Skip to contentSkip to conversation

ASoC: add rt5665 codec driver

PATCH Superseded Bard Liao 2 participants 4 messages Oct 31, 2016 – Nov 7, 2016 2 messages not in the archive

Subsystems (4)

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore

Bard Liao

author
#0 lore raw

PATCH 1/1 ASoC: add rt5665 codec driver

This is the initial codec driver for rt5665.

---
 Documentation/devicetree/bindings/sound/rt5665.txt |   68 +
 include/sound/rt5665.h                             |   51 +
 sound/soc/codecs/Kconfig                           |    6 +
 sound/soc/codecs/Makefile                          |    2 +
 sound/soc/codecs/rt5665.c                          | 4643 ++++++++++++++++++++
 sound/soc/codecs/rt5665.h                          | 2017 +++++++++
 6 files changed, 6787 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/sound/rt5665.txt
 create mode 100644 include/sound/rt5665.h
 create mode 100644 sound/soc/codecs/rt5665.c
 create mode 100644 sound/soc/codecs/rt5665.h
Documentation/devicetree/bindings/sound/rt5665.txt new file +68 -0
new file mode 100644
index 0000000..419c892
@@ -0,0 +1,68 @@
+RT5665/RT5666/RT5668 audio CODEC
+
+This device supports I2C only.
+
+Required properties:
+
+- compatible : One of "realtek,rt5665", "realtek,rt5666" or "realtek,rt5668".
+
+- reg : The I2C address of the device.
+
+- interrupts : The CODEC's interrupt output.
+
+Optional properties:
+
+- realtek,in1-differential
+- realtek,in2-differential
+- realtek,in3-differential
+- realtek,in4-differential
+  Boolean. Indicate MIC1/2/3/4 input are differential, rather than single-ended.
+
+- realtek,dmic1-data-pin
+  0: dmic1 is not used
+  1: using GPIO4 pin as dmic1 data pin
+  2: using IN2N pin as dmic2 data pin
+
+- realtek,dmic2-data-pin
+  0: dmic2 is not used
+  1: using GPIO5 pin as dmic2 data pin
+  2: using IN2P pin as dmic2 data pin
+
+- realtek,jd-src
+  0: No JD is used
+  1: using JD1 as JD source
+
+- realtek,ldo1-en-gpios : The GPIO that controls the CODEC's LDO1_EN pin.
+
+Pins on the device (for linking into audio routes) for RT5659/RT5658:
+
+  * DMIC L1
+  * DMIC R1
+  * DMIC L2
+  * DMIC R2
+  * IN1P
+  * IN1N
+  * IN2P
+  * IN2N
+  * IN3P
+  * IN3N
+  * IN4P
+  * IN4N
+  * HPOL
+  * HPOR
+  * LOUTL
+  * LOUTR
+  * MONOOUT
+  * PDML
+  * PDMR
+
+Example:
+
+rt5659 {
+	compatible = "realtek,rt5665";
+	reg = <0x1b>;
+	interrupt-parent = <&gpio>;
+	interrupts = <TEGRA_GPIO(W, 3) GPIO_ACTIVE_HIGH>;
+	realtek,ldo1-en-gpios =
+		<&gpio TEGRA_GPIO(V, 3) GPIO_ACTIVE_HIGH>;
+};
include/sound/rt5665.h new file +51 -0

51 lines not on this page · raw message

sound/soc/codecs/Kconfig +6 -0
index 0b1be61..8b9b0729 100644
@@ -116,6 +116,7 @@ config SND_SOC_ALL_CODECS
 	select SND_SOC_RT5651 if I2C
 	select SND_SOC_RT5659 if I2C
 	select SND_SOC_RT5660 if I2C
+	select SND_SOC_RT5665 if I2C
 	select SND_SOC_RT5663 if I2C
 	select SND_SOC_RT5670 if I2C
 	select SND_SOC_RT5677 if I2C && SPI_MASTER
@@ -662,6 +663,7 @@ config SND_SOC_RL6231
 	default y if SND_SOC_RT5651=y
 	default y if SND_SOC_RT5659=y
 	default y if SND_SOC_RT5660=y
+	default y if SND_SOC_RT5665=y
 	default y if SND_SOC_RT5663=y
 	default y if SND_SOC_RT5670=y
 	default y if SND_SOC_RT5677=y
@@ -672,6 +674,7 @@ config SND_SOC_RL6231
 	default m if SND_SOC_RT5651=m
 	default m if SND_SOC_RT5659=m
 	default m if SND_SOC_RT5660=m
+	default m if SND_SOC_RT5665=m
 	default m if SND_SOC_RT5663=m
 	default m if SND_SOC_RT5670=m
 	default m if SND_SOC_RT5677=m
@@ -721,6 +724,9 @@ config SND_SOC_RT5659
 config SND_SOC_RT5660
 	tristate
 
+config SND_SOC_RT5665
+	tristate
+
 config SND_SOC_RT5663
 	tristate
 
sound/soc/codecs/Makefile +2 -0
index d8984eb..4ab8dac 100644
@@ -117,6 +117,7 @@ snd-soc-rt5645-objs := rt5645.o
 snd-soc-rt5651-objs := rt5651.o
 snd-soc-rt5659-objs := rt5659.o
 snd-soc-rt5660-objs := rt5660.o
+snd-soc-rt5665-objs := rt5665.o
 snd-soc-rt5663-objs := rt5663.o
 snd-soc-rt5670-objs := rt5670.o
 snd-soc-rt5677-objs := rt5677.o
@@ -343,6 +344,7 @@ obj-$(CONFIG_SND_SOC_RT5645)	+= snd-soc-rt5645.o
 obj-$(CONFIG_SND_SOC_RT5651)	+= snd-soc-rt5651.o
 obj-$(CONFIG_SND_SOC_RT5659)	+= snd-soc-rt5659.o
 obj-$(CONFIG_SND_SOC_RT5660)	+= snd-soc-rt5660.o
+obj-$(CONFIG_SND_SOC_RT5665)	+= snd-soc-rt5665.o
 obj-$(CONFIG_SND_SOC_RT5663)	+= snd-soc-rt5663.o
 obj-$(CONFIG_SND_SOC_RT5670)	+= snd-soc-rt5670.o
 obj-$(CONFIG_SND_SOC_RT5677)	+= snd-soc-rt5677.o
sound/soc/codecs/rt5665.c new file +4643 -0

2,837 lines not on this page · raw message

sound/soc/codecs/rt5665.h new file +2017 -0

2,017 lines not on this page · raw message

The rest of this diff is not shown. View the raw message.

Signature · 1 line
1.8.1.1.439.g50a6b54

Part of the diff is not shown here. View the raw message.

maintainer
#1 lore raw

sound/soc/codecs/rt5665.c line 3387

Quoted 2 lines · Bard Liao
sound/soc/codecs/rt5665.c new file +2 -0
+	{"TDM2 Data Mux", "1324", "IF1_2 ADC TDM"},
+	{"TDM2 Data Mux", "1342", "IF1_2 ADC TDM"},

These aren't really done properly - they're squashing all the data paths down into one. Ideally there'd be a separate AIF widget for each channel. It's not super urgent right now since it won't make a practical difference.

sound/soc/codecs/rt5665.c line 4430

Quoted 6 lines
sound/soc/codecs/rt5665.c new file +6 -0
+	regulator_1v8 = devm_regulator_get_optional(&i2c->dev,
+						rt5665->pdata.regulator_1v8);
+	if (IS_ERR(regulator_1v8))
+		dev_err(&i2c->dev, "Fail to get regulator_1v8\n");
+	else if (regulator_enable(regulator_1v8))
+		dev_err(&i2c->dev, "Fail to enable regulator_1v8\n");

This isn't good, while we log if we fail to enable the regulator we don't otherwise handle the error. If the we failed to power on a power supply the chances are that the device isn't going to work so I'd expect us to fail the proble.

sound/soc/codecs/rt5665.c line 4440

Quoted 5 lines
sound/soc/codecs/rt5665.c new file +4 -0
+	regulator_3v3 = devm_regulator_get_optional(&i2c->dev,
+						rt5665->pdata.regulator_3v3);
 
+	regulator_5v = devm_regulator_get_optional(&i2c->dev,
+						rt5665->pdata.regulator_5v);

I also just don't believe that *all* of the regulators the device has are optional.

A message that is not in the archive (sent off-list or not yet received): ABFD875FF5FB574BA706497D987D48D7011D28D9@RTITMBSV03.realtek.com.tw

maintainer
#3 lore raw
Quoted 7 lines · Bard Liao
Quoted 3 lines

These aren't really done properly - they're squashing all the data paths down into one. Ideally there'd be a separate AIF widget for each channel. It's not super urgent right now since it won't make a practical difference.

I will modify that. Something like below, right? SND_SOC_DAPM_AIF_OUT("AIFOUTL", "Capture", 0, SND_SOC_NOPM, 0, 0), SND_SOC_DAPM_AIF_OUT("AIFOUTR", "Capture", 1, SND_SOC_NOPM, 0, 0),

Yes, exactly

Quoted 11 lines
Quoted 7 lines

sound/soc/codecs/rt5665.c line 4440

Quoted 5 lines
sound/soc/codecs/rt5665.c new file +4 -0
+	regulator_3v3 = devm_regulator_get_optional(&i2c->dev,
+						rt5665->pdata.regulator_3v3);
 
+	regulator_5v = devm_regulator_get_optional(&i2c->dev,
+						rt5665->pdata.regulator_5v);

I also just don't believe that *all* of the regulators the device has are optional.

That is because we don't know how customers design their schematic. All regulators can be connected by hardware pin. It means we don't need to use a regular to supply it.

No, that's something the regulator API will handle for you - individual drivers should be written with sensible error handling, the regulator API will handle stubbing out things that are needed.

A message that is not in the archive (sent off-list or not yet received): ABFD875FF5FB574BA706497D987D48D7011D5502@RTITMBSV03.realtek.com.tw

maintainer
#5 lore raw
Quoted 9 lines · Bard Liao
Quoted 3 lines

No, that's something the regulator API will handle for you - individual drivers should be written with sensible error handling, the regulator API will handle stubbing out things that are needed.

Sorry, maybe I didn't describe it clear. What I mean is all power pins needed by rt5665 codec can be connected by hardware directly without SW control. There is no specific power sequence for these power pins. So, we don't need to control them by driver. That's why we think they are optional.

That's exactly what I understood you to mean. The regulator API will hide this for you.