

[Sriram Periyasamy]

The series contains few bug fixes and debug updates

Pardha Saradhi K (4):
  ALSA: hda: Make sure DMA is stopped by reading back the RUN bit
  ALSA: hda: Make sure DMA is started by reading back the RUN bit
  ALSA: hda: Log HDA Hardware related errors
  ALSA: hda: check if stream is stopped in snd_hdac_stream_clear

 sound/hda/hdac_stream.c | 43 +++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 43 insertions(+)

-- 
2.7.4

[PATCH 1/4] ALSA: hda: Make sure DMA is stopped by reading back the RUN bit

[Sriram Periyasamy]

From: Pardha Saradhi K <pardha.saradhi.kesapragada@intel.com>

As per HW recommendation, after clearing the RUN bit,
software must read a 0 from the RUN bit, before modifying
related control registers or re-starting the DMA engine.

Signed-off-by: Pardha Saradhi K <pardha.saradhi.kesapragada@intel.com>
Signed-off-by: Sriram Periyasamy <sriramx.periyasamy@intel.com>
---
 sound/hda/hdac_stream.c | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/sound/hda/hdac_stream.c b/sound/hda/hdac_stream.c
index e1472c7ab6c1..2000ea6f48fa 100644
--- a/sound/hda/hdac_stream.c
+++ b/sound/hda/hdac_stream.c
@@ -70,8 +70,23 @@ EXPORT_SYMBOL_GPL(snd_hdac_stream_start);
  */
 void snd_hdac_stream_clear(struct hdac_stream *azx_dev)
 {
+	int timeout = 300;
+	unsigned char val;
+
 	snd_hdac_stream_updateb(azx_dev, SD_CTL,
 				SD_CTL_DMA_START | SD_INT_MASK, 0);
+
+	do {
+		udelay(3);
+		val = snd_hdac_stream_readb(azx_dev, SD_CTL) &
+				SD_CTL_DMA_START;
+		if (!val)
+			break;
+	} while (--timeout);
+
+	if (!timeout)
+		dev_err(azx_dev->bus->dev, "unable to stop the stream\n");
+
 	snd_hdac_stream_writeb(azx_dev, SD_STS, SD_INT_MASK); /* to be sure */
 	azx_dev->running = false;
 }
-- 
2.7.4

Re: [PATCH 1/4] ALSA: hda: Make sure DMA is stopped by reading back the RUN bit

[Takashi Iwai]

On Tue, 15 May 2018 08:40:53 +0200,
Sriram Periyasamy wrote:
> 
> From: Pardha Saradhi K <pardha.saradhi.kesapragada@intel.com>
> 
> As per HW recommendation, after clearing the RUN bit,
> software must read a 0 from the RUN bit, before modifying
> related control registers or re-starting the DMA engine.

This is already done in snd_hdac_stream_sync().

Doing this in snd_hdac_stream_clear() may delay the sync'ed operations
unnecessarily.


thanks,

Takashi

> 
> Signed-off-by: Pardha Saradhi K <pardha.saradhi.kesapragada@intel.com>
> Signed-off-by: Sriram Periyasamy <sriramx.periyasamy@intel.com>
> ---
>  sound/hda/hdac_stream.c | 15 +++++++++++++++
>  1 file changed, 15 insertions(+)
> 
> diff --git a/sound/hda/hdac_stream.c b/sound/hda/hdac_stream.c
> index e1472c7ab6c1..2000ea6f48fa 100644
> --- a/sound/hda/hdac_stream.c
> +++ b/sound/hda/hdac_stream.c
> @@ -70,8 +70,23 @@ EXPORT_SYMBOL_GPL(snd_hdac_stream_start);
>   */
>  void snd_hdac_stream_clear(struct hdac_stream *azx_dev)
>  {
> +	int timeout = 300;
> +	unsigned char val;
> +
>  	snd_hdac_stream_updateb(azx_dev, SD_CTL,
>  				SD_CTL_DMA_START | SD_INT_MASK, 0);
> +
> +	do {
> +		udelay(3);
> +		val = snd_hdac_stream_readb(azx_dev, SD_CTL) &
> +				SD_CTL_DMA_START;
> +		if (!val)
> +			break;
> +	} while (--timeout);
> +
> +	if (!timeout)
> +		dev_err(azx_dev->bus->dev, "unable to stop the stream\n");
> +
>  	snd_hdac_stream_writeb(azx_dev, SD_STS, SD_INT_MASK); /* to be sure */
>  	azx_dev->running = false;
>  }
> -- 
> 2.7.4
> 

[PATCH 2/4] ALSA: hda: Make sure DMA is started by reading back the RUN bit

[Sriram Periyasamy]

From: Pardha Saradhi K <pardha.saradhi.kesapragada@intel.com>

As per HW recommendation, after setting the RUN bit,
software must read a 1 from the RUN bit, before modifying
related control registers/re-starting the DMA engine.

Signed-off-by: Pardha Saradhi K <pardha.saradhi.kesapragada@intel.com>
Signed-off-by: Sriram Periyasamy <sriramx.periyasamy@intel.com>
---
 sound/hda/hdac_stream.c | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)

diff --git a/sound/hda/hdac_stream.c b/sound/hda/hdac_stream.c
index 2000ea6f48fa..33c8ced528f6 100644
--- a/sound/hda/hdac_stream.c
+++ b/sound/hda/hdac_stream.c
@@ -48,6 +48,8 @@ EXPORT_SYMBOL_GPL(snd_hdac_stream_init);
 void snd_hdac_stream_start(struct hdac_stream *azx_dev, bool fresh_start)
 {
 	struct hdac_bus *bus = azx_dev->bus;
+	int timeout = 300;
+	unsigned char val;
 
 	trace_snd_hdac_stream_start(bus, azx_dev);
 
@@ -60,6 +62,20 @@ void snd_hdac_stream_start(struct hdac_stream *azx_dev, bool fresh_start)
 	/* set DMA start and interrupt mask */
 	snd_hdac_stream_updateb(azx_dev, SD_CTL,
 				0, SD_CTL_DMA_START | SD_INT_MASK);
+
+	do {
+		udelay(3);
+		val = snd_hdac_stream_readb(azx_dev, SD_CTL) & SD_CTL_DMA_START;
+		if (val)
+			break;
+	} while (--timeout);
+
+	if (!timeout) {
+		azx_dev->running = false;
+		dev_err(azx_dev->bus->dev, "unable to start the stream\n");
+		return;
+	}
+
 	azx_dev->running = true;
 }
 EXPORT_SYMBOL_GPL(snd_hdac_stream_start);
-- 
2.7.4

Re: [PATCH 2/4] ALSA: hda: Make sure DMA is started by reading back the RUN bit

[Takashi Iwai]

On Tue, 15 May 2018 08:40:54 +0200,
Sriram Periyasamy wrote:
> 
> From: Pardha Saradhi K <pardha.saradhi.kesapragada@intel.com>
> 
> As per HW recommendation, after setting the RUN bit,
> software must read a 1 from the RUN bit, before modifying
> related control registers/re-starting the DMA engine.

Rather FIFO is checked in snd_hdac_stream_sync(), so I guess it
already suffices.  But we can add this sanity check there, too, if it
really matters.


thanks,

Takashi

[PATCH 3/4] ALSA: hda: Log HDA Hardware related errors

[Sriram Periyasamy]

From: Pardha Saradhi K <pardha.saradhi.kesapragada@intel.com>

Detect the timeout while modifying HDA DMA related Registers
for stream reset and print them to console for user information

Signed-off-by: Pardha Saradhi K <pardha.saradhi.kesapragada@intel.com>
Signed-off-by: Sriram Periyasamy <sriramx.periyasamy@intel.com>
---
 sound/hda/hdac_stream.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/sound/hda/hdac_stream.c b/sound/hda/hdac_stream.c
index 33c8ced528f6..aa8a2884289f 100644
--- a/sound/hda/hdac_stream.c
+++ b/sound/hda/hdac_stream.c
@@ -144,6 +144,10 @@ void snd_hdac_stream_reset(struct hdac_stream *azx_dev)
 		if (val)
 			break;
 	} while (--timeout);
+
+	if (!timeout)
+		dev_err(azx_dev->bus->dev, "timeout on stream reset entry\n");
+
 	val &= ~SD_CTL_STREAM_RESET;
 	snd_hdac_stream_writeb(azx_dev, SD_CTL, val);
 	udelay(3);
@@ -157,6 +161,9 @@ void snd_hdac_stream_reset(struct hdac_stream *azx_dev)
 			break;
 	} while (--timeout);
 
+	if (!timeout)
+		dev_err(azx_dev->bus->dev, "timeout on stream reset exit\n");
+
 	/* reset first position - may not be synced with hw at this time */
 	if (azx_dev->posbuf)
 		*azx_dev->posbuf = 0;
-- 
2.7.4

Re: [PATCH 3/4] ALSA: hda: Log HDA Hardware related errors

[Takashi Iwai]

On Tue, 15 May 2018 08:40:55 +0200,
Sriram Periyasamy wrote:
> 
> From: Pardha Saradhi K <pardha.saradhi.kesapragada@intel.com>
> 
> Detect the timeout while modifying HDA DMA related Registers
> for stream reset and print them to console for user information
> 
> Signed-off-by: Pardha Saradhi K <pardha.saradhi.kesapragada@intel.com>
> Signed-off-by: Sriram Periyasamy <sriramx.periyasamy@intel.com>
> ---
>  sound/hda/hdac_stream.c | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/sound/hda/hdac_stream.c b/sound/hda/hdac_stream.c
> index 33c8ced528f6..aa8a2884289f 100644
> --- a/sound/hda/hdac_stream.c
> +++ b/sound/hda/hdac_stream.c
> @@ -144,6 +144,10 @@ void snd_hdac_stream_reset(struct hdac_stream *azx_dev)
>  		if (val)
>  			break;
>  	} while (--timeout);
> +
> +	if (!timeout)
> +		dev_err(azx_dev->bus->dev, "timeout on stream reset entry\n");
> +
>  	val &= ~SD_CTL_STREAM_RESET;
>  	snd_hdac_stream_writeb(azx_dev, SD_CTL, val);
>  	udelay(3);
> @@ -157,6 +161,9 @@ void snd_hdac_stream_reset(struct hdac_stream *azx_dev)
>  			break;
>  	} while (--timeout);
>  
> +	if (!timeout)
> +		dev_err(azx_dev->bus->dev, "timeout on stream reset exit\n");

Just from curiosity: did you ever hit these?  Or just because of
consistency?


Takashi

[PATCH 4/4] ALSA: hda: check if stream is stopped in snd_hdac_stream_clear

[Sriram Periyasamy]

From: Pardha Saradhi K <pardha.saradhi.kesapragada@intel.com>

Check if the DMA Channel is already stopped. There is no need
to stop it again if stopped.

Signed-off-by: Pardha Saradhi K <pardha.saradhi.kesapragada@intel.com>
Signed-off-by: Sriram Periyasamy <sriramx.periyasamy@intel.com>
---
 sound/hda/hdac_stream.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/sound/hda/hdac_stream.c b/sound/hda/hdac_stream.c
index aa8a2884289f..c1810eb7212b 100644
--- a/sound/hda/hdac_stream.c
+++ b/sound/hda/hdac_stream.c
@@ -89,6 +89,11 @@ void snd_hdac_stream_clear(struct hdac_stream *azx_dev)
 	int timeout = 300;
 	unsigned char val;
 
+	/* check if the DMA is already stopped */
+	val = snd_hdac_stream_readb(azx_dev, SD_CTL) & SD_CTL_DMA_START;
+	if (!val)
+		return;
+
 	snd_hdac_stream_updateb(azx_dev, SD_CTL,
 				SD_CTL_DMA_START | SD_INT_MASK, 0);
 
-- 
2.7.4

Re: [PATCH 4/4] ALSA: hda: check if stream is stopped in snd_hdac_stream_clear

[Takashi Iwai]

On Tue, 15 May 2018 08:40:56 +0200,
Sriram Periyasamy wrote:
> 
> From: Pardha Saradhi K <pardha.saradhi.kesapragada@intel.com>
> 
> Check if the DMA Channel is already stopped. There is no need
> to stop it again if stopped.
> 
> Signed-off-by: Pardha Saradhi K <pardha.saradhi.kesapragada@intel.com>
> Signed-off-by: Sriram Periyasamy <sriramx.periyasamy@intel.com>
> ---
>  sound/hda/hdac_stream.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/sound/hda/hdac_stream.c b/sound/hda/hdac_stream.c
> index aa8a2884289f..c1810eb7212b 100644
> --- a/sound/hda/hdac_stream.c
> +++ b/sound/hda/hdac_stream.c
> @@ -89,6 +89,11 @@ void snd_hdac_stream_clear(struct hdac_stream *azx_dev)
>  	int timeout = 300;
>  	unsigned char val;
>  
> +	/* check if the DMA is already stopped */
> +	val = snd_hdac_stream_readb(azx_dev, SD_CTL) & SD_CTL_DMA_START;
> +	if (!val)
> +		return;
> +
>  	snd_hdac_stream_updateb(azx_dev, SD_CTL,
>  				SD_CTL_DMA_START | SD_INT_MASK, 0);

Here we call snd_hdac_stream_updateb(), so we don't write
unnecessarily.

And, if any shortcut is wanted, rather check azx_dev->running flag
instead of the h/w register.  The former is significantly cheaper.


thanks,

Takashi