Skip to contentSkip to conversation

Add DAPM/ASoC helpers to create SDCA drivers

PATCH Superseded Charles Keepax 2 participants 26 messages May 12, 2025 – May 19, 2025

Subsystems (3)

v5 of 6 Superseded Latest is v6 →

lore
Share & tools
lore
author maintainer
#3 lore raw

PATCH v5 3/6 ASoC: dapm: Add component level pin switches

In reply to Charles Keepax

The core currently supports pin switches for source/sink widgets, but only at the card level. SDCA components specify the fabric at the level of the individual components, to support this add helpers to allow component level pin switches.

---

No changes since v4.

 include/sound/soc-dapm.h |  4 ++
 sound/soc/soc-dapm.c     | 84 +++++++++++++++++++++++++++++++++-------
 2 files changed, 74 insertions(+), 14 deletions(-)
include/sound/soc-dapm.h +4 -0
index af802ef536e73..400584474bc8b 100644
@@ -445,6 +445,10 @@ int snd_soc_dapm_get_pin_switch(struct snd_kcontrol *kcontrol,
 	struct snd_ctl_elem_value *uncontrol);
 int snd_soc_dapm_put_pin_switch(struct snd_kcontrol *kcontrol,
 	struct snd_ctl_elem_value *uncontrol);
+int snd_soc_dapm_get_component_pin_switch(struct snd_kcontrol *kcontrol,
+					  struct snd_ctl_elem_value *uncontrol);
+int snd_soc_dapm_put_component_pin_switch(struct snd_kcontrol *kcontrol,
+					  struct snd_ctl_elem_value *uncontrol);
 int snd_soc_dapm_new_controls(struct snd_soc_dapm_context *dapm,
 	const struct snd_soc_dapm_widget *widget, unsigned int num);
 struct snd_soc_dapm_widget *snd_soc_dapm_new_control(struct snd_soc_dapm_context *dapm,
sound/soc/soc-dapm.c +70 -14

121 lines not on this page · raw message

Signature · 1 line
2.39.5
author maintainer
#14 lore raw

PATCH v5 5/6 ASoC: SDCA: Create ALSA controls from DisCo

In reply to Charles Keepax

Use the previously parsed DisCo information from ACPI to create the ALSA controls required by an SDCA Function. This maps all User and Application level SDCA Controls to ALSA controls. Typically controls marked with those access levels are just volumes and mutes.

SDCA defines volume controls as an integer in 1/256ths of a dB and then provides a mechanism to specify what values are valid (range templates). Currently only a simple case of a single linear volume range with a power of 2 step size is supported. This allows the code to expose the volume control using a simple shift. This will need expanded in the future, to support more complex ranges and probably also some additional control types but this should be sufficient to for a first pass.

For non-dataport terminal widgets also add a pin switch to allow that endpoint to be turned on/off.

---

No changes since v4.

 include/sound/sdca_asoc.h     |   6 +-
 include/sound/sdca_function.h |  10 ++
 sound/soc/sdca/sdca_asoc.c    | 241 +++++++++++++++++++++++++++++++++-
 3 files changed, 252 insertions(+), 5 deletions(-)
include/sound/sdca_asoc.h +5 -1
index 414d461b6fc4a..d19e7e969283a 100644
@@ -12,16 +12,20 @@
 
 struct device;
 struct sdca_function_data;
+struct snd_kcontrol_new;
 struct snd_soc_component_driver;
 struct snd_soc_dapm_route;
 struct snd_soc_dapm_widget;
 
 int sdca_asoc_count_component(struct device *dev, struct sdca_function_data *function,
-			      int *num_widgets, int *num_routes);
+			      int *num_widgets, int *num_routes, int *num_controls);
 
 int sdca_asoc_populate_dapm(struct device *dev, struct sdca_function_data *function,
 			    struct snd_soc_dapm_widget *widgets,
 			    struct snd_soc_dapm_route *routes);
+int sdca_asoc_populate_controls(struct device *dev,
+				struct sdca_function_data *function,
+				struct snd_kcontrol_new *kctl);
 
 int sdca_asoc_populate_component(struct device *dev,
 				 struct sdca_function_data *function,
include/sound/sdca_function.h +10 -0
index 83fedc39cf714..77ffb1f4e1ca9 100644
@@ -206,6 +206,16 @@ enum sdca_fu_controls {
 	SDCA_CTL_FU_LATENCY				= 0x10,
 };
 
+/**
+ * enum sdca_volume_range - Column definitions for Q7.8dB volumes/gains
+ */
+enum sdca_volume_range {
+	SDCA_VOLUME_LINEAR_MIN				= 0,
+	SDCA_VOLUME_LINEAR_MAX				= 1,
+	SDCA_VOLUME_LINEAR_STEP				= 2,
+	SDCA_VOLUME_LINEAR_NCOLS			= 3,
+};
+
 /**
  * enum sdca_xu_controls - SDCA Controls for Extension Unit
  *
sound/soc/sdca/sdca_asoc.c +237 -4

309 lines not on this page · raw message

Signature · 1 line
2.39.5
author maintainer
#4 lore raw

PATCH v5 4/6 ASoC: SDCA: Create DAPM widgets and routes from DisCo

In reply to Charles Keepax

Use the previously parsed DisCo information from ACPI to create DAPM widgets and routes representing a SDCA Function. For the most part SDCA maps well to the DAPM abstractions.

The primary point of interest is the SDCA Power Domain Entities (PDEs), which actually control the power status of the device. Whilst these PDEs are the primary widgets the other parts of the SDCA graph are added to maintain a consistency with the hardware abstract, and allow routing to take effect. As for the PDEs themselves the code currently only handle PS0 and PS3 (basically on and off), the two intermediate power states are not commonly used and don't map well to ASoC/DAPM.

Other minor points of slightly complexity include, the Group Entities (GEs) these set the value of several other controls, typically Selector Units (SUs) for enabling a cetain jack configuration. Multiple SUs being controlled by a GE are easily modelled creating a single control and sharing it among the controlled muxes.

SDCA also has a slight habit of having fully connected paths, relying more on activating the PDEs to enable functionality. This doesn't map quite so perfectly to DAPM which considers the path a reason to power the PDE. Whilst in the current specification Mixer Units are defined as fixed-function, in DAPM we create a virtual control for each input (which defaults to connected). This allows paths to be connected/disconnected, providing a more ASoC style approach to managing the power. PIN_SWITCHs will also be added for non-dataport terminal entities in a later patch along with the other ALSA controls, providing greater flexibility in power management.

A top level helper sdca_asoc_populate_component() is exported that counts and allocates everything, however, the intermediate counting and population functions are also exported. This will allow end drivers to do allocation and add custom handling, which is probably fairly likely for the early SDCA devices.

Clock muxes are currently not fully supported, so some future work will also be required there.

---

Changes sinve v4:
- Remove unnecessary MODULE_LICENSE() and MODULE_DESCRIPTION() macros.
- Move the error message into selector_find_control to better match the other search functions.

 include/sound/sdca_asoc.h     |  30 ++
 include/sound/sdca_function.h |  36 ++
 sound/soc/sdca/Makefile       |   2 +-
 sound/soc/sdca/sdca_asoc.c    | 839 ++++++++++++++++++++++++++++++++++
 4 files changed, 906 insertions(+), 1 deletion(-)
 create mode 100644 include/sound/sdca_asoc.h
 create mode 100644 sound/soc/sdca/sdca_asoc.c
include/sound/sdca_asoc.h new file +30 -0
new file mode 100644
index 0000000000000..414d461b6fc4a
@@ -0,0 +1,30 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * The MIPI SDCA specification is available for public downloads at
+ * https://www.mipi.org/mipi-sdca-v1-0-download
+ *
+ * Copyright (C) 2025 Cirrus Logic, Inc. and
+ *                    Cirrus Logic International Semiconductor Ltd.
+ */
+
+#ifndef __SDCA_ASOC_H__
+#define __SDCA_ASOC_H__
+
+struct device;
+struct sdca_function_data;
+struct snd_soc_component_driver;
+struct snd_soc_dapm_route;
+struct snd_soc_dapm_widget;
+
+int sdca_asoc_count_component(struct device *dev, struct sdca_function_data *function,
+			      int *num_widgets, int *num_routes);
+
+int sdca_asoc_populate_dapm(struct device *dev, struct sdca_function_data *function,
+			    struct snd_soc_dapm_widget *widgets,
+			    struct snd_soc_dapm_route *routes);
+
+int sdca_asoc_populate_component(struct device *dev,
+				 struct sdca_function_data *function,
+				 struct snd_soc_component_driver *component_drv);
+
+#endif // __SDCA_ASOC_H__
include/sound/sdca_function.h +36 -0
index 253654568a41e..83fedc39cf714 100644
@@ -257,6 +257,14 @@ enum sdca_pde_controls {
 	SDCA_CTL_PDE_ACTUAL_PS				= 0x10,
 };
 
+/**
+ * enum sdca_requested_ps_range - Column definitions for Requested PS
+ */
+enum sdca_requested_ps_range {
+	SDCA_REQUESTED_PS_STATE				= 0,
+	SDCA_REQUESTED_PS_NCOLS				= 1,
+};
+
 /**
  * enum sdca_ge_controls - SDCA Controls for Group Unit
  *
@@ -268,6 +276,15 @@ enum sdca_ge_controls {
 	SDCA_CTL_GE_DETECTED_MODE			= 0x02,
 };
 
+/**
+ * enum sdca_selected_mode_range - Column definitions for Selected Mode
+ */
+enum sdca_selected_mode_range {
+	SDCA_SELECTED_MODE_INDEX			= 0,
+	SDCA_SELECTED_MODE_TERM_TYPE			= 1,
+	SDCA_SELECTED_MODE_NCOLS			= 2,
+};
+
 /**
  * enum sdca_spe_controls - SDCA Controls for Security & Privacy Unit
  *
@@ -773,6 +790,25 @@ enum sdca_terminal_type {
 	SDCA_TERM_TYPE_PRIVACY_INDICATORS		= 0x747,
 };
 
+#define SDCA_TERM_TYPE_LINEIN_STEREO_NAME		"LineIn Stereo"
+#define SDCA_TERM_TYPE_LINEIN_FRONT_LR_NAME		"LineIn Front-LR"
+#define SDCA_TERM_TYPE_LINEIN_CENTER_LFE_NAME		"LineIn Center-LFE"
+#define SDCA_TERM_TYPE_LINEIN_SURROUND_LR_NAME		"LineIn Surround-LR"
+#define SDCA_TERM_TYPE_LINEIN_REAR_LR_NAME		"LineIn Rear-LR"
+#define SDCA_TERM_TYPE_LINEOUT_STEREO_NAME		"LineOut Stereo"
+#define SDCA_TERM_TYPE_LINEOUT_FRONT_LR_NAME		"LineOut Front-LR"
+#define SDCA_TERM_TYPE_LINEOUT_CENTER_LFE_NAME		"LineOut Center-LFE"
+#define SDCA_TERM_TYPE_LINEOUT_SURROUND_LR_NAME		"LineOut Surround-LR"
+#define SDCA_TERM_TYPE_LINEOUT_REAR_LR_NAME		"LineOut Rear-LR"
+#define SDCA_TERM_TYPE_MIC_JACK_NAME			"Microphone"
+#define SDCA_TERM_TYPE_STEREO_JACK_NAME			"Speaker Stereo"
+#define SDCA_TERM_TYPE_FRONT_LR_JACK_NAME		"Speaker Front-LR"
+#define SDCA_TERM_TYPE_CENTER_LFE_JACK_NAME		"Speaker Center-LFE"
+#define SDCA_TERM_TYPE_SURROUND_LR_JACK_NAME		"Speaker Surround-LR"
+#define SDCA_TERM_TYPE_REAR_LR_JACK_NAME		"Speaker Rear-LR"
+#define SDCA_TERM_TYPE_HEADPHONE_JACK_NAME		"Headphone"
+#define SDCA_TERM_TYPE_HEADSET_JACK_NAME		"Headset"
+
 /**
  * enum sdca_connector_type - SDCA Connector Types
  *
sound/soc/sdca/Makefile +1 -1
index dddc3e6942569..53344f108ca67 100644
@@ -1,5 +1,5 @@
 # SPDX-License-Identifier: GPL-2.0-only
 
-snd-soc-sdca-y	:= sdca_functions.o sdca_device.o sdca_regmap.o
+snd-soc-sdca-y	:= sdca_functions.o sdca_device.o sdca_regmap.o sdca_asoc.o
 
 obj-$(CONFIG_SND_SOC_SDCA)	+= snd-soc-sdca.o
sound/soc/sdca/sdca_asoc.c new file +839 -0

839 lines not on this page · raw message

Signature · 1 line
2.39.5

Charles Keepax

author maintainer
#0 lore raw

v5 cover Add DAPM/ASoC helpers to create SDCA drivers

Add helper functions to add DAPM widgets, routes, ALSA controls, and DAI drivers, these will be used to create SDCA function device drivers.

This series should provide most of the core functionality needed to get a device registered and have a working DAPM graph within the device. There are some features that still need additional work, these are marked with FIXMEs in the code. The two main things are SDCA Clock Muxes (not used in our devices and needs some ASoC core work), and better support for more complex SDCA volume control definitions (our parts have fairly simple volumes, and SDCA has a large amount of flexibility in how the volume control is specified).

The next steps in the process are to add helpers for the DAI ops themselves, some IRQ handling, and firmware download. And finally we should be able to actually add the SDCA class driver itself.

Thanks,
Charles

Changes since v4:
- Add include of soc-dai.h, this should be included directly.
- Remove unnecessary MODULE_LICENSE() and MODULE_DESCRIPTION() macros.
- Move the error message into selector_find_control to better match the other search functions.

Charles Keepax (6):
ASoC: SDCA: Remove regmap module macros
ASoC: SDCA: Move allocation of PDE delays array
ASoC: dapm: Add component level pin switches
ASoC: SDCA: Create DAPM widgets and routes from DisCo
ASoC: SDCA: Create ALSA controls from DisCo
ASoC: SDCA: Create DAI drivers from DisCo

 include/sound/sdca_asoc.h       |   42 +
 include/sound/sdca_function.h   |   69 ++
 include/sound/soc-dapm.h        |    4 +
 sound/soc/sdca/Makefile         |    2 +-
 sound/soc/sdca/sdca_asoc.c      | 1291 +++++++++++++++++++++++++++++++
 sound/soc/sdca/sdca_functions.c |   10 +-
 sound/soc/sdca/sdca_regmap.c    |    3 -
 sound/soc/soc-dapm.c            |   84 +-
 8 files changed, 1481 insertions(+), 24 deletions(-)
 create mode 100644 include/sound/sdca_asoc.h
 create mode 100644 sound/soc/sdca/sdca_asoc.c
Signature · 1 line
2.39.5
author maintainer
#2 lore raw

PATCH v5 2/6 ASoC: SDCA: Move allocation of PDE delays array

Move the allocation of the PDE delays array until after the size has been adjusted, this saves an additional division and simplifies the code slightly.

---

No changes since v4.

 sound/soc/sdca/sdca_functions.c | 10 ++++------
 1 file changed, 4 insertions(+), 6 deletions(-)
sound/soc/sdca/sdca_functions.c +4 -6
index 493f390f087ad..64ac264438907 100644
@@ -1105,12 +1105,6 @@ static int find_sdca_entity_pde(struct device *dev,
 		return -EINVAL;
 	}
 
-	/* There are 3 values per delay */
-	delays = devm_kcalloc(dev, num_delays / mult_delay,
-			      sizeof(*delays), GFP_KERNEL);
-	if (!delays)
-		return -ENOMEM;
-
 	delay_list = kcalloc(num_delays, sizeof(*delay_list), GFP_KERNEL);
 	if (!delay_list)
 		return -ENOMEM;
@@ -1121,6 +1115,10 @@ static int find_sdca_entity_pde(struct device *dev,
 
 	num_delays /= mult_delay;
 
+	delays = devm_kcalloc(dev, num_delays, sizeof(*delays), GFP_KERNEL);
+	if (!delays)
+		return -ENOMEM;
+
 	for (i = 0, j = 0; i < num_delays; i++) {
 		delays[i].from_ps = delay_list[j++];
 		delays[i].to_ps = delay_list[j++];
Signature · 1 line
2.39.5
author maintainer
#21 lore raw

PATCH v5 6/6 ASoC: SDCA: Create DAI drivers from DisCo

In reply to Charles Keepax

Use the previously parsed DisCo information from ACPI to create the DAI drivers required to connect an SDCA Function into an ASoC soundcard.

Create DAI driver structures and populate the supported sample rates and sample widths into them based on the Input/Output Terminal and any attach Clock Source entities. More complex relationships with channels etc. will be added later as constraints as part of the DAI startup.

---

Changes since v4:
- Add include of soc-dai.h, this should be included directly.

 include/sound/sdca_asoc.h     |  12 +-
 include/sound/sdca_function.h |  23 ++++
 sound/soc/sdca/sdca_asoc.c    | 227 +++++++++++++++++++++++++++++++++-
 3 files changed, 256 insertions(+), 6 deletions(-)
include/sound/sdca_asoc.h +10 -2
index d19e7e969283a..9121531f08260 100644
@@ -14,11 +14,14 @@ struct device;
 struct sdca_function_data;
 struct snd_kcontrol_new;
 struct snd_soc_component_driver;
+struct snd_soc_dai_driver;
+struct snd_soc_dai_ops;
 struct snd_soc_dapm_route;
 struct snd_soc_dapm_widget;
 
 int sdca_asoc_count_component(struct device *dev, struct sdca_function_data *function,
-			      int *num_widgets, int *num_routes, int *num_controls);
+			      int *num_widgets, int *num_routes, int *num_controls,
+			      int *num_dais);
 
 int sdca_asoc_populate_dapm(struct device *dev, struct sdca_function_data *function,
 			    struct snd_soc_dapm_widget *widgets,
@@ -26,9 +29,14 @@ int sdca_asoc_populate_dapm(struct device *dev, struct sdca_function_data *funct
 int sdca_asoc_populate_controls(struct device *dev,
 				struct sdca_function_data *function,
 				struct snd_kcontrol_new *kctl);
+int sdca_asoc_populate_dais(struct device *dev, struct sdca_function_data *function,
+			    struct snd_soc_dai_driver *dais,
+			    const struct snd_soc_dai_ops *ops);
 
 int sdca_asoc_populate_component(struct device *dev,
 				 struct sdca_function_data *function,
-				 struct snd_soc_component_driver *component_drv);
+				 struct snd_soc_component_driver *component_drv,
+				 struct snd_soc_dai_driver **dai_drv, int *num_dai_drv,
+				 const struct snd_soc_dai_ops *ops);
 
 #endif // __SDCA_ASOC_H__
include/sound/sdca_function.h +23 -0
index 77ffb1f4e1ca9..be7e4a88cbed0 100644
@@ -168,6 +168,20 @@ enum sdca_ot_controls {
 	SDCA_CTL_OT_NDAI_PACKETTYPE			= 0x17,
 };
 
+/**
+ * enum sdca_usage_range - Column definitions for Usage
+ */
+enum sdca_usage_range {
+	SDCA_USAGE_NUMBER				= 0,
+	SDCA_USAGE_CBN					= 1,
+	SDCA_USAGE_SAMPLE_RATE				= 2,
+	SDCA_USAGE_SAMPLE_WIDTH				= 3,
+	SDCA_USAGE_FULL_SCALE				= 4,
+	SDCA_USAGE_NOISE_FLOOR				= 5,
+	SDCA_USAGE_TAG					= 6,
+	SDCA_USAGE_NCOLS				= 7,
+};
+
 /**
  * enum sdca_mu_controls - SDCA Controls for Mixer Unit
  *
@@ -246,6 +260,15 @@ enum sdca_cs_controls {
 	SDCA_CTL_CS_SAMPLERATEINDEX			= 0x10,
 };
 
+/**
+ * enum sdca_samplerateindex_range - Column definitions for SampleRateIndex
+ */
+enum sdca_samplerateindex_range {
+	SDCA_SAMPLERATEINDEX_INDEX			= 0,
+	SDCA_SAMPLERATEINDEX_RATE			= 1,
+	SDCA_SAMPLERATEINDEX_NCOLS			= 2,
+};
+
 /**
  * enum sdca_cx_controls - SDCA Controls for Clock Selector
  *
sound/soc/sdca/sdca_asoc.c +223 -4

294 lines not on this page · raw message

Signature · 1 line
2.39.5
author maintainer
#1 lore raw

PATCH v5 1/6 ASoC: SDCA: Remove regmap module macros

In reply to Charles Keepax

There is no need to include MODULE_LICENSE() and MODULE_DESCRIPTION() in sdca_regmap.c as this file is part of a larger module that already defines these.

Fixes e3f7caf74b79 ("ASoC: SDCA: Add generic regmap SDCA helpers")

---

New since v4.

 sound/soc/sdca/sdca_regmap.c | 3 ---
 1 file changed, 3 deletions(-)
sound/soc/sdca/sdca_regmap.c +0 -3
index 4b78188cfcebd..66e7eee7d7f49 100644
@@ -316,6 +316,3 @@ int sdca_regmap_write_defaults(struct device *dev, struct regmap *regmap,
 	return 0;
 }
 EXPORT_SYMBOL_NS(sdca_regmap_write_defaults, "SND_SOC_SDCA");
-
-MODULE_LICENSE("GPL");
-MODULE_DESCRIPTION("SDCA library");
Signature · 1 line
2.39.5
#5 lore raw

Re: [PATCH v5 4/6] ASoC: SDCA: Create DAPM widgets and routes from DisCo

In reply to Charles Keepax
Quoted 19 lines · Charles Keepax

Use the previously parsed DisCo information from ACPI to create DAPM widgets and routes representing a SDCA Function. For the most part SDCA maps well to the DAPM abstractions.

The primary point of interest is the SDCA Power Domain Entities (PDEs), which actually control the power status of the device. Whilst these PDEs are the primary widgets the other parts of the SDCA graph are added to maintain a consistency with the hardware abstract, and allow routing to take effect. As for the PDEs themselves the code currently only handle PS0 and PS3 (basically on and off), the two intermediate power states are not commonly used and don't map well to ASoC/DAPM.

Other minor points of slightly complexity include, the Group Entities (GEs) these set the value of several other controls, typically Selector Units (SUs) for enabling a cetain jack configuration. Multiple SUs being controlled by a GE are easily modelled creating a single control and sharing it among the controlled muxes.

Quoted 10 lines

SDCA also has a slight habit of having fully connected paths, relying more on activating the PDEs to enable functionality. This doesn't map quite so perfectly to DAPM which considers the path a reason to power the PDE. Whilst in the current specification Mixer Units are defined as fixed-function, in DAPM we create a virtual control for each input (which defaults to connected). This allows paths to be connected/disconnected, providing a more ASoC style approach to managing the power. PIN_SWITCHs will also be added for non-dataport terminal entities in a later patch along with the other ALSA controls, providing greater flexibility in power management.

It's been a while since I reviewed an earlier versions so now I am confused.
- Is the proposal to have a control to change the PDE state directly, as well as a virtual control to describe a path activation? That would lead to more flexibility but also to confusion, e.g. with 'no sound' when the paths are active but the PDE still in PS3.
- Or do those virtual controls manage the PDE state, without userspace interaction? In that case the 'connected' default would higher power consumption until userspace decides which paths it really wants active.

Quoted 20 lines
sound/soc/sdca/sdca_asoc.c new file +20 -0
+/**
+ * sdca_asoc_count_component - count the various component parts
+ * @function: Pointer to the Function information.
+ * @num_widgets: Output integer pointer, will be filled with the
+ * required number of DAPM widgets for the Function.
+ * @num_routes: Output integer pointer, will be filled with the
+ * required number of DAPM routes for the Function.
+ *
+ * This function counts various things within the SDCA Function such
+ * that the calling driver can allocate appropriate space before
+ * calling the appropriate population functions.
+ *
+ * Return: Returns zero on success, and a negative error code on failure.
+ */
+int sdca_asoc_count_component(struct device *dev, struct sdca_function_data *function,
+			      int *num_widgets, int *num_routes)
+{
+	int i;
+
+	*num_widgets = function->num_entities - 1;

sound/soc/sdca/sdca_asoc.c line 101

Quoted 10 lines
sound/soc/sdca/sdca_asoc.c new file +10 -0
+	*num_routes = 0;
+
+	for (i = 0; i < function->num_entities - 1; i++) {
+		struct sdca_entity *entity = &function->entities[i];
+
+		switch (entity->type) {
+		case SDCA_ENTITY_TYPE_IT:
+		case SDCA_ENTITY_TYPE_OT:
+			*num_routes += !!entity->iot.clock;
+			*num_routes += !!entity->iot.is_dataport;

I couldn't follow those two lines.
Why does the clock play a role, for now it's not really managed?
And the difference between streaming and transducer terminals shouldn't change the accounting of routes?

sound/soc/sdca/sdca_asoc.c line 104

Quoted 3 lines
sound/soc/sdca/sdca_asoc.c new file +3 -0
+			break;
+		case SDCA_ENTITY_TYPE_PDE:
+			*num_routes += entity->pde.num_managed;

same here, isn't there a risk of the same route being counted multiple times?

sound/soc/sdca/sdca_asoc.c line 113

Quoted 9 lines
sound/soc/sdca/sdca_asoc.c new file +9 -0
+			break;
+		default:
+			break;
+		}
+
+		*num_routes += entity->num_sources;
+
+		if (entity->group)
+			(*num_routes)++;

And here as well, in case an entity is controlled by a GE does the number of routes really change?

I guess the meaning of 'routes' isn't really clear in my head, it's not just data paths but control as well that's taken into account?

Quoted 54 lines
sound/soc/sdca/sdca_asoc.c new file +53 -0
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_NS(sdca_asoc_count_component, "SND_SOC_SDCA");
 
+static int entity_early_parse_ge(struct device *dev,
+				 struct sdca_function_data *function,
+				 struct sdca_entity *entity)
+{
+	struct sdca_control_range *range;
+	struct sdca_control *control;
+	struct snd_kcontrol_new *kctl;
+	struct soc_enum *soc_enum;
+	const char *control_name;
+	unsigned int *values;
+	const char **texts;
+	int i;
+
+	control = selector_find_control(dev, entity, SDCA_CTL_GE_SELECTED_MODE);
+	if (!control)
+		return -EINVAL;
+
+	if (control->layers != SDCA_ACCESS_LAYER_CLASS)
+		dev_warn(dev, "%s: unexpected access layer: %x\n",
+			 entity->label, control->layers);
+
+	range = control_find_range(dev, entity, control, SDCA_SELECTED_MODE_NCOLS, 0);
+	if (!range)
+		return -EINVAL;
+
+	control_name = devm_kasprintf(dev, GFP_KERNEL, "%s %s",
+				      entity->label, control->label);
+	if (!control_name)
+		return -ENOMEM;
+
+	kctl = devm_kmalloc(dev, sizeof(*kctl), GFP_KERNEL);
+	if (!kctl)
+		return -ENOMEM;
+
+	soc_enum = devm_kmalloc(dev, sizeof(*soc_enum), GFP_KERNEL);
+	if (!soc_enum)
+		return -ENOMEM;
+
+	texts = devm_kcalloc(dev, range->rows + 3, sizeof(*texts), GFP_KERNEL);
+	if (!texts)
+		return -ENOMEM;
+
+	values = devm_kcalloc(dev, range->rows + 3, sizeof(*values), GFP_KERNEL);
+	if (!values)
+		return -ENOMEM;
+
+	texts[0] = "No Jack";
+	texts[1] = "Jack Unknown";

sound/soc/sdca/sdca_asoc.c line 221

Quoted 10 lines
sound/soc/sdca/sdca_asoc.c new file +10 -0
+	texts[2] = "Detection in Progress";
+	values[0] = 0;
+	values[1] = 1;
+	values[2] = 2;
+	for (i = 0; i < range->rows; i++) {
+		enum sdca_terminal_type type;
+
+		type = sdca_range(range, SDCA_SELECTED_MODE_TERM_TYPE, i);
+
+		values[i + 3] = sdca_range(range, SDCA_SELECTED_MODE_INDEX, i);

Humm, my memory of GEs is that the hardware reports a 'DETECTED_MODE' and then software (kernel and/or userspace) can choose a "SELECTED_MODE". here we only deal with SELECTED_MODE, is this intentional?

Quoted 35 lines
sound/soc/sdca/sdca_asoc.c new file +34 -0
+		texts[i + 3] = get_terminal_name(type);
+		if (!texts[i + 3]) {
+			dev_err(dev, "%s: unrecognised terminal type: %#x\n",
+				entity->label, type);
+			return -EINVAL;
+		}
+	}
+
+	soc_enum->reg = SDW_SDCA_CTL(function->desc->adr, entity->id, control->sel, 0);
+	soc_enum->items = range->rows + 3;
+	soc_enum->mask = roundup_pow_of_two(soc_enum->items) - 1;
+	soc_enum->texts = texts;
+	soc_enum->values = values;
+
+	kctl->iface = SNDRV_CTL_ELEM_IFACE_MIXER;
+	kctl->name = control_name;
+	kctl->info = snd_soc_info_enum_double;
+	kctl->get = snd_soc_dapm_get_enum_double;
+	kctl->put = snd_soc_dapm_put_enum_double;
+	kctl->private_value = (unsigned long)soc_enum;
+
+	entity->ge.kctl = kctl;
+
+	return 0;
+}
 
+static int entity_parse_it(struct device *dev,
+			   struct sdca_function_data *function,
+			   struct sdca_entity *entity,
+			   struct snd_soc_dapm_widget **widget,
+			   struct snd_soc_dapm_route **route)
+{
+	int i;
+
+	if (entity->iot.is_dataport) {

sound/soc/sdca/sdca_asoc.c line 293

Quoted 10 lines
sound/soc/sdca/sdca_asoc.c new file +10 -0
+		const char *aif_name = devm_kasprintf(dev, GFP_KERNEL, "%s %s",
+						      entity->label, "Playback");
+		if (!aif_name)
+			return -ENOMEM;
+
+		(*widget)->id = snd_soc_dapm_aif_in;
+
+		add_route(route, entity->label, NULL, aif_name);
+	} else {
+		(*widget)->id = snd_soc_dapm_mic;

so for non-dataport terminals there's no route added, but didn't the intro say that there was a virtual control for such terminals? that seems like a route definition if a control can alter the data path, no?

Quoted 9 lines
sound/soc/sdca/sdca_asoc.c new file +9 -0
+	}
+
+	if (entity->iot.clock)
+		add_route(route, entity->label, NULL, entity->iot.clock->label);
+
+	for (i = 0; i < entity->num_sources; i++)
+		add_route(route, entity->label, NULL, entity->sources[i]->label);
+
+	(*widget)++;

sound/soc/sdca/sdca_asoc.c line 344

Quoted 10 lines
sound/soc/sdca/sdca_asoc.c new file +9 -0
+
+	return 0;
+}
 
+static int entity_pde_event(struct snd_soc_dapm_widget *widget,
+			    struct snd_kcontrol *kctl, int event)
+{
+	struct snd_soc_component *component = widget->dapm->component;
+	struct sdca_entity *entity = widget->priv;
+	static const int poll_us = 10000;

This should be a #define, no?

Quoted 25 lines
sound/soc/sdca/sdca_asoc.c new file +25 -0
+	int polls = 1;
+	unsigned int reg, val;
+	int from, to, i;
+	int ret;
+
+	if (!component)
+		return -EIO;
+
+	switch (event) {
+	case SND_SOC_DAPM_POST_PMD:
+		from = widget->on_val;
+		to = widget->off_val;
+		break;
+	case SND_SOC_DAPM_POST_PMU:
+		from = widget->off_val;
+		to = widget->on_val;
+		break;
+	}
+
+	for (i = 0; i < entity->pde.num_max_delay; i++) {
+		struct sdca_pde_delay *delay = &entity->pde.max_delay[i];
+
+		if (delay->from_ps == from && delay->to_ps == to) {
+			polls = max(polls, delay->us / poll_us);
+			break;

sound/soc/sdca/sdca_asoc.c line 379

Quoted 10 lines
sound/soc/sdca/sdca_asoc.c new file +10 -0
+		}
+	}
+
+	reg = SDW_SDCA_CTL(SDW_SDCA_CTL_FUNC(widget->reg),
+			   SDW_SDCA_CTL_ENT(widget->reg),
+			   SDCA_CTL_PDE_ACTUAL_PS, 0);
+
+	for (i = 0; i < polls; i++) {
+		if (i)
+			fsleep(poll_us);

The logic seems inverted? I would first sleep by the amount of time required for a transition before reading the status register. Here we read it first and then sleep for following iterations. The first read is almost guaranteed to fail.

Quoted 20 lines
sound/soc/sdca/sdca_asoc.c new file +20 -0
+
+		ret = regmap_read(component->regmap, reg, &val);
+		if (ret)
+			return ret;
+		else if (val == to)
+			return 0;
+	}
+
+	dev_err(component->dev, "%s: power transition failed: %x\n",
+		entity->label, val);
+	return -ETIMEDOUT;
+}
+
+static int entity_parse_pde(struct device *dev,
+			    struct sdca_function_data *function,
+			    struct sdca_entity *entity,
+			    struct snd_soc_dapm_widget **widget,
+			    struct snd_soc_dapm_route **route)
+{
+	unsigned int target = (1 << SDCA_PDE_PS0) | (1 << SDCA_PDE_PS3);

sound/soc/sdca/sdca_asoc.c line 409

Quoted 10 lines
sound/soc/sdca/sdca_asoc.c new file +10 -0
+	struct sdca_control_range *range;
+	struct sdca_control *control;
+	unsigned int mask = 0;
+	int i;
+
+	control = selector_find_control(dev, entity, SDCA_CTL_PDE_REQUESTED_PS);
+	if (!control)
+		return -EINVAL;
+
+	/* Power should only be controlled by the driver */

is it though? With the virtual controls you were referring to earlier, it's possible that the path is not used and the PDE in PS3. I guess I am still in the dark on what controls the PDE.

Quoted 65 lines
sound/soc/sdca/sdca_asoc.c new file +64 -0
+	if (control->layers != SDCA_ACCESS_LAYER_CLASS)
+		dev_warn(dev, "%s: unexpected access layer: %x\n",
+			 entity->label, control->layers);
+
+	range = control_find_range(dev, entity, control, SDCA_REQUESTED_PS_NCOLS, 0);
+	if (!range)
+		return -EINVAL;
+
+	for (i = 0; i < range->rows; i++)
+		mask |= 1 << sdca_range(range, SDCA_REQUESTED_PS_STATE, i);
+
+	if ((mask & target) != target) {
+		dev_err(dev, "%s: power control missing states\n", entity->label);
+		return -EINVAL;
+	}
+
+	(*widget)->id = snd_soc_dapm_supply;
+	(*widget)->reg = SDW_SDCA_CTL(function->desc->adr, entity->id, control->sel, 0);
+	(*widget)->mask = GENMASK(control->nbits - 1, 0);
+	(*widget)->on_val = SDCA_PDE_PS0;
+	(*widget)->off_val = SDCA_PDE_PS3;
+	(*widget)->event_flags = SND_SOC_DAPM_POST_PMU | SND_SOC_DAPM_POST_PMD;
+	(*widget)->event = entity_pde_event;
+	(*widget)->priv = entity;
+	(*widget)++;
+
+	for (i = 0; i < entity->pde.num_managed; i++)
+		add_route(route, entity->pde.managed[i]->label, NULL, entity->label);
+
+	for (i = 0; i < entity->num_sources; i++)
+		add_route(route, entity->label, NULL, entity->sources[i]->label);
+
+	return 0;
+}
 
+/* Class selector units will be exported as an ALSA control */
+static int entity_parse_su_class(struct device *dev,
+				 struct sdca_function_data *function,
+				 struct sdca_entity *entity,
+				 struct sdca_control *control,
+				 struct snd_soc_dapm_widget **widget,
+				 struct snd_soc_dapm_route **route)
+{
+	struct snd_kcontrol_new *kctl;
+	struct soc_enum *soc_enum;
+	const char **texts;
+	int i;
+
+	kctl = devm_kmalloc(dev, sizeof(*kctl), GFP_KERNEL);
+	if (!kctl)
+		return -ENOMEM;
+
+	soc_enum = devm_kmalloc(dev, sizeof(*soc_enum), GFP_KERNEL);
+	if (!soc_enum)
+		return -ENOMEM;
+
+	texts = devm_kcalloc(dev, entity->num_sources + 1, sizeof(*texts), GFP_KERNEL);
+	if (!texts)
+		return -ENOMEM;
+
+	texts[0] = "No Signal";
+	for (i = 0; i < entity->num_sources; i++)
+		texts[i + 1] = entity->sources[i]->label;
+
+	soc_enum->reg = SDW_SDCA_CTL(function->desc->adr, entity->id, control->sel, 0);

sound/soc/sdca/sdca_asoc.c line 549

Quoted 10 lines
sound/soc/sdca/sdca_asoc.c new file +10 -0
+	soc_enum->items = entity->num_sources + 1;
+	soc_enum->mask = roundup_pow_of_two(soc_enum->items) - 1;
+	soc_enum->texts = texts;
+
+	kctl->iface = SNDRV_CTL_ELEM_IFACE_MIXER;
+	kctl->name = "Route";
+	kctl->info = snd_soc_info_enum_double;
+	kctl->get = snd_soc_dapm_get_enum_double;
+	kctl->put = snd_soc_dapm_put_enum_double;
+	kctl->private_value = (unsigned long)soc_enum;

Can you remind me in which cases we have a Selector Unit not controlled by a GE, and what userspace would do with such kcontrols?

Quoted 26 lines
sound/soc/sdca/sdca_asoc.c new file +25 -0
+
+	(*widget)->id = snd_soc_dapm_mux;
+	(*widget)->kcontrol_news = kctl;
+	(*widget)->num_kcontrols = 1;
+	(*widget)++;
+
+	for (i = 0; i < entity->num_sources; i++)
+		add_route(route, entity->label, texts[i + 1], entity->sources[i]->label);
+
+	return 0;
+}
 
+static int entity_parse_mu(struct device *dev,
+			   struct sdca_function_data *function,
+			   struct sdca_entity *entity,
+			   struct snd_soc_dapm_widget **widget,
+			   struct snd_soc_dapm_route **route)
+{
+	struct sdca_control *control;
+	struct snd_kcontrol_new *kctl;
+	int cn;
+	int i;
+
+	if (!entity->num_sources) {
+		dev_err(dev, "%s: selector 1 or more inputs\n", entity->label);
+		return -EINVAL;

sound/soc/sdca/sdca_asoc.c line 612

Quoted 10 lines
sound/soc/sdca/sdca_asoc.c new file +10 -0
+	}
+
+	control = selector_find_control(dev, entity, SDCA_CTL_MU_MIXER);
+	if (!control)
+		return -EINVAL;
+
+	/* MU control should be through DAPM */
+	if (control->layers != SDCA_ACCESS_LAYER_CLASS)
+		dev_warn(dev, "%s: unexpected access layer: %x\n",
+			 entity->label, control->layers);

isn't this an error? How would DAPM deal with this case?

Quoted 69 lines
sound/soc/sdca/sdca_asoc.c new file +68 -0
+
+	if (entity->num_sources != hweight64(control->cn_list)) {
+		dev_err(dev, "%s: mismatched control and sources\n", entity->label);
+		return -EINVAL;
+	}
+
+	kctl = devm_kcalloc(dev, entity->num_sources, sizeof(*kctl), GFP_KERNEL);
+	if (!kctl)
+		return -ENOMEM;
+
+	i = 0;
+	for_each_set_bit(cn, (unsigned long *)&control->cn_list,
+			 BITS_PER_TYPE(control->cn_list)) {
+		const char *control_name;
+		struct soc_mixer_control *mc;
+
+		control_name = devm_kasprintf(dev, GFP_KERNEL, "%s %d",
+					      control->label, i + 1);
+		if (!control_name)
+			return -ENOMEM;
+
+		mc = devm_kmalloc(dev, sizeof(*mc), GFP_KERNEL);
+		if (!mc)
+			return -ENOMEM;
+
+		mc->reg = SND_SOC_NOPM;
+		mc->rreg = SND_SOC_NOPM;
+		mc->invert = 1; // Ensure default is connected
+		mc->min = 0;
+		mc->max = 1;
+
+		kctl[i].name = control_name;
+		kctl[i].private_value = (unsigned long)mc;
+		kctl[i].iface = SNDRV_CTL_ELEM_IFACE_MIXER;
+		kctl[i].info = snd_soc_info_volsw;
+		kctl[i].get = snd_soc_dapm_get_volsw;
+		kctl[i].put = snd_soc_dapm_put_volsw;
+		i++;
+	}
+
+	(*widget)->id = snd_soc_dapm_mixer;
+	(*widget)->kcontrol_news = kctl;
+	(*widget)->num_kcontrols = entity->num_sources;
+	(*widget)++;
+
+	for (i = 0; i < entity->num_sources; i++)
+		add_route(route, entity->label, kctl[i].name, entity->sources[i]->label);
+
+	return 0;
+}
 
+/**
+ * sdca_asoc_populate_dapm - fill in arrays of DAPM widgets and routes
+ * @dev: Pointer to the device against which allocations will be done.
+ * @function: Pointer to the Function information.
+ * @widget: Array of DAPM widgets to be populated.
+ * @route: Array of DAPM routes to be populated.
+ *
+ * This function populates arrays of DAPM widgets and routes from the
+ * DisCo information for a particular SDCA Function. Typically,
+ * snd_soc_asoc_count_component will be used to allocate appropriately
+ * sized arrays before calling this function.
+ *
+ * Return: Returns zero on success, and a negative error code on failure.
+ */
+int sdca_asoc_populate_dapm(struct device *dev, struct sdca_function_data *function,
+			    struct snd_soc_dapm_widget *widget,
+			    struct snd_soc_dapm_route *route)
+{

sound/soc/sdca/sdca_asoc.c line 727

Quoted 10 lines
sound/soc/sdca/sdca_asoc.c new file +10 -0
+	int ret;
+	int i;
+
+	/* Some entities need to add controls referenced by other entities */
+	for (i = 0; i < function->num_entities - 1; i++) {
+		struct sdca_entity *entity = &function->entities[i];
+
+		switch (entity->type) {
+		case SDCA_ENTITY_TYPE_GE:
+			ret = entity_early_parse_ge(dev, function, entity);

what does 'early' mean here? is there a 'late_parse_ge', or is this to say that GEs are parsed first?

Quoted 7 lines
sound/soc/sdca/sdca_asoc.c new file +7 -0
+			if (ret)
+				return ret;
+			break;
+		default:
+			break;
+		}
+	}
#15 lore raw

Re: [PATCH v5 5/6] ASoC: SDCA: Create ALSA controls from DisCo

In reply to Charles Keepax

sound/soc/sdca/sdca_asoc.c static struct sdca_control_range *selector_find_range(struct device *dev, · line 77

Quoted 6 lines
sound/soc/sdca/sdca_asoc.c +5 -0
+static bool exported_control(struct sdca_control *control)
+{
+	/* No need to export control for something that only has one value */
+	if (control->has_fixed)
+		return false;

why is that? It'd simplify userspace if there was always a way to get a value, even if it's fixed?

sound/soc/sdca/sdca_asoc.c static struct sdca_control_range *selector_find_range(struct device *dev, · line 80

Quoted 3 lines
sound/soc/sdca/sdca_asoc.c +3 -0
+
+	return control->layers & (SDCA_ACCESS_LAYER_USER |
+				  SDCA_ACCESS_LAYER_APPLICATION);

That filter on access layer does make sense however.

Quoted 31 lines
sound/soc/sdca/sdca_asoc.c +7 -2
+}
+
 /**
  * sdca_asoc_count_component - count the various component parts
  * @function: Pointer to the Function information.
@@ -76,6 +87,8 @@ static struct sdca_control_range *selector_find_range(struct device *dev,
  * required number of DAPM widgets for the Function.
  * @num_routes: Output integer pointer, will be filled with the
  * required number of DAPM routes for the Function.
+ * @num_controls: Output integer pointer, will be filled with the
+ * required number of ALSA controls for the Function.
  *
  * This function counts various things within the SDCA Function such
  * that the calling driver can allocate appropriate space before
@@ -84,12 +97,13 @@ static struct sdca_control_range *selector_find_range(struct device *dev,
  * Return: Returns zero on success, and a negative error code on failure.
  */
 int sdca_asoc_count_component(struct device *dev, struct sdca_function_data *function,
-			      int *num_widgets, int *num_routes)
+			      int *num_widgets, int *num_routes, int *num_controls)
 {
-	int i;
+	int i, j;
 
 	*num_widgets = function->num_entities - 1;
 	*num_routes = 0;
+	*num_controls = 0;
 
 	for (i = 0; i < function->num_entities - 1; i++) {
 		struct sdca_entity *entity = &function->entities[i];

sound/soc/sdca/sdca_asoc.c int sdca_asoc_count_component(struct device *dev, struct sdca_function_data *fun · line 116

Quoted 4 lines
sound/soc/sdca/sdca_asoc.c +1 -0
@@ -99,6 +113,7 @@ int sdca_asoc_count_component(struct device *dev, struct sdca_function_data *fun
 		case SDCA_ENTITY_TYPE_OT:
 			*num_routes += !!entity->iot.clock;
 			*num_routes += !!entity->iot.is_dataport;
+			*num_controls += !entity->iot.is_dataport;

shouldn't the same line be added for TYPE_IT?

Quoted 15 lines
sound/soc/sdca/sdca_asoc.c +5 -0
 			break;
 		case SDCA_ENTITY_TYPE_PDE:
 			*num_routes += entity->pde.num_managed;
@@ -111,6 +126,11 @@ int sdca_asoc_count_component(struct device *dev, struct sdca_function_data *fun
 
 		if (entity->group)
 			(*num_routes)++;
+
+		for (j = 0; j < entity->num_controls; j++) {
+			if (exported_control(&entity->controls[j]))
+				(*num_controls)++;
+		}
 	}
 
 	return 0;
#22 lore raw

Re: [PATCH v5 6/6] ASoC: SDCA: Create DAI drivers from DisCo

In reply to Charles Keepax
Quoted 18 lines
sound/soc/sdca/sdca_asoc.c +17 -0
+int sdca_asoc_populate_dais(struct device *dev, struct sdca_function_data *function,
+			    struct snd_soc_dai_driver *dais,
+			    const struct snd_soc_dai_ops *ops)
+{
+	int i, j;
+	int ret;
+
+	for (i = 0, j = 0; i < function->num_entities - 1; i++) {
+		struct sdca_entity *entity = &function->entities[i];
+		struct snd_soc_pcm_stream *stream;
+		const char *stream_suffix;
+
+		switch (entity->type) {
+		case SDCA_ENTITY_TYPE_IT:
+			stream = &dais[j].playback;
+			stream_suffix = "Playback";
+			break;

sound/soc/sdca/sdca_asoc.c int sdca_asoc_populate_controls(struct device *dev, · line 1197

Quoted 10 lines
sound/soc/sdca/sdca_asoc.c +10 -0
+		case SDCA_ENTITY_TYPE_OT:
+			stream = &dais[j].capture;
+			stream_suffix = "Capture";
+			break;
+		default:
+			continue;
+		}
+
+		if (!entity->iot.is_dataport)
+			continue;

should this be moved as the first test in the for() loop? The switch cases for IT/OT only makes sense for streaming/dataport terminals, no?

Quoted 22 lines
sound/soc/sdca/sdca_asoc.c +22 -0
+
+		stream->stream_name = devm_kasprintf(dev, GFP_KERNEL, "%s %s",
+						     entity->label, stream_suffix);
+		if (!stream->stream_name)
+			return -ENOMEM;
+		/* Channels will be further limited by constraints */
+		stream->channels_min = 1;
+		stream->channels_max = SDCA_MAX_CHANNEL_COUNT;
+
+		ret = populate_rate_format(dev, function, entity, stream);
+		if (ret)
+			return ret;
+
+		dais[j].id = i;
+		dais[j].name = entity->label;
+		dais[j].ops = ops;
+		j++;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_NS(sdca_asoc_populate_dais, "SND_SOC_SDCA");
author maintainer
#6 lore raw

Re: [PATCH v5 4/6] ASoC: SDCA: Create DAPM widgets and routes from DisCo

In reply to Pierre-Louis Bossart
Quoted 12 lines · Pierre-Louis Bossart
Quoted 10 lines · Charles Keepax

SDCA also has a slight habit of having fully connected paths, relying more on activating the PDEs to enable functionality. This doesn't map quite so perfectly to DAPM which considers the path a reason to power the PDE. Whilst in the current specification Mixer Units are defined as fixed-function, in DAPM we create a virtual control for each input (which defaults to connected). This allows paths to be connected/disconnected, providing a more ASoC style approach to managing the power. PIN_SWITCHs will also be added for non-dataport terminal entities in a later patch along with the other ALSA controls, providing greater flexibility in power management.

Quoted 7 lines

It's been a while since I reviewed an earlier versions so now I am confused.
- Is the proposal to have a control to change the PDE state directly, as well as a virtual control to describe a path activation? That would lead to more flexibility but also to confusion, e.g. with 'no sound' when the paths are active but the PDE still in PS3.

No the PDEs are their own DAPM widgets and are controlled by DAPM, they power up when they are on an active path. User-space has no direct control over the power state.

The two methods user-space has for affecting the power state of the system. Firstly, the virtual switches on the mixer units, I initially added. These let you connect/disconnect a route into the MU. Secondly, the pin switched on the endpoints that you requested, these let you enable/disable a non-dataport endpoint.

Quoted 4 lines

- Or do those virtual controls manage the PDE state, without userspace interaction? In that case the 'connected' default would higher power consumption until userspace decides which paths it really wants active.

All the pin switches default to on, but that is normal and probably makes the most sense since that way paths will generally work without any interaction.

Quoted 8 lines

sound/soc/sdca/sdca_asoc.c line 101

Quoted 5 lines
sound/soc/sdca/sdca_asoc.c new file +5 -0
+		switch (entity->type) {
+		case SDCA_ENTITY_TYPE_IT:
+		case SDCA_ENTITY_TYPE_OT:
+			*num_routes += !!entity->iot.clock;
+			*num_routes += !!entity->iot.is_dataport;

I couldn't follow those two lines.
Why does the clock play a role, for now it's not really managed?

Clocks are managed, it is simply the clock muxes that are not fully supported at the moment. Which is because clocks are supply widgets and DAPM doesn't really have a concept of selectively connecting supply widgets at the moment. Which is a little fiddly to implement and given our parts don't use any clock muxes not high on my priority list.

Quoted 2 lines

And the difference between streaming and transducer terminals shouldn't change the accounting of routes?

Yes it does, remember everything is being mapped to DAPM widgets here, so streaming terminals need a DAPM route in from the DAI widget.

Quoted 6 lines

sound/soc/sdca/sdca_asoc.c line 104

Quoted 3 lines
sound/soc/sdca/sdca_asoc.c new file +3 -0
+			break;
+		case SDCA_ENTITY_TYPE_PDE:
+			*num_routes += entity->pde.num_managed;

same here, isn't there a risk of the same route being counted multiple times?

The PDE is implemented as a supply widget and it will be connected to every widget it controls with a DAPM route to signify that connection, this adds space for each of those routes.

Quoted 4 lines
Quoted 4 lines
sound/soc/sdca/sdca_asoc.c new file +4 -0
+			break;
+		default:
+			break;
+		}
Quoted 8 lines

sound/soc/sdca/sdca_asoc.c line 113

Quoted 5 lines
sound/soc/sdca/sdca_asoc.c new file +5 -0
+
+		*num_routes += entity->num_sources;
+
+		if (entity->group)
+			(*num_routes)++;

And here as well, in case an entity is controlled by a GE does the number of routes really change?

Yes because the GE is also implemented as a supply widget and linked to all the widgets it controls.

Quoted 3 lines

I guess the meaning of 'routes' isn't really clear in my head, it's not just data paths but control as well that's taken into account?

I mean the hard definition here is DAPM routes. But roughly speaking each line on the SDCA function diagram will correspond to a route here, with a few tweaks like the GEs only show a single line going to a red box around a bunch of other entities that will be multiple routes one for each entity in the box.

Quoted 8 lines
Quoted 8 lines
sound/soc/sdca/sdca_asoc.c new file +8 -0
+	texts[0] = "No Jack";
+	texts[1] = "Jack Unknown";
+	texts[2] = "Detection in Progress";
+	values[0] = 0;
+	values[1] = 1;
+	values[2] = 2;
+	for (i = 0; i < range->rows; i++) {
+		enum sdca_terminal_type type;
Quoted 9 lines

sound/soc/sdca/sdca_asoc.c line 221

Quoted 4 lines
sound/soc/sdca/sdca_asoc.c new file +4 -0
+
+		type = sdca_range(range, SDCA_SELECTED_MODE_TERM_TYPE, i);
+
+		values[i + 3] = sdca_range(range, SDCA_SELECTED_MODE_INDEX, i);

Humm, my memory of GEs is that the hardware reports a 'DETECTED_MODE' and then software (kernel and/or userspace) can choose a "SELECTED_MODE". here we only deal with SELECTED_MODE, is this intentional?

This code is creating the ALSA control for the SELECTED_MODE control, the DETECTED_MODE is all handled internally and there is no need to export it.

Quoted 7 lines
Quoted 7 lines
sound/soc/sdca/sdca_asoc.c new file +7 -0
+	if (entity->iot.is_dataport) {
+		const char *aif_name = devm_kasprintf(dev, GFP_KERNEL, "%s %s",
+						      entity->label, "Playback");
+		if (!aif_name)
+			return -ENOMEM;
+
+		(*widget)->id = snd_soc_dapm_aif_in;
Quoted 9 lines

sound/soc/sdca/sdca_asoc.c line 293

Quoted 4 lines
sound/soc/sdca/sdca_asoc.c new file +4 -0
+
+		add_route(route, entity->label, NULL, aif_name);
+	} else {
+		(*widget)->id = snd_soc_dapm_mic;

so for non-dataport terminals there's no route added, but didn't the intro say that there was a virtual control for such terminals? that seems like a route definition if a control can alter the data path, no?

This add route is for connecting the DAI widget, the actual routes specified by the SDCA topology are added in the for loop below.

And the PIN_SWITCH's are added in the next patch which adds the non-DAPM ALSA controls, as noted in the commit message.

Quoted 13 lines
Quoted 12 lines
sound/soc/sdca/sdca_asoc.c new file +12 -0
+	}
+
+	if (entity->iot.clock)
+		add_route(route, entity->label, NULL, entity->iot.clock->label);
+
+	for (i = 0; i < entity->num_sources; i++)
+		add_route(route, entity->label, NULL, entity->sources[i]->label);
+
+	(*widget)++;
+
+	return 0;
+}
Quoted 9 lines

sound/soc/sdca/sdca_asoc.c line 344

Quoted 6 lines
sound/soc/sdca/sdca_asoc.c new file +6 -0
+static int entity_pde_event(struct snd_soc_dapm_widget *widget,
+			    struct snd_kcontrol *kctl, int event)
+{
+	struct snd_soc_component *component = widget->dapm->component;
+	struct sdca_entity *entity = widget->priv;
+	static const int poll_us = 10000;

This should be a #define, no?

I could make it a define if you feel strong but I generally quite like if it just a local delay being using in one place to define it locally, means you don't have to look things up when reading the code.

Quoted 9 lines

sound/soc/sdca/sdca_asoc.c line 379

Quoted 3 lines
sound/soc/sdca/sdca_asoc.c new file +3 -0
+	for (i = 0; i < polls; i++) {
+		if (i)
+			fsleep(poll_us);

The logic seems inverted? I would first sleep by the amount of time required for a transition before reading the status register. Here we read it first and then sleep for following iterations. The first read is almost guaranteed to fail.

I mean it would be fine to sleep first but often there is enough slack in the system that things are ready as soon as you get here. Often these transitions are way way way faster than the max time we are parsing from DisCo implies.

Quoted 13 lines
Quoted 13 lines
sound/soc/sdca/sdca_asoc.c new file +13 -0
+static int entity_parse_pde(struct device *dev,
+			    struct sdca_function_data *function,
+			    struct sdca_entity *entity,
+			    struct snd_soc_dapm_widget **widget,
+			    struct snd_soc_dapm_route **route)
+{
+	unsigned int target = (1 << SDCA_PDE_PS0) | (1 << SDCA_PDE_PS3);
+	struct sdca_control_range *range;
+	struct sdca_control *control;
+	unsigned int mask = 0;
+	int i;
+
+	control = selector_find_control(dev, entity, SDCA_CTL_PDE_REQUESTED_PS);
Quoted 8 lines

sound/soc/sdca/sdca_asoc.c line 409

Quoted 4 lines
sound/soc/sdca/sdca_asoc.c new file +4 -0
+	if (!control)
+		return -EINVAL;
+
+	/* Power should only be controlled by the driver */

is it though? With the virtual controls you were referring to earlier, it's possible that the path is not used and the PDE in PS3. I guess I am still in the dark on what controls the PDE.

The PDE is controlled by the DAPM widget created at the bottom of this function and yes power should only be directly controlled by drivers.

Quoted 10 lines

sound/soc/sdca/sdca_asoc.c line 549

Quoted 6 lines
sound/soc/sdca/sdca_asoc.c new file +6 -0
+	kctl->iface = SNDRV_CTL_ELEM_IFACE_MIXER;
+	kctl->name = "Route";
+	kctl->info = snd_soc_info_enum_double;
+	kctl->get = snd_soc_dapm_get_enum_double;
+	kctl->put = snd_soc_dapm_put_enum_double;
+	kctl->private_value = (unsigned long)soc_enum;

Can you remind me in which cases we have a Selector Unit not controlled by a GE, and what userspace would do with such kcontrols?

A selector unit not tidied to a GE is just a DAPM mux, it provides an ALSA control which allows user-space to select which input it wants routed through the selector.

Quoted 6 lines

sound/soc/sdca/sdca_asoc.c line 612

Quoted 4 lines
sound/soc/sdca/sdca_asoc.c new file +4 -0
+	/* MU control should be through DAPM */
+	if (control->layers != SDCA_ACCESS_LAYER_CLASS)
+		dev_warn(dev, "%s: unexpected access layer: %x\n",
+			 entity->label, control->layers);

isn't this an error? How would DAPM deal with this case?

In general the control will work the same no matter what access specifier it is given, it still sets the control. So I didn't really see the need to hard error out here. Mostly likely if one gets these its a sign of slightly dodgy DisCo but we should be fine to carry one.

Quoted 9 lines
Quoted 9 lines
sound/soc/sdca/sdca_asoc.c new file +9 -0
+int sdca_asoc_populate_dapm(struct device *dev, struct sdca_function_data *function,
+			    struct snd_soc_dapm_widget *widget,
+			    struct snd_soc_dapm_route *route)
+{
+	int ret;
+	int i;
+
+	/* Some entities need to add controls referenced by other entities */
+	for (i = 0; i < function->num_entities - 1; i++) {
Quoted 8 lines

sound/soc/sdca/sdca_asoc.c line 727

Quoted 5 lines
sound/soc/sdca/sdca_asoc.c new file +5 -0
+		struct sdca_entity *entity = &function->entities[i];
+
+		switch (entity->type) {
+		case SDCA_ENTITY_TYPE_GE:
+			ret = entity_early_parse_ge(dev, function, entity);

what does 'early' mean here? is there a 'late_parse_ge', or is this to say that GEs are parsed first?

Yeah we need to parse them first such that the GE controls are ready when the SUs are parsed since they will link to the control generated by the GE.

Thanks,
Charles

author maintainer
#16 lore raw

Re: [PATCH v5 5/6] ASoC: SDCA: Create ALSA controls from DisCo

In reply to Pierre-Louis Bossart
Quoted 9 lines · Pierre-Louis Bossart

sound/soc/sdca/sdca_asoc.c static struct sdca_control_range *selector_find_range(struct device *dev, · line 77

Quoted 6 lines
sound/soc/sdca/sdca_asoc.c +5 -0
+static bool exported_control(struct sdca_control *control)
+{
+	/* No need to export control for something that only has one value */
+	if (control->has_fixed)
+		return false;

why is that? It'd simplify userspace if there was always a way to get a value, even if it's fixed?

Well mostly it is because various things in the stack get sad with a single value control you can't set. If we wanted to change that I would suggest we add a FIXME. It isn't really critical and I would really like to get more of this class driver stuff moving, it has been out for review for a really long time now.

Quoted 7 lines

sound/soc/sdca/sdca_asoc.c int sdca_asoc_count_component(struct device *dev, struct sdca_function_data *fun · line 116

Quoted 5 lines
sound/soc/sdca/sdca_asoc.c +1 -0
@@ -99,6 +113,7 @@ int sdca_asoc_count_component(struct device *dev, struct sdca_function_data *fun
 		case SDCA_ENTITY_TYPE_OT:
 			*num_routes += !!entity->iot.clock;
 			*num_routes += !!entity->iot.is_dataport;
+			*num_controls += !entity->iot.is_dataport;

shouldn't the same line be added for TYPE_IT?

This is just the cutting on the patch generation, case
SDCA_ENTITY_TYPE_IT: is the line right above so this runs for both.

Thanks,
Charles

author maintainer
#23 lore raw

Re: [PATCH v5 6/6] ASoC: SDCA: Create DAI drivers from DisCo

In reply to Pierre-Louis Bossart
Quoted 10 lines · Pierre-Louis Bossart
Quoted 10 lines
sound/soc/sdca/sdca_asoc.c +10 -0
+		switch (entity->type) {
+		case SDCA_ENTITY_TYPE_IT:
+			stream = &dais[j].playback;
+			stream_suffix = "Playback";
+			break;
+		case SDCA_ENTITY_TYPE_OT:
+			stream = &dais[j].capture;
+			stream_suffix = "Capture";
+			break;
+		default:
Quoted 8 lines

sound/soc/sdca/sdca_asoc.c int sdca_asoc_populate_controls(struct device *dev, · line 1197

Quoted 5 lines
sound/soc/sdca/sdca_asoc.c +5 -0
+			continue;
+		}
+
+		if (!entity->iot.is_dataport)
+			continue;

should this be moved as the first test in the for() loop? The switch cases for IT/OT only makes sense for streaming/dataport terminals, no?

Can't be first in the loop since the entities can be different types and the iot is part of a union so it is only safe to access is_dataport once you know it is a terminal entity.

Thanks,
Charles

author maintainer
#11 lore raw

Re: [PATCH v5 4/6] ASoC: SDCA: Create DAPM widgets and routes from DisCo

In reply to Pierre-Louis Bossart
Quoted 2 lines · Pierre-Louis Bossart

On 5/12/25 14:42, Charles Keepax wrote:
It's been a while since I reviewed an earlier versions so now I am confused.

I have attached the output of the dapm-graph tool for the UAJ component for one of our devices, whilst a headset is connected. Hopefully this will help to make it a little more clear what the generated graph from this code looks like. Hopefully the attachment makes it through the ether, let me know if it doesn't and I can send directly.

The PDE supply widgets are what control the PDE power registers, they will power up when they are attached to an active path, ie. a valid link between an active source and an active sink.

For the streaming terminals IT41,OT36 they are active when the audio stream is open. For the non-streaming ones IT31,IT32,IT33,OT43,OT44,OT45 these are active when the pin switch is enabled which it is by default, without pin switches these would be always active. The pin switches are not shown on the output of dapm-graph.

You can see the effect of the GE control, where it has set the input of SU35 to FU33, by the selection of "Headset". And you can see the virtual mixer switches going into MU35, currently both Mixer 1 and 2 are on, so you can see the paths connected.

But as you can see the correlation to the function diagrams is the spec is pretty good. As I noted with a few minor exceptions, like the GE having more paths, since it must connect to all the widgets it controls separately. And the DAI widgets that are created by the DAPM core.

Thanks,
Charles

author maintainer
#17 lore raw

Re: [PATCH v5 5/6] ASoC: SDCA: Create ALSA controls from DisCo

In reply to Charles Keepax
Quoted 11 lines · Charles Keepax
Quoted 9 lines · Pierre-Louis Bossart

sound/soc/sdca/sdca_asoc.c static struct sdca_control_range *selector_find_range(struct device *dev, · line 77

Quoted 6 lines
sound/soc/sdca/sdca_asoc.c +5 -0
+static bool exported_control(struct sdca_control *control)
+{
+	/* No need to export control for something that only has one value */
+	if (control->has_fixed)
+		return false;

why is that? It'd simplify userspace if there was always a way to get a value, even if it's fixed?

Quoted 7 lines

Well mostly it is because various things in the stack get sad with a single value control you can't set. If we wanted to change that I would suggest we add a FIXME. It isn't really critical and I would really like to get more of this class driver stuff moving, it has been out for review for a really long time now.

Thinking a little more about this user-space is going to have to handle controls not being there as different controls will be exported across different parts anyway, based of which options are implemented. Also we have seen a bit of people hiding controls by changing the access layer. Although sometimes that does result in things that are technically not class conformant but they do work through the Microsoft drivers, so we will likely need to support it anyway.

Thanks,
Charles

#7 lore raw

Re: [PATCH v5 4/6] ASoC: SDCA: Create DAPM widgets and routes from DisCo

In reply to Charles Keepax

Thanks for the comments Charles, I am fine with most of your points except the DETECTED_MODE and a question on the non-GE-managed-SUs, see below -Pierre

Quoted 23 lines
Quoted 11 lines
Quoted 10 lines

SDCA also has a slight habit of having fully connected paths, relying more on activating the PDEs to enable functionality. This doesn't map quite so perfectly to DAPM which considers the path a reason to power the PDE. Whilst in the current specification Mixer Units are defined as fixed-function, in DAPM we create a virtual control for each input (which defaults to connected). This allows paths to be connected/disconnected, providing a more ASoC style approach to managing the power. PIN_SWITCHs will also be added for non-dataport terminal entities in a later patch along with the other ALSA controls, providing greater flexibility in power management.

Quoted 7 lines

It's been a while since I reviewed an earlier versions so now I am confused.
- Is the proposal to have a control to change the PDE state directly, as well as a virtual control to describe a path activation? That would lead to more flexibility but also to confusion, e.g. with 'no sound' when the paths are active but the PDE still in PS3.

No the PDEs are their own DAPM widgets and are controlled by DAPM, they power up when they are on an active path. User-space has no direct control over the power state.

Quoted 5 lines

The two methods user-space has for affecting the power state of the system. Firstly, the virtual switches on the mixer units, I initially added. These let you connect/disconnect a route into the MU. Secondly, the pin switched on the endpoints that you requested, these let you enable/disable a non-dataport endpoint.

ok

Quoted 8 lines
Quoted 4 lines

- Or do those virtual controls manage the PDE state, without userspace interaction? In that case the 'connected' default would higher power consumption until userspace decides which paths it really wants active.

All the pin switches default to on, but that is normal and probably makes the most sense since that way paths will generally work without any interaction.

ok

Quoted 9 lines
Quoted 8 lines

sound/soc/sdca/sdca_asoc.c line 101

Quoted 5 lines
sound/soc/sdca/sdca_asoc.c new file +5 -0
+		switch (entity->type) {
+		case SDCA_ENTITY_TYPE_IT:
+		case SDCA_ENTITY_TYPE_OT:
+			*num_routes += !!entity->iot.clock;
+			*num_routes += !!entity->iot.is_dataport;

I couldn't follow those two lines.
Why does the clock play a role, for now it's not really managed?

Quoted 6 lines

Clocks are managed, it is simply the clock muxes that are not fully supported at the moment. Which is because clocks are supply widgets and DAPM doesn't really have a concept of selectively connecting supply widgets at the moment. Which is a little fiddly to implement and given our parts don't use any clock muxes not high on my priority list.

ok, makes sense.

Quoted 6 lines
Quoted 2 lines

And the difference between streaming and transducer terminals shouldn't change the accounting of routes?

Yes it does, remember everything is being mapped to DAPM widgets here, so streaming terminals need a DAPM route in from the DAI widget.

ah yes, you added one layer. I naively thought the DAI was the input/output terminal directly.

Quoted 11 lines
Quoted 6 lines

sound/soc/sdca/sdca_asoc.c line 104

Quoted 3 lines
sound/soc/sdca/sdca_asoc.c new file +3 -0
+			break;
+		case SDCA_ENTITY_TYPE_PDE:
+			*num_routes += entity->pde.num_managed;

same here, isn't there a risk of the same route being counted multiple times?

The PDE is implemented as a supply widget and it will be connected to every widget it controls with a DAPM route to signify that connection, this adds space for each of those routes.

ok.

Quoted 20 lines
Quoted 4 lines
Quoted 4 lines
sound/soc/sdca/sdca_asoc.c new file +4 -0
+			break;
+		default:
+			break;
+		}
Quoted 8 lines

sound/soc/sdca/sdca_asoc.c line 113

Quoted 5 lines
sound/soc/sdca/sdca_asoc.c new file +5 -0
+
+		*num_routes += entity->num_sources;
+
+		if (entity->group)
+			(*num_routes)++;

And here as well, in case an entity is controlled by a GE does the number of routes really change?

Yes because the GE is also implemented as a supply widget and linked to all the widgets it controls.

Quoted 3 lines

I guess the meaning of 'routes' isn't really clear in my head, it's not just data paths but control as well that's taken into account?

Quoted 5 lines

I mean the hard definition here is DAPM routes. But roughly speaking each line on the SDCA function diagram will correspond to a route here, with a few tweaks like the GEs only show a single line going to a red box around a bunch of other entities that will be multiple routes one for each entity in the box.

This could be described in comments. The SDCA spec shows lines of different color for pure data paths and 'control' links. Here we are bringing everything back to the same concept of DAPM routes, but those extra routes only connect to supply widgets.

Quoted 18 lines
Quoted 8 lines
Quoted 8 lines
sound/soc/sdca/sdca_asoc.c new file +8 -0
+	texts[0] = "No Jack";
+	texts[1] = "Jack Unknown";
+	texts[2] = "Detection in Progress";
+	values[0] = 0;
+	values[1] = 1;
+	values[2] = 2;
+	for (i = 0; i < range->rows; i++) {
+		enum sdca_terminal_type type;
Quoted 9 lines

sound/soc/sdca/sdca_asoc.c line 221

Quoted 4 lines
sound/soc/sdca/sdca_asoc.c new file +4 -0
+
+		type = sdca_range(range, SDCA_SELECTED_MODE_TERM_TYPE, i);
+
+		values[i + 3] = sdca_range(range, SDCA_SELECTED_MODE_INDEX, i);

Humm, my memory of GEs is that the hardware reports a 'DETECTED_MODE' and then software (kernel and/or userspace) can choose a "SELECTED_MODE". here we only deal with SELECTED_MODE, is this intentional?

Quoted 3 lines

This code is creating the ALSA control for the SELECTED_MODE control, the DETECTED_MODE is all handled internally and there is no need to export it.

Humm, that one doesn't seem right. There could be cases where the DETECTED_MODE is inconclusive, or that additional vendor-specific steps are needed to decide which SELECTED_MODE makes sense. I don't think we should assume everything is handled internally, or we have to define what 'internally' means. This cannot be generic SDCA stuff, it has to be configurable for specific implementations/vendors.

Quoted 17 lines
Quoted 7 lines
Quoted 7 lines
sound/soc/sdca/sdca_asoc.c new file +7 -0
+	if (entity->iot.is_dataport) {
+		const char *aif_name = devm_kasprintf(dev, GFP_KERNEL, "%s %s",
+						      entity->label, "Playback");
+		if (!aif_name)
+			return -ENOMEM;
+
+		(*widget)->id = snd_soc_dapm_aif_in;
Quoted 9 lines

sound/soc/sdca/sdca_asoc.c line 293

Quoted 4 lines
sound/soc/sdca/sdca_asoc.c new file +4 -0
+
+		add_route(route, entity->label, NULL, aif_name);
+	} else {
+		(*widget)->id = snd_soc_dapm_mic;

so for non-dataport terminals there's no route added, but didn't the intro say that there was a virtual control for such terminals? that seems like a route definition if a control can alter the data path, no?

Quoted 6 lines

This add route is for connecting the DAI widget, the actual routes specified by the SDCA topology are added in the for loop below.

And the PIN_SWITCH's are added in the next patch which adds the non-DAPM ALSA controls, as noted in the commit message.

ok

Quoted 23 lines
Quoted 13 lines
Quoted 12 lines
sound/soc/sdca/sdca_asoc.c new file +12 -0
+	}
+
+	if (entity->iot.clock)
+		add_route(route, entity->label, NULL, entity->iot.clock->label);
+
+	for (i = 0; i < entity->num_sources; i++)
+		add_route(route, entity->label, NULL, entity->sources[i]->label);
+
+	(*widget)++;
+
+	return 0;
+}
Quoted 9 lines

sound/soc/sdca/sdca_asoc.c line 344

Quoted 6 lines
sound/soc/sdca/sdca_asoc.c new file +6 -0
+static int entity_pde_event(struct snd_soc_dapm_widget *widget,
+			    struct snd_kcontrol *kctl, int event)
+{
+	struct snd_soc_component *component = widget->dapm->component;
+	struct sdca_entity *entity = widget->priv;
+	static const int poll_us = 10000;

This should be a #define, no?

Quoted 4 lines

I could make it a define if you feel strong but I generally quite like if it just a local delay being using in one place to define it locally, means you don't have to look things up when reading the code.

I guess my main point was to explain where this 10ms value comes from. I am not sure what to make of it, it's either very large or very small depending on the initial and desired power states. Plus I would expect the sleep times to vary by orders of magnitude depending on the type of function managed.

Quoted 10 lines
Quoted 9 lines

sound/soc/sdca/sdca_asoc.c line 379

Quoted 3 lines
sound/soc/sdca/sdca_asoc.c new file +3 -0
+	for (i = 0; i < polls; i++) {
+		if (i)
+			fsleep(poll_us);

The logic seems inverted? I would first sleep by the amount of time required for a transition before reading the status register. Here we read it first and then sleep for following iterations. The first read is almost guaranteed to fail.

Quoted 4 lines

I mean it would be fine to sleep first but often there is enough slack in the system that things are ready as soon as you get here. Often these transitions are way way way faster than the max time we are parsing from DisCo implies.

Sorry, my bad. I thought the DisCo definitions were minimum sleep times, as for HDaudio. But they are worst-case times, so it makes sense to try one's luck early and wait later.

Quoted 22 lines
Quoted 13 lines
Quoted 13 lines
sound/soc/sdca/sdca_asoc.c new file +13 -0
+static int entity_parse_pde(struct device *dev,
+			    struct sdca_function_data *function,
+			    struct sdca_entity *entity,
+			    struct snd_soc_dapm_widget **widget,
+			    struct snd_soc_dapm_route **route)
+{
+	unsigned int target = (1 << SDCA_PDE_PS0) | (1 << SDCA_PDE_PS3);
+	struct sdca_control_range *range;
+	struct sdca_control *control;
+	unsigned int mask = 0;
+	int i;
+
+	control = selector_find_control(dev, entity, SDCA_CTL_PDE_REQUESTED_PS);
Quoted 8 lines

sound/soc/sdca/sdca_asoc.c line 409

Quoted 4 lines
sound/soc/sdca/sdca_asoc.c new file +4 -0
+	if (!control)
+		return -EINVAL;
+
+	/* Power should only be controlled by the driver */

is it though? With the virtual controls you were referring to earlier, it's possible that the path is not used and the PDE in PS3. I guess I am still in the dark on what controls the PDE.

Quoted 3 lines

The PDE is controlled by the DAPM widget created at the bottom of this function and yes power should only be directly controlled by drivers.

ok

Quoted 11 lines
Quoted 10 lines

sound/soc/sdca/sdca_asoc.c line 549

Quoted 6 lines
sound/soc/sdca/sdca_asoc.c new file +6 -0
+	kctl->iface = SNDRV_CTL_ELEM_IFACE_MIXER;
+	kctl->name = "Route";
+	kctl->info = snd_soc_info_enum_double;
+	kctl->get = snd_soc_dapm_get_enum_double;
+	kctl->put = snd_soc_dapm_put_enum_double;
+	kctl->private_value = (unsigned long)soc_enum;

Can you remind me in which cases we have a Selector Unit not controlled by a GE, and what userspace would do with such kcontrols?

Quoted 3 lines

A selector unit not tidied to a GE is just a DAPM mux, it provides an ALSA control which allows user-space to select which input it wants routed through the selector.

I meant in which topologies is this used/needed?

Quoted 7 lines
Quoted 6 lines

sound/soc/sdca/sdca_asoc.c line 612

Quoted 4 lines
sound/soc/sdca/sdca_asoc.c new file +4 -0
+	/* MU control should be through DAPM */
+	if (control->layers != SDCA_ACCESS_LAYER_CLASS)
+		dev_warn(dev, "%s: unexpected access layer: %x\n",
+			 entity->label, control->layers);

isn't this an error? How would DAPM deal with this case?

Quoted 5 lines

In general the control will work the same no matter what access specifier it is given, it still sets the control. So I didn't really see the need to hard error out here. Mostly likely if one gets these its a sign of slightly dodgy DisCo but we should be fine to carry one.

ok, probably best to deal with flaky DSDT at a later time.

Quoted 18 lines
Quoted 9 lines
Quoted 9 lines
sound/soc/sdca/sdca_asoc.c new file +9 -0
+int sdca_asoc_populate_dapm(struct device *dev, struct sdca_function_data *function,
+			    struct snd_soc_dapm_widget *widget,
+			    struct snd_soc_dapm_route *route)
+{
+	int ret;
+	int i;
+
+	/* Some entities need to add controls referenced by other entities */
+	for (i = 0; i < function->num_entities - 1; i++) {
Quoted 8 lines

sound/soc/sdca/sdca_asoc.c line 727

Quoted 5 lines
sound/soc/sdca/sdca_asoc.c new file +5 -0
+		struct sdca_entity *entity = &function->entities[i];
+
+		switch (entity->type) {
+		case SDCA_ENTITY_TYPE_GE:
+			ret = entity_early_parse_ge(dev, function, entity);

what does 'early' mean here? is there a 'late_parse_ge', or is this to say that GEs are parsed first?

Quoted 3 lines

Yeah we need to parse them first such that the GE controls are ready when the SUs are parsed since they will link to the control generated by the GE.

ok, worthy of a comment to make this dependency clearer.

#20 lore raw

Re: [PATCH v5 5/6] ASoC: SDCA: Create ALSA controls from DisCo

In reply to Charles Keepax
Quoted 11 lines · Charles Keepax
Quoted 9 lines · Pierre-Louis Bossart

sound/soc/sdca/sdca_asoc.c static struct sdca_control_range *selector_find_range(struct device *dev, · line 77

Quoted 6 lines
sound/soc/sdca/sdca_asoc.c +5 -0
+static bool exported_control(struct sdca_control *control)
+{
+	/* No need to export control for something that only has one value */
+	if (control->has_fixed)
+		return false;

why is that? It'd simplify userspace if there was always a way to get a value, even if it's fixed?

Quoted 6 lines

Well mostly it is because various things in the stack get sad with a single value control you can't set. If we wanted to change that I would suggest we add a FIXME. It isn't really critical and I would really like to get more of this class driver stuff moving, it has been out for review for a really long time now.

Not following what parts of the stack you are referring to. We will have controls which are sometimes non-existent (DC), fixed (ROM) and some RW. Userspace has to be tolerant of those variations, no?

I am fine with the FIXME and I understand the frustration with limited reviews over a long time.

Quoted 11 lines
Quoted 7 lines

sound/soc/sdca/sdca_asoc.c int sdca_asoc_count_component(struct device *dev, struct sdca_function_data *fun · line 116

Quoted 5 lines
sound/soc/sdca/sdca_asoc.c +1 -0
@@ -99,6 +113,7 @@ int sdca_asoc_count_component(struct device *dev, struct sdca_function_data *fun
 		case SDCA_ENTITY_TYPE_OT:
 			*num_routes += !!entity->iot.clock;
 			*num_routes += !!entity->iot.is_dataport;
+			*num_controls += !entity->iot.is_dataport;

shouldn't the same line be added for TYPE_IT?

This is just the cutting on the patch generation, case
SDCA_ENTITY_TYPE_IT: is the line right above so this runs for both.

ok.

#12 lore raw

Re: [PATCH v5 4/6] ASoC: SDCA: Create DAPM widgets and routes from DisCo

In reply to Charles Keepax
Quoted 27 lines · Charles Keepax
Quoted 2 lines · Pierre-Louis Bossart

On 5/12/25 14:42, Charles Keepax wrote:
It's been a while since I reviewed an earlier versions so now I am confused.

I have attached the output of the dapm-graph tool for the UAJ component for one of our devices, whilst a headset is connected. Hopefully this will help to make it a little more clear what the generated graph from this code looks like. Hopefully the attachment makes it through the ether, let me know if it doesn't and I can send directly.

The PDE supply widgets are what control the PDE power registers, they will power up when they are attached to an active path, ie. a valid link between an active source and an active sink.

For the streaming terminals IT41,OT36 they are active when the audio stream is open. For the non-streaming ones IT31,IT32,IT33,OT43,OT44,OT45 these are active when the pin switch is enabled which it is by default, without pin switches these would be always active. The pin switches are not shown on the output of dapm-graph.

You can see the effect of the GE control, where it has set the input of SU35 to FU33, by the selection of "Headset". And you can see the virtual mixer switches going into MU35, currently both Mixer 1 and 2 are on, so you can see the paths connected.

Quoted 5 lines

But as you can see the correlation to the function diagrams is the spec is pretty good. As I noted with a few minor exceptions, like the GE having more paths, since it must connect to all the widgets it controls separately. And the DAI widgets that are created by the DAPM core.

Thanks for the diagram, very useful indeed.

The only comment I'd have is that all SUs are represented as MUXes in this DAPM graph, but in practice most of them are simple on/off switches. The case of SU35 is one of the exceptions to the rule where it can have multiple inputs. I don't really mind if all SUs are represented as MUX, but it's not a direct mapping from the SDCA spec.

#24 lore raw

Re: [PATCH v5 6/6] ASoC: SDCA: Create DAI drivers from DisCo

In reply to Charles Keepax
Quoted 20 lines · Charles Keepax
Quoted 10 lines · Pierre-Louis Bossart
Quoted 10 lines
sound/soc/sdca/sdca_asoc.c +10 -0
+		switch (entity->type) {
+		case SDCA_ENTITY_TYPE_IT:
+			stream = &dais[j].playback;
+			stream_suffix = "Playback";
+			break;
+		case SDCA_ENTITY_TYPE_OT:
+			stream = &dais[j].capture;
+			stream_suffix = "Capture";
+			break;
+		default:
Quoted 8 lines

sound/soc/sdca/sdca_asoc.c int sdca_asoc_populate_controls(struct device *dev, · line 1197

Quoted 5 lines
sound/soc/sdca/sdca_asoc.c +5 -0
+			continue;
+		}
+
+		if (!entity->iot.is_dataport)
+			continue;

should this be moved as the first test in the for() loop? The switch cases for IT/OT only makes sense for streaming/dataport terminals, no?

Quoted 3 lines

Can't be first in the loop since the entities can be different types and the iot is part of a union so it is only safe to access is_dataport once you know it is a terminal entity.

ok, it's be nice to add a comment for the rest of us who don't remember the low-level details of the data structures :-)

#18 lore raw

Re: [PATCH v5 5/6] ASoC: SDCA: Create ALSA controls from DisCo

In reply to Charles Keepax
Quoted 20 lines · Charles Keepax
Quoted 11 lines · Charles Keepax
Quoted 9 lines · Pierre-Louis Bossart

sound/soc/sdca/sdca_asoc.c static struct sdca_control_range *selector_find_range(struct device *dev, · line 77

Quoted 6 lines
sound/soc/sdca/sdca_asoc.c +5 -0
+static bool exported_control(struct sdca_control *control)
+{
+	/* No need to export control for something that only has one value */
+	if (control->has_fixed)
+		return false;

why is that? It'd simplify userspace if there was always a way to get a value, even if it's fixed?

Quoted 7 lines

Well mostly it is because various things in the stack get sad with a single value control you can't set. If we wanted to change that I would suggest we add a FIXME. It isn't really critical and I would really like to get more of this class driver stuff moving, it has been out for review for a really long time now.

Quoted 8 lines

Thinking a little more about this user-space is going to have to handle controls not being there as different controls will be exported across different parts anyway, based of which options are implemented. Also we have seen a bit of people hiding controls by changing the access layer. Although sometimes that does result in things that are technically not class conformant but they do work through the Microsoft drivers, so we will likely need to support it anyway.

Sorry, I missed what the conclusion was, would this filter stay or be removed?

author maintainer
#8 lore raw

Re: [PATCH v5 4/6] ASoC: SDCA: Create DAPM widgets and routes from DisCo

In reply to Pierre-Louis Bossart
Quoted 10 lines · Pierre-Louis Bossart
Quoted 5 lines

I mean the hard definition here is DAPM routes. But roughly speaking each line on the SDCA function diagram will correspond to a route here, with a few tweaks like the GEs only show a single line going to a red box around a bunch of other entities that will be multiple routes one for each entity in the box.

This could be described in comments. The SDCA spec shows lines of different color for pure data paths and 'control' links. Here we are bringing everything back to the same concept of DAPM routes, but those extra routes only connect to supply widgets.

Yeah I am happy to add extra comments for this.

Quoted 10 lines
Quoted 9 lines
Quoted 5 lines

sound/soc/sdca/sdca_asoc.c line 221

Quoted 1 line
sound/soc/sdca/sdca_asoc.c new file +1 -0
+		values[i + 3] = sdca_range(range, SDCA_SELECTED_MODE_INDEX, i);

Humm, my memory of GEs is that the hardware reports a 'DETECTED_MODE' and then software (kernel and/or userspace) can choose a "SELECTED_MODE". here we only deal with SELECTED_MODE, is this intentional?

This code is creating the ALSA control for the SELECTED_MODE control, the DETECTED_MODE is all handled internally and there is no need to export it.

Quoted 7 lines

Humm, that one doesn't seem right. There could be cases where the DETECTED_MODE is inconclusive, or that additional vendor-specific steps are needed to decide which SELECTED_MODE makes sense. I don't think we should assume everything is handled internally, or we have to define what 'internally' means. This cannot be generic SDCA stuff, it has to be configurable for specific implementations/vendors.

Ok yeah I see, the current code reads the DETECTED_MODE in the IRQ handler and then updates the selected mode. I guess currently what is missing is the handling for the JACK_UNKNOWN case, which could probably require user-space. I think it would be pretty easy to export the DETECTED_MODE as well I will have a look.

Quoted 15 lines
Quoted 10 lines
Quoted 9 lines

sound/soc/sdca/sdca_asoc.c line 344

Quoted 6 lines
sound/soc/sdca/sdca_asoc.c new file +6 -0
+static int entity_pde_event(struct snd_soc_dapm_widget *widget,
+			    struct snd_kcontrol *kctl, int event)
+{
+	struct snd_soc_component *component = widget->dapm->component;
+	struct sdca_entity *entity = widget->priv;
+	static const int poll_us = 10000;

This should be a #define, no?

Quoted 4 lines

I could make it a define if you feel strong but I generally quite like if it just a local delay being using in one place to define it locally, means you don't have to look things up when reading the code.

Quoted 5 lines

I guess my main point was to explain where this 10ms value comes from. I am not sure what to make of it, it's either very large or very small depending on the initial and desired power states. Plus I would expect the sleep times to vary by orders of magnitude depending on the type of function managed.

The trouble is there isn't a great way to parse from the DisCo what a good poll interval would be. I guess we could change to say 1/100th of the maximum, clipped to a minimum time? My concern is the maximums can be quite long so something based on them definitely runs the risk of really under polling and introducing a lot of lag.

Quoted 9 lines
Quoted 7 lines
Quoted 3 lines

Can you remind me in which cases we have a Selector Unit not controlled by a GE, and what userspace would do with such kcontrols?

A selector unit not tidied to a GE is just a DAPM mux, it provides an ALSA control which allows user-space to select which input it wants routed through the selector.

I meant in which topologies is this used/needed?

Its also nice to support topologies above the set ones in the spec, but the current in spec users of these would be:

UAJ:
SU136 - Which lets you mux the second jack capture onto the primary capture path.
SU137 - Which lets you sidetone from the second jack capture.

Admittedly our chips don't currently have those but since I implemented SU support before GE support we had code for this anyway.

Quoted 11 lines
Quoted 9 lines
Quoted 5 lines

sound/soc/sdca/sdca_asoc.c line 727

Quoted 2 lines
sound/soc/sdca/sdca_asoc.c new file +2 -0
+		case SDCA_ENTITY_TYPE_GE:
+			ret = entity_early_parse_ge(dev, function, entity);

what does 'early' mean here? is there a 'late_parse_ge', or is this to say that GEs are parsed first?

Yeah we need to parse them first such that the GE controls are ready when the SUs are parsed since they will link to the control generated by the GE.

ok, worthy of a comment to make this dependency clearer.

Yeah no problem to add a comment for that.

Thanks,
Charles

author maintainer
#13 lore raw

Re: [PATCH v5 4/6] ASoC: SDCA: Create DAPM widgets and routes from DisCo

In reply to Pierre-Louis Bossart
Quoted 9 lines · Pierre-Louis Bossart
Quoted 7 lines · Charles Keepax
Quoted 1 line · Pierre-Louis Bossart

On 5/12/25 14:42, Charles Keepax wrote:

But as you can see the correlation to the function diagrams is the spec is pretty good. As I noted with a few minor exceptions, like the GE having more paths, since it must connect to all the widgets it controls separately. And the DAI widgets that are created by the DAPM core.

Quoted 8 lines

Thanks for the diagram, very useful indeed.

The only comment I'd have is that all SUs are represented as MUXes in this DAPM graph, but in practice most of them are simple on/off switches. The case of SU35 is one of the exceptions to the rule where it can have multiple inputs. I don't really mind if all SUs are represented as MUX, but it's not a direct mapping from the SDCA spec.

I mean I think we could special case 1 input versions as switches but a) I am slightly concerned it might make sharing them across a GE harder and b) I am not sure I see what it really buys us. Generally I prefer the code to not require special casing for certain situations and there shouldn't be any difference in the behaviour from switches or muxes.

Thanks,
Charles

author maintainer
#19 lore raw

Re: [PATCH v5 5/6] ASoC: SDCA: Create ALSA controls from DisCo

In reply to Pierre-Louis Bossart
Quoted 30 lines · Pierre-Louis Bossart
Quoted 20 lines · Charles Keepax
Quoted 11 lines · Charles Keepax
Quoted 9 lines · Pierre-Louis Bossart

sound/soc/sdca/sdca_asoc.c static struct sdca_control_range *selector_find_range(struct device *dev, · line 77

Quoted 6 lines
sound/soc/sdca/sdca_asoc.c +5 -0
+static bool exported_control(struct sdca_control *control)
+{
+	/* No need to export control for something that only has one value */
+	if (control->has_fixed)
+		return false;

why is that? It'd simplify userspace if there was always a way to get a value, even if it's fixed?

Quoted 7 lines

Well mostly it is because various things in the stack get sad with a single value control you can't set. If we wanted to change that I would suggest we add a FIXME. It isn't really critical and I would really like to get more of this class driver stuff moving, it has been out for review for a really long time now.

Quoted 8 lines

Thinking a little more about this user-space is going to have to handle controls not being there as different controls will be exported across different parts anyway, based of which options are implemented. Also we have seen a bit of people hiding controls by changing the access layer. Although sometimes that does result in things that are technically not class conformant but they do work through the Microsoft drivers, so we will likely need to support it anyway.

Quoted 1 line

Sorry, I missed what the conclusion was, would this filter stay or be removed?

Not sure I was clear either :-)

I have been poking this a little more today and I think I might have overstated some of the problems here. I think it might be ok to just mark the controls are read only. Will keep poking it and will update soon, but kinda hopeful we could actually export these in the next rev.

Thanks,
Charles

author maintainer
#25 lore raw

Re: [PATCH v5 6/6] ASoC: SDCA: Create DAI drivers from DisCo

In reply to Pierre-Louis Bossart
Quoted 21 lines · Pierre-Louis Bossart
Quoted 20 lines · Charles Keepax
Quoted 18 lines · Pierre-Louis Bossart
Quoted 5 lines
sound/soc/sdca/sdca_asoc.c +5 -0
+		switch (entity->type) {
+		case SDCA_ENTITY_TYPE_IT:
+			stream = &dais[j].playback;
+			stream_suffix = "Playback";
+			break;

sound/soc/sdca/sdca_asoc.c int sdca_asoc_populate_controls(struct device *dev, · line 1197

Quoted 10 lines
sound/soc/sdca/sdca_asoc.c +10 -0
+		case SDCA_ENTITY_TYPE_OT:
+			stream = &dais[j].capture;
+			stream_suffix = "Capture";
+			break;
+		default:
+			continue;
+		}
+
+		if (!entity->iot.is_dataport)
+			continue;

should this be moved as the first test in the for() loop? The switch cases for IT/OT only makes sense for streaming/dataport terminals, no?

Quoted 6 lines
Quoted 3 lines

Can't be first in the loop since the entities can be different types and the iot is part of a union so it is only safe to access is_dataport once you know it is a terminal entity.

ok, it's be nice to add a comment for the rest of us who don't remember the low-level details of the data structures :-)

Yeah never mind putting a comment in, I will add one for the next spin.

Thanks,
Charles

author maintainer
#9 lore raw

Re: [PATCH v5 4/6] ASoC: SDCA: Create DAPM widgets and routes from DisCo

In reply to Charles Keepax
Quoted 9 lines · Charles Keepax
Quoted 7 lines · Pierre-Louis Bossart

Humm, that one doesn't seem right. There could be cases where the DETECTED_MODE is inconclusive, or that additional vendor-specific steps are needed to decide which SELECTED_MODE makes sense. I don't think we should assume everything is handled internally, or we have to define what 'internally' means. This cannot be generic SDCA stuff, it has to be configurable for specific implementations/vendors.

Quoted 6 lines

Ok yeah I see, the current code reads the DETECTED_MODE in the IRQ handler and then updates the selected mode. I guess currently what is missing is the handling for the JACK_UNKNOWN case, which could probably require user-space. I think it would be pretty easy to export the DETECTED_MODE as well I will have a look.

Should hopefully be able to send a new spin tomorrow. I have managed to include all the comments I think.

I started being a little more unsure on this DETECTED_MODE again, but I am going to go with forcing the export of DETECTED_MODE for now and we can revise later if we have a better plan.

My concerns really are most custom handling will actually want to live in the kernel, except the case where the system wants to present a dialog to the user to select the jack device manually. In that manual case I also can't see any reason why just setting the selected mode to unknown would not make the most sense, but the spec does not say the device should do that.

Thanks,
Charles

#10 lore raw

Re: [PATCH v5 4/6] ASoC: SDCA: Create DAPM widgets and routes from DisCo

Quoted 24 lines · Charles Keepax
Quoted 9 lines · Charles Keepax
Quoted 7 lines · Pierre-Louis Bossart

Humm, that one doesn't seem right. There could be cases where the DETECTED_MODE is inconclusive, or that additional vendor-specific steps are needed to decide which SELECTED_MODE makes sense. I don't think we should assume everything is handled internally, or we have to define what 'internally' means. This cannot be generic SDCA stuff, it has to be configurable for specific implementations/vendors.

Quoted 6 lines

Ok yeah I see, the current code reads the DETECTED_MODE in the IRQ handler and then updates the selected mode. I guess currently what is missing is the handling for the JACK_UNKNOWN case, which could probably require user-space. I think it would be pretty easy to export the DETECTED_MODE as well I will have a look.

Should hopefully be able to send a new spin tomorrow. I have managed to include all the comments I think.

I started being a little more unsure on this DETECTED_MODE again, but I am going to go with forcing the export of DETECTED_MODE for now and we can revise later if we have a better plan.

Quoted 6 lines

My concerns really are most custom handling will actually want to live in the kernel, except the case where the system wants to present a dialog to the user to select the jack device manually. In that manual case I also can't see any reason why just setting the selected mode to unknown would not make the most sense, but the spec does not say the device should do that.

I am fine with a TODO on the DETECTED/SELECTED mode handling for now. I do recall discussions in the Audio Subgroup for the Retaskable Jack where there are just too many options and a basic driver would probably give up and say "undefined" for the DETECTED_MODE. For the UAJ mirroring the DETECTED_MODE and/or asking the user for input are probably the two most likely options.