alsa-devel.alsa-project.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v3 0/5] MFD/ASoC: Add support for Intel Bay Trail boards with WM5102 codec
@ 2021-01-17 21:22 Hans de Goede
  2021-01-17 21:22 ` [PATCH v3 1/5] mfd: arizona: Add MODULE_SOFTDEP("pre: arizona_ldo1") Hans de Goede
                   ` (4 more replies)
  0 siblings, 5 replies; 15+ messages in thread
From: Hans de Goede @ 2021-01-17 21:22 UTC (permalink / raw)
  To: Lee Jones, Cezary Rojewski, Pierre-Louis Bossart, Liam Girdwood,
	Jie Yang, Mark Brown
  Cc: alsa-devel, Charles Keepax, patches, linux-kernel, Hans de Goede,
	Andy Shevchenko

Hi All,

Here is v3 of my series to add support for Intel Bay Trail based devices
which use a WM5102 codec for audio output/input. New in v3 is that the
compile error when CONFIG_ACPI is unset should be gone. Everything else
is the same as in v2, see the cover-letter of v2 below:

This was developed and tested on a Lenovo Yoga Tablet 1051L.

The biggest difference from v1 is that, based on all the discussions about
the jack-detect stuff, I've decided to split this into 2 series.
One series adding the basic support for this setup, which is this series.

And a second series which reworks the extcon driver into an arizona-jackdet
library and then modifies the codec drivers to use that directly, replacing
the old separate extcon child-device and extcon-driver.

The are 2 reasons for the split:

1. With the new jack-det rework, the series really address 2 separate
(but related) enhancements.

2. I expect this series to be ready for merging, while the jack-det stuff
likely will need a couple more revisions.

Other then the split there are some minor changes addressing various review
comments, see the individual patch changelogs.

The MFD and ASoC parts can be merged independent from each-other (both must
be merged to get working sound on these boards, but that is only a runtime
dependency and a part missing won't have any bad side-effects). Or the
entire series could be merged through the MFD tree if people prefer that.

Regards,

Hans



^ permalink raw reply	[flat|nested] 15+ messages in thread

* [PATCH v3 1/5] mfd: arizona: Add MODULE_SOFTDEP("pre: arizona_ldo1")
  2021-01-17 21:22 [PATCH v3 0/5] MFD/ASoC: Add support for Intel Bay Trail boards with WM5102 codec Hans de Goede
@ 2021-01-17 21:22 ` Hans de Goede
  2021-01-17 21:22 ` [PATCH v3 2/5] mfd: arizona: Replace arizona_of_get_type() with device_get_match_data() Hans de Goede
                   ` (3 subsequent siblings)
  4 siblings, 0 replies; 15+ messages in thread
From: Hans de Goede @ 2021-01-17 21:22 UTC (permalink / raw)
  To: Lee Jones, Cezary Rojewski, Pierre-Louis Bossart, Liam Girdwood,
	Jie Yang, Mark Brown
  Cc: alsa-devel, Charles Keepax, patches, linux-kernel, Hans de Goede,
	Andy Shevchenko

The (shared) probing code of the arizona-i2c and arizona-spi modules
takes the following steps during init:

1. Call mfd_add_devices() for a set of early child-devices, this
includes the arizona_ldo1 device which provides one of the
core-regulators.

2. Bulk enable the core-regulators.

3. Read the device id.

4. Call mfd_add_devices() for the other child-devices.

This sequence depends on 1. leading to not only the child-device
being created, but also the driver for the child-device binding
to it and registering its regulator.

This requires the arizona_ldo1 driver to be loaded before the
shared probing code runs. Add a softdep for this to both modules to
ensure that this requirement is met.

Note this mirrors the existing MODULE_SOFTDEP("pre: wm8994_regulator")
in the wm8994 code, which has a similar init sequence.

Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
 drivers/mfd/arizona-i2c.c | 1 +
 drivers/mfd/arizona-spi.c | 1 +
 2 files changed, 2 insertions(+)

diff --git a/drivers/mfd/arizona-i2c.c b/drivers/mfd/arizona-i2c.c
index 4b58e3ad6eb6..2a4a3a164d0a 100644
--- a/drivers/mfd/arizona-i2c.c
+++ b/drivers/mfd/arizona-i2c.c
@@ -115,6 +115,7 @@ static struct i2c_driver arizona_i2c_driver = {
 
 module_i2c_driver(arizona_i2c_driver);
 
+MODULE_SOFTDEP("pre: arizona_ldo1");
 MODULE_DESCRIPTION("Arizona I2C bus interface");
 MODULE_AUTHOR("Mark Brown <broonie@opensource.wolfsonmicro.com>");
 MODULE_LICENSE("GPL");
diff --git a/drivers/mfd/arizona-spi.c b/drivers/mfd/arizona-spi.c
index 2633e147b76c..704f214d2614 100644
--- a/drivers/mfd/arizona-spi.c
+++ b/drivers/mfd/arizona-spi.c
@@ -110,6 +110,7 @@ static struct spi_driver arizona_spi_driver = {
 
 module_spi_driver(arizona_spi_driver);
 
+MODULE_SOFTDEP("pre: arizona_ldo1");
 MODULE_DESCRIPTION("Arizona SPI bus interface");
 MODULE_AUTHOR("Mark Brown <broonie@opensource.wolfsonmicro.com>");
 MODULE_LICENSE("GPL");
-- 
2.28.0


^ permalink raw reply related	[flat|nested] 15+ messages in thread

* [PATCH v3 2/5] mfd: arizona: Replace arizona_of_get_type() with device_get_match_data()
  2021-01-17 21:22 [PATCH v3 0/5] MFD/ASoC: Add support for Intel Bay Trail boards with WM5102 codec Hans de Goede
  2021-01-17 21:22 ` [PATCH v3 1/5] mfd: arizona: Add MODULE_SOFTDEP("pre: arizona_ldo1") Hans de Goede
@ 2021-01-17 21:22 ` Hans de Goede
  2021-01-18 11:51   ` Andy Shevchenko
  2021-01-20 10:02   ` Charles Keepax
  2021-01-17 21:22 ` [PATCH v3 3/5] mfd: arizona: Add support for ACPI enumeration of WM5102 connected over SPI Hans de Goede
                   ` (2 subsequent siblings)
  4 siblings, 2 replies; 15+ messages in thread
From: Hans de Goede @ 2021-01-17 21:22 UTC (permalink / raw)
  To: Lee Jones, Cezary Rojewski, Pierre-Louis Bossart, Liam Girdwood,
	Jie Yang, Mark Brown
  Cc: alsa-devel, Charles Keepax, patches, linux-kernel, Hans de Goede,
	Andy Shevchenko

Replace the custom arizona_of_get_type() function with the generic
device_get_match_data() helper. Besides being a nice cleanup this
also makes it easier to add support for binding to ACPI enumerated
devices.

While at it also fix a possible NULL pointer deref of the id
argument to the probe functions (this could happen on e.g. manual
driver binding through sysfs).

Suggested-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
Changes in v2:
- New patch in v2 of this patchset
---
 drivers/mfd/arizona-core.c | 11 -----------
 drivers/mfd/arizona-i2c.c  | 10 ++++++----
 drivers/mfd/arizona-spi.c  | 10 ++++++----
 drivers/mfd/arizona.h      |  9 ---------
 4 files changed, 12 insertions(+), 28 deletions(-)

diff --git a/drivers/mfd/arizona-core.c b/drivers/mfd/arizona-core.c
index 000cb82023e3..75f1bc671d59 100644
--- a/drivers/mfd/arizona-core.c
+++ b/drivers/mfd/arizona-core.c
@@ -797,17 +797,6 @@ const struct dev_pm_ops arizona_pm_ops = {
 EXPORT_SYMBOL_GPL(arizona_pm_ops);
 
 #ifdef CONFIG_OF
-unsigned long arizona_of_get_type(struct device *dev)
-{
-	const struct of_device_id *id = of_match_device(arizona_of_match, dev);
-
-	if (id)
-		return (unsigned long)id->data;
-	else
-		return 0;
-}
-EXPORT_SYMBOL_GPL(arizona_of_get_type);
-
 static int arizona_of_get_core_pdata(struct arizona *arizona)
 {
 	struct arizona_pdata *pdata = &arizona->pdata;
diff --git a/drivers/mfd/arizona-i2c.c b/drivers/mfd/arizona-i2c.c
index 2a4a3a164d0a..5e83b730c4ce 100644
--- a/drivers/mfd/arizona-i2c.c
+++ b/drivers/mfd/arizona-i2c.c
@@ -23,14 +23,16 @@
 static int arizona_i2c_probe(struct i2c_client *i2c,
 			     const struct i2c_device_id *id)
 {
+	const void *match_data;
 	struct arizona *arizona;
 	const struct regmap_config *regmap_config = NULL;
-	unsigned long type;
+	unsigned long type = 0;
 	int ret;
 
-	if (i2c->dev.of_node)
-		type = arizona_of_get_type(&i2c->dev);
-	else
+	match_data = device_get_match_data(&i2c->dev);
+	if (match_data)
+		type = (unsigned long)match_data;
+	else if (id)
 		type = id->driver_data;
 
 	switch (type) {
diff --git a/drivers/mfd/arizona-spi.c b/drivers/mfd/arizona-spi.c
index 704f214d2614..798b88295c77 100644
--- a/drivers/mfd/arizona-spi.c
+++ b/drivers/mfd/arizona-spi.c
@@ -23,14 +23,16 @@
 static int arizona_spi_probe(struct spi_device *spi)
 {
 	const struct spi_device_id *id = spi_get_device_id(spi);
+	const void *match_data;
 	struct arizona *arizona;
 	const struct regmap_config *regmap_config = NULL;
-	unsigned long type;
+	unsigned long type = 0;
 	int ret;
 
-	if (spi->dev.of_node)
-		type = arizona_of_get_type(&spi->dev);
-	else
+	match_data = device_get_match_data(&spi->dev);
+	if (match_data)
+		type = (unsigned long)match_data;
+	else if (id)
 		type = id->driver_data;
 
 	switch (type) {
diff --git a/drivers/mfd/arizona.h b/drivers/mfd/arizona.h
index 995efc6d7f32..801cbbcd71cb 100644
--- a/drivers/mfd/arizona.h
+++ b/drivers/mfd/arizona.h
@@ -50,13 +50,4 @@ int arizona_dev_exit(struct arizona *arizona);
 int arizona_irq_init(struct arizona *arizona);
 int arizona_irq_exit(struct arizona *arizona);
 
-#ifdef CONFIG_OF
-unsigned long arizona_of_get_type(struct device *dev);
-#else
-static inline unsigned long arizona_of_get_type(struct device *dev)
-{
-	return 0;
-}
-#endif
-
 #endif
-- 
2.28.0


^ permalink raw reply related	[flat|nested] 15+ messages in thread

* [PATCH v3 3/5] mfd: arizona: Add support for ACPI enumeration of WM5102 connected over SPI
  2021-01-17 21:22 [PATCH v3 0/5] MFD/ASoC: Add support for Intel Bay Trail boards with WM5102 codec Hans de Goede
  2021-01-17 21:22 ` [PATCH v3 1/5] mfd: arizona: Add MODULE_SOFTDEP("pre: arizona_ldo1") Hans de Goede
  2021-01-17 21:22 ` [PATCH v3 2/5] mfd: arizona: Replace arizona_of_get_type() with device_get_match_data() Hans de Goede
@ 2021-01-17 21:22 ` Hans de Goede
  2021-01-18 13:02   ` Mark Brown
  2021-01-17 21:22 ` [PATCH v3 4/5] ASoC: Intel: Add DMI quirk table to soc_intel_is_byt_cr() Hans de Goede
  2021-01-17 21:22 ` [PATCH v3 5/5] ASoC: Intel: bytcr_wm5102: Add machine driver for BYT/WM5102 Hans de Goede
  4 siblings, 1 reply; 15+ messages in thread
From: Hans de Goede @ 2021-01-17 21:22 UTC (permalink / raw)
  To: Lee Jones, Cezary Rojewski, Pierre-Louis Bossart, Liam Girdwood,
	Jie Yang, Mark Brown
  Cc: alsa-devel, Charles Keepax, Christian Hartmann, patches,
	linux-kernel, Hans de Goede, Andy Shevchenko

The Intel Bay Trail (x86/ACPI) based Lenovo Yoga Tablet 2 series use
a WM5102 codec connected over SPI.

Add support for ACPI enumeration to arizona-spi so that arizona-spi can
bind to the codec on these tablets.

This is loosely based on an earlier attempt (for Android-x86) at this by
Christian Hartmann, combined with insights in things like the speaker GPIO
from the android-x86 android port for the Lenovo Yoga Tablet 2 1051F/L [1].

[1] https://github.com/Kitsune2222/Android_Yoga_Tablet_2-1051F_Kernel

Cc: Christian Hartmann <cornogle@googlemail.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
Changes in v3:
- Fix compilation error when CONFIG_ACPI is not set

Changes in v2:
- Minor coding style tweaks
- Use memcpy instead of for loop to copy gpiod_lookup-s
- Log a warning when the ACPI "CLKE" call fails
- Drop addition of acpi_device_get_match_data() call, as the code was
  moved over to use the generic device_get_match_data() helper in a
  (new in v2) preparation patch
---
 drivers/mfd/arizona-spi.c | 121 ++++++++++++++++++++++++++++++++++++++
 1 file changed, 121 insertions(+)

diff --git a/drivers/mfd/arizona-spi.c b/drivers/mfd/arizona-spi.c
index 798b88295c77..c652b733b2cb 100644
--- a/drivers/mfd/arizona-spi.c
+++ b/drivers/mfd/arizona-spi.c
@@ -7,7 +7,10 @@
  * Author: Mark Brown <broonie@opensource.wolfsonmicro.com>
  */
 
+#include <linux/acpi.h>
 #include <linux/err.h>
+#include <linux/gpio/consumer.h>
+#include <linux/gpio/machine.h>
 #include <linux/module.h>
 #include <linux/pm_runtime.h>
 #include <linux/regmap.h>
@@ -15,11 +18,122 @@
 #include <linux/slab.h>
 #include <linux/spi/spi.h>
 #include <linux/of.h>
+#include <uapi/linux/input-event-codes.h>
 
 #include <linux/mfd/arizona/core.h>
 
 #include "arizona.h"
 
+#ifdef CONFIG_ACPI
+const struct acpi_gpio_params reset_gpios = { 1, 0, false };
+const struct acpi_gpio_params ldoena_gpios = { 2, 0, false };
+
+static const struct acpi_gpio_mapping arizona_acpi_gpios[] = {
+	{ "reset-gpios", &reset_gpios, 1, },
+	{ "wlf,ldoena-gpios", &ldoena_gpios, 1 },
+	{ }
+};
+
+/*
+ * The ACPI resources for the device only describe external GPIO-s. They do
+ * not provide mappings for the GPIO-s coming from the Arizona codec itself.
+ */
+static const struct gpiod_lookup arizona_soc_gpios[] = {
+	{ "arizona", 2, "wlf,spkvdd-ena", 0, GPIO_ACTIVE_HIGH },
+	{ "arizona", 4, "wlf,micd-pol", 0, GPIO_ACTIVE_LOW },
+};
+
+/*
+ * The AOSP 3.5 mm Headset: Accessory Specification gives the following values:
+ * Function A Play/Pause:           0 ohm
+ * Function D Voice assistant:    135 ohm
+ * Function B Volume Up           240 ohm
+ * Function C Volume Down         470 ohm
+ * Minimum Mic DC resistance     1000 ohm
+ * Minimum Ear speaker impedance   16 ohm
+ * Note the first max value below must be less then the min. speaker impedance,
+ * to allow CTIA/OMTP detection to work. The other max values are the closest
+ * value from extcon-arizona.c:arizona_micd_levels halfway 2 button resistances.
+ */
+static const struct arizona_micd_range arizona_micd_aosp_ranges[] = {
+	{ .max =  11, .key = KEY_PLAYPAUSE },
+	{ .max = 186, .key = KEY_VOICECOMMAND },
+	{ .max = 348, .key = KEY_VOLUMEUP },
+	{ .max = 752, .key = KEY_VOLUMEDOWN },
+};
+
+static void arizona_spi_acpi_remove_lookup(void *lookup)
+{
+	gpiod_remove_lookup_table(lookup);
+}
+
+static int arizona_spi_acpi_probe(struct arizona *arizona)
+{
+	struct gpiod_lookup_table *lookup;
+	acpi_status status;
+	int ret;
+
+	/* Add mappings for the 2 ACPI declared GPIOs used for reset and ldo-ena */
+	devm_acpi_dev_add_driver_gpios(arizona->dev, arizona_acpi_gpios);
+
+	/* Add lookups for the SoCs own GPIOs used for micdet-polarity and spkVDD-enable */
+	lookup = devm_kzalloc(arizona->dev,
+			      struct_size(lookup, table, ARRAY_SIZE(arizona_soc_gpios) + 1),
+			      GFP_KERNEL);
+	if (!lookup)
+		return -ENOMEM;
+
+	lookup->dev_id = dev_name(arizona->dev);
+	memcpy(lookup->table, arizona_soc_gpios, sizeof(arizona_soc_gpios));
+
+	gpiod_add_lookup_table(lookup);
+	ret = devm_add_action_or_reset(arizona->dev, arizona_spi_acpi_remove_lookup, lookup);
+	if (ret)
+		return ret;
+
+	/* Enable 32KHz clock from SoC to codec for jack-detect */
+	status = acpi_evaluate_object(ACPI_HANDLE(arizona->dev), "CLKE", NULL, NULL);
+	if (ACPI_FAILURE(status))
+		dev_warn(arizona->dev, "Failed to enable 32KHz clk ACPI error %d\n", status);
+
+	/*
+	 * Some DSDTs wrongly declare the IRQ trigger-type as IRQF_TRIGGER_FALLING
+	 * The IRQ line will stay low when a new IRQ event happens between reading
+	 * the IRQ status flags and acknowledging them. When the IRQ line stays
+	 * low like this the IRQ will never trigger again when its type is set
+	 * to IRQF_TRIGGER_FALLING. Correct the IRQ trigger-type to fix this.
+	 */
+	arizona->pdata.irq_flags = IRQF_TRIGGER_LOW;
+
+	/* Wait 200 ms after jack insertion */
+	arizona->pdata.micd_detect_debounce = 200;
+
+	/* Use standard AOSP values for headset-button mappings */
+	arizona->pdata.micd_ranges = arizona_micd_aosp_ranges;
+	arizona->pdata.num_micd_ranges = ARRAY_SIZE(arizona_micd_aosp_ranges);
+
+	return 0;
+}
+
+static const struct acpi_device_id arizona_acpi_match[] = {
+	{
+		.id = "WM510204",
+		.driver_data = WM5102,
+	},
+	{
+		.id = "WM510205",
+		.driver_data = WM5102,
+	},
+	{ }
+};
+MODULE_DEVICE_TABLE(acpi, arizona_acpi_match);
+#else
+static int arizona_spi_acpi_probe(struct arizona *arizona)
+{
+	return -ENODEV;
+}
+#endif
+
 static int arizona_spi_probe(struct spi_device *spi)
 {
 	const struct spi_device_id *id = spi_get_device_id(spi);
@@ -77,6 +191,12 @@ static int arizona_spi_probe(struct spi_device *spi)
 	arizona->dev = &spi->dev;
 	arizona->irq = spi->irq;
 
+	if (has_acpi_companion(&spi->dev)) {
+		ret = arizona_spi_acpi_probe(arizona);
+		if (ret)
+			return ret;
+	}
+
 	return arizona_dev_init(arizona);
 }
 
@@ -104,6 +224,7 @@ static struct spi_driver arizona_spi_driver = {
 		.name	= "arizona",
 		.pm	= &arizona_pm_ops,
 		.of_match_table	= of_match_ptr(arizona_of_match),
+		.acpi_match_table = ACPI_PTR(arizona_acpi_match),
 	},
 	.probe		= arizona_spi_probe,
 	.remove		= arizona_spi_remove,
-- 
2.28.0


^ permalink raw reply related	[flat|nested] 15+ messages in thread

* [PATCH v3 4/5] ASoC: Intel: Add DMI quirk table to soc_intel_is_byt_cr()
  2021-01-17 21:22 [PATCH v3 0/5] MFD/ASoC: Add support for Intel Bay Trail boards with WM5102 codec Hans de Goede
                   ` (2 preceding siblings ...)
  2021-01-17 21:22 ` [PATCH v3 3/5] mfd: arizona: Add support for ACPI enumeration of WM5102 connected over SPI Hans de Goede
@ 2021-01-17 21:22 ` Hans de Goede
  2021-01-17 21:22 ` [PATCH v3 5/5] ASoC: Intel: bytcr_wm5102: Add machine driver for BYT/WM5102 Hans de Goede
  4 siblings, 0 replies; 15+ messages in thread
From: Hans de Goede @ 2021-01-17 21:22 UTC (permalink / raw)
  To: Lee Jones, Cezary Rojewski, Pierre-Louis Bossart, Liam Girdwood,
	Jie Yang, Mark Brown
  Cc: alsa-devel, Charles Keepax, patches, linux-kernel, Hans de Goede,
	Andy Shevchenko

Some Bay Trail systems:
1. Use a non CR version of the Bay Trail SoC
2. Contain at least 6 interrupt resources so that the
   platform_get_resource(pdev, IORESOURCE_IRQ, 5) check to workaround
   non CR systems which list their IPC IRQ at index 0 despite being
   non CR does not work
3. Despite 1. and 2. still have their IPC IRQ at index 0 rather then 5

Add a DMI quirk table to check for the few known models with this issue,
so that the right IPC IRQ index is used on these systems.

Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
 sound/soc/intel/common/soc-intel-quirks.h | 25 +++++++++++++++++++++++
 1 file changed, 25 insertions(+)

diff --git a/sound/soc/intel/common/soc-intel-quirks.h b/sound/soc/intel/common/soc-intel-quirks.h
index b07df3059926..a93987ab7f4d 100644
--- a/sound/soc/intel/common/soc-intel-quirks.h
+++ b/sound/soc/intel/common/soc-intel-quirks.h
@@ -11,6 +11,7 @@
 
 #if IS_ENABLED(CONFIG_X86)
 
+#include <linux/dmi.h>
 #include <asm/cpu_device_id.h>
 #include <asm/intel-family.h>
 #include <asm/iosf_mbi.h>
@@ -38,12 +39,36 @@ SOC_INTEL_IS_CPU(cml, KABYLAKE_L);
 
 static inline bool soc_intel_is_byt_cr(struct platform_device *pdev)
 {
+	/*
+	 * List of systems which:
+	 * 1. Use a non CR version of the Bay Trail SoC
+	 * 2. Contain at least 6 interrupt resources so that the
+	 *    platform_get_resource(pdev, IORESOURCE_IRQ, 5) check below
+	 *    succeeds
+	 * 3. Despite 1. and 2. still have their IPC IRQ at index 0 rather then 5
+	 *
+	 * This needs to be here so that it can be shared between the SST and
+	 * SOF drivers. We rely on the compiler to optimize this out in files
+	 * where soc_intel_is_byt_cr is not used.
+	 */
+	static const struct dmi_system_id force_bytcr_table[] = {
+		{	/* Lenovo Yoga Tablet 2 series */
+			.matches = {
+				DMI_MATCH(DMI_SYS_VENDOR, "LENOVO"),
+				DMI_MATCH(DMI_PRODUCT_FAMILY, "YOGATablet2"),
+			},
+		},
+		{}
+	};
 	struct device *dev = &pdev->dev;
 	int status = 0;
 
 	if (!soc_intel_is_byt())
 		return false;
 
+	if (dmi_check_system(force_bytcr_table))
+		return true;
+
 	if (iosf_mbi_available()) {
 		u32 bios_status;
 
-- 
2.28.0


^ permalink raw reply related	[flat|nested] 15+ messages in thread

* [PATCH v3 5/5] ASoC: Intel: bytcr_wm5102: Add machine driver for BYT/WM5102
  2021-01-17 21:22 [PATCH v3 0/5] MFD/ASoC: Add support for Intel Bay Trail boards with WM5102 codec Hans de Goede
                   ` (3 preceding siblings ...)
  2021-01-17 21:22 ` [PATCH v3 4/5] ASoC: Intel: Add DMI quirk table to soc_intel_is_byt_cr() Hans de Goede
@ 2021-01-17 21:22 ` Hans de Goede
  4 siblings, 0 replies; 15+ messages in thread
From: Hans de Goede @ 2021-01-17 21:22 UTC (permalink / raw)
  To: Lee Jones, Cezary Rojewski, Pierre-Louis Bossart, Liam Girdwood,
	Jie Yang, Mark Brown
  Cc: alsa-devel, Charles Keepax, patches, linux-kernel, Hans de Goede,
	Andy Shevchenko

From: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

Add a new ASoc Machine driver for Intel Baytrail platforms with a
Wolfson Microelectronics WM5102 codec.

This is based on a past contributions [1] from Paulo Sergio Travaglia
<pstglia@gmail.com> based on the Levono kernel [2] combined with
insights in things like the speaker GPIO from the android-x86 android
port for the Lenovo Yoga Tablet 2 1051F/L [3].

[1] https://patchwork.kernel.org/project/alsa-devel/patch/593313f5.3636c80a.50e05.47e9@mx.google.com/
[2] https://github.com/lenovo-yt2-dev/android_kernel_lenovo_baytrail/blob/cm-12.1/sound/soc/intel/board/byt_bl_wm5102.c
[3] https://github.com/Kitsune2222/Android_Yoga_Tablet_2-1051F_Kernel

The original machine driver from the Android ports was a crude modified
copy of bytcr_rt5640.c adjusted to work with the WM5102 codec.
This version has been extensively reworked to:

1. Remove all rt5640 related quirk handling. to the best of my knowledge
this setup is only used on the Lenovo Yoga Tablet 2 series (8, 10 and 13
inch models) which all use the same setup. So there is no need to deal
with all the variations with which we need to deal on rt5640 boards.

2. Rework clock handling, properly turn off the FLL and the platform-clock
when they are no longer necessary and don't reconfigure the FLL
unnecessarily when it is already running. This fixes a number of:
"Timed out waiting for lock" warnings being logged.

3. Add the GPIO controlled Speaker-VDD regulator as a DAPM_SUPPLY

This only adds the machine driver and ACPI hooks, the BYT-CR detection
quirk which these devices need will be added in a separate patch.

BugLink: https://github.com/thesofproject/linux/issues/2485
Co-authored-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
Changes in v2:
- Drop unnecessary configuring of OPCLK
- Fix error-msg when setting the SYSCLK fails
---
 sound/soc/intel/boards/Kconfig                |  12 +
 sound/soc/intel/boards/Makefile               |   2 +
 sound/soc/intel/boards/bytcr_wm5102.c         | 465 ++++++++++++++++++
 .../intel/common/soc-acpi-intel-byt-match.c   |  16 +
 4 files changed, 495 insertions(+)
 create mode 100644 sound/soc/intel/boards/bytcr_wm5102.c

diff --git a/sound/soc/intel/boards/Kconfig b/sound/soc/intel/boards/Kconfig
index b58b9b60d37e..d1d28129a32b 100644
--- a/sound/soc/intel/boards/Kconfig
+++ b/sound/soc/intel/boards/Kconfig
@@ -111,6 +111,18 @@ config SND_SOC_INTEL_BYTCR_RT5651_MACH
 	  Say Y or m if you have such a device. This is a recommended option.
 	  If unsure select "N".
 
+config SND_SOC_INTEL_BYTCR_WM5102_MACH
+	tristate "Baytrail and Baytrail-CR with WM5102 codec"
+	depends on MFD_ARIZONA && MFD_WM5102 && SPI_MASTER && ACPI
+	depends on X86_INTEL_LPSS || COMPILE_TEST
+	select SND_SOC_ACPI
+	select SND_SOC_WM5102
+	help
+	  This adds support for ASoC machine driver for Intel(R) Baytrail and Baytrail-CR
+	  platforms with WM5102 audio codec.
+	  Say Y if you have such a device.
+	  If unsure select "N".
+
 config SND_SOC_INTEL_CHT_BSW_RT5672_MACH
 	tristate "Cherrytrail & Braswell with RT5672 codec"
 	depends on I2C && ACPI
diff --git a/sound/soc/intel/boards/Makefile b/sound/soc/intel/boards/Makefile
index 5f03e484b215..616c5fbab7d5 100644
--- a/sound/soc/intel/boards/Makefile
+++ b/sound/soc/intel/boards/Makefile
@@ -10,6 +10,7 @@ snd-soc-sst-sof-wm8804-objs := sof_wm8804.o
 snd-soc-sst-glk-rt5682_max98357a-objs := glk_rt5682_max98357a.o hda_dsp_common.o
 snd-soc-sst-bytcr-rt5640-objs := bytcr_rt5640.o
 snd-soc-sst-bytcr-rt5651-objs := bytcr_rt5651.o
+snd-soc-sst-bytcr-wm5102-objs := bytcr_wm5102.o
 snd-soc-sst-cht-bsw-rt5672-objs := cht_bsw_rt5672.o
 snd-soc-sst-cht-bsw-rt5645-objs := cht_bsw_rt5645.o
 snd-soc-sst-cht-bsw-max98090_ti-objs := cht_bsw_max98090_ti.o
@@ -51,6 +52,7 @@ obj-$(CONFIG_SND_SOC_INTEL_BDW_RT5650_MACH) += snd-soc-sst-bdw-rt5650-mach.o
 obj-$(CONFIG_SND_SOC_INTEL_BDW_RT5677_MACH) += snd-soc-sst-bdw-rt5677-mach.o
 obj-$(CONFIG_SND_SOC_INTEL_BYTCR_RT5640_MACH) += snd-soc-sst-bytcr-rt5640.o
 obj-$(CONFIG_SND_SOC_INTEL_BYTCR_RT5651_MACH) += snd-soc-sst-bytcr-rt5651.o
+obj-$(CONFIG_SND_SOC_INTEL_BYTCR_WM5102_MACH) += snd-soc-sst-bytcr-wm5102.o
 obj-$(CONFIG_SND_SOC_INTEL_CHT_BSW_RT5672_MACH) += snd-soc-sst-cht-bsw-rt5672.o
 obj-$(CONFIG_SND_SOC_INTEL_CHT_BSW_RT5645_MACH) += snd-soc-sst-cht-bsw-rt5645.o
 obj-$(CONFIG_SND_SOC_INTEL_CHT_BSW_MAX98090_TI_MACH) += snd-soc-sst-cht-bsw-max98090_ti.o
diff --git a/sound/soc/intel/boards/bytcr_wm5102.c b/sound/soc/intel/boards/bytcr_wm5102.c
new file mode 100644
index 000000000000..f38850eb2eaf
--- /dev/null
+++ b/sound/soc/intel/boards/bytcr_wm5102.c
@@ -0,0 +1,465 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ *  bytcr_wm5102.c - ASoc Machine driver for Intel Baytrail platforms with a
+ *                   Wolfson Microelectronics WM5102 codec
+ *
+ *  Copyright (C) 2020 Hans de Goede <hdegoede@redhat.com>
+ *  Loosely based on bytcr_rt5640.c which is:
+ *  Copyright (C) 2014-2020 Intel Corp
+ *  Author: Subhransu S. Prusty <subhransu.s.prusty@intel.com>
+ */
+
+#include <linux/acpi.h>
+#include <linux/clk.h>
+#include <linux/device.h>
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+#include <linux/spi/spi.h>
+#include <sound/pcm.h>
+#include <sound/pcm_params.h>
+#include <sound/soc.h>
+#include <sound/soc-acpi.h>
+#include "../../codecs/wm5102.h"
+#include "../atom/sst-atom-controls.h"
+
+#define MCLK_FREQ		25000000
+
+#define WM5102_MAX_SYSCLK_4K	49152000 /* max sysclk for 4K family */
+#define WM5102_MAX_SYSCLK_11025	45158400 /* max sysclk for 11.025K family */
+
+struct byt_wm5102_private {
+	struct clk *mclk;
+	struct gpio_desc *spkvdd_en_gpio;
+};
+
+static int byt_wm5102_spkvdd_power_event(struct snd_soc_dapm_widget *w,
+	struct snd_kcontrol *kcontrol, int event)
+{
+	struct snd_soc_card *card = w->dapm->card;
+	struct byt_wm5102_private *priv = snd_soc_card_get_drvdata(card);
+
+	gpiod_set_value_cansleep(priv->spkvdd_en_gpio,
+				 !!SND_SOC_DAPM_EVENT_ON(event));
+
+	return 0;
+}
+
+static int byt_wm5102_prepare_and_enable_pll1(struct snd_soc_dai *codec_dai, int rate)
+{
+	struct snd_soc_component *codec_component = codec_dai->component;
+	int sr_mult = ((rate % 4000) == 0) ?
+		(WM5102_MAX_SYSCLK_4K / rate) :
+		(WM5102_MAX_SYSCLK_11025 / rate);
+	int ret;
+
+	/* Reset FLL1 */
+	snd_soc_dai_set_pll(codec_dai, WM5102_FLL1_REFCLK, ARIZONA_FLL_SRC_NONE, 0, 0);
+	snd_soc_dai_set_pll(codec_dai, WM5102_FLL1, ARIZONA_FLL_SRC_NONE, 0, 0);
+
+	/* Configure the FLL1 PLL before selecting it */
+	ret = snd_soc_dai_set_pll(codec_dai, WM5102_FLL1, ARIZONA_CLK_SRC_MCLK1,
+				  MCLK_FREQ, rate * sr_mult);
+	if (ret) {
+		dev_err(codec_component->dev, "Error setting PLL: %d\n", ret);
+		return ret;
+	}
+
+	ret = snd_soc_component_set_sysclk(codec_component, ARIZONA_CLK_SYSCLK,
+					   ARIZONA_CLK_SRC_FLL1, rate * sr_mult,
+					   SND_SOC_CLOCK_IN);
+	if (ret) {
+		dev_err(codec_component->dev, "Error setting SYSCLK: %d\n", ret);
+		return ret;
+	}
+
+	ret = snd_soc_dai_set_sysclk(codec_dai, ARIZONA_CLK_SYSCLK,
+				     rate * 512, SND_SOC_CLOCK_IN);
+	if (ret) {
+		dev_err(codec_component->dev, "Error setting clock: %d\n", ret);
+		return ret;
+	}
+
+	return 0;
+}
+
+static int platform_clock_control(struct snd_soc_dapm_widget *w,
+				  struct snd_kcontrol *k, int event)
+{
+	struct snd_soc_dapm_context *dapm = w->dapm;
+	struct snd_soc_card *card = dapm->card;
+	struct snd_soc_dai *codec_dai;
+	struct byt_wm5102_private *priv = snd_soc_card_get_drvdata(card);
+	int ret;
+
+	codec_dai = snd_soc_card_get_codec_dai(card, "wm5102-aif1");
+	if (!codec_dai) {
+		dev_err(card->dev, "Error codec DAI not found\n");
+		return -EIO;
+	}
+
+	if (SND_SOC_DAPM_EVENT_ON(event)) {
+		ret = clk_prepare_enable(priv->mclk);
+		if (ret) {
+			dev_err(card->dev, "Error enabling MCLK: %d\n", ret);
+			return ret;
+		}
+		ret = byt_wm5102_prepare_and_enable_pll1(codec_dai, 48000);
+		if (ret) {
+			dev_err(card->dev, "Error setting codec sysclk: %d\n", ret);
+			return ret;
+		}
+	} else {
+		/*
+		 * The WM5102 has a separate 32KHz clock for jack-detect
+		 * so we can disable the PLL, followed by disabling the
+		 * platform clock which is the source-clock for the PLL.
+		 */
+		snd_soc_dai_set_pll(codec_dai, WM5102_FLL1, ARIZONA_FLL_SRC_NONE, 0, 0);
+		clk_disable_unprepare(priv->mclk);
+	}
+
+	return 0;
+}
+
+static const struct snd_soc_dapm_widget byt_wm5102_widgets[] = {
+	SND_SOC_DAPM_HP("Headphone", NULL),
+	SND_SOC_DAPM_MIC("Headset Mic", NULL),
+	SND_SOC_DAPM_MIC("Internal Mic", NULL),
+	SND_SOC_DAPM_SPK("Speaker", NULL),
+	SND_SOC_DAPM_SUPPLY("Platform Clock", SND_SOC_NOPM, 0, 0,
+			    platform_clock_control, SND_SOC_DAPM_PRE_PMU |
+			    SND_SOC_DAPM_POST_PMD),
+	SND_SOC_DAPM_SUPPLY("Speaker VDD", SND_SOC_NOPM, 0, 0,
+			    byt_wm5102_spkvdd_power_event,
+			    SND_SOC_DAPM_PRE_PMD | SND_SOC_DAPM_POST_PMU),
+};
+
+static const struct snd_soc_dapm_route byt_wm5102_audio_map[] = {
+	{"Headphone", NULL, "Platform Clock"},
+	{"Headset Mic", NULL, "Platform Clock"},
+	{"Internal Mic", NULL, "Platform Clock"},
+	{"Speaker", NULL, "Platform Clock"},
+
+	{"Speaker", NULL, "SPKOUTLP"},
+	{"Speaker", NULL, "SPKOUTLN"},
+	{"Speaker", NULL, "SPKOUTRP"},
+	{"Speaker", NULL, "SPKOUTRN"},
+	{"Speaker", NULL, "Speaker VDD"},
+
+	{"Headphone", NULL, "HPOUT1L"},
+	{"Headphone", NULL, "HPOUT1R"},
+
+	{"Internal Mic", NULL, "MICBIAS3"},
+	{"IN3L", NULL, "Internal Mic"},
+
+	/*
+	 * The Headset Mix uses MICBIAS1 or 2 depending on if a CTIA/OMTP Headset
+	 * is connected, as the MICBIAS is applied after the CTIA/OMTP cross-switch.
+	 */
+	{"Headset Mic", NULL, "MICBIAS1"},
+	{"Headset Mic", NULL, "MICBIAS2"},
+	{"IN1L", NULL, "Headset Mic"},
+
+	{"AIF1 Playback", NULL, "ssp0 Tx"},
+	{"ssp0 Tx", NULL, "modem_out"},
+
+	{"modem_in", NULL, "ssp0 Rx"},
+	{"ssp0 Rx", NULL, "AIF1 Capture"},
+};
+
+static const struct snd_kcontrol_new byt_wm5102_controls[] = {
+	SOC_DAPM_PIN_SWITCH("Headphone"),
+	SOC_DAPM_PIN_SWITCH("Headset Mic"),
+	SOC_DAPM_PIN_SWITCH("Internal Mic"),
+	SOC_DAPM_PIN_SWITCH("Speaker"),
+};
+
+static int byt_wm5102_init(struct snd_soc_pcm_runtime *runtime)
+{
+	struct snd_soc_card *card = runtime->card;
+	struct byt_wm5102_private *priv = snd_soc_card_get_drvdata(card);
+	int ret;
+
+	card->dapm.idle_bias_off = true;
+
+	ret = snd_soc_add_card_controls(card, byt_wm5102_controls,
+					ARRAY_SIZE(byt_wm5102_controls));
+	if (ret) {
+		dev_err(card->dev, "Error adding card controls: %d\n", ret);
+		return ret;
+	}
+
+	/*
+	 * The firmware might enable the clock at boot (this information
+	 * may or may not be reflected in the enable clock register).
+	 * To change the rate we must disable the clock first to cover these
+	 * cases. Due to common clock framework restrictions that do not allow
+	 * to disable a clock that has not been enabled, we need to enable
+	 * the clock first.
+	 */
+	ret = clk_prepare_enable(priv->mclk);
+	if (!ret)
+		clk_disable_unprepare(priv->mclk);
+
+	ret = clk_set_rate(priv->mclk, MCLK_FREQ);
+	if (ret) {
+		dev_err(card->dev, "Error setting MCLK rate: %d\n", ret);
+		return ret;
+	}
+
+	return 0;
+}
+
+static const struct snd_soc_pcm_stream byt_wm5102_dai_params = {
+	.formats = SNDRV_PCM_FMTBIT_S16_LE,
+	.rate_min = 48000,
+	.rate_max = 48000,
+	.channels_min = 2,
+	.channels_max = 2,
+};
+
+static int byt_wm5102_codec_fixup(struct snd_soc_pcm_runtime *rtd,
+				  struct snd_pcm_hw_params *params)
+{
+	struct snd_interval *rate = hw_param_interval(params,
+						      SNDRV_PCM_HW_PARAM_RATE);
+	struct snd_interval *channels = hw_param_interval(params,
+							  SNDRV_PCM_HW_PARAM_CHANNELS);
+	int ret;
+
+	/* The DSP will covert the FE rate to 48k, stereo */
+	rate->min = 48000;
+	rate->max = 48000;
+	channels->min = 2;
+	channels->max = 2;
+
+	/* set SSP0 to 16-bit */
+	params_set_format(params, SNDRV_PCM_FORMAT_S16_LE);
+
+	/*
+	 * Default mode for SSP configuration is TDM 4 slot, override config
+	 * with explicit setting to I2S 2ch 16-bit. The word length is set with
+	 * dai_set_tdm_slot() since there is no other API exposed
+	 */
+	ret = snd_soc_dai_set_fmt(asoc_rtd_to_cpu(rtd, 0),
+				  SND_SOC_DAIFMT_I2S     |
+				  SND_SOC_DAIFMT_NB_NF   |
+				  SND_SOC_DAIFMT_CBS_CFS);
+	if (ret) {
+		dev_err(rtd->dev, "Error setting format to I2S: %d\n", ret);
+		return ret;
+	}
+
+	ret = snd_soc_dai_set_tdm_slot(asoc_rtd_to_cpu(rtd, 0), 0x3, 0x3, 2, 16);
+	if (ret) {
+		dev_err(rtd->dev, "Error setting I2S config: %d\n", ret);
+		return ret;
+	}
+
+	return 0;
+}
+
+static int byt_wm5102_aif1_startup(struct snd_pcm_substream *substream)
+{
+	return snd_pcm_hw_constraint_single(substream->runtime,
+					    SNDRV_PCM_HW_PARAM_RATE, 48000);
+}
+
+static const struct snd_soc_ops byt_wm5102_aif1_ops = {
+	.startup = byt_wm5102_aif1_startup,
+};
+
+SND_SOC_DAILINK_DEF(dummy,
+	DAILINK_COMP_ARRAY(COMP_DUMMY()));
+
+SND_SOC_DAILINK_DEF(media,
+	DAILINK_COMP_ARRAY(COMP_CPU("media-cpu-dai")));
+
+SND_SOC_DAILINK_DEF(deepbuffer,
+	DAILINK_COMP_ARRAY(COMP_CPU("deepbuffer-cpu-dai")));
+
+SND_SOC_DAILINK_DEF(ssp0_port,
+	DAILINK_COMP_ARRAY(COMP_CPU("ssp0-port")));
+
+SND_SOC_DAILINK_DEF(ssp0_codec,
+	DAILINK_COMP_ARRAY(COMP_CODEC(
+	/*
+	 * Note there is no need to overwrite the codec-name as is done in
+	 * other bytcr machine drivers, because the codec is a MFD child-dev.
+	 */
+	"wm5102-codec",
+	"wm5102-aif1")));
+
+SND_SOC_DAILINK_DEF(platform,
+	DAILINK_COMP_ARRAY(COMP_PLATFORM("sst-mfld-platform")));
+
+static struct snd_soc_dai_link byt_wm5102_dais[] = {
+	[MERR_DPCM_AUDIO] = {
+		.name = "Baytrail Audio Port",
+		.stream_name = "Baytrail Audio",
+		.nonatomic = true,
+		.dynamic = 1,
+		.dpcm_playback = 1,
+		.dpcm_capture = 1,
+		.ops = &byt_wm5102_aif1_ops,
+		SND_SOC_DAILINK_REG(media, dummy, platform),
+
+	},
+	[MERR_DPCM_DEEP_BUFFER] = {
+		.name = "Deep-Buffer Audio Port",
+		.stream_name = "Deep-Buffer Audio",
+		.nonatomic = true,
+		.dynamic = 1,
+		.dpcm_playback = 1,
+		.ops = &byt_wm5102_aif1_ops,
+		SND_SOC_DAILINK_REG(deepbuffer, dummy, platform),
+	},
+		/* back ends */
+	{
+		/*
+		 * This must be named SSP2-Codec even though this machine driver
+		 * always uses SSP0. Most machine drivers support both and dynamically
+		 * update the dailink to point to SSP0 or SSP2, while keeping the name
+		 * as "SSP2-Codec". The SOF tplg files hardcode the "SSP2-Codec" even
+		 * in the byt-foo-ssp0.tplg versions because the other machine-drivers
+		 * use "SSP2-Codec" even when SSP0 is used.
+		 */
+		.name = "SSP2-Codec",
+		.id = 0,
+		.no_pcm = 1,
+		.dai_fmt = SND_SOC_DAIFMT_I2S | SND_SOC_DAIFMT_NB_NF
+						| SND_SOC_DAIFMT_CBS_CFS,
+		.be_hw_params_fixup = byt_wm5102_codec_fixup,
+		.nonatomic = true,
+		.dpcm_playback = 1,
+		.dpcm_capture = 1,
+		.init = byt_wm5102_init,
+		SND_SOC_DAILINK_REG(ssp0_port, ssp0_codec, platform),
+	},
+};
+
+/* use space before codec name to simplify card ID, and simplify driver name */
+#define SOF_CARD_NAME "bytcht wm5102" /* card name will be 'sof-bytcht wm5102' */
+#define SOF_DRIVER_NAME "SOF"
+
+#define CARD_NAME "bytcr-wm5102"
+#define DRIVER_NAME NULL /* card name will be used for driver name */
+
+/* SoC card */
+static struct snd_soc_card byt_wm5102_card = {
+	.owner = THIS_MODULE,
+	.dai_link = byt_wm5102_dais,
+	.num_links = ARRAY_SIZE(byt_wm5102_dais),
+	.dapm_widgets = byt_wm5102_widgets,
+	.num_dapm_widgets = ARRAY_SIZE(byt_wm5102_widgets),
+	.dapm_routes = byt_wm5102_audio_map,
+	.num_dapm_routes = ARRAY_SIZE(byt_wm5102_audio_map),
+	.fully_routed = true,
+};
+
+static int snd_byt_wm5102_mc_probe(struct platform_device *pdev)
+{
+	char codec_name[SND_ACPI_I2C_ID_LEN];
+	struct device *dev = &pdev->dev;
+	struct byt_wm5102_private *priv;
+	struct snd_soc_acpi_mach *mach;
+	const char *platform_name;
+	struct acpi_device *adev;
+	struct device *codec_dev;
+	bool sof_parent;
+	int ret;
+
+	priv = devm_kzalloc(dev, sizeof(*priv), GFP_ATOMIC);
+	if (!priv)
+		return -ENOMEM;
+
+	/* Get MCLK */
+	priv->mclk = devm_clk_get(dev, "pmc_plt_clk_3");
+	if (IS_ERR(priv->mclk))
+		return dev_err_probe(dev, PTR_ERR(priv->mclk), "getting pmc_plt_clk_3\n");
+
+	/*
+	 * Get speaker VDD enable GPIO:
+	 * 1. Get codec-device-name
+	 * 2. Get codec-device
+	 * 3. Get GPIO from codec-device
+	 */
+	mach = dev->platform_data;
+	adev = acpi_dev_get_first_match_dev(mach->id, NULL, -1);
+	if (!adev) {
+		dev_err(dev, "Error cannot find acpi-dev for codec\n");
+		return -ENOENT;
+	}
+	snprintf(codec_name, sizeof(codec_name), "spi-%s", acpi_dev_name(adev));
+	put_device(&adev->dev);
+
+	codec_dev = bus_find_device_by_name(&spi_bus_type, NULL, codec_name);
+	if (!codec_dev)
+		return -EPROBE_DEFER;
+
+	/* Note no devm_ here since we call gpiod_get on codec_dev rather then dev */
+	priv->spkvdd_en_gpio = gpiod_get(codec_dev, "wlf,spkvdd-ena", GPIOD_OUT_LOW);
+	put_device(codec_dev);
+
+	if (IS_ERR(priv->spkvdd_en_gpio))
+		return dev_err_probe(dev, PTR_ERR(priv->spkvdd_en_gpio), "getting spkvdd-GPIO\n");
+
+	/* override platform name, if required */
+	byt_wm5102_card.dev = dev;
+	platform_name = mach->mach_params.platform;
+	ret = snd_soc_fixup_dai_links_platform_name(&byt_wm5102_card, platform_name);
+	if (ret)
+		goto out_put_gpio;
+
+	/* set card and driver name and pm-ops */
+	sof_parent = snd_soc_acpi_sof_parent(dev);
+	if (sof_parent) {
+		byt_wm5102_card.name = SOF_CARD_NAME;
+		byt_wm5102_card.driver_name = SOF_DRIVER_NAME;
+		dev->driver->pm = &snd_soc_pm_ops;
+	} else {
+		byt_wm5102_card.name = CARD_NAME;
+		byt_wm5102_card.driver_name = DRIVER_NAME;
+	}
+
+	snd_soc_card_set_drvdata(&byt_wm5102_card, priv);
+	ret = devm_snd_soc_register_card(dev, &byt_wm5102_card);
+	if (ret) {
+		dev_err_probe(dev, ret, "registering card\n");
+		goto out_put_gpio;
+	}
+
+	platform_set_drvdata(pdev, &byt_wm5102_card);
+	return 0;
+
+out_put_gpio:
+	gpiod_put(priv->spkvdd_en_gpio);
+	return ret;
+}
+
+static int snd_byt_wm5102_mc_remove(struct platform_device *pdev)
+{
+	struct snd_soc_card *card = platform_get_drvdata(pdev);
+	struct byt_wm5102_private *priv = snd_soc_card_get_drvdata(card);
+
+	gpiod_put(priv->spkvdd_en_gpio);
+	return 0;
+}
+
+static struct platform_driver snd_byt_wm5102_mc_driver = {
+	.driver = {
+		.name = "bytcr_wm5102",
+	},
+	.probe = snd_byt_wm5102_mc_probe,
+	.remove = snd_byt_wm5102_mc_remove,
+};
+
+module_platform_driver(snd_byt_wm5102_mc_driver);
+
+MODULE_DESCRIPTION("ASoC Baytrail with WM5102 codec machine driver");
+MODULE_AUTHOR("Hans de Goede <hdegoede@redhat.com>");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS("platform:bytcr_wm5102");
diff --git a/sound/soc/intel/common/soc-acpi-intel-byt-match.c b/sound/soc/intel/common/soc-acpi-intel-byt-match.c
index c348607b49a5..ec7932549655 100644
--- a/sound/soc/intel/common/soc-acpi-intel-byt-match.c
+++ b/sound/soc/intel/common/soc-acpi-intel-byt-match.c
@@ -154,6 +154,22 @@ struct snd_soc_acpi_mach  snd_soc_acpi_intel_baytrail_machines[] = {
 		.sof_fw_filename = "sof-byt.ri",
 		.sof_tplg_filename = "sof-byt-rt5651.tplg",
 	},
+	{
+		.id = "WM510204",
+		.drv_name = "bytcr_wm5102",
+		.fw_filename = "intel/fw_sst_0f28.bin",
+		.board = "bytcr_wm5102",
+		.sof_fw_filename = "sof-byt.ri",
+		.sof_tplg_filename = "sof-byt-wm5102.tplg",
+	},
+	{
+		.id = "WM510205",
+		.drv_name = "bytcr_wm5102",
+		.fw_filename = "intel/fw_sst_0f28.bin",
+		.board = "bytcr_wm5102",
+		.sof_fw_filename = "sof-byt.ri",
+		.sof_tplg_filename = "sof-byt-wm5102.tplg",
+	},
 	{
 		.id = "DLGS7212",
 		.drv_name = "bytcht_da7213",
-- 
2.28.0


^ permalink raw reply related	[flat|nested] 15+ messages in thread

* Re: [PATCH v3 2/5] mfd: arizona: Replace arizona_of_get_type() with device_get_match_data()
  2021-01-17 21:22 ` [PATCH v3 2/5] mfd: arizona: Replace arizona_of_get_type() with device_get_match_data() Hans de Goede
@ 2021-01-18 11:51   ` Andy Shevchenko
  2021-01-20 10:02   ` Charles Keepax
  1 sibling, 0 replies; 15+ messages in thread
From: Andy Shevchenko @ 2021-01-18 11:51 UTC (permalink / raw)
  To: Hans de Goede
  Cc: Cezary Rojewski, Charles Keepax, ALSA Development Mailing List,
	patches, Jie Yang, Pierre-Louis Bossart,
	Linux Kernel Mailing List, Liam Girdwood, Mark Brown, Lee Jones

On Sun, Jan 17, 2021 at 11:23 PM Hans de Goede <hdegoede@redhat.com> wrote:
>
> Replace the custom arizona_of_get_type() function with the generic
> device_get_match_data() helper. Besides being a nice cleanup this
> also makes it easier to add support for binding to ACPI enumerated
> devices.
>
> While at it also fix a possible NULL pointer deref of the id
> argument to the probe functions (this could happen on e.g. manual
> driver binding through sysfs).

Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>

> Suggested-by: Andy Shevchenko <andy.shevchenko@gmail.com>
> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
> ---
> Changes in v2:
> - New patch in v2 of this patchset
> ---
>  drivers/mfd/arizona-core.c | 11 -----------
>  drivers/mfd/arizona-i2c.c  | 10 ++++++----
>  drivers/mfd/arizona-spi.c  | 10 ++++++----
>  drivers/mfd/arizona.h      |  9 ---------
>  4 files changed, 12 insertions(+), 28 deletions(-)
>
> diff --git a/drivers/mfd/arizona-core.c b/drivers/mfd/arizona-core.c
> index 000cb82023e3..75f1bc671d59 100644
> --- a/drivers/mfd/arizona-core.c
> +++ b/drivers/mfd/arizona-core.c
> @@ -797,17 +797,6 @@ const struct dev_pm_ops arizona_pm_ops = {
>  EXPORT_SYMBOL_GPL(arizona_pm_ops);
>
>  #ifdef CONFIG_OF
> -unsigned long arizona_of_get_type(struct device *dev)
> -{
> -       const struct of_device_id *id = of_match_device(arizona_of_match, dev);
> -
> -       if (id)
> -               return (unsigned long)id->data;
> -       else
> -               return 0;
> -}
> -EXPORT_SYMBOL_GPL(arizona_of_get_type);
> -
>  static int arizona_of_get_core_pdata(struct arizona *arizona)
>  {
>         struct arizona_pdata *pdata = &arizona->pdata;
> diff --git a/drivers/mfd/arizona-i2c.c b/drivers/mfd/arizona-i2c.c
> index 2a4a3a164d0a..5e83b730c4ce 100644
> --- a/drivers/mfd/arizona-i2c.c
> +++ b/drivers/mfd/arizona-i2c.c
> @@ -23,14 +23,16 @@
>  static int arizona_i2c_probe(struct i2c_client *i2c,
>                              const struct i2c_device_id *id)
>  {
> +       const void *match_data;
>         struct arizona *arizona;
>         const struct regmap_config *regmap_config = NULL;
> -       unsigned long type;
> +       unsigned long type = 0;
>         int ret;
>
> -       if (i2c->dev.of_node)
> -               type = arizona_of_get_type(&i2c->dev);
> -       else
> +       match_data = device_get_match_data(&i2c->dev);
> +       if (match_data)
> +               type = (unsigned long)match_data;
> +       else if (id)
>                 type = id->driver_data;
>
>         switch (type) {
> diff --git a/drivers/mfd/arizona-spi.c b/drivers/mfd/arizona-spi.c
> index 704f214d2614..798b88295c77 100644
> --- a/drivers/mfd/arizona-spi.c
> +++ b/drivers/mfd/arizona-spi.c
> @@ -23,14 +23,16 @@
>  static int arizona_spi_probe(struct spi_device *spi)
>  {
>         const struct spi_device_id *id = spi_get_device_id(spi);
> +       const void *match_data;
>         struct arizona *arizona;
>         const struct regmap_config *regmap_config = NULL;
> -       unsigned long type;
> +       unsigned long type = 0;
>         int ret;
>
> -       if (spi->dev.of_node)
> -               type = arizona_of_get_type(&spi->dev);
> -       else
> +       match_data = device_get_match_data(&spi->dev);
> +       if (match_data)
> +               type = (unsigned long)match_data;
> +       else if (id)
>                 type = id->driver_data;
>
>         switch (type) {
> diff --git a/drivers/mfd/arizona.h b/drivers/mfd/arizona.h
> index 995efc6d7f32..801cbbcd71cb 100644
> --- a/drivers/mfd/arizona.h
> +++ b/drivers/mfd/arizona.h
> @@ -50,13 +50,4 @@ int arizona_dev_exit(struct arizona *arizona);
>  int arizona_irq_init(struct arizona *arizona);
>  int arizona_irq_exit(struct arizona *arizona);
>
> -#ifdef CONFIG_OF
> -unsigned long arizona_of_get_type(struct device *dev);
> -#else
> -static inline unsigned long arizona_of_get_type(struct device *dev)
> -{
> -       return 0;
> -}
> -#endif
> -
>  #endif
> --
> 2.28.0
>


-- 
With Best Regards,
Andy Shevchenko

^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: [PATCH v3 3/5] mfd: arizona: Add support for ACPI enumeration of WM5102 connected over SPI
  2021-01-17 21:22 ` [PATCH v3 3/5] mfd: arizona: Add support for ACPI enumeration of WM5102 connected over SPI Hans de Goede
@ 2021-01-18 13:02   ` Mark Brown
  2021-01-18 13:13     ` Hans de Goede
  0 siblings, 1 reply; 15+ messages in thread
From: Mark Brown @ 2021-01-18 13:02 UTC (permalink / raw)
  To: Hans de Goede
  Cc: Cezary Rojewski, Charles Keepax, Christian Hartmann, alsa-devel,
	patches, Jie Yang, Pierre-Louis Bossart, linux-kernel,
	Liam Girdwood, Andy Shevchenko, Lee Jones

[-- Attachment #1: Type: text/plain, Size: 2492 bytes --]

On Sun, Jan 17, 2021 at 10:22:50PM +0100, Hans de Goede wrote:

> +	/*
> +	 * Some DSDTs wrongly declare the IRQ trigger-type as IRQF_TRIGGER_FALLING
> +	 * The IRQ line will stay low when a new IRQ event happens between reading
> +	 * the IRQ status flags and acknowledging them. When the IRQ line stays
> +	 * low like this the IRQ will never trigger again when its type is set
> +	 * to IRQF_TRIGGER_FALLING. Correct the IRQ trigger-type to fix this.
> +	 */
> +	arizona->pdata.irq_flags = IRQF_TRIGGER_LOW;

Are you sure that all the relevant interrupt controllers support active
low interrupts?  There were issues on some systems with missing support
for active low interrupts (see the bodge in wm8994-irq.c to work around
them) - it's entirely likely that there are DSDTs that are just plain
buggy here but if someone's copying and pasting it smells like there may
be some systems that actually need an edge triggered interrupt that
they're getting it from.

> +
> +	/* Wait 200 ms after jack insertion */
> +	arizona->pdata.micd_detect_debounce = 200;
> +
> +	/* Use standard AOSP values for headset-button mappings */
> +	arizona->pdata.micd_ranges = arizona_micd_aosp_ranges;
> +	arizona->pdata.num_micd_ranges = ARRAY_SIZE(arizona_micd_aosp_ranges);
> +
> +	return 0;
> +}
> +
> +static const struct acpi_device_id arizona_acpi_match[] = {
> +	{
> +		.id = "WM510204",
> +		.driver_data = WM5102,
> +	},
> +	{
> +		.id = "WM510205",
> +		.driver_data = WM5102,
> +	},
> +	{ }
> +};
> +MODULE_DEVICE_TABLE(acpi, arizona_acpi_match);
> +#else
> +static int arizona_spi_acpi_probe(struct arizona *arizona)
> +{
> +	return -ENODEV;
> +}
> +#endif
> +
>  static int arizona_spi_probe(struct spi_device *spi)
>  {
>  	const struct spi_device_id *id = spi_get_device_id(spi);
> @@ -77,6 +191,12 @@ static int arizona_spi_probe(struct spi_device *spi)
>  	arizona->dev = &spi->dev;
>  	arizona->irq = spi->irq;
>  
> +	if (has_acpi_companion(&spi->dev)) {
> +		ret = arizona_spi_acpi_probe(arizona);
> +		if (ret)
> +			return ret;
> +	}
> +
>  	return arizona_dev_init(arizona);
>  }
>  
> @@ -104,6 +224,7 @@ static struct spi_driver arizona_spi_driver = {
>  		.name	= "arizona",
>  		.pm	= &arizona_pm_ops,
>  		.of_match_table	= of_match_ptr(arizona_of_match),
> +		.acpi_match_table = ACPI_PTR(arizona_acpi_match),
>  	},
>  	.probe		= arizona_spi_probe,
>  	.remove		= arizona_spi_remove,
> -- 
> 2.28.0
> 

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: [PATCH v3 3/5] mfd: arizona: Add support for ACPI enumeration of WM5102 connected over SPI
  2021-01-18 13:02   ` Mark Brown
@ 2021-01-18 13:13     ` Hans de Goede
  2021-01-18 13:34       ` Mark Brown
  0 siblings, 1 reply; 15+ messages in thread
From: Hans de Goede @ 2021-01-18 13:13 UTC (permalink / raw)
  To: Mark Brown
  Cc: Cezary Rojewski, Charles Keepax, Christian Hartmann, alsa-devel,
	patches, Jie Yang, Pierre-Louis Bossart, linux-kernel,
	Liam Girdwood, Andy Shevchenko, Lee Jones

Hi,

On 1/18/21 2:02 PM, Mark Brown wrote:
> On Sun, Jan 17, 2021 at 10:22:50PM +0100, Hans de Goede wrote:
> 
>> +	/*
>> +	 * Some DSDTs wrongly declare the IRQ trigger-type as IRQF_TRIGGER_FALLING
>> +	 * The IRQ line will stay low when a new IRQ event happens between reading
>> +	 * the IRQ status flags and acknowledging them. When the IRQ line stays
>> +	 * low like this the IRQ will never trigger again when its type is set
>> +	 * to IRQF_TRIGGER_FALLING. Correct the IRQ trigger-type to fix this.
>> +	 */
>> +	arizona->pdata.irq_flags = IRQF_TRIGGER_LOW;
> 
> Are you sure that all the relevant interrupt controllers support active
> low interrupts?  There were issues on some systems with missing support
> for active low interrupts (see the bodge in wm8994-irq.c to work around
> them) - it's entirely likely that there are DSDTs that are just plain
> buggy here but if someone's copying and pasting it smells like there may
> be some systems that actually need an edge triggered interrupt that
> they're getting it from.

I'm only aware of one series of devices / models which actually
use the combination of ACPI enumeration and the WM5102 codec, and that
is the Lenovo Yoga Tablet 2 series (in 8, 10 and 13 inch versions
shipping with both Windows and Android). These all use a Bay Trail
SoC which is capable of using active low interrupts.

More in general I'm not aware of any (recent-ish) x86 GPIO controllers
not being able to do active low interrupts. In theory we could hit this
code path on ARM devices using ACPI enumeration, but I don't think it
is likely we will see a combination of ARM + ACPI enumeration +
WM5102 + GPIO controller not capable of active-low interrupts.

This overriding of the flags definitely is necessary on the Lenovo
devices in question. I could add a
"if (dmi_name_in_vendors("LENOVO"))" guard around it, but that
seems unnecessary.

Regards,

Hans



> 
>> +
>> +	/* Wait 200 ms after jack insertion */
>> +	arizona->pdata.micd_detect_debounce = 200;
>> +
>> +	/* Use standard AOSP values for headset-button mappings */
>> +	arizona->pdata.micd_ranges = arizona_micd_aosp_ranges;
>> +	arizona->pdata.num_micd_ranges = ARRAY_SIZE(arizona_micd_aosp_ranges);
>> +
>> +	return 0;
>> +}
>> +
>> +static const struct acpi_device_id arizona_acpi_match[] = {
>> +	{
>> +		.id = "WM510204",
>> +		.driver_data = WM5102,
>> +	},
>> +	{
>> +		.id = "WM510205",
>> +		.driver_data = WM5102,
>> +	},
>> +	{ }
>> +};
>> +MODULE_DEVICE_TABLE(acpi, arizona_acpi_match);
>> +#else
>> +static int arizona_spi_acpi_probe(struct arizona *arizona)
>> +{
>> +	return -ENODEV;
>> +}
>> +#endif
>> +
>>  static int arizona_spi_probe(struct spi_device *spi)
>>  {
>>  	const struct spi_device_id *id = spi_get_device_id(spi);
>> @@ -77,6 +191,12 @@ static int arizona_spi_probe(struct spi_device *spi)
>>  	arizona->dev = &spi->dev;
>>  	arizona->irq = spi->irq;
>>  
>> +	if (has_acpi_companion(&spi->dev)) {
>> +		ret = arizona_spi_acpi_probe(arizona);
>> +		if (ret)
>> +			return ret;
>> +	}
>> +
>>  	return arizona_dev_init(arizona);
>>  }
>>  
>> @@ -104,6 +224,7 @@ static struct spi_driver arizona_spi_driver = {
>>  		.name	= "arizona",
>>  		.pm	= &arizona_pm_ops,
>>  		.of_match_table	= of_match_ptr(arizona_of_match),
>> +		.acpi_match_table = ACPI_PTR(arizona_acpi_match),
>>  	},
>>  	.probe		= arizona_spi_probe,
>>  	.remove		= arizona_spi_remove,
>> -- 
>> 2.28.0
>>


^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: [PATCH v3 3/5] mfd: arizona: Add support for ACPI enumeration of WM5102 connected over SPI
  2021-01-18 13:13     ` Hans de Goede
@ 2021-01-18 13:34       ` Mark Brown
  2021-01-18 13:38         ` Hans de Goede
  2021-01-20 19:18         ` Hans de Goede
  0 siblings, 2 replies; 15+ messages in thread
From: Mark Brown @ 2021-01-18 13:34 UTC (permalink / raw)
  To: Hans de Goede
  Cc: Cezary Rojewski, Charles Keepax, Christian Hartmann, alsa-devel,
	patches, Jie Yang, Pierre-Louis Bossart, linux-kernel,
	Liam Girdwood, Andy Shevchenko, Lee Jones

[-- Attachment #1: Type: text/plain, Size: 769 bytes --]

On Mon, Jan 18, 2021 at 02:13:50PM +0100, Hans de Goede wrote:

> More in general I'm not aware of any (recent-ish) x86 GPIO controllers
> not being able to do active low interrupts. In theory we could hit this
> code path on ARM devices using ACPI enumeration, but I don't think it
> is likely we will see a combination of ARM + ACPI enumeration +
> WM5102 + GPIO controller not capable of active-low interrupts.

I've not seen this issue on any ARM based systems.

> This overriding of the flags definitely is necessary on the Lenovo
> devices in question. I could add a
> "if (dmi_name_in_vendors("LENOVO"))" guard around it, but that
> seems unnecessary.

Possibly just an update to the comment to make it clear that some
firmwares might legitimately set the flag?

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: [PATCH v3 3/5] mfd: arizona: Add support for ACPI enumeration of WM5102 connected over SPI
  2021-01-18 13:34       ` Mark Brown
@ 2021-01-18 13:38         ` Hans de Goede
  2021-01-20 19:18         ` Hans de Goede
  1 sibling, 0 replies; 15+ messages in thread
From: Hans de Goede @ 2021-01-18 13:38 UTC (permalink / raw)
  To: Mark Brown
  Cc: Cezary Rojewski, Charles Keepax, Christian Hartmann, alsa-devel,
	patches, Jie Yang, Pierre-Louis Bossart, linux-kernel,
	Liam Girdwood, Andy Shevchenko, Lee Jones

Hi,

On 1/18/21 2:34 PM, Mark Brown wrote:
> On Mon, Jan 18, 2021 at 02:13:50PM +0100, Hans de Goede wrote:
> 
>> More in general I'm not aware of any (recent-ish) x86 GPIO controllers
>> not being able to do active low interrupts. In theory we could hit this
>> code path on ARM devices using ACPI enumeration, but I don't think it
>> is likely we will see a combination of ARM + ACPI enumeration +
>> WM5102 + GPIO controller not capable of active-low interrupts.
> 
> I've not seen this issue on any ARM based systems.
> 
>> This overriding of the flags definitely is necessary on the Lenovo
>> devices in question. I could add a
>> "if (dmi_name_in_vendors("LENOVO"))" guard around it, but that
>> seems unnecessary.
> 
> Possibly just an update to the comment to make it clear that some
> firmwares might legitimately set the flag
That seems sensible, I will wait a bit to so if you (or someone)
has any more review remarks to this series and then send out
a v4 with the comment updated.

Regards,

Hans


^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: [PATCH v3 2/5] mfd: arizona: Replace arizona_of_get_type() with device_get_match_data()
  2021-01-17 21:22 ` [PATCH v3 2/5] mfd: arizona: Replace arizona_of_get_type() with device_get_match_data() Hans de Goede
  2021-01-18 11:51   ` Andy Shevchenko
@ 2021-01-20 10:02   ` Charles Keepax
  1 sibling, 0 replies; 15+ messages in thread
From: Charles Keepax @ 2021-01-20 10:02 UTC (permalink / raw)
  To: Hans de Goede
  Cc: Cezary Rojewski, alsa-devel, patches, Jie Yang,
	Pierre-Louis Bossart, linux-kernel, Liam Girdwood,
	Andy Shevchenko, Mark Brown, Lee Jones

On Sun, Jan 17, 2021 at 10:22:49PM +0100, Hans de Goede wrote:
> Replace the custom arizona_of_get_type() function with the generic
> device_get_match_data() helper. Besides being a nice cleanup this
> also makes it easier to add support for binding to ACPI enumerated
> devices.
> 
> While at it also fix a possible NULL pointer deref of the id
> argument to the probe functions (this could happen on e.g. manual
> driver binding through sysfs).
> 
> Suggested-by: Andy Shevchenko <andy.shevchenko@gmail.com>
> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
> ---

Acked-by: Charles Keepax <ckeepax@opensource.cirrus.com>

Thanks,
Charles

^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: [PATCH v3 3/5] mfd: arizona: Add support for ACPI enumeration of WM5102 connected over SPI
  2021-01-18 13:34       ` Mark Brown
  2021-01-18 13:38         ` Hans de Goede
@ 2021-01-20 19:18         ` Hans de Goede
  2021-01-20 19:59           ` Andy Shevchenko
  1 sibling, 1 reply; 15+ messages in thread
From: Hans de Goede @ 2021-01-20 19:18 UTC (permalink / raw)
  To: Mark Brown
  Cc: Cezary Rojewski, Charles Keepax, Christian Hartmann, alsa-devel,
	patches, Jie Yang, Pierre-Louis Bossart, linux-kernel,
	Liam Girdwood, Andy Shevchenko, Lee Jones

Hi,

On 1/18/21 2:34 PM, Mark Brown wrote:
> On Mon, Jan 18, 2021 at 02:13:50PM +0100, Hans de Goede wrote:
> 
>> More in general I'm not aware of any (recent-ish) x86 GPIO controllers
>> not being able to do active low interrupts. In theory we could hit this
>> code path on ARM devices using ACPI enumeration, but I don't think it
>> is likely we will see a combination of ARM + ACPI enumeration +
>> WM5102 + GPIO controller not capable of active-low interrupts.
> 
> I've not seen this issue on any ARM based systems.
> 
>> This overriding of the flags definitely is necessary on the Lenovo
>> devices in question. I could add a
>> "if (dmi_name_in_vendors("LENOVO"))" guard around it, but that
>> seems unnecessary.
> 
> Possibly just an update to the comment to make it clear that some
> firmwares might legitimately set the flag?

Ok, I've extended the comment above the override of the irq-flags with
the following paragraph for v4 of this patch-set:

         * Note theoretically it is possible that some boards are not capable
         * of handling active low level interrupts. In that case setting the
         * flag to IRQF_TRIGGER_FALLING would not be a bug (and we would need
         * to work around this) but sofar all known usages of IRQF_TRIGGER_FALLING
         * are a bug in the boards DSDT.

Regards,

Hans


^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: [PATCH v3 3/5] mfd: arizona: Add support for ACPI enumeration of WM5102 connected over SPI
  2021-01-20 19:18         ` Hans de Goede
@ 2021-01-20 19:59           ` Andy Shevchenko
  2021-01-20 21:38             ` Hans de Goede
  0 siblings, 1 reply; 15+ messages in thread
From: Andy Shevchenko @ 2021-01-20 19:59 UTC (permalink / raw)
  To: Hans de Goede
  Cc: Cezary Rojewski, Charles Keepax, Christian Hartmann,
	ALSA Development Mailing List, patches, Jie Yang,
	Pierre-Louis Bossart, Linux Kernel Mailing List, Liam Girdwood,
	Mark Brown, Lee Jones

On Wed, Jan 20, 2021 at 9:18 PM Hans de Goede <hdegoede@redhat.com> wrote:
> On 1/18/21 2:34 PM, Mark Brown wrote:
> > On Mon, Jan 18, 2021 at 02:13:50PM +0100, Hans de Goede wrote:
> >
> >> More in general I'm not aware of any (recent-ish) x86 GPIO controllers
> >> not being able to do active low interrupts. In theory we could hit this
> >> code path on ARM devices using ACPI enumeration, but I don't think it
> >> is likely we will see a combination of ARM + ACPI enumeration +
> >> WM5102 + GPIO controller not capable of active-low interrupts.
> >
> > I've not seen this issue on any ARM based systems.
> >
> >> This overriding of the flags definitely is necessary on the Lenovo
> >> devices in question. I could add a
> >> "if (dmi_name_in_vendors("LENOVO"))" guard around it, but that
> >> seems unnecessary.
> >
> > Possibly just an update to the comment to make it clear that some
> > firmwares might legitimately set the flag?
>
> Ok, I've extended the comment above the override of the irq-flags with
> the following paragraph for v4 of this patch-set:
>
>          * Note theoretically it is possible that some boards are not capable
>          * of handling active low level interrupts. In that case setting the
>          * flag to IRQF_TRIGGER_FALLING would not be a bug (and we would need
>          * to work around this) but sofar all known usages of IRQF_TRIGGER_FALLING

so far

>          * are a bug in the boards DSDT.

board's

-- 
With Best Regards,
Andy Shevchenko

^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: [PATCH v3 3/5] mfd: arizona: Add support for ACPI enumeration of WM5102 connected over SPI
  2021-01-20 19:59           ` Andy Shevchenko
@ 2021-01-20 21:38             ` Hans de Goede
  0 siblings, 0 replies; 15+ messages in thread
From: Hans de Goede @ 2021-01-20 21:38 UTC (permalink / raw)
  To: Andy Shevchenko
  Cc: Cezary Rojewski, Charles Keepax, Christian Hartmann,
	ALSA Development Mailing List, patches, Jie Yang,
	Pierre-Louis Bossart, Linux Kernel Mailing List, Liam Girdwood,
	Mark Brown, Lee Jones

Hi,

On 1/20/21 8:59 PM, Andy Shevchenko wrote:
> On Wed, Jan 20, 2021 at 9:18 PM Hans de Goede <hdegoede@redhat.com> wrote:
>> On 1/18/21 2:34 PM, Mark Brown wrote:
>>> On Mon, Jan 18, 2021 at 02:13:50PM +0100, Hans de Goede wrote:
>>>
>>>> More in general I'm not aware of any (recent-ish) x86 GPIO controllers
>>>> not being able to do active low interrupts. In theory we could hit this
>>>> code path on ARM devices using ACPI enumeration, but I don't think it
>>>> is likely we will see a combination of ARM + ACPI enumeration +
>>>> WM5102 + GPIO controller not capable of active-low interrupts.
>>>
>>> I've not seen this issue on any ARM based systems.
>>>
>>>> This overriding of the flags definitely is necessary on the Lenovo
>>>> devices in question. I could add a
>>>> "if (dmi_name_in_vendors("LENOVO"))" guard around it, but that
>>>> seems unnecessary.
>>>
>>> Possibly just an update to the comment to make it clear that some
>>> firmwares might legitimately set the flag?
>>
>> Ok, I've extended the comment above the override of the irq-flags with
>> the following paragraph for v4 of this patch-set:
>>
>>          * Note theoretically it is possible that some boards are not capable
>>          * of handling active low level interrupts. In that case setting the
>>          * flag to IRQF_TRIGGER_FALLING would not be a bug (and we would need
>>          * to work around this) but sofar all known usages of IRQF_TRIGGER_FALLING
> 
> so far
> 
>>          * are a bug in the boards DSDT.
> 
> board's
> 

Thank you for the quick review, I've fixed both spelling errors for the upcoming v4.

Regards,

Hans


^ permalink raw reply	[flat|nested] 15+ messages in thread

end of thread, other threads:[~2021-01-20 21:39 UTC | newest]

Thread overview: 15+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-01-17 21:22 [PATCH v3 0/5] MFD/ASoC: Add support for Intel Bay Trail boards with WM5102 codec Hans de Goede
2021-01-17 21:22 ` [PATCH v3 1/5] mfd: arizona: Add MODULE_SOFTDEP("pre: arizona_ldo1") Hans de Goede
2021-01-17 21:22 ` [PATCH v3 2/5] mfd: arizona: Replace arizona_of_get_type() with device_get_match_data() Hans de Goede
2021-01-18 11:51   ` Andy Shevchenko
2021-01-20 10:02   ` Charles Keepax
2021-01-17 21:22 ` [PATCH v3 3/5] mfd: arizona: Add support for ACPI enumeration of WM5102 connected over SPI Hans de Goede
2021-01-18 13:02   ` Mark Brown
2021-01-18 13:13     ` Hans de Goede
2021-01-18 13:34       ` Mark Brown
2021-01-18 13:38         ` Hans de Goede
2021-01-20 19:18         ` Hans de Goede
2021-01-20 19:59           ` Andy Shevchenko
2021-01-20 21:38             ` Hans de Goede
2021-01-17 21:22 ` [PATCH v3 4/5] ASoC: Intel: Add DMI quirk table to soc_intel_is_byt_cr() Hans de Goede
2021-01-17 21:22 ` [PATCH v3 5/5] ASoC: Intel: bytcr_wm5102: Add machine driver for BYT/WM5102 Hans de Goede

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).