[RESEND RFC PATCH 0/3] ASoC: Mediatek: Add support for MT8173 SOC

[RESEND RFC PATCH 0/3] ASoC: Mediatek: Add support for MT8173 SOC

[Koro Chen]

Refactor and resend this series because of the 60KB limitation.
 
This adds basic support for the Mediatek AFE unit for MT8173 SoC.
This patch is based on Linux 4.0-rc1, Sascha's clk patch [1], 
and his SCPSYS power patch [2].

The AFE unit comprises several memory interfaces that communicate with CPU,
a multi input multi output digital audio interconnect, and several external
interfaces (e.g. I2S). 

The AFE unit is a DMA master, so no external DMA engine is needed.

Each external interface is presented as a DAI to ASoC. A memory interface 
must be connected via the interconnect to an external interface. 
The connection pathes are configured through the device tree.

[1] http://lists.infradead.org/pipermail/linux-mediatek/2015-March/000079.html
[2] http://lists.infradead.org/pipermail/linux-mediatek/2015-March/000132.html

Koro Chen (3):
  ASoC: mediatek: Add binding support for AFE driver
  ASoC: mediatek: Add AFE connection control
  ASoC: mediatek: Add AFE platform driver

 .../devicetree/bindings/sound/mtk-afe-pcm.txt      |  105 ++
 include/dt-bindings/sound/mtk-afe.h                |   36 +
 sound/soc/Kconfig                                  |    1 +
 sound/soc/Makefile                                 |    1 +
 sound/soc/mediatek/Kconfig                         |    8 +
 sound/soc/mediatek/Makefile                        |    2 +
 sound/soc/mediatek/mtk-afe-common.h                |  105 ++
 sound/soc/mediatek/mtk-afe-connection.c            |  416 ++++++
 sound/soc/mediatek/mtk-afe-connection.h            |   30 +
 sound/soc/mediatek/mtk-afe-pcm.c                   | 1497 ++++++++++++++++++++
 10 files changed, 2201 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/sound/mtk-afe-pcm.txt
 create mode 100644 include/dt-bindings/sound/mtk-afe.h
 create mode 100644 sound/soc/mediatek/Kconfig
 create mode 100644 sound/soc/mediatek/Makefile
 create mode 100644 sound/soc/mediatek/mtk-afe-common.h
 create mode 100644 sound/soc/mediatek/mtk-afe-connection.c
 create mode 100644 sound/soc/mediatek/mtk-afe-connection.h
 create mode 100644 sound/soc/mediatek/mtk-afe-pcm.c

--

[RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Koro Chen]

Add documentation and header file to support binding of Mediatek's AFE driver

Signed-off-by: Koro Chen <koro.chen@mediatek.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
---
 .../devicetree/bindings/sound/mtk-afe-pcm.txt      | 105 +++++++++++++++++++++
 include/dt-bindings/sound/mtk-afe.h                |  36 +++++++
 2 files changed, 141 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/sound/mtk-afe-pcm.txt
 create mode 100644 include/dt-bindings/sound/mtk-afe.h

diff --git a/Documentation/devicetree/bindings/sound/mtk-afe-pcm.txt b/Documentation/devicetree/bindings/sound/mtk-afe-pcm.txt
new file mode 100644
index 0000000..11fc8ba
--- /dev/null
+++ b/Documentation/devicetree/bindings/sound/mtk-afe-pcm.txt
@@ -0,0 +1,105 @@
+Mediatek AFE PCM controller
+
+The AFE unit can be illustrated by this figure:
+
+| MEMIF    |      AFE      |     IO     |
+           *****************
+DL1------> *I5           O3* <----I2S--->
+           *I6           O4*
+           *             I3*
+           *             I4*
+           *             O0* <--2ndI2S-->
+           *             O1*
+           *             I0*
+           *             I1*
+AWB<------ *O5             * <--MTKIF--->
+           *O6             *
+VUL<------ *O9             * <---HDMI--->
+           *O10            *
+           *      AFE      *
+HDMI-----> * inter-connect *
+           *****************
+AFE comprises several memory interfaces (DL1, DL2, VUL, DAI, AWB, MOD_DAI
+and HDMI) that communicate with CPU, a multi input multi output digital audio
+interconnect, and several external interfaces (I2S, proprietary MTKIF, HDMI).
+Each external interface (called "IO" in this driver) is presented as a
+DAI to ASoC. An IO must be connected via the interconnect to a memif.
+The connection paths are configured through the device tree.
+
+Required properties:
+- compatible = "mediatek,mt8173-afe-pcm";
+- reg: array of register and sram location and size:
+       <register base address, size>,
+       <sram base address, size>;
+- interrupts: Should contain AFE interrupt
+- clock-names: should have these clock names:
+		"infra_sys_audio_clk",
+		"top_pdn_audio",
+		"top_pdn_aud_intbus",
+		"bck0",
+		"bck1",
+		"i2s0_m",
+		"i2s1_m",
+		"i2s2_m",
+		"i2s3_m",
+		"i2s3_b";
+
+DAI subnodes:
+  A DAI subnode describes which io connects to which memif.
+
+Required subnode properties:
+- io: which I/O to be used
+      (defined in include/dt-bindings/sound/mtk-afe.h)
+- connections: AFE connection pairs definition of this dai
+	       For example, <5 3 6 4> means I5->O3, I6->O4
+	       check SoC datasheet for a complete description
+- mem-interface-playback:
+  mem-interface-capture: property of memif, format is: <memif irq use_sram>;
+	                 memif: which memif to be used
+			        (defined in include/dt-bindings/sound/mtk-afe.h)
+		         irq: which irq to be used
+			      (defined in include/dt-bindings/sound/mtk-afe.h)
+		         use_sram: 1 is yes, 0 is no
+
+  Each DAI should describes at least playback or capture
+
+Example:
+
+	afe: mt8173-afe-pcm@11220000  {
+		compatible = "mediatek,mt8173-afe-pcm";
+		reg = <0 0x11220000 0 0x1000>,
+		      <0 0x11221000 0 0x9000>;
+		interrupts = <GIC_SPI 134 IRQ_TYPE_EDGE_FALLING>;
+		clocks = <&infracfg INFRA_AUDIO>,
+			<&topckgen TOP_AUDIO_SEL>,
+			<&topckgen TOP_AUD_INTBUS_SEL>,
+			<&topckgen TOP_APLL1_DIV0>,
+			<&topckgen TOP_APLL2_DIV0>,
+			<&topckgen TOP_I2S0_M_CK_SEL>,
+			<&topckgen TOP_I2S1_M_CK_SEL>,
+			<&topckgen TOP_I2S2_M_CK_SEL>,
+			<&topckgen TOP_I2S3_M_CK_SEL>,
+			<&topckgen TOP_I2S3_B_CK_SEL>;
+		clock-names = "infra_sys_audio_clk",
+				"top_pdn_audio",
+				"top_pdn_aud_intbus",
+				"bck0",
+				"bck1",
+				"i2s0_m",
+				"i2s1_m",
+				"i2s2_m",
+				"i2s3_m",
+				"i2s3_b";
+		dai@0 {
+			io = <MTK_AFE_IO_I2S>;
+			connections = <5 3 6 4 3 9 4 10>;
+			mem-interface-playback = <MTK_AFE_MEMIF_DL1 MTK_AFE_IRQ_1 1>;
+			mem-interface-capture = <MTK_AFE_MEMIF_VUL MTK_AFE_IRQ_2 0>;
+		};
+
+		dai@1 {
+			io = <MTK_AFE_IO_HDMI>;
+			connections = <36 36 37 37 34 32 35 33 32 34 33 35 30 30 31 31>;
+			mem-interface-playback = <MTK_AFE_MEMIF_HDMI MTK_AFE_IRQ_5 0>;
+		};
+	};
diff --git a/include/dt-bindings/sound/mtk-afe.h b/include/dt-bindings/sound/mtk-afe.h
new file mode 100644
index 0000000..e6da18e
--- /dev/null
+++ b/include/dt-bindings/sound/mtk-afe.h
@@ -0,0 +1,36 @@
+#ifndef __DT_MTK_AFE_H
+#define __DT_MTK_AFE_H
+
+#define MTK_AFE_MEMIF_DL1		0
+#define MTK_AFE_MEMIF_DL2		1
+#define MTK_AFE_MEMIF_VUL		2
+#define MTK_AFE_MEMIF_DAI		3
+#define MTK_AFE_MEMIF_AWB		4
+#define MTK_AFE_MEMIF_MOD_DAI		5
+#define MTK_AFE_MEMIF_HDMI		6
+#define MTK_AFE_MEMIF_NUM		7
+
+#define MTK_AFE_IO_MOD_PCM1		0 /* connection to int main modem */
+#define MTK_AFE_IO_MOD_PCM2		1 /* connection to extrt/int modem */
+#define MTK_AFE_IO_PMIC		2 /* MTKIF for DAC and ADC */
+#define MTK_AFE_IO_I2S			3 /* I2S */
+#define MTK_AFE_IO_2ND_I2S		4 /* 2nd I2S */
+#define MTK_AFE_IO_HW_GAIN1		5 /* HW gain control */
+#define MTK_AFE_IO_HW_GAIN2		6
+#define MTK_AFE_IO_MRG_O		7 /* merge interface */
+#define MTK_AFE_IO_MRG_I		8
+#define MTK_AFE_IO_DAIBT		9
+#define MTK_AFE_IO_HDMI		10
+#define MTK_AFE_IO_NUM			11
+
+#define MTK_AFE_IRQ_1			0
+#define MTK_AFE_IRQ_2			1
+#define MTK_AFE_IRQ_3			2
+#define MTK_AFE_IRQ_4			3
+#define MTK_AFE_IRQ_5			4
+#define MTK_AFE_IRQ_6			5
+#define MTK_AFE_IRQ_7			6
+#define MTK_AFE_IRQ_8			7
+#define MTK_AFE_IRQ_NUM			8
+
+#endif /* __DT_MTK_AFE_H */
-- 
1.8.1.1.dirty

[RESEND RFC PATCH 2/3] ASoC: mediatek: Add AFE connection control

[Koro Chen]

This is the AFE inter-connection control APIs.

Signed-off-by: Koro Chen <koro.chen@mediatek.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
---
 sound/soc/mediatek/mtk-afe-common.h     | 105 ++++++++
 sound/soc/mediatek/mtk-afe-connection.c | 416 ++++++++++++++++++++++++++++++++
 sound/soc/mediatek/mtk-afe-connection.h |  30 +++
 3 files changed, 551 insertions(+)
 create mode 100644 sound/soc/mediatek/mtk-afe-common.h
 create mode 100644 sound/soc/mediatek/mtk-afe-connection.c
 create mode 100644 sound/soc/mediatek/mtk-afe-connection.h

diff --git a/sound/soc/mediatek/mtk-afe-common.h b/sound/soc/mediatek/mtk-afe-common.h
new file mode 100644
index 0000000..71b426d
--- /dev/null
+++ b/sound/soc/mediatek/mtk-afe-common.h
@@ -0,0 +1,105 @@
+/*
+ * mtk_afe_common.h  --  Mediatek audio driver common definitions
+ *
+ * Copyright (c) 2015 MediaTek Inc.
+ * Author: Koro Chen <koro.chen@mediatek.com>
+ *             Sascha Hauer <s.hauer@pengutronix.de>
+ *             Hidalgo Huang <hidalgo.huang@mediatek.com>
+ *             Ir Lian <ir.lian@mediatek.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#ifndef _MTK_AFE_COMMON_H_
+#define _MTK_AFE_COMMON_H_
+
+#include <linux/clk.h>
+#include <linux/regmap.h>
+#include <dt-bindings/sound/mtk-afe.h>
+
+enum {
+	MTK_CLK_INFRASYS_AUD,
+	MTK_CLK_TOP_PDN_AUD,
+	MTK_CLK_TOP_PDN_AUD_BUS,
+	MTK_CLK_I2S0_M,
+	MTK_CLK_I2S1_M,
+	MTK_CLK_I2S2_M,
+	MTK_CLK_I2S3_M,
+	MTK_CLK_I2S3_B,
+	MTK_CLK_BCK0,
+	MTK_CLK_BCK1,
+	MTK_CLK_NUM
+};
+
+struct mtk_afe;
+struct snd_pcm_substream;
+
+struct mtk_afe_io_data {
+	int num;
+	const char *name;
+	int (*startup)(struct mtk_afe *, struct snd_pcm_substream *);
+	void (*shutdown)(struct mtk_afe *, struct snd_pcm_substream *);
+	int (*prepare)(struct mtk_afe *, struct snd_pcm_substream *);
+	int (*start)(struct mtk_afe *, struct snd_pcm_substream *);
+	void (*pause)(struct mtk_afe *, struct snd_pcm_substream *);
+};
+
+struct mtk_afe_io {
+	const struct mtk_afe_io_data *data;
+	struct clk *m_ck;
+	struct clk *b_ck;
+	u32 *connections;
+	int num_connections;
+	int mem[2]; /* playback and capture */
+};
+
+struct mtk_afe_irq_data {
+	int reg_cnt;
+	int cnt_shift;
+	int en_shift;
+	int fs_shift;
+	int clr_shift;
+};
+
+struct mtk_afe_memif_data {
+	int id;
+	const char *name;
+	int reg_ofs_base;
+	int reg_ofs_cur;
+	int fs_shift;
+	int mono_shift;
+	int enable_shift;
+};
+
+struct mtk_afe_memif {
+	unsigned int phys_buf_addr;
+	int buffer_size;
+	unsigned int hw_ptr;		/* Previous IRQ's HW ptr */
+	bool use_sram;
+	struct snd_pcm_substream *substream;
+	const struct mtk_afe_memif_data *data;
+	const struct mtk_afe_irq_data *irqdata;
+};
+
+struct mtk_afe {
+	/* address for ioremap audio hardware register */
+	void __iomem *base_addr;
+	void __iomem *sram_address;
+	u32 sram_phy_address;
+	u32 sram_size;
+	struct device *dev;
+	struct regmap *regmap;
+
+	struct mtk_afe_memif memif[MTK_AFE_MEMIF_NUM];
+	struct mtk_afe_io ios[MTK_AFE_IO_NUM];
+
+	struct clk *clocks[MTK_CLK_NUM];
+};
+#endif
diff --git a/sound/soc/mediatek/mtk-afe-connection.c b/sound/soc/mediatek/mtk-afe-connection.c
new file mode 100644
index 0000000..714e9cd
--- /dev/null
+++ b/sound/soc/mediatek/mtk-afe-connection.c
@@ -0,0 +1,416 @@
+/*
+ * Mediatek AFE audio interconnect support
+ *
+ * Copyright (c) 2015 MediaTek Inc.
+ * Author: Koro Chen <koro.chen@mediatek.com>
+ *             Sascha Hauer <s.hauer@pengutronix.de>
+ *             Hidalgo Huang <hidalgo.huang@mediatek.com>
+ *             Ir Lian <ir.lian@mediatek.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/regmap.h>
+#include <linux/device.h>
+#include "mtk-afe-common.h"
+#include "mtk-afe-connection.h"
+
+#define MTK_AFE_INTERCONN_NUM_INPUT	21
+#define MTK_AFE_INTERCONN_NUM_OUTPUT	23
+
+#define MTK_AFE_HDMI_CONN_INPUT_BASE	30
+#define MTK_AFE_HDMI_CONN_INPUT_MAX	37
+#define MTK_AFE_NUM_HDMI_INPUT		(37 - 30 + 1)
+
+#define MTK_AFE_HDMI_CONN_OUTPUT_BASE	30
+#define MTK_AFE_HDMI_CONN_OUTPUT_MAX	41
+#define MTK_AFE_NUM_HDMI_OUTPUT		(41 - 30 + 1)
+
+struct mtk_afe_connection {
+	short creg, sreg;
+	char cshift, sshift;
+};
+
+/*
+ * The MTK AFE unit has a audio interconnect with MTK_AFE_INTERCONN_NUM_INPUT
+ * inputs and MTK_AFE_INTERCONN_NUM_OUTPUT outputs. Below table holds the
+ * register/bits to set to connect an input with an output.
+ */
+static const struct mtk_afe_connection
+	connections[MTK_AFE_INTERCONN_NUM_INPUT][MTK_AFE_INTERCONN_NUM_OUTPUT] = {
+	[0][0] =   { .creg = 0x020, .cshift =  0, .sreg = 0x020, .sshift = 10},
+	[0][1] =   { .creg = 0x020, .cshift = 16, .sreg = 0x020, .sshift = 26},
+	[0][2] =   { .creg = 0x024, .cshift =  0, .sreg = 0x024, .sshift = 10},
+	[0][3] =   { .creg = 0x024, .cshift = 16, .sreg = 0x024, .sshift = 26},
+	[0][4] =   { .creg = 0x028, .cshift =  0, .sreg = 0x028, .sshift = 10},
+	[0][5] =   { .creg = 0x028, .cshift = 16, .sreg = 0x030, .sshift = 19},
+	[0][7] =   { .creg = 0x05c, .cshift =  2, },
+	[0][9] =   { .creg = 0x05c, .cshift =  8, },
+	[0][10] =  { .creg = 0x05c, .cshift = 12, },
+	[0][13] =  { .creg = 0x448, .cshift =  2, },
+	[0][14] =  { .creg = 0x448, .cshift = 15, },
+	[0][15] =  { .creg = 0x438, .cshift = 16, .sreg = 0x438, .sshift = 31},
+	[0][16] =  { .creg = 0x438, .cshift = 22, .sreg = 0x440, .sshift = 25},
+	[0][19] =  { .creg = 0x464, .cshift =  8, .sreg = 0x464, .sshift =  9},
+	[0][20] =  { .creg = 0x464, .cshift = 24, .sreg = 0x464, .sshift = 25},
+	[1][0] =   { .creg = 0x020, .cshift =  1, .sreg = 0x020, .sshift = 11},
+	[1][1] =   { .creg = 0x020, .cshift = 17, .sreg = 0x020, .sshift = 27},
+	[1][2] =   { .creg = 0x024, .cshift =  1, .sreg = 0x024, .sshift = 11},
+	[1][3] =   { .creg = 0x024, .cshift = 17, .sreg = 0x024, .sshift = 27},
+	[1][4] =   { .creg = 0x028, .cshift =  1, .sreg = 0x028, .sshift = 11},
+	[1][6] =   { .creg = 0x028, .cshift = 22, .sreg = 0x030, .sshift = 20},
+	[1][7] =   { .creg = 0x05c, .cshift =  3, },
+	[1][8] =   { .creg = 0x05c, .cshift =  6, },
+	[1][9] =   { .creg = 0x05c, .cshift =  9, },
+	[1][10] =  { .creg = 0x05c, .cshift = 13, },
+	[1][13] =  { .creg = 0x448, .cshift =  3, },
+	[1][14] =  { .creg = 0x448, .cshift = 16, },
+	[1][15] =  { .creg = 0x438, .cshift = 17, .sreg = 0x440, .sshift = 16},
+	[1][16] =  { .creg = 0x438, .cshift = 23, .sreg = 0x440, .sshift =  4},
+	[1][19] =  { .creg = 0x464, .cshift = 10, .sreg = 0x464, .sshift = 11},
+	[1][20] =  { .creg = 0x464, .cshift = 26, .sreg = 0x464, .sshift = 27},
+	[1][22] =  { .creg = 0x0bc, .cshift =  2, },
+	[2][1] =   { .creg = 0x020, .cshift = 18, },
+	[2][2] =   { .creg = 0x024, .cshift =  2, },
+	[2][3] =   { .creg = 0x024, .cshift = 18, },
+	[2][7] =   { .creg = 0x030, .cshift = 21, },
+	[2][11] =  { .creg = 0x02c, .cshift =  6, },
+	[2][13] =  { .creg = 0x448, .cshift =  4, },
+	[2][14] =  { .creg = 0x448, .cshift = 17, },
+	[3][0] =   { .creg = 0x020, .cshift =  3, },
+	[3][1] =   { .creg = 0x020, .cshift = 19, },
+	[3][2] =   { .creg = 0x024, .cshift =  3, .sreg = 0x030, .sshift = 25},
+	[3][3] =   { .creg = 0x024, .cshift = 19, },
+	[3][4] =   { .creg = 0x028, .cshift =  3, },
+	[3][5] =   { .creg = 0x028, .cshift = 18, },
+	[3][7] =   { .creg = 0x028, .cshift = 26, },
+	[3][9] =   { .creg = 0x02c, .cshift =  0, },
+	[3][13] =  { .creg = 0x448, .cshift =  5, },
+	[3][14] =  { .creg = 0x448, .cshift = 18, },
+	[3][15] =  { .creg = 0x438, .cshift = 19, },
+	[3][16] =  { .creg = 0x438, .cshift = 25, },
+	[3][19] =  { .creg = 0x464, .cshift = 12, },
+	[3][20] =  { .creg = 0x464, .cshift = 28, },
+	[3][21] =  { .creg = 0x05c, .cshift = 31, },
+	[4][0] =   { .creg = 0x020, .cshift =  4, },
+	[4][1] =   { .creg = 0x020, .cshift = 20, },
+	[4][2] =   { .creg = 0x024, .cshift =  4, .sreg = 0x030, .sshift = 26},
+	[4][3] =   { .creg = 0x024, .cshift = 20, },
+	[4][4] =   { .creg = 0x028, .cshift =  4, },
+	[4][6] =   { .creg = 0x028, .cshift = 23, },
+	[4][8] =   { .creg = 0x028, .cshift = 29, },
+	[4][10] =  { .creg = 0x02c, .cshift =  3, },
+	[4][13] =  { .creg = 0x448, .cshift =  6, },
+	[4][14] =  { .creg = 0x448, .cshift = 19, },
+	[4][15] =  { .creg = 0x438, .cshift = 20, },
+	[4][16] =  { .creg = 0x438, .cshift = 26, },
+	[4][19] =  { .creg = 0x464, .cshift = 13, },
+	[4][20] =  { .creg = 0x464, .cshift = 29, },
+	[4][22] =  { .creg = 0x0bc, .cshift =  3, },
+	[5][0] =   { .creg = 0x020, .cshift =  5, .sreg = 0x020, .sshift = 12},
+	[5][1] =   { .creg = 0x020, .cshift = 21, .sreg = 0x020, .sshift = 28},
+	[5][2] =   { .creg = 0x024, .cshift =  5, .sreg = 0x024, .sshift = 12},
+	[5][3] =   { .creg = 0x024, .cshift = 21, .sreg = 0x024, .sshift = 28},
+	[5][4] =   { .creg = 0x028, .cshift =  5, .sreg = 0x028, .sshift = 12},
+	[5][5] =   { .creg = 0x028, .cshift = 19, },
+	[5][7] =   { .creg = 0x028, .cshift = 27, },
+	[5][9] =   { .creg = 0x02c, .cshift =  1, },
+	[5][13] =  { .creg = 0x420, .cshift = 16, },
+	[5][14] =  { .creg = 0x420, .cshift = 20, },
+	[5][19] =  { .creg = 0x464, .cshift = 14, .sreg = 0x464, .sshift = 15},
+	[5][20] =  { .creg = 0x464, .cshift = 31, .sreg = 0x464, .sshift = 30},
+	[6][0] =   { .creg = 0x020, .cshift =  6, .sreg = 0x020, .sshift = 13},
+	[6][1] =   { .creg = 0x020, .cshift = 22, .sreg = 0x020, .sshift = 29},
+	[6][2] =   { .creg = 0x024, .cshift =  6, .sreg = 0x024, .sshift = 13},
+	[6][3] =   { .creg = 0x024, .cshift = 22, .sreg = 0x024, .sshift = 29},
+	[6][4] =   { .creg = 0x028, .cshift =  6, .sreg = 0x028, .sshift = 13},
+	[6][6] =   { .creg = 0x028, .cshift = 24, },
+	[6][8] =   { .creg = 0x028, .cshift = 30, },
+	[6][10] =  { .creg = 0x02c, .cshift =  4, },
+	[6][12] =  { .creg = 0x02c, .cshift =  9, },
+	[6][13] =  { .creg = 0x420, .cshift = 17, },
+	[6][14] =  { .creg = 0x420, .cshift = 21, },
+	[6][19] =  { .creg = 0x464, .cshift = 16, .sreg = 0x464, .sshift = 17},
+	[7][0] =   { .creg = 0x020, .cshift =  7, .sreg = 0x020, .sshift = 14},
+	[7][1] =   { .creg = 0x020, .cshift = 23, .sreg = 0x020, .sshift = 30},
+	[7][2] =   { .creg = 0x024, .cshift =  7, .sreg = 0x024, .sshift = 14},
+	[7][3] =   { .creg = 0x024, .cshift = 23, .sreg = 0x024, .sshift = 30},
+	[7][4] =   { .creg = 0x028, .cshift =  7, .sreg = 0x028, .sshift = 14},
+	[7][5] =   { .creg = 0x028, .cshift = 20, },
+	[7][7] =   { .creg = 0x028, .cshift = 28, },
+	[7][9] =   { .creg = 0x02c, .cshift =  2, },
+	[7][13] =  { .creg = 0x420, .cshift = 18, },
+	[7][14] =  { .creg = 0x420, .cshift = 22, },
+	[7][19] =  { .creg = 0x464, .cshift = 18, .sreg = 0x464, .sshift = 19},
+	[8][0] =   { .creg = 0x020, .cshift =  8, .sreg = 0x020, .sshift = 15},
+	[8][1] =   { .creg = 0x020, .cshift = 24, .sreg = 0x020, .sshift = 31},
+	[8][2] =   { .creg = 0x024, .cshift =  8, .sreg = 0x024, .sshift = 15},
+	[8][3] =   { .creg = 0x024, .cshift = 24, .sreg = 0x024, .sshift = 31},
+	[8][4] =   { .creg = 0x028, .cshift =  8, .sreg = 0x028, .sshift = 15},
+	[8][6] =   { .creg = 0x028, .cshift = 25, },
+	[8][8] =   { .creg = 0x028, .cshift = 31, },
+	[8][10] =  { .creg = 0x02c, .cshift =  5, },
+	[8][12] =  { .creg = 0x02c, .cshift = 10, },
+	[8][13] =  { .creg = 0x420, .cshift = 19, },
+	[8][14] =  { .creg = 0x420, .cshift = 23, },
+	[8][19] =  { .creg = 0x464, .cshift = 20, .sreg = 0x464, .sshift = 21},
+	[9][0] =   { .creg = 0x020, .cshift =  9, },
+	[9][1] =   { .creg = 0x020, .cshift = 25, },
+	[9][2] =   { .creg = 0x024, .cshift =  9, },
+	[9][3] =   { .creg = 0x024, .cshift = 25, },
+	[9][4] =   { .creg = 0x028, .cshift =  9, },
+	[9][5] =   { .creg = 0x028, .cshift = 21, },
+	[9][9] =   { .creg = 0x05c, .cshift = 10, },
+	[9][12] =  { .creg = 0x02c, .cshift = 11, },
+	[9][13] =  { .creg = 0x448, .cshift =  7, },
+	[9][14] =  { .creg = 0x448, .cshift = 20, },
+	[9][15] =  { .creg = 0x438, .cshift = 21, },
+	[9][16] =  { .creg = 0x438, .cshift = 27, },
+	[9][19] =  { .creg = 0x05c, .cshift = 22, },
+	[9][20] =  { .creg = 0x05c, .cshift =  6, },
+	[10][0] =  { .creg = 0x420, .cshift =  0, .sreg = 0x420, .sshift =  1},
+	[10][3] =  { .creg = 0x420, .cshift =  8, .sreg = 0x420, .sshift =  9},
+	[10][5] =  { .creg = 0x420, .cshift = 12, },
+	[10][7] =  { .creg = 0x420, .cshift = 14, },
+	[10][12] = { .creg = 0x448, .cshift =  0, },
+	[10][19] = { .creg = 0x448, .cshift = 28, },
+	[10][21] = { .creg = 0x448, .cshift =  0, },
+	[10][22] = { .creg = 0x44c, .cshift =  0, },
+	[11][1] =  { .creg = 0x420, .cshift =  2, .sreg = 0x420, .sshift =  3},
+	[11][4] =  { .creg = 0x420, .cshift = 10, .sreg = 0x420, .sshift = 11},
+	[11][6] =  { .creg = 0x420, .cshift = 13, },
+	[11][8] =  { .creg = 0x420, .cshift = 15, },
+	[11][12] = { .creg = 0x448, .cshift =  1, },
+	[11][20] = { .creg = 0x448, .cshift = 29, },
+	[11][21] = { .creg = 0x448, .cshift = 31, },
+	[11][22] = { .creg = 0x44c, .cshift =  1, },
+	[12][0] =  { .creg = 0x438, .cshift =  0, .sreg = 0x438, .sshift =  1},
+	[12][3] =  { .creg = 0x438, .cshift =  8, .sreg = 0x438, .sshift =  9},
+	[12][5] =  { .creg = 0x438, .cshift = 12, },
+	[12][7] =  { .creg = 0x438, .cshift = 14, },
+	[12][19] = { .creg = 0x444, .cshift =  2, },
+	[12][21] = { .creg = 0x444, .cshift =  4, },
+	[12][22] = { .creg = 0x444, .cshift =  6, },
+	[13][1] =  { .creg = 0x438, .cshift =  2, .sreg = 0x438, .sshift =  3},
+	[13][4] =  { .creg = 0x438, .cshift = 10, .sreg = 0x438, .sshift = 11},
+	[13][6] =  { .creg = 0x438, .cshift = 13, },
+	[13][8] =  { .creg = 0x438, .cshift = 15, },
+	[13][20] = { .creg = 0x444, .cshift =  3, },
+	[13][21] = { .creg = 0x444, .cshift =  4, },
+	[13][22] = { .creg = 0x444, .cshift =  7, },
+	[15][0] =  { .creg = 0x02c, .cshift = 13, .sreg = 0x02c, .sshift = 15},
+	[15][1] =  { .creg = 0x02c, .cshift = 18, .sreg = 0x02c, .sshift = 20},
+	[15][3] =  { .creg = 0x02c, .cshift = 28, .sreg = 0x02c, .sshift = 30},
+	[15][4] =  { .creg = 0x030, .cshift =  1, .sreg = 0x030, .sshift =  3},
+	[15][5] =  { .creg = 0x030, .cshift =  6, .sreg = 0x030, .sshift =  7},
+	[15][9] =  { .creg = 0x030, .cshift = 10, },
+	[15][13] = { .creg = 0x448, .cshift =  9, },
+	[15][14] = { .creg = 0x448, .cshift = 22, },
+	[15][15] = { .creg = 0x438, .cshift = 29, .sreg = 0x440, .sshift =  0},
+	[15][16] = { .creg = 0x440, .cshift =  2, },
+	[16][0] =  { .creg = 0x02c, .cshift = 14, .sreg = 0x02c, .sshift = 16},
+	[16][1] =  { .creg = 0x02c, .cshift = 19, .sreg = 0x02c, .sshift = 21},
+	[16][2] =  { .creg = 0x02c, .cshift = 24, .sreg = 0x02c, .sshift = 26},
+	[16][3] =  { .creg = 0x02c, .cshift = 29, .sreg = 0x02c, .sshift = 31},
+	[16][4] =  { .creg = 0x030, .cshift =  2, .sreg = 0x030, .sshift =  4},
+	[16][6] =  { .creg = 0x030, .cshift =  8, .sreg = 0x030, .sshift =  9},
+	[16][10] = { .creg = 0x030, .cshift = 11, },
+	[16][13] = { .creg = 0x448, .cshift = 10, },
+	[16][14] = { .creg = 0x448, .cshift = 23, },
+	[16][15] = { .creg = 0x438, .cshift = 30, },
+	[16][16] = { .creg = 0x440, .cshift =  3, .sreg = 0x440, .sshift =  5},
+	[17][0] =  { .creg = 0x460, .cshift =  0, },
+	[17][2] =  { .creg = 0x02c, .cshift = 27, .sreg = 0x0bc, .sshift = 30},
+	[17][3] =  { .creg = 0x05c, .cshift =  0, },
+	[17][5] =  { .creg = 0x460, .cshift = 22, },
+	[17][7] =  { .creg = 0x460, .cshift = 26, },
+	[17][9] =  { .creg = 0x460, .cshift = 30, },
+	[17][11] = { .creg = 0x464, .cshift =  2, },
+	[17][13] = { .creg = 0x448, .cshift = 11, },
+	[17][14] = { .creg = 0x448, .cshift = 24, },
+	[17][15] = { .creg = 0x440, .cshift = 21, },
+	[17][16] = { .creg = 0x440, .cshift = 30, },
+	[17][19] = { .creg = 0x464, .cshift = 22, },
+	[17][21] = { .creg = 0x0bc, .cshift =  0, },
+	[18][1] =  { .creg = 0x460, .cshift =  3, },
+	[18][2] =  { .creg = 0x02c, .cshift = 28, .sreg = 0x0bc, .sshift = 31},
+	[18][4] =  { .creg = 0x05c, .cshift =  1, },
+	[18][6] =  { .creg = 0x460, .cshift = 24, },
+	[18][8] =  { .creg = 0x460, .cshift = 28, },
+	[18][10] = { .creg = 0x464, .cshift =  0, },
+	[18][12] = { .creg = 0x464, .cshift =  4, },
+	[18][13] = { .creg = 0x448, .cshift = 12, },
+	[18][14] = { .creg = 0x448, .cshift = 25, },
+	[18][15] = { .creg = 0x440, .cshift = 22, },
+	[18][16] = { .creg = 0x440, .cshift = 31, },
+	[18][19] = { .creg = 0x464, .cshift = 23, },
+	[18][22] = { .creg = 0x0bc, .cshift =  4, },
+	[19][0] =  { .creg = 0x460, .cshift =  1, .sreg = 0x460, .sshift =  2},
+	[19][2] =  { .creg = 0x460, .cshift = 10, .sreg = 0x460, .sshift = 11},
+	[19][3] =  { .creg = 0x460, .cshift = 14, .sreg = 0x460, .sshift = 16},
+	[19][4] =  { .creg = 0x460, .cshift = 18, .sreg = 0x460, .sshift = 19},
+	[19][5] =  { .creg = 0x460, .cshift = 23, },
+	[19][7] =  { .creg = 0x460, .cshift = 27, },
+	[19][9] =  { .creg = 0x460, .cshift = 31, },
+	[19][11] = { .creg = 0x464, .cshift =  3, },
+	[19][13] = { .creg = 0x448, .cshift = 13, },
+	[19][14] = { .creg = 0x448, .cshift = 26, },
+	[19][15] = { .creg = 0x440, .cshift = 23, },
+	[19][16] = { .creg = 0x444, .cshift =  0, },
+	[19][19] = { .creg = 0x05c, .cshift = 24, },
+	[19][20] = { .creg = 0x05c, .cshift = 28, },
+	[19][21] = { .creg = 0x0bc, .cshift =  1, },
+	[20][1] =  { .creg = 0x460, .cshift =  4, .sreg = 0x460, .sshift =  5},
+	[20][2] =  { .creg = 0x460, .cshift = 12, .sreg = 0x460, .sshift = 13},
+	[20][3] =  { .creg = 0x460, .cshift = 16, .sreg = 0x460, .sshift = 17},
+	[20][4] =  { .creg = 0x460, .cshift = 20, .sreg = 0x460, .sshift = 21},
+	[20][6] =  { .creg = 0x460, .cshift = 25, },
+	[20][8] =  { .creg = 0x460, .cshift = 29, },
+	[20][10] = { .creg = 0x464, .cshift =  1, },
+	[20][12] = { .creg = 0x464, .cshift =  5, },
+	[20][13] = { .creg = 0x448, .cshift = 14, },
+	[20][14] = { .creg = 0x448, .cshift = 27, },
+	[20][15] = { .creg = 0x440, .cshift = 24, },
+	[20][16] = { .creg = 0x444, .cshift =  1, },
+	[20][19] = { .creg = 0x05c, .cshift = 25, },
+	[20][20] = { .creg = 0x05c, .cshift = 29, },
+	[20][22] = { .creg = 0x0bc, .cshift =  5, },
+};
+
+struct mtk_afe_hdmi_connection {
+	short reg;
+	char shift;
+};
+
+static const struct mtk_afe_hdmi_connection
+	hdmi_connections[MTK_AFE_NUM_HDMI_OUTPUT] = {
+	{ .reg = 0x390, .shift = 0 },
+	{ .reg = 0x390, .shift = 3 },
+	{ .reg = 0x390, .shift = 6 },
+	{ .reg = 0x390, .shift = 9 },
+
+	{ .reg = 0x390, .shift = 12 },
+	{ .reg = 0x390, .shift = 15 },
+	{ .reg = 0x390, .shift = 18 },
+	{ .reg = 0x390, .shift = 21 },
+
+	{ .reg = 0x390, .shift = 24 },
+	{ .reg = 0x390, .shift = 27 },
+	{ .reg = 0x398, .shift = 0 },
+	{ .reg = 0x398, .shift = 2 },
+};
+
+static int mtk_afe_interconn_hdmi(struct mtk_afe *afe_info, u32 in,
+				  u32 out)
+{
+	const struct mtk_afe_hdmi_connection *con;
+
+	if (in < MTK_AFE_HDMI_CONN_INPUT_BASE ||
+	    out < MTK_AFE_HDMI_CONN_OUTPUT_BASE)
+		return -EINVAL;
+
+	in -= MTK_AFE_HDMI_CONN_INPUT_BASE;
+	out -= MTK_AFE_HDMI_CONN_OUTPUT_BASE;
+
+	if (out >= MTK_AFE_NUM_HDMI_OUTPUT)
+		return -EINVAL;
+
+	if (in >= MTK_AFE_NUM_HDMI_INPUT)
+		return -EINVAL;
+
+	con = &hdmi_connections[out];
+
+	regmap_update_bits(afe_info->regmap, con->reg,
+			   0x7 << con->shift, in << con->shift);
+	return 0;
+}
+
+static const struct mtk_afe_connection *mtk_afe_get_connection(
+		struct mtk_afe *afe_info, u32 in, u32 out)
+{
+	if (in >= MTK_AFE_INTERCONN_NUM_INPUT ||
+	    out >= MTK_AFE_INTERCONN_NUM_OUTPUT) {
+		dev_err(afe_info->dev,
+			"out of bound mpConnectionTable[%d][%d]\n", in, out);
+		return NULL;
+	}
+
+	if (connections[in][out].creg == 0) {
+		dev_err(afe_info->dev,
+			"No connection between I%02d and O%02d\n", in, out);
+		return NULL;
+	}
+
+	return &connections[in][out];
+}
+
+/*
+ * mtk_afe_interconn_connect - Connect an input with an output
+ * @afe_info:	Context
+ * @in:		Input number, as in the SoC datasheet
+ * @out:	Output number, as in the SoC datasheet
+ * @rightshift:	Apply a rightshift on the input data
+ *
+ * This function connects an input of the audio interconnect with an
+ * output.
+ */
+int mtk_afe_interconn_connect(struct mtk_afe *afe_info, unsigned int in,
+			      unsigned int out, bool rightshift)
+{
+	const struct mtk_afe_connection *con;
+
+	if (in >= MTK_AFE_HDMI_CONN_INPUT_BASE ||
+	    out >= MTK_AFE_HDMI_CONN_OUTPUT_BASE)
+		return mtk_afe_interconn_hdmi(afe_info, in, out);
+
+	con = mtk_afe_get_connection(afe_info, in, out);
+	if (!con)
+		return -EINVAL;
+
+	regmap_update_bits(afe_info->regmap, con->creg,
+			   1 << con->cshift, 1 << con->cshift);
+
+	if (!con->sreg)
+		return 0;
+
+	if (rightshift)
+		regmap_update_bits(afe_info->regmap, con->sreg,
+				   1 << con->sshift, 1 << con->sshift);
+	else
+		regmap_update_bits(afe_info->regmap, con->sreg,
+				   1 << con->sshift, 0);
+
+	return 0;
+}
+
+/*
+ * mtk_afe_interconn_disconnect - Disconnect an input from an output
+ * @afe_info:	Context
+ * @in:		Input number, as in the SoC datasheet
+ * @out:	Output number, as in the SoC datasheet
+ *
+ * This function disconnects an input of the audio interconnect from an
+ * output.
+ */
+int mtk_afe_interconn_disconnect(struct mtk_afe *afe_info, unsigned int in,
+				 unsigned int out)
+{
+	const struct mtk_afe_connection *con;
+
+	con = mtk_afe_get_connection(afe_info, in, out);
+	if (!con)
+		return -EINVAL;
+
+	regmap_update_bits(afe_info->regmap, con->creg, 1 << con->cshift, 0);
+
+	return 0;
+}
diff --git a/sound/soc/mediatek/mtk-afe-connection.h b/sound/soc/mediatek/mtk-afe-connection.h
new file mode 100644
index 0000000..eac2555
--- /dev/null
+++ b/sound/soc/mediatek/mtk-afe-connection.h
@@ -0,0 +1,30 @@
+/*
+ * mtk_afe_connection.h  --  Mediatek AFE connection support
+ *
+ * Copyright (c) 2015 MediaTek Inc.
+ * Author: Koro Chen <koro.chen@mediatek.com>
+ *             Sascha Hauer <s.hauer@pengutronix.de>
+ *             Hidalgo Huang <hidalgo.huang@mediatek.com>
+ *             Ir Lian <ir.lian@mediatek.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#ifndef _MTK_AFE_CONNECTION_H_
+#define _MTK_AFE_CONNECTION_H_
+
+struct mtk_afe;
+
+int mtk_afe_interconn_connect(struct mtk_afe *afe_info, unsigned int in,
+			      unsigned int out, bool rightshift);
+int mtk_afe_interconn_disconnect(struct mtk_afe *afe_info, unsigned int in,
+				 unsigned int out);
+
+#endif
-- 
1.8.1.1.dirty

[RESEND RFC PATCH 3/3] ASoC: mediatek: Add AFE platform driver

[Koro Chen]

This is the platform driver of Mediatek's AFE (Audio Front End) unit

Signed-off-by: Koro Chen <koro.chen@mediatek.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
---
 sound/soc/Kconfig                |    1 +
 sound/soc/Makefile               |    1 +
 sound/soc/mediatek/Kconfig       |    8 +
 sound/soc/mediatek/Makefile      |    2 +
 sound/soc/mediatek/mtk-afe-pcm.c | 1497 ++++++++++++++++++++++++++++++++++++++
 5 files changed, 1509 insertions(+)
 create mode 100644 sound/soc/mediatek/Kconfig
 create mode 100644 sound/soc/mediatek/Makefile
 create mode 100644 sound/soc/mediatek/mtk-afe-pcm.c

diff --git a/sound/soc/Kconfig b/sound/soc/Kconfig
index dcc79aa..c022374 100644
--- a/sound/soc/Kconfig
+++ b/sound/soc/Kconfig
@@ -45,6 +45,7 @@ source "sound/soc/nuc900/Kconfig"
 source "sound/soc/omap/Kconfig"
 source "sound/soc/kirkwood/Kconfig"
 source "sound/soc/intel/Kconfig"
+source "sound/soc/mediatek/Kconfig"
 source "sound/soc/mxs/Kconfig"
 source "sound/soc/pxa/Kconfig"
 source "sound/soc/rockchip/Kconfig"
diff --git a/sound/soc/Makefile b/sound/soc/Makefile
index 5b3c8f6..6f69f01d 100644
--- a/sound/soc/Makefile
+++ b/sound/soc/Makefile
@@ -23,6 +23,7 @@ obj-$(CONFIG_SND_SOC)	+= dwc/
 obj-$(CONFIG_SND_SOC)	+= fsl/
 obj-$(CONFIG_SND_SOC)	+= jz4740/
 obj-$(CONFIG_SND_SOC)	+= intel/
+obj-$(CONFIG_SND_SOC)	+= mediatek/
 obj-$(CONFIG_SND_SOC)	+= mxs/
 obj-$(CONFIG_SND_SOC)	+= nuc900/
 obj-$(CONFIG_SND_SOC)	+= omap/
diff --git a/sound/soc/mediatek/Kconfig b/sound/soc/mediatek/Kconfig
new file mode 100644
index 0000000..7d97cf1
--- /dev/null
+++ b/sound/soc/mediatek/Kconfig
@@ -0,0 +1,8 @@
+config SND_SOC_MEDIATEK
+	bool "ASoC support for Mediatek chip"
+	depends on ARCH_MEDIATEK
+	help
+	  This adds ASoC platform driver support for Mediatek chip
+	  that can be used with other codecs.
+	  Select Y if you have such device.
+	  Ex: MT8173
diff --git a/sound/soc/mediatek/Makefile b/sound/soc/mediatek/Makefile
new file mode 100644
index 0000000..12003bb
--- /dev/null
+++ b/sound/soc/mediatek/Makefile
@@ -0,0 +1,2 @@
+# MTK Platform Support
+obj-$(CONFIG_SND_SOC_MEDIATEK) += mtk-afe-connection.o mtk-afe-pcm.o
diff --git a/sound/soc/mediatek/mtk-afe-pcm.c b/sound/soc/mediatek/mtk-afe-pcm.c
new file mode 100644
index 0000000..a89c753
--- /dev/null
+++ b/sound/soc/mediatek/mtk-afe-pcm.c
@@ -0,0 +1,1497 @@
+/*
+ * Mediatek ALSA SoC Afe platform driver
+ *
+ * Copyright (c) 2015 MediaTek Inc.
+ * Author: Koro Chen <koro.chen@mediatek.com>
+ *             Sascha Hauer <s.hauer@pengutronix.de>
+ *             Hidalgo Huang <hidalgo.huang@mediatek.com>
+ *             Ir Lian <ir.lian@mediatek.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include <linux/delay.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_address.h>
+#include <linux/pm_runtime.h>
+#include <sound/soc.h>
+#include "mtk-afe-common.h"
+#include "mtk-afe-connection.h"
+#include <dt-bindings/sound/mtk-afe.h>
+
+/*****************************************************************************
+ *                  R E G I S T E R       D E F I N I T I O N
+ *****************************************************************************/
+#define AUDIO_TOP_CON0		0x0000
+#define AUDIO_TOP_CON1		0x0004
+#define AFE_DAC_CON0		0x0010
+#define AFE_DAC_CON1		0x0014
+#define AFE_I2S_CON		0x0018
+#define AFE_I2S_CON1		0x0034
+#define AFE_I2S_CON2		0x0038
+#define AFE_I2S_CON3		0x004c
+#define AFE_CONN_24BIT		0x006c
+
+/* Memory interface */
+#define AFE_DL1_BASE		0x0040
+#define AFE_DL1_CUR		0x0044
+#define AFE_DL1_END		0x0048
+#define AFE_DL2_BASE		0x0050
+#define AFE_DL2_CUR		0x0054
+#define AFE_DL2_END		0x0058
+#define AFE_AWB_BASE		0x0070
+#define AFE_AWB_END		0x0078
+#define AFE_AWB_CUR		0x007c
+#define AFE_VUL_BASE		0x0080
+#define AFE_VUL_END		0x0088
+#define AFE_VUL_CUR		0x008c
+#define AFE_DAI_BASE		0x0090
+#define AFE_DAI_END		0x0098
+#define AFE_DAI_CUR		0x009c
+#define AFE_MOD_PCM_BASE	0x0330
+#define AFE_MOD_PCM_END		0x0338
+#define AFE_MOD_PCM_CUR		0x033C
+#define AFE_HDMI_OUT_BASE	0x0374
+#define AFE_HDMI_OUT_CUR	0x0378
+#define AFE_HDMI_OUT_END	0x037c
+
+#define AFE_ADDA_DL_SRC2_CON0	0x0108
+#define AFE_ADDA_DL_SRC2_CON1	0x010c
+#define AFE_ADDA_UL_SRC_CON0	0x0114
+#define AFE_ADDA_UL_SRC_CON1	0x0118
+#define AFE_ADDA_TOP_CON0	0x0120
+#define AFE_ADDA_UL_DL_CON0	0x0124
+#define AFE_ADDA_NEWIF_CFG0     0x0138
+#define AFE_ADDA_NEWIF_CFG1     0x013C
+#define AFE_ADDA2_TOP_CON0	0x0600
+
+#define AFE_HDMI_OUT_CON0	0x0370
+
+#define AFE_IRQ_MCU_CON		0x03a0
+#define AFE_IRQ_STATUS		0x03a4
+#define AFE_IRQ_CLR		0x03a8
+#define AFE_IRQ_CNT1		0x03ac
+#define AFE_IRQ_CNT2		0x03b0
+#define AFE_IRQ_MCU_EN		0x03b4
+#define AFE_IRQ_CNT5		0x03bc
+#define AFE_IRQ_CNT7		0x03dc
+
+#define AFE_TDM_CON1		0x0548
+#define AFE_TDM_CON2		0x054c
+
+#define AFE_BASE_END_OFFSET	8
+#define AFE_IRQ_STATUS_BITS	0xff
+
+#define AFE_TDM_CON1_BCK_INV			(1 << 1)
+#define AFE_TDM_CON1_LRCK_INV			(1 << 2)
+#define AFE_TDM_CON1_1_BCK_DELAY		(1 << 3)
+#define AFE_TDM_CON1_MSB_ALIGNED		(1 << 4)
+#define AFE_TDM_CON1_WLEN_16BIT			(1 << 8)
+#define AFE_TDM_CON1_WLEN_32BIT			(2 << 8)
+#define AFE_TDM_CON1_2CH_SDATA			(0 << 10)
+#define AFE_TDM_CON1_4CH_SDATA			(1 << 10)
+#define AFE_TDM_CON1_8CH_SDATA			(2 << 10)
+#define AFE_TDM_CON1_16_BCK_CYCLES		(0 << 12)
+#define AFE_TDM_CON1_24_BCK_CYCLES		(1 << 12)
+#define AFE_TDM_CON1_32_BCK_CYCLES		(2 << 12)
+#define AFE_TDM_CON1_LRCK_TDM_WIDTH(x)		(((x) - 1) << 24)
+
+#define MTK_AFE_I2S_CON_PHASE_SHIFT_FIX		(1 << 31)
+#define MTK_AFE_I2S_CON_BCK_INVERT		(1 << 29)
+#define MTK_AFE_I2S_CON_FROM_IO_MUX		(1 << 28)
+#define MTK_AFE_I2S_CON_LOW_JITTER_CLOCK	(1 << 12)
+#define MTK_AFE_I2S_CON_INV_LRCK		(1 << 5)
+#define MTK_AFE_I2S_CON_FORMAT_I2S		(1 << 3)
+#define MTK_AFE_I2S_CON_SRC_SLAVE		(1 << 2)
+#define MTK_AFE_I2S_CON_WLEN_32BITS		(1 << 1)
+#define MTK_AFE_I2S_CON_ENABLE			(1 << 0)
+
+#define MTK_AFE_I2S_CON3_LR_SWAP		(1 << 31)
+#define MTK_AFE_I2S_CON3_LOW_JITTER_CLOCK	(1 << 12)
+#define MTK_AFE_I2S_CON3_RATE(x)		(((x) & 0xf) << 8)
+#define MTK_AFE_I2S_CON3_INV_LRCK		(1 << 5)
+#define MTK_AFE_I2S_CON3_FORMAT_I2S		(1 << 3)
+#define MTK_AFE_I2S_CON3_WLEN_32BITS		(1 << 1)
+#define MTK_AFE_I2S_CON3_ENABLE			(1 << 0)
+
+#define MTK_AFE_I2S_CON2_LR_SWAP		(1 << 31)
+#define MTK_AFE_I2S_CON2_BCK_INVERT		(1 << 23)
+#define MTK_AFE_I2S_CON2_LOW_JITTER_CLOCK	(1 << 12)
+#define MTK_AFE_I2S_CON2_RATE(x)		(((x) & 0xf) << 8)
+#define MTK_AFE_I2S_CON2_FORMAT_I2S		(1 << 3)
+#define MTK_AFE_I2S_CON2_WLEN_32BITS		(1 << 1)
+#define MTK_AFE_I2S_CON2_ENABLE			(1 << 0)
+
+#define MTK_AFE_I2S_CON1_LR_SWAP		(1 << 31)
+#define MTK_AFE_I2S_CON1_LOW_JITTER_CLOCK	(1 << 12)
+#define MTK_AFE_I2S_CON1_RATE(x)		(((x) & 0xf) << 8)
+#define MTK_AFE_I2S_CON1_INV_LRCK		(1 << 5)
+#define MTK_AFE_I2S_CON1_FORMAT_I2S		(1 << 3)
+#define MTK_AFE_I2S_CON1_WLEN_32BITS		(1 << 1)
+#define MTK_AFE_I2S_CON1_ENABLE			(1 << 0)
+
+enum mtk_afe_tdm_ch_start {
+	MTK_AFE_TDM_CH_START_O30_O31 = 0,
+	MTK_AFE_TDM_CH_START_O32_O33,
+	MTK_AFE_TDM_CH_START_O34_O35,
+	MTK_AFE_TDM_CH_START_O36_O37,
+	MTK_AFE_TDM_CH_ZERO,
+};
+
+enum mtk_afe_i2s_rate {
+	MTK_AFE_I2S_RATE_8K = 0,
+	MTK_AFE_I2S_RATE_11K = 1,
+	MTK_AFE_I2S_RATE_12K = 2,
+	MTK_AFE_I2S_RATE_16K = 4,
+	MTK_AFE_I2S_RATE_22K = 5,
+	MTK_AFE_I2S_RATE_24K = 6,
+	MTK_AFE_I2S_RATE_32K = 8,
+	MTK_AFE_I2S_RATE_44K = 9,
+	MTK_AFE_I2S_RATE_48K = 10,
+	MTK_AFE_I2S_RATE_88K = 11,
+	MTK_AFE_I2S_RATE_96K = 12,
+	MTK_AFE_I2S_RATE_174K = 13,
+	MTK_AFE_I2S_RATE_192K = 14,
+};
+
+static const struct snd_pcm_hardware mtk_afe_hardware = {
+	.info = (SNDRV_PCM_INFO_MMAP | SNDRV_PCM_INFO_INTERLEAVED |
+		 SNDRV_PCM_INFO_RESUME | SNDRV_PCM_INFO_MMAP_VALID),
+	.buffer_bytes_max = 256 * 1024,
+	.period_bytes_min = 512,
+	.period_bytes_max = 128 * 1024,
+	.periods_min = 2,
+	.periods_max = 256,
+	.fifo_size = 0,
+};
+
+static int mtk_afe_pcm_open(struct snd_pcm_substream *substream)
+{
+	struct snd_soc_pcm_runtime *rtd = substream->private_data;
+	struct snd_pcm_runtime *runtime = substream->runtime;
+	struct mtk_afe *afe = snd_soc_platform_get_drvdata(rtd->platform);
+	struct mtk_afe_io *io = &afe->ios[rtd->cpu_dai->id];
+	struct mtk_afe_memif *memif;
+	int ret;
+
+	if (io->mem[substream->stream] < 0)
+		/* memif is not specified */
+		return -ENXIO;
+	memif = &afe->memif[io->mem[substream->stream]];
+	memif->substream = substream;
+
+	snd_soc_set_runtime_hwparams(substream, &mtk_afe_hardware);
+	ret = snd_pcm_hw_constraint_integer(runtime,
+					    SNDRV_PCM_HW_PARAM_PERIODS);
+	if (ret < 0) {
+		dev_err(afe->dev, "snd_pcm_hw_constraint_integer failed\n");
+		return ret;
+	}
+
+	if (!memif->use_sram)
+		return 0;
+
+	return snd_pcm_hw_constraint_minmax(runtime,
+					    SNDRV_PCM_HW_PARAM_BUFFER_BYTES,
+					    mtk_afe_hardware.period_bytes_min,
+					    afe->sram_size);
+}
+
+static int mtk_afe_pcm_close(struct snd_pcm_substream *substream)
+{
+	struct snd_soc_pcm_runtime *rtd = substream->private_data;
+	struct mtk_afe *afe = snd_soc_platform_get_drvdata(rtd->platform);
+	struct mtk_afe_io *io = &afe->ios[rtd->cpu_dai->id];
+
+	afe->memif[io->mem[substream->stream]].substream = NULL;
+
+	return 0;
+}
+
+static int mtk_afe_pcm_hw_params(struct snd_pcm_substream *substream,
+				 struct snd_pcm_hw_params *params)
+{
+	struct snd_soc_pcm_runtime *rtd = substream->private_data;
+	struct mtk_afe *afe = snd_soc_platform_get_drvdata(rtd->platform);
+	struct mtk_afe_io *io = &afe->ios[rtd->cpu_dai->id];
+	struct mtk_afe_memif *memif = &afe->memif[io->mem[substream->stream]];
+	int ret;
+
+	dev_dbg(afe->dev,
+		"%s period = %u,rate= %u, channels=%u\n",
+		__func__, params_period_size(params), params_rate(params),
+		params_channels(params));
+
+	if (memif->use_sram) {
+		struct snd_dma_buffer *dma_buf = &substream->dma_buffer;
+		int size = params_buffer_bytes(params);
+
+		memif->buffer_size = size;
+		memif->phys_buf_addr = afe->sram_phy_address;
+
+		dma_buf->bytes = size;
+		dma_buf->area = (unsigned char *)afe->sram_address;
+		dma_buf->addr = afe->sram_phy_address;
+		dma_buf->dev.type = SNDRV_DMA_TYPE_DEV;
+		dma_buf->dev.dev = substream->pcm->card->dev;
+		snd_pcm_set_runtime_buffer(substream, dma_buf);
+	} else {
+		ret = snd_pcm_lib_malloc_pages(substream,
+					       params_buffer_bytes(params));
+		if (ret < 0)
+			return ret;
+
+		memif->phys_buf_addr = substream->runtime->dma_addr;
+		memif->buffer_size = substream->runtime->dma_bytes;
+	}
+	memif->hw_ptr = 0;
+
+	/* start */
+	regmap_write(afe->regmap,
+		     memif->data->reg_ofs_base, memif->phys_buf_addr);
+	/* end */
+	regmap_write(afe->regmap,
+		     memif->data->reg_ofs_base + AFE_BASE_END_OFFSET,
+		     memif->phys_buf_addr + memif->buffer_size - 1);
+
+	return 0;
+}
+
+static int mtk_afe_pcm_hw_free(struct snd_pcm_substream *substream)
+{
+	struct snd_soc_pcm_runtime *rtd = substream->private_data;
+	struct mtk_afe *afe = snd_soc_platform_get_drvdata(rtd->platform);
+	struct mtk_afe_io *io = &afe->ios[rtd->cpu_dai->id];
+	struct mtk_afe_memif *memif = &afe->memif[io->mem[substream->stream]];
+
+	if (memif->use_sram)
+		return 0;
+
+	return snd_pcm_lib_free_pages(substream);
+}
+
+static snd_pcm_uframes_t mtk_afe_pcm_pointer
+			 (struct snd_pcm_substream *substream)
+{
+	struct snd_soc_pcm_runtime *rtd = substream->private_data;
+	struct mtk_afe *afe = snd_soc_platform_get_drvdata(rtd->platform);
+	struct mtk_afe_io *io = &afe->ios[rtd->cpu_dai->id];
+	struct mtk_afe_memif *memif = &afe->memif[io->mem[substream->stream]];
+
+	return bytes_to_frames(substream->runtime, memif->hw_ptr);
+}
+
+static const struct snd_pcm_ops mtk_afe_pcm_ops = {
+	.open = mtk_afe_pcm_open,
+	.close = mtk_afe_pcm_close,
+	.ioctl = snd_pcm_lib_ioctl,
+	.hw_params = mtk_afe_pcm_hw_params,
+	.hw_free = mtk_afe_pcm_hw_free,
+	.pointer = mtk_afe_pcm_pointer,
+};
+
+static int mtk_afe_pcm_new(struct snd_soc_pcm_runtime *rtd)
+{
+	size_t size;
+	struct snd_card *card = rtd->card->snd_card;
+	struct snd_pcm *pcm = rtd->pcm;
+
+	size = mtk_afe_hardware.buffer_bytes_max;
+
+	return snd_pcm_lib_preallocate_pages_for_all(pcm, SNDRV_DMA_TYPE_DEV,
+						     card->dev, size, size);
+}
+
+static void mtk_afe_pcm_free(struct snd_pcm *pcm)
+{
+	snd_pcm_lib_preallocate_free_for_all(pcm);
+}
+
+static const struct snd_soc_platform_driver mtk_afe_pcm_platform = {
+	.ops = &mtk_afe_pcm_ops,
+	.pcm_new = mtk_afe_pcm_new,
+	.pcm_free = mtk_afe_pcm_free,
+};
+
+struct mtk_afe_rate {
+	int rate;
+	int regvalue;
+};
+
+static const struct mtk_afe_rate mtk_afe_adda_rates[] = {
+	{ .rate = 8000, .regvalue = 0 },
+	{ .rate = 11025, .regvalue = 1 },
+	{ .rate = 12000, .regvalue = 2 },
+	{ .rate = 16000, .regvalue = 3 },
+	{ .rate = 22050, .regvalue = 4 },
+	{ .rate = 24000, .regvalue = 5 },
+	{ .rate = 32000, .regvalue = 6 },
+	{ .rate = 44100, .regvalue = 7 },
+	{ .rate = 48000, .regvalue = 8 },
+};
+
+static int mtk_afe_adda_fs(u32 sample_rate)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(mtk_afe_adda_rates); i++)
+		if (mtk_afe_adda_rates[i].rate == sample_rate)
+			return mtk_afe_adda_rates[i].regvalue;
+
+	return -EINVAL;
+}
+
+static const struct mtk_afe_rate mtk_afe_i2s_rates[] = {
+	{ .rate = 8000, .regvalue = MTK_AFE_I2S_RATE_8K },
+	{ .rate = 11025, .regvalue = MTK_AFE_I2S_RATE_11K },
+	{ .rate = 12000, .regvalue = MTK_AFE_I2S_RATE_12K },
+	{ .rate = 16000, .regvalue = MTK_AFE_I2S_RATE_16K },
+	{ .rate = 22050, .regvalue = MTK_AFE_I2S_RATE_22K },
+	{ .rate = 24000, .regvalue = MTK_AFE_I2S_RATE_24K },
+	{ .rate = 32000, .regvalue = MTK_AFE_I2S_RATE_32K },
+	{ .rate = 44100, .regvalue = MTK_AFE_I2S_RATE_44K },
+	{ .rate = 48000, .regvalue = MTK_AFE_I2S_RATE_48K },
+	{ .rate = 88000, .regvalue = MTK_AFE_I2S_RATE_88K },
+	{ .rate = 96000, .regvalue = MTK_AFE_I2S_RATE_96K },
+	{ .rate = 174000, .regvalue = MTK_AFE_I2S_RATE_174K },
+	{ .rate = 192000, .regvalue = MTK_AFE_I2S_RATE_192K },
+};
+
+static int mtk_afe_i2s_fs(u32 sample_rate)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(mtk_afe_i2s_rates); i++)
+		if (mtk_afe_i2s_rates[i].rate == sample_rate)
+			return mtk_afe_i2s_rates[i].regvalue;
+
+	return -EINVAL;
+}
+
+static int mtk_afe_set_adda_dac_out(struct mtk_afe *afe, uint32_t rate)
+{
+	u32 audio_i2s_dac = 0;
+	u32 con0, con1;
+
+	/* set dl src2 */
+	con0 = (mtk_afe_adda_fs(rate) << 28) | (0x03 << 24) | (0x03 << 11);
+
+	/* 8k or 16k voice mode */
+	if (con0 == 0 || con0 == 3)
+		con0 |= 0x01 << 5;
+
+	/* SA suggests to apply -0.3db to audio/speech path */
+	con0 = con0 | (0x01 << 1);
+	con1 = 0xf74f0000;
+
+	/* set and turn off */
+	regmap_write(afe->regmap, AFE_ADDA_DL_SRC2_CON0, con0);
+
+	regmap_write(afe->regmap, AFE_ADDA_DL_SRC2_CON1, 0xf74f0000);
+
+	audio_i2s_dac =	MTK_AFE_I2S_CON1_FORMAT_I2S |
+		MTK_AFE_I2S_CON1_RATE(mtk_afe_i2s_fs(rate));
+
+	/* set and turn off */
+	regmap_write(afe->regmap, AFE_I2S_CON1, audio_i2s_dac);
+	return 0;
+}
+
+static void mtk_afe_pmic_shutdown(struct mtk_afe *afe,
+				  struct snd_pcm_substream *substream)
+{
+	/* output */
+	regmap_update_bits(afe->regmap, AFE_ADDA_DL_SRC2_CON0, 1, 0);
+	regmap_update_bits(afe->regmap, AFE_I2S_CON1, 1, 0);
+
+	/* input */
+	regmap_update_bits(afe->regmap, AFE_ADDA_UL_SRC_CON0, 1, 0);
+	/* disable ADDA */
+	regmap_update_bits(afe->regmap, AFE_ADDA_UL_DL_CON0, 1, false);
+}
+
+static int mtk_afe_pmic_prepare(struct mtk_afe *afe,
+				struct snd_pcm_substream *substream)
+{
+	if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) {
+		mtk_afe_set_adda_dac_out(afe, substream->runtime->rate);
+		regmap_update_bits(afe->regmap, AFE_ADDA_DL_SRC2_CON0, 1, 1);
+		regmap_update_bits(afe->regmap, AFE_I2S_CON1, 1, 1);
+
+	} else {
+		u32 rate = mtk_afe_i2s_fs(substream->runtime->rate);
+		u32 voice_mode = 0;
+
+		/* Use Int ADC */
+		regmap_update_bits(afe->regmap, AFE_ADDA_TOP_CON0, 1, 0);
+
+		if (rate == MTK_AFE_I2S_RATE_8K)
+			voice_mode = 0;
+		else if (rate == MTK_AFE_I2S_RATE_16K)
+			voice_mode = 1;
+		else if (rate == MTK_AFE_I2S_RATE_32K)
+			voice_mode = 2;
+		else if (rate == MTK_AFE_I2S_RATE_48K)
+			voice_mode = 3;
+
+		/* set and turn off */
+		regmap_write(afe->regmap, AFE_ADDA_UL_SRC_CON0,
+			     ((voice_mode << 2) | voice_mode) << 17);
+
+		regmap_update_bits(afe->regmap, AFE_ADDA_NEWIF_CFG1,
+				   0xc00, ((voice_mode < 3) ? 1 : 3) << 10);
+
+		regmap_update_bits(afe->regmap, AFE_ADDA_UL_SRC_CON0, 1, 1);
+	}
+	/* enable ADDA */
+	regmap_update_bits(afe->regmap, AFE_ADDA_UL_DL_CON0, 1, true);
+	return 0;
+}
+
+static int mtk_afe_set_2nd_i2s(struct mtk_afe *afe, bool doublew,
+			       uint32_t rate, bool bck_inv, bool slave)
+{
+	uint32_t val;
+	unsigned int fs = mtk_afe_i2s_fs(rate);
+
+	regmap_update_bits(afe->regmap, AFE_DAC_CON1, 0xf00, fs << 8);
+
+	val = MTK_AFE_I2S_CON_PHASE_SHIFT_FIX |
+	      MTK_AFE_I2S_CON_FROM_IO_MUX |
+	      MTK_AFE_I2S_CON_LOW_JITTER_CLOCK |
+	      MTK_AFE_I2S_CON_FORMAT_I2S;
+
+	if (bck_inv)
+		val |= MTK_AFE_I2S_CON_BCK_INVERT;
+	if (slave)
+		val |= MTK_AFE_I2S_CON_SRC_SLAVE;
+	if (doublew)
+		val |= MTK_AFE_I2S_CON_WLEN_32BITS;
+
+	regmap_update_bits(afe->regmap,
+			   AFE_I2S_CON, 0xfffffffe, val);
+
+	/* set output */
+	val = MTK_AFE_I2S_CON3_LOW_JITTER_CLOCK |
+	      MTK_AFE_I2S_CON3_RATE(fs) |
+	      MTK_AFE_I2S_CON3_FORMAT_I2S;
+
+	if (doublew)
+		val |= MTK_AFE_I2S_CON3_WLEN_32BITS;
+
+	regmap_update_bits(afe->regmap,
+			   AFE_I2S_CON3, 0xfffffffe, val);
+	return 0;
+}
+
+static int mtk_afe_set_i2s(struct mtk_afe *afe, uint32_t rate)
+{
+	uint32_t val;
+	unsigned int fs = mtk_afe_i2s_fs(rate);
+
+	/* from external ADC */
+	regmap_update_bits(afe->regmap, AFE_ADDA2_TOP_CON0, 0x1, 0x1);
+
+	/* set input */
+	val = MTK_AFE_I2S_CON2_LOW_JITTER_CLOCK |
+	      MTK_AFE_I2S_CON2_RATE(fs) |
+	      MTK_AFE_I2S_CON2_FORMAT_I2S;
+
+	regmap_update_bits(afe->regmap,
+			   AFE_I2S_CON2, 0xfffffffe, val);
+
+	/* set output */
+	val = MTK_AFE_I2S_CON1_LOW_JITTER_CLOCK |
+	      MTK_AFE_I2S_CON1_RATE(fs) |
+	      MTK_AFE_I2S_CON1_FORMAT_I2S;
+
+	regmap_update_bits(afe->regmap,
+			   AFE_I2S_CON1, 0xfffffffe, val);
+	return 0;
+}
+
+static void mtk_afe_set_2nd_i2s_enable(struct mtk_afe *afe,
+				       uint32_t rate, bool enable)
+{
+	int val;
+
+	regmap_read(afe->regmap, AFE_I2S_CON, &val);
+	if (!!(val & MTK_AFE_I2S_CON_ENABLE) == enable)
+		return; /* must skip soft reset */
+
+	/* 2nd I2S soft reset begin */
+	regmap_update_bits(afe->regmap, AUDIO_TOP_CON1, 2, 2);
+
+	/* input */
+	regmap_update_bits(afe->regmap, AFE_I2S_CON, 1, enable);
+
+	/* output */
+	regmap_update_bits(afe->regmap, AFE_I2S_CON3, 1, enable);
+
+	/* 2nd I2S soft reset end */
+	udelay(1);
+	regmap_update_bits(afe->regmap, AUDIO_TOP_CON1, 2, 0);
+}
+
+static void mtk_afe_set_i2s_enable(struct mtk_afe *afe,
+				   uint32_t rate, bool enable)
+{
+	int val;
+
+	regmap_read(afe->regmap, AFE_I2S_CON2, &val);
+	if (!!(val & MTK_AFE_I2S_CON2_ENABLE) == enable)
+		return; /* must skip soft reset */
+
+	/* I2S soft reset begin */
+	regmap_update_bits(afe->regmap, AUDIO_TOP_CON1, 4, 4);
+
+	/* input */
+	regmap_update_bits(afe->regmap, AFE_I2S_CON2, 1, enable);
+
+	/* output */
+	regmap_update_bits(afe->regmap, AFE_I2S_CON1, 1, enable);
+
+	/* I2S soft reset end */
+	udelay(1);
+	regmap_update_bits(afe->regmap, AUDIO_TOP_CON1, 4, 0);
+}
+
+static void mtk_afe_i2s_shutdown(struct mtk_afe *afe,
+				 struct snd_pcm_substream *substream)
+{
+	mtk_afe_set_i2s_enable(afe, substream->runtime->rate, false);
+}
+
+static int mtk_afe_i2s_prepare(struct mtk_afe *afe,
+			       struct snd_pcm_substream *substream)
+{
+	/* config I2S */
+	mtk_afe_set_i2s(afe, substream->runtime->rate);
+	mtk_afe_set_i2s_enable(afe, substream->runtime->rate, true);
+
+	return 0;
+}
+
+static void mtk_afe_i2s2_shutdown(struct mtk_afe *afe,
+				  struct snd_pcm_substream *substream)
+{
+	mtk_afe_set_2nd_i2s_enable(afe, substream->runtime->rate, false);
+}
+
+static int mtk_afe_i2s2_prepare(struct mtk_afe *afe,
+				struct snd_pcm_substream *substream)
+{
+	/* config I2S */
+	mtk_afe_set_2nd_i2s(afe, false, substream->runtime->rate, false, false);
+	mtk_afe_set_2nd_i2s_enable(afe, substream->runtime->rate, true);
+	return 0;
+}
+
+static int mtk_afe_hdmi_start(struct mtk_afe *afe,
+			      struct snd_pcm_substream *substream)
+{
+	regmap_update_bits(afe->regmap, AUDIO_TOP_CON0,
+			   (1 << 20) | (1 << 21), 0);
+
+	/* enable Out control */
+	regmap_update_bits(afe->regmap, AFE_HDMI_OUT_CON0, 1, 1);
+
+	/* enable tdm */
+	regmap_update_bits(afe->regmap, AFE_TDM_CON1, 1, 1);
+
+	return 0;
+}
+
+static void mtk_afe_hdmi_pause(struct mtk_afe *afe,
+			       struct snd_pcm_substream *substream)
+{
+	/* disable tdm */
+	regmap_update_bits(afe->regmap, AFE_TDM_CON1, 1, 0);
+
+	/* disable Out control */
+	regmap_update_bits(afe->regmap, AFE_HDMI_OUT_CON0, 1, 0);
+
+	regmap_update_bits(afe->regmap, AUDIO_TOP_CON0,
+			   (1 << 20) | (1 << 21), (1 << 20) | (1 << 21));
+}
+
+static int mtk_afe_hdmi_prepare(struct mtk_afe *afe,
+				struct snd_pcm_substream *substream)
+{
+	struct snd_pcm_runtime *runtime = substream->runtime;
+	u32 val;
+
+	val = AFE_TDM_CON1_BCK_INV |
+		AFE_TDM_CON1_1_BCK_DELAY |
+		AFE_TDM_CON1_MSB_ALIGNED | /* I2S mode */
+		AFE_TDM_CON1_WLEN_32BIT |
+		AFE_TDM_CON1_32_BCK_CYCLES |
+		AFE_TDM_CON1_LRCK_TDM_WIDTH(32);
+
+	regmap_update_bits(afe->regmap, AFE_TDM_CON1, ~1, val);
+
+	/* set tdm2 config */
+	switch (runtime->channels) {
+	case 7:
+	case 8:
+		val |= MTK_AFE_TDM_CH_START_O30_O31;
+		val |= (MTK_AFE_TDM_CH_START_O32_O33 << 4);
+		val |= (MTK_AFE_TDM_CH_START_O34_O35 << 8);
+		val |= (MTK_AFE_TDM_CH_START_O36_O37 << 12);
+		break;
+	case 5:
+	case 6:
+		val |= MTK_AFE_TDM_CH_START_O30_O31;
+		val |= (MTK_AFE_TDM_CH_START_O32_O33 << 4);
+		val |= (MTK_AFE_TDM_CH_START_O34_O35 << 8);
+		val |= (MTK_AFE_TDM_CH_ZERO << 12);
+		break;
+	case 3:
+	case 4:
+		val |= MTK_AFE_TDM_CH_START_O30_O31;
+		val |= (MTK_AFE_TDM_CH_START_O32_O33 << 4);
+		val |= (MTK_AFE_TDM_CH_ZERO << 8);
+		val |= (MTK_AFE_TDM_CH_ZERO << 12);
+		break;
+	case 1:
+	case 2:
+		val |= MTK_AFE_TDM_CH_START_O30_O31;
+		val |= (MTK_AFE_TDM_CH_ZERO << 4);
+		val |= (MTK_AFE_TDM_CH_ZERO << 8);
+		val |= (MTK_AFE_TDM_CH_ZERO << 12);
+		break;
+	default:
+		val = 0;
+	}
+	regmap_update_bits(afe->regmap, AFE_TDM_CON2,
+			   0x0000ffff, val);
+
+	regmap_update_bits(afe->regmap, AFE_HDMI_OUT_CON0,
+			   0x000000f0, runtime->channels << 4);
+	return 0;
+}
+
+static int mtk_set_connections(struct mtk_afe *afe, struct mtk_afe_io *io)
+{
+	int i, ret;
+
+	for (i = 0; i < io->num_connections; i++) {
+		ret = mtk_afe_interconn_connect(afe, io->connections[i * 2],
+						io->connections[i * 2 + 1],
+						0);
+		if (ret) {
+			dev_err(afe->dev, "Cannot create connection %d <-> %d\n",
+				io->connections[i * 2],
+				io->connections[i * 2 + 1]);
+			return ret;
+		}
+	}
+
+	return 0;
+}
+
+static int mtk_afe_dais_startup(struct snd_pcm_substream *substream,
+				struct snd_soc_dai *dai)
+{
+	struct snd_soc_pcm_runtime *rtd = substream->private_data;
+	struct mtk_afe *afe = snd_soc_platform_get_drvdata(rtd->platform);
+	struct mtk_afe_io *io = &afe->ios[rtd->cpu_dai->id];
+	int ret;
+
+	dev_dbg(afe->dev, "%s\n", __func__);
+
+	if (dai->active)
+		return 0;
+
+	ret = mtk_set_connections(afe, io);
+	if (ret)
+		return ret;
+
+	dev_dbg(afe->dev, "enabling io %s\n", io->data->name);
+	if (io->m_ck) {
+		ret = clk_prepare_enable(io->m_ck);
+		if (ret) {
+			dev_err(afe->dev, "Failed to enable m_ck\n");
+			return ret;
+		}
+		regmap_update_bits(afe->regmap, AUDIO_TOP_CON0,
+				   (1 << 8) | (1 << 9), 0);
+	}
+
+	if (io->b_ck) {
+		ret = clk_prepare_enable(io->b_ck);
+		if (ret) {
+			dev_err(afe->dev, "Failed to enable b_ck\n");
+			return ret;
+		}
+	}
+
+	return 0;
+}
+
+static void mtk_afe_dais_shutdown(struct snd_pcm_substream *substream,
+				  struct snd_soc_dai *dai)
+{
+	struct snd_soc_pcm_runtime *rtd = substream->private_data;
+	struct mtk_afe *afe = snd_soc_platform_get_drvdata(rtd->platform);
+	struct mtk_afe_io *io = &afe->ios[rtd->cpu_dai->id];
+
+	dev_dbg(afe->dev, "%s\n", __func__);
+
+	if (dai->active)
+		return;
+
+	if (io->data->shutdown)
+		io->data->shutdown(afe, substream);
+
+	if (io->m_ck) {
+		regmap_update_bits(afe->regmap, AUDIO_TOP_CON0,
+				   (1 << 8) | (1 << 9),
+				   (1 << 8) | (1 << 9));
+		clk_disable_unprepare(io->m_ck);
+	}
+	if (io->b_ck)
+		clk_disable_unprepare(io->b_ck);
+	/* disable AFE */
+	regmap_update_bits(afe->regmap, AFE_DAC_CON0, 1, 0);
+}
+
+static int mtk_afe_dais_hw_params(struct snd_pcm_substream *substream,
+				  struct snd_pcm_hw_params *params,
+				  struct snd_soc_dai *dai)
+{
+	struct snd_soc_pcm_runtime *rtd = substream->private_data;
+	struct mtk_afe *afe = snd_soc_platform_get_drvdata(rtd->platform);
+	struct mtk_afe_io *io = &afe->ios[rtd->cpu_dai->id];
+	struct mtk_afe_memif *memif = &afe->memif[io->mem[substream->stream]];
+	unsigned int mono;
+	int sample_rate = params_rate(params);
+	int fs_shift = memif->data->fs_shift;
+	u32 val;
+
+	/* set channel */
+	if (memif->data->mono_shift >= 0) {
+		mono = (params_channels(params) == 1) ? 1 : 0;
+		regmap_update_bits(afe->regmap, AFE_DAC_CON1,
+				   1 << memif->data->mono_shift,
+				   mono << memif->data->mono_shift);
+	}
+
+	/* set rate */
+	if (fs_shift < 0)
+		return 0;
+	if (memif->data->id == MTK_AFE_MEMIF_DAI ||
+	    memif->data->id == MTK_AFE_MEMIF_MOD_DAI) {
+		if (sample_rate == 8000)
+			val = 0;
+		else if (sample_rate == 16000)
+			val = 1;
+		else if (sample_rate == 32000)
+			val = 2;
+		else
+			return -EINVAL;
+		if (memif->data->id == MTK_AFE_MEMIF_DAI)
+			regmap_update_bits(afe->regmap, AFE_DAC_CON0,
+					   0x3 << fs_shift, val << fs_shift);
+		else
+			regmap_update_bits(afe->regmap, AFE_DAC_CON1,
+					   0x3 << fs_shift, val << fs_shift);
+
+	} else {
+		regmap_update_bits(afe->regmap, AFE_DAC_CON1,
+				   0xf << fs_shift,
+				   mtk_afe_i2s_fs(sample_rate) << fs_shift);
+	}
+
+	return 0;
+}
+
+static int mtk_afe_dais_prepare(struct snd_pcm_substream *substream,
+				struct snd_soc_dai *dai)
+{
+	struct snd_soc_pcm_runtime *rtd = substream->private_data;
+	struct snd_pcm_runtime * const runtime = substream->runtime;
+	struct mtk_afe *afe = snd_soc_platform_get_drvdata(rtd->platform);
+	struct mtk_afe_io *io = &afe->ios[rtd->cpu_dai->id];
+	int ret;
+
+	dev_dbg(afe->dev, "%s: %d\n", __func__, substream->runtime->rate);
+
+	if (io->m_ck) {
+		ret = clk_set_rate(io->m_ck, runtime->rate * 256);
+		if (ret) {
+			dev_err(afe->dev, "Failed to set m_ck rate\n");
+			return ret;
+		}
+	}
+
+	if (io->b_ck) {
+		ret = clk_set_rate(io->b_ck, runtime->rate *
+				   runtime->channels * 32 * 16);
+		if (ret) {
+			dev_err(afe->dev, "Failed to set b_ck rate\n");
+			return ret;
+		}
+	}
+
+	if (io->data->prepare)
+		ret = io->data->prepare(afe, substream);
+
+	/* enable AFE */
+	regmap_update_bits(afe->regmap, AFE_DAC_CON0, 1, 1);
+	return 0;
+}
+
+static int mtk_afe_dais_trigger(struct snd_pcm_substream *substream, int cmd,
+				struct snd_soc_dai *dai)
+{
+	struct snd_soc_pcm_runtime *rtd = substream->private_data;
+	struct snd_pcm_runtime * const runtime = substream->runtime;
+	struct mtk_afe *afe = snd_soc_platform_get_drvdata(rtd->platform);
+	struct mtk_afe_io *io = &afe->ios[rtd->cpu_dai->id];
+	struct mtk_afe_memif *memif = &afe->memif[io->mem[substream->stream]];
+	unsigned int counter = runtime->period_size;
+	int ret;
+	int enable_shift = memif->data->enable_shift;
+
+	dev_info(afe->dev, "%s %s cmd=%d\n", __func__, memif->data->name, cmd);
+
+	if (!memif->irqdata) {
+		dev_err(afe->dev, "IRQ for memif %d is not described\n",
+			rtd->cpu_dai->id);
+		return -ENXIO;
+	}
+
+	switch (cmd) {
+	case SNDRV_PCM_TRIGGER_START:
+	case SNDRV_PCM_TRIGGER_RESUME:
+		if (enable_shift >= 0)
+			regmap_update_bits(afe->regmap, AFE_DAC_CON0,
+					   1 << enable_shift, 0xffffffff);
+
+		/* set irq counter */
+		regmap_update_bits(afe->regmap,
+				   memif->irqdata->reg_cnt,
+				   0x3ffff << memif->irqdata->cnt_shift,
+				   counter << memif->irqdata->cnt_shift);
+
+		/* set irq fs */
+		if (memif->irqdata->fs_shift >= 0)
+			regmap_update_bits(afe->regmap,
+					   AFE_IRQ_MCU_CON,
+					   0xf << memif->irqdata->fs_shift,
+					   mtk_afe_i2s_fs(runtime->rate) <<
+					       memif->irqdata->fs_shift);
+		/* enable interrupt */
+		regmap_update_bits(afe->regmap, AFE_IRQ_MCU_CON,
+				   1 << memif->irqdata->en_shift,
+				   1 << memif->irqdata->en_shift);
+
+		if (io->data->start) {
+			ret = io->data->start(afe, substream);
+			if (ret)
+				return ret;
+		}
+		return 0;
+	case SNDRV_PCM_TRIGGER_STOP:
+	case SNDRV_PCM_TRIGGER_SUSPEND:
+		if (io->data->pause)
+			io->data->pause(afe, substream);
+
+		if (enable_shift >= 0)
+			regmap_update_bits(afe->regmap, AFE_DAC_CON0,
+					   1 << enable_shift, 0);
+		/* disable interrupt */
+		regmap_update_bits(afe->regmap, AFE_IRQ_MCU_CON,
+				   1 << memif->irqdata->en_shift,
+				   0 << memif->irqdata->en_shift);
+		/* and clear pending IRQ */
+		regmap_write(afe->regmap, AFE_IRQ_CLR,
+			     1 << memif->irqdata->clr_shift);
+		memif->hw_ptr = 0;
+		return 0;
+	default:
+		return -EINVAL;
+	}
+}
+
+static const struct snd_soc_dai_ops mtk_afe_dai_ops = {
+	.startup	= mtk_afe_dais_startup,
+	.shutdown	= mtk_afe_dais_shutdown,
+	.hw_params	= mtk_afe_dais_hw_params,
+	.prepare	= mtk_afe_dais_prepare,
+	.trigger	= mtk_afe_dais_trigger,
+};
+
+static const struct mtk_afe_io_data mtk_afe_io_data[MTK_AFE_IO_NUM] = {
+	[MTK_AFE_IO_PMIC] = {
+		.num = MTK_AFE_IO_PMIC,
+		.name = "pmic",
+		.shutdown = mtk_afe_pmic_shutdown,
+		.prepare = mtk_afe_pmic_prepare,
+	},
+	[MTK_AFE_IO_I2S] = {
+		.num = MTK_AFE_IO_I2S,
+		.name = "i2s",
+		.shutdown = mtk_afe_i2s_shutdown,
+		.prepare = mtk_afe_i2s_prepare,
+	},
+	[MTK_AFE_IO_2ND_I2S] = {
+		.num = MTK_AFE_IO_2ND_I2S,
+		.name = "i2s2",
+		.shutdown = mtk_afe_i2s2_shutdown,
+		.prepare = mtk_afe_i2s2_prepare,
+	},
+	[MTK_AFE_IO_HDMI] = {
+		.num = MTK_AFE_IO_HDMI,
+		.name = "hdmi",
+		.prepare = mtk_afe_hdmi_prepare,
+		.start = mtk_afe_hdmi_start,
+		.pause = mtk_afe_hdmi_pause,
+	},
+};
+
+static struct snd_soc_dai_driver mtk_afe_pcm_dais[] = {
+	{
+		.name  = "I2S",
+		.id  = MTK_AFE_IO_I2S,
+		.playback = {
+			.stream_name = "I2S Playback",
+			.channels_min = 1,
+			.channels_max = 2,
+			.rates = SNDRV_PCM_RATE_8000_48000,
+			.formats = SNDRV_PCM_FMTBIT_S16_LE,
+		},
+		.capture = {
+			.stream_name = "I2S Capture",
+			.channels_min = 1,
+			.channels_max = 2,
+			.rates = SNDRV_PCM_RATE_8000_48000,
+			.formats = SNDRV_PCM_FMTBIT_S16_LE,
+		},
+		.ops = &mtk_afe_dai_ops,
+		.symmetric_rates = 1,
+	}, {
+		.name = "2nd I2S",
+		.id  = MTK_AFE_IO_2ND_I2S,
+		.playback = {
+			.stream_name = "2nd I2S Playback",
+			.channels_min = 1,
+			.channels_max = 2,
+			.rates = SNDRV_PCM_RATE_8000_48000,
+			.formats = SNDRV_PCM_FMTBIT_S16_LE,
+		},
+		.capture = {
+			.stream_name = "2nd I2S Capture",
+			.channels_min = 1,
+			.channels_max = 2,
+			.rates = SNDRV_PCM_RATE_8000_48000,
+			.formats = SNDRV_PCM_FMTBIT_S16_LE,
+		},
+		.ops = &mtk_afe_dai_ops,
+		.symmetric_rates = 1,
+	}, {
+		.name = "HDMI",
+		.id  = MTK_AFE_IO_HDMI,
+		.playback = {
+			.stream_name = "HDMI Playback",
+			.channels_min = 2,
+			.channels_max = 8,
+			.rates = SNDRV_PCM_RATE_32000 | SNDRV_PCM_RATE_44100 |
+				SNDRV_PCM_RATE_48000 | SNDRV_PCM_RATE_88200 |
+				SNDRV_PCM_RATE_96000 | SNDRV_PCM_RATE_176400 |
+				SNDRV_PCM_RATE_192000,
+			.formats = SNDRV_PCM_FMTBIT_S16_LE,
+		},
+		.ops = &mtk_afe_dai_ops,
+	}, {
+		.name = "PMIC",
+		.id  = MTK_AFE_IO_PMIC,
+		.playback = {
+			.stream_name = "PMIC Playback",
+			.channels_min = 1,
+			.channels_max = 2,
+			.rates = SNDRV_PCM_RATE_8000_48000,
+			.formats = SNDRV_PCM_FMTBIT_S16_LE,
+		},
+		.capture = {
+			.stream_name = "PMIC Capture",
+			.channels_min = 1,
+			.channels_max = 2,
+			.rates = SNDRV_PCM_RATE_8000_48000,
+			.formats = SNDRV_PCM_FMTBIT_S16_LE,
+		},
+		.ops = &mtk_afe_dai_ops,
+	},
+};
+
+static const struct snd_soc_component_driver mtk_afe_pcm_dai_component = {
+	.name		= "mtk-afe-pcm-dai",
+};
+
+static const char *aud_clks[MTK_CLK_NUM] = {
+	[MTK_CLK_INFRASYS_AUD] = "infra_sys_audio_clk",
+	[MTK_CLK_TOP_PDN_AUD] = "top_pdn_audio",
+	[MTK_CLK_TOP_PDN_AUD_BUS] = "top_pdn_aud_intbus",
+	[MTK_CLK_I2S0_M] =  "i2s0_m",
+	[MTK_CLK_I2S1_M] =  "i2s1_m",
+	[MTK_CLK_I2S2_M] =  "i2s2_m",
+	[MTK_CLK_I2S3_M] =  "i2s3_m",
+	[MTK_CLK_I2S3_B] =  "i2s3_b",
+	[MTK_CLK_BCK0] =  "bck0",
+	[MTK_CLK_BCK1] =  "bck1",
+};
+
+static const struct mtk_afe_irq_data irq_data[MTK_AFE_IRQ_NUM] = {
+	[MTK_AFE_IRQ_1] = {
+		.reg_cnt = AFE_IRQ_CNT1,
+		.cnt_shift = 0,
+		.en_shift = 0,
+		.fs_shift = 4,
+		.clr_shift = 0,
+	},
+	[MTK_AFE_IRQ_2] = {
+		.reg_cnt = AFE_IRQ_CNT2,
+		.cnt_shift = 0,
+		.en_shift = 1,
+		.fs_shift = 8,
+		.clr_shift = 1,
+	},
+	[MTK_AFE_IRQ_3] = {
+		.reg_cnt = AFE_IRQ_CNT1,
+		.cnt_shift = 20,
+		.en_shift = 2,
+		.fs_shift = 16,
+		.clr_shift = 2,
+	},
+	[MTK_AFE_IRQ_4] = {
+		.reg_cnt = AFE_IRQ_CNT2,
+		.cnt_shift = 20,
+		.en_shift = 3,
+		.fs_shift = 20,
+		.clr_shift = 3,
+	},
+	[MTK_AFE_IRQ_5] = {
+		.reg_cnt = AFE_IRQ_CNT5,
+		.cnt_shift = 0,
+		.en_shift = 12,
+		.fs_shift = -1,
+		.clr_shift = 4,
+	},
+	[MTK_AFE_IRQ_7] = {
+		.reg_cnt = AFE_IRQ_CNT7,
+		.cnt_shift = 0,
+		.en_shift = 14,
+		.fs_shift = 24,
+		.clr_shift = 6,
+	},
+};
+
+static const struct mtk_afe_memif_data memif_data[MTK_AFE_MEMIF_NUM] = {
+	{
+		.name = "DL1",
+		.id = MTK_AFE_MEMIF_DL1,
+		.reg_ofs_base = AFE_DL1_BASE,
+		.reg_ofs_cur = AFE_DL1_CUR,
+		.fs_shift = 0,
+		.mono_shift = 21,
+		.enable_shift = 1,
+	}, {
+		.name = "DL2",
+		.id = MTK_AFE_MEMIF_DL2,
+		.reg_ofs_base = AFE_DL2_BASE,
+		.reg_ofs_cur = AFE_DL2_CUR,
+		.fs_shift = 4,
+		.mono_shift = 22,
+		.enable_shift = 2,
+	}, {
+		.name = "VUL",
+		.id = MTK_AFE_MEMIF_VUL,
+		.reg_ofs_base = AFE_VUL_BASE,
+		.reg_ofs_cur = AFE_VUL_CUR,
+		.fs_shift = 16,
+		.mono_shift = 27,
+		.enable_shift = 3,
+	}, {
+		.name = "DAI",
+		.id = MTK_AFE_MEMIF_DAI,
+		.reg_ofs_base = AFE_DAI_BASE,
+		.reg_ofs_cur = AFE_DAI_CUR,
+		.fs_shift = 24,
+		.mono_shift = -1,
+		.enable_shift = 4,
+	}, {
+		.name = "AWB",
+		.id = MTK_AFE_MEMIF_AWB,
+		.reg_ofs_base = AFE_AWB_BASE,
+		.reg_ofs_cur = AFE_AWB_CUR,
+		.fs_shift = 12,
+		.mono_shift = 24,
+		.enable_shift = 6,
+	}, {
+		.name = "MOD_DAI",
+		.id = MTK_AFE_MEMIF_MOD_DAI,
+		.reg_ofs_base = AFE_MOD_PCM_BASE,
+		.reg_ofs_cur = AFE_MOD_PCM_CUR,
+		.fs_shift = 30,
+		.mono_shift = 30,
+		.enable_shift = 7,
+	}, {
+		.name = "HDMI",
+		.id = MTK_AFE_MEMIF_HDMI,
+		.reg_ofs_base = AFE_HDMI_OUT_BASE,
+		.reg_ofs_cur = AFE_HDMI_OUT_CUR,
+		.fs_shift = -1,
+		.mono_shift = -1,
+		.enable_shift = -1,
+	},
+};
+
+static const struct regmap_config mtk_afe_regmap_config = {
+	.reg_bits = 32,
+	.reg_stride = 4,
+	.val_bits = 32,
+	.max_register = AFE_ADDA2_TOP_CON0,
+	.cache_type = REGCACHE_NONE,
+};
+
+static irqreturn_t mtk_afe_irq_handler(int irq, void *dev_id)
+{
+	struct mtk_afe *afe = dev_id;
+	unsigned int reg_value, hw_ptr;
+	int i, ret;
+
+	ret = regmap_read(afe->regmap, AFE_IRQ_STATUS, &reg_value);
+	if (ret) {
+		dev_err(afe->dev, "%s irq status err\n", __func__);
+		reg_value = AFE_IRQ_STATUS_BITS;
+		goto err_irq;
+	}
+
+	for (i = 0; i < MTK_AFE_MEMIF_NUM; i++) {
+		struct mtk_afe_memif *memif = &afe->memif[i];
+
+		if (!memif->irqdata)
+			continue;
+		if (!(reg_value & (1 << memif->irqdata->clr_shift)))
+			continue;
+
+		ret = regmap_read(afe->regmap, memif->data->reg_ofs_cur,
+				  &hw_ptr);
+		if (ret || hw_ptr == 0) {
+			dev_err(afe->dev, "%s hw_ptr err\n", __func__);
+			hw_ptr = memif->phys_buf_addr;
+		}
+		memif->hw_ptr = hw_ptr - memif->phys_buf_addr;
+		snd_pcm_period_elapsed(memif->substream);
+	}
+
+err_irq:
+	/* clear irq */
+	regmap_write(afe->regmap, AFE_IRQ_CLR, reg_value & AFE_IRQ_STATUS_BITS);
+
+	return IRQ_HANDLED;
+}
+
+static int mtk_afe_runtime_suspend(struct device *dev)
+{
+	struct mtk_afe *afe = dev_get_drvdata(dev);
+
+	regmap_update_bits(afe->regmap, AUDIO_TOP_CON0, 1 << 2, 1 << 2);
+
+	clk_disable_unprepare(afe->clocks[MTK_CLK_BCK0]);
+	clk_disable_unprepare(afe->clocks[MTK_CLK_BCK1]);
+	clk_disable_unprepare(afe->clocks[MTK_CLK_TOP_PDN_AUD]);
+	clk_disable_unprepare(afe->clocks[MTK_CLK_TOP_PDN_AUD_BUS]);
+	clk_disable_unprepare(afe->clocks[MTK_CLK_INFRASYS_AUD]);
+	return 0;
+}
+
+static int mtk_afe_runtime_resume(struct device *dev)
+{
+	struct mtk_afe *afe = dev_get_drvdata(dev);
+	int ret;
+
+	ret = clk_prepare_enable(afe->clocks[MTK_CLK_INFRASYS_AUD]);
+	if (ret)
+		return ret;
+
+	ret = clk_prepare_enable(afe->clocks[MTK_CLK_TOP_PDN_AUD_BUS]);
+	if (ret)
+		goto err_infra;
+
+	ret = clk_prepare_enable(afe->clocks[MTK_CLK_TOP_PDN_AUD]);
+	if (ret)
+		goto err_top_aud_bus;
+
+	ret = clk_prepare_enable(afe->clocks[MTK_CLK_BCK0]);
+	if (ret)
+		goto err_top_aud;
+	ret = clk_prepare_enable(afe->clocks[MTK_CLK_BCK1]);
+	if (ret)
+		goto err_bck0;
+
+	/* enable AFE clk */
+	regmap_update_bits(afe->regmap, AUDIO_TOP_CON0, 1 << 2, 0);
+
+	/* set O3/O4 16bits */
+	regmap_update_bits(afe->regmap, AFE_CONN_24BIT, (1 << 3) | (1 << 4), 0);
+
+	/* unmask all IRQs */
+	regmap_update_bits(afe->regmap, AFE_IRQ_MCU_EN, 0xff, 0xff);
+	return 0;
+
+err_bck0:
+	clk_disable_unprepare(afe->clocks[MTK_CLK_BCK0]);
+err_top_aud:
+	clk_disable_unprepare(afe->clocks[MTK_CLK_TOP_PDN_AUD]);
+err_top_aud_bus:
+	clk_disable_unprepare(afe->clocks[MTK_CLK_TOP_PDN_AUD_BUS]);
+err_infra:
+	clk_disable_unprepare(afe->clocks[MTK_CLK_INFRASYS_AUD]);
+	return ret;
+}
+
+static int mtk_afe_init_audio_clk(struct mtk_afe *afe)
+{
+	size_t i;
+
+	for (i = 0; i < ARRAY_SIZE(aud_clks); i++) {
+		afe->clocks[i] = devm_clk_get(afe->dev, aud_clks[i]);
+		if (IS_ERR(afe->clocks[i])) {
+			dev_err(afe->dev, "%s devm_clk_get %s fail\n",
+				__func__, aud_clks[i]);
+			return PTR_ERR(afe->clocks[i]);
+		}
+	}
+	afe->ios[MTK_AFE_IO_2ND_I2S].m_ck = afe->clocks[MTK_CLK_I2S0_M];
+	afe->ios[MTK_AFE_IO_I2S].m_ck = afe->clocks[MTK_CLK_I2S1_M];
+	afe->ios[MTK_AFE_IO_HDMI].m_ck = afe->clocks[MTK_CLK_I2S3_M];
+	afe->ios[MTK_AFE_IO_HDMI].b_ck = afe->clocks[MTK_CLK_I2S3_B];
+
+	clk_set_rate(afe->clocks[MTK_CLK_BCK0], 22579200);
+	clk_set_rate(afe->clocks[MTK_CLK_BCK1], 24576000);
+	return 0;
+}
+
+static int mtk_afe_parse_memif(struct mtk_afe *afe, struct device_node *np,
+			       const char *propname, int *mem)
+{
+	struct mtk_afe_memif *memif;
+	int ret;
+	u32 val;
+	static bool use_sram;
+	*mem = -1;
+	ret = of_property_read_u32_index(np, propname, 0, &val);
+	if (ret)
+		/* to be check later since we may have only play or capture */
+		return 0;
+	if (val > MTK_AFE_MEMIF_NUM) {
+		dev_err(afe->dev, "invalid memif %d\n", val);
+		return -EINVAL;
+	}
+	*mem = val;
+	memif = &afe->memif[val];
+
+	ret = of_property_read_u32_index(np, propname, 1, &val);
+	if (ret)
+		return ret;
+	if (val >= MTK_AFE_IRQ_NUM) {
+		dev_err(afe->dev, "invalid mem irq %d\n", val);
+		return -EINVAL;
+	}
+	memif->irqdata = &irq_data[val];
+
+	ret = of_property_read_u32_index(np, propname, 2, &val);
+	if (ret)
+		return ret;
+	if (val && use_sram)
+		return -EINVAL; /* only one memif can use SRAM */
+	memif->use_sram = val ? true : false;
+	use_sram = memif->use_sram;
+	return 0;
+}
+
+static int mtk_afe_init_subnode(struct mtk_afe *afe, struct device_node *np)
+{
+	struct mtk_afe_io *io;
+	u32 val;
+	int ret, n, i;
+
+	ret = of_property_read_u32(np, "io", &val);
+	if (ret)
+		return ret;
+	if (val >= MTK_AFE_IO_NUM) {
+		dev_err(afe->dev, "invalid io %d\n", val);
+		return -EINVAL;
+	}
+	io = &afe->ios[val];
+
+	n = of_property_count_u32_elems(np, "connections");
+	if (n <= 0)
+		return n;
+	io->connections = devm_kzalloc(afe->dev, n * sizeof(u32),
+			GFP_KERNEL);
+	if (!io->connections)
+		return -ENOMEM;
+	io->num_connections = n / 2;
+	ret = of_property_read_u32_array(np, "connections", io->connections,
+					 n);
+	if (ret)
+		return ret;
+	for (i = 0; i < io->num_connections; i++) {
+		dev_dbg(afe->dev, "connection %d: %d -> %d\n", i,
+			io->connections[i * 2],
+			io->connections[i * 2 + 1]);
+	}
+
+	ret = mtk_afe_parse_memif(afe, np, "mem-interface-playback", io->mem);
+	if (ret)
+		return ret;
+	ret = mtk_afe_parse_memif(afe, np, "mem-interface-capture",
+				  io->mem + 1);
+	if (ret)
+		return ret;
+	if (io->mem[0] < 0 && io->mem[1] < 0) {
+		dev_err(afe->dev, "no mem interface given\n");
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static int mtk_afe_pcm_dev_probe(struct platform_device *pdev)
+{
+	int ret, i;
+	unsigned int irq_id;
+	struct mtk_afe *afe;
+	struct device_node *dainode;
+	struct resource *res;
+
+	afe = devm_kzalloc(&pdev->dev, sizeof(*afe), GFP_KERNEL);
+	if (!afe)
+		return -ENOMEM;
+
+	afe->dev = &pdev->dev;
+
+	irq_id = platform_get_irq(pdev, 0);
+	if (!irq_id) {
+		dev_err(afe->dev, "np %s no irq\n", afe->dev->of_node->name);
+		return -ENXIO;
+	}
+	ret = devm_request_irq(afe->dev, irq_id, mtk_afe_irq_handler,
+			       0, "Afe_ISR_Handle", (void *)afe);
+	if (ret) {
+		dev_err(afe->dev, "could not request_irq\n");
+		return ret;
+	}
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	afe->base_addr = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(afe->base_addr))
+		return PTR_ERR(afe->base_addr);
+
+	afe->regmap = devm_regmap_init_mmio(&pdev->dev, afe->base_addr,
+		&mtk_afe_regmap_config);
+	if (IS_ERR(afe->regmap))
+		return PTR_ERR(afe->regmap);
+
+	/* get SRAM base */
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 1);
+	afe->sram_address = devm_ioremap_resource(&pdev->dev, res);
+	if (!IS_ERR(afe->sram_address)) {
+		afe->sram_phy_address = res->start;
+		afe->sram_size = resource_size(res);
+	}
+
+	/* initial audio related clock */
+	ret = mtk_afe_init_audio_clk(afe);
+	if (ret) {
+		dev_err(afe->dev, "mtk_afe_init_audio_clk fail\n");
+		return ret;
+	}
+
+	for (i = 0; i < MTK_AFE_MEMIF_NUM; i++)
+		afe->memif[i].data = &memif_data[i];
+
+	for (i = 0; i < MTK_AFE_IO_NUM; i++)
+		afe->ios[i].data = &mtk_afe_io_data[i];
+
+	for_each_child_of_node(pdev->dev.of_node, dainode) {
+		ret = mtk_afe_init_subnode(afe, dainode);
+		if (ret)
+			return ret;
+	}
+
+	platform_set_drvdata(pdev, afe);
+
+	pm_runtime_enable(&pdev->dev);
+	if (!pm_runtime_enabled(&pdev->dev)) {
+		ret = mtk_afe_runtime_resume(&pdev->dev);
+		if (ret)
+			goto err_pm_disable;
+	}
+
+	ret = snd_soc_register_platform(&pdev->dev, &mtk_afe_pcm_platform);
+	if (ret)
+		goto err_pm_disable;
+
+	ret = snd_soc_register_component(&pdev->dev,
+					 &mtk_afe_pcm_dai_component,
+					 mtk_afe_pcm_dais,
+					 ARRAY_SIZE(mtk_afe_pcm_dais));
+	if (ret)
+		goto err_platform;
+
+	dev_info(&pdev->dev, "MTK AFE driver initialized.\n");
+	return 0;
+
+err_platform:
+	snd_soc_unregister_platform(&pdev->dev);
+err_pm_disable:
+	pm_runtime_disable(&pdev->dev);
+	return ret;
+}
+
+static int mtk_afe_pcm_dev_remove(struct platform_device *pdev)
+{
+	pm_runtime_disable(&pdev->dev);
+	snd_soc_unregister_component(&pdev->dev);
+	snd_soc_unregister_platform(&pdev->dev);
+	return 0;
+}
+
+static const struct of_device_id mtk_afe_pcm_dt_match[] = {
+	{ .compatible = "mediatek,mt8173-afe-pcm", },
+	{ }
+};
+
+static const struct dev_pm_ops mtk_afe_pm_ops = {
+	SET_RUNTIME_PM_OPS(mtk_afe_runtime_suspend, mtk_afe_runtime_resume,
+			   NULL)
+};
+
+static struct platform_driver mtk_afe_pcm_driver = {
+	.driver = {
+		   .name = "mtk-afe-pcm",
+		   .owner = THIS_MODULE,
+		   .of_match_table = mtk_afe_pcm_dt_match,
+		   .pm = &mtk_afe_pm_ops,
+	},
+	.probe = mtk_afe_pcm_dev_probe,
+	.remove = mtk_afe_pcm_dev_remove,
+};
+
+module_platform_driver(mtk_afe_pcm_driver);
+
+MODULE_DESCRIPTION("Mediatek ALSA SoC Afe platform driver");
+MODULE_AUTHOR("Koro Chen <koro.chen@mediatek.com>");
+MODULE_LICENSE("GPL v2");
+MODULE_DEVICE_TABLE(of, mtk_afe_pcm_dt_match);
-- 
1.8.1.1.dirty

Re: [RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 1306 bytes --]

On Fri, Apr 10, 2015 at 04:14:07PM +0800, Koro Chen wrote:

> +Each external interface (called "IO" in this driver) is presented as a
> +DAI to ASoC. An IO must be connected via the interconnect to a memif.
> +The connection paths are configured through the device tree.

Why are these connection paths configured via device tree?  I would
expect that either there would be runtime configurability of these
things (particularly if loopback configurations within the hardware are
possible) or we'd just allocate memory interfaces to DAIs automatically
as DAIs come into use.

> +- mem-interface-playback:
> +  mem-interface-capture: property of memif, format is: <memif irq use_sram>;
> +	                 memif: which memif to be used
> +			        (defined in include/dt-bindings/sound/mtk-afe.h)
> +		         irq: which irq to be used
> +			      (defined in include/dt-bindings/sound/mtk-afe.h)
> +		         use_sram: 1 is yes, 0 is no

Again, this looks like stuff we should be able to figure out at runtime
- the use of SRAM in particular looks like something we might want to
change depending on use case.  Assuming it adds buffering then for a
VoIP application we might not want to use SRAM to minimize latency but
during music playback we might want to enable SRAM to minimize power
consumption.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

[RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Mark Brown]

On Fri, Apr 10, 2015 at 04:14:07PM +0800, Koro Chen wrote:

> +Each external interface (called "IO" in this driver) is presented as a
> +DAI to ASoC. An IO must be connected via the interconnect to a memif.
> +The connection paths are configured through the device tree.

Why are these connection paths configured via device tree?  I would
expect that either there would be runtime configurability of these
things (particularly if loopback configurations within the hardware are
possible) or we'd just allocate memory interfaces to DAIs automatically
as DAIs come into use.

> +- mem-interface-playback:
> +  mem-interface-capture: property of memif, format is: <memif irq use_sram>;
> +	                 memif: which memif to be used
> +			        (defined in include/dt-bindings/sound/mtk-afe.h)
> +		         irq: which irq to be used
> +			      (defined in include/dt-bindings/sound/mtk-afe.h)
> +		         use_sram: 1 is yes, 0 is no

Again, this looks like stuff we should be able to figure out at runtime
- the use of SRAM in particular looks like something we might want to
change depending on use case.  Assuming it adds buffering then for a
VoIP application we might not want to use SRAM to minimize latency but
during music playback we might want to enable SRAM to minimize power
consumption.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20150418/f9a9b312/attachment.sig>

Re: [RESEND RFC PATCH 2/3] ASoC: mediatek: Add AFE connection control

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 925 bytes --]

On Fri, Apr 10, 2015 at 04:14:08PM +0800, Koro Chen wrote:

> +/*
> + * The MTK AFE unit has a audio interconnect with MTK_AFE_INTERCONN_NUM_INPUT
> + * inputs and MTK_AFE_INTERCONN_NUM_OUTPUT outputs. Below table holds the
> + * register/bits to set to connect an input with an output.
> + */
> +static const struct mtk_afe_connection
> +	connections[MTK_AFE_INTERCONN_NUM_INPUT][MTK_AFE_INTERCONN_NUM_OUTPUT] = {
> +	[0][0] =   { .creg = 0x020, .cshift =  0, .sreg = 0x020, .sshift = 10},
> +	[0][1] =   { .creg = 0x020, .cshift = 16, .sreg = 0x020, .sshift = 26},

What are the constraints on using this - is it one input to one output
only or can we do mixing or duplication?  The register interface
definitely does look like something asking for runtime configuration.

It'd also be nice to have less magic numbers in the table, at least for
the indexes (which I guess correspond to some of the defines in the
headers)?

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

Re: [RESEND RFC PATCH 2/3] ASoC: mediatek: Add AFE connection control

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 925 bytes --]

On Fri, Apr 10, 2015 at 04:14:08PM +0800, Koro Chen wrote:

> +/*
> + * The MTK AFE unit has a audio interconnect with MTK_AFE_INTERCONN_NUM_INPUT
> + * inputs and MTK_AFE_INTERCONN_NUM_OUTPUT outputs. Below table holds the
> + * register/bits to set to connect an input with an output.
> + */
> +static const struct mtk_afe_connection
> +	connections[MTK_AFE_INTERCONN_NUM_INPUT][MTK_AFE_INTERCONN_NUM_OUTPUT] = {
> +	[0][0] =   { .creg = 0x020, .cshift =  0, .sreg = 0x020, .sshift = 10},
> +	[0][1] =   { .creg = 0x020, .cshift = 16, .sreg = 0x020, .sshift = 26},

What are the constraints on using this - is it one input to one output
only or can we do mixing or duplication?  The register interface
definitely does look like something asking for runtime configuration.

It'd also be nice to have less magic numbers in the table, at least for
the indexes (which I guess correspond to some of the defines in the
headers)?

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

[RESEND RFC PATCH 2/3] ASoC: mediatek: Add AFE connection control

[Mark Brown]

On Fri, Apr 10, 2015 at 04:14:08PM +0800, Koro Chen wrote:

> +/*
> + * The MTK AFE unit has a audio interconnect with MTK_AFE_INTERCONN_NUM_INPUT
> + * inputs and MTK_AFE_INTERCONN_NUM_OUTPUT outputs. Below table holds the
> + * register/bits to set to connect an input with an output.
> + */
> +static const struct mtk_afe_connection
> +	connections[MTK_AFE_INTERCONN_NUM_INPUT][MTK_AFE_INTERCONN_NUM_OUTPUT] = {
> +	[0][0] =   { .creg = 0x020, .cshift =  0, .sreg = 0x020, .sshift = 10},
> +	[0][1] =   { .creg = 0x020, .cshift = 16, .sreg = 0x020, .sshift = 26},

What are the constraints on using this - is it one input to one output
only or can we do mixing or duplication?  The register interface
definitely does look like something asking for runtime configuration.

It'd also be nice to have less magic numbers in the table, at least for
the indexes (which I guess correspond to some of the defines in the
headers)?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20150418/28fcacb5/attachment.sig>

Re: [RESEND RFC PATCH 3/3] ASoC: mediatek: Add AFE platform driver

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 2683 bytes --]

On Fri, Apr 10, 2015 at 04:14:09PM +0800, Koro Chen wrote:

> +	if (memif->use_sram) {
> +		struct snd_dma_buffer *dma_buf = &substream->dma_buffer;
> +		int size = params_buffer_bytes(params);
> +
> +		memif->buffer_size = size;
> +		memif->phys_buf_addr = afe->sram_phy_address;
> +
> +		dma_buf->bytes = size;
> +		dma_buf->area = (unsigned char *)afe->sram_address;
> +		dma_buf->addr = afe->sram_phy_address;
> +		dma_buf->dev.type = SNDRV_DMA_TYPE_DEV;
> +		dma_buf->dev.dev = substream->pcm->card->dev;
> +		snd_pcm_set_runtime_buffer(substream, dma_buf);
> +	} else {
> +		ret = snd_pcm_lib_malloc_pages(substream,
> +					       params_buffer_bytes(params));
> +		if (ret < 0)
> +			return ret;
> +
> +		memif->phys_buf_addr = substream->runtime->dma_addr;
> +		memif->buffer_size = substream->runtime->dma_bytes;
> +	}

Ah, so the SRAM is directly memory mappable.  Nice.  But we have a
limited amount of it so need to allocate it to a device somehow based on
some factor I guess?

> +static int mtk_afe_set_adda_dac_out(struct mtk_afe *afe, uint32_t rate)
> +{
> +	u32 audio_i2s_dac = 0;
> +	u32 con0, con1;
> +
> +	/* set dl src2 */
> +	con0 = (mtk_afe_adda_fs(rate) << 28) | (0x03 << 24) | (0x03 << 11);
> +
> +	/* 8k or 16k voice mode */
> +	if (con0 == 0 || con0 == 3)
> +		con0 |= 0x01 << 5;

This all looks a bit magic, some defines would not go amiss here.

> +	/* SA suggests to apply -0.3db to audio/speech path */
> +	con0 = con0 | (0x01 << 1);
> +	con1 = 0xf74f0000;

More magic numbers!  This also suggests that there is a volume control
lurking in here which could usefully be exposed to applications?

> +static void mtk_afe_pmic_shutdown(struct mtk_afe *afe,
> +				  struct snd_pcm_substream *substream)
> +{
> +	/* output */
> +	regmap_update_bits(afe->regmap, AFE_ADDA_DL_SRC2_CON0, 1, 0);
> +	regmap_update_bits(afe->regmap, AFE_I2S_CON1, 1, 0);
> +
> +	/* input */
> +	regmap_update_bits(afe->regmap, AFE_ADDA_UL_SRC_CON0, 1, 0);
> +	/* disable ADDA */
> +	regmap_update_bits(afe->regmap, AFE_ADDA_UL_DL_CON0, 1, false);
> +}

This is looking like exposing the routing and using DAPM might save a
bunch of code?  Overall my main thought looking at the code here and
what the hardware was described as doing is that it'd all be simpler if
it were a DPCM based thing using DAPM for power.  I think I'd like to
see a strong reason for not using at least DPCM.

> +		if (rate == MTK_AFE_I2S_RATE_8K)
> +			voice_mode = 0;
> +		else if (rate == MTK_AFE_I2S_RATE_16K)
> +			voice_mode = 1;
> +		else if (rate == MTK_AFE_I2S_RATE_32K)
> +			voice_mode = 2;
> +		else if (rate == MTK_AFE_I2S_RATE_48K)
> +			voice_mode = 3;

This should be a switch statement.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

[RESEND RFC PATCH 3/3] ASoC: mediatek: Add AFE platform driver

[Mark Brown]

On Fri, Apr 10, 2015 at 04:14:09PM +0800, Koro Chen wrote:

> +	if (memif->use_sram) {
> +		struct snd_dma_buffer *dma_buf = &substream->dma_buffer;
> +		int size = params_buffer_bytes(params);
> +
> +		memif->buffer_size = size;
> +		memif->phys_buf_addr = afe->sram_phy_address;
> +
> +		dma_buf->bytes = size;
> +		dma_buf->area = (unsigned char *)afe->sram_address;
> +		dma_buf->addr = afe->sram_phy_address;
> +		dma_buf->dev.type = SNDRV_DMA_TYPE_DEV;
> +		dma_buf->dev.dev = substream->pcm->card->dev;
> +		snd_pcm_set_runtime_buffer(substream, dma_buf);
> +	} else {
> +		ret = snd_pcm_lib_malloc_pages(substream,
> +					       params_buffer_bytes(params));
> +		if (ret < 0)
> +			return ret;
> +
> +		memif->phys_buf_addr = substream->runtime->dma_addr;
> +		memif->buffer_size = substream->runtime->dma_bytes;
> +	}

Ah, so the SRAM is directly memory mappable.  Nice.  But we have a
limited amount of it so need to allocate it to a device somehow based on
some factor I guess?

> +static int mtk_afe_set_adda_dac_out(struct mtk_afe *afe, uint32_t rate)
> +{
> +	u32 audio_i2s_dac = 0;
> +	u32 con0, con1;
> +
> +	/* set dl src2 */
> +	con0 = (mtk_afe_adda_fs(rate) << 28) | (0x03 << 24) | (0x03 << 11);
> +
> +	/* 8k or 16k voice mode */
> +	if (con0 == 0 || con0 == 3)
> +		con0 |= 0x01 << 5;

This all looks a bit magic, some defines would not go amiss here.

> +	/* SA suggests to apply -0.3db to audio/speech path */
> +	con0 = con0 | (0x01 << 1);
> +	con1 = 0xf74f0000;

More magic numbers!  This also suggests that there is a volume control
lurking in here which could usefully be exposed to applications?

> +static void mtk_afe_pmic_shutdown(struct mtk_afe *afe,
> +				  struct snd_pcm_substream *substream)
> +{
> +	/* output */
> +	regmap_update_bits(afe->regmap, AFE_ADDA_DL_SRC2_CON0, 1, 0);
> +	regmap_update_bits(afe->regmap, AFE_I2S_CON1, 1, 0);
> +
> +	/* input */
> +	regmap_update_bits(afe->regmap, AFE_ADDA_UL_SRC_CON0, 1, 0);
> +	/* disable ADDA */
> +	regmap_update_bits(afe->regmap, AFE_ADDA_UL_DL_CON0, 1, false);
> +}

This is looking like exposing the routing and using DAPM might save a
bunch of code?  Overall my main thought looking at the code here and
what the hardware was described as doing is that it'd all be simpler if
it were a DPCM based thing using DAPM for power.  I think I'd like to
see a strong reason for not using at least DPCM.

> +		if (rate == MTK_AFE_I2S_RATE_8K)
> +			voice_mode = 0;
> +		else if (rate == MTK_AFE_I2S_RATE_16K)
> +			voice_mode = 1;
> +		else if (rate == MTK_AFE_I2S_RATE_32K)
> +			voice_mode = 2;
> +		else if (rate == MTK_AFE_I2S_RATE_48K)
> +			voice_mode = 3;

This should be a switch statement.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20150418/ca74b09e/attachment.sig>

Re: [RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Sascha Hauer]

On Sat, Apr 18, 2015 at 06:34:07PM +0100, Mark Brown wrote:
> On Fri, Apr 10, 2015 at 04:14:07PM +0800, Koro Chen wrote:
> 
> > +Each external interface (called "IO" in this driver) is presented as a
> > +DAI to ASoC. An IO must be connected via the interconnect to a memif.
> > +The connection paths are configured through the device tree.
> 
> Why are these connection paths configured via device tree?  I would
> expect that either there would be runtime configurability of these
> things (particularly if loopback configurations within the hardware are
> possible) or we'd just allocate memory interfaces to DAIs automatically
> as DAIs come into use.

There is a crossbar switch between the memory interfaces and the DAIs.
Not every connection is possible, so not every memory interface can be
used for every DAI. An algorithm choosing a suitable memory interface
must be quite clever, complicated and also SoC dependent (the same but
different hardware is used on MT8135 aswell), so I thought offering a
static configuration via device tree is a good start. Should there be
runtime configuration possible later the device tree settings could
provide a good default.

> 
> > +- mem-interface-playback:
> > +  mem-interface-capture: property of memif, format is: <memif irq use_sram>;
> > +	                 memif: which memif to be used
> > +			        (defined in include/dt-bindings/sound/mtk-afe.h)
> > +		         irq: which irq to be used
> > +			      (defined in include/dt-bindings/sound/mtk-afe.h)
> > +		         use_sram: 1 is yes, 0 is no
> 
> Again, this looks like stuff we should be able to figure out at runtime
> - the use of SRAM in particular looks like something we might want to
> change depending on use case.  Assuming it adds buffering then for a
> VoIP application we might not want to use SRAM to minimize latency but
> during music playback we might want to enable SRAM to minimize power
> consumption.

That's exactly the usecase. How could such a runtime configurability
look like? sysfs? Or something based on the buffer sizes?

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

Re: [RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Sascha Hauer]

On Sat, Apr 18, 2015 at 06:34:07PM +0100, Mark Brown wrote:
> On Fri, Apr 10, 2015 at 04:14:07PM +0800, Koro Chen wrote:
> 
> > +Each external interface (called "IO" in this driver) is presented as a
> > +DAI to ASoC. An IO must be connected via the interconnect to a memif.
> > +The connection paths are configured through the device tree.
> 
> Why are these connection paths configured via device tree?  I would
> expect that either there would be runtime configurability of these
> things (particularly if loopback configurations within the hardware are
> possible) or we'd just allocate memory interfaces to DAIs automatically
> as DAIs come into use.

There is a crossbar switch between the memory interfaces and the DAIs.
Not every connection is possible, so not every memory interface can be
used for every DAI. An algorithm choosing a suitable memory interface
must be quite clever, complicated and also SoC dependent (the same but
different hardware is used on MT8135 aswell), so I thought offering a
static configuration via device tree is a good start. Should there be
runtime configuration possible later the device tree settings could
provide a good default.

> 
> > +- mem-interface-playback:
> > +  mem-interface-capture: property of memif, format is: <memif irq use_sram>;
> > +	                 memif: which memif to be used
> > +			        (defined in include/dt-bindings/sound/mtk-afe.h)
> > +		         irq: which irq to be used
> > +			      (defined in include/dt-bindings/sound/mtk-afe.h)
> > +		         use_sram: 1 is yes, 0 is no
> 
> Again, this looks like stuff we should be able to figure out at runtime
> - the use of SRAM in particular looks like something we might want to
> change depending on use case.  Assuming it adds buffering then for a
> VoIP application we might not want to use SRAM to minimize latency but
> during music playback we might want to enable SRAM to minimize power
> consumption.

That's exactly the usecase. How could such a runtime configurability
look like? sysfs? Or something based on the buffer sizes?

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

[RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Sascha Hauer]

On Sat, Apr 18, 2015 at 06:34:07PM +0100, Mark Brown wrote:
> On Fri, Apr 10, 2015 at 04:14:07PM +0800, Koro Chen wrote:
> 
> > +Each external interface (called "IO" in this driver) is presented as a
> > +DAI to ASoC. An IO must be connected via the interconnect to a memif.
> > +The connection paths are configured through the device tree.
> 
> Why are these connection paths configured via device tree?  I would
> expect that either there would be runtime configurability of these
> things (particularly if loopback configurations within the hardware are
> possible) or we'd just allocate memory interfaces to DAIs automatically
> as DAIs come into use.

There is a crossbar switch between the memory interfaces and the DAIs.
Not every connection is possible, so not every memory interface can be
used for every DAI. An algorithm choosing a suitable memory interface
must be quite clever, complicated and also SoC dependent (the same but
different hardware is used on MT8135 aswell), so I thought offering a
static configuration via device tree is a good start. Should there be
runtime configuration possible later the device tree settings could
provide a good default.

> 
> > +- mem-interface-playback:
> > +  mem-interface-capture: property of memif, format is: <memif irq use_sram>;
> > +	                 memif: which memif to be used
> > +			        (defined in include/dt-bindings/sound/mtk-afe.h)
> > +		         irq: which irq to be used
> > +			      (defined in include/dt-bindings/sound/mtk-afe.h)
> > +		         use_sram: 1 is yes, 0 is no
> 
> Again, this looks like stuff we should be able to figure out at runtime
> - the use of SRAM in particular looks like something we might want to
> change depending on use case.  Assuming it adds buffering then for a
> VoIP application we might not want to use SRAM to minimize latency but
> during music playback we might want to enable SRAM to minimize power
> consumption.

That's exactly the usecase. How could such a runtime configurability
look like? sysfs? Or something based on the buffer sizes?

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

Re: [RESEND RFC PATCH 2/3] ASoC: mediatek: Add AFE connection control

[Sascha Hauer]

On Sat, Apr 18, 2015 at 06:37:40PM +0100, Mark Brown wrote:
> On Fri, Apr 10, 2015 at 04:14:08PM +0800, Koro Chen wrote:
> 
> > +/*
> > + * The MTK AFE unit has a audio interconnect with MTK_AFE_INTERCONN_NUM_INPUT
> > + * inputs and MTK_AFE_INTERCONN_NUM_OUTPUT outputs. Below table holds the
> > + * register/bits to set to connect an input with an output.
> > + */
> > +static const struct mtk_afe_connection
> > +	connections[MTK_AFE_INTERCONN_NUM_INPUT][MTK_AFE_INTERCONN_NUM_OUTPUT] = {
> > +	[0][0] =   { .creg = 0x020, .cshift =  0, .sreg = 0x020, .sshift = 10},
> > +	[0][1] =   { .creg = 0x020, .cshift = 16, .sreg = 0x020, .sshift = 26},
> 
> What are the constraints on using this - is it one input to one output
> only or can we do mixing or duplication?  The register interface
> definitely does look like something asking for runtime configuration.
> 
> It'd also be nice to have less magic numbers in the table, at least for
> the indexes (which I guess correspond to some of the defines in the
> headers)?

With defines the above two lines would become something like:

#define AFE_CONN0		0x20

#define CONN0_I00_O00_S		0
...
#define CONN0_I00_O00_R		10
...
#define CONN0_I00_O01_s		16
...
#define CONN0_I00_O01_S		26

[0][0] = { .creg = AFE_CONN0, .cshift = CONN0_I00_O00_S, .sreg = AFE_CONN0, .sshift = CONN0_I00_O00_R },
[0][1] = { .creg = AFE_CONN0, .cshift = CONN0_I00_O01_S, .sreg = AFE_CONN0, .sshift = CONN0_I00_O01_S },

For the registers we could use defines, but I think using defines for
the shifts doesn't add much value given they are only used once.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

Re: [RESEND RFC PATCH 2/3] ASoC: mediatek: Add AFE connection control

[Sascha Hauer]

On Sat, Apr 18, 2015 at 06:37:40PM +0100, Mark Brown wrote:
> On Fri, Apr 10, 2015 at 04:14:08PM +0800, Koro Chen wrote:
> 
> > +/*
> > + * The MTK AFE unit has a audio interconnect with MTK_AFE_INTERCONN_NUM_INPUT
> > + * inputs and MTK_AFE_INTERCONN_NUM_OUTPUT outputs. Below table holds the
> > + * register/bits to set to connect an input with an output.
> > + */
> > +static const struct mtk_afe_connection
> > +	connections[MTK_AFE_INTERCONN_NUM_INPUT][MTK_AFE_INTERCONN_NUM_OUTPUT] = {
> > +	[0][0] =   { .creg = 0x020, .cshift =  0, .sreg = 0x020, .sshift = 10},
> > +	[0][1] =   { .creg = 0x020, .cshift = 16, .sreg = 0x020, .sshift = 26},
> 
> What are the constraints on using this - is it one input to one output
> only or can we do mixing or duplication?  The register interface
> definitely does look like something asking for runtime configuration.
> 
> It'd also be nice to have less magic numbers in the table, at least for
> the indexes (which I guess correspond to some of the defines in the
> headers)?

With defines the above two lines would become something like:

#define AFE_CONN0		0x20

#define CONN0_I00_O00_S		0
...
#define CONN0_I00_O00_R		10
...
#define CONN0_I00_O01_s		16
...
#define CONN0_I00_O01_S		26

[0][0] = { .creg = AFE_CONN0, .cshift = CONN0_I00_O00_S, .sreg = AFE_CONN0, .sshift = CONN0_I00_O00_R },
[0][1] = { .creg = AFE_CONN0, .cshift = CONN0_I00_O01_S, .sreg = AFE_CONN0, .sshift = CONN0_I00_O01_S },

For the registers we could use defines, but I think using defines for
the shifts doesn't add much value given they are only used once.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[RESEND RFC PATCH 2/3] ASoC: mediatek: Add AFE connection control

[Sascha Hauer]

On Sat, Apr 18, 2015 at 06:37:40PM +0100, Mark Brown wrote:
> On Fri, Apr 10, 2015 at 04:14:08PM +0800, Koro Chen wrote:
> 
> > +/*
> > + * The MTK AFE unit has a audio interconnect with MTK_AFE_INTERCONN_NUM_INPUT
> > + * inputs and MTK_AFE_INTERCONN_NUM_OUTPUT outputs. Below table holds the
> > + * register/bits to set to connect an input with an output.
> > + */
> > +static const struct mtk_afe_connection
> > +	connections[MTK_AFE_INTERCONN_NUM_INPUT][MTK_AFE_INTERCONN_NUM_OUTPUT] = {
> > +	[0][0] =   { .creg = 0x020, .cshift =  0, .sreg = 0x020, .sshift = 10},
> > +	[0][1] =   { .creg = 0x020, .cshift = 16, .sreg = 0x020, .sshift = 26},
> 
> What are the constraints on using this - is it one input to one output
> only or can we do mixing or duplication?  The register interface
> definitely does look like something asking for runtime configuration.
> 
> It'd also be nice to have less magic numbers in the table, at least for
> the indexes (which I guess correspond to some of the defines in the
> headers)?

With defines the above two lines would become something like:

#define AFE_CONN0		0x20

#define CONN0_I00_O00_S		0
...
#define CONN0_I00_O00_R		10
...
#define CONN0_I00_O01_s		16
...
#define CONN0_I00_O01_S		26

[0][0] = { .creg = AFE_CONN0, .cshift = CONN0_I00_O00_S, .sreg = AFE_CONN0, .sshift = CONN0_I00_O00_R },
[0][1] = { .creg = AFE_CONN0, .cshift = CONN0_I00_O01_S, .sreg = AFE_CONN0, .sshift = CONN0_I00_O01_S },

For the registers we could use defines, but I think using defines for
the shifts doesn't add much value given they are only used once.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

Re: [RESEND RFC PATCH 3/3] ASoC: mediatek: Add AFE platform driver

[Koro Chen]

On Sat, 2015-04-18 at 18:51 +0100, Mark Brown wrote:
> On Fri, Apr 10, 2015 at 04:14:09PM +0800, Koro Chen wrote:
> 
> > +	if (memif->use_sram) {
> > +		struct snd_dma_buffer *dma_buf = &substream->dma_buffer;
> > +		int size = params_buffer_bytes(params);
> > +
> > +		memif->buffer_size = size;
> > +		memif->phys_buf_addr = afe->sram_phy_address;
> > +
> > +		dma_buf->bytes = size;
> > +		dma_buf->area = (unsigned char *)afe->sram_address;
> > +		dma_buf->addr = afe->sram_phy_address;
> > +		dma_buf->dev.type = SNDRV_DMA_TYPE_DEV;
> > +		dma_buf->dev.dev = substream->pcm->card->dev;
> > +		snd_pcm_set_runtime_buffer(substream, dma_buf);
> > +	} else {
> > +		ret = snd_pcm_lib_malloc_pages(substream,
> > +					       params_buffer_bytes(params));
> > +		if (ret < 0)
> > +			return ret;
> > +
> > +		memif->phys_buf_addr = substream->runtime->dma_addr;
> > +		memif->buffer_size = substream->runtime->dma_bytes;
> > +	}
> 
> Ah, so the SRAM is directly memory mappable.  Nice.  But we have a
> limited amount of it so need to allocate it to a device somehow based on
> some factor I guess?
Yes, actually SRAM is only used for the main playback path (which is
memif "DL1") to achieve low power in real use case. Maybe you think it's
better to not describe this in the device tree, but to choose SRAM
automatically if memif "DL1" is chosen?
> 
> > +static int mtk_afe_set_adda_dac_out(struct mtk_afe *afe, uint32_t rate)
> > +{
> > +	u32 audio_i2s_dac = 0;
> > +	u32 con0, con1;
> > +
> > +	/* set dl src2 */
> > +	con0 = (mtk_afe_adda_fs(rate) << 28) | (0x03 << 24) | (0x03 << 11);
> > +
> > +	/* 8k or 16k voice mode */
> > +	if (con0 == 0 || con0 == 3)
> > +		con0 |= 0x01 << 5;
> 
> This all looks a bit magic, some defines would not go amiss here.
> 
> > +	/* SA suggests to apply -0.3db to audio/speech path */
> > +	con0 = con0 | (0x01 << 1);
> > +	con1 = 0xf74f0000;
> 
> More magic numbers!  This also suggests that there is a volume control
> lurking in here which could usefully be exposed to applications?
> 
Sorry, I will fix these magic numbers.
It is actually not a real volume control and not that suitable for
application control because it cannot be changed in runtime; its value
is fixed and was decided off-lined by experiment that can have best
audio quality when using with our proprietary codec (not for I2S)

> > +static void mtk_afe_pmic_shutdown(struct mtk_afe *afe,
> > +				  struct snd_pcm_substream *substream)
> > +{
> > +	/* output */
> > +	regmap_update_bits(afe->regmap, AFE_ADDA_DL_SRC2_CON0, 1, 0);
> > +	regmap_update_bits(afe->regmap, AFE_I2S_CON1, 1, 0);
> > +
> > +	/* input */
> > +	regmap_update_bits(afe->regmap, AFE_ADDA_UL_SRC_CON0, 1, 0);
> > +	/* disable ADDA */
> > +	regmap_update_bits(afe->regmap, AFE_ADDA_UL_DL_CON0, 1, false);
> > +}
> 
> This is looking like exposing the routing and using DAPM might save a
> bunch of code?  Overall my main thought looking at the code here and
> what the hardware was described as doing is that it'd all be simpler if
> it were a DPCM based thing using DAPM for power.  I think I'd like to
> see a strong reason for not using at least DPCM.
Thank you very much for mentioning the DPCM. I didn't know much about
DPCM and I will definitely study and check if it is suitable for our HW.
> 
> > +		if (rate == MTK_AFE_I2S_RATE_8K)
> > +			voice_mode = 0;
> > +		else if (rate == MTK_AFE_I2S_RATE_16K)
> > +			voice_mode = 1;
> > +		else if (rate == MTK_AFE_I2S_RATE_32K)
> > +			voice_mode = 2;
> > +		else if (rate == MTK_AFE_I2S_RATE_48K)
> > +			voice_mode = 3;
> 
> This should be a switch statement.

Re: [RESEND RFC PATCH 3/3] ASoC: mediatek: Add AFE platform driver

[Koro Chen]

On Sat, 2015-04-18 at 18:51 +0100, Mark Brown wrote:
> On Fri, Apr 10, 2015 at 04:14:09PM +0800, Koro Chen wrote:
> 
> > +	if (memif->use_sram) {
> > +		struct snd_dma_buffer *dma_buf = &substream->dma_buffer;
> > +		int size = params_buffer_bytes(params);
> > +
> > +		memif->buffer_size = size;
> > +		memif->phys_buf_addr = afe->sram_phy_address;
> > +
> > +		dma_buf->bytes = size;
> > +		dma_buf->area = (unsigned char *)afe->sram_address;
> > +		dma_buf->addr = afe->sram_phy_address;
> > +		dma_buf->dev.type = SNDRV_DMA_TYPE_DEV;
> > +		dma_buf->dev.dev = substream->pcm->card->dev;
> > +		snd_pcm_set_runtime_buffer(substream, dma_buf);
> > +	} else {
> > +		ret = snd_pcm_lib_malloc_pages(substream,
> > +					       params_buffer_bytes(params));
> > +		if (ret < 0)
> > +			return ret;
> > +
> > +		memif->phys_buf_addr = substream->runtime->dma_addr;
> > +		memif->buffer_size = substream->runtime->dma_bytes;
> > +	}
> 
> Ah, so the SRAM is directly memory mappable.  Nice.  But we have a
> limited amount of it so need to allocate it to a device somehow based on
> some factor I guess?
Yes, actually SRAM is only used for the main playback path (which is
memif "DL1") to achieve low power in real use case. Maybe you think it's
better to not describe this in the device tree, but to choose SRAM
automatically if memif "DL1" is chosen?
> 
> > +static int mtk_afe_set_adda_dac_out(struct mtk_afe *afe, uint32_t rate)
> > +{
> > +	u32 audio_i2s_dac = 0;
> > +	u32 con0, con1;
> > +
> > +	/* set dl src2 */
> > +	con0 = (mtk_afe_adda_fs(rate) << 28) | (0x03 << 24) | (0x03 << 11);
> > +
> > +	/* 8k or 16k voice mode */
> > +	if (con0 == 0 || con0 == 3)
> > +		con0 |= 0x01 << 5;
> 
> This all looks a bit magic, some defines would not go amiss here.
> 
> > +	/* SA suggests to apply -0.3db to audio/speech path */
> > +	con0 = con0 | (0x01 << 1);
> > +	con1 = 0xf74f0000;
> 
> More magic numbers!  This also suggests that there is a volume control
> lurking in here which could usefully be exposed to applications?
> 
Sorry, I will fix these magic numbers.
It is actually not a real volume control and not that suitable for
application control because it cannot be changed in runtime; its value
is fixed and was decided off-lined by experiment that can have best
audio quality when using with our proprietary codec (not for I2S)

> > +static void mtk_afe_pmic_shutdown(struct mtk_afe *afe,
> > +				  struct snd_pcm_substream *substream)
> > +{
> > +	/* output */
> > +	regmap_update_bits(afe->regmap, AFE_ADDA_DL_SRC2_CON0, 1, 0);
> > +	regmap_update_bits(afe->regmap, AFE_I2S_CON1, 1, 0);
> > +
> > +	/* input */
> > +	regmap_update_bits(afe->regmap, AFE_ADDA_UL_SRC_CON0, 1, 0);
> > +	/* disable ADDA */
> > +	regmap_update_bits(afe->regmap, AFE_ADDA_UL_DL_CON0, 1, false);
> > +}
> 
> This is looking like exposing the routing and using DAPM might save a
> bunch of code?  Overall my main thought looking at the code here and
> what the hardware was described as doing is that it'd all be simpler if
> it were a DPCM based thing using DAPM for power.  I think I'd like to
> see a strong reason for not using at least DPCM.
Thank you very much for mentioning the DPCM. I didn't know much about
DPCM and I will definitely study and check if it is suitable for our HW.
> 
> > +		if (rate == MTK_AFE_I2S_RATE_8K)
> > +			voice_mode = 0;
> > +		else if (rate == MTK_AFE_I2S_RATE_16K)
> > +			voice_mode = 1;
> > +		else if (rate == MTK_AFE_I2S_RATE_32K)
> > +			voice_mode = 2;
> > +		else if (rate == MTK_AFE_I2S_RATE_48K)
> > +			voice_mode = 3;
> 
> This should be a switch statement.


--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[RESEND RFC PATCH 3/3] ASoC: mediatek: Add AFE platform driver

[Koro Chen]

On Sat, 2015-04-18 at 18:51 +0100, Mark Brown wrote:
> On Fri, Apr 10, 2015 at 04:14:09PM +0800, Koro Chen wrote:
> 
> > +	if (memif->use_sram) {
> > +		struct snd_dma_buffer *dma_buf = &substream->dma_buffer;
> > +		int size = params_buffer_bytes(params);
> > +
> > +		memif->buffer_size = size;
> > +		memif->phys_buf_addr = afe->sram_phy_address;
> > +
> > +		dma_buf->bytes = size;
> > +		dma_buf->area = (unsigned char *)afe->sram_address;
> > +		dma_buf->addr = afe->sram_phy_address;
> > +		dma_buf->dev.type = SNDRV_DMA_TYPE_DEV;
> > +		dma_buf->dev.dev = substream->pcm->card->dev;
> > +		snd_pcm_set_runtime_buffer(substream, dma_buf);
> > +	} else {
> > +		ret = snd_pcm_lib_malloc_pages(substream,
> > +					       params_buffer_bytes(params));
> > +		if (ret < 0)
> > +			return ret;
> > +
> > +		memif->phys_buf_addr = substream->runtime->dma_addr;
> > +		memif->buffer_size = substream->runtime->dma_bytes;
> > +	}
> 
> Ah, so the SRAM is directly memory mappable.  Nice.  But we have a
> limited amount of it so need to allocate it to a device somehow based on
> some factor I guess?
Yes, actually SRAM is only used for the main playback path (which is
memif "DL1") to achieve low power in real use case. Maybe you think it's
better to not describe this in the device tree, but to choose SRAM
automatically if memif "DL1" is chosen?
> 
> > +static int mtk_afe_set_adda_dac_out(struct mtk_afe *afe, uint32_t rate)
> > +{
> > +	u32 audio_i2s_dac = 0;
> > +	u32 con0, con1;
> > +
> > +	/* set dl src2 */
> > +	con0 = (mtk_afe_adda_fs(rate) << 28) | (0x03 << 24) | (0x03 << 11);
> > +
> > +	/* 8k or 16k voice mode */
> > +	if (con0 == 0 || con0 == 3)
> > +		con0 |= 0x01 << 5;
> 
> This all looks a bit magic, some defines would not go amiss here.
> 
> > +	/* SA suggests to apply -0.3db to audio/speech path */
> > +	con0 = con0 | (0x01 << 1);
> > +	con1 = 0xf74f0000;
> 
> More magic numbers!  This also suggests that there is a volume control
> lurking in here which could usefully be exposed to applications?
> 
Sorry, I will fix these magic numbers.
It is actually not a real volume control and not that suitable for
application control because it cannot be changed in runtime; its value
is fixed and was decided off-lined by experiment that can have best
audio quality when using with our proprietary codec (not for I2S)

> > +static void mtk_afe_pmic_shutdown(struct mtk_afe *afe,
> > +				  struct snd_pcm_substream *substream)
> > +{
> > +	/* output */
> > +	regmap_update_bits(afe->regmap, AFE_ADDA_DL_SRC2_CON0, 1, 0);
> > +	regmap_update_bits(afe->regmap, AFE_I2S_CON1, 1, 0);
> > +
> > +	/* input */
> > +	regmap_update_bits(afe->regmap, AFE_ADDA_UL_SRC_CON0, 1, 0);
> > +	/* disable ADDA */
> > +	regmap_update_bits(afe->regmap, AFE_ADDA_UL_DL_CON0, 1, false);
> > +}
> 
> This is looking like exposing the routing and using DAPM might save a
> bunch of code?  Overall my main thought looking at the code here and
> what the hardware was described as doing is that it'd all be simpler if
> it were a DPCM based thing using DAPM for power.  I think I'd like to
> see a strong reason for not using at least DPCM.
Thank you very much for mentioning the DPCM. I didn't know much about
DPCM and I will definitely study and check if it is suitable for our HW.
> 
> > +		if (rate == MTK_AFE_I2S_RATE_8K)
> > +			voice_mode = 0;
> > +		else if (rate == MTK_AFE_I2S_RATE_16K)
> > +			voice_mode = 1;
> > +		else if (rate == MTK_AFE_I2S_RATE_32K)
> > +			voice_mode = 2;
> > +		else if (rate == MTK_AFE_I2S_RATE_48K)
> > +			voice_mode = 3;
> 
> This should be a switch statement.

Re: [RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 2813 bytes --]

On Mon, Apr 20, 2015 at 06:37:47AM +0200, Sascha Hauer wrote:
> On Sat, Apr 18, 2015 at 06:34:07PM +0100, Mark Brown wrote:
> > On Fri, Apr 10, 2015 at 04:14:07PM +0800, Koro Chen wrote:

> > > +Each external interface (called "IO" in this driver) is presented as a
> > > +DAI to ASoC. An IO must be connected via the interconnect to a memif.
> > > +The connection paths are configured through the device tree.

> > Why are these connection paths configured via device tree?  I would
> > expect that either there would be runtime configurability of these
> > things (particularly if loopback configurations within the hardware are
> > possible) or we'd just allocate memory interfaces to DAIs automatically
> > as DAIs come into use.

> There is a crossbar switch between the memory interfaces and the DAIs.
> Not every connection is possible, so not every memory interface can be
> used for every DAI. An algorithm choosing a suitable memory interface
> must be quite clever, complicated and also SoC dependent (the same but
> different hardware is used on MT8135 aswell), so I thought offering a
> static configuration via device tree is a good start. Should there be
> runtime configuration possible later the device tree settings could
> provide a good default.

What exactly do the restrictions look like and how often do they vary in
practice (can we get away with just doing a single static setup in the
driver)?  I'd have thought it should be fairly straightforward to have a
table of valid mappings and just pick the first free memory interface?
I'd rather not get stuck with the tables in the DT.  It's partly to
avoid setting bad precendents, we really don't want everyone coming
along hard coding this stuff, and partly because the hardware you
described didn't seem that hard to handle.

> > > +- mem-interface-playback:
> > > +  mem-interface-capture: property of memif, format is: <memif irq use_sram>;
> > > +	                 memif: which memif to be used
> > > +			        (defined in include/dt-bindings/sound/mtk-afe.h)
> > > +		         irq: which irq to be used
> > > +			      (defined in include/dt-bindings/sound/mtk-afe.h)
> > > +		         use_sram: 1 is yes, 0 is no

> > Again, this looks like stuff we should be able to figure out at runtime
> > - the use of SRAM in particular looks like something we might want to
> > change depending on use case.  Assuming it adds buffering then for a
> > VoIP application we might not want to use SRAM to minimize latency but
> > during music playback we might want to enable SRAM to minimize power
> > consumption.

> That's exactly the usecase. How could such a runtime configurability
> look like? sysfs? Or something based on the buffer sizes?

Yeah, one of those :) but probably an ALSA control is going to be the
easiest for applications.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

Re: [RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 2813 bytes --]

On Mon, Apr 20, 2015 at 06:37:47AM +0200, Sascha Hauer wrote:
> On Sat, Apr 18, 2015 at 06:34:07PM +0100, Mark Brown wrote:
> > On Fri, Apr 10, 2015 at 04:14:07PM +0800, Koro Chen wrote:

> > > +Each external interface (called "IO" in this driver) is presented as a
> > > +DAI to ASoC. An IO must be connected via the interconnect to a memif.
> > > +The connection paths are configured through the device tree.

> > Why are these connection paths configured via device tree?  I would
> > expect that either there would be runtime configurability of these
> > things (particularly if loopback configurations within the hardware are
> > possible) or we'd just allocate memory interfaces to DAIs automatically
> > as DAIs come into use.

> There is a crossbar switch between the memory interfaces and the DAIs.
> Not every connection is possible, so not every memory interface can be
> used for every DAI. An algorithm choosing a suitable memory interface
> must be quite clever, complicated and also SoC dependent (the same but
> different hardware is used on MT8135 aswell), so I thought offering a
> static configuration via device tree is a good start. Should there be
> runtime configuration possible later the device tree settings could
> provide a good default.

What exactly do the restrictions look like and how often do they vary in
practice (can we get away with just doing a single static setup in the
driver)?  I'd have thought it should be fairly straightforward to have a
table of valid mappings and just pick the first free memory interface?
I'd rather not get stuck with the tables in the DT.  It's partly to
avoid setting bad precendents, we really don't want everyone coming
along hard coding this stuff, and partly because the hardware you
described didn't seem that hard to handle.

> > > +- mem-interface-playback:
> > > +  mem-interface-capture: property of memif, format is: <memif irq use_sram>;
> > > +	                 memif: which memif to be used
> > > +			        (defined in include/dt-bindings/sound/mtk-afe.h)
> > > +		         irq: which irq to be used
> > > +			      (defined in include/dt-bindings/sound/mtk-afe.h)
> > > +		         use_sram: 1 is yes, 0 is no

> > Again, this looks like stuff we should be able to figure out at runtime
> > - the use of SRAM in particular looks like something we might want to
> > change depending on use case.  Assuming it adds buffering then for a
> > VoIP application we might not want to use SRAM to minimize latency but
> > during music playback we might want to enable SRAM to minimize power
> > consumption.

> That's exactly the usecase. How could such a runtime configurability
> look like? sysfs? Or something based on the buffer sizes?

Yeah, one of those :) but probably an ALSA control is going to be the
easiest for applications.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

[RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Mark Brown]

On Mon, Apr 20, 2015 at 06:37:47AM +0200, Sascha Hauer wrote:
> On Sat, Apr 18, 2015 at 06:34:07PM +0100, Mark Brown wrote:
> > On Fri, Apr 10, 2015 at 04:14:07PM +0800, Koro Chen wrote:

> > > +Each external interface (called "IO" in this driver) is presented as a
> > > +DAI to ASoC. An IO must be connected via the interconnect to a memif.
> > > +The connection paths are configured through the device tree.

> > Why are these connection paths configured via device tree?  I would
> > expect that either there would be runtime configurability of these
> > things (particularly if loopback configurations within the hardware are
> > possible) or we'd just allocate memory interfaces to DAIs automatically
> > as DAIs come into use.

> There is a crossbar switch between the memory interfaces and the DAIs.
> Not every connection is possible, so not every memory interface can be
> used for every DAI. An algorithm choosing a suitable memory interface
> must be quite clever, complicated and also SoC dependent (the same but
> different hardware is used on MT8135 aswell), so I thought offering a
> static configuration via device tree is a good start. Should there be
> runtime configuration possible later the device tree settings could
> provide a good default.

What exactly do the restrictions look like and how often do they vary in
practice (can we get away with just doing a single static setup in the
driver)?  I'd have thought it should be fairly straightforward to have a
table of valid mappings and just pick the first free memory interface?
I'd rather not get stuck with the tables in the DT.  It's partly to
avoid setting bad precendents, we really don't want everyone coming
along hard coding this stuff, and partly because the hardware you
described didn't seem that hard to handle.

> > > +- mem-interface-playback:
> > > +  mem-interface-capture: property of memif, format is: <memif irq use_sram>;
> > > +	                 memif: which memif to be used
> > > +			        (defined in include/dt-bindings/sound/mtk-afe.h)
> > > +		         irq: which irq to be used
> > > +			      (defined in include/dt-bindings/sound/mtk-afe.h)
> > > +		         use_sram: 1 is yes, 0 is no

> > Again, this looks like stuff we should be able to figure out at runtime
> > - the use of SRAM in particular looks like something we might want to
> > change depending on use case.  Assuming it adds buffering then for a
> > VoIP application we might not want to use SRAM to minimize latency but
> > during music playback we might want to enable SRAM to minimize power
> > consumption.

> That's exactly the usecase. How could such a runtime configurability
> look like? sysfs? Or something based on the buffer sizes?

Yeah, one of those :) but probably an ALSA control is going to be the
easiest for applications.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20150420/c4e99187/attachment.sig>

Re: [RESEND RFC PATCH 2/3] ASoC: mediatek: Add AFE connection control

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 1013 bytes --]

On Mon, Apr 20, 2015 at 06:50:17AM +0200, Sascha Hauer wrote:
> On Sat, Apr 18, 2015 at 06:37:40PM +0100, Mark Brown wrote:
> > On Fri, Apr 10, 2015 at 04:14:08PM +0800, Koro Chen wrote:

> > > +	[0][0] =   { .creg = 0x020, .cshift =  0, .sreg = 0x020, .sshift = 10},
> > > +	[0][1] =   { .creg = 0x020, .cshift = 16, .sreg = 0x020, .sshift = 26},

> > It'd also be nice to have less magic numbers in the table, at least for
> > the indexes (which I guess correspond to some of the defines in the
> > headers)?

> With defines the above two lines would become something like:

> [0][0] = { .creg = AFE_CONN0, .cshift = CONN0_I00_O00_S, .sreg = AFE_CONN0, .sshift = CONN0_I00_O00_R },
> [0][1] = { .creg = AFE_CONN0, .cshift = CONN0_I00_O01_S, .sreg = AFE_CONN0, .sshift = CONN0_I00_O01_S },

> For the registers we could use defines, but I think using defines for
> the shifts doesn't add much value given they are only used once.

By indexes I actually meant the [0][0] and so on - they seem the more
magic bit.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

Re: [RESEND RFC PATCH 2/3] ASoC: mediatek: Add AFE connection control

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 1013 bytes --]

On Mon, Apr 20, 2015 at 06:50:17AM +0200, Sascha Hauer wrote:
> On Sat, Apr 18, 2015 at 06:37:40PM +0100, Mark Brown wrote:
> > On Fri, Apr 10, 2015 at 04:14:08PM +0800, Koro Chen wrote:

> > > +	[0][0] =   { .creg = 0x020, .cshift =  0, .sreg = 0x020, .sshift = 10},
> > > +	[0][1] =   { .creg = 0x020, .cshift = 16, .sreg = 0x020, .sshift = 26},

> > It'd also be nice to have less magic numbers in the table, at least for
> > the indexes (which I guess correspond to some of the defines in the
> > headers)?

> With defines the above two lines would become something like:

> [0][0] = { .creg = AFE_CONN0, .cshift = CONN0_I00_O00_S, .sreg = AFE_CONN0, .sshift = CONN0_I00_O00_R },
> [0][1] = { .creg = AFE_CONN0, .cshift = CONN0_I00_O01_S, .sreg = AFE_CONN0, .sshift = CONN0_I00_O01_S },

> For the registers we could use defines, but I think using defines for
> the shifts doesn't add much value given they are only used once.

By indexes I actually meant the [0][0] and so on - they seem the more
magic bit.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

[RESEND RFC PATCH 2/3] ASoC: mediatek: Add AFE connection control

[Mark Brown]

On Mon, Apr 20, 2015 at 06:50:17AM +0200, Sascha Hauer wrote:
> On Sat, Apr 18, 2015 at 06:37:40PM +0100, Mark Brown wrote:
> > On Fri, Apr 10, 2015 at 04:14:08PM +0800, Koro Chen wrote:

> > > +	[0][0] =   { .creg = 0x020, .cshift =  0, .sreg = 0x020, .sshift = 10},
> > > +	[0][1] =   { .creg = 0x020, .cshift = 16, .sreg = 0x020, .sshift = 26},

> > It'd also be nice to have less magic numbers in the table, at least for
> > the indexes (which I guess correspond to some of the defines in the
> > headers)?

> With defines the above two lines would become something like:

> [0][0] = { .creg = AFE_CONN0, .cshift = CONN0_I00_O00_S, .sreg = AFE_CONN0, .sshift = CONN0_I00_O00_R },
> [0][1] = { .creg = AFE_CONN0, .cshift = CONN0_I00_O01_S, .sreg = AFE_CONN0, .sshift = CONN0_I00_O01_S },

> For the registers we could use defines, but I think using defines for
> the shifts doesn't add much value given they are only used once.

By indexes I actually meant the [0][0] and so on - they seem the more
magic bit.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20150420/5bcf6617/attachment.sig>

Re: [RESEND RFC PATCH 3/3] ASoC: mediatek: Add AFE platform driver

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 865 bytes --]

On Mon, Apr 20, 2015 at 02:22:24PM +0800, Koro Chen wrote:
> On Sat, 2015-04-18 at 18:51 +0100, Mark Brown wrote:
> > On Fri, Apr 10, 2015 at 04:14:09PM +0800, Koro Chen wrote:

> > Ah, so the SRAM is directly memory mappable.  Nice.  But we have a
> > limited amount of it so need to allocate it to a device somehow based on
> > some factor I guess?

> Yes, actually SRAM is only used for the main playback path (which is
> memif "DL1") to achieve low power in real use case. Maybe you think it's
> better to not describe this in the device tree, but to choose SRAM
> automatically if memif "DL1" is chosen?

Since it's directly memory mappable is there actually any cost in
latency terms from using the SRAM in low latency cases (or did I misread
what the code was doing there)?  If it can only be used with one
interface and there's no downside from using it...

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

Re: [RESEND RFC PATCH 3/3] ASoC: mediatek: Add AFE platform driver

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 865 bytes --]

On Mon, Apr 20, 2015 at 02:22:24PM +0800, Koro Chen wrote:
> On Sat, 2015-04-18 at 18:51 +0100, Mark Brown wrote:
> > On Fri, Apr 10, 2015 at 04:14:09PM +0800, Koro Chen wrote:

> > Ah, so the SRAM is directly memory mappable.  Nice.  But we have a
> > limited amount of it so need to allocate it to a device somehow based on
> > some factor I guess?

> Yes, actually SRAM is only used for the main playback path (which is
> memif "DL1") to achieve low power in real use case. Maybe you think it's
> better to not describe this in the device tree, but to choose SRAM
> automatically if memif "DL1" is chosen?

Since it's directly memory mappable is there actually any cost in
latency terms from using the SRAM in low latency cases (or did I misread
what the code was doing there)?  If it can only be used with one
interface and there's no downside from using it...

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

[RESEND RFC PATCH 3/3] ASoC: mediatek: Add AFE platform driver

[Mark Brown]

On Mon, Apr 20, 2015 at 02:22:24PM +0800, Koro Chen wrote:
> On Sat, 2015-04-18 at 18:51 +0100, Mark Brown wrote:
> > On Fri, Apr 10, 2015 at 04:14:09PM +0800, Koro Chen wrote:

> > Ah, so the SRAM is directly memory mappable.  Nice.  But we have a
> > limited amount of it so need to allocate it to a device somehow based on
> > some factor I guess?

> Yes, actually SRAM is only used for the main playback path (which is
> memif "DL1") to achieve low power in real use case. Maybe you think it's
> better to not describe this in the device tree, but to choose SRAM
> automatically if memif "DL1" is chosen?

Since it's directly memory mappable is there actually any cost in
latency terms from using the SRAM in low latency cases (or did I misread
what the code was doing there)?  If it can only be used with one
interface and there's no downside from using it...
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20150420/071b0bdc/attachment-0001.sig>

Re: [RESEND RFC PATCH 3/3] ASoC: mediatek: Add AFE platform driver

[Koro Chen]

On Mon, 2015-04-20 at 21:55 +0100, Mark Brown wrote:
> On Mon, Apr 20, 2015 at 02:22:24PM +0800, Koro Chen wrote:
> > On Sat, 2015-04-18 at 18:51 +0100, Mark Brown wrote:
> > > On Fri, Apr 10, 2015 at 04:14:09PM +0800, Koro Chen wrote:
> 
> > > Ah, so the SRAM is directly memory mappable.  Nice.  But we have a
> > > limited amount of it so need to allocate it to a device somehow based on
> > > some factor I guess?
> 
> > Yes, actually SRAM is only used for the main playback path (which is
> > memif "DL1") to achieve low power in real use case. Maybe you think it's
> > better to not describe this in the device tree, but to choose SRAM
> > automatically if memif "DL1" is chosen?
> 
> Since it's directly memory mappable is there actually any cost in
> latency terms from using the SRAM in low latency cases (or did I misread
> what the code was doing there)?  If it can only be used with one
> interface and there's no downside from using it...
The SRAM size to be used is defined by params_buffer_bytes(params), not
fixed (of course limited by the actual available SRAM size on HW), so
the latency should be the same compared to a DRAM having the same size. 

The SRAM can be used by any memif, and that's why the plan was let DT
make the decision.

Re: [RESEND RFC PATCH 3/3] ASoC: mediatek: Add AFE platform driver

[Koro Chen]

On Mon, 2015-04-20 at 21:55 +0100, Mark Brown wrote:
> On Mon, Apr 20, 2015 at 02:22:24PM +0800, Koro Chen wrote:
> > On Sat, 2015-04-18 at 18:51 +0100, Mark Brown wrote:
> > > On Fri, Apr 10, 2015 at 04:14:09PM +0800, Koro Chen wrote:
> 
> > > Ah, so the SRAM is directly memory mappable.  Nice.  But we have a
> > > limited amount of it so need to allocate it to a device somehow based on
> > > some factor I guess?
> 
> > Yes, actually SRAM is only used for the main playback path (which is
> > memif "DL1") to achieve low power in real use case. Maybe you think it's
> > better to not describe this in the device tree, but to choose SRAM
> > automatically if memif "DL1" is chosen?
> 
> Since it's directly memory mappable is there actually any cost in
> latency terms from using the SRAM in low latency cases (or did I misread
> what the code was doing there)?  If it can only be used with one
> interface and there's no downside from using it...
The SRAM size to be used is defined by params_buffer_bytes(params), not
fixed (of course limited by the actual available SRAM size on HW), so
the latency should be the same compared to a DRAM having the same size. 

The SRAM can be used by any memif, and that's why the plan was let DT
make the decision.

[RESEND RFC PATCH 3/3] ASoC: mediatek: Add AFE platform driver

[Koro Chen]

On Mon, 2015-04-20 at 21:55 +0100, Mark Brown wrote:
> On Mon, Apr 20, 2015 at 02:22:24PM +0800, Koro Chen wrote:
> > On Sat, 2015-04-18 at 18:51 +0100, Mark Brown wrote:
> > > On Fri, Apr 10, 2015 at 04:14:09PM +0800, Koro Chen wrote:
> 
> > > Ah, so the SRAM is directly memory mappable.  Nice.  But we have a
> > > limited amount of it so need to allocate it to a device somehow based on
> > > some factor I guess?
> 
> > Yes, actually SRAM is only used for the main playback path (which is
> > memif "DL1") to achieve low power in real use case. Maybe you think it's
> > better to not describe this in the device tree, but to choose SRAM
> > automatically if memif "DL1" is chosen?
> 
> Since it's directly memory mappable is there actually any cost in
> latency terms from using the SRAM in low latency cases (or did I misread
> what the code was doing there)?  If it can only be used with one
> interface and there's no downside from using it...
The SRAM size to be used is defined by params_buffer_bytes(params), not
fixed (of course limited by the actual available SRAM size on HW), so
the latency should be the same compared to a DRAM having the same size. 

The SRAM can be used by any memif, and that's why the plan was let DT
make the decision.

Re: [RESEND RFC PATCH 2/3] ASoC: mediatek: Add AFE connection control

[Sascha Hauer]

On Mon, Apr 20, 2015 at 09:52:30PM +0100, Mark Brown wrote:
> On Mon, Apr 20, 2015 at 06:50:17AM +0200, Sascha Hauer wrote:
> > On Sat, Apr 18, 2015 at 06:37:40PM +0100, Mark Brown wrote:
> > > On Fri, Apr 10, 2015 at 04:14:08PM +0800, Koro Chen wrote:
> 
> > > > +	[0][0] =   { .creg = 0x020, .cshift =  0, .sreg = 0x020, .sshift = 10},
> > > > +	[0][1] =   { .creg = 0x020, .cshift = 16, .sreg = 0x020, .sshift = 26},
> 
> > > It'd also be nice to have less magic numbers in the table, at least for
> > > the indexes (which I guess correspond to some of the defines in the
> > > headers)?
> 
> > With defines the above two lines would become something like:
> 
> > [0][0] = { .creg = AFE_CONN0, .cshift = CONN0_I00_O00_S, .sreg = AFE_CONN0, .sshift = CONN0_I00_O00_R },
> > [0][1] = { .creg = AFE_CONN0, .cshift = CONN0_I00_O01_S, .sreg = AFE_CONN0, .sshift = CONN0_I00_O01_S },
> 
> > For the registers we could use defines, but I think using defines for
> > the shifts doesn't add much value given they are only used once.
> 
> By indexes I actually meant the [0][0] and so on - they seem the more
> magic bit.

Oh, that's not magic at all, the crossbar switch has inputs and outputs
numbered from 0 to MTK_AFE_INTERCONN_NUM_INPUT /
MTK_AFE_INTERCONN_NUM_OUTPUT (they have the same numbers in the
datasheet. To connect input x with output y look at index [x][y] in the
table and write the register values found at that place. If .creg is 0x0
then it's not possible to connect the given input with the given output.

Sascha



-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

Re: [RESEND RFC PATCH 2/3] ASoC: mediatek: Add AFE connection control

[Sascha Hauer]

On Mon, Apr 20, 2015 at 09:52:30PM +0100, Mark Brown wrote:
> On Mon, Apr 20, 2015 at 06:50:17AM +0200, Sascha Hauer wrote:
> > On Sat, Apr 18, 2015 at 06:37:40PM +0100, Mark Brown wrote:
> > > On Fri, Apr 10, 2015 at 04:14:08PM +0800, Koro Chen wrote:
> 
> > > > +	[0][0] =   { .creg = 0x020, .cshift =  0, .sreg = 0x020, .sshift = 10},
> > > > +	[0][1] =   { .creg = 0x020, .cshift = 16, .sreg = 0x020, .sshift = 26},
> 
> > > It'd also be nice to have less magic numbers in the table, at least for
> > > the indexes (which I guess correspond to some of the defines in the
> > > headers)?
> 
> > With defines the above two lines would become something like:
> 
> > [0][0] = { .creg = AFE_CONN0, .cshift = CONN0_I00_O00_S, .sreg = AFE_CONN0, .sshift = CONN0_I00_O00_R },
> > [0][1] = { .creg = AFE_CONN0, .cshift = CONN0_I00_O01_S, .sreg = AFE_CONN0, .sshift = CONN0_I00_O01_S },
> 
> > For the registers we could use defines, but I think using defines for
> > the shifts doesn't add much value given they are only used once.
> 
> By indexes I actually meant the [0][0] and so on - they seem the more
> magic bit.

Oh, that's not magic at all, the crossbar switch has inputs and outputs
numbered from 0 to MTK_AFE_INTERCONN_NUM_INPUT /
MTK_AFE_INTERCONN_NUM_OUTPUT (they have the same numbers in the
datasheet. To connect input x with output y look at index [x][y] in the
table and write the register values found at that place. If .creg is 0x0
then it's not possible to connect the given input with the given output.

Sascha



-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

[RESEND RFC PATCH 2/3] ASoC: mediatek: Add AFE connection control

[Sascha Hauer]

On Mon, Apr 20, 2015 at 09:52:30PM +0100, Mark Brown wrote:
> On Mon, Apr 20, 2015 at 06:50:17AM +0200, Sascha Hauer wrote:
> > On Sat, Apr 18, 2015 at 06:37:40PM +0100, Mark Brown wrote:
> > > On Fri, Apr 10, 2015 at 04:14:08PM +0800, Koro Chen wrote:
> 
> > > > +	[0][0] =   { .creg = 0x020, .cshift =  0, .sreg = 0x020, .sshift = 10},
> > > > +	[0][1] =   { .creg = 0x020, .cshift = 16, .sreg = 0x020, .sshift = 26},
> 
> > > It'd also be nice to have less magic numbers in the table, at least for
> > > the indexes (which I guess correspond to some of the defines in the
> > > headers)?
> 
> > With defines the above two lines would become something like:
> 
> > [0][0] = { .creg = AFE_CONN0, .cshift = CONN0_I00_O00_S, .sreg = AFE_CONN0, .sshift = CONN0_I00_O00_R },
> > [0][1] = { .creg = AFE_CONN0, .cshift = CONN0_I00_O01_S, .sreg = AFE_CONN0, .sshift = CONN0_I00_O01_S },
> 
> > For the registers we could use defines, but I think using defines for
> > the shifts doesn't add much value given they are only used once.
> 
> By indexes I actually meant the [0][0] and so on - they seem the more
> magic bit.

Oh, that's not magic at all, the crossbar switch has inputs and outputs
numbered from 0 to MTK_AFE_INTERCONN_NUM_INPUT /
MTK_AFE_INTERCONN_NUM_OUTPUT (they have the same numbers in the
datasheet. To connect input x with output y look at index [x][y] in the
table and write the register values found at that place. If .creg is 0x0
then it's not possible to connect the given input with the given output.

Sascha



-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

Re: [RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Sascha Hauer]

On Mon, Apr 20, 2015 at 09:48:49PM +0100, Mark Brown wrote:
> On Mon, Apr 20, 2015 at 06:37:47AM +0200, Sascha Hauer wrote:
> > On Sat, Apr 18, 2015 at 06:34:07PM +0100, Mark Brown wrote:
> > > On Fri, Apr 10, 2015 at 04:14:07PM +0800, Koro Chen wrote:
> 
> > > > +Each external interface (called "IO" in this driver) is presented as a
> > > > +DAI to ASoC. An IO must be connected via the interconnect to a memif.
> > > > +The connection paths are configured through the device tree.
> 
> > > Why are these connection paths configured via device tree?  I would
> > > expect that either there would be runtime configurability of these
> > > things (particularly if loopback configurations within the hardware are
> > > possible) or we'd just allocate memory interfaces to DAIs automatically
> > > as DAIs come into use.
> 
> > There is a crossbar switch between the memory interfaces and the DAIs.
> > Not every connection is possible, so not every memory interface can be
> > used for every DAI. An algorithm choosing a suitable memory interface
> > must be quite clever, complicated and also SoC dependent (the same but
> > different hardware is used on MT8135 aswell), so I thought offering a
> > static configuration via device tree is a good start. Should there be
> > runtime configuration possible later the device tree settings could
> > provide a good default.
> 
> What exactly do the restrictions look like and how often do they vary in
> practice (can we get away with just doing a single static setup in the
> driver)? I'd have thought it should be fairly straightforward to have a
> table of valid mappings and just pick the first free memory interface?

I think this could be done. I checked the possible connections in the
crossbar switch and it seems all memory interfaces can be connected with
all relevant external interfaces. So indeed the memory interfaces could
be dynamically allocated instead of statically associated to an
external interface. There are two problems I see: Some memory interfaces
are limited in the rates they support, they can only do 8k/16k/32k (for
speech). How can we know such memory interface should be used? Also
there are two programmable hardware gain blocks which can be inserted to
the digital audio path using the crossbar switch. There must be some
mechanism to configure them into different places.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

Re: [RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Sascha Hauer]

On Mon, Apr 20, 2015 at 09:48:49PM +0100, Mark Brown wrote:
> On Mon, Apr 20, 2015 at 06:37:47AM +0200, Sascha Hauer wrote:
> > On Sat, Apr 18, 2015 at 06:34:07PM +0100, Mark Brown wrote:
> > > On Fri, Apr 10, 2015 at 04:14:07PM +0800, Koro Chen wrote:
> 
> > > > +Each external interface (called "IO" in this driver) is presented as a
> > > > +DAI to ASoC. An IO must be connected via the interconnect to a memif.
> > > > +The connection paths are configured through the device tree.
> 
> > > Why are these connection paths configured via device tree?  I would
> > > expect that either there would be runtime configurability of these
> > > things (particularly if loopback configurations within the hardware are
> > > possible) or we'd just allocate memory interfaces to DAIs automatically
> > > as DAIs come into use.
> 
> > There is a crossbar switch between the memory interfaces and the DAIs.
> > Not every connection is possible, so not every memory interface can be
> > used for every DAI. An algorithm choosing a suitable memory interface
> > must be quite clever, complicated and also SoC dependent (the same but
> > different hardware is used on MT8135 aswell), so I thought offering a
> > static configuration via device tree is a good start. Should there be
> > runtime configuration possible later the device tree settings could
> > provide a good default.
> 
> What exactly do the restrictions look like and how often do they vary in
> practice (can we get away with just doing a single static setup in the
> driver)? I'd have thought it should be fairly straightforward to have a
> table of valid mappings and just pick the first free memory interface?

I think this could be done. I checked the possible connections in the
crossbar switch and it seems all memory interfaces can be connected with
all relevant external interfaces. So indeed the memory interfaces could
be dynamically allocated instead of statically associated to an
external interface. There are two problems I see: Some memory interfaces
are limited in the rates they support, they can only do 8k/16k/32k (for
speech). How can we know such memory interface should be used? Also
there are two programmable hardware gain blocks which can be inserted to
the digital audio path using the crossbar switch. There must be some
mechanism to configure them into different places.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

[RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Sascha Hauer]

On Mon, Apr 20, 2015 at 09:48:49PM +0100, Mark Brown wrote:
> On Mon, Apr 20, 2015 at 06:37:47AM +0200, Sascha Hauer wrote:
> > On Sat, Apr 18, 2015 at 06:34:07PM +0100, Mark Brown wrote:
> > > On Fri, Apr 10, 2015 at 04:14:07PM +0800, Koro Chen wrote:
> 
> > > > +Each external interface (called "IO" in this driver) is presented as a
> > > > +DAI to ASoC. An IO must be connected via the interconnect to a memif.
> > > > +The connection paths are configured through the device tree.
> 
> > > Why are these connection paths configured via device tree?  I would
> > > expect that either there would be runtime configurability of these
> > > things (particularly if loopback configurations within the hardware are
> > > possible) or we'd just allocate memory interfaces to DAIs automatically
> > > as DAIs come into use.
> 
> > There is a crossbar switch between the memory interfaces and the DAIs.
> > Not every connection is possible, so not every memory interface can be
> > used for every DAI. An algorithm choosing a suitable memory interface
> > must be quite clever, complicated and also SoC dependent (the same but
> > different hardware is used on MT8135 aswell), so I thought offering a
> > static configuration via device tree is a good start. Should there be
> > runtime configuration possible later the device tree settings could
> > provide a good default.
> 
> What exactly do the restrictions look like and how often do they vary in
> practice (can we get away with just doing a single static setup in the
> driver)? I'd have thought it should be fairly straightforward to have a
> table of valid mappings and just pick the first free memory interface?

I think this could be done. I checked the possible connections in the
crossbar switch and it seems all memory interfaces can be connected with
all relevant external interfaces. So indeed the memory interfaces could
be dynamically allocated instead of statically associated to an
external interface. There are two problems I see: Some memory interfaces
are limited in the rates they support, they can only do 8k/16k/32k (for
speech). How can we know such memory interface should be used? Also
there are two programmable hardware gain blocks which can be inserted to
the digital audio path using the crossbar switch. There must be some
mechanism to configure them into different places.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

Re: [RESEND RFC PATCH 3/3] ASoC: mediatek: Add AFE platform driver

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 632 bytes --]

On Tue, Apr 21, 2015 at 10:27:13AM +0800, Koro Chen wrote:

> The SRAM size to be used is defined by params_buffer_bytes(params), not
> fixed (of course limited by the actual available SRAM size on HW), so
> the latency should be the same compared to a DRAM having the same size. 

Right, some systems have the SRAM as essentially a big FIFO but this
doesn't have that problem.

> The SRAM can be used by any memif, and that's why the plan was let DT
> make the decision.

OK, if it's for any interface rather than just DL1 like Sascha said then
it does need to be selectable, but DT doesn't seem the platce to do it.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

[RESEND RFC PATCH 3/3] ASoC: mediatek: Add AFE platform driver

[Mark Brown]

On Tue, Apr 21, 2015 at 10:27:13AM +0800, Koro Chen wrote:

> The SRAM size to be used is defined by params_buffer_bytes(params), not
> fixed (of course limited by the actual available SRAM size on HW), so
> the latency should be the same compared to a DRAM having the same size. 

Right, some systems have the SRAM as essentially a big FIFO but this
doesn't have that problem.

> The SRAM can be used by any memif, and that's why the plan was let DT
> make the decision.

OK, if it's for any interface rather than just DL1 like Sascha said then
it does need to be selectable, but DT doesn't seem the platce to do it.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20150421/9f438000/attachment.sig>

Re: [RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 939 bytes --]

On Tue, Apr 21, 2015 at 11:49:26AM +0200, Sascha Hauer wrote:

> I think this could be done. I checked the possible connections in the
> crossbar switch and it seems all memory interfaces can be connected with
> all relevant external interfaces. So indeed the memory interfaces could
> be dynamically allocated instead of statically associated to an
> external interface. There are two problems I see: Some memory interfaces
> are limited in the rates they support, they can only do 8k/16k/32k (for
> speech). How can we know such memory interface should be used? Also
> there are two programmable hardware gain blocks which can be inserted to
> the digital audio path using the crossbar switch. There must be some
> mechanism to configure them into different places.

This (particularly the gain controls) sounds like you want to expose the
routing to userspace and use DPCM, the code also seemed to look like it
was a good fit for DPCM.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

[RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Mark Brown]

On Tue, Apr 21, 2015 at 11:49:26AM +0200, Sascha Hauer wrote:

> I think this could be done. I checked the possible connections in the
> crossbar switch and it seems all memory interfaces can be connected with
> all relevant external interfaces. So indeed the memory interfaces could
> be dynamically allocated instead of statically associated to an
> external interface. There are two problems I see: Some memory interfaces
> are limited in the rates they support, they can only do 8k/16k/32k (for
> speech). How can we know such memory interface should be used? Also
> there are two programmable hardware gain blocks which can be inserted to
> the digital audio path using the crossbar switch. There must be some
> mechanism to configure them into different places.

This (particularly the gain controls) sounds like you want to expose the
routing to userspace and use DPCM, the code also seemed to look like it
was a good fit for DPCM.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20150421/e006ea93/attachment.sig>

Re: [RESEND RFC PATCH 2/3] ASoC: mediatek: Add AFE connection control

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 518 bytes --]

On Tue, Apr 21, 2015 at 07:50:41AM +0200, Sascha Hauer wrote:

> Oh, that's not magic at all, the crossbar switch has inputs and outputs
> numbered from 0 to MTK_AFE_INTERCONN_NUM_INPUT /
> MTK_AFE_INTERCONN_NUM_OUTPUT (they have the same numbers in the
> datasheet. To connect input x with output y look at index [x][y] in the
> table and write the register values found at that place. If .creg is 0x0
> then it's not possible to connect the given input with the given output.

A comment saying that would help then.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

Re: [RESEND RFC PATCH 2/3] ASoC: mediatek: Add AFE connection control

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 518 bytes --]

On Tue, Apr 21, 2015 at 07:50:41AM +0200, Sascha Hauer wrote:

> Oh, that's not magic at all, the crossbar switch has inputs and outputs
> numbered from 0 to MTK_AFE_INTERCONN_NUM_INPUT /
> MTK_AFE_INTERCONN_NUM_OUTPUT (they have the same numbers in the
> datasheet. To connect input x with output y look at index [x][y] in the
> table and write the register values found at that place. If .creg is 0x0
> then it's not possible to connect the given input with the given output.

A comment saying that would help then.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

[RESEND RFC PATCH 2/3] ASoC: mediatek: Add AFE connection control

[Mark Brown]

On Tue, Apr 21, 2015 at 07:50:41AM +0200, Sascha Hauer wrote:

> Oh, that's not magic at all, the crossbar switch has inputs and outputs
> numbered from 0 to MTK_AFE_INTERCONN_NUM_INPUT /
> MTK_AFE_INTERCONN_NUM_OUTPUT (they have the same numbers in the
> datasheet. To connect input x with output y look at index [x][y] in the
> table and write the register values found at that place. If .creg is 0x0
> then it's not possible to connect the given input with the given output.

A comment saying that would help then.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20150421/563f9fc2/attachment.sig>

Re: [RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Koro Chen]

On Tue, 2015-04-21 at 11:49 +0200, Sascha Hauer wrote:
> On Mon, Apr 20, 2015 at 09:48:49PM +0100, Mark Brown wrote:
> > On Mon, Apr 20, 2015 at 06:37:47AM +0200, Sascha Hauer wrote:
> > > On Sat, Apr 18, 2015 at 06:34:07PM +0100, Mark Brown wrote:
> > > > On Fri, Apr 10, 2015 at 04:14:07PM +0800, Koro Chen wrote:
> > 
> > > > > +Each external interface (called "IO" in this driver) is presented as a
> > > > > +DAI to ASoC. An IO must be connected via the interconnect to a memif.
> > > > > +The connection paths are configured through the device tree.
> > 
> > > > Why are these connection paths configured via device tree?  I would
> > > > expect that either there would be runtime configurability of these
> > > > things (particularly if loopback configurations within the hardware are
> > > > possible) or we'd just allocate memory interfaces to DAIs automatically
> > > > as DAIs come into use.
> > 
> > > There is a crossbar switch between the memory interfaces and the DAIs.
> > > Not every connection is possible, so not every memory interface can be
> > > used for every DAI. An algorithm choosing a suitable memory interface
> > > must be quite clever, complicated and also SoC dependent (the same but
> > > different hardware is used on MT8135 aswell), so I thought offering a
> > > static configuration via device tree is a good start. Should there be
> > > runtime configuration possible later the device tree settings could
> > > provide a good default.
> > 
> > What exactly do the restrictions look like and how often do they vary in
> > practice (can we get away with just doing a single static setup in the
> > driver)? I'd have thought it should be fairly straightforward to have a
> > table of valid mappings and just pick the first free memory interface?
> 
> I think this could be done. I checked the possible connections in the
> crossbar switch and it seems all memory interfaces can be connected with
> all relevant external interfaces. So indeed the memory interfaces could
> be dynamically allocated instead of statically associated to an
> external interface. There are two problems I see: Some memory interfaces
> are limited in the rates they support, they can only do 8k/16k/32k (for
> speech). How can we know such memory interface should be used? Also
The 2 memif are "DAI" and "MOD_DAI", designed for speech cases, and they
should be only connected to corresponding external interface "DAI/BT"
and "modem", respectively. We don't need to put them into dynamic
allocation. 
> there are two programmable hardware gain blocks which can be inserted to
> the digital audio path using the crossbar switch. There must be some
> mechanism to configure them into different places.
Maybe in DPCM, they can be "widgets", and we can define "routes" and
corresponding controls for them. 
> 
> Sascha
> 

Re: [RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Koro Chen]

On Tue, 2015-04-21 at 11:49 +0200, Sascha Hauer wrote:
> On Mon, Apr 20, 2015 at 09:48:49PM +0100, Mark Brown wrote:
> > On Mon, Apr 20, 2015 at 06:37:47AM +0200, Sascha Hauer wrote:
> > > On Sat, Apr 18, 2015 at 06:34:07PM +0100, Mark Brown wrote:
> > > > On Fri, Apr 10, 2015 at 04:14:07PM +0800, Koro Chen wrote:
> > 
> > > > > +Each external interface (called "IO" in this driver) is presented as a
> > > > > +DAI to ASoC. An IO must be connected via the interconnect to a memif.
> > > > > +The connection paths are configured through the device tree.
> > 
> > > > Why are these connection paths configured via device tree?  I would
> > > > expect that either there would be runtime configurability of these
> > > > things (particularly if loopback configurations within the hardware are
> > > > possible) or we'd just allocate memory interfaces to DAIs automatically
> > > > as DAIs come into use.
> > 
> > > There is a crossbar switch between the memory interfaces and the DAIs.
> > > Not every connection is possible, so not every memory interface can be
> > > used for every DAI. An algorithm choosing a suitable memory interface
> > > must be quite clever, complicated and also SoC dependent (the same but
> > > different hardware is used on MT8135 aswell), so I thought offering a
> > > static configuration via device tree is a good start. Should there be
> > > runtime configuration possible later the device tree settings could
> > > provide a good default.
> > 
> > What exactly do the restrictions look like and how often do they vary in
> > practice (can we get away with just doing a single static setup in the
> > driver)? I'd have thought it should be fairly straightforward to have a
> > table of valid mappings and just pick the first free memory interface?
> 
> I think this could be done. I checked the possible connections in the
> crossbar switch and it seems all memory interfaces can be connected with
> all relevant external interfaces. So indeed the memory interfaces could
> be dynamically allocated instead of statically associated to an
> external interface. There are two problems I see: Some memory interfaces
> are limited in the rates they support, they can only do 8k/16k/32k (for
> speech). How can we know such memory interface should be used? Also
The 2 memif are "DAI" and "MOD_DAI", designed for speech cases, and they
should be only connected to corresponding external interface "DAI/BT"
and "modem", respectively. We don't need to put them into dynamic
allocation. 
> there are two programmable hardware gain blocks which can be inserted to
> the digital audio path using the crossbar switch. There must be some
> mechanism to configure them into different places.
Maybe in DPCM, they can be "widgets", and we can define "routes" and
corresponding controls for them. 
> 
> Sascha
> 

[RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Koro Chen]

On Tue, 2015-04-21 at 11:49 +0200, Sascha Hauer wrote:
> On Mon, Apr 20, 2015 at 09:48:49PM +0100, Mark Brown wrote:
> > On Mon, Apr 20, 2015 at 06:37:47AM +0200, Sascha Hauer wrote:
> > > On Sat, Apr 18, 2015 at 06:34:07PM +0100, Mark Brown wrote:
> > > > On Fri, Apr 10, 2015 at 04:14:07PM +0800, Koro Chen wrote:
> > 
> > > > > +Each external interface (called "IO" in this driver) is presented as a
> > > > > +DAI to ASoC. An IO must be connected via the interconnect to a memif.
> > > > > +The connection paths are configured through the device tree.
> > 
> > > > Why are these connection paths configured via device tree?  I would
> > > > expect that either there would be runtime configurability of these
> > > > things (particularly if loopback configurations within the hardware are
> > > > possible) or we'd just allocate memory interfaces to DAIs automatically
> > > > as DAIs come into use.
> > 
> > > There is a crossbar switch between the memory interfaces and the DAIs.
> > > Not every connection is possible, so not every memory interface can be
> > > used for every DAI. An algorithm choosing a suitable memory interface
> > > must be quite clever, complicated and also SoC dependent (the same but
> > > different hardware is used on MT8135 aswell), so I thought offering a
> > > static configuration via device tree is a good start. Should there be
> > > runtime configuration possible later the device tree settings could
> > > provide a good default.
> > 
> > What exactly do the restrictions look like and how often do they vary in
> > practice (can we get away with just doing a single static setup in the
> > driver)? I'd have thought it should be fairly straightforward to have a
> > table of valid mappings and just pick the first free memory interface?
> 
> I think this could be done. I checked the possible connections in the
> crossbar switch and it seems all memory interfaces can be connected with
> all relevant external interfaces. So indeed the memory interfaces could
> be dynamically allocated instead of statically associated to an
> external interface. There are two problems I see: Some memory interfaces
> are limited in the rates they support, they can only do 8k/16k/32k (for
> speech). How can we know such memory interface should be used? Also
The 2 memif are "DAI" and "MOD_DAI", designed for speech cases, and they
should be only connected to corresponding external interface "DAI/BT"
and "modem", respectively. We don't need to put them into dynamic
allocation. 
> there are two programmable hardware gain blocks which can be inserted to
> the digital audio path using the crossbar switch. There must be some
> mechanism to configure them into different places.
Maybe in DPCM, they can be "widgets", and we can define "routes" and
corresponding controls for them. 
> 
> Sascha
> 

Re: [RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 491 bytes --]

On Tue, Apr 21, 2015 at 06:15:06PM +0800, Koro Chen wrote:
> On Tue, 2015-04-21 at 11:49 +0200, Sascha Hauer wrote:

> > there are two programmable hardware gain blocks which can be inserted to
> > the digital audio path using the crossbar switch. There must be some
> > mechanism to configure them into different places.

> Maybe in DPCM, they can be "widgets", and we can define "routes" and
> corresponding controls for them. 

Yes, exactly - it's basically just like a CODEC.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

[RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Mark Brown]

On Tue, Apr 21, 2015 at 06:15:06PM +0800, Koro Chen wrote:
> On Tue, 2015-04-21 at 11:49 +0200, Sascha Hauer wrote:

> > there are two programmable hardware gain blocks which can be inserted to
> > the digital audio path using the crossbar switch. There must be some
> > mechanism to configure them into different places.

> Maybe in DPCM, they can be "widgets", and we can define "routes" and
> corresponding controls for them. 

Yes, exactly - it's basically just like a CODEC.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20150421/8bf8f506/attachment.sig>

Re: [RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Koro Chen]

On Mon, 2015-04-20 at 21:48 +0100, Mark Brown wrote:
> On Mon, Apr 20, 2015 at 06:37:47AM +0200, Sascha Hauer wrote:
> > On Sat, Apr 18, 2015 at 06:34:07PM +0100, Mark Brown wrote:
> > > On Fri, Apr 10, 2015 at 04:14:07PM +0800, Koro Chen wrote:
> 
> > > > +Each external interface (called "IO" in this driver) is presented as a
> > > > +DAI to ASoC. An IO must be connected via the interconnect to a memif.
> > > > +The connection paths are configured through the device tree.
> 
> > > Why are these connection paths configured via device tree?  I would
> > > expect that either there would be runtime configurability of these
> > > things (particularly if loopback configurations within the hardware are
> > > possible) or we'd just allocate memory interfaces to DAIs automatically
> > > as DAIs come into use.
> 
> > There is a crossbar switch between the memory interfaces and the DAIs.
> > Not every connection is possible, so not every memory interface can be
> > used for every DAI. An algorithm choosing a suitable memory interface
> > must be quite clever, complicated and also SoC dependent (the same but
> > different hardware is used on MT8135 aswell), so I thought offering a
> > static configuration via device tree is a good start. Should there be
> > runtime configuration possible later the device tree settings could
> > provide a good default.
> 
> What exactly do the restrictions look like and how often do they vary in
> practice (can we get away with just doing a single static setup in the
> driver)?  I'd have thought it should be fairly straightforward to have a
> table of valid mappings and just pick the first free memory interface?
> I'd rather not get stuck with the tables in the DT.  It's partly to
> avoid setting bad precendents, we really don't want everyone coming
> along hard coding this stuff, and partly because the hardware you
> described didn't seem that hard to handle.
> 
If using DPCM, it seems the most suitable FE DAIs will be memif, and
external interface like I2S should be BE DAIs. Do you think it is
suitable for our memif to be FE DAIs? Then which memif to used can be
described in a machine driver's FE DAIs. But I think this has a problem
that our memif is one-direction, playback or capture. So the binded pcm
device cannot have playback and capture capability together. May it
cause trouble for user space apps that assumes there will be pcmC0D0p,
pcmC0D0c?

> > > > +- mem-interface-playback:
> > > > +  mem-interface-capture: property of memif, format is: <memif irq use_sram>;
> > > > +	                 memif: which memif to be used
> > > > +			        (defined in include/dt-bindings/sound/mtk-afe.h)
> > > > +		         irq: which irq to be used
> > > > +			      (defined in include/dt-bindings/sound/mtk-afe.h)
> > > > +		         use_sram: 1 is yes, 0 is no
> 
> > > Again, this looks like stuff we should be able to figure out at runtime
> > > - the use of SRAM in particular looks like something we might want to
> > > change depending on use case.  Assuming it adds buffering then for a
> > > VoIP application we might not want to use SRAM to minimize latency but
> > > during music playback we might want to enable SRAM to minimize power
> > > consumption.
> 
> > That's exactly the usecase. How could such a runtime configurability
> > look like? sysfs? Or something based on the buffer sizes?
> 
> Yeah, one of those :) but probably an ALSA control is going to be the
> easiest for applications.

Re: [RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Koro Chen]

On Mon, 2015-04-20 at 21:48 +0100, Mark Brown wrote:
> On Mon, Apr 20, 2015 at 06:37:47AM +0200, Sascha Hauer wrote:
> > On Sat, Apr 18, 2015 at 06:34:07PM +0100, Mark Brown wrote:
> > > On Fri, Apr 10, 2015 at 04:14:07PM +0800, Koro Chen wrote:
> 
> > > > +Each external interface (called "IO" in this driver) is presented as a
> > > > +DAI to ASoC. An IO must be connected via the interconnect to a memif.
> > > > +The connection paths are configured through the device tree.
> 
> > > Why are these connection paths configured via device tree?  I would
> > > expect that either there would be runtime configurability of these
> > > things (particularly if loopback configurations within the hardware are
> > > possible) or we'd just allocate memory interfaces to DAIs automatically
> > > as DAIs come into use.
> 
> > There is a crossbar switch between the memory interfaces and the DAIs.
> > Not every connection is possible, so not every memory interface can be
> > used for every DAI. An algorithm choosing a suitable memory interface
> > must be quite clever, complicated and also SoC dependent (the same but
> > different hardware is used on MT8135 aswell), so I thought offering a
> > static configuration via device tree is a good start. Should there be
> > runtime configuration possible later the device tree settings could
> > provide a good default.
> 
> What exactly do the restrictions look like and how often do they vary in
> practice (can we get away with just doing a single static setup in the
> driver)?  I'd have thought it should be fairly straightforward to have a
> table of valid mappings and just pick the first free memory interface?
> I'd rather not get stuck with the tables in the DT.  It's partly to
> avoid setting bad precendents, we really don't want everyone coming
> along hard coding this stuff, and partly because the hardware you
> described didn't seem that hard to handle.
> 
If using DPCM, it seems the most suitable FE DAIs will be memif, and
external interface like I2S should be BE DAIs. Do you think it is
suitable for our memif to be FE DAIs? Then which memif to used can be
described in a machine driver's FE DAIs. But I think this has a problem
that our memif is one-direction, playback or capture. So the binded pcm
device cannot have playback and capture capability together. May it
cause trouble for user space apps that assumes there will be pcmC0D0p,
pcmC0D0c?

> > > > +- mem-interface-playback:
> > > > +  mem-interface-capture: property of memif, format is: <memif irq use_sram>;
> > > > +	                 memif: which memif to be used
> > > > +			        (defined in include/dt-bindings/sound/mtk-afe.h)
> > > > +		         irq: which irq to be used
> > > > +			      (defined in include/dt-bindings/sound/mtk-afe.h)
> > > > +		         use_sram: 1 is yes, 0 is no
> 
> > > Again, this looks like stuff we should be able to figure out at runtime
> > > - the use of SRAM in particular looks like something we might want to
> > > change depending on use case.  Assuming it adds buffering then for a
> > > VoIP application we might not want to use SRAM to minimize latency but
> > > during music playback we might want to enable SRAM to minimize power
> > > consumption.
> 
> > That's exactly the usecase. How could such a runtime configurability
> > look like? sysfs? Or something based on the buffer sizes?
> 
> Yeah, one of those :) but probably an ALSA control is going to be the
> easiest for applications.


--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Koro Chen]

On Mon, 2015-04-20 at 21:48 +0100, Mark Brown wrote:
> On Mon, Apr 20, 2015 at 06:37:47AM +0200, Sascha Hauer wrote:
> > On Sat, Apr 18, 2015 at 06:34:07PM +0100, Mark Brown wrote:
> > > On Fri, Apr 10, 2015 at 04:14:07PM +0800, Koro Chen wrote:
> 
> > > > +Each external interface (called "IO" in this driver) is presented as a
> > > > +DAI to ASoC. An IO must be connected via the interconnect to a memif.
> > > > +The connection paths are configured through the device tree.
> 
> > > Why are these connection paths configured via device tree?  I would
> > > expect that either there would be runtime configurability of these
> > > things (particularly if loopback configurations within the hardware are
> > > possible) or we'd just allocate memory interfaces to DAIs automatically
> > > as DAIs come into use.
> 
> > There is a crossbar switch between the memory interfaces and the DAIs.
> > Not every connection is possible, so not every memory interface can be
> > used for every DAI. An algorithm choosing a suitable memory interface
> > must be quite clever, complicated and also SoC dependent (the same but
> > different hardware is used on MT8135 aswell), so I thought offering a
> > static configuration via device tree is a good start. Should there be
> > runtime configuration possible later the device tree settings could
> > provide a good default.
> 
> What exactly do the restrictions look like and how often do they vary in
> practice (can we get away with just doing a single static setup in the
> driver)?  I'd have thought it should be fairly straightforward to have a
> table of valid mappings and just pick the first free memory interface?
> I'd rather not get stuck with the tables in the DT.  It's partly to
> avoid setting bad precendents, we really don't want everyone coming
> along hard coding this stuff, and partly because the hardware you
> described didn't seem that hard to handle.
> 
If using DPCM, it seems the most suitable FE DAIs will be memif, and
external interface like I2S should be BE DAIs. Do you think it is
suitable for our memif to be FE DAIs? Then which memif to used can be
described in a machine driver's FE DAIs. But I think this has a problem
that our memif is one-direction, playback or capture. So the binded pcm
device cannot have playback and capture capability together. May it
cause trouble for user space apps that assumes there will be pcmC0D0p,
pcmC0D0c?

> > > > +- mem-interface-playback:
> > > > +  mem-interface-capture: property of memif, format is: <memif irq use_sram>;
> > > > +	                 memif: which memif to be used
> > > > +			        (defined in include/dt-bindings/sound/mtk-afe.h)
> > > > +		         irq: which irq to be used
> > > > +			      (defined in include/dt-bindings/sound/mtk-afe.h)
> > > > +		         use_sram: 1 is yes, 0 is no
> 
> > > Again, this looks like stuff we should be able to figure out at runtime
> > > - the use of SRAM in particular looks like something we might want to
> > > change depending on use case.  Assuming it adds buffering then for a
> > > VoIP application we might not want to use SRAM to minimize latency but
> > > during music playback we might want to enable SRAM to minimize power
> > > consumption.
> 
> > That's exactly the usecase. How could such a runtime configurability
> > look like? sysfs? Or something based on the buffer sizes?
> 
> Yeah, one of those :) but probably an ALSA control is going to be the
> easiest for applications.

Re: [RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 902 bytes --]

On Wed, Apr 22, 2015 at 11:17:20AM +0800, Koro Chen wrote:

> If using DPCM, it seems the most suitable FE DAIs will be memif, and
> external interface like I2S should be BE DAIs. Do you think it is
> suitable for our memif to be FE DAIs? Then which memif to used can be
> described in a machine driver's FE DAIs. But I think this has a problem
> that our memif is one-direction, playback or capture. So the binded pcm
> device cannot have playback and capture capability together. May it
> cause trouble for user space apps that assumes there will be pcmC0D0p,
> pcmC0D0c?

I think the only userspace application you really have to worry about
here is PulseAudio which (at least with UCM) should be able to cope
since it needs to handle things like USB microphones.  I think
everything else I can think of can either cope or has to handle the same
use cases as PulseAudio does so should have support.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

Re: [RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 902 bytes --]

On Wed, Apr 22, 2015 at 11:17:20AM +0800, Koro Chen wrote:

> If using DPCM, it seems the most suitable FE DAIs will be memif, and
> external interface like I2S should be BE DAIs. Do you think it is
> suitable for our memif to be FE DAIs? Then which memif to used can be
> described in a machine driver's FE DAIs. But I think this has a problem
> that our memif is one-direction, playback or capture. So the binded pcm
> device cannot have playback and capture capability together. May it
> cause trouble for user space apps that assumes there will be pcmC0D0p,
> pcmC0D0c?

I think the only userspace application you really have to worry about
here is PulseAudio which (at least with UCM) should be able to cope
since it needs to handle things like USB microphones.  I think
everything else I can think of can either cope or has to handle the same
use cases as PulseAudio does so should have support.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

[RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Mark Brown]

On Wed, Apr 22, 2015 at 11:17:20AM +0800, Koro Chen wrote:

> If using DPCM, it seems the most suitable FE DAIs will be memif, and
> external interface like I2S should be BE DAIs. Do you think it is
> suitable for our memif to be FE DAIs? Then which memif to used can be
> described in a machine driver's FE DAIs. But I think this has a problem
> that our memif is one-direction, playback or capture. So the binded pcm
> device cannot have playback and capture capability together. May it
> cause trouble for user space apps that assumes there will be pcmC0D0p,
> pcmC0D0c?

I think the only userspace application you really have to worry about
here is PulseAudio which (at least with UCM) should be able to cope
since it needs to handle things like USB microphones.  I think
everything else I can think of can either cope or has to handle the same
use cases as PulseAudio does so should have support.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20150430/243c1ad7/attachment.sig>

Re: [RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Koro Chen]

On Thu, 2015-04-30 at 21:12 +0100, Mark Brown wrote:
> On Wed, Apr 22, 2015 at 11:17:20AM +0800, Koro Chen wrote:
> 
> > If using DPCM, it seems the most suitable FE DAIs will be memif, and
> > external interface like I2S should be BE DAIs. Do you think it is
> > suitable for our memif to be FE DAIs? Then which memif to used can be
> > described in a machine driver's FE DAIs. But I think this has a problem
> > that our memif is one-direction, playback or capture. So the binded pcm
> > device cannot have playback and capture capability together. May it
> > cause trouble for user space apps that assumes there will be pcmC0D0p,
> > pcmC0D0c?
> 
> I think the only userspace application you really have to worry about
> here is PulseAudio which (at least with UCM) should be able to cope
> since it needs to handle things like USB microphones.  I think
> everything else I can think of can either cope or has to handle the same
> use cases as PulseAudio does so should have support.
Thank you very much for the feedback, I will try DPCM.
  

Re: [RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Koro Chen]

On Thu, 2015-04-30 at 21:12 +0100, Mark Brown wrote:
> On Wed, Apr 22, 2015 at 11:17:20AM +0800, Koro Chen wrote:
> 
> > If using DPCM, it seems the most suitable FE DAIs will be memif, and
> > external interface like I2S should be BE DAIs. Do you think it is
> > suitable for our memif to be FE DAIs? Then which memif to used can be
> > described in a machine driver's FE DAIs. But I think this has a problem
> > that our memif is one-direction, playback or capture. So the binded pcm
> > device cannot have playback and capture capability together. May it
> > cause trouble for user space apps that assumes there will be pcmC0D0p,
> > pcmC0D0c?
> 
> I think the only userspace application you really have to worry about
> here is PulseAudio which (at least with UCM) should be able to cope
> since it needs to handle things like USB microphones.  I think
> everything else I can think of can either cope or has to handle the same
> use cases as PulseAudio does so should have support.
Thank you very much for the feedback, I will try DPCM.
  


--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[RESEND RFC PATCH 1/3] ASoC: mediatek: Add binding support for AFE driver

[Koro Chen]

On Thu, 2015-04-30 at 21:12 +0100, Mark Brown wrote:
> On Wed, Apr 22, 2015 at 11:17:20AM +0800, Koro Chen wrote:
> 
> > If using DPCM, it seems the most suitable FE DAIs will be memif, and
> > external interface like I2S should be BE DAIs. Do you think it is
> > suitable for our memif to be FE DAIs? Then which memif to used can be
> > described in a machine driver's FE DAIs. But I think this has a problem
> > that our memif is one-direction, playback or capture. So the binded pcm
> > device cannot have playback and capture capability together. May it
> > cause trouble for user space apps that assumes there will be pcmC0D0p,
> > pcmC0D0c?
> 
> I think the only userspace application you really have to worry about
> here is PulseAudio which (at least with UCM) should be able to cope
> since it needs to handle things like USB microphones.  I think
> everything else I can think of can either cope or has to handle the same
> use cases as PulseAudio does so should have support.
Thank you very much for the feedback, I will try DPCM.