ASoC: rt5575: Add the codec driver for the ALC5575

PATCH Superseded Oder Chiou 4 participants 6 messages Nov 21, 2025 – Jan 6

Subsystems (4)

Device tree bindings Sound ASoC Realtek audio codecs

lore

Share & tools

lore

Oder Chiou

author maintainer

2025-11-21 08:40 UTC #0 lore raw

PATCH v7 1/2 ASoC: rt5575: Add the codec driver for the ALC5575

The ALC5575 integrates an audio DSP that typically loads its firmware from an external flash via its own SPI host interface. In certain hardware configurations, the firmware can alternatively be loaded through the SPI client interface. The driver provides basic mute and volume control functions. When the SPI client interface is enabled, firmware loading is handled by the SPI driver.

Signed-off-by Oder Chiou <[email protected]>

---
 sound/soc/codecs/Kconfig      |  10 +
 sound/soc/codecs/Makefile     |   4 +
 sound/soc/codecs/rt5575-spi.c |  86 ++++++++
 sound/soc/codecs/rt5575-spi.h |  16 ++
 sound/soc/codecs/rt5575.c     | 370 ++++++++++++++++++++++++++++++++++
 sound/soc/codecs/rt5575.h     |  54 +++++
 6 files changed, 540 insertions(+)
 create mode 100644 sound/soc/codecs/rt5575-spi.c
 create mode 100644 sound/soc/codecs/rt5575-spi.h
 create mode 100644 sound/soc/codecs/rt5575.c
 create mode 100644 sound/soc/codecs/rt5575.h

sound/soc/codecs/Kconfig +10 -0

index a0dfef57200c..a3ea5febd1e0 100644
@@ -211,6 +211,7 @@ config SND_SOC_ALL_CODECS
 	imply SND_SOC_RT1305
 	imply SND_SOC_RT1308
 	imply SND_SOC_RT5514
+	imply SND_SOC_RT5575
 	imply SND_SOC_RT5616
 	imply SND_SOC_RT5631
 	imply SND_SOC_RT5640

@@ -1767,6 +1768,15 @@ config SND_SOC_RT5514_SPI_BUILTIN
 	bool # force RT5514_SPI to be built-in to avoid link errors
 	default SND_SOC_RT5514=y && SND_SOC_RT5514_SPI=m
 
+config SND_SOC_RT5575
+	tristate "Realtek ALC5575 Codec - I2C"
+	depends on I2C
+
+config SND_SOC_RT5575_SPI
+	tristate "Realtek ALC5575 Codec - SPI"
+	depends on SPI_MASTER
+	select SND_SOC_RT5575
+
 config SND_SOC_RT5616
 	tristate "Realtek RT5616 CODEC"
 	depends on I2C

sound/soc/codecs/Makefile +4 -0

index 39138d96a720..82f660cbe8ec 100644
@@ -252,6 +252,8 @@ snd-soc-rt286-y := rt286.o
 snd-soc-rt298-y := rt298.o
 snd-soc-rt5514-y := rt5514.o
 snd-soc-rt5514-spi-y := rt5514-spi.o
+snd-soc-rt5575-y := rt5575.o
+snd-soc-rt5575-spi-y := rt5575-spi.o
 snd-soc-rt5616-y := rt5616.o
 snd-soc-rt5631-y := rt5631.o
 snd-soc-rt5640-y := rt5640.o

@@ -684,6 +686,8 @@ obj-$(CONFIG_SND_SOC_RT298)	+= snd-soc-rt298.o
 obj-$(CONFIG_SND_SOC_RT5514)	+= snd-soc-rt5514.o
 obj-$(CONFIG_SND_SOC_RT5514_SPI)	+= snd-soc-rt5514-spi.o
 obj-$(CONFIG_SND_SOC_RT5514_SPI_BUILTIN)	+= snd-soc-rt5514-spi.o
+obj-$(CONFIG_SND_SOC_RT5575)	+= snd-soc-rt5575.o
+obj-$(CONFIG_SND_SOC_RT5575_SPI)	+= snd-soc-rt5575-spi.o
 obj-$(CONFIG_SND_SOC_RT5616)	+= snd-soc-rt5616.o
 obj-$(CONFIG_SND_SOC_RT5631)	+= snd-soc-rt5631.o
 obj-$(CONFIG_SND_SOC_RT5640)	+= snd-soc-rt5640.o

sound/soc/codecs/rt5575-spi.c new file +86 -0

86 lines not on this page · raw message

sound/soc/codecs/rt5575-spi.h new file +16 -0

new file mode 100644
index 000000000000..cafe49a7ecc2
@@ -0,0 +1,16 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * rt5575-spi.h  --  ALC5575 SPI driver
+ *
+ * Copyright(c) 2025 Realtek Semiconductor Corp.
+ *
+ */
+
+#ifndef __RT5575_SPI_H__
+#define __RT5575_SPI_H__
+
+extern bool rt5575_spi_ready;
+
+int rt5575_spi_burst_write(u32 addr, const u8 *txbuf, size_t len);
+
+#endif /* __RT5575_SPI_H__ */

sound/soc/codecs/rt5575.c new file +370 -0

370 lines not on this page · raw message

sound/soc/codecs/rt5575.h new file +54 -0

54 lines not on this page · raw message

Signature · 1 line

2.51.2

Oder Chiou

author maintainer

2025-11-21 08:40 UTC #1 lore raw

PATCH v7 2/2 ASoC: dt-bindings: realtek,rt5575: add bindings for ALC5575

Realtek ALC5575 is a highly advanced DSP and microphone CODEC that has been designed for AI audio technology. Its impressive features include an advanced HiFi-5 DSP core, a Neural Network Processing Unit (NPU) owned by Realtek, and embedded 4MB memory, which enables it to operate highly advanced AI audio algorithms. The ALC5575 supports 4xA-mic input and 8xD-mic input, as well as a rich set of interfaces such as I2S, I2C, SPI, and UART.

Signed-off-by Oder Chiou <[email protected]>

---
 .../bindings/sound/realtek,rt5575.yaml        | 42 +++++++++++++++++++
 1 file changed, 42 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/sound/realtek,rt5575.yaml

Documentation/devicetree/bindings/sound/realtek,rt5575.yaml new file +42 -0

new file mode 100644
index 000000000000..cc7540278ad9
@@ -0,0 +1,42 @@
+# SPDX-License-Identifier: GPL-2.0-only OR BSD-2-Clause
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/sound/realtek,rt5575.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: ALC5575 audio CODEC
+
+maintainers:
+  - Oder Chiou <[email protected]>
+
+description:
+  The device supports both I2C and SPI. I2C is mandatory, while SPI is
+  optional depending on the hardware configuration.
+
+allOf:
+  - $ref: /schemas/spi/spi-peripheral-props.yaml#
+  - $ref: dai-common.yaml#
+
+properties:
+  compatible:
+    const: realtek,rt5575
+
+  reg:
+    maxItems: 1
+
+required:
+  - compatible
+  - reg
+
+unevaluatedProperties: false
+
+examples:
+  - |
+    i2c {
+        #address-cells = <1>;
+        #size-cells = <0>;
+        codec@57 {
+            compatible = "realtek,rt5575";
+            reg = <0x57>;
+        };
+    };

Signature · 1 line

2.51.2

Krzysztof Kozlowski

2025-11-22 10:55 UTC #2 lore raw

Re: [PATCH v7 2/2] ASoC: dt-bindings: realtek,rt5575: add bindings for ALC5575

Quoted 10 lines · Oder Chiou

Realtek ALC5575 is a highly advanced DSP and microphone CODEC that has been designed for AI audio technology. Its impressive features include an advanced HiFi-5 DSP core, a Neural Network Processing Unit (NPU) owned by Realtek, and embedded 4MB memory, which enables it to operate highly advanced AI audio algorithms. The ALC5575 supports 4xA-mic input and 8xD-mic input, as well as a rich set of interfaces such as I2S, I2C, SPI, and UART.
Signed-off-by Oder Chiou <[email protected]>
---

You still did not respond to my comments from v5. You did not implemented some of them and I already asked for improvements.

I will not review this and I mark it as changes requested in patchwork.

Best regards,
Krzysztof

Cezary Rojewski

2025-11-24 10:45 UTC #3 lore raw

In reply to Oder Chiou

Quoted 8 lines · Oder Chiou

The ALC5575 integrates an audio DSP that typically loads its firmware from an external flash via its own SPI host interface. In certain hardware configurations, the firmware can alternatively be loaded through the SPI client interface. The driver provides basic mute and volume control functions. When the SPI client interface is enabled, firmware loading is handled by the SPI driver.
Signed-off-by Oder Chiou <[email protected]>

Please CC the reviewers who actively review your patches, makes it easier not miss the follow ups.

Not talking about the tag-area, just the email Cc :)

Quoted 83 lines

---
  sound/soc/codecs/Kconfig      |  10 +
  sound/soc/codecs/Makefile     |   4 +
  sound/soc/codecs/rt5575-spi.c |  86 ++++++++
  sound/soc/codecs/rt5575-spi.h |  16 ++
  sound/soc/codecs/rt5575.c     | 370 ++++++++++++++++++++++++++++++++++
  sound/soc/codecs/rt5575.h     |  54 +++++
  6 files changed, 540 insertions(+)

create mode 100644 sound/soc/codecs/rt5575-spi.c create mode 100644 sound/soc/codecs/rt5575-spi.h create mode 100644 sound/soc/codecs/rt5575.c create mode 100644 sound/soc/codecs/rt5575.h

sound/soc/codecs/Kconfig +10 -0

index a0dfef57200c..a3ea5febd1e0 100644
@@ -211,6 +211,7 @@ config SND_SOC_ALL_CODECS
  	imply SND_SOC_RT1305
  	imply SND_SOC_RT1308
  	imply SND_SOC_RT5514
+	imply SND_SOC_RT5575
  	imply SND_SOC_RT5616
  	imply SND_SOC_RT5631
  	imply SND_SOC_RT5640

@@ -1767,6 +1768,15 @@ config SND_SOC_RT5514_SPI_BUILTIN
  	bool # force RT5514_SPI to be built-in to avoid link errors
  	default SND_SOC_RT5514=y && SND_SOC_RT5514_SPI=m
 
+config SND_SOC_RT5575
+	tristate "Realtek ALC5575 Codec - I2C"
+	depends on I2C
+
+config SND_SOC_RT5575_SPI
+	tristate "Realtek ALC5575 Codec - SPI"
+	depends on SPI_MASTER
+	select SND_SOC_RT5575
+
  config SND_SOC_RT5616
  	tristate "Realtek RT5616 CODEC"
  	depends on I2C

sound/soc/codecs/Makefile +4 -0

index 39138d96a720..82f660cbe8ec 100644
@@ -252,6 +252,8 @@ snd-soc-rt286-y := rt286.o
  snd-soc-rt298-y := rt298.o
  snd-soc-rt5514-y := rt5514.o
  snd-soc-rt5514-spi-y := rt5514-spi.o
+snd-soc-rt5575-y := rt5575.o
+snd-soc-rt5575-spi-y := rt5575-spi.o
  snd-soc-rt5616-y := rt5616.o
  snd-soc-rt5631-y := rt5631.o
  snd-soc-rt5640-y := rt5640.o

@@ -684,6 +686,8 @@ obj-$(CONFIG_SND_SOC_RT298)	+= snd-soc-rt298.o
  obj-$(CONFIG_SND_SOC_RT5514)	+= snd-soc-rt5514.o
  obj-$(CONFIG_SND_SOC_RT5514_SPI)	+= snd-soc-rt5514-spi.o
  obj-$(CONFIG_SND_SOC_RT5514_SPI_BUILTIN)	+= snd-soc-rt5514-spi.o
+obj-$(CONFIG_SND_SOC_RT5575)	+= snd-soc-rt5575.o
+obj-$(CONFIG_SND_SOC_RT5575_SPI)	+= snd-soc-rt5575-spi.o
  obj-$(CONFIG_SND_SOC_RT5616)	+= snd-soc-rt5616.o
  obj-$(CONFIG_SND_SOC_RT5631)	+= snd-soc-rt5631.o
  obj-$(CONFIG_SND_SOC_RT5640)	+= snd-soc-rt5640.o

sound/soc/codecs/rt5575-spi.c new file +14 -0

new file mode 100644
index 000000000000..9dbc8170fb76
@@ -0,0 +1,86 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * rt5575-spi.c  --  ALC5575 SPI driver
+ *
+ * Copyright(c) 2025 Realtek Semiconductor Corp.
+ *
+ */
+
+#include <linux/of.h>
+#include <linux/spi/spi.h>
+
+#include "rt5575-spi.h"
+
+#define RT5575_SPI_CMD_BURST_WRITE	5

sound/soc/codecs/rt5575-spi.c line 24

Quoted 10 lines

sound/soc/codecs/rt5575-spi.c new file +10 -0

+#define RT5575_SPI_BUF_LEN		240
+
+struct rt5575_spi_burst_write {
+	u8 cmd;
+	u32 addr;
+	u8 data[RT5575_SPI_BUF_LEN];
+	u8 dummy;
+} __packed;
+
+bool rt5575_spi_ready;

I believe the patch is dead on arrival, unfortunately as it won't compile as =m (module) when CONFIG_SND_SOC_RT5575_SPI is enabled as =m simultaneously. EXPORT_SYMBOL_GPL() or a friend of his is missing.

Quoted 30 lines

sound/soc/codecs/rt5575-spi.c new file +30 -0

+static struct spi_device *rt5575_spi;
+
+/**
+ * rt5575_spi_burst_write - Write data to SPI by rt5575 address.
+ * @addr: Start address.
+ * @txbuf: Data buffer for writing.
+ * @len: Data length.
+ *
+ */
+int rt5575_spi_burst_write(u32 addr, const u8 *txbuf, size_t len)
+{
+	struct rt5575_spi_burst_write buf = {
+		.cmd = RT5575_SPI_CMD_BURST_WRITE
+	};
+	unsigned int end, offset = 0;
+
+	while (offset < len) {
+		if (offset + RT5575_SPI_BUF_LEN <= len)
+			end = RT5575_SPI_BUF_LEN;
+		else
+			end = len % RT5575_SPI_BUF_LEN;
+
+		buf.addr = cpu_to_le32(addr + offset);
+
+		memcpy(&buf.data, &txbuf[offset], end);
+
+		spi_write(rt5575_spi, &buf, sizeof(buf));
+
+		offset += RT5575_SPI_BUF_LEN;
+	}

sound/soc/codecs/rt5575-spi.c line 64

Quoted 10 lines

sound/soc/codecs/rt5575-spi.c new file +10 -0

+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(rt5575_spi_burst_write);
+
+static int rt5575_spi_probe(struct spi_device *spi)
+{
+	rt5575_spi = spi;
+
+	rt5575_spi_ready = true;

I still do not understand the logic behind having rt5575_spi _and_ rt5575_spi_ready. The latter is used just one, in the I2C-device probe. rt5575_spi is static and will be by default initialized to NULL. Statement 'if (!rt5575_spi)' is enough to cover that single occurrence in your code.

Quoted 3 lines

sound/soc/codecs/rt5575-spi.c new file +3 -0
+
+	return 0;
+}

...

Quoted 5 lines

sound/soc/codecs/rt5575.c new file +0 -0
new file mode 100644
index 000000000000..58508d643273

...

sound/soc/codecs/rt5575.c line 159

Quoted 11 lines

sound/soc/codecs/rt5575.c new file +11 -0

+static int rt5575_probe(struct snd_soc_component *component)
+{
+	struct rt5575_priv *rt5575 = snd_soc_component_get_drvdata(component);
+
+	rt5575->component = component;
+
+	dev_info(component->dev, "Private ID: %llx\n", rt5575_get_priv_id(rt5575));
+
+#if IS_ENABLED(CONFIG_SND_SOC_RT5575_SPI)
+	request_firmware_nowait(THIS_MODULE, FW_ACTION_UEVENT, "realtek/rt5575/rt5575_fw1.bin",
+		component->dev, GFP_KERNEL, rt5575, rt5575_fw_load_by_spi);

I do not believe this is OK. If the SPI component exists, from the code-perspective it looks like the firmware-loading procedure is essential and failing to load the firmware should halt driver's operations.

See below as this touches on the driver design which looks off and I haven't found any valid argumentation in your previous message that would clarify the situation.

Quoted 4 lines

sound/soc/codecs/rt5575.c new file +4 -0
+#endif
+
+	return 0;
+}

...

Quoted 5 lines

sound/soc/codecs/rt5575.c new file +5 -0

+static const struct i2c_device_id rt5575_i2c_id[] = {
+	{ "rt5575" },
+	{ }
+};
+MODULE_DEVICE_TABLE(i2c, rt5575_i2c_id);

sound/soc/codecs/rt5575.c line 310

Quoted 10 lines

sound/soc/codecs/rt5575.c new file +10 -0

+
+static int rt5575_i2c_probe(struct i2c_client *i2c)
+{
+	struct rt5575_priv *rt5575;
+	int ret, val;
+
+#if IS_ENABLED(CONFIG_SND_SOC_RT5575_SPI)
+	if (!rt5575_spi_ready)
+		return -EPROBE_DEFER;
+#endif

I finally understood what's off in the design, and this part is crucial to the subject.

If CONFIG_SND_SOC_RT5575_SPI is enabled, and the SPI kept failing for whatever reason, the I2C device probe would get deferred continuously. As the kconfig looks optional, the question is: Does rt5575 codec exist in a "without SPI flash" version or, is it always shipped with SPI flash?

If it does exist as a standalone component - without SPI - then enabling CONFIG_SND_SOC_RT5575_SPI effectively renders the driver dysfunctional.

Quoted 60 lines

sound/soc/codecs/rt5575.c new file +60 -0

+
+	rt5575 = devm_kzalloc(&i2c->dev, sizeof(struct rt5575_priv),
+				GFP_KERNEL);
+	if (rt5575 == NULL)
+		return -ENOMEM;
+
+	i2c_set_clientdata(i2c, rt5575);
+
+	rt5575->i2c = i2c;
+
+	rt5575->dsp_regmap = devm_regmap_init_i2c(i2c, &rt5575_dsp_regmap);
+	if (IS_ERR(rt5575->dsp_regmap)) {
+		ret = PTR_ERR(rt5575->dsp_regmap);
+		dev_err(&i2c->dev, "Failed to allocate register map: %d\n", ret);
+		return ret;
+	}
+
+	rt5575->regmap = devm_regmap_init(&i2c->dev, NULL, i2c, &rt5575_regmap);
+	if (IS_ERR(rt5575->regmap)) {
+		ret = PTR_ERR(rt5575->regmap);
+		dev_err(&i2c->dev, "Failed to allocate register map: %d\n", ret);
+		return ret;
+	}
+
+	regmap_read(rt5575->regmap, RT5575_ID, &val);
+	if (val != RT5575_DEVICE_ID) {
+		dev_err(&i2c->dev, "Device with ID register %08x is not rt5575\n", val);
+		return -ENODEV;
+	}
+
+	regmap_read(rt5575->regmap, RT5575_ID_1, &val);
+	if (!val) {
+		dev_err(&i2c->dev, "This is not formal version\n");
+		return -ENODEV;
+	}
+
+	return devm_snd_soc_register_component(&i2c->dev, &rt5575_soc_component_dev, rt5575_dai,
+		ARRAY_SIZE(rt5575_dai));
+}
+
+static const struct of_device_id rt5575_of_match[] = {
+	{ .compatible = "realtek,rt5575" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, rt5575_of_match);
+
+static struct i2c_driver rt5575_i2c_driver = {
+	.driver = {
+		.name = "rt5575",
+		.owner = THIS_MODULE,
+		.of_match_table = of_match_ptr(rt5575_of_match),
+	},
+	.probe = rt5575_i2c_probe,
+	.id_table = rt5575_i2c_id,
+};
+module_i2c_driver(rt5575_i2c_driver);
+
+MODULE_DESCRIPTION("ASoC ALC5575 driver");
+MODULE_AUTHOR("Oder Chiou <[email protected]>");
+MODULE_LICENSE("GPL");

Oder Chiou

author maintainer

2025-11-26 11:09 UTC #4 lore raw

Quoted 175 lines

-----Original Message-----
From: Cezary Rojewski <[email protected]>
Sent: Monday, November 24, 2025 6:45 PM
To: Oder Chiou <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; Flove(HsinFu) <[email protected]>; Shuming [范書銘] <[email protected]>; Jack Yu <[email protected]>; Derek [方德義] <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]
Subject: Re: [PATCH v7 1/2] ASoC: rt5575: Add the codec driver for the ALC5575
External mail : This email originated from outside the organization. Do not reply, click links, or open attachments unless you recognize the sender and know the content is safe.
Quoted 8 lines · Oder Chiou
The ALC5575 integrates an audio DSP that typically loads its firmware from an external flash via its own SPI host interface. In certain hardware configurations, the firmware can alternatively be loaded through the SPI client interface. The driver provides basic mute and volume control functions. When the SPI client interface is enabled, firmware loading is handled by the SPI driver.
Signed-off-by Oder Chiou <[email protected]>
Please CC the reviewers who actively review your patches, makes it easier not miss the follow ups.
Not talking about the tag-area, just the email Cc :)
Quoted 6 lines
---
  sound/soc/codecs/Kconfig      |  10 +
  sound/soc/codecs/Makefile     |   4 +
  sound/soc/codecs/rt5575-spi.c |  86 ++++++++
  sound/soc/codecs/rt5575-spi.h |  16 ++
  sound/soc/codecs/rt5575.c     | 370
++++++++++++++++++++++++++++++++++
sound/soc/codecs/Makefile obj-$(CONFIG_SND_SOC_RT298) +=
Quoted 49 lines
  sound/soc/codecs/rt5575.h     |  54 +++++
  6 files changed, 540 insertions(+)
create mode 100644 sound/soc/codecs/rt5575-spi.c create mode 100644 sound/soc/codecs/rt5575-spi.h create mode 100644 sound/soc/codecs/rt5575.c create mode 100644 sound/soc/codecs/rt5575.h
sound/soc/codecs/Kconfig +10 -0
index a0dfef57200c..a3ea5febd1e0 100644
@@ -211,6 +211,7 @@ config SND_SOC_ALL_CODECS
      imply SND_SOC_RT1305
      imply SND_SOC_RT1308
      imply SND_SOC_RT5514
+     imply SND_SOC_RT5575
      imply SND_SOC_RT5616
      imply SND_SOC_RT5631
      imply SND_SOC_RT5640
@@ -1767,6 +1768,15 @@ config SND_SOC_RT5514_SPI_BUILTIN
      bool # force RT5514_SPI to be built-in to avoid link errors
      default SND_SOC_RT5514=y && SND_SOC_RT5514_SPI=m
 
+config SND_SOC_RT5575
+     tristate "Realtek ALC5575 Codec - I2C"
+     depends on I2C
+
+config SND_SOC_RT5575_SPI
+     tristate "Realtek ALC5575 Codec - SPI"
+     depends on SPI_MASTER
+     select SND_SOC_RT5575
+
  config SND_SOC_RT5616
      tristate "Realtek RT5616 CODEC"
      depends on I2C
sound/soc/codecs/Makefile +2 -0
index 39138d96a720..82f660cbe8ec 100644
@@ -252,6 +252,8 @@ snd-soc-rt286-y := rt286.o
  snd-soc-rt298-y := rt298.o
  snd-soc-rt5514-y := rt5514.o
  snd-soc-rt5514-spi-y := rt5514-spi.o
+snd-soc-rt5575-y := rt5575.o
+snd-soc-rt5575-spi-y := rt5575-spi.o
  snd-soc-rt5616-y := rt5616.o
  snd-soc-rt5631-y := rt5631.o
  snd-soc-rt5640-y := rt5640.o
@@ -684,6 +686,8 @@ obj-$(CONFIG_SND_SOC_RT298)       +=
snd-soc-rt298.o
sound/soc/codecs/Makefile obj-$(CONFIG_SND_SOC_RT298) += snd-soc-rt298.o
Quoted 3 lines
sound/soc/codecs/Makefile +0 -0
  obj-$(CONFIG_SND_SOC_RT5514)        += snd-soc-rt5514.o
  obj-$(CONFIG_SND_SOC_RT5514_SPI)    += snd-soc-rt5514-spi.o
  obj-$(CONFIG_SND_SOC_RT5514_SPI_BUILTIN)    +=
snd-soc-rt5514-spi.o
Quoted 25 lines
sound/soc/codecs/Makefile +2 -0
+obj-$(CONFIG_SND_SOC_RT5575) += snd-soc-rt5575.o
+obj-$(CONFIG_SND_SOC_RT5575_SPI)     += snd-soc-rt5575-spi.o
  obj-$(CONFIG_SND_SOC_RT5616)        += snd-soc-rt5616.o
  obj-$(CONFIG_SND_SOC_RT5631)        += snd-soc-rt5631.o
  obj-$(CONFIG_SND_SOC_RT5640)        += snd-soc-rt5640.o
sound/soc/codecs/rt5575-spi.c new file +14 -0
new file mode 100644
index 000000000000..9dbc8170fb76
@@ -0,0 +1,86 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * rt5575-spi.c  --  ALC5575 SPI driver
+ *
+ * Copyright(c) 2025 Realtek Semiconductor Corp.
+ *
+ */
+
+#include <linux/of.h>
+#include <linux/spi/spi.h>
+
+#include "rt5575-spi.h"
+
+#define RT5575_SPI_CMD_BURST_WRITE   5
sound/soc/codecs/rt5575-spi.c line 24
Quoted 10 lines
sound/soc/codecs/rt5575-spi.c new file +10 -0
+#define RT5575_SPI_BUF_LEN           240
+
+struct rt5575_spi_burst_write {
+     u8 cmd;
+     u32 addr;
+     u8 data[RT5575_SPI_BUF_LEN];
+     u8 dummy;
+} __packed;
+
+bool rt5575_spi_ready;
I believe the patch is dead on arrival, unfortunately as it won't compile as =m (module) when CONFIG_SND_SOC_RT5575_SPI is enabled as =m
simultaneously. EXPORT_SYMBOL_GPL() or a friend of his is missing.
Quoted 39 lines
sound/soc/codecs/rt5575-spi.c new file +39 -0
+static struct spi_device *rt5575_spi;
+
+/**
+ * rt5575_spi_burst_write - Write data to SPI by rt5575 address.
+ * @addr: Start address.
+ * @txbuf: Data buffer for writing.
+ * @len: Data length.
+ *
+ */
+int rt5575_spi_burst_write(u32 addr, const u8 *txbuf, size_t len)
+{
+     struct rt5575_spi_burst_write buf = {
+             .cmd = RT5575_SPI_CMD_BURST_WRITE
+     };
+     unsigned int end, offset = 0;
+
+     while (offset < len) {
+             if (offset + RT5575_SPI_BUF_LEN <= len)
+                     end = RT5575_SPI_BUF_LEN;
+             else
+                     end = len % RT5575_SPI_BUF_LEN;
+
+             buf.addr = cpu_to_le32(addr + offset);
+
+             memcpy(&buf.data, &txbuf[offset], end);
+
+             spi_write(rt5575_spi, &buf, sizeof(buf));
+
+             offset += RT5575_SPI_BUF_LEN;
+     }
+
+     return 0;
+}
+EXPORT_SYMBOL_GPL(rt5575_spi_burst_write);
+
+static int rt5575_spi_probe(struct spi_device *spi)
+{
+     rt5575_spi = spi;
+

Quoted 8 lines

sound/soc/codecs/rt5575-spi.c line 64
Quoted 1 line
sound/soc/codecs/rt5575-spi.c new file +1 -0
+     rt5575_spi_ready = true;
I still do not understand the logic behind having rt5575_spi _and_ rt5575_spi_ready. The latter is used just one, in the I2C-device probe. rt5575_spi is static and will be by default initialized to NULL. Statement 'if (!rt5575_spi)' is enough to cover that single occurrence in your code.

The variable rt5575_spi_ready will be removed.
Declare the rt5575_spi as extern variable, adds it in EXPORT_SYMBOL_GPL(), and uses the statement 'if (!rt5575_spi)' in the i2c_probe.

Quoted 65 lines

Quoted 4 lines
sound/soc/codecs/rt5575-spi.c new file +3 -0
+
+     return 0;
+}
...
Quoted 5 lines
sound/soc/codecs/rt5575.c new file +0 -0
new file mode 100644
index 000000000000..58508d643273
...
sound/soc/codecs/rt5575.c
Quoted 3 lines
sound/soc/codecs/rt5575.c new file +3 -0
+static int rt5575_probe(struct snd_soc_component *component)
+{
+     struct rt5575_priv *rt5575 =
snd_soc_component_get_drvdata(component);
sound/soc/codecs/rt5575.c
Quoted 4 lines
sound/soc/codecs/rt5575.c new file +4 -0
+
+     rt5575->component = component;
+
+     dev_info(component->dev, "Private ID: %llx\n",
rt5575_get_priv_id(rt5575));
sound/soc/codecs/rt5575.c
Quoted 3 lines
sound/soc/codecs/rt5575.c new file +3 -0
+
+#if IS_ENABLED(CONFIG_SND_SOC_RT5575_SPI)
+     request_firmware_nowait(THIS_MODULE, FW_ACTION_UEVENT,
"realtek/rt5575/rt5575_fw1.bin",
Quoted 1 line
+ component->dev, GFP_KERNEL, rt5575,
rt5575_fw_load_by_spi);
I do not believe this is OK. If the SPI component exists, from the code-perspective it looks like the firmware-loading procedure is essential and failing to load the firmware should halt driver's operations.
See below as this touches on the driver design which looks off and I haven't found any valid argumentation in your previous message that would clarify the situation.
Quoted 4 lines
sound/soc/codecs/rt5575.c new file +4 -0
+#endif
+
+     return 0;
+}
...
Quoted 5 lines
sound/soc/codecs/rt5575.c new file +5 -0
+static const struct i2c_device_id rt5575_i2c_id[] = {
+     { "rt5575" },
+     { }
+};
+MODULE_DEVICE_TABLE(i2c, rt5575_i2c_id);
sound/soc/codecs/rt5575.c line 310
Quoted 10 lines
sound/soc/codecs/rt5575.c new file +10 -0
+
+static int rt5575_i2c_probe(struct i2c_client *i2c)
+{
+     struct rt5575_priv *rt5575;
+     int ret, val;
+
+#if IS_ENABLED(CONFIG_SND_SOC_RT5575_SPI)
+     if (!rt5575_spi_ready)
+             return -EPROBE_DEFER;
+#endif
I finally understood what's off in the design, and this part is crucial to the subject.

Quoted 5 lines

If CONFIG_SND_SOC_RT5575_SPI is enabled, and the SPI kept failing for whatever reason, the I2C device probe would get deferred continuously. As the kconfig looks optional, the question is: Does rt5575 codec exist in a "without SPI flash" version or, is it always shipped with SPI flash?

Yes, the customers can choose 2 different hardware configurations in w/wo external SPI flash.

Quoted 2 lines

If it does exist as a standalone component - without SPI - then enabling CONFIG_SND_SOC_RT5575_SPI effectively renders the driver dysfunctional.

The multiple compatible names can distinguish between w/wo flash to avoid it.

static const struct of_device_id rt5575_of_match[] = {
	{ .compatible = "realtek,rt5575" },
	{ .compatible = "realtek,rt5575-with-spi" },
	{ }
};

#if IS_ENABLED(CONFIG_SND_SOC_RT5575_SPI)
	if (!rt5575_spi && of_device_is_compatible(dev->of_node, rt5575_of_match[1].compatible))
		return -EPROBE_DEFER;

#endif

Quoted 66 lines

sound/soc/codecs/rt5575.c

Quoted 11 lines

sound/soc/codecs/rt5575.c new file +11 -0

+
+     rt5575 = devm_kzalloc(&i2c->dev, sizeof(struct rt5575_priv),
+                             GFP_KERNEL);
+     if (rt5575 == NULL)
+             return -ENOMEM;
+
+     i2c_set_clientdata(i2c, rt5575);
+
+     rt5575->i2c = i2c;
+
+     rt5575->dsp_regmap = devm_regmap_init_i2c(i2c,

&rt5575_dsp_regmap);

sound/soc/codecs/rt5575.c

Quoted 3 lines

sound/soc/codecs/rt5575.c new file +3 -0

+     if (IS_ERR(rt5575->dsp_regmap)) {
+             ret = PTR_ERR(rt5575->dsp_regmap);
+             dev_err(&i2c->dev, "Failed to allocate register map: %d\n",

ret);

sound/soc/codecs/rt5575.c

Quoted 4 lines

sound/soc/codecs/rt5575.c new file +4 -0

+             return ret;
+     }
+
+     rt5575->regmap = devm_regmap_init(&i2c->dev, NULL, i2c,

&rt5575_regmap);

sound/soc/codecs/rt5575.c

Quoted 3 lines

sound/soc/codecs/rt5575.c new file +3 -0

+     if (IS_ERR(rt5575->regmap)) {
+             ret = PTR_ERR(rt5575->regmap);
+             dev_err(&i2c->dev, "Failed to allocate register map: %d\n",

ret);

sound/soc/codecs/rt5575.c

Quoted 6 lines

sound/soc/codecs/rt5575.c new file +6 -0

+             return ret;
+     }
+
+     regmap_read(rt5575->regmap, RT5575_ID, &val);
+     if (val != RT5575_DEVICE_ID) {
+             dev_err(&i2c->dev, "Device with ID register %08x is not

rt5575\n", val);

sound/soc/codecs/rt5575.c

Quoted 10 lines

sound/soc/codecs/rt5575.c new file +10 -0

+             return -ENODEV;
+     }
+
+     regmap_read(rt5575->regmap, RT5575_ID_1, &val);
+     if (!val) {
+             dev_err(&i2c->dev, "This is not formal version\n");
+             return -ENODEV;
+     }
+
+     return devm_snd_soc_register_component(&i2c->dev,

&rt5575_soc_component_dev, rt5575_dai,

Quoted 23 lines

sound/soc/codecs/rt5575.c new file +23 -0

+             ARRAY_SIZE(rt5575_dai));
+}
+
+static const struct of_device_id rt5575_of_match[] = {
+     { .compatible = "realtek,rt5575" },
+     { }
+};
+MODULE_DEVICE_TABLE(of, rt5575_of_match);
+
+static struct i2c_driver rt5575_i2c_driver = {
+     .driver = {
+             .name = "rt5575",
+             .owner = THIS_MODULE,
+             .of_match_table = of_match_ptr(rt5575_of_match),
+     },
+     .probe = rt5575_i2c_probe,
+     .id_table = rt5575_i2c_id,
+};
+module_i2c_driver(rt5575_i2c_driver);
+
+MODULE_DESCRIPTION("ASoC ALC5575 driver");
+MODULE_AUTHOR("Oder Chiou <[email protected]>");
+MODULE_LICENSE("GPL");

Mark Brown

maintainer

2026-01-06 19:42 UTC #5 lore raw

In reply to Oder Chiou

Quoted 8 lines · Oder Chiou

The ALC5575 integrates an audio DSP that typically loads its firmware from an external flash via its own SPI host interface. In certain hardware configurations, the firmware can alternatively be loaded through the SPI client interface. The driver provides basic mute and volume control functions. When the SPI client interface is enabled, firmware loading is handled by the SPI driver.
[...]

Applied to

https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound.git for-next

Thanks!

[1/2] ASoC: rt5575: Add the codec driver for the ALC5575

      commit: c7ac7499ac5bb50ab3e00add121822c5c904bc91

[2/2] ASoC: dt-bindings: realtek,rt5575: add bindings for ALC5575
(no commit info)

All being well this means that it will be integrated into the linux-next tree (usually sometime in the next 24 hours) and sent to Linus during the next merge window (or sooner if it is a bug fix), however if problems are discovered then the patch may be dropped or reverted.

You may get further e-mails resulting from automated or manual testing and review of the tree, please engage with people reporting problems and send followup patches addressing any issues that are reported if needed.

If any updates are required or you are submitting further changes they should be sent as incremental updates against current git, existing patches will not be replaced.

Please add any relevant lists and maintainers to the CCs when replying to this mail.

Thanks,
Mark