[PATCH 0/2] ASoC/SoundWire: Intel: add sdw BE dai trigger

[PATCH 0/2] ASoC/SoundWire: Intel: add sdw BE dai trigger

[Bard Liao]

For SOF IPC4, we need to set pipeline state in BE DAI trigger.
@Vinod, Could you ack if the soundwire patch looks good to you?
And we can go through the ASoC tree since the change is mainly on ASoC.

Bard Liao (2):
  soundwire: Intel: add trigger callback
  ASoC: SOF: Intel: add trigger callback into sdw_callback

 drivers/soundwire/intel.c           |  8 ++++++++
 include/linux/soundwire/sdw_intel.h |  1 +
 sound/soc/sof/intel/hda-dai.c       | 15 ++++++++++++---
 sound/soc/sof/intel/hda.c           |  2 +-
 sound/soc/sof/intel/hda.h           |  1 +
 5 files changed, 23 insertions(+), 4 deletions(-)

-- 
2.17.1

[PATCH 1/2] soundwire: Intel: add trigger callback

[Bard Liao]

When a pipeline is split into FE and BE parts, the BE pipeline may need to
be triggered separately in the BE trigger op. So add the trigger callback
in the link_res ops that will be invoked during BE DAI trigger.

Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Rander Wang <rander.wang@intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
---
 drivers/soundwire/intel.c           | 8 ++++++++
 include/linux/soundwire/sdw_intel.h | 1 +
 2 files changed, 9 insertions(+)

diff --git a/drivers/soundwire/intel.c b/drivers/soundwire/intel.c
index 505c5ef061e3..2e7c27d303b4 100644
--- a/drivers/soundwire/intel.c
+++ b/drivers/soundwire/intel.c
@@ -1004,9 +1004,17 @@ static int intel_trigger(struct snd_pcm_substream *substream, int cmd, struct sn
 {
 	struct sdw_cdns *cdns = snd_soc_dai_get_drvdata(dai);
 	struct sdw_intel *sdw = cdns_to_intel(cdns);
+	struct sdw_intel_link_res *res = sdw->link_res;
 	struct sdw_cdns_dma_data *dma;
 	int ret = 0;
 
+	/* The .trigger callback is used to send required IPC to audio
+	 * firmware. The .free_stream callback will still be called
+	 * by intel_free_stream() in the TRIGGER_SUSPEND case.
+	 */
+	if (res->ops && res->ops->trigger)
+		res->ops->trigger(dai, cmd, substream->stream);
+
 	dma = snd_soc_dai_get_dma_data(dai, substream);
 	if (!dma) {
 		dev_err(dai->dev, "failed to get dma data in %s\n",
diff --git a/include/linux/soundwire/sdw_intel.h b/include/linux/soundwire/sdw_intel.h
index 67e0d3e750b5..f638707fd712 100644
--- a/include/linux/soundwire/sdw_intel.h
+++ b/include/linux/soundwire/sdw_intel.h
@@ -119,6 +119,7 @@ struct sdw_intel_ops {
 			     struct sdw_intel_stream_params_data *params_data);
 	int (*free_stream)(struct device *dev,
 			   struct sdw_intel_stream_free_data *free_data);
+	int (*trigger)(struct snd_soc_dai *dai, int cmd, int stream);
 };
 
 /**
-- 
2.17.1

[PATCH 2/2] ASoC: SOF: Intel: add trigger callback into sdw_callback

[Bard Liao]

For IPC4, we need to set pipeline state in BE DAI trigger.

Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Rander Wang <rander.wang@intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
---
 sound/soc/sof/intel/hda-dai.c | 15 ++++++++++++---
 sound/soc/sof/intel/hda.c     |  2 +-
 sound/soc/sof/intel/hda.h     |  1 +
 3 files changed, 14 insertions(+), 4 deletions(-)

diff --git a/sound/soc/sof/intel/hda-dai.c b/sound/soc/sof/intel/hda-dai.c
index 228079a52c3d..6ed99fdc5793 100644
--- a/sound/soc/sof/intel/hda-dai.c
+++ b/sound/soc/sof/intel/hda-dai.c
@@ -713,8 +713,7 @@ static const struct snd_soc_dai_ops ipc3_ssp_dai_ops = {
 	.shutdown = ssp_dai_shutdown,
 };
 
-static int ipc4_be_dai_trigger(struct snd_pcm_substream *substream,
-			       int cmd, struct snd_soc_dai *dai)
+static int ipc4_be_dai_common_trigger(struct snd_soc_dai *dai, int cmd, int stream)
 {
 	struct snd_sof_widget *pipe_widget;
 	struct sof_ipc4_pipeline *pipeline;
@@ -723,7 +722,7 @@ static int ipc4_be_dai_trigger(struct snd_pcm_substream *substream,
 	struct snd_sof_dev *sdev;
 	int ret;
 
-	w = snd_soc_dai_get_widget(dai, substream->stream);
+	w = snd_soc_dai_get_widget(dai, stream);
 	swidget = w->dobj.private;
 	pipe_widget = swidget->pipe_widget;
 	pipeline = pipe_widget->private;
@@ -758,6 +757,12 @@ static int ipc4_be_dai_trigger(struct snd_pcm_substream *substream,
 	return 0;
 }
 
+static int ipc4_be_dai_trigger(struct snd_pcm_substream *substream,
+			       int cmd, struct snd_soc_dai *dai)
+{
+	return ipc4_be_dai_common_trigger(dai, cmd, substream->stream);
+}
+
 static const struct snd_soc_dai_ops ipc4_dmic_dai_ops = {
 	.trigger = ipc4_be_dai_trigger,
 };
@@ -809,6 +814,10 @@ void hda_set_dai_drv_ops(struct snd_sof_dev *sdev, struct snd_sof_dsp_ops *ops)
 		if (!hda_use_tplg_nhlt)
 			ipc4_data->nhlt = intel_nhlt_init(sdev->dev);
 
+#if IS_ENABLED(CONFIG_SND_SOC_SOF_INTEL_SOUNDWIRE)
+		sdw_callback.trigger = ipc4_be_dai_common_trigger;
+#endif
+
 		break;
 	}
 	default:
diff --git a/sound/soc/sof/intel/hda.c b/sound/soc/sof/intel/hda.c
index bc07df1fc39f..cdd3601e84f5 100644
--- a/sound/soc/sof/intel/hda.c
+++ b/sound/soc/sof/intel/hda.c
@@ -147,7 +147,7 @@ static int sdw_free_stream(struct device *dev,
 	return hda_ctrl_dai_widget_free(w, SOF_DAI_CONFIG_FLAGS_NONE, &data);
 }
 
-static const struct sdw_intel_ops sdw_callback = {
+struct sdw_intel_ops sdw_callback = {
 	.params_stream = sdw_params_stream,
 	.free_stream = sdw_free_stream,
 };
diff --git a/sound/soc/sof/intel/hda.h b/sound/soc/sof/intel/hda.h
index f4e4cd7d7406..ec7a2d947eb6 100644
--- a/sound/soc/sof/intel/hda.h
+++ b/sound/soc/sof/intel/hda.h
@@ -771,5 +771,6 @@ irqreturn_t cnl_ipc4_irq_thread(int irq, void *context);
 int cnl_ipc4_send_msg(struct snd_sof_dev *sdev, struct snd_sof_ipc_msg *msg);
 irqreturn_t hda_dsp_ipc4_irq_thread(int irq, void *context);
 int hda_dsp_ipc4_send_msg(struct snd_sof_dev *sdev, struct snd_sof_ipc_msg *msg);
+extern struct sdw_intel_ops sdw_callback;
 
 #endif
-- 
2.17.1

Re: [PATCH 2/2] ASoC: SOF: Intel: add trigger callback into sdw_callback

[Greg KH]

On Tue, Jun 14, 2022 at 03:08:17PM +0800, Bard Liao wrote:
> For IPC4, we need to set pipeline state in BE DAI trigger.
> 
> Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>
> Reviewed-by: Rander Wang <rander.wang@intel.com>
> Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
> Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
> ---
>  sound/soc/sof/intel/hda-dai.c | 15 ++++++++++++---
>  sound/soc/sof/intel/hda.c     |  2 +-
>  sound/soc/sof/intel/hda.h     |  1 +
>  3 files changed, 14 insertions(+), 4 deletions(-)
> 
> diff --git a/sound/soc/sof/intel/hda-dai.c b/sound/soc/sof/intel/hda-dai.c
> index 228079a52c3d..6ed99fdc5793 100644
> --- a/sound/soc/sof/intel/hda-dai.c
> +++ b/sound/soc/sof/intel/hda-dai.c
> @@ -713,8 +713,7 @@ static const struct snd_soc_dai_ops ipc3_ssp_dai_ops = {
>  	.shutdown = ssp_dai_shutdown,
>  };
>  
> -static int ipc4_be_dai_trigger(struct snd_pcm_substream *substream,
> -			       int cmd, struct snd_soc_dai *dai)
> +static int ipc4_be_dai_common_trigger(struct snd_soc_dai *dai, int cmd, int stream)
>  {
>  	struct snd_sof_widget *pipe_widget;
>  	struct sof_ipc4_pipeline *pipeline;
> @@ -723,7 +722,7 @@ static int ipc4_be_dai_trigger(struct snd_pcm_substream *substream,
>  	struct snd_sof_dev *sdev;
>  	int ret;
>  
> -	w = snd_soc_dai_get_widget(dai, substream->stream);
> +	w = snd_soc_dai_get_widget(dai, stream);
>  	swidget = w->dobj.private;
>  	pipe_widget = swidget->pipe_widget;
>  	pipeline = pipe_widget->private;
> @@ -758,6 +757,12 @@ static int ipc4_be_dai_trigger(struct snd_pcm_substream *substream,
>  	return 0;
>  }
>  
> +static int ipc4_be_dai_trigger(struct snd_pcm_substream *substream,
> +			       int cmd, struct snd_soc_dai *dai)
> +{
> +	return ipc4_be_dai_common_trigger(dai, cmd, substream->stream);
> +}
> +
>  static const struct snd_soc_dai_ops ipc4_dmic_dai_ops = {
>  	.trigger = ipc4_be_dai_trigger,
>  };
> @@ -809,6 +814,10 @@ void hda_set_dai_drv_ops(struct snd_sof_dev *sdev, struct snd_sof_dsp_ops *ops)
>  		if (!hda_use_tplg_nhlt)
>  			ipc4_data->nhlt = intel_nhlt_init(sdev->dev);
>  
> +#if IS_ENABLED(CONFIG_SND_SOC_SOF_INTEL_SOUNDWIRE)
> +		sdw_callback.trigger = ipc4_be_dai_common_trigger;
> +#endif

#if should not be in .c files if at all possible.  Surely there's a
better way here...

thanks,

greg k-h

Re: [PATCH 2/2] ASoC: SOF: Intel: add trigger callback into sdw_callback

[Pierre-Louis Bossart]

>> @@ -809,6 +814,10 @@ void hda_set_dai_drv_ops(struct snd_sof_dev *sdev, struct snd_sof_dsp_ops *ops)
>>  		if (!hda_use_tplg_nhlt)
>>  			ipc4_data->nhlt = intel_nhlt_init(sdev->dev);
>>  
>> +#if IS_ENABLED(CONFIG_SND_SOC_SOF_INTEL_SOUNDWIRE)
>> +		sdw_callback.trigger = ipc4_be_dai_common_trigger;
>> +#endif
> 
> #if should not be in .c files if at all possible.  Surely there's a
> better way here...

we could use

if (IS_ENABLED(CONFIG_SND_SOC_SOF_INTEL_SOUNDWIRE))
    sdw_callback.trigger = ipc4_be_dai_common_trigger;

would that work?

We try to keep this driver configurable, not all platforms require
SoundWire or HDaudio, and that 'sdw_callback' ops structure is
conditionally declared.

Re: [PATCH 2/2] ASoC: SOF: Intel: add trigger callback into sdw_callback

[Greg KH]

On Tue, Jun 14, 2022 at 09:55:41AM -0500, Pierre-Louis Bossart wrote:
> 
> >> @@ -809,6 +814,10 @@ void hda_set_dai_drv_ops(struct snd_sof_dev *sdev, struct snd_sof_dsp_ops *ops)
> >>  		if (!hda_use_tplg_nhlt)
> >>  			ipc4_data->nhlt = intel_nhlt_init(sdev->dev);
> >>  
> >> +#if IS_ENABLED(CONFIG_SND_SOC_SOF_INTEL_SOUNDWIRE)
> >> +		sdw_callback.trigger = ipc4_be_dai_common_trigger;
> >> +#endif
> > 
> > #if should not be in .c files if at all possible.  Surely there's a
> > better way here...
> 
> we could use
> 
> if (IS_ENABLED(CONFIG_SND_SOC_SOF_INTEL_SOUNDWIRE))
>     sdw_callback.trigger = ipc4_be_dai_common_trigger;
> 
> would that work?

It's more readable, right?  Also easier to maintain over time.

> We try to keep this driver configurable, not all platforms require
> SoundWire or HDaudio, and that 'sdw_callback' ops structure is
> conditionally declared.

Perhaps don't conditionally declare that?  How much does it really save
to do that?

thanks,

greg k-h

Re: [PATCH 2/2] ASoC: SOF: Intel: add trigger callback into sdw_callback

[Pierre-Louis Bossart]

>>>> @@ -809,6 +814,10 @@ void hda_set_dai_drv_ops(struct snd_sof_dev *sdev, struct snd_sof_dsp_ops *ops)
>>>>  		if (!hda_use_tplg_nhlt)
>>>>  			ipc4_data->nhlt = intel_nhlt_init(sdev->dev);
>>>>  
>>>> +#if IS_ENABLED(CONFIG_SND_SOC_SOF_INTEL_SOUNDWIRE)
>>>> +		sdw_callback.trigger = ipc4_be_dai_common_trigger;
>>>> +#endif
>>>
>>> #if should not be in .c files if at all possible.  Surely there's a
>>> better way here...
>>
>> we could use
>>
>> if (IS_ENABLED(CONFIG_SND_SOC_SOF_INTEL_SOUNDWIRE))
>>     sdw_callback.trigger = ipc4_be_dai_common_trigger;
>>
>> would that work?
> 
> It's more readable, right?  Also easier to maintain over time.

yes, it's more readable, and that's no problem to update.

There's another #if IS_ENABLED( that we can replace by a if (IS_ENABLED
in the same routine, that would make this routine less of an eyesore. I
am not sure why we added all these #if, we can cleanup.

 	case SOF_INTEL_IPC4:
	{
		struct sof_ipc4_fw_data *ipc4_data = sdev->private;

		for (i = 0; i < ops->num_drv; i++) {
			if (strstr(ops->drv[i].name, "DMIC")) {
				ops->drv[i].ops = &ipc4_dmic_dai_ops;
				continue;
			}
			if (strstr(ops->drv[i].name, "SSP")) {
				ops->drv[i].ops = &ipc4_ssp_dai_ops;
				continue;
			}
#if IS_ENABLED(CONFIG_SND_SOC_SOF_HDA)
			if (strstr(ops->drv[i].name, "iDisp") ||
			    strstr(ops->drv[i].name, "Analog") ||
			    strstr(ops->drv[i].name, "Digital"))
				ops->drv[i].ops = &ipc4_hda_dai_ops;
#endif
		}

		if (!hda_use_tplg_nhlt)
			ipc4_data->nhlt = intel_nhlt_init(sdev->dev);

#if IS_ENABLED(CONFIG_SND_SOC_SOF_INTEL_SOUNDWIRE)
		sdw_callback.trigger = ipc4_be_dai_common_trigger;
#endif

		break;
	}
	default:
		break;
	}
}
>> We try to keep this driver configurable, not all platforms require
>> SoundWire or HDaudio, and that 'sdw_callback' ops structure is
>> conditionally declared.
> 
> Perhaps don't conditionally declare that?  How much does it really save
> to do that?

It would force us to expose additional things that are only relevant for
SoundWire, see the large block of code in hda.c

https://elixir.bootlin.com/linux/latest/source/sound/soc/sof/intel/hda.c#L108

We've been burned in the past by having too many things in a single
configuration, so we try to allow for minimal builds that don't depend
on other modules in sound/soc/codecs/, sound/pci/hda and
drivers/soundwire - it also forces us to get the Kconfig dependencies right.

if (IS_ENABLED()) is less invasive in that case.

Thanks for your feedback
-Pierre