Merge series "Tegra210 audio graph card" from Sameer Pujar <spujar@nvidia.com>:

This series adds audio graph based sound card support for Tegra210
platforms like Jetson-TX1 an Jetson-Nano. The following preparatory
audio graph enhancement series is already merged.
 * https://patchwork.kernel.org/project/alsa-devel/list/?series=375629&state=*

Following are the summary of changes:
 * Add graph/audio-graph based schemas or schema updates for Tegra210
   component and machine drivers.
 * Add Tegra audio graph machine driver.
 * Add required DT support for Jetson-TX1/Nano.

This work is based on earlier discussion of DPCM usage for Tegra
and simple card driver updates.
 * https://lkml.org/lkml/2020/4/30/519
 * https://lkml.org/lkml/2020/6/27/4

Original v6 series was sent about 6-7 weeks back. The dependency commit,
https://lore.kernel.org/alsa-devel/1610948585-16286-1-git-send-email-spujar@nvidia.com/
is now merged. Resending this now to appear in the top of the mail list.

Changelog
=========

v5 -> v6
--------
 * Added ports or port description in YAML docs for Tegra AHUB
   devices and graph card in patch 1/6 and 2/6. Reference of
   audio-graph-port.yaml is used for AHUB devices.
 * Dropped redundant NULL check return for of_device_get_match_data()
   in patch 3/6.
 * Added 'Reviewed-by' tag from Jon Hunter.
 * No changes in remaining patches.

v4 -> v5
--------
 * Audio graph related changes were sent in separate v5 series as
   mentioned above and are dropped from current series.
 * Graph and audio graph doc patches are dropped from this series
   and are sent separately as mentioned above.
 * Minor change with phandle label for TX1 and Nano platform DT files.
 * No changes in other patches.

v3 -> v4
--------
 * Added new patches to convert graph.txt and audio-graph-card.txt
   to corresponding json-schema files. Later these references
   are used in Tegra audio graph schema.

 * AHUB component binding docs are updated to reflect the usage
   of ports/port/endpoint

 * More common stuff is moved into graph_parse_of() and this is
   used by both generic and Tegra audio graph.

 * DT binding for Tegra audio graph is updated to included "ports { }"

 * As per the suggestion 'void *data' member is dropped from
   'asoc_simple_priv' and instead container method is used to
   maintain required custom data internal to Tegra audio graph.

v2 -> v3
--------
 * Dropped new compatible addition in generic graph driver
   after reviewing it with Morimoto-san. Instead added Tegra
   audio graph driver and new compatibles are added in the same.
 * Added new patches to expose new members for customization
   in audio graph driver.
 * Added new patch for Tegra audio graph driver and related
   documentation.
 * Minor change in below commit where mutex version of helper is used
   "ASoC: audio-graph: Identify 'no_pcm' DAI links for DPCM"
 * DT binding is updated to use the newly exposed compatibles
 * No changes in other patches

v1 -> v2
--------
 * Re-organized ports/endpoints description for ADMAIF and XBAR.
   Updated DT patches accordingly.
 * After above change, multiple Codec endpoint support is not
   required and hence dropped for now. This will be considered
   separately if at all required in future.
 * Re-ordered patches in the series.

Sameer Pujar (6):
  ASoC: dt-bindings: tegra: Add graph bindings
  ASoC: dt-bindings: tegra: Add json-schema for Tegra audio graph card
  ASoC: tegra: Add audio graph based card driver
  arm64: defconfig: Enable Tegra audio graph card driver
  arm64: tegra: Audio graph header for Tegra210
  arm64: tegra: Audio graph sound card for Jetson Nano and TX1

 .../sound/nvidia,tegra-audio-graph-card.yaml       | 187 +++++++++++++++
 .../bindings/sound/nvidia,tegra186-dspk.yaml       |  18 +-
 .../bindings/sound/nvidia,tegra210-admaif.yaml     |  13 +-
 .../bindings/sound/nvidia,tegra210-ahub.yaml       |  13 +-
 .../bindings/sound/nvidia,tegra210-dmic.yaml       |  18 +-
 .../bindings/sound/nvidia,tegra210-i2s.yaml        |  18 +-
 .../boot/dts/nvidia/tegra210-audio-graph.dtsi      | 153 ++++++++++++
 arch/arm64/boot/dts/nvidia/tegra210-p2371-2180.dts | 262 +++++++++++++++++++++
 arch/arm64/boot/dts/nvidia/tegra210-p3450-0000.dts | 146 ++++++++++++
 arch/arm64/configs/defconfig                       |   1 +
 sound/soc/tegra/Kconfig                            |   9 +
 sound/soc/tegra/Makefile                           |   2 +
 sound/soc/tegra/tegra_audio_graph_card.c           | 251 ++++++++++++++++++++
 13 files changed, 1085 insertions(+), 6 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/sound/nvidia,tegra-audio-graph-card.yaml
 create mode 100644 arch/arm64/boot/dts/nvidia/tegra210-audio-graph.dtsi
 create mode 100644 sound/soc/tegra/tegra_audio_graph_card.c

--
2.7.4

3 files changed, 262 insertions, 0 deletions

diff --git a/sound/soc/tegra/Kconfig b/sound/soc/tegra/Kconfig
index a62cc87551ac..6dc83ad8607a 100644
--- a/sound/soc/tegra/Kconfig
+++ b/sound/soc/tegra/Kconfig
@@ -117,6 +117,15 @@ config SND_SOC_TEGRA210_ADMAIF
 	  channel. Buffer size is configurable for each ADMAIIF channel.
 	  Say Y or M if you want to add support for Tegra210 ADMAIF module.
 
+config SND_SOC_TEGRA_AUDIO_GRAPH_CARD
+	tristate "Audio Graph Card based Tegra driver"
+	depends on SND_AUDIO_GRAPH_CARD
+	help
+	  Config to enable Tegra audio machine driver based on generic
+	  audio graph driver. It is a thin driver written to customize
+	  few things for Tegra audio. Most of the code is re-used from
+	  audio graph driver and the same DT bindings are used.
+
 config SND_SOC_TEGRA_RT5640
 	tristate "SoC Audio support for Tegra boards using an RT5640 codec"
 	depends on SND_SOC_TEGRA && I2C && GPIOLIB
diff --git a/sound/soc/tegra/Makefile b/sound/soc/tegra/Makefile
index 60040a06b814..b17dd6eef92a 100644
--- a/sound/soc/tegra/Makefile
+++ b/sound/soc/tegra/Makefile
@@ -38,6 +38,7 @@ snd-soc-tegra-trimslice-objs := trimslice.o
 snd-soc-tegra-alc5632-objs := tegra_alc5632.o
 snd-soc-tegra-max98090-objs := tegra_max98090.o
 snd-soc-tegra-sgtl5000-objs := tegra_sgtl5000.o
+snd-soc-tegra-audio-graph-card-objs := tegra_audio_graph_card.o
 
 obj-$(CONFIG_SND_SOC_TEGRA_RT5640) += snd-soc-tegra-rt5640.o
 obj-$(CONFIG_SND_SOC_TEGRA_RT5677) += snd-soc-tegra-rt5677.o
@@ -48,3 +49,4 @@ obj-$(CONFIG_SND_SOC_TEGRA_TRIMSLICE) += snd-soc-tegra-trimslice.o
 obj-$(CONFIG_SND_SOC_TEGRA_ALC5632) += snd-soc-tegra-alc5632.o
 obj-$(CONFIG_SND_SOC_TEGRA_MAX98090) += snd-soc-tegra-max98090.o
 obj-$(CONFIG_SND_SOC_TEGRA_SGTL5000) += snd-soc-tegra-sgtl5000.o
+obj-$(CONFIG_SND_SOC_TEGRA_AUDIO_GRAPH_CARD) += snd-soc-tegra-audio-graph-card.o
diff --git a/sound/soc/tegra/tegra_audio_graph_card.c b/sound/soc/tegra/tegra_audio_graph_card.c
new file mode 100644
index 000000000000..9e43f16f9336
--- /dev/null
+++ b/sound/soc/tegra/tegra_audio_graph_card.c
@@ -0,0 +1,251 @@
+// SPDX-License-Identifier: GPL-2.0-only
+//
+// tegra_audio_graph_card.c - Audio Graph based Tegra Machine Driver
+//
+// Copyright (c) 2020 NVIDIA CORPORATION.  All rights reserved.
+
+#include <linux/math64.h>
+#include <linux/module.h>
+#include <linux/of_device.h>
+#include <linux/platform_device.h>
+#include <sound/graph_card.h>
+#include <sound/pcm_params.h>
+
+#define MAX_PLLA_OUT0_DIV 128
+
+#define simple_to_tegra_priv(simple) \
+		container_of(simple, struct tegra_audio_priv, simple)
+
+enum srate_type {
+	/*
+	 * Sample rates multiple of 8000 Hz and below are supported:
+	 * ( 8000, 16000, 32000, 48000, 96000, 192000 Hz )
+	 */
+	x8_RATE,
+
+	/*
+	 * Sample rates multiple of 11025 Hz and below are supported:
+	 * ( 11025, 22050, 44100, 88200, 176400 Hz )
+	 */
+	x11_RATE,
+
+	NUM_RATE_TYPE,
+};
+
+struct tegra_audio_priv {
+	struct asoc_simple_priv simple;
+	struct clk *clk_plla_out0;
+	struct clk *clk_plla;
+};
+
+/* Tegra audio chip data */
+struct tegra_audio_cdata {
+	unsigned int plla_rates[NUM_RATE_TYPE];
+	unsigned int plla_out0_rates[NUM_RATE_TYPE];
+};
+
+/* Setup PLL clock as per the given sample rate */
+static int tegra_audio_graph_update_pll(struct snd_pcm_substream *substream,
+					struct snd_pcm_hw_params *params)
+{
+	struct snd_soc_pcm_runtime *rtd = asoc_substream_to_rtd(substream);
+	struct asoc_simple_priv *simple = snd_soc_card_get_drvdata(rtd->card);
+	struct tegra_audio_priv *priv = simple_to_tegra_priv(simple);
+	struct device *dev = rtd->card->dev;
+	const struct tegra_audio_cdata *data = of_device_get_match_data(dev);
+	unsigned int plla_rate, plla_out0_rate, bclk;
+	unsigned int srate = params_rate(params);
+	int err;
+
+	switch (srate) {
+	case 11025:
+	case 22050:
+	case 44100:
+	case 88200:
+	case 176400:
+		plla_out0_rate = data->plla_out0_rates[x11_RATE];
+		plla_rate = data->plla_rates[x11_RATE];
+		break;
+	case 8000:
+	case 16000:
+	case 32000:
+	case 48000:
+	case 96000:
+	case 192000:
+		plla_out0_rate = data->plla_out0_rates[x8_RATE];
+		plla_rate = data->plla_rates[x8_RATE];
+		break;
+	default:
+		dev_err(rtd->card->dev, "Unsupported sample rate %u\n",
+			srate);
+		return -EINVAL;
+	}
+
+	/*
+	 * Below is the clock relation:
+	 *
+	 *	PLLA
+	 *	  |
+	 *	  |--> PLLA_OUT0
+	 *		  |
+	 *		  |---> I2S modules
+	 *		  |
+	 *		  |---> DMIC modules
+	 *		  |
+	 *		  |---> DSPK modules
+	 *
+	 *
+	 * Default PLLA_OUT0 rate might be too high when I/O is running
+	 * at minimum PCM configurations. This may result in incorrect
+	 * clock rates and glitchy audio. The maximum divider is 128
+	 * and any thing higher than that won't work. Thus reduce PLLA_OUT0
+	 * to work for lower configurations.
+	 *
+	 * This problem is seen for I2S only, as DMIC and DSPK minimum
+	 * clock requirements are under allowed divider limits.
+	 */
+	bclk = srate * params_channels(params) * params_width(params);
+	if (div_u64(plla_out0_rate, bclk) > MAX_PLLA_OUT0_DIV)
+		plla_out0_rate >>= 1;
+
+	dev_dbg(rtd->card->dev,
+		"Update clock rates: PLLA(= %u Hz) and PLLA_OUT0(= %u Hz)\n",
+		plla_rate, plla_out0_rate);
+
+	/* Set PLLA rate */
+	err = clk_set_rate(priv->clk_plla, plla_rate);
+	if (err) {
+		dev_err(rtd->card->dev,
+			"Can't set plla rate for %u, err: %d\n",
+			plla_rate, err);
+		return err;
+	}
+
+	/* Set PLLA_OUT0 rate */
+	err = clk_set_rate(priv->clk_plla_out0, plla_out0_rate);
+	if (err) {
+		dev_err(rtd->card->dev,
+			"Can't set plla_out0 rate %u, err: %d\n",
+			plla_out0_rate, err);
+		return err;
+	}
+
+	return err;
+}
+
+static int tegra_audio_graph_hw_params(struct snd_pcm_substream *substream,
+				       struct snd_pcm_hw_params *params)
+{
+	struct snd_soc_pcm_runtime *rtd = asoc_substream_to_rtd(substream);
+	struct snd_soc_dai *cpu_dai = asoc_rtd_to_cpu(rtd, 0);
+	int err;
+
+	/*
+	 * This gets called for each DAI link (FE or BE) when DPCM is used.
+	 * We may not want to update PLLA rate for each call. So PLLA update
+	 * must be restricted to external I/O links (I2S, DMIC or DSPK) since
+	 * they actually depend on it. I/O modules update their clocks in
+	 * hw_param() of their respective component driver and PLLA rate
+	 * update here helps them to derive appropriate rates.
+	 *
+	 * TODO: When more HW accelerators get added (like sample rate
+	 * converter, volume gain controller etc., which don't really
+	 * depend on PLLA) we need a better way to filter here.
+	 */
+	if (cpu_dai->driver->ops && rtd->dai_link->no_pcm) {
+		err = tegra_audio_graph_update_pll(substream, params);
+		if (err)
+			return err;
+	}
+
+	return asoc_simple_hw_params(substream, params);
+}
+
+static const struct snd_soc_ops tegra_audio_graph_ops = {
+	.startup	= asoc_simple_startup,
+	.shutdown	= asoc_simple_shutdown,
+	.hw_params	= tegra_audio_graph_hw_params,
+};
+
+static int tegra_audio_graph_card_probe(struct snd_soc_card *card)
+{
+	struct asoc_simple_priv *simple = snd_soc_card_get_drvdata(card);
+	struct tegra_audio_priv *priv = simple_to_tegra_priv(simple);
+
+	priv->clk_plla = devm_clk_get(card->dev, "pll_a");
+	if (IS_ERR(priv->clk_plla)) {
+		dev_err(card->dev, "Can't retrieve clk pll_a\n");
+		return PTR_ERR(priv->clk_plla);
+	}
+
+	priv->clk_plla_out0 = devm_clk_get(card->dev, "plla_out0");
+	if (IS_ERR(priv->clk_plla_out0)) {
+		dev_err(card->dev, "Can't retrieve clk plla_out0\n");
+		return PTR_ERR(priv->clk_plla_out0);
+	}
+
+	return graph_card_probe(card);
+}
+
+static int tegra_audio_graph_probe(struct platform_device *pdev)
+{
+	struct tegra_audio_priv *priv;
+	struct device *dev = &pdev->dev;
+	struct snd_soc_card *card;
+
+	priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	card = simple_priv_to_card(&priv->simple);
+
+	card->probe = tegra_audio_graph_card_probe;
+
+	/* graph_parse_of() depends on below */
+	card->component_chaining = 1;
+	priv->simple.ops = &tegra_audio_graph_ops;
+	priv->simple.force_dpcm = 1;
+
+	return graph_parse_of(&priv->simple, dev);
+}
+
+static const struct tegra_audio_cdata tegra210_data = {
+	/* PLLA */
+	.plla_rates[x8_RATE] = 368640000,
+	.plla_rates[x11_RATE] = 338688000,
+	/* PLLA_OUT0 */
+	.plla_out0_rates[x8_RATE] = 49152000,
+	.plla_out0_rates[x11_RATE] = 45158400,
+};
+
+static const struct tegra_audio_cdata tegra186_data = {
+	/* PLLA */
+	.plla_rates[x8_RATE] = 245760000,
+	.plla_rates[x11_RATE] = 270950400,
+	/* PLLA_OUT0 */
+	.plla_out0_rates[x8_RATE] = 49152000,
+	.plla_out0_rates[x11_RATE] = 45158400,
+};
+
+static const struct of_device_id graph_of_tegra_match[] = {
+	{ .compatible = "nvidia,tegra210-audio-graph-card",
+	  .data = &tegra210_data },
+	{ .compatible = "nvidia,tegra186-audio-graph-card",
+	  .data = &tegra186_data },
+	{},
+};
+MODULE_DEVICE_TABLE(of, graph_of_tegra_match);
+
+static struct platform_driver tegra_audio_graph_card = {
+	.driver = {
+		.name = "tegra-audio-graph-card",
+		.pm = &snd_soc_pm_ops,
+		.of_match_table = graph_of_tegra_match,
+	},
+	.probe = tegra_audio_graph_probe,
+};
+module_platform_driver(tegra_audio_graph_card);
+
+MODULE_LICENSE("GPL v2");
+MODULE_DESCRIPTION("ASoC Tegra Audio Graph Sound Card");
+MODULE_AUTHOR("Sameer Pujar <spujar@nvidia.com>");