[PATCH v3 0/6] ASoC: q6apm: fix under runs and fragment sizes

[PATCH v3 0/6] ASoC: q6apm: fix under runs and fragment sizes

[srinivas.kandagatla]

From: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>

On Qualcomm Audioreach setup, some of the audio artifacts are seen in
both recording and playback. These patches fix issues by
1. Adjusting the fragment size that dsp can service.
2. schedule available playback buffers in time for dsp to not hit under runs 
3. remove some of the manual calculations done to get hardware pointer.

With these patches, am able to see Audio quality improvements.

Any testing would be appreciated.

thanks,
Srini

Changes since v2:
	- dropped patch which is causing regression with pluseaudio.
	- setup period sizes only for capture path
	- fix underruns/overruns in dsp pipelines.
	- add fixes tag
	- add patch to fix buffer alignment

Changes since v1:
	- added new patches to fix the fragment size, pointer
	  calculations
	- updated to schedule only available buffers.

Srinivas Kandagatla (6):
  ASoC: q6apm-dai: schedule all available frames to avoid dsp under-runs
  ASoC: q6apm: add q6apm_get_hw_pointer helper
  ASoC: q6apm-dai: make use of q6apm_get_hw_pointer
  ASoC: qdsp6: q6apm-dai: set correct period and buffer alignment.
  ASoC: qdsp6: q6apm-dai: fix capture pipeline overruns.
  ASoC: qdsp6: q6apm-dai: fix playback dsp pipeline underruns

 sound/soc/qcom/qdsp6/q6apm-dai.c | 63 +++++++++++++++++---------------
 sound/soc/qcom/qdsp6/q6apm.c     | 18 ++++++++-
 sound/soc/qcom/qdsp6/q6apm.h     |  3 ++
 3 files changed, 53 insertions(+), 31 deletions(-)

-- 
2.39.5

[PATCH v3 1/6] ASoC: q6apm-dai: schedule all available frames to avoid dsp under-runs

[srinivas.kandagatla]

From: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>

With the existing code, we are only setting up one period at a time, in a
ping-pong buffer style. This triggers lot of underruns in the dsp
leading to jitter noise during audio playback.

Fix this by scheduling all available periods, this will ensure that the dsp
has enough buffer feed and ultimatley fixing the underruns and audio
distortion.

Fixes: 9b4fe0f1cd79 ("ASoC: qdsp6: audioreach: add q6apm-dai support")
Reported-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
---
 sound/soc/qcom/qdsp6/q6apm-dai.c | 28 +++++++++++++++++++++++-----
 1 file changed, 23 insertions(+), 5 deletions(-)

diff --git a/sound/soc/qcom/qdsp6/q6apm-dai.c b/sound/soc/qcom/qdsp6/q6apm-dai.c
index c9404b5934c7..9d8e8e37c6de 100644
--- a/sound/soc/qcom/qdsp6/q6apm-dai.c
+++ b/sound/soc/qcom/qdsp6/q6apm-dai.c
@@ -70,6 +70,7 @@ struct q6apm_dai_rtd {
 	unsigned int bytes_received;
 	unsigned int copied_total;
 	uint16_t bits_per_sample;
+	snd_pcm_uframes_t queue_ptr;
 	bool next_track;
 	enum stream_state state;
 	struct q6apm_graph *graph;
@@ -134,8 +135,6 @@ static void event_handler(uint32_t opcode, uint32_t token, void *payload, void *
 		prtd->pos += prtd->pcm_count;
 		spin_unlock_irqrestore(&prtd->lock, flags);
 		snd_pcm_period_elapsed(substream);
-		if (prtd->state == Q6APM_STREAM_RUNNING)
-			q6apm_write_async(prtd->graph, prtd->pcm_count, 0, 0, 0);
 
 		break;
 	case APM_CLIENT_EVENT_DATA_READ_DONE:
@@ -294,6 +293,27 @@ static int q6apm_dai_prepare(struct snd_soc_component *component,
 	return 0;
 }
 
+static int q6apm_dai_ack(struct snd_soc_component *component, struct snd_pcm_substream *substream)
+{
+	struct snd_pcm_runtime *runtime = substream->runtime;
+	struct q6apm_dai_rtd *prtd = runtime->private_data;
+	int i, ret = 0, avail_periods;
+
+	if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) {
+		avail_periods = (runtime->control->appl_ptr - prtd->queue_ptr)/runtime->period_size;
+		for (i = 0; i < avail_periods; i++) {
+			ret = q6apm_write_async(prtd->graph, prtd->pcm_count, 0, 0, NO_TIMESTAMP);
+			if (ret < 0) {
+				dev_err(component->dev, "Error queuing playback buffer %d\n", ret);
+				return ret;
+			}
+			prtd->queue_ptr += runtime->period_size;
+		}
+	}
+
+	return ret;
+}
+
 static int q6apm_dai_trigger(struct snd_soc_component *component,
 			     struct snd_pcm_substream *substream, int cmd)
 {
@@ -305,9 +325,6 @@ static int q6apm_dai_trigger(struct snd_soc_component *component,
 	case SNDRV_PCM_TRIGGER_START:
 	case SNDRV_PCM_TRIGGER_RESUME:
 	case SNDRV_PCM_TRIGGER_PAUSE_RELEASE:
-		 /* start writing buffers for playback only as we already queued capture buffers */
-		if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK)
-			ret = q6apm_write_async(prtd->graph, prtd->pcm_count, 0, 0, 0);
 		break;
 	case SNDRV_PCM_TRIGGER_STOP:
 		/* TODO support be handled via SoftPause Module */
@@ -836,6 +853,7 @@ static const struct snd_soc_component_driver q6apm_fe_dai_component = {
 	.hw_params	= q6apm_dai_hw_params,
 	.pointer	= q6apm_dai_pointer,
 	.trigger	= q6apm_dai_trigger,
+	.ack		= q6apm_dai_ack,
 	.compress_ops	= &q6apm_dai_compress_ops,
 	.use_dai_pcm_id = true,
 };
-- 
2.39.5

[PATCH v3 2/6] ASoC: q6apm: add q6apm_get_hw_pointer helper

[srinivas.kandagatla]

From: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>

Implement an helper function in q6apm to be able to read the current
hardware pointer for both read and write buffers.

This should help q6apm-dai to get the hardware pointer consistently
without it doing manual calculation, which could go wrong in some race
conditions.

Fixes: 9b4fe0f1cd79 ("ASoC: qdsp6: audioreach: add q6apm-dai support")
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
---
 sound/soc/qcom/qdsp6/q6apm.c | 18 +++++++++++++++++-
 sound/soc/qcom/qdsp6/q6apm.h |  3 +++
 2 files changed, 20 insertions(+), 1 deletion(-)

diff --git a/sound/soc/qcom/qdsp6/q6apm.c b/sound/soc/qcom/qdsp6/q6apm.c
index 2a2a5bd98110..ca57413cb784 100644
--- a/sound/soc/qcom/qdsp6/q6apm.c
+++ b/sound/soc/qcom/qdsp6/q6apm.c
@@ -494,6 +494,19 @@ int q6apm_read(struct q6apm_graph *graph)
 }
 EXPORT_SYMBOL_GPL(q6apm_read);
 
+int q6apm_get_hw_pointer(struct q6apm_graph *graph, int dir)
+{
+	struct audioreach_graph_data *data;
+
+	if (dir == SNDRV_PCM_STREAM_PLAYBACK)
+		data = &graph->rx_data;
+	else
+		data = &graph->tx_data;
+
+	return (int)atomic_read(&data->hw_ptr);
+}
+EXPORT_SYMBOL_GPL(q6apm_get_hw_pointer);
+
 static int graph_callback(struct gpr_resp_pkt *data, void *priv, int op)
 {
 	struct data_cmd_rsp_rd_sh_mem_ep_data_buffer_done_v2 *rd_done;
@@ -520,7 +533,8 @@ static int graph_callback(struct gpr_resp_pkt *data, void *priv, int op)
 		done = data->payload;
 		phys = graph->rx_data.buf[token].phys;
 		mutex_unlock(&graph->lock);
-
+		/* token numbering starts at 0 */
+		atomic_set(&graph->rx_data.hw_ptr, token + 1);
 		if (lower_32_bits(phys) == done->buf_addr_lsw &&
 		    upper_32_bits(phys) == done->buf_addr_msw) {
 			graph->result.opcode = hdr->opcode;
@@ -553,6 +567,8 @@ static int graph_callback(struct gpr_resp_pkt *data, void *priv, int op)
 		rd_done = data->payload;
 		phys = graph->tx_data.buf[hdr->token].phys;
 		mutex_unlock(&graph->lock);
+		/* token numbering starts at 0 */
+		atomic_set(&graph->tx_data.hw_ptr, hdr->token + 1);
 
 		if (upper_32_bits(phys) == rd_done->buf_addr_msw &&
 		    lower_32_bits(phys) == rd_done->buf_addr_lsw) {
diff --git a/sound/soc/qcom/qdsp6/q6apm.h b/sound/soc/qcom/qdsp6/q6apm.h
index c248c8d2b1ab..7ce08b401e31 100644
--- a/sound/soc/qcom/qdsp6/q6apm.h
+++ b/sound/soc/qcom/qdsp6/q6apm.h
@@ -2,6 +2,7 @@
 #ifndef __Q6APM_H__
 #define __Q6APM_H__
 #include <linux/types.h>
+#include <linux/atomic.h>
 #include <linux/slab.h>
 #include <linux/wait.h>
 #include <linux/kernel.h>
@@ -77,6 +78,7 @@ struct audioreach_graph_data {
 	uint32_t num_periods;
 	uint32_t dsp_buf;
 	uint32_t mem_map_handle;
+	atomic_t hw_ptr;
 };
 
 struct audioreach_graph {
@@ -150,4 +152,5 @@ int q6apm_enable_compress_module(struct device *dev, struct q6apm_graph *graph,
 int q6apm_remove_initial_silence(struct device *dev, struct q6apm_graph *graph, uint32_t samples);
 int q6apm_remove_trailing_silence(struct device *dev, struct q6apm_graph *graph, uint32_t samples);
 int q6apm_set_real_module_id(struct device *dev, struct q6apm_graph *graph, uint32_t codec_id);
+int q6apm_get_hw_pointer(struct q6apm_graph *graph, int dir);
 #endif /* __APM_GRAPH_ */
-- 
2.39.5

[PATCH v3 3/6] ASoC: q6apm-dai: make use of q6apm_get_hw_pointer

[srinivas.kandagatla]

From: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>

With the existing code, the buffer position is only reset in pointer
callback, which leaves the possiblity of it going over the size of
buffer size and reporting incorrect position to userspace.

Without this patch, its possible to see errors like:
snd-x1e80100 sound: invalid position: pcmC0D0p:0, pos = 12288, buffer size = 12288, period size = 1536
snd-x1e80100 sound: invalid position: pcmC0D0p:0, pos = 12288, buffer size = 12288, period size = 1536

Fixes: 9b4fe0f1cd79 ("ASoC: qdsp6: audioreach: add q6apm-dai support")
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
---
 sound/soc/qcom/qdsp6/q6apm-dai.c | 23 ++++-------------------
 1 file changed, 4 insertions(+), 19 deletions(-)

diff --git a/sound/soc/qcom/qdsp6/q6apm-dai.c b/sound/soc/qcom/qdsp6/q6apm-dai.c
index 9d8e8e37c6de..90cb24947f31 100644
--- a/sound/soc/qcom/qdsp6/q6apm-dai.c
+++ b/sound/soc/qcom/qdsp6/q6apm-dai.c
@@ -64,7 +64,6 @@ struct q6apm_dai_rtd {
 	phys_addr_t phys;
 	unsigned int pcm_size;
 	unsigned int pcm_count;
-	unsigned int pos;       /* Buffer position */
 	unsigned int periods;
 	unsigned int bytes_sent;
 	unsigned int bytes_received;
@@ -124,23 +123,16 @@ static void event_handler(uint32_t opcode, uint32_t token, void *payload, void *
 {
 	struct q6apm_dai_rtd *prtd = priv;
 	struct snd_pcm_substream *substream = prtd->substream;
-	unsigned long flags;
 
 	switch (opcode) {
 	case APM_CLIENT_EVENT_CMD_EOS_DONE:
 		prtd->state = Q6APM_STREAM_STOPPED;
 		break;
 	case APM_CLIENT_EVENT_DATA_WRITE_DONE:
-		spin_lock_irqsave(&prtd->lock, flags);
-		prtd->pos += prtd->pcm_count;
-		spin_unlock_irqrestore(&prtd->lock, flags);
 		snd_pcm_period_elapsed(substream);
 
 		break;
 	case APM_CLIENT_EVENT_DATA_READ_DONE:
-		spin_lock_irqsave(&prtd->lock, flags);
-		prtd->pos += prtd->pcm_count;
-		spin_unlock_irqrestore(&prtd->lock, flags);
 		snd_pcm_period_elapsed(substream);
 		if (prtd->state == Q6APM_STREAM_RUNNING)
 			q6apm_read(prtd->graph);
@@ -247,7 +239,6 @@ static int q6apm_dai_prepare(struct snd_soc_component *component,
 	}
 
 	prtd->pcm_count = snd_pcm_lib_period_bytes(substream);
-	prtd->pos = 0;
 	/* rate and channels are sent to audio driver */
 	ret = q6apm_graph_media_format_shmem(prtd->graph, &cfg);
 	if (ret < 0) {
@@ -445,16 +436,12 @@ static snd_pcm_uframes_t q6apm_dai_pointer(struct snd_soc_component *component,
 	struct snd_pcm_runtime *runtime = substream->runtime;
 	struct q6apm_dai_rtd *prtd = runtime->private_data;
 	snd_pcm_uframes_t ptr;
-	unsigned long flags;
 
-	spin_lock_irqsave(&prtd->lock, flags);
-	if (prtd->pos == prtd->pcm_size)
-		prtd->pos = 0;
-
-	ptr =  bytes_to_frames(runtime, prtd->pos);
-	spin_unlock_irqrestore(&prtd->lock, flags);
+	ptr = q6apm_get_hw_pointer(prtd->graph, substream->stream) * runtime->period_size;
+	if (ptr)
+		return ptr - 1;
 
-	return ptr;
+	return 0;
 }
 
 static int q6apm_dai_hw_params(struct snd_soc_component *component,
@@ -669,8 +656,6 @@ static int q6apm_dai_compr_set_params(struct snd_soc_component *component,
 	prtd->pcm_size = runtime->fragments * runtime->fragment_size;
 	prtd->bits_per_sample = 16;
 
-	prtd->pos = 0;
-
 	if (prtd->next_track != true) {
 		memcpy(&prtd->codec, codec, sizeof(*codec));
 
-- 
2.39.5

[PATCH v3 4/6] ASoC: qdsp6: q6apm-dai: set correct period and buffer alignment.

[srinivas.kandagatla]

From: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>

DSP expects the buffers to be aligned to 64 bytes. Having a 32 byte
alignment can result in fast playback or capture in some configurations.

Fixes: 9b4fe0f1cd79 ("ASoC: qdsp6: audioreach: add q6apm-dai support")
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
---
 sound/soc/qcom/qdsp6/q6apm-dai.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sound/soc/qcom/qdsp6/q6apm-dai.c b/sound/soc/qcom/qdsp6/q6apm-dai.c
index 90cb24947f31..bb761ee0164b 100644
--- a/sound/soc/qcom/qdsp6/q6apm-dai.c
+++ b/sound/soc/qcom/qdsp6/q6apm-dai.c
@@ -385,13 +385,13 @@ static int q6apm_dai_open(struct snd_soc_component *component,
 		}
 	}
 
-	ret = snd_pcm_hw_constraint_step(runtime, 0, SNDRV_PCM_HW_PARAM_PERIOD_BYTES, 32);
+	ret = snd_pcm_hw_constraint_step(runtime, 0, SNDRV_PCM_HW_PARAM_PERIOD_BYTES, 64);
 	if (ret < 0) {
 		dev_err(dev, "constraint for period bytes step ret = %d\n", ret);
 		goto err;
 	}
 
-	ret = snd_pcm_hw_constraint_step(runtime, 0, SNDRV_PCM_HW_PARAM_BUFFER_BYTES, 32);
+	ret = snd_pcm_hw_constraint_step(runtime, 0, SNDRV_PCM_HW_PARAM_BUFFER_BYTES, 64);
 	if (ret < 0) {
 		dev_err(dev, "constraint for buffer bytes step ret = %d\n", ret);
 		goto err;
-- 
2.39.5

[PATCH v3 5/6] ASoC: qdsp6: q6apm-dai: fix capture pipeline overruns.

[srinivas.kandagatla]

From: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>

Period sizes less than 6k for capture path triggers overruns in the
dsp capture pipeline.

Change the period size and number of periods to value which DSP is happy with.

Fixes: 9b4fe0f1cd79 ("ASoC: qdsp6: audioreach: add q6apm-dai support")
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
---
 sound/soc/qcom/qdsp6/q6apm-dai.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sound/soc/qcom/qdsp6/q6apm-dai.c b/sound/soc/qcom/qdsp6/q6apm-dai.c
index bb761ee0164b..aca0a98d58a5 100644
--- a/sound/soc/qcom/qdsp6/q6apm-dai.c
+++ b/sound/soc/qcom/qdsp6/q6apm-dai.c
@@ -24,8 +24,8 @@
 #define PLAYBACK_MIN_PERIOD_SIZE	128
 #define CAPTURE_MIN_NUM_PERIODS		2
 #define CAPTURE_MAX_NUM_PERIODS		8
-#define CAPTURE_MAX_PERIOD_SIZE		4096
-#define CAPTURE_MIN_PERIOD_SIZE		320
+#define CAPTURE_MAX_PERIOD_SIZE		65536
+#define CAPTURE_MIN_PERIOD_SIZE		6144
 #define BUFFER_BYTES_MAX (PLAYBACK_MAX_NUM_PERIODS * PLAYBACK_MAX_PERIOD_SIZE)
 #define BUFFER_BYTES_MIN (PLAYBACK_MIN_NUM_PERIODS * PLAYBACK_MIN_PERIOD_SIZE)
 #define COMPR_PLAYBACK_MAX_FRAGMENT_SIZE (128 * 1024)
-- 
2.39.5

[PATCH v3 6/6] ASoC: qdsp6: q6apm-dai: fix playback dsp pipeline underruns

[srinivas.kandagatla]

From: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>

QDSP has latency of 10 plus milli seconds, Which is difficult to acheive
with just 2 periods and fragment size starting at 128. Increase the number
of fragments to 8 so that it fulfils the dsp requirements

Without this patch, DSP playback pipeline seems to hit few underruns.

Fixes: 9b4fe0f1cd79 ("ASoC: qdsp6: audioreach: add q6apm-dai support")
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
---
 sound/soc/qcom/qdsp6/q6apm-dai.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sound/soc/qcom/qdsp6/q6apm-dai.c b/sound/soc/qcom/qdsp6/q6apm-dai.c
index aca0a98d58a5..aa6bd359300b 100644
--- a/sound/soc/qcom/qdsp6/q6apm-dai.c
+++ b/sound/soc/qcom/qdsp6/q6apm-dai.c
@@ -18,8 +18,8 @@
 
 #define DRV_NAME "q6apm-dai"
 
-#define PLAYBACK_MIN_NUM_PERIODS	2
-#define PLAYBACK_MAX_NUM_PERIODS	8
+#define PLAYBACK_MIN_NUM_PERIODS	8
+#define PLAYBACK_MAX_NUM_PERIODS	16
 #define PLAYBACK_MAX_PERIOD_SIZE	65536
 #define PLAYBACK_MIN_PERIOD_SIZE	128
 #define CAPTURE_MIN_NUM_PERIODS		2
-- 
2.39.5

Re: [PATCH v3 0/6] ASoC: q6apm: fix under runs and fragment sizes

[Johan Hovold]

On Tue, Mar 04, 2025 at 10:57:17AM +0000, Srinivas Kandagatla wrote:
> From: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
> 
> On Qualcomm Audioreach setup, some of the audio artifacts are seen in
> both recording and playback. These patches fix issues by
> 1. Adjusting the fragment size that dsp can service.
> 2. schedule available playback buffers in time for dsp to not hit under runs 
> 3. remove some of the manual calculations done to get hardware pointer.
> 
> With these patches, am able to see Audio quality improvements.
> 
> Any testing would be appreciated.

This fixes the playback and capture issues with Pipewire, including the
heavily distorted, choppy playback when pavucontrol is open. Turns out
that the pavucontrol volume meters reduces the output sink quantum size
from the default 1024 to 256, which was too low with v2 but now seems to
work.

Unfortunately, this series still regresses Pulseaudio as capture is now
choppy (e.g. when recording using parecord).

During our off-list debugging sessions you suggested reducing the max
capture period size (that this series increases) to 6144 (same as min)
and that fixed the Pulseaudio capture issue. In v3 the, the max period
is again increased to 65536 which appears to break Pulseaudio capture.

> Changes since v2:
> 	- dropped patch which is causing regression with pluseaudio.
> 	- setup period sizes only for capture path
> 	- fix underruns/overruns in dsp pipelines.
> 	- add fixes tag
> 	- add patch to fix buffer alignment

Johan

Re: [PATCH v3 0/6] ASoC: q6apm: fix under runs and fragment sizes

[Srinivas Kandagatla]

On 04/03/2025 17:04, Johan Hovold wrote:
> On Tue, Mar 04, 2025 at 10:57:17AM +0000, Srinivas Kandagatla wrote:
>> From: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
>>
>> On Qualcomm Audioreach setup, some of the audio artifacts are seen in
>> both recording and playback. These patches fix issues by
>> 1. Adjusting the fragment size that dsp can service.
>> 2. schedule available playback buffers in time for dsp to not hit under runs
>> 3. remove some of the manual calculations done to get hardware pointer.
>>
>> With these patches, am able to see Audio quality improvements.
>>
>> Any testing would be appreciated.
> 
> This fixes the playback and capture issues with Pipewire, including the
> heavily distorted, choppy playback when pavucontrol is open. Turns out
> that the pavucontrol volume meters reduces the output sink quantum size
> from the default 1024 to 256, which was too low with v2 but now seems to
> work.
> 
> Unfortunately, this series still regresses Pulseaudio as capture is now
> choppy (e.g. when recording using parecord).

> 
> During our off-list debugging sessions you suggested reducing the max
> capture period size (that this series increases) to 6144 (same as min)
> and that fixed the Pulseaudio capture issue. In v3 the, the max period
> is again increased to 65536 which appears to break Pulseaudio capture.

I have verified pulseaudio on my side as well, the 64 byte alignment 
should have fixed the issue on the pulseaudio side.

Did you pick up the [PATCH v3 4/6] ASoC: qdsp6: q6apm-dai: set correct 
period and buffer alignment. patch in your tests?

thanks,
Srini
> 
>> Changes since v2:
>> 	- dropped patch which is causing regression with pluseaudio.
>> 	- setup period sizes only for capture path
>> 	- fix underruns/overruns in dsp pipelines.
>> 	- add fixes tag
>> 	- add patch to fix buffer alignment
> 
> Johan

Re: [PATCH v3 0/6] ASoC: q6apm: fix under runs and fragment sizes

[Johan Hovold]

On Tue, Mar 04, 2025 at 05:12:20PM +0000, Srinivas Kandagatla wrote:
> On 04/03/2025 17:04, Johan Hovold wrote:

> > Unfortunately, this series still regresses Pulseaudio as capture is now
> > choppy (e.g. when recording using parecord).
> 
> > 
> > During our off-list debugging sessions you suggested reducing the max
> > capture period size (that this series increases) to 6144 (same as min)
> > and that fixed the Pulseaudio capture issue. In v3 the, the max period
> > is again increased to 65536 which appears to break Pulseaudio capture.
> 
> I have verified pulseaudio on my side as well, the 64 byte alignment 
> should have fixed the issue on the pulseaudio side.
> 
> Did you pick up the [PATCH v3 4/6] ASoC: qdsp6: q6apm-dai: set correct 
> period and buffer alignment. patch in your tests?

Yes, I'm running with the full series applied. I tested PA on the X13s
(and PW on the T14s).

Johan

Re: [PATCH v3 0/6] ASoC: q6apm: fix under runs and fragment sizes

[Johan Hovold]

On Tue, Mar 04, 2025 at 06:19:53PM +0100, Johan Hovold wrote:
> On Tue, Mar 04, 2025 at 05:12:20PM +0000, Srinivas Kandagatla wrote:
> > On 04/03/2025 17:04, Johan Hovold wrote:
> 
> > > Unfortunately, this series still regresses Pulseaudio as capture is now
> > > choppy (e.g. when recording using parecord).
> > > 
> > > During our off-list debugging sessions you suggested reducing the max
> > > capture period size (that this series increases) to 6144 (same as min)
> > > and that fixed the Pulseaudio capture issue. In v3 the, the max period
> > > is again increased to 65536 which appears to break Pulseaudio capture.
> > 
> > I have verified pulseaudio on my side as well, the 64 byte alignment 
> > should have fixed the issue on the pulseaudio side.
> > 
> > Did you pick up the [PATCH v3 4/6] ASoC: qdsp6: q6apm-dai: set correct 
> > period and buffer alignment. patch in your tests?
> 
> Yes, I'm running with the full series applied. I tested PA on the X13s
> (and PW on the T14s).

I just tested PA on the T14s and it has the same problem with
choppy recording with this series applied.

As with the (amended) v2, setting the max capture period size to 6144
appears to make the problem go away.

Johan

Re: [PATCH v3 0/6] ASoC: q6apm: fix under runs and fragment sizes

[Johan Hovold]

On Wed, Mar 05, 2025 at 10:06:52AM +0100, Johan Hovold wrote:
> On Tue, Mar 04, 2025 at 06:19:53PM +0100, Johan Hovold wrote:
> > On Tue, Mar 04, 2025 at 05:12:20PM +0000, Srinivas Kandagatla wrote:
> > > On 04/03/2025 17:04, Johan Hovold wrote:
> > 
> > > > Unfortunately, this series still regresses Pulseaudio as capture is now
> > > > choppy (e.g. when recording using parecord).
> > > > 
> > > > During our off-list debugging sessions you suggested reducing the max
> > > > capture period size (that this series increases) to 6144 (same as min)
> > > > and that fixed the Pulseaudio capture issue. In v3 the, the max period
> > > > is again increased to 65536 which appears to break Pulseaudio capture.
> > > 
> > > I have verified pulseaudio on my side as well, the 64 byte alignment 
> > > should have fixed the issue on the pulseaudio side.
> > > 
> > > Did you pick up the [PATCH v3 4/6] ASoC: qdsp6: q6apm-dai: set correct 
> > > period and buffer alignment. patch in your tests?
> > 
> > Yes, I'm running with the full series applied. I tested PA on the X13s
> > (and PW on the T14s).
> 
> I just tested PA on the T14s and it has the same problem with
> choppy recording with this series applied.
> 
> As with the (amended) v2, setting the max capture period size to 6144
> appears to make the problem go away.

Actually, there may still be some distortion also with the max period
size capped at 6144 with PA (there's a hint of it breaking up as with
the non-modified v3). At least the PW recording sounded cleaner, even if
I guess there could be other causes for this.
 
Johan

Re: [PATCH v3 0/6] ASoC: q6apm: fix under runs and fragment sizes

[Caleb Connolly]

Hi Srini,

On 3/4/25 10:57, srinivas.kandagatla@linaro.org wrote:
> From: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
> 
> On Qualcomm Audioreach setup, some of the audio artifacts are seen in
> both recording and playback. These patches fix issues by
> 1. Adjusting the fragment size that dsp can service.
> 2. schedule available playback buffers in time for dsp to not hit under runs
> 3. remove some of the manual calculations done to get hardware pointer.
> 
> With these patches, am able to see Audio quality improvements.
> 
> Any testing would be appreciated.

This totally breaks audio on SDM845, and often results in a hard-crash 
to crashdump mode on my OnePlus 6.

Unfortunately I can't get UART logs right now. But I imagine this is 
reproducible on RB3.

Kind regards,
> 
> thanks,
> Srini
> 
> Changes since v2:
> 	- dropped patch which is causing regression with pluseaudio.
> 	- setup period sizes only for capture path
> 	- fix underruns/overruns in dsp pipelines.
> 	- add fixes tag
> 	- add patch to fix buffer alignment
> 
> Changes since v1:
> 	- added new patches to fix the fragment size, pointer
> 	  calculations
> 	- updated to schedule only available buffers.
> 
> Srinivas Kandagatla (6):
>    ASoC: q6apm-dai: schedule all available frames to avoid dsp under-runs
>    ASoC: q6apm: add q6apm_get_hw_pointer helper
>    ASoC: q6apm-dai: make use of q6apm_get_hw_pointer
>    ASoC: qdsp6: q6apm-dai: set correct period and buffer alignment.
>    ASoC: qdsp6: q6apm-dai: fix capture pipeline overruns.
>    ASoC: qdsp6: q6apm-dai: fix playback dsp pipeline underruns
> 
>   sound/soc/qcom/qdsp6/q6apm-dai.c | 63 +++++++++++++++++---------------
>   sound/soc/qcom/qdsp6/q6apm.c     | 18 ++++++++-
>   sound/soc/qcom/qdsp6/q6apm.h     |  3 ++
>   3 files changed, 53 insertions(+), 31 deletions(-)
> 

-- 
Caleb (they/them)

Re: [PATCH v3 0/6] ASoC: q6apm: fix under runs and fragment sizes

[Srinivas Kandagatla]

On 05/03/2025 22:17, Caleb Connolly wrote:
> Hi Srini,
> 
> On 3/4/25 10:57, srinivas.kandagatla@linaro.org wrote:
>> From: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
>>
>> On Qualcomm Audioreach setup, some of the audio artifacts are seen in
>> both recording and playback. These patches fix issues by
>> 1. Adjusting the fragment size that dsp can service.
>> 2. schedule available playback buffers in time for dsp to not hit 
>> under runs
>> 3. remove some of the manual calculations done to get hardware pointer.
>>
>> With these patches, am able to see Audio quality improvements.
>>
>> Any testing would be appreciated.
> 
> This totally breaks audio on SDM845, and often results in a hard-crash 
> to crashdump mode on my OnePlus 6.

Its not possible.. as SDM845 is Elite DSP architecture, all of these 
patches changes are very specific to Audioreach DSP architecture.

Could you please share the crash logs if you have any. It highly likely 
that something else is triggering the hard-crash that you reported.

thanks,
Srini
> 
> Unfortunately I can't get UART logs right now. But I imagine this is 
> reproducible on RB3.
> 
> Kind regards,
>>
>> thanks,
>> Srini
>>
>> Changes since v2:
>>     - dropped patch which is causing regression with pluseaudio.
>>     - setup period sizes only for capture path
>>     - fix underruns/overruns in dsp pipelines.
>>     - add fixes tag
>>     - add patch to fix buffer alignment
>>
>> Changes since v1:
>>     - added new patches to fix the fragment size, pointer
>>       calculations
>>     - updated to schedule only available buffers.
>>
>> Srinivas Kandagatla (6):
>>    ASoC: q6apm-dai: schedule all available frames to avoid dsp under-runs
>>    ASoC: q6apm: add q6apm_get_hw_pointer helper
>>    ASoC: q6apm-dai: make use of q6apm_get_hw_pointer
>>    ASoC: qdsp6: q6apm-dai: set correct period and buffer alignment.
>>    ASoC: qdsp6: q6apm-dai: fix capture pipeline overruns.
>>    ASoC: qdsp6: q6apm-dai: fix playback dsp pipeline underruns
>>
>>   sound/soc/qcom/qdsp6/q6apm-dai.c | 63 +++++++++++++++++---------------
>>   sound/soc/qcom/qdsp6/q6apm.c     | 18 ++++++++-
>>   sound/soc/qcom/qdsp6/q6apm.h     |  3 ++
>>   3 files changed, 53 insertions(+), 31 deletions(-)
>>
> 

Re: [PATCH v3 0/6] ASoC: q6apm: fix under runs and fragment sizes

[Dmitry Baryshkov]

On Thu, Mar 06, 2025 at 01:15:43PM +0000, Srinivas Kandagatla wrote:
> 
> 
> On 05/03/2025 22:17, Caleb Connolly wrote:
> > Hi Srini,
> > 
> > On 3/4/25 10:57, srinivas.kandagatla@linaro.org wrote:
> > > From: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
> > > 
> > > On Qualcomm Audioreach setup, some of the audio artifacts are seen in
> > > both recording and playback. These patches fix issues by
> > > 1. Adjusting the fragment size that dsp can service.
> > > 2. schedule available playback buffers in time for dsp to not hit
> > > under runs
> > > 3. remove some of the manual calculations done to get hardware pointer.
> > > 
> > > With these patches, am able to see Audio quality improvements.
> > > 
> > > Any testing would be appreciated.
> > 
> > This totally breaks audio on SDM845, and often results in a hard-crash
> > to crashdump mode on my OnePlus 6.
> 
> Its not possible.. as SDM845 is Elite DSP architecture, all of these patches
> changes are very specific to Audioreach DSP architecture.
> 
> Could you please share the crash logs if you have any. It highly likely that
> something else is triggering the hard-crash that you reported.

I will try reproducing this on RB3 today in the evening.

> 
> thanks,
> Srini
> > 
> > Unfortunately I can't get UART logs right now. But I imagine this is
> > reproducible on RB3.
> > 
> > Kind regards,
> > > 
> > > thanks,
> > > Srini
> > > 
> > > Changes since v2:
> > >     - dropped patch which is causing regression with pluseaudio.
> > >     - setup period sizes only for capture path
> > >     - fix underruns/overruns in dsp pipelines.
> > >     - add fixes tag
> > >     - add patch to fix buffer alignment
> > > 
> > > Changes since v1:
> > >     - added new patches to fix the fragment size, pointer
> > >       calculations
> > >     - updated to schedule only available buffers.
> > > 
> > > Srinivas Kandagatla (6):
> > >    ASoC: q6apm-dai: schedule all available frames to avoid dsp under-runs
> > >    ASoC: q6apm: add q6apm_get_hw_pointer helper
> > >    ASoC: q6apm-dai: make use of q6apm_get_hw_pointer
> > >    ASoC: qdsp6: q6apm-dai: set correct period and buffer alignment.
> > >    ASoC: qdsp6: q6apm-dai: fix capture pipeline overruns.
> > >    ASoC: qdsp6: q6apm-dai: fix playback dsp pipeline underruns
> > > 
> > >   sound/soc/qcom/qdsp6/q6apm-dai.c | 63 +++++++++++++++++---------------
> > >   sound/soc/qcom/qdsp6/q6apm.c     | 18 ++++++++-
> > >   sound/soc/qcom/qdsp6/q6apm.h     |  3 ++
> > >   3 files changed, 53 insertions(+), 31 deletions(-)
> > > 
> > 

-- 
With best wishes
Dmitry

Re: [PATCH v3 0/6] ASoC: q6apm: fix under runs and fragment sizes

[Krzysztof Kozlowski]

On 04/03/2025 11:57, srinivas.kandagatla@linaro.org wrote:
> From: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
> 
> On Qualcomm Audioreach setup, some of the audio artifacts are seen in
> both recording and playback. These patches fix issues by
> 1. Adjusting the fragment size that dsp can service.
> 2. schedule available playback buffers in time for dsp to not hit under runs 
> 3. remove some of the manual calculations done to get hardware pointer.
> 
> With these patches, am able to see Audio quality improvements.
> 
> Any testing would be appreciated.

Worked well for my pipewire setup on T14s and X1E8100 CRD.

Tested-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>

Best regards,
Krzysztof