Skip to contentSkip to conversation

ASoC: SDCA: Add PDE verification reusable helper

PATCH Superseded Niranjan H Y 3 participants 12 messages Apr 17 – Apr 21

Subsystems (4)

v9 of 16 Superseded Latest is v16 →

lore
Share & tools
lore
#0 lore raw

PATCH v9 1/4 ASoC: SDCA: Add PDE verification reusable helper

Implement sdca_asoc_pde_ensure_ps() helper function to poll for PDE
power state transitions. Per SDCA specification, after writing REQUESTED_PS, drivers must poll ACTUAL_PS until the target power state is reached.

Changes include:
- Add sdca_asoc_pde_ensure_ps() to handle ACTUAL_PS polling with support for device-specific delay tables or default intervals
- Export function via sdca_asoc.h for use by SDCA-compliant drivers
- Refactor entity_pde_event() in sdca_asoc.c to use the helper

---
v9:
- newly added interface

---
 include/sound/sdca_asoc.h  |   9 ++++
 sound/soc/sdca/sdca_asoc.c | 105 ++++++++++++++++++++++++++-----------
 2 files changed, 82 insertions(+), 32 deletions(-)
include/sound/sdca_asoc.h +9 -0
index 46a61a52decc5..6c1c8af514122 100644
@@ -99,4 +99,13 @@ int sdca_asoc_q78_put_volsw(struct snd_kcontrol *kcontrol,
 			    struct snd_ctl_elem_value *ucontrol);
 int sdca_asoc_q78_get_volsw(struct snd_kcontrol *kcontrol,
 			    struct snd_ctl_elem_value *ucontrol);
+
+struct device;
+struct sdca_pde_delay;
+
+int sdca_asoc_pde_ensure_ps(struct device *dev, struct regmap *regmap,
+			    int function_id, int entity_id,
+			    int from_ps, int to_ps,
+			    const struct sdca_pde_delay *pde_delays,
+			    int num_delays);
 #endif // __SDCA_ASOC_H__
sound/soc/sdca/sdca_asoc.c +73 -32

122 lines not on this page · raw message

Signature · 1 line
2.34.1
#1 lore raw

PATCH v9 2/4 ASoC: tac5xx2-sdw: add soundwire based codec driver

Add codec driver for tac5xx2 family of devices. This includes the support for
1. tac5572 - DAC, PDM, UAJ and HID
2. tas2883 - Amplifier with DSP
3. tac5672 - Similar to tac5572 with feedback
4. tac5682 - Similar to tac5672 with Amplifier DSP.

---
v9:
- Update tac5xx2-sdw.c hw_params() and hw_free() to use new SDCA helper sdca_asoc_pde_ensure_ps
- Remove redundant PPU11 initialization as it is there as part of the initialisation table.
- remove unnecessary default value writes for latch clear as the register.
- Improve jack detection and PDE power state transition comments

v8:
- disable interrupts in .set_jack if jack is set to NULL
- hold uaj_lock in .set_jack
- remove unwanted interrupt enable in initialisation

v7:
- refactor remove the cx11, cs18 and cs11 they are not required to be configured for mic
- fix error handling and interrupt cleanup in .set_jack
- fix typo in Kconfig

v6:
- move firmware caching trigger from soundwire bus callback to soundwire probe and make it asynchronous
- validate the firmware fully before starting firmware.
- drop mixer control for cx11 configuration
- check the actual power state in .hw_free after requesting power state transition to power off
- use dev_is_pci() and remove CONFIG_PCI

v5:
- add missing pde_lock while accessing the current power state.
- add uaj_lock to serialize uaj detection update
- return error, if any, returned by sdw_stream_add_slave in hw_parms
- drop snd_soc_unregister_component as we are using devm_snd_soc_register_component
- fix issue in error handling for invalid firmware
- remove unsupported line in/out detection in jack detection
- refactor the jack and button detection logic to be based on their respective interrupt

v4:
- make volume controls as stereo controls for amp gain, playback, dmic and uaj playback.
- better error handling for "CX11 CS Select" mixer control.
- use switch statement for scalability in hw_params and hw_free
- remove unnecessary else in tac_io_init
- change error log for unsupported sample rates to debug logs
- make firmware binary parsing robust by checking the fw size during caching and parsing

v3:
- use macro SDCA_SINGLE_Q78_TLV for the volume mixer controls
- replace magic number with macro for current owner
- fix debug log message about the current owner for hid event

v2:
- Define and use consistent macros for register access to improve code readability and maintainability
- Replace complex event handlers with simpler DAPM widget implementations using direct register control for muting/unmuting and enabling/disabling components
- Replace custom volume controls with standard Q7.8 controls from the SDCA framework
- Fix PDE power state transitions with proper verification of actual power state
- Combine separate locks into a single ops\_lock to protect both firmware operations and power state transitions
- Clean up interrupt handling by removing unnecessary register accesses
- Fix headset microphone control by using the correct mono channel value
- Add comments to clarify firmware handling and critical sections

---
 sound/soc/codecs/Kconfig       |   11 +
 sound/soc/codecs/Makefile      |    2 +
 sound/soc/codecs/tac5xx2-sdw.c | 2023 ++++++++++++++++++++++++++++++++
 sound/soc/codecs/tac5xx2.h     |  259 ++++
 4 files changed, 2295 insertions(+)
 create mode 100644 sound/soc/codecs/tac5xx2-sdw.c
 create mode 100644 sound/soc/codecs/tac5xx2.h
sound/soc/codecs/Kconfig +11 -0
index cf94a1c756e09..176f1b061f66d 100644
@@ -264,6 +264,7 @@ config SND_SOC_ALL_CODECS
 	imply SND_SOC_STA529
 	imply SND_SOC_STAC9766
 	imply SND_SOC_STI_SAS
+	imply SND_SOC_TAC5XX2_SDW
 	imply SND_SOC_TAS2552
 	imply SND_SOC_TAS2562
 	imply SND_SOC_TAS2764
@@ -2139,6 +2140,16 @@ config SND_SOC_STAC9766
 config SND_SOC_STI_SAS
 	tristate "codec Audio support for STI SAS codec"
 
+config SND_SOC_TAC5XX2_SDW
+	tristate "Texas Instruments TAC5XX2 SoundWire Smart Amplifier"
+	depends on SOUNDWIRE
+	depends on SND_SOC_SDCA
+	help
+	  This option enables support for Texas Instruments TAC5XX2 family
+	  of SoundWire Smart Amplifiers. This includes TAC5572, TAC5672,
+	  TAC5682 and TAS2883. To compile this driver as a module, choose
+	  M here: the module will be called snd-soc-tac5xx2.
+
 config SND_SOC_TAS2552
 	tristate "Texas Instruments TAS2552 Mono Audio amplifier"
 	depends on I2C
sound/soc/codecs/Makefile +2 -0
index 172861d17cfd0..f64e75dd2f35d 100644
@@ -313,6 +313,7 @@ snd-soc-sta350-y := sta350.o
 snd-soc-sta529-y := sta529.o
 snd-soc-stac9766-y := stac9766.o
 snd-soc-sti-sas-y := sti-sas.o
+snd-soc-tac5xx2-sdw-y := tac5xx2-sdw.o
 snd-soc-tas5086-y := tas5086.o
 snd-soc-tas571x-y := tas571x.o
 snd-soc-tas5720-y := tas5720.o
@@ -746,6 +747,7 @@ obj-$(CONFIG_SND_SOC_STA350)   += snd-soc-sta350.o
 obj-$(CONFIG_SND_SOC_STA529)   += snd-soc-sta529.o
 obj-$(CONFIG_SND_SOC_STAC9766)	+= snd-soc-stac9766.o
 obj-$(CONFIG_SND_SOC_STI_SAS)	+= snd-soc-sti-sas.o
+obj-$(CONFIG_SND_SOC_TAC5XX2_SDW)	+= snd-soc-tac5xx2-sdw.o
 obj-$(CONFIG_SND_SOC_TAS2552)	+= snd-soc-tas2552.o
 obj-$(CONFIG_SND_SOC_TAS2562)	+= snd-soc-tas2562.o
 obj-$(CONFIG_SND_SOC_TAS2764)	+= snd-soc-tas2764.o
sound/soc/codecs/tac5xx2-sdw.c new file +2023 -0

2,023 lines not on this page · raw message

sound/soc/codecs/tac5xx2.h new file +259 -0

259 lines not on this page · raw message

Signature · 1 line
2.34.1
#2 lore raw

Re: [PATCH v9 2/4] ASoC: tac5xx2-sdw: add soundwire based codec driver

Quoted 4 lines
sound/soc/codecs/tac5xx2-sdw.c new file +3 -0
+struct tac5xx2_prv {
+	struct snd_soc_component *component;
+	struct sdw_slave *sdw_peripheral;

sound/soc/codecs/tac5xx2-sdw.c line 139

Quoted 10 lines
sound/soc/codecs/tac5xx2-sdw.c new file +10 -0
+	struct sdca_function_data *sa_func_data;
+	struct sdca_function_data *sm_func_data;
+	struct sdca_function_data *uaj_func_data;
+	struct sdca_function_data *hid_func_data;
+	enum sdw_slave_status status;
+	/* Lock for firmware download and PDE state transitions.
+	 * Serializes FW caching/download and DAPM-driven power
+	 * state changes to prevent PDE operations during firmware load.
+	 */
+	struct mutex pde_lock;

that's a lot of stuff that's protected with this lock. See below for one question...

Quoted 46 lines
sound/soc/codecs/tac5xx2-sdw.c new file +46 -0
+static int tac_sdw_hw_params(struct snd_pcm_substream *substream,
+			     struct snd_pcm_hw_params *params,
+			     struct snd_soc_dai *dai)
+{
+	struct snd_soc_component *component = dai->component;
+	struct tac5xx2_prv *tac_dev = snd_soc_component_get_drvdata(component);
+	struct sdw_stream_config stream_config = {0};
+	struct sdw_port_config port_config = {0};
+	struct sdw_stream_runtime *sdw_stream;
+	struct sdw_slave *sdw_peripheral = tac_dev->sdw_peripheral;
+	unsigned long time;
+	int ret;
+	int function_id;
+	int pde_entity;
+	int port_num;
+	u8 sample_rate_idx = 0;
+
+	time = wait_for_completion_timeout(&sdw_peripheral->initialization_complete,
+					   msecs_to_jiffies(TAC5XX2_PROBE_TIMEOUT_MS));
+	if (!time) {
+		dev_warn(tac_dev->dev, "%s: hw initialization timeout\n", __func__);
+		return -ETIMEDOUT;
+	}
+	if (!tac_dev->hw_init) {
+		dev_err(tac_dev->dev,
+			"error: operation without hw initialization");
+		return -EINVAL;
+	}
+
+	sdw_stream = snd_soc_dai_get_dma_data(dai, substream);
+	if (!sdw_stream) {
+		dev_err(tac_dev->dev, "failed to get dma data");
+		return -EINVAL;
+	}
+
+	ret = tac_clear_latch(tac_dev);
+	if (ret)
+		dev_warn(tac_dev->dev, "clear latch failed, err=%d", ret);
+
+	switch (dai->id) {
+	case TAC5XX2_DMIC:
+		function_id = TAC_FUNCTION_ID_SM;
+		pde_entity = TAC_SDCA_ENT_PDE11;
+		port_num = TAC_SDW_PORT_NUM_DMIC;
+		break;
+	case TAC5XX2_UAJ:

sound/soc/codecs/tac5xx2-sdw.c line 732

Quoted 10 lines
sound/soc/codecs/tac5xx2-sdw.c new file +10 -0
+		function_id = TAC_FUNCTION_ID_UAJ;
+		pde_entity = substream->stream == SNDRV_PCM_STREAM_PLAYBACK ?
+				TAC_SDCA_ENT_PDE47 : TAC_SDCA_ENT_PDE34;
+		port_num = substream->stream == SNDRV_PCM_STREAM_PLAYBACK ?
+				TAC_SDW_PORT_NUM_UAJ_PLAYBACK :
+				TAC_SDW_PORT_NUM_UAJ_CAPTURE;
+		/* Detect and set jack type for UAJ path before playback.
+		 * This is required as jack detection does not trigger interrupt
+		 * when device is in runtime_pm suspend with bus in clock stop mode.
+		 */

so here we have an interesting logic - or I misunderstood the comment?

If a headset is inserted when the device is in runtime_pm suspend, how would applications modify the routing and select playback on the headset, which would then ripple down to this hw_params() call?

IOW to play on a headset you first have to know there's a headset.

Quoted 74 lines
sound/soc/codecs/tac5xx2-sdw.c new file +74 -0
+		mutex_lock(&tac_dev->uaj_lock);
+		tac5xx2_sdca_headset_detect(tac_dev);
+		mutex_unlock(&tac_dev->uaj_lock);
+		break;
+	case TAC5XX2_SPK:
+		function_id = TAC_FUNCTION_ID_SA;
+		pde_entity = TAC_SDCA_ENT_PDE23;
+		port_num = substream->stream == SNDRV_PCM_STREAM_PLAYBACK ?
+				TAC_SDW_PORT_NUM_SPK_PLAYBACK :
+				TAC_SDW_PORT_NUM_SPK_CAPTURE;
+		break;
+	default:
+		dev_err(tac_dev->dev, "Invalid dai id: %d", dai->id);
+		return -EINVAL;
+	}
+
+	snd_sdw_params_to_config(substream, params, &stream_config, &port_config);
+	port_config.num = port_num;
+	ret = sdw_stream_add_slave(sdw_peripheral, &stream_config,
+				   &port_config, 1, sdw_stream);
+	if (ret) {
+		dev_err(dai->dev,
+			"Unable to configure port %d: %d\n", port_num, ret);
+		return ret;
+	}
+
+	switch (params_rate(params)) {
+	case 48000:
+		sample_rate_idx = 0x01;
+		break;
+	case 44100:
+		sample_rate_idx = 0x02;
+		break;
+	case 96000:
+		sample_rate_idx = 0x03;
+		break;
+	case 88200:
+		sample_rate_idx = 0x04;
+		break;
+	default:
+		dev_dbg(tac_dev->dev, "Unsupported sample rate: %d Hz",
+			params_rate(params));
+		return -EINVAL;
+	}
+
+	switch (function_id) {
+	case TAC_FUNCTION_ID_SM:
+		ret = regmap_write(tac_dev->regmap,
+				   SDW_SDCA_CTL(function_id, TAC_SDCA_ENT_CS113,
+						TAC_SDCA_CTL_CS_SAMP_RATE_IDX, 0),
+			sample_rate_idx);
+		if (ret) {
+			dev_err(tac_dev->dev, "Failed to set CS113 sample rate: %d", ret);
+			return ret;
+		}
+
+		break;
+	case TAC_FUNCTION_ID_UAJ:
+		if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) {
+			ret = regmap_write(tac_dev->regmap,
+					   SDW_SDCA_CTL(function_id, TAC_SDCA_ENT_CS41,
+							TAC_SDCA_CTL_CS_SAMP_RATE_IDX, 0),
+					sample_rate_idx);
+			if (ret) {
+				dev_err(tac_dev->dev, "Failed to set CS41 sample rate: %d", ret);
+				return ret;
+			}
+		} else {
+			ret = regmap_write(tac_dev->regmap,
+					   SDW_SDCA_CTL(function_id, TAC_SDCA_ENT_CS36,
+							TAC_SDCA_CTL_CS_SAMP_RATE_IDX, 0),
+					sample_rate_idx);
+			if (ret) {
+				dev_err(tac_dev->dev, "Failed to set CS36 sample rate: %d", ret);

sound/soc/codecs/tac5xx2-sdw.c line 816

Quoted 10 lines
sound/soc/codecs/tac5xx2-sdw.c new file +10 -0
+				return ret;
+			}
+		}
+		break;
+	case TAC_FUNCTION_ID_SA:
+		/* SmartAmp: no additional sample rate configuration needed */
+		break;
+	}
+
+	guard(mutex)(&tac_dev->pde_lock);

question I mentioned above: when you reach the hw_params phase, do you really have a potential race with firmware download? What does this specific use of pde_lock protect against?

Quoted 38 lines
sound/soc/codecs/tac5xx2-sdw.c new file +38 -0
+	ret = regmap_write(tac_dev->regmap,
+			   SDW_SDCA_CTL(function_id, pde_entity,
+					TAC_SDCA_REQUESTED_PS, 0), 0);
+	if (ret) {
+		dev_err(tac_dev->dev, "failed to set PS to 0: %d\n", ret);
+		return ret;
+	}
+
+	ret = sdca_asoc_pde_ensure_ps(tac_dev->dev, tac_dev->regmap,
+				      function_id, pde_entity,
+				      SDCA_PDE_PS3, SDCA_PDE_PS0,
+				      NULL, 0);
+	if (ret)
+		dev_err(tac_dev->dev,
+			"failed to transition to PS0, err= %d\n", ret);
+	return ret;
+}
+
+static s32 tac_sdw_pcm_hw_free(struct snd_pcm_substream *substream,
+			       struct snd_soc_dai *dai)
+{
+	s32 ret;
+	struct snd_soc_component *component = dai->component;
+	struct tac5xx2_prv *tac_dev =
+		snd_soc_component_get_drvdata(component);
+	struct sdw_stream_runtime *sdw_stream =
+		snd_soc_dai_get_dma_data(dai, substream);
+	int pde_entity, function_id;
+
+	sdw_stream_remove_slave(tac_dev->sdw_peripheral, sdw_stream);
+
+	switch (dai->id) {
+	case TAC5XX2_DMIC:
+		pde_entity = TAC_SDCA_ENT_PDE11;
+		function_id = TAC_FUNCTION_ID_SM;
+		break;
+	case TAC5XX2_UAJ:
+		function_id = TAC_FUNCTION_ID_UAJ;

sound/soc/codecs/tac5xx2-sdw.c line 864

Quoted 10 lines
sound/soc/codecs/tac5xx2-sdw.c new file +10 -0
+		pde_entity = substream->stream == SNDRV_PCM_STREAM_PLAYBACK ?
+				TAC_SDCA_ENT_PDE47 : TAC_SDCA_ENT_PDE34;
+		break;
+	default:
+		function_id = TAC_FUNCTION_ID_SA;
+		pde_entity = TAC_SDCA_ENT_PDE23;
+		break;
+	}
+
+	guard(mutex)(&tac_dev->pde_lock);

same here, do you really have a race with firmware download?

Or is this a case of dependencies between functions that requires all power state transitions to be serialized?

Quoted 92 lines
sound/soc/codecs/tac5xx2-sdw.c new file +92 -0
+	ret = regmap_write(tac_dev->regmap,
+			   SDW_SDCA_CTL(function_id, pde_entity, TAC_SDCA_REQUESTED_PS, 0),
+			   SDCA_PDE_PS3);
+	if (ret)
+		return ret;
+
+	ret = sdca_asoc_pde_ensure_ps(tac_dev->dev, tac_dev->regmap,
+				      function_id, pde_entity,
+				      SDCA_PDE_PS0, SDCA_PDE_PS3,
+				      NULL, 0);
+	if (ret)
+		dev_err(tac_dev->dev, "failed to trasition from PS0 to PS3");
+	return ret;
+}
+
+static const struct snd_soc_dai_ops tac_dai_ops = {
+	.hw_params = tac_sdw_hw_params,
+	.hw_free = tac_sdw_pcm_hw_free,
+	.set_stream = tac_set_sdw_stream,
+	.shutdown = tac_sdw_shutdown,
+};
+
+static int tac5xx2_sdca_btn_type(unsigned char *buffer, struct tac5xx2_prv *tac_dev)
+{
+	switch (*buffer) {
+	case 1: /* play pause */
+		return SND_JACK_BTN_0;
+	case 10: /* vol down */
+		return SND_JACK_BTN_3;
+	case 8: /* vol up */
+		return SND_JACK_BTN_2;
+	case 4: /* long press */
+		return SND_JACK_BTN_1;
+	case 2: /* next song */
+	case 32: /* next song */
+		return SND_JACK_BTN_4;
+	default:
+		return 0;
+	}
+}
+
+static int tac5xx2_sdca_button_detect(struct tac5xx2_prv *tac_dev)
+{
+	unsigned int btn_type, offset, idx;
+	int ret, value, owner;
+	u8 buf[2];
+
+	ret = regmap_read(tac_dev->regmap,
+			  SDW_SDCA_CTL(TAC_FUNCTION_ID_HID, TAC_SDCA_ENT_HID1,
+				       TAC_SDCA_CTL_HIDTX_CURRENT_OWNER, 0), &owner);
+	if (ret) {
+		dev_err(tac_dev->dev,
+			"Failed to read current UMP message owner 0x%x", ret);
+		return ret;
+	}
+
+	if (owner == SDCA_UMP_OWNER_DEVICE) {
+		dev_dbg(tac_dev->dev, "skip button detect as current owner is not host\n");
+		return 0;
+	}
+
+	ret = regmap_read(tac_dev->regmap,
+			  SDW_SDCA_CTL(TAC_FUNCTION_ID_HID, TAC_SDCA_ENT_HID1,
+				       TAC_SDCA_CTL_HIDTX_MESSAGE_OFFSET, 0), &offset);
+	if (ret) {
+		dev_err(tac_dev->dev,
+			"Failed to read current UMP message offset: %d", ret);
+		goto end_btn_det;
+	}
+
+	dev_dbg(tac_dev->dev, "button detect: message offset = %x", offset);
+
+	for (idx = 0; idx < sizeof(buf); idx++) {
+		ret = regmap_read(tac_dev->regmap,
+				  TAC_BUF_ADDR_HID1 + offset + idx, &value);
+		if (ret) {
+			dev_err(tac_dev->dev,
+				"Failed to read HID buffer: %d", ret);
+			goto end_btn_det;
+		}
+		buf[idx] = value & 0xff;
+	}
+
+	if (buf[0] == 0x1) {
+		btn_type = tac5xx2_sdca_btn_type(&buf[1], tac_dev);
+		ret = btn_type;
+	}
+
+end_btn_det:
+	regmap_write(tac_dev->regmap,
+		     SDW_SDCA_CTL(TAC_FUNCTION_ID_HID, TAC_SDCA_ENT_HID1,
+				  TAC_SDCA_CTL_HIDTX_CURRENT_OWNER, 0), 0x01);

sound/soc/codecs/tac5xx2-sdw.c line 966

Quoted 10 lines
sound/soc/codecs/tac5xx2-sdw.c new file +10 -0
+
+	return ret;
+}
+
+static int tac5xx2_sdca_headset_detect(struct tac5xx2_prv *tac_dev)
+{
+	int val, ret;
+
+	if (!tac_has_uaj_support(tac_dev))
+		return 0;

can this really happen? usually you try to detect a headset if the device is capable of dealing with headsets, no? Should this test be moved at a higher level before you enable low-level handling of headset stuff?

Quoted 31 lines
sound/soc/codecs/tac5xx2-sdw.c new file +31 -0
+	ret = regmap_read(tac_dev->regmap,
+			  SDW_SDCA_CTL(TAC_FUNCTION_ID_UAJ, TAC_SDCA_ENT_GE35,
+				       TAC_SDCA_CTL_DET_MODE, 0), &val);
+	if (ret) {
+		dev_err(tac_dev->dev, "Failed to read the detect mode");
+		return ret;
+	}
+
+	switch (val) {
+	case 4:
+		tac_dev->jack_type = SND_JACK_MICROPHONE;
+		break;
+	case 5:
+		tac_dev->jack_type = SND_JACK_HEADPHONE;
+		break;
+	case 6:
+		tac_dev->jack_type = SND_JACK_HEADSET;
+		break;
+	case 0:
+	default:
+		tac_dev->jack_type = 0;
+		break;
+	}
+
+	ret = regmap_write(tac_dev->regmap,
+			   SDW_SDCA_CTL(TAC_FUNCTION_ID_UAJ, TAC_SDCA_ENT_GE35,
+					TAC_SDCA_CTL_SEL_MODE, 0), val);
+	if (ret)
+		dev_err(tac_dev->dev, "Failed to update the jack type to device");
+
+	return 0;

sound/soc/codecs/tac5xx2-sdw.c line 1008

Quoted 10 lines
sound/soc/codecs/tac5xx2-sdw.c new file +10 -0
+}
+
+static int tac5xx2_set_jack(struct snd_soc_component *component,
+			    struct snd_soc_jack *hs_jack, void *data)
+{
+	struct tac5xx2_prv *tac_dev = snd_soc_component_get_drvdata(component);
+	int ret;
+
+	if (!tac_has_uaj_support(tac_dev))
+		return 0;

same here, shouldn't set_jack() be added to the component callbacks before probe?

Quoted 44 lines
sound/soc/codecs/tac5xx2-sdw.c new file +43 -0
+	guard(mutex)(&tac_dev->uaj_lock);
+	if (!hs_jack) {
+		if (tac_dev->hs_jack) {
+			tac_dev->hs_jack = NULL;
+			ret = 0;
+			goto disable_interrupts;
+		}
+		return 0;
+	}
+
+	tac_dev->hs_jack = hs_jack;
+	if (!tac_dev->hw_init) {
+		dev_err(tac_dev->dev, "jack init failed, hw not initialized");
+		return 0;
+	}
+
+	ret = regmap_write(tac_dev->regmap, SDW_SCP_SDCA_INTMASK2,
+			   SDW_SCP_SDCA_INTMASK_SDCA_11);
+	if (ret) {
+		dev_warn(tac_dev->dev,
+			 "Failed to register jack detection interrupt");
+		goto disable_interrupts;
+	}
+
+	ret = regmap_write(tac_dev->regmap, SDW_SCP_SDCA_INTMASK3,
+			   SDW_SCP_SDCA_INTMASK_SDCA_16);
+	if (ret) {
+		dev_warn(tac_dev->dev,
+			 "Failed to register for button detect interrupt");
+		goto disable_interrupts;
+	}
+
+	return 0;
+
+disable_interrupts:
+	/* ignore errors while disabling interrupts */
+	regmap_write(tac_dev->regmap, SDW_SCP_SDCA_INTMASK2, 0);
+	regmap_write(tac_dev->regmap, SDW_SCP_SDCA_INTMASK3, 0);
+
+	return ret;
+}
 
+static const struct snd_soc_component_driver soc_codec_driver_tacdevice = {
+	.probe = tac_component_probe,

sound/soc/codecs/tac5xx2-sdw.c

Quoted 10 lines
sound/soc/codecs/tac5xx2-sdw.c new file +10 -0
+	.remove = tac_component_remove,
+	.controls = tac5xx2_snd_controls,
+	.num_controls = ARRAY_SIZE(tac5xx2_snd_controls),
+	.dapm_widgets = tac5xx2_common_widgets,
+	.num_dapm_widgets = ARRAY_SIZE(tac5xx2_common_widgets),
+	.dapm_routes = tac5xx2_common_routes,
+	.num_dapm_routes = ARRAY_SIZE(tac5xx2_common_routes),
+	.idle_bias_on = 0,
+	.endianness = 1,
+	.set_jack = tac5xx2_set_jack,

maybe make this dynamic and only populate .set_jack in tac_init() below when you can deal with a jack?

Quoted 44 lines
sound/soc/codecs/tac5xx2-sdw.c new file +44 -0
+};
+
+static s32 tac_init(struct tac5xx2_prv *tac_dev)
+{
+	s32 ret;
+	struct snd_soc_dai_driver *dai_drv;
+	int num_dais;
+
+	dev_set_drvdata(tac_dev->dev, tac_dev);
+
+	switch (tac_dev->part_id) {
+	case 0x5572:
+		dai_drv = tac5572_dai_driver;
+		num_dais = ARRAY_SIZE(tac5572_dai_driver);
+		break;
+	case 0x5672:
+		dai_drv = tac5672_dai_driver;
+		num_dais = ARRAY_SIZE(tac5672_dai_driver);
+		break;
+	case 0x5682:
+		dai_drv = tac5682_dai_driver;
+		num_dais = ARRAY_SIZE(tac5682_dai_driver);
+		break;
+	case 0x2883:
+		dai_drv = tas2883_dai_driver;
+		num_dais = ARRAY_SIZE(tas2883_dai_driver);
+		break;
+	default:
+		dev_err(tac_dev->dev, "Unsupported device: 0x%x\n",
+			tac_dev->part_id);
+		return -EINVAL;
+	}
+
+	ret = devm_snd_soc_register_component(tac_dev->dev,
+					      &soc_codec_driver_tacdevice,
+					      dai_drv, num_dais);
+	if (ret) {
+		dev_err(tac_dev->dev, "%s: codec register error:%d.\n",
+			__func__, ret);
+		return ret;
+	}
+
+	return 0;
+}
#3 lore raw

Re: [PATCH v9 2/4] ASoC: tac5xx2-sdw: add soundwire based codec driver

Quoted 41 lines · Pierre-Louis Bossart

Subject: Re: [PATCH v9 2/4] ASoC: tac5xx2-sdw: add soundwire based codec driver

sound/soc/codecs/tac5xx2-sdw.c line 139

Quoted 13 lines
sound/soc/codecs/tac5xx2-sdw.c new file +13 -0
+struct tac5xx2_prv {
+	struct snd_soc_component *component;
+	struct sdw_slave *sdw_peripheral;
+	struct sdca_function_data *sa_func_data;
+	struct sdca_function_data *sm_func_data;
+	struct sdca_function_data *uaj_func_data;
+	struct sdca_function_data *hid_func_data;
+	enum sdw_slave_status status;
+	/* Lock for firmware download and PDE state transitions.
+	 * Serializes FW caching/download and DAPM-driven power
+	 * state changes to prevent PDE operations during firmware load.
+	 */
+	struct mutex pde_lock;

that's a lot of stuff that's protected with this lock. See below for one question...

sound/soc/codecs/tac5xx2-sdw.c

Quoted 6 lines
sound/soc/codecs/tac5xx2-sdw.c new file +6 -0
+static int tac_sdw_hw_params(struct snd_pcm_substream *substream,
+			     struct snd_pcm_hw_params *params,
+			     struct snd_soc_dai *dai)
+{
+	struct snd_soc_component *component = dai->component;
+	struct tac5xx2_prv *tac_dev =

snd_soc_component_get_drvdata(component);

Quoted 14 lines
sound/soc/codecs/tac5xx2-sdw.c new file +12 -0
+	struct sdw_stream_config stream_config = {0};
+	struct sdw_port_config port_config = {0};
+	struct sdw_stream_runtime *sdw_stream;
+	struct sdw_slave *sdw_peripheral = tac_dev->sdw_peripheral;
+	unsigned long time;
+	int ret;
+	int function_id;
+	int pde_entity;
+	int port_num;
+	u8 sample_rate_idx = 0;
+
+	time = wait_for_completion_timeout(&sdw_peripheral-

initialization_complete,
+

....

Quoted 8 lines

sound/soc/codecs/tac5xx2-sdw.c

Quoted 3 lines
sound/soc/codecs/tac5xx2-sdw.c new file +3 -0
+		/* Detect and set jack type for UAJ path before playback.
+		 * This is required as jack detection does not trigger interrupt
+		 * when device is in runtime_pm suspend with bus in clock stop

mode.

Review

Quoted 1 line
(quoted diff) +1 -0
+		 */

so here we have an interesting logic - or I misunderstood the comment?

Quoted 5 lines

If a headset is inserted when the device is in runtime_pm suspend, how would applications modify the routing and select playback on the headset, which would then ripple down to this hw_params() call?

IOW to play on a headset you first have to know there's a headset.

This is limitation in the current HW revision. So keeping this as temporary workaround to make uaj playback work.

Quoted 93 lines

sound/soc/codecs/tac5xx2-sdw.c

Quoted 9 lines
sound/soc/codecs/tac5xx2-sdw.c new file +8 -0
+		mutex_lock(&tac_dev->uaj_lock);
+		tac5xx2_sdca_headset_detect(tac_dev);
+		mutex_unlock(&tac_dev->uaj_lock);
+		break;
+	case TAC5XX2_SPK:
+		function_id = TAC_FUNCTION_ID_SA;
+		pde_entity = TAC_SDCA_ENT_PDE23;
+		port_num = substream->stream ==

SNDRV_PCM_STREAM_PLAYBACK ?

sound/soc/codecs/tac5xx2-sdw.c

Quoted 9 lines
sound/soc/codecs/tac5xx2-sdw.c new file +9 -0
+				TAC_SDW_PORT_NUM_SPK_PLAYBACK :
+				TAC_SDW_PORT_NUM_SPK_CAPTURE;
+		break;
+	default:
+		dev_err(tac_dev->dev, "Invalid dai id: %d", dai->id);
+		return -EINVAL;
+	}
+
+	snd_sdw_params_to_config(substream, params, &stream_config,

&port_config);

Quoted 22 lines
sound/soc/codecs/tac5xx2-sdw.c new file +22 -0
+	port_config.num = port_num;
+	ret = sdw_stream_add_slave(sdw_peripheral, &stream_config,
+				   &port_config, 1, sdw_stream);
+	if (ret) {
+		dev_err(dai->dev,
+			"Unable to configure port %d: %d\n", port_num, ret);
+		return ret;
+	}
+
+	switch (params_rate(params)) {
+	case 48000:
+		sample_rate_idx = 0x01;
+		break;
+	case 44100:
+		sample_rate_idx = 0x02;
+		break;
+	case 96000:
+		sample_rate_idx = 0x03;
+		break;
+	case 88200:
+		sample_rate_idx = 0x04;
+		break;

sound/soc/codecs/tac5xx2-sdw.c

Quoted 10 lines
sound/soc/codecs/tac5xx2-sdw.c new file +10 -0
+	default:
+		dev_dbg(tac_dev->dev, "Unsupported sample rate: %d Hz",
+			params_rate(params));
+		return -EINVAL;
+	}
+
+	switch (function_id) {
+	case TAC_FUNCTION_ID_SM:
+		ret = regmap_write(tac_dev->regmap,
+				   SDW_SDCA_CTL(function_id,

TAC_SDCA_ENT_CS113,

Quoted 1 line

+

	TAC_SDCA_CTL_CS_SAMP_RATE_IDX, 0),

sound/soc/codecs/tac5xx2-sdw.c

Quoted 3 lines
sound/soc/codecs/tac5xx2-sdw.c new file +3 -0
+			sample_rate_idx);
+		if (ret) {
+			dev_err(tac_dev->dev, "Failed to set CS113 sample

rate: %d", ret);

sound/soc/codecs/tac5xx2-sdw.c

Quoted 8 lines
sound/soc/codecs/tac5xx2-sdw.c new file +8 -0
+			return ret;
+		}
+
+		break;
+	case TAC_FUNCTION_ID_UAJ:
+		if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) {
+			ret = regmap_write(tac_dev->regmap,
+					   SDW_SDCA_CTL(function_id,

TAC_SDCA_ENT_CS41,

Quoted 1 line

+

	TAC_SDCA_CTL_CS_SAMP_RATE_IDX, 0),

sound/soc/codecs/tac5xx2-sdw.c

Quoted 3 lines
sound/soc/codecs/tac5xx2-sdw.c new file +3 -0
+					sample_rate_idx);
+			if (ret) {
+				dev_err(tac_dev->dev, "Failed to set CS41

sample rate: %d", ret);

sound/soc/codecs/tac5xx2-sdw.c

Quoted 5 lines
sound/soc/codecs/tac5xx2-sdw.c new file +5 -0
+				return ret;
+			}
+		} else {
+			ret = regmap_write(tac_dev->regmap,
+					   SDW_SDCA_CTL(function_id,

TAC_SDCA_ENT_CS36,

Quoted 1 line

+

	TAC_SDCA_CTL_CS_SAMP_RATE_IDX, 0),

sound/soc/codecs/tac5xx2-sdw.c

Quoted 3 lines
sound/soc/codecs/tac5xx2-sdw.c new file +3 -0
+					sample_rate_idx);
+			if (ret) {
+				dev_err(tac_dev->dev, "Failed to set CS36

sample rate: %d", ret);

sound/soc/codecs/tac5xx2-sdw.c

Quoted 6 lines
sound/soc/codecs/tac5xx2-sdw.c new file +6 -0
+				return ret;
+			}
+		}
+		break;
+	case TAC_FUNCTION_ID_SA:
+		/* SmartAmp: no additional sample rate configuration needed

*/

Quoted 8 lines

sound/soc/codecs/tac5xx2-sdw.c line 816

Quoted 4 lines
sound/soc/codecs/tac5xx2-sdw.c new file +4 -0
+		break;
+	}
+
+	guard(mutex)(&tac_dev->pde_lock);

question I mentioned above: when you reach the hw_params phase, do you really have a potential race with firmware download? What does this specific use of pde_lock protect against?

answer below ...

Quoted 9 lines

sound/soc/codecs/tac5xx2-sdw.c line 864

Quoted 4 lines
sound/soc/codecs/tac5xx2-sdw.c new file +4 -0
+static s32 tac_sdw_pcm_hw_free(struct snd_pcm_substream *substream,
+			       struct snd_soc_dai *dai)
+
+	guard(mutex)(&tac_dev->pde_lock);

same here, do you really have a race with firmware download?

Or is this a case of dependencies between functions that requires all power state transitions to be serialized?

... during my testing, I have noted when the playback starts (from clock stop mode),
device gets "detached" first, tries to resume, .hw_params is called, then gets "attached".
I am checking slave's initialization completion variable in .hw_params to make sure that the firmware download is complete before .hw_params proceeds.
On the "master" side, looks like the slave status is handled via workqueues.
So, still I wanted to keep one more protection, just in case, because of the above sequence, for serialization.

Quoted 11 lines

sound/soc/codecs/tac5xx2-sdw.c line 966

Quoted 6 lines
sound/soc/codecs/tac5xx2-sdw.c new file +6 -0
+static int tac5xx2_sdca_headset_detect(struct tac5xx2_prv *tac_dev)
+{
+	int val, ret;
+
+	if (!tac_has_uaj_support(tac_dev))
+		return 0;

can this really happen? usually you try to detect a headset if the device is capable of dealing with headsets, no?
Should this test be moved at a higher level before you enable low-level handling of headset stuff?

This is currently there because of workaround in .hw_params as it is called for all devices.

Quoted 6 lines

sound/soc/codecs/tac5xx2-sdw.c

Quoted 5 lines
sound/soc/codecs/tac5xx2-sdw.c new file +5 -0
+
+static int tac5xx2_set_jack(struct snd_soc_component *component,
+			    struct snd_soc_jack *hs_jack, void *data)
+{
+	struct tac5xx2_prv *tac_dev =

snd_soc_component_get_drvdata(component);

Quoted 7 lines

sound/soc/codecs/tac5xx2-sdw.c line 1008

Quoted 4 lines
sound/soc/codecs/tac5xx2-sdw.c new file +4 -0
+	int ret;
+
+	if (!tac_has_uaj_support(tac_dev))
+		return 0;

same here, shouldn't set_jack() be added to the component callbacks before probe?

Yes, currently it is. ( .set_jack = tac5xx2_set_jack )

Quoted 16 lines
Quoted 1 line

+static const struct snd_soc_component_driver soc_codec_driver_tacdevice =

{

sound/soc/codecs/tac5xx2-sdw.c line 1376

Quoted 11 lines
sound/soc/codecs/tac5xx2-sdw.c new file +11 -0
+	.probe = tac_component_probe,
+	.remove = tac_component_remove,
+	.controls = tac5xx2_snd_controls,
+	.num_controls = ARRAY_SIZE(tac5xx2_snd_controls),
+	.dapm_widgets = tac5xx2_common_widgets,
+	.num_dapm_widgets = ARRAY_SIZE(tac5xx2_common_widgets),
+	.dapm_routes = tac5xx2_common_routes,
+	.num_dapm_routes = ARRAY_SIZE(tac5xx2_common_routes),
+	.idle_bias_on = 0,
+	.endianness = 1,
+	.set_jack = tac5xx2_set_jack,

maybe make this dynamic and only populate .set_jack in tac_init() below when you can deal with a jack?

Let me test this case. Thanks for the suggestion.

Regards
Niranjan

#4 lore raw

Re: [PATCH v9 2/4] ASoC: tac5xx2-sdw: add soundwire based codec driver

Quoted 10 lines
Quoted 10 lines

sound/soc/codecs/tac5xx2-sdw.c

Quoted 5 lines
sound/soc/codecs/tac5xx2-sdw.c new file +3 -0
+		/* Detect and set jack type for UAJ path before playback.
+		 * This is required as jack detection does not trigger interrupt
+		 * when device is in runtime_pm suspend with bus in clock stop

mode.

Review

Quoted 1 line
(quoted diff) +1 -0
+		 */

so here we have an interesting logic - or I misunderstood the comment?

Quoted 8 lines
Quoted 5 lines

If a headset is inserted when the device is in runtime_pm suspend, how would applications modify the routing and select playback on the headset, which would then ripple down to this hw_params() call?

IOW to play on a headset you first have to know there's a headset.

This is limitation in the current HW revision. So keeping this as temporary workaround to make uaj playback work.

That doesn't really address my point. If you don't have an interrupt to trigger a runtime_pm resume, I am not sure how audio routing would work. You could detect the headset when playing on *another* endpoint but that's a different function/GE entity to deal with...

In any case why not do the detection in the resume callback instead of hw_params? see more on this below...

Quoted 14 lines
Quoted 5 lines

Review

Quoted 1 line
(quoted diff) +1 -0
+	guard(mutex)(&tac_dev->pde_lock);

question I mentioned above: when you reach the hw_params phase, do you really have a potential race with firmware download? What does this specific use of pde_lock protect against?

answer below ...

Quoted 7 lines

sound/soc/codecs/tac5xx2-sdw.c line 864

Quoted 4 lines
sound/soc/codecs/tac5xx2-sdw.c new file +4 -0
+static s32 tac_sdw_pcm_hw_free(struct snd_pcm_substream *substream,
+			       struct snd_soc_dai *dai)
+
+	guard(mutex)(&tac_dev->pde_lock);

same here, do you really have a race with firmware download?

Quoted 5 lines
Quoted 2 lines

Or is this a case of dependencies between functions that requires all power state transitions to be serialized?

... during my testing, I have noted when the playback starts (from clock stop mode), device gets "detached" first, tries to resume, .hw_params is called, then gets "attached".

That seems very very odd. The resume should wait on the device being completely operational. Only then would you deal with hw_params.

Note that there are devices that require a soft reset after firmware download, and that does cause an extra detach-attach sequence, but again the resume has to complete fully before hw_params is reached. See examples from Cirrus Logic.

Quoted 2 lines

I am checking slave's initialization completion variable in .hw_params to make sure that the firmware download is complete before .hw_params proceeds.

that's not the expected sequence, see all other codec drivers, they wait on the initialization_complete in resume callbacks.

Quoted 2 lines

On the "master" side, looks like the slave status is handled via workqueues.
So, still I wanted to keep one more protection, just in case, because of the above sequence, for serialization.

I would recommend you start with the expected sequences, I really don't see what's different in your case that would require such a fundamental change in programming sequences?

#5 lore raw

PATCH v9 3/4 ASoC: sdw_utils: TI amp utility for tac5xx2 family

In reply to Niranjan H Y

Add TI amp utility for supporting the tac5xx2 family
of devices to support tac5572, tac5672, tac5682 and tas2883

---
v9:
- no change

v8:
- no change

---
 include/sound/soc_sdw_utils.h        |   4 +
 sound/soc/sdw_utils/soc_sdw_ti_amp.c | 144 ++++++++++++++++++++++++-
 sound/soc/sdw_utils/soc_sdw_utils.c  | 151 +++++++++++++++++++++++++++
 3 files changed, 298 insertions(+), 1 deletion(-)
include/sound/soc_sdw_utils.h +4 -0
index 4890831836734..d713ab2f66203 100644
@@ -272,7 +272,11 @@ int asoc_sdw_ti_amp_init(struct snd_soc_card *card,
 			 struct asoc_sdw_codec_info *info,
 			 bool playback);
 int asoc_sdw_ti_spk_rtd_init(struct snd_soc_pcm_runtime *rtd, struct snd_soc_dai *dai);
+int asoc_sdw_ti_tac5xx2_spk_rtd_init(struct snd_soc_pcm_runtime *rtd,
+				     struct snd_soc_dai *dai);
 int asoc_sdw_ti_amp_initial_settings(struct snd_soc_card *card,
 				     const char *name_prefix);
+int asoc_sdw_ti_dmic_rtd_init(struct snd_soc_pcm_runtime *rtd, struct snd_soc_dai *dai);
+int asoc_sdw_ti_sdca_jack_rtd_init(struct snd_soc_pcm_runtime *rtd, struct snd_soc_dai *dai);
 
 #endif
sound/soc/sdw_utils/soc_sdw_ti_amp.c +143 -1

158 lines not on this page · raw message

sound/soc/sdw_utils/soc_sdw_utils.c +151 -0

157 lines not on this page · raw message

Signature · 1 line
2.34.1
#6 lore raw

PATCH v9 4/4 ASoC: tac5xx2-sdw: ACPI match for intel mtl platform

In reply to Niranjan H Y

Add acpi match entries to support TI's tac5572,
tas2883, tac5672 and tac5682 on link 0 on MTL machine.

---
v9:
- no change

v8:
- no change

---
 .../intel/common/soc-acpi-intel-mtl-match.c   | 132 ++++++++++++++++++
 1 file changed, 132 insertions(+)
sound/soc/intel/common/soc-acpi-intel-mtl-match.c +132 -0
index 72c35e73078e3..2e4222456f27f 100644
@@ -122,6 +122,42 @@ static const struct snd_soc_acpi_endpoint spk_r_endpoint = {
 	.group_id = 1,
 };
 
+static const struct snd_soc_acpi_endpoint tac5xx2_endpoints[] = {
+	{ /* Playback Endpoint */
+		.num = 0,
+		.aggregated = 0,
+		.group_position = 0,
+		.group_id = 0,
+	},
+	{ /* Mic Capture Endpoint */
+		.num = 1,
+		.aggregated = 0,
+		.group_position = 0,
+		.group_id = 0,
+	},
+	{ /* UAJ-HP with Mic Endpoint */
+		.num = 2,
+		.aggregated = 0,
+		.group_position = 0,
+		.group_id = 0,
+	},
+};
+
+static const struct snd_soc_acpi_endpoint tas2883_endpoints[] = {
+	{ /* Playback Endpoint */
+		.num = 0,
+		.aggregated = 0,
+		.group_position = 0,
+		.group_id = 0,
+	},
+	{ /* Mic Capture Endpoint */
+		.num = 1,
+		.aggregated = 0,
+		.group_position = 0,
+		.group_id = 0,
+	},
+};
+
 static const struct snd_soc_acpi_endpoint rt712_endpoints[] = {
 	{
 		.num = 0,
@@ -1011,6 +1047,33 @@ static const struct snd_soc_acpi_adr_device cs42l42_0_adr[] = {
 	}
 };
 
+static const struct snd_soc_acpi_adr_device tac5572_0_adr[] = {
+	{
+		.adr = 0x0000300102557201ull,
+		.num_endpoints = ARRAY_SIZE(tac5xx2_endpoints),
+		.endpoints = tac5xx2_endpoints,
+		.name_prefix = "tac5572"
+	}
+};
+
+static const struct snd_soc_acpi_adr_device tac5672_0_adr[] = {
+	{
+		.adr = 0x0000300102567201ull,
+		.num_endpoints = ARRAY_SIZE(tac5xx2_endpoints),
+		.endpoints = tac5xx2_endpoints,
+		.name_prefix = "tac5672"
+	}
+};
+
+static const struct snd_soc_acpi_adr_device tac5682_0_adr[] = {
+	{
+		.adr = 0x0000300102568201ull,
+		.num_endpoints = ARRAY_SIZE(tac5xx2_endpoints),
+		.endpoints = tac5xx2_endpoints,
+		.name_prefix = "tac5682"
+	}
+};
+
 static const struct snd_soc_acpi_adr_device tas2783_0_adr[] = {
 	{
 		.adr = 0x00003c0102000001ull,
@@ -1035,9 +1098,45 @@ static const struct snd_soc_acpi_adr_device tas2783_0_adr[] = {
 		.num_endpoints = 1,
 		.endpoints = &spk_r_endpoint,
 		.name_prefix = "tas2783-4"
+	},
+};
+
+static const struct snd_soc_acpi_adr_device tas2883_0_adr[] = {
+	{
+		.adr = 0x0000300102288301ull,
+		.num_endpoints = ARRAY_SIZE(tas2883_endpoints),
+		.endpoints = tas2883_endpoints,
+		.name_prefix = "tas2883"
 	}
 };
 
+static const struct snd_soc_acpi_link_adr tac5572_l0[] = {
+	{
+		.mask = BIT(0),
+		.num_adr = ARRAY_SIZE(tac5572_0_adr),
+		.adr_d = tac5572_0_adr,
+	},
+	{}
+};
+
+static const struct snd_soc_acpi_link_adr tac5672_l0[] = {
+	{
+		.mask = BIT(0),
+		.num_adr = ARRAY_SIZE(tac5672_0_adr),
+		.adr_d = tac5672_0_adr,
+	},
+	{}
+};
+
+static const struct snd_soc_acpi_link_adr tac5682_l0[] = {
+	{
+		.mask = BIT(0),
+		.num_adr = ARRAY_SIZE(tac5682_0_adr),
+		.adr_d = tac5682_0_adr,
+	},
+	{}
+};
+
 static const struct snd_soc_acpi_link_adr tas2783_link0[] = {
 	{
 		.mask = BIT(0),
@@ -1047,6 +1146,15 @@ static const struct snd_soc_acpi_link_adr tas2783_link0[] = {
 	{}
 };
 
+static const struct snd_soc_acpi_link_adr tas2883_l0[] = {
+	{
+		.mask = BIT(0),
+		.num_adr = ARRAY_SIZE(tas2883_0_adr),
+		.adr_d = tas2883_0_adr,
+	},
+	{}
+};
+
 static const struct snd_soc_acpi_link_adr cs42l42_link0_max98363_link2[] = {
 	/* Expected order: jack -> amp */
 	{
@@ -1208,12 +1316,36 @@ struct snd_soc_acpi_mach snd_soc_acpi_intel_mtl_sdw_machines[] = {
 		.drv_name = "sof_sdw",
 		.sof_tplg_filename = "sof-mtl-rt715-rt711-rt1308-mono.tplg",
 	},
+	{
+		.link_mask = BIT(0),
+		.links = tac5572_l0,
+		.drv_name = "sof_sdw",
+		.sof_tplg_filename = "sof-mtl-tac5572.tplg",
+	},
+	{
+		.link_mask = BIT(0),
+		.links = tac5672_l0,
+		.drv_name = "sof_sdw",
+		.sof_tplg_filename = "sof-mtl-tac5672.tplg",
+	},
+	{
+		.link_mask = BIT(0),
+		.links = tac5682_l0,
+		.drv_name = "sof_sdw",
+		.sof_tplg_filename = "sof-mtl-tac5682.tplg",
+	},
 	{
 		.link_mask = BIT(0),
 		.links = tas2783_link0,
 		.drv_name = "sof_sdw",
 		.sof_tplg_filename = "sof-mtl-tas2783.tplg",
 	},
+	{
+		.link_mask = BIT(0),
+		.links = tas2883_l0,
+		.drv_name = "sof_sdw",
+		.sof_tplg_filename = "sof-mtl-tas2883.tplg",
+	},
 	{
 		.link_mask = GENMASK(3, 0),
 		.links = mtl_rt713_l0_rt1316_l12_rt1713_l3,
Signature · 1 line
2.34.1
maintainer
#7 lore raw
In reply to Niranjan H Y
Quoted 16 lines · Niranjan H Y

Implement sdca_asoc_pde_ensure_ps() helper function to poll for PDE
power state transitions. Per SDCA specification, after writing REQUESTED_PS, drivers must poll ACTUAL_PS until the target power state is reached.

Changes include:
- Add sdca_asoc_pde_ensure_ps() to handle ACTUAL_PS polling with support for device-specific delay tables or default intervals
- Export function via sdca_asoc.h for use by SDCA-compliant drivers
- Refactor entity_pde_event() in sdca_asoc.c to use the helper

---

include/sound/sdca_asoc.h int sdca_asoc_q78_put_volsw(struct snd_kcontrol *kcontrol, · line 104

Quoted 6 lines
include/sound/sdca_asoc.h +3 -0
@@ -99,4 +99,13 @@ int sdca_asoc_q78_put_volsw(struct snd_kcontrol *kcontrol,
 			    struct snd_ctl_elem_value *ucontrol);
 int sdca_asoc_q78_get_volsw(struct snd_kcontrol *kcontrol,
 			    struct snd_ctl_elem_value *ucontrol);
+
+struct device;
+struct sdca_pde_delay;

sdca_pde_delay should go at the top of the file with the others, device is already there so can be removed.

include/sound/sdca_asoc.h int sdca_asoc_q78_put_volsw(struct snd_kcontrol *kcontrol, · line 105

Quoted 1 line
include/sound/sdca_asoc.h +1 -0
+int sdca_asoc_pde_ensure_ps(struct device *dev, struct regmap *regmap,

Perhaps sdca_asoc_pde_poll_ps, but I don't feel super strongly on that.

include/sound/sdca_asoc.h int sdca_asoc_q78_put_volsw(struct snd_kcontrol *kcontrol,

Quoted 9 lines
include/sound/sdca_asoc.h +9 -0
+			    int function_id, int entity_id,
+			    int from_ps, int to_ps,
+			    const struct sdca_pde_delay *pde_delays,
+			    int num_delays)
+{
+	static const int polls = 100;
+	static const int default_poll_us = 1000;
+	unsigned int reg, val;
+	int i, poll_us = default_poll_us;

Put poll_us on its own line, and do all the assigned variables first. I would combine i and ret on the same line instead.

Other than those minor nitpicks this looks good to me though.

Thanks,
Charles

#8 lore raw
In reply to Niranjan H Y
Quoted 4 lines · Niranjan H Y

Implement sdca_asoc_pde_ensure_ps() helper function to poll for PDE
power state transitions. Per SDCA specification, after writing REQUESTED_PS, drivers must poll ACTUAL_PS until the target power state is reached.

Good initiative to introduce a new common helper...

Quoted 10 lines

+/**
+ * sdca_asoc_pde_ensure_ps - Verify PDE power state reached target state
+ * @dev: Pointer to the device for error logging.
+ * @regmap: Register map for reading ACTUAL_PS register.
+ * @function_id: SDCA function identifier.
+ * @entity_id: SDCA entity identifier for the power domain.
+ * @from_ps: Source power state (SDCA_PDE_PSn value).
+ * @to_ps: Target power state (SDCA_PDE_PSn value).
+ * @pde_delays: Pointer to array of PDE delay specifications for this device,
+ * or NULL to use default polling interval.

Quoted 8 lines

+ * @num_delays: Number of entries in pde_delays array.
+ *
+ * This function polls the ACTUAL_PS register to verify that a PDE power state
+ * transition has completed. Per SDCA specification, after writing REQUESTED_PS,
+ * the caller must poll ACTUAL_PS until it reflects the requested state.
+ *
+ * This function implements the polling logic but does NOT modify the power state.
+ * The caller is responsible for writing REQUESTED_PS before invoking this function.

Erm, why not dealing with the write to REQUESTED_PS in this helper? You have all the 'to' and 'from' information in the parameters.

Quoted 25 lines

+ *
+ * If a delay table is provided, appropriate polling intervals are extracted based
+ * on the from_ps and to_ps transition. If no table is provided or no matching entry
+ * is found, a default polling interval is used.
+ *
+ * Return: Returns zero when ACTUAL_PS reaches the target state, -ETIMEDOUT if the
+ * polling times out before reaching the target state, or a negative error code if
+ * a register read fails.
+ */

sound/soc/sdca/sdca_asoc.c +16 -0
+int sdca_asoc_pde_ensure_ps(struct device *dev, struct regmap *regmap,
+			    int function_id, int entity_id,
+			    int from_ps, int to_ps,
+			    const struct sdca_pde_delay *pde_delays,
+			    int num_delays)
+{
+	static const int polls = 100;
+	static const int default_poll_us = 1000;
+	unsigned int reg, val;
+	int i, poll_us = default_poll_us;
+	int ret;
+
+	if (pde_delays && num_delays > 0) {
+		for (i = 0; i < num_delays; i++) {
+			if (pde_delays[i].from_ps == from_ps && pde_delays[i].to_ps == to_ps) {
+				poll_us = pde_delays[i].us / polls;

sound/soc/sdca/sdca_asoc.c static int entity_parse_ot(struct device *dev, · line 414

Quoted 10 lines
sound/soc/sdca/sdca_asoc.c +10 -0
+				break;
+			}
+		}
+	}
+
+	reg = SDW_SDCA_CTL(function_id, entity_id, SDCA_CTL_PDE_ACTUAL_PS, 0);
+
+	for (i = 0; i < polls; i++) {
+		if (i)
+			fsleep(poll_us);

This solution will loop for up to 100 times, and the sleep duration could be questionable.

Say for example you have a 10ms transition, do you really want to read ACTUAL_PS every 100us?

If the pde_delay is 1ms then a read every 10us makes no sense, the SoundWire command protocol would not be able to handle such reads.

A minimum threshold on poll_us would make sense IMHO.

Quoted 12 lines
sound/soc/sdca/sdca_asoc.c +12 -0
+
+		ret = regmap_read(regmap, reg, &val);
+		if (ret)
+			return ret;
+		else if (val == to_ps)
+			return 0;
+	}
+
+	dev_err(dev, "PDE power transition failed: expected 0x%x, got 0x%x\n", to_ps, val);
+	return -ETIMEDOUT;
+}
+EXPORT_SYMBOL(sdca_asoc_pde_ensure_ps);
maintainer
#9 lore raw
Quoted 7 lines · Pierre-Louis Bossart
Quoted 2 lines · Niranjan H Y

+ * This function implements the polling logic but does NOT modify the power state.
+ * The caller is responsible for writing REQUESTED_PS before invoking this function.

Erm, why not dealing with the write to REQUESTED_PS in this helper? You have all the 'to' and 'from' information in the parameters.

I have no objections to moving that into the helper as well.

Quoted 15 lines
Quoted 15 lines
sound/soc/sdca/sdca_asoc.c +15 -0
+	static const int polls = 100;
+	static const int default_poll_us = 1000;
+	unsigned int reg, val;
+	int i, poll_us = default_poll_us;
+	int ret;
+
+	if (pde_delays && num_delays > 0) {
+		for (i = 0; i < num_delays; i++) {
+			if (pde_delays[i].from_ps == from_ps && pde_delays[i].to_ps == to_ps) {
+				poll_us = pde_delays[i].us / polls;
+				break;
+			}
+		}
+	}
+
Quoted 8 lines

sound/soc/sdca/sdca_asoc.c static int entity_parse_ot(struct device *dev, · line 414

Quoted 5 lines
sound/soc/sdca/sdca_asoc.c +5 -0
+	reg = SDW_SDCA_CTL(function_id, entity_id, SDCA_CTL_PDE_ACTUAL_PS, 0);
+
+	for (i = 0; i < polls; i++) {
+		if (i)
+			fsleep(poll_us);

This solution will loop for up to 100 times, and the sleep duration could be questionable.

The duration doesn't have to be precise here, as long as the result is longer than the requested time everything is fine.

Quoted 2 lines

Say for example you have a 10ms transition, do you really want to read ACTUAL_PS every 100us?

Quite potentially, I imagine it will be fairly common for parts to change PS a lot faster than the actual timeouts they provide, due to corner cases and people just being conservative in the DisCo. So its quite possible something that says 10mS typically switches in a couple 100uS.

Quoted 5 lines

If the pde_delay is 1ms then a read every 10us makes no sense, the SoundWire command protocol would not be able to handle such reads.

A minimum threshold on poll_us would make sense IMHO.

I guess you do reach a point where the soundwire command makes the delay effectively meaningless. What would you suggest for a minimum? Something like 100uS feels kinda reasonable to me, I would lean towards quite a small value here. Other options might be to look at some sort of exponential back off, doing the first few polls faster than later ones.

This is definitely one of those situations where SDCA is a little too vague for its own good. But I would also say making a change like this should at a minimum be a separate patch rather than part of this one. And I am not convinced we need to block this series on updating it, although if we just wanted to go with a simple minimum that seems easy enough to add.

Thanks,
Charles

#10 lore raw
Quoted 47 lines · Charles Keepax
Quoted 7 lines · Pierre-Louis Bossart
Quoted 2 lines · Niranjan H Y

+ * This function implements the polling logic but does NOT modify the power state.
+ * The caller is responsible for writing REQUESTED_PS before invoking this function.

Erm, why not dealing with the write to REQUESTED_PS in this helper? You have all the 'to' and 'from' information in the parameters.

I have no objections to moving that into the helper as well.

Quoted 15 lines
Quoted 15 lines
sound/soc/sdca/sdca_asoc.c +15 -0
+	static const int polls = 100;
+	static const int default_poll_us = 1000;
+	unsigned int reg, val;
+	int i, poll_us = default_poll_us;
+	int ret;
+
+	if (pde_delays && num_delays > 0) {
+		for (i = 0; i < num_delays; i++) {
+			if (pde_delays[i].from_ps == from_ps && pde_delays[i].to_ps == to_ps) {
+				poll_us = pde_delays[i].us / polls;
+				break;
+			}
+		}
+	}
+
Quoted 8 lines

sound/soc/sdca/sdca_asoc.c static int entity_parse_ot(struct device *dev, · line 414

Quoted 5 lines
sound/soc/sdca/sdca_asoc.c +5 -0
+	reg = SDW_SDCA_CTL(function_id, entity_id, SDCA_CTL_PDE_ACTUAL_PS, 0);
+
+	for (i = 0; i < polls; i++) {
+		if (i)
+			fsleep(poll_us);

This solution will loop for up to 100 times, and the sleep duration could be questionable.

The duration doesn't have to be precise here, as long as the result is longer than the requested time everything is fine.

Quoted 2 lines

Say for example you have a 10ms transition, do you really want to read ACTUAL_PS every 100us?

Quite potentially, I imagine it will be fairly common for parts to change PS a lot faster than the actual timeouts they provide, due to corner cases and people just being conservative in the DisCo. So its quite possible something that says 10mS typically switches in a couple 100uS.

Quoted 8 lines
Quoted 5 lines

If the pde_delay is 1ms then a read every 10us makes no sense, the SoundWire command protocol would not be able to handle such reads.

A minimum threshold on poll_us would make sense IMHO.

I guess you do reach a point where the soundwire command makes the delay effectively meaningless. What would you suggest for a

yep, that was the main point.

Quoted 11 lines

minimum? Something like 100uS feels kinda reasonable to me, I would lean towards quite a small value here. Other options might be to look at some sort of exponential back off, doing the first few polls faster than later ones.

This is definitely one of those situations where SDCA is a little too vague for its own good. But I would also say making a change like this should at a minimum be a separate patch rather than part of this one. And I am not convinced we need to block this series on updating it, although if we just wanted to go with a simple minimum that seems easy enough to add.

A minimum of 100us would be fine, we can always optimize for long delays later.

#11 lore raw
Quoted 14 lines · Pierre-Louis Bossart

Sent: Monday, April 20, 2026 4:57 PM
Subject: Re: [PATCH v9 1/4] ASoC: SDCA: Add PDE verification reusable helper

Quoted 3 lines · Charles Keepax
Quoted 2 lines · Pierre-Louis Bossart
Quoted 1 line · Niranjan H Y

+ * The caller is responsible for writing REQUESTED_PS before invoking this

function.

Quoted 6 lines
Quoted 4 lines

Erm, why not dealing with the write to REQUESTED_PS in this helper? You have all the 'to' and 'from' information in the parameters.

I have no objections to moving that into the helper as well.

Ok, I did not want to disturb the widget's event handler in this patch. I will add this in the next patch.

Quoted 62 lines
Quoted 10 lines
Quoted 10 lines

sound/soc/sdca/sdca_asoc.c static int entity_parse_ot(struct device *dev,

Quoted 10 lines
sound/soc/sdca/sdca_asoc.c +9 -0
+	static const int polls = 100;
+	static const int default_poll_us = 1000;
+	unsigned int reg, val;
+	int i, poll_us = default_poll_us;
+	int ret;
+
+	if (pde_delays && num_delays > 0) {
+		for (i = 0; i < num_delays; i++) {
+			if (pde_delays[i].from_ps == from_ps &&

pde_delays[i].to_ps == to_ps) {

Quoted 7 lines
Quoted 7 lines

sound/soc/sdca/sdca_asoc.c static int entity_parse_ot(struct device *dev,

Quoted 7 lines
sound/soc/sdca/sdca_asoc.c +7 -0
+				poll_us = pde_delays[i].us / polls;
+				break;
+			}
+		}
+	}
+
+	reg = SDW_SDCA_CTL(function_id, entity_id,

SDCA_CTL_PDE_ACTUAL_PS, 0);

Quoted 20 lines
Quoted 7 lines

sound/soc/sdca/sdca_asoc.c static int entity_parse_ot(struct device *dev, · line 414

Quoted 4 lines
sound/soc/sdca/sdca_asoc.c +4 -0
+
+	for (i = 0; i < polls; i++) {
+		if (i)
+			fsleep(poll_us);

This solution will loop for up to 100 times, and the sleep duration could be questionable.

The duration doesn't have to be precise here, as long as the result is longer than the requested time everything is fine.

Quoted 2 lines

Say for example you have a 10ms transition, do you really want to read ACTUAL_PS every 100us?

Quite potentially, I imagine it will be fairly common for parts to change PS a lot faster than the actual timeouts they provide, due to corner cases and people just being conservative in the DisCo. So its quite possible something that says 10mS typically switches in a couple 100uS.

Quoted 8 lines
Quoted 5 lines

If the pde_delay is 1ms then a read every 10us makes no sense, the SoundWire command protocol would not be able to handle such reads.

A minimum threshold on poll_us would make sense IMHO.

I guess you do reach a point where the soundwire command makes the delay effectively meaningless. What would you suggest for a

yep, that was the main point.

Quoted 11 lines

minimum? Something like 100uS feels kinda reasonable to me, I would lean towards quite a small value here. Other options might be to look at some sort of exponential back off, doing the first few polls faster than later ones.

This is definitely one of those situations where SDCA is a little too vague for its own good. But I would also say making a change like this should at a minimum be a separate patch rather than part of this one. And I am not convinced we need to block this series on updating it, although if we just wanted to go with a simple minimum that seems easy enough to add.

Quoted 2 lines

A minimum of 100us would be fine, we can always optimize for long delays later.

Understood, I will change it to 100us and send new patch. I realized that at @4.8MHz to 6MHz, 50x10, I think the read itself takes roughly ~42us to 50us.
And with 100us it is 10ms before timeout happens
(practically maybe it is roughly 15ms).

Regards
Niranjan