Skip to contentSkip to conversation

arm: sun9i: Support AXP808 PMIC and Sunchip CX-A99 board

PATCH New Rask Ingemann Lambertsen 3 participants 19 messages Jan 22, 2017 – Feb 21, 2017

Subsystems (6)
lore
Share & tools
lore
#4 lore raw

PATCH v2 4/5 regulator: axp20x: Add support for the AXP808 PMIC

In reply to Rask Ingemann Lambertsen

The regulators are the same as on the AXP806.

---

No changes in v2.

 drivers/regulator/axp20x-regulator.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
drivers/regulator/axp20x-regulator.c +5 -2
index a3ade9e..034dce8 100644
@@ -370,6 +370,7 @@ static int axp20x_set_dcdc_freq(struct platform_device *pdev, u32 dcdcfreq)
 		step = 75;
 		break;
 	case AXP806_ID:
+	case AXP808_ID:
 		/*
 		 * AXP806 DCDC work frequency setting has the same range and
 		 * step as AXP22X, but at a different register.
@@ -459,6 +460,7 @@ static int axp20x_set_dcdc_workmode(struct regulator_dev *rdev, int id, u32 work
 		break;
 
 	case AXP806_ID:
+	case AXP808_ID:
 		reg = AXP806_DCDC_MODE_CTRL2;
 		/*
 		 * AXP806 DCDC regulator IDs have the same range as AXP22X.
@@ -492,8 +494,8 @@ static bool axp20x_is_polyphase_slave(struct axp20x_dev *axp20x, int id)
 {
 	u32 reg = 0;
 
-	/* Only AXP806 has poly-phase outputs */
-	if (axp20x->variant != AXP806_ID)
+	/* Only AXP806 and AXP808 have poly-phase outputs */
+	if (axp20x->variant != AXP806_ID && axp20x->variant != AXP808_ID)
 		return false;
 
 	regmap_read(axp20x->regmap, AXP806_DCDC_MODE_CTRL2, &reg);
@@ -541,6 +543,7 @@ static int axp20x_regulator_probe(struct platform_device *pdev)
 						  "x-powers,drive-vbus-en");
 		break;
 	case AXP806_ID:
+	case AXP808_ID:
 		regulators = axp806_regulators;
 		nregulators = AXP806_REG_ID_MAX;
 		break;
Signature · 1 line
2.10.2
#0 lore raw

v2 cover arm: sun9i: Support AXP808 PMIC and Sunchip CX-A99 board

This patch series add support for the Sunchip CX-A99 board, which is used in a few media players, making it the third device with an Allwinner A80 SoC to be supported. The board uses the X-Powers AXP808 PMIC which this patch series also adds support for.

Patch 1 adds a documentation line which was forgotten when device tree bindings for the axp20x family were updated for the AXP806 PMIC. A line immediately below the missing line will be inserted by patch 2, so I'm including patch 1 in this series to avoid conflicts between patches.

Patch 2 adds the AXP808 to the axp20x device tree bindings.

Patch 3 adds support for the AXP808 to the axp20x MFD driver.

Patch 4 adds support for the AXP808 to the axp20x regulator driver.

Patch 5 add the device tree for the Sunchip CX-A99 board. This patch is v6 of the only remaining patch from v4 of the patch series posted back in August. Support for the AXP808 has made major improvements possible, notably Wifi.

Before trying out this patch series, make sure your tree contains patch "regulator: axp20x: AXP806: Fix dcdcb being set instead of dcdce" [1] to avoid damage to the DRAM chips through over-voltage.

[1] https://patchwork.kernel.org/patch/9530319/

Rask Ingemann Lambertsen (5):
dts: mfd: axp20x: Add AXP806 to list of current AXP20x family members
dts: mfd: axp20x: Add binding for the AXP808
mfd: axp20x: Add support for the AXP808 PMIC
regulator: axp20x: Add support for the AXP808 PMIC
ARM: dts: sun9i: Initial support for the Sunchip CX-A99 board

 Documentation/devicetree/bindings/mfd/axp20x.txt |   8 +-
 arch/arm/boot/dts/Makefile                       |   3 +-
 arch/arm/boot/dts/sun9i-a80-cx-a99.dts           | 409 +++++++++++++++++++++++
 drivers/mfd/axp20x-rsb.c                         |   1 +
 drivers/mfd/axp20x.c                             |  26 ++
 drivers/regulator/axp20x-regulator.c             |   7 +-
 include/linux/mfd/axp20x.h                       |   1 +
 7 files changed, 449 insertions(+), 6 deletions(-)
 create mode 100644 arch/arm/boot/dts/sun9i-a80-cx-a99.dts
Signature · 1 line
2.10.2
#1 lore raw

PATCH v2 1/5 dts: mfd: axp20x: Add AXP806 to list of current AXP20x family members

An entry for the AXP806 was forgotten, so add one.

Fixes 204ae2963e10 ("mfd: axp20x: Add bindings for AXP806 PMIC")

---

No changes in v2.

 Documentation/devicetree/bindings/mfd/axp20x.txt | 1 +
 1 file changed, 1 insertion(+)
Documentation/devicetree/bindings/mfd/axp20x.txt +1 -0
index 8f3ad9a..86ae540 100644
@@ -6,6 +6,7 @@ axp202 (X-Powers)
 axp209 (X-Powers)
 axp221 (X-Powers)
 axp223 (X-Powers)
+axp806 (X-Powers)
 axp809 (X-Powers)
 
 Required properties:
Signature · 1 line
2.10.2
#2 lore raw

PATCH v2 2/5 dts: mfd: axp20x: Add binding for the AXP808

In reply to Rask Ingemann Lambertsen

The AXP808 does not support address space extension, but is otherwise identical to the AXP806, including the chip ID, so add a compatible string for it to the binding.

---

No changes in v2.

 Documentation/devicetree/bindings/mfd/axp20x.txt | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
Documentation/devicetree/bindings/mfd/axp20x.txt +4 -3
index 86ae540..8195b06 100644
@@ -7,12 +7,13 @@ axp209 (X-Powers)
 axp221 (X-Powers)
 axp223 (X-Powers)
 axp806 (X-Powers)
+axp808 (X-Powers)
 axp809 (X-Powers)
 
 Required properties:
 - compatible: "x-powers,axp152", "x-powers,axp202", "x-powers,axp209",
 	      "x-powers,axp221", "x-powers,axp223", "x-powers,axp806",
-	      "x-powers,axp809"
+	      "x-powers,axp808", "x-powers,axp809"
 - reg: The I2C slave address or RSB hardware address for the AXP chip
 - interrupt-parent: The parent interrupt controller
 - interrupts: SoC NMI / GPIO interrupt connected to the PMIC's IRQ pin
@@ -87,7 +88,7 @@ LDO_IO1		: LDO		: ips-supply		: GPIO 1
 RTC_LDO		: LDO		: ips-supply		: always on
 DRIVEVBUS	: Enable output	: drivevbus-supply	: external regulator
 
-AXP806 regulators, type, and corresponding input supply names:
+AXP806/AXP808 regulators, type, and corresponding input supply names:
 
 Regulator	  Type		  Supply Name		  Notes
 ---------	  ----		  -----------		  -----
@@ -108,7 +109,7 @@ CLDO2		: LDO		: cldoin-supply		: shared supply
 CLDO3		: LDO		: cldoin-supply		: shared supply
 SW		: On/Off Switch : swin-supply
 
-Additionally, the AXP806 DC-DC regulators support poly-phase arrangements
+Additionally, the AXP806/AXP808 DC-DC regulators support poly-phase arrangements
 for higher output current. The possible groupings are: A+B, A+B+C, D+E.
 
 AXP809 regulators, type, and corresponding input supply names:
Signature · 1 line
2.10.2
#3 lore raw

PATCH v2 3/5 mfd: axp20x: Add support for the AXP808 PMIC

In reply to Rask Ingemann Lambertsen

The X-Powers AXP808 is a PMIC which, like the very similar AXP806, is used on boards featuring Allwinner's A80 SoC. Unlike the AXP806, it doesn't support address space extension and its associated registers, but the two are otherwise identical (including the chip ID). After commit b101829a029a ("mfd: axp20x: Fix AXP806 access errors on cold boot") those registers are now used, causing the driver to no longer work with an AXP808.

This patch adds support for the interrupts on the AXP808 and enables the regulator sub-device. The next patch in the series adds the actual regulator support.

---

Changes in v2:
- Noted in the commit message that another commit for better AXP806 support means the driver no longer works with an AXP808.
- axp808_writeable_table was changed to use the common part (between AXP806 and AXP808) of axp806_writeable_ranges instead of duplicating the entries. Suggested by Chen-Yu Tsai.

 drivers/mfd/axp20x-rsb.c   |  1 +
 drivers/mfd/axp20x.c       | 26 ++++++++++++++++++++++++++
 include/linux/mfd/axp20x.h |  1 +
 3 files changed, 28 insertions(+)
drivers/mfd/axp20x-rsb.c +1 -0
index a732cb5..96ea2e9 100644
@@ -62,6 +62,7 @@ static int axp20x_rsb_remove(struct sunxi_rsb_device *rdev)
 static const struct of_device_id axp20x_rsb_of_match[] = {
 	{ .compatible = "x-powers,axp223", .data = (void *)AXP223_ID },
 	{ .compatible = "x-powers,axp806", .data = (void *)AXP806_ID },
+	{ .compatible = "x-powers,axp808", .data = (void *)AXP808_ID },
 	{ .compatible = "x-powers,axp809", .data = (void *)AXP809_ID },
 	{ },
 };
drivers/mfd/axp20x.c +26 -0
index 25115fe..125b470 100644
@@ -41,6 +41,7 @@ static const char * const axp20x_model_names[] = {
 	"AXP223",
 	"AXP288",
 	"AXP806",
+	"AXP808",
 	"AXP809",
 };
 
@@ -140,6 +141,11 @@ static const struct regmap_access_table axp288_volatile_table = {
 	.n_yes_ranges	= ARRAY_SIZE(axp288_volatile_ranges),
 };
 
+/*
+ * These ranges are used by both axp806_writeable_table and
+ * axp808_writeable_table. This is possible as long as those not supported
+ * by the AXP808 (currently only the last one) are kept at an end.
+ */
 static const struct regmap_range axp806_writeable_ranges[] = {
 	regmap_reg_range(AXP20X_DATACACHE(0), AXP20X_DATACACHE(3)),
 	regmap_reg_range(AXP806_PWR_OUT_CTRL1, AXP806_CLDO3_V_CTRL),
@@ -162,6 +168,11 @@ static const struct regmap_access_table axp806_volatile_table = {
 	.n_yes_ranges	= ARRAY_SIZE(axp806_volatile_ranges),
 };
 
+static const struct regmap_access_table axp808_writeable_table = {
+	.yes_ranges	= axp806_writeable_ranges,
+	.n_yes_ranges	= ARRAY_SIZE(axp806_writeable_ranges) - 1,
+};
+
 static struct resource axp152_pek_resources[] = {
 	DEFINE_RES_IRQ_NAMED(AXP152_IRQ_PEK_RIS_EDGE, "PEK_DBR"),
 	DEFINE_RES_IRQ_NAMED(AXP152_IRQ_PEK_FAL_EDGE, "PEK_DBF"),
@@ -320,6 +331,15 @@ static const struct regmap_config axp806_regmap_config = {
 	.cache_type	= REGCACHE_RBTREE,
 };
 
+static const struct regmap_config axp808_regmap_config = {
+	.reg_bits	= 8,
+	.val_bits	= 8,
+	.wr_table	= &axp808_writeable_table,
+	.volatile_table	= &axp806_volatile_table,
+	.max_register	= AXP806_VREF_TEMP_WARN_L,
+	.cache_type	= REGCACHE_RBTREE,
+};
+
 #define INIT_REGMAP_IRQ(_variant, _irq, _off, _mask)			\
 	[_variant##_IRQ_##_irq] = { .reg_offset = (_off), .mask = BIT(_mask) }
 
@@ -842,6 +862,12 @@ int axp20x_match_device(struct axp20x_dev *axp20x)
 		axp20x->regmap_cfg = &axp806_regmap_config;
 		axp20x->regmap_irq_chip = &axp806_regmap_irq_chip;
 		break;
+	case AXP808_ID:
+		axp20x->nr_cells = ARRAY_SIZE(axp806_cells);
+		axp20x->cells = axp806_cells;
+		axp20x->regmap_cfg = &axp808_regmap_config;
+		axp20x->regmap_irq_chip = &axp806_regmap_irq_chip;
+		break;
 	case AXP809_ID:
 		axp20x->nr_cells = ARRAY_SIZE(axp809_cells);
 		axp20x->cells = axp809_cells;
include/linux/mfd/axp20x.h +1 -0
index 0d9a1ff..48c1788 100644
@@ -21,6 +21,7 @@ enum axp20x_variants {
 	AXP223_ID,
 	AXP288_ID,
 	AXP806_ID,
+	AXP808_ID,
 	AXP809_ID,
 	NR_AXP20X_VARIANTS,
 };
Signature · 1 line
2.10.2
#5 lore raw

PATCH v6 5/5 ARM: dts: sun9i: Initial support for the Sunchip CX-A99 board

In reply to Rask Ingemann Lambertsen

The Suncip CX-A99 board is found in at least four brands of media players. It features an Allwinner A80 ARM SoC and is found in two models:

1) 2 GiB DDR3 DRAM and 16 GB eMMC
2) 4 GiB DDR3 DRAM and 32 GB eMMC

For details of the board, see the linux-sunxi page <URL:https://linux-sunxi.org/Sunchip_CX-A99>.

Supported features (+ means tested):
+ One Cortex-A7 CPU core (or four with experimental U-Boot PSCI patches)
+ AXP808 power management chip
+ OZ80120 voltage regulator
+ Serial console port (internal)
+ eMMC and SD card slot
+ USB 2.0 host ports on on-board USB hub
+ SATA port on on-board SATA-to-USB bridge *
+ IEEE 802.11 a/b/g/n/ac SDIO Wifi
+ Real-time clock
+ LEDs
- IR receiver for remote control

* Only shows up when a SATA device is connected. Also, if a power source is connected to the USB 3.0 connector across power cycles (e.g. FEL boot), the bridge may not properly reset and not show up on the USB bus. The vendor U-Boot performs some unknown magic which resets the bridge.

So far unsupported features:
- Using any of the Cortex-A15 CPU cores
- USB 3.0 port (except for supplying 5 V power)
- IEEE 802.3 10/100/1000 megabit Ethernet
- HDMI connector
- S/PDIF audio output
- Jack socket with composite video and analog stereo audio
- Bluetooth
- FM radio receiver (assuming it is even wired on the board)

---

Changes in v6:
- Updated commit message description of SATA-to-USB bridge quirk and added note about experimental U-Boot PSCI support for up to four CPU cores.
- The blue LED is no longer on by default as its meaning is not documented.
- Removed "regulator-boot-on" from regulators having "regulator-always-on".
- Removed misleading mention of "OTG connector" which the device doesn't have.
- More detailed explanation for the need for "broken-cd" on mmc0.
- Several regulators have had their voltage range relaxed a little to match the permissible range according to the data sheets of the consumers. This is similar to what is used for the Cubieboard4 and Merrii A80 Optimus.
- Shortened regulator dcdce name as per v5 comments. A comment now lists the pin groups supplied by dcdce.

Changes in v5:
- Switched pinmux modes to generic properties and dropped #include <dt-bindings/pinctrl/sun4i-a10.h> as a consequence.
- Dropped pinctrl properties from GPIO nodes and dropped the pinmux nodes for them.
- AXP808 regulators added.
- Dropped the now unused #include <sunxi-common-regulators.dtsi>.
- Ampak AP6335 SDIO-Wifi added.
- USB Vbus changes as per v4 comments.
- Added "broken-cd" to mmc0 because GPIO interrupts don't work.

Changes in v4:
- Node names had underscores changed to hyphens.
- Changed formatting of the ac100/rtc node's clock output name list to match that of the same node in the cubieboard4 and a80-optimus device trees.

Changes in v3:
None.

Changes in v2:
- Fixed formatting and style issues found by scripts/checkpatch.pl.

 arch/arm/boot/dts/Makefile             |   3 +-
 arch/arm/boot/dts/sun9i-a80-cx-a99.dts | 409 +++++++++++++++++++++++++++++++++
 2 files changed, 411 insertions(+), 1 deletion(-)
 create mode 100644 arch/arm/boot/dts/sun9i-a80-cx-a99.dts
arch/arm/boot/dts/Makefile +2 -1
index 8553bd7..40546fa 100644
@@ -862,7 +862,8 @@ dtb-$(CONFIG_MACH_SUN8I) += \
 	sun8i-v3s-licheepi-zero.dtb
 dtb-$(CONFIG_MACH_SUN9I) += \
 	sun9i-a80-optimus.dtb \
-	sun9i-a80-cubieboard4.dtb
+	sun9i-a80-cubieboard4.dtb \
+	sun9i-a80-cx-a99.dtb
 dtb-$(CONFIG_ARCH_TANGO) += \
 	tango4-vantage-1172.dtb
 dtb-$(CONFIG_ARCH_TEGRA_2x_SOC) += \
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +409 -0

409 lines not on this page · raw message

Signature · 1 line
2.10.2
#6 lore raw

Re: [PATCH v6 5/5] ARM: dts: sun9i: Initial support for the Sunchip CX-A99 board

Hi,

Quoted 8 lines · Rask Ingemann Lambertsen

The Suncip CX-A99 board is found in at least four brands of media players. It features an Allwinner A80 ARM SoC and is found in two models:

1) 2 GiB DDR3 DRAM and 16 GB eMMC
2) 4 GiB DDR3 DRAM and 32 GB eMMC

For details of the board, see the linux-sunxi page <URL:https://linux-sunxi.org/Sunchip_CX-A99>.

Please don't put URLs in commit logs (and the DT).

Quoted 14 lines

Supported features (+ means tested):
+ One Cortex-A7 CPU core (or four with experimental U-Boot PSCI patches)
+ AXP808 power management chip
+ OZ80120 voltage regulator
+ Serial console port (internal)
+ eMMC and SD card slot
+ USB 2.0 host ports on on-board USB hub
+ SATA port on on-board SATA-to-USB bridge *
+ IEEE 802.11 a/b/g/n/ac SDIO Wifi
+ Real-time clock
+ LEDs
- IR receiver for remote control

Quoted 4 lines

* Only shows up when a SATA device is connected. Also, if a power source is connected to the USB 3.0 connector across power cycles (e.g. FEL boot), the bridge may not properly reset and not show up on the USB bus. The vendor U-Boot performs some unknown magic which resets the bridge.

Is that magic at the USB level, or specific to the bridge itself?

Quoted 9 lines

So far unsupported features:
- Using any of the Cortex-A15 CPU cores
- USB 3.0 port (except for supplying 5 V power)
- IEEE 802.3 10/100/1000 megabit Ethernet
- HDMI connector
- S/PDIF audio output
- Jack socket with composite video and analog stereo audio
- Bluetooth
- FM radio receiver (assuming it is even wired on the board)

I guess that should be in your cover letter.

This is not found in your DT, so no one really expects it to work :)

Quoted 150 lines

---

Changes in v6:
- Updated commit message description of SATA-to-USB bridge quirk and added note about experimental U-Boot PSCI support for up to four CPU cores.
- The blue LED is no longer on by default as its meaning is not documented.
- Removed "regulator-boot-on" from regulators having "regulator-always-on".
- Removed misleading mention of "OTG connector" which the device doesn't have.
- More detailed explanation for the need for "broken-cd" on mmc0.
- Several regulators have had their voltage range relaxed a little to match the permissible range according to the data sheets of the consumers. This is similar to what is used for the Cubieboard4 and Merrii A80 Optimus.
- Shortened regulator dcdce name as per v5 comments. A comment now lists the pin groups supplied by dcdce.

Changes in v5:
- Switched pinmux modes to generic properties and dropped #include <dt-bindings/pinctrl/sun4i-a10.h> as a consequence.
- Dropped pinctrl properties from GPIO nodes and dropped the pinmux nodes for them.
- AXP808 regulators added.
- Dropped the now unused #include <sunxi-common-regulators.dtsi>.
- Ampak AP6335 SDIO-Wifi added.
- USB Vbus changes as per v4 comments.
- Added "broken-cd" to mmc0 because GPIO interrupts don't work.

Changes in v4:
- Node names had underscores changed to hyphens.
- Changed formatting of the ac100/rtc node's clock output name list to match that of the same node in the cubieboard4 and a80-optimus device trees.

Changes in v3:
None.

Changes in v2:
- Fixed formatting and style issues found by scripts/checkpatch.pl.

 arch/arm/boot/dts/Makefile             |   3 +-
 arch/arm/boot/dts/sun9i-a80-cx-a99.dts | 409 +++++++++++++++++++++++++++++++++
 2 files changed, 411 insertions(+), 1 deletion(-)
 create mode 100644 arch/arm/boot/dts/sun9i-a80-cx-a99.dts
arch/arm/boot/dts/Makefile +2 -1
index 8553bd7..40546fa 100644
@@ -862,7 +862,8 @@ dtb-$(CONFIG_MACH_SUN8I) += \
 	sun8i-v3s-licheepi-zero.dtb
 dtb-$(CONFIG_MACH_SUN9I) += \
 	sun9i-a80-optimus.dtb \
-	sun9i-a80-cubieboard4.dtb
+	sun9i-a80-cubieboard4.dtb \
+	sun9i-a80-cx-a99.dtb
 dtb-$(CONFIG_ARCH_TANGO) += \
 	tango4-vantage-1172.dtb
 dtb-$(CONFIG_ARCH_TEGRA_2x_SOC) += \
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +87 -0
new file mode 100644
index 0000000..f5496d2
@@ -0,0 +1,409 @@
+/*
+ * sun9i-a80-cx-a99.dts - Device Tree file for the Sunchip CX-A99 board.
+ *
+ * Copyright (C) 2017 Rask Ingemann Lambertsen <[email protected]>
+ *
+ * This file is dual-licensed: you can use it either under the terms
+ * of the GPL or the X11 license, at your option. Note that this dual
+ * licensing only applies to this file, and not this project as a
+ * whole.
+ *
+ *  a) This file is free software; you can redistribute it and/or
+ *     modify it under the terms of the GNU General Public License as
+ *     published by the Free Software Foundation; either version 2 of the
+ *     License, or (at your option) any later version.
+ *
+ *     This file 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.
+ *
+ * Or, alternatively,
+ *
+ *  b) Permission is hereby granted, free of charge, to any person
+ *     obtaining a copy of this software and associated documentation
+ *     files (the "Software"), to deal in the Software without
+ *     restriction, including without limitation the rights to use,
+ *     copy, modify, merge, publish, distribute, sublicense, and/or
+ *     sell copies of the Software, and to permit persons to whom the
+ *     Software is furnished to do so, subject to the following
+ *     conditions:
+ *
+ *     The above copyright notice and this permission notice shall be
+ *     included in all copies or substantial portions of the Software.
+ *
+ *     THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ *     EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES
+ *     OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ *     NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
+ *     HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
+ *     WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ *     FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ *     OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+/*
+ * The Sunchip CX-A99 board is found in several similar Android media
+ * players, such as:
+ *
+ * Instabox Fantasy A8 (no external antenna)
+ * Jesurun CS-Q8 (ships with larger remote control)
+ * Jesurun Maxone
+ * Rikomagic (RKM) MK80/MK80LE
+ * Tronsmart Draco AW80 Meta/Telos
+ *
+ * See <URL:https://linux-sunxi.org/Sunchip_CX-A99> for more information.
+ */
+
+/dts-v1/;
+#include "sun9i-a80.dtsi"
+
+#include <dt-bindings/gpio/gpio.h>
+
+/ {
+	model = "Sunchip CX-A99";
+	compatible = "sunchip,cx-a99", "allwinner,sun9i-a80";
+
+	aliases {
+		serial0 = &uart0;
+	};
+
+	chosen {
+		stdout-path = "serial0:115200n8";
+	};
+
+	leds {
+		compatible = "gpio-leds";
+
+		blue {
+			gpios = <&pio 6 10 GPIO_ACTIVE_HIGH>;	/* PG10 */
+			label = "cx-a99:blue:status";
+		};
+
+		red {
+			gpios = <&pio 6 11 GPIO_ACTIVE_HIGH>;	/* PG11 */
+			label = "cx-a99:red:status";
+		};
+	};

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 97

Quoted 10 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +10 -0
+
+	powerseq_wifi: powerseq-wifi {
+		compatible = "mmc-pwrseq-simple";
+		clocks = <&ac100_rtc 1>;
+		clock-names = "ext_clock";
+		reset-gpios = <&r_pio 1 0 GPIO_ACTIVE_LOW>;	/* PM0 */
+		post-power-on-delay-ms = <1>;	/* Minimum 2 cycles. */
+	};
+
+	/* USB 3.0 standard-A receptacle. For now, only Vbus is supported. */

I'm not sure what you mean by "only VBUS is supported"? Is there any other signal?

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 104

Quoted 7 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +7 -0
+	reg_usb0_vbus: regulator-usb0-vbus {
+		compatible = "regulator-fixed";
+		regulator-name = "usb0-vbus";
+		regulator-min-microvolt = <5000000>;
+		regulator-max-microvolt = <5000000>;
+		gpio = <&pio 7 15 GPIO_ACTIVE_HIGH>;	/* PH15 */
+		enable-active-high;

This is redundant with the GPIO flag

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 105

Quoted 1 line
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +1 -0
+		regulator-always-on;

And it shouldn't be always on. The USB driver will enable it if needs be.

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 117

Quoted 12 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +12 -0
+	};
+
+	/*
+	 * A GL850G hub with two external USB connectors is connected
+	 * to ehci0. Each has a Vbus regulator controlled by a GPIO:
+	 * PL7 for port 1, closest to the 12 V power connector, and
+	 * PL8 for port 2, next to the SD card slot.
+	 * Because regulator-fixed doesn't support a GPIO list, and
+	 * allwinner,sun9i-a80-usb-phy doesn't support more than one
+	 * supply, we have to use regulator-always-on on usb1-2-vbus.
+	 * Note that the GPIO pins also need cldo1 to be enabled.
+	 */

What is the source of those regulators connected then? Some PMIC regulator? AC-IN?

Quoted 7 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +7 -0
+	reg_usb1_1_vbus: regulator-usb1-1-vbus {
+		compatible = "regulator-fixed";
+		regulator-name = "usb1-1-vbus";
+		regulator-min-microvolt = <5000000>;
+		regulator-max-microvolt = <5000000>;
+		gpio = <&r_pio 0 7 GPIO_ACTIVE_HIGH>;	/* PL7 */
+		enable-active-high;

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 134

Quoted 10 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +10 -0
+	};
+
+	reg_usb1_2_vbus: regulator-usb1-2-vbus {
+		compatible = "regulator-fixed";
+		regulator-name = "usb1-2-vbus";
+		regulator-min-microvolt = <5000000>;
+		regulator-max-microvolt = <5000000>;
+		gpio = <&r_pio 0 8 GPIO_ACTIVE_HIGH>;	/* PL8 */
+		enable-active-high;
+		regulator-always-on;

Same comment about always on. If the driver needs fixing to grab an additional regulator, fix it, but this shouldn't be left that way.

Quoted 16 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +16 -0
+	};
+
+	/* OZ80120 voltage regulator for the four Cortex-A15 cores. */
+	reg_vdd_cpub: regulator-vdd-cpub {
+		compatible = "regulator-gpio";
+
+		regulator-always-on;
+		regulator-min-microvolt = < 800000>;
+		regulator-max-microvolt = <1100000>;
+		regulator-name = "vdd-cpub";
+
+		enable-gpio = <&r_pio 0 2 GPIO_ACTIVE_HIGH>;	/* PL2 */
+		enable-active-high;
+		gpios = <&r_pio 0 3 GPIO_ACTIVE_HIGH>,		/* PL3 */
+			<&r_pio 0 4 GPIO_ACTIVE_HIGH>,		/* PL4 */
+			<&r_pio 0 5 GPIO_ACTIVE_HIGH>;		/* PL5 */

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 160

Quoted 10 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +10 -0
+
+		gpios-states = <1 0 0>;
+		states = <	 750000 0x7
+				 800000 0x3
+				 850000 0x5
+				 900000 0x1
+				 950000 0x6
+				1000000 0x2
+				1100000 0x4
+				1200000 0x0>;

You're listing a minimum state of 750mv, yet your minimum voltage is 800mV.

Quoted 18 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +18 -0
+	};
+};
+
+&ehci0 {
+	status = "okay";
+};
+
+&ehci2 {
+	status = "okay";
+};
+
+/*
+ * SD card slot. Although the GPIO pin for card detection is listed as capable
+ * of generating interrupts in the "A80 User Manual", this doesn't work for
+ * some unknown reason, so poll the GPIO for card detection. This is also what
+ * the vendor sys_config.fex file specifies.
+ */
+&mmc0 {

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 188

Quoted 10 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +10 -0
+	bus-width = <4>;
+	cd-gpios = <&pio 7 17 GPIO_ACTIVE_LOW>;	/* PH17 */
+	broken-cd;				/* Poll. */
+	pinctrl-names = "default";
+	pinctrl-0 = <&mmc0_pins>;
+	vmmc-supply = <&reg_dcdce>;
+	status = "okay";
+};
+
+/* Ampak AP6335 IEEE 802.11 a/b/g/n/ac "Wifi". */

Why "wifi" ? It's not implementing true wifi?

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 195

Quoted 7 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +7 -0
+&mmc1 {
+	bus-width = <4>;
+	non-removable;
+	pinctrl-names = "default";
+	pinctrl-0 = <&mmc1_pins>;
+	vmmc-supply = <&reg_cldo3>;	/* See cldo2,cldo3 note. */
+	vqmmc-supply = <&reg_aldo2>;

So it's able to support 1.2 or 1.8V IO modes? Surely you want to enable those modes here to.

Quoted 48 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +48 -0
+	mmc-pwrseq = <&powerseq_wifi>;
+	status = "okay";
+};
+
+/* On-board eMMC card. */
+&mmc2 {
+	bus-width = <8>;
+	non-removable;
+	pinctrl-names = "default";
+	pinctrl-0 = <&mmc2_8bit_pins>;
+	vmmc-supply = <&reg_dcdce>;
+	status = "okay";
+};
+
+&osc32k {
+	clocks = <&ac100_rtc 0>;
+};
+
+&r_ir {
+	status = "okay";
+};
+
+&r_rsb {
+	status = "okay";
+
+	ac100: codec at e89 {
+		compatible = "x-powers,ac100";
+		reg = <0xe89>;
+
+		ac100_codec: codec {
+			compatible = "x-powers,ac100-codec";
+			interrupt-parent = <&r_pio>;
+			interrupts = <0 9 IRQ_TYPE_LEVEL_LOW>;	/* PL9 */
+			#clock-cells = <0>;
+			clock-output-names = "4M_adda";
+		};
+
+		ac100_rtc: rtc {
+			compatible = "x-powers,ac100-rtc";
+			interrupt-parent = <&nmi_intc>;
+			interrupts = <0 IRQ_TYPE_LEVEL_LOW>;
+			clocks = <&ac100_codec>;
+			#clock-cells = <1>;
+			clock-output-names = "cko1_rtc",
+					     "cko2_rtc",
+					     "cko3_rtc";
+		};
+	};

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 253

Quoted 10 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +10 -0
+
+	pmic at 745 {
+		compatible = "x-powers,axp808", "x-powers,axp806";
+		reg = <0x745>;
+		interrupt-parent = <&nmi_intc>;
+		interrupts = <0 IRQ_TYPE_LEVEL_LOW>;
+		interrupt-controller;
+		#interrupt-cells = <1>;
+
+		swin-supply = <&reg_dcdce>;

Please use an incude for that PMIC.

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 257

Quoted 4 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +4 -0
+
+		/* In comments: Initial setup from vendor sys_config.fex. */
+		regulators {
+			/* 3.0 V (enabled). */

This might be disabled though.

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 270

Quoted 13 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +13 -0
+			reg_aldo1: aldo1 {
+				regulator-boot-on;
+				regulator-min-microvolt = <3000000>;
+				regulator-max-microvolt = <3000000>;
+				regulator-name = "vcc-3v0";
+			};
+
+			/* 1.8 V (enabled). */
+			reg_aldo2: aldo2 {
+				regulator-boot-on;
+				regulator-min-microvolt = <1800000>;
+				regulator-max-microvolt = <3600000>;
+				regulator-name = "vcc-pg-pm-wifi+btio-audio";

Usually, there is simpler names available on the schematics, or at least simpler names we can come up with.

Something like vcc-wifi would be enough her.

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 278

Quoted 8 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +8 -0
+			};
+
+			/* 2.5 V (enabled). */
+			reg_aldo3: aldo3 {
+				regulator-boot-on;
+				regulator-min-microvolt = <2500000>;
+				regulator-max-microvolt = <2500000>;
+				regulator-name = "vcc-pa-gmac2v5";

vcc-gmac

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 286

Quoted 8 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +8 -0
+			};
+
+			/* 1.8 V (enabled). */
+			reg_bldo1: bldo1 {
+				regulator-always-on;	/* Hang if disabled */
+				regulator-min-microvolt = <1700000>;
+				regulator-max-microvolt = <1900000>;
+				regulator-name = "vdd18-dll-vcc18-pll";

vdd-dll

Quoted 5 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +5 -0
+			};
+
+			/* 0.9 V (enabled). */
+			reg_bldo2: bldo2 {
+				regulator-always-on;	/* Hang if disabled */

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 301

Quoted 10 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +10 -0
+				regulator-min-microvolt = < 800000>;
+				regulator-max-microvolt = <1100000>;
+				regulator-name = "vdd-cpus";
+			};
+
+			/* 1.2 V (disabled). */
+			reg_bldo3: bldo3 {
+				regulator-min-microvolt = <1100000>;
+				regulator-max-microvolt = <1300000>;
+				regulator-name = "vcc12-hsic";

vcc-hsic

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 309

Quoted 8 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +8 -0
+			};
+
+			/* 1.1 V (enabled). */
+			reg_bldo4: bldo4 {
+				regulator-boot-on;
+				regulator-min-microvolt = < 800000>;
+				regulator-max-microvolt = <1100000>;
+				regulator-name = "vdd09-hdmi";

vdd-hdmi

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 317

Quoted 8 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +8 -0
+			};
+
+			/* 3.3 V (enabled). PLx pins control some regulators. */
+			reg_cldo1: cldo1 {
+				regulator-always-on;
+				regulator-min-microvolt = <3300000>;
+				regulator-max-microvolt = <3300000>;
+				regulator-name = "vcc-pl-led";

vcc-led, etc...

Thanks,
Maxime

Signature · 10 lines
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 801 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20170210/ec920ad5/attachment-0001.sig>
#7 lore raw

Re: [PATCH v6 5/5] ARM: dts: sun9i: Initial support for the Sunchip CX-A99 board

Quoted 380 lines · Maxime Ripard

Hi,

Quoted 8 lines · Rask Ingemann Lambertsen

The Suncip CX-A99 board is found in at least four brands of media players. It features an Allwinner A80 ARM SoC and is found in two models:

1) 2 GiB DDR3 DRAM and 16 GB eMMC
2) 4 GiB DDR3 DRAM and 32 GB eMMC

For details of the board, see the linux-sunxi page <URL:https://linux-sunxi.org/Sunchip_CX-A99>.

Please don't put URLs in commit logs (and the DT).

Quoted 14 lines

Supported features (+ means tested):
+ One Cortex-A7 CPU core (or four with experimental U-Boot PSCI patches)
+ AXP808 power management chip
+ OZ80120 voltage regulator
+ Serial console port (internal)
+ eMMC and SD card slot
+ USB 2.0 host ports on on-board USB hub
+ SATA port on on-board SATA-to-USB bridge *
+ IEEE 802.11 a/b/g/n/ac SDIO Wifi
+ Real-time clock
+ LEDs
- IR receiver for remote control

Quoted 4 lines

* Only shows up when a SATA device is connected. Also, if a power source is connected to the USB 3.0 connector across power cycles (e.g. FEL boot), the bridge may not properly reset and not show up on the USB bus. The vendor U-Boot performs some unknown magic which resets the bridge.

Is that magic at the USB level, or specific to the bridge itself?

Quoted 9 lines

So far unsupported features:
- Using any of the Cortex-A15 CPU cores
- USB 3.0 port (except for supplying 5 V power)
- IEEE 802.3 10/100/1000 megabit Ethernet
- HDMI connector
- S/PDIF audio output
- Jack socket with composite video and analog stereo audio
- Bluetooth
- FM radio receiver (assuming it is even wired on the board)

I guess that should be in your cover letter.

This is not found in your DT, so no one really expects it to work :)

Quoted 150 lines

---

Changes in v6:
- Updated commit message description of SATA-to-USB bridge quirk and added note about experimental U-Boot PSCI support for up to four CPU cores.
- The blue LED is no longer on by default as its meaning is not documented.
- Removed "regulator-boot-on" from regulators having "regulator-always-on".
- Removed misleading mention of "OTG connector" which the device doesn't have.
- More detailed explanation for the need for "broken-cd" on mmc0.
- Several regulators have had their voltage range relaxed a little to match the permissible range according to the data sheets of the consumers. This is similar to what is used for the Cubieboard4 and Merrii A80 Optimus.
- Shortened regulator dcdce name as per v5 comments. A comment now lists the pin groups supplied by dcdce.

Changes in v5:
- Switched pinmux modes to generic properties and dropped #include <dt-bindings/pinctrl/sun4i-a10.h> as a consequence.
- Dropped pinctrl properties from GPIO nodes and dropped the pinmux nodes for them.
- AXP808 regulators added.
- Dropped the now unused #include <sunxi-common-regulators.dtsi>.
- Ampak AP6335 SDIO-Wifi added.
- USB Vbus changes as per v4 comments.
- Added "broken-cd" to mmc0 because GPIO interrupts don't work.

Changes in v4:
- Node names had underscores changed to hyphens.
- Changed formatting of the ac100/rtc node's clock output name list to match that of the same node in the cubieboard4 and a80-optimus device trees.

Changes in v3:
None.

Changes in v2:
- Fixed formatting and style issues found by scripts/checkpatch.pl.

 arch/arm/boot/dts/Makefile             |   3 +-
 arch/arm/boot/dts/sun9i-a80-cx-a99.dts | 409 +++++++++++++++++++++++++++++++++
 2 files changed, 411 insertions(+), 1 deletion(-)
 create mode 100644 arch/arm/boot/dts/sun9i-a80-cx-a99.dts
arch/arm/boot/dts/Makefile +2 -1
index 8553bd7..40546fa 100644
@@ -862,7 +862,8 @@ dtb-$(CONFIG_MACH_SUN8I) += \
      sun8i-v3s-licheepi-zero.dtb
 dtb-$(CONFIG_MACH_SUN9I) += \
      sun9i-a80-optimus.dtb \
-     sun9i-a80-cubieboard4.dtb
+     sun9i-a80-cubieboard4.dtb \
+     sun9i-a80-cx-a99.dtb
 dtb-$(CONFIG_ARCH_TANGO) += \
      tango4-vantage-1172.dtb
 dtb-$(CONFIG_ARCH_TEGRA_2x_SOC) += \
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +87 -0
new file mode 100644
index 0000000..f5496d2
@@ -0,0 +1,409 @@
+/*
+ * sun9i-a80-cx-a99.dts - Device Tree file for the Sunchip CX-A99 board.
+ *
+ * Copyright (C) 2017 Rask Ingemann Lambertsen <[email protected]>
+ *
+ * This file is dual-licensed: you can use it either under the terms
+ * of the GPL or the X11 license, at your option. Note that this dual
+ * licensing only applies to this file, and not this project as a
+ * whole.
+ *
+ *  a) This file is free software; you can redistribute it and/or
+ *     modify it under the terms of the GNU General Public License as
+ *     published by the Free Software Foundation; either version 2 of the
+ *     License, or (at your option) any later version.
+ *
+ *     This file 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.
+ *
+ * Or, alternatively,
+ *
+ *  b) Permission is hereby granted, free of charge, to any person
+ *     obtaining a copy of this software and associated documentation
+ *     files (the "Software"), to deal in the Software without
+ *     restriction, including without limitation the rights to use,
+ *     copy, modify, merge, publish, distribute, sublicense, and/or
+ *     sell copies of the Software, and to permit persons to whom the
+ *     Software is furnished to do so, subject to the following
+ *     conditions:
+ *
+ *     The above copyright notice and this permission notice shall be
+ *     included in all copies or substantial portions of the Software.
+ *
+ *     THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ *     EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES
+ *     OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ *     NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
+ *     HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
+ *     WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ *     FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ *     OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+/*
+ * The Sunchip CX-A99 board is found in several similar Android media
+ * players, such as:
+ *
+ * Instabox Fantasy A8 (no external antenna)
+ * Jesurun CS-Q8 (ships with larger remote control)
+ * Jesurun Maxone
+ * Rikomagic (RKM) MK80/MK80LE
+ * Tronsmart Draco AW80 Meta/Telos
+ *
+ * See <URL:https://linux-sunxi.org/Sunchip_CX-A99> for more information.
+ */
+
+/dts-v1/;
+#include "sun9i-a80.dtsi"
+
+#include <dt-bindings/gpio/gpio.h>
+
+/ {
+     model = "Sunchip CX-A99";
+     compatible = "sunchip,cx-a99", "allwinner,sun9i-a80";
+
+     aliases {
+             serial0 = &uart0;
+     };
+
+     chosen {
+             stdout-path = "serial0:115200n8";
+     };
+
+     leds {
+             compatible = "gpio-leds";
+
+             blue {
+                     gpios = <&pio 6 10 GPIO_ACTIVE_HIGH>;   /* PG10 */
+                     label = "cx-a99:blue:status";
+             };
+
+             red {
+                     gpios = <&pio 6 11 GPIO_ACTIVE_HIGH>;   /* PG11 */
+                     label = "cx-a99:red:status";
+             };
+     };

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 97

Quoted 10 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +10 -0
+
+     powerseq_wifi: powerseq-wifi {
+             compatible = "mmc-pwrseq-simple";
+             clocks = <&ac100_rtc 1>;
+             clock-names = "ext_clock";
+             reset-gpios = <&r_pio 1 0 GPIO_ACTIVE_LOW>;     /* PM0 */
+             post-power-on-delay-ms = <1>;   /* Minimum 2 cycles. */
+     };
+
+     /* USB 3.0 standard-A receptacle. For now, only Vbus is supported. */

I'm not sure what you mean by "only VBUS is supported"? Is there any other signal?

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 104

Quoted 7 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +7 -0
+     reg_usb0_vbus: regulator-usb0-vbus {
+             compatible = "regulator-fixed";
+             regulator-name = "usb0-vbus";
+             regulator-min-microvolt = <5000000>;
+             regulator-max-microvolt = <5000000>;
+             gpio = <&pio 7 15 GPIO_ACTIVE_HIGH>;    /* PH15 */
+             enable-active-high;

This is redundant with the GPIO flag

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 105

Quoted 1 line
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +1 -0
+             regulator-always-on;

And it shouldn't be always on. The USB driver will enable it if needs be.

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 117

Quoted 12 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +12 -0
+     };
+
+     /*
+      * A GL850G hub with two external USB connectors is connected
+      * to ehci0. Each has a Vbus regulator controlled by a GPIO:
+      * PL7 for port 1, closest to the 12 V power connector, and
+      * PL8 for port 2, next to the SD card slot.
+      * Because regulator-fixed doesn't support a GPIO list, and
+      * allwinner,sun9i-a80-usb-phy doesn't support more than one
+      * supply, we have to use regulator-always-on on usb1-2-vbus.
+      * Note that the GPIO pins also need cldo1 to be enabled.
+      */

What is the source of those regulators connected then? Some PMIC regulator? AC-IN?

Quoted 7 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +7 -0
+     reg_usb1_1_vbus: regulator-usb1-1-vbus {
+             compatible = "regulator-fixed";
+             regulator-name = "usb1-1-vbus";
+             regulator-min-microvolt = <5000000>;
+             regulator-max-microvolt = <5000000>;
+             gpio = <&r_pio 0 7 GPIO_ACTIVE_HIGH>;   /* PL7 */
+             enable-active-high;

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 134

Quoted 10 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +10 -0
+     };
+
+     reg_usb1_2_vbus: regulator-usb1-2-vbus {
+             compatible = "regulator-fixed";
+             regulator-name = "usb1-2-vbus";
+             regulator-min-microvolt = <5000000>;
+             regulator-max-microvolt = <5000000>;
+             gpio = <&r_pio 0 8 GPIO_ACTIVE_HIGH>;   /* PL8 */
+             enable-active-high;
+             regulator-always-on;

Same comment about always on. If the driver needs fixing to grab an additional regulator, fix it, but this shouldn't be left that way.

Quoted 16 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +16 -0
+     };
+
+     /* OZ80120 voltage regulator for the four Cortex-A15 cores. */
+     reg_vdd_cpub: regulator-vdd-cpub {
+             compatible = "regulator-gpio";
+
+             regulator-always-on;
+             regulator-min-microvolt = < 800000>;
+             regulator-max-microvolt = <1100000>;
+             regulator-name = "vdd-cpub";
+
+             enable-gpio = <&r_pio 0 2 GPIO_ACTIVE_HIGH>;    /* PL2 */
+             enable-active-high;
+             gpios = <&r_pio 0 3 GPIO_ACTIVE_HIGH>,          /* PL3 */
+                     <&r_pio 0 4 GPIO_ACTIVE_HIGH>,          /* PL4 */
+                     <&r_pio 0 5 GPIO_ACTIVE_HIGH>;          /* PL5 */

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 160

Quoted 10 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +10 -0
+
+             gpios-states = <1 0 0>;
+             states = <       750000 0x7
+                              800000 0x3
+                              850000 0x5
+                              900000 0x1
+                              950000 0x6
+                             1000000 0x2
+                             1100000 0x4
+                             1200000 0x0>;

You're listing a minimum state of 750mv, yet your minimum voltage is 800mV.

Quoted 18 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +18 -0
+     };
+};
+
+&ehci0 {
+     status = "okay";
+};
+
+&ehci2 {
+     status = "okay";
+};
+
+/*
+ * SD card slot. Although the GPIO pin for card detection is listed as capable
+ * of generating interrupts in the "A80 User Manual", this doesn't work for
+ * some unknown reason, so poll the GPIO for card detection. This is also what
+ * the vendor sys_config.fex file specifies.
+ */
+&mmc0 {

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 188

Quoted 10 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +10 -0
+     bus-width = <4>;
+     cd-gpios = <&pio 7 17 GPIO_ACTIVE_LOW>; /* PH17 */
+     broken-cd;                              /* Poll. */
+     pinctrl-names = "default";
+     pinctrl-0 = <&mmc0_pins>;
+     vmmc-supply = <&reg_dcdce>;
+     status = "okay";
+};
+
+/* Ampak AP6335 IEEE 802.11 a/b/g/n/ac "Wifi". */

Why "wifi" ? It's not implementing true wifi?

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 195

Quoted 7 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +7 -0
+&mmc1 {
+     bus-width = <4>;
+     non-removable;
+     pinctrl-names = "default";
+     pinctrl-0 = <&mmc1_pins>;
+     vmmc-supply = <&reg_cldo3>;     /* See cldo2,cldo3 note. */
+     vqmmc-supply = <&reg_aldo2>;

So it's able to support 1.2 or 1.8V IO modes? Surely you want to enable those modes here to.

Quoted 43 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +43 -0
+     mmc-pwrseq = <&powerseq_wifi>;
+     status = "okay";
+};
+
+/* On-board eMMC card. */
+&mmc2 {
+     bus-width = <8>;
+     non-removable;
+     pinctrl-names = "default";
+     pinctrl-0 = <&mmc2_8bit_pins>;
+     vmmc-supply = <&reg_dcdce>;
+     status = "okay";
+};
+
+&osc32k {
+     clocks = <&ac100_rtc 0>;
+};
+
+&r_ir {
+     status = "okay";
+};
+
+&r_rsb {
+     status = "okay";
+
+     ac100: codec at e89 {
+             compatible = "x-powers,ac100";
+             reg = <0xe89>;
+
+             ac100_codec: codec {
+                     compatible = "x-powers,ac100-codec";
+                     interrupt-parent = <&r_pio>;
+                     interrupts = <0 9 IRQ_TYPE_LEVEL_LOW>;  /* PL9 */
+                     #clock-cells = <0>;
+                     clock-output-names = "4M_adda";
+             };
+
+             ac100_rtc: rtc {
+                     compatible = "x-powers,ac100-rtc";
+                     interrupt-parent = <&nmi_intc>;
+                     interrupts = <0 IRQ_TYPE_LEVEL_LOW>;
+                     clocks = <&ac100_codec>;
+                     #clock-cells = <1>;
Quoted 8 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 246

Quoted 8 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +8 -0
+                     clock-output-names = "cko1_rtc",
+                                          "cko2_rtc",
+                                          "cko3_rtc";
+             };
+     };
+
+     pmic at 745 {
+             compatible = "x-powers,axp808", "x-powers,axp806";

As you mentioned elsewhere, they are not really compatible. You should drop the latter compatible.

Quoted 9 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 253

Quoted 7 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +7 -0
+             reg = <0x745>;
+             interrupt-parent = <&nmi_intc>;
+             interrupts = <0 IRQ_TYPE_LEVEL_LOW>;
+             interrupt-controller;
+             #interrupt-cells = <1>;
+
+             swin-supply = <&reg_dcdce>;

Please use an incude for that PMIC.

Hmm, there isn't one for the AXP806 either.

Some more info about the regulator names:

Quoted 21 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 257

Quoted 4 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +4 -0
+
+             /* In comments: Initial setup from vendor sys_config.fex. */
+             regulators {
+                     /* 3.0 V (enabled). */

This might be disabled though.

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 270

Quoted 13 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +13 -0
+                     reg_aldo1: aldo1 {
+                             regulator-boot-on;
+                             regulator-min-microvolt = <3000000>;
+                             regulator-max-microvolt = <3000000>;
+                             regulator-name = "vcc-3v0";
+                     };
+
+                     /* 1.8 V (enabled). */
+                     reg_aldo2: aldo2 {
+                             regulator-boot-on;
+                             regulator-min-microvolt = <1800000>;
+                             regulator-max-microvolt = <3600000>;
+                             regulator-name = "vcc-pg-pm-wifi+btio-audio";
Quoted 4 lines

Usually, there is simpler names available on the schematics, or at least simpler names we can come up with.

Something like vcc-wifi would be enough her.

It should be vddio-wifi. Looking at the pin groups it names, it likely just drives the I/O pins on both ends.

Quoted 11 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 278

Quoted 9 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +8 -0
+                     };
+
+                     /* 2.5 V (enabled). */
+                     reg_aldo3: aldo3 {
+                             regulator-boot-on;
+                             regulator-min-microvolt = <2500000>;
+                             regulator-max-microvolt = <2500000>;
+                             regulator-name = "vcc-pa-gmac2v5";

vcc-gmac

vcc-pa or vddio-gmac. 2.5V is for RGMII I/O. vcc-gmac is the "sw" regulator @ 3.3V.

Quoted 11 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 286

Quoted 9 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +8 -0
+                     };
+
+                     /* 1.8 V (enabled). */
+                     reg_bldo1: bldo1 {
+                             regulator-always-on;    /* Hang if disabled */
+                             regulator-min-microvolt = <1700000>;
+                             regulator-max-microvolt = <1900000>;
+                             regulator-name = "vdd18-dll-vcc18-pll";

vdd-dll

PLLs and DLLs are different though. Maybe vcc-pll-dll?

Quoted 10 lines
Quoted 10 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +9 -0
+                     };
+
+                     /* 0.9 V (enabled). */
+                     reg_bldo2: bldo2 {
+                             regulator-always-on;    /* Hang if disabled */
+                             regulator-min-microvolt = < 800000>;
+                             regulator-max-microvolt = <1100000>;
+                             regulator-name = "vdd-cpus";
+                     };
Quoted 8 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 301

Quoted 6 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +6 -0
+
+                     /* 1.2 V (disabled). */
+                     reg_bldo3: bldo3 {
+                             regulator-min-microvolt = <1100000>;
+                             regulator-max-microvolt = <1300000>;
+                             regulator-name = "vcc12-hsic";

vcc-hsic

vcc12-hsic is actually the name listed in the SoC datasheet.

Quoted 11 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 309

Quoted 9 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +8 -0
+                     };
+
+                     /* 1.1 V (enabled). */
+                     reg_bldo4: bldo4 {
+                             regulator-boot-on;
+                             regulator-min-microvolt = < 800000>;
+                             regulator-max-microvolt = <1100000>;
+                             regulator-name = "vdd09-hdmi";

vdd-hdmi

vdd09-hdmi is actually the name listed in the SoC datasheet.

Quoted 11 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 317

Quoted 9 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +8 -0
+                     };
+
+                     /* 3.3 V (enabled). PLx pins control some regulators. */
+                     reg_cldo1: cldo1 {
+                             regulator-always-on;
+                             regulator-min-microvolt = <3300000>;
+                             regulator-max-microvolt = <3300000>;
+                             regulator-name = "vcc-pl-led";

vcc-led, etc...

vcc-pl is probably better... One can figure out the LEDs are connected to the PL group and maybe realize they are powered this way. Not that easy the other way around.

I used really long names for the other 2 A80 boards though.

Regards
ChenYu

Quoted 8 lines

Thanks,
Maxime

--
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com

#8 lore raw

Re: [PATCH v6 5/5] ARM: dts: sun9i: Initial support for the Sunchip CX-A99 board

Let me just answer the questions about the regulator names and get back to you about your other points.

Quoted 5 lines · Chen-Yu Tsai
Quoted 3 lines · Maxime Ripard

Hi,

On Thu, Feb 09, 2017 at 12:34:06AM +0100, Rask Ingemann Lambertsen wrote:

[...]

Quoted 7 lines
Quoted 7 lines
Quoted 6 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +6 -0
+                     /* 1.8 V (enabled). */
+                     reg_aldo2: aldo2 {
+                             regulator-boot-on;
+                             regulator-min-microvolt = <1800000>;
+                             regulator-max-microvolt = <3600000>;
+                             regulator-name = "vcc-pg-pm-wifi+btio-audio";
Quoted 7 lines
Quoted 4 lines

Usually, there is simpler names available on the schematics, or at least simpler names we can come up with.

Something like vcc-wifi would be enough her.

It should be vddio-wifi. Looking at the pin groups it names, it likely just drives the I/O pins on both ends.

Sys_config.fex lists is as supplying the audio codec as well, so "vddio-wifi-codec" perhaps?

Quoted 11 lines
Quoted 8 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 278

Quoted 6 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +6 -0
+                     /* 2.5 V (enabled). */
+                     reg_aldo3: aldo3 {
+                             regulator-boot-on;
+                             regulator-min-microvolt = <2500000>;
+                             regulator-max-microvolt = <2500000>;
+                             regulator-name = "vcc-pa-gmac2v5";

vcc-gmac

vcc-pa or vddio-gmac. 2.5V is for RGMII I/O. vcc-gmac is the "sw" regulator @ 3.3V.

I think nothing else is connected to pin group A, so I'll call it "vddio-gmac".

Quoted 10 lines
Quoted 8 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 286

Quoted 6 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +6 -0
+                     /* 1.8 V (enabled). */
+                     reg_bldo1: bldo1 {
+                             regulator-always-on;    /* Hang if disabled */
+                             regulator-min-microvolt = <1700000>;
+                             regulator-max-microvolt = <1900000>;
+                             regulator-name = "vdd18-dll-vcc18-pll";

vdd-dll

PLLs and DLLs are different though. Maybe vcc-pll-dll?

The SoC data sheet uses the names "VDD18-DLL" and "VCC18-PLL". However, it's quite likely that more of the SoC's 1.8 V consumers are connected to bldo1, as I see no other likely 1.8 V supply for them. I'll call it "vdd18".

Quoted 7 lines
Quoted 7 lines
Quoted 6 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +6 -0
+                     /* 3.3 V (enabled). PLx pins control some regulators. */
+                     reg_cldo1: cldo1 {
+                             regulator-always-on;
+                             regulator-min-microvolt = <3300000>;
+                             regulator-max-microvolt = <3300000>;
+                             regulator-name = "vcc-pl-led";
Quoted 5 lines
Quoted 1 line

vcc-led, etc...

vcc-pl is probably better... One can figure out the LEDs are connected to the PL group and maybe realize they are powered this way. Not that easy the other way around.

The LEDs are controlled by PG10 and PG11 but get their power from cldo1, which is connected to the (common) anode of the red and blue LED. Cldo1 also supplies the SoCs "VCC-PL", which powers pin group L that controls the USB 2.0 Vbus and vdd-cpub regulators. Pin group L has nothing to do with the LEDs. Thus I think the regulator name is appropriate in this case.

Signature · 1 line
Rask Ingemann Lambertsen
#9 lore raw

Re: [PATCH v6 5/5] ARM: dts: sun9i: Initial support for the Sunchip CX-A99 board

In reply to Chen-Yu Tsai
Quoted 5 lines · Chen-Yu Tsai
Quoted 3 lines · Maxime Ripard

Hi,

On Thu, Feb 09, 2017 at 12:34:06AM +0100, Rask Ingemann Lambertsen wrote:

[...]

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +0 -0
new file mode 100644
index 0000000..f5496d2

[...]

Quoted 5 lines
Quoted 2 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 246

Quoted 2 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +2 -0
+     pmic at 745 {
+             compatible = "x-powers,axp808", "x-powers,axp806";

As you mentioned elsewhere, they are not really compatible. You should drop the latter compatible.

The compatibility is good enough that it works fine with the driver that went into kernel 4.9 and 4.10. Using this device tree file, I have built and booted 4.9.9 and 4.10-rc7 kernels as per the instructions here;
https://linux-sunxi.org/Sunchip_CX-A99#Linux_kernel
Likewise any other AXP806 driver which doesn't touch the register at address 0xff will work fine.

So, the disadvantage of removing the "x-powers,axp806" compatible is that it breaks on two kernel versions where it works fine with the compatible. What is the advantage of removing the "x-powers,axp806" compatible?

As to touching the register at 0xff, it would be interesting to know more precisely what goes wrong. The symptom is these messages in the log:

[    3.209573] axp20x-rsb sunxi-rsb-745: AXP20x variant AXP806 found
[    3.210153] axp20x-rsb sunxi-rsb-745: Failed to set masks in 0x40: -5
[    3.210178] axp20x-rsb sunxi-rsb-745: failed to add irq chip: -5
[    3.210306] axp20x-rsb: probe of sunxi-rsb-745 failed with error -5

I haven't yet looked further into what happens at the RSB bus level.

Signature · 1 line
Rask Ingemann Lambertsen
#14 lore raw

Re: [PATCH v6 5/5] ARM: dts: sun9i: Initial support for the Sunchip CX-A99 board

In reply to Maxime Ripard
Quoted 6 lines · Maxime Ripard

Hi,

Quoted 3 lines · Rask Ingemann Lambertsen

The Suncip CX-A99 board is found in at least four brands of media players. It features an Allwinner A80 ARM SoC and is found in two models:

Quoted 7 lines
Quoted 5 lines

1) 2 GiB DDR3 DRAM and 16 GB eMMC
2) 4 GiB DDR3 DRAM and 32 GB eMMC

For details of the board, see the linux-sunxi page <URL:https://linux-sunxi.org/Sunchip_CX-A99>.

Please don't put URLs in commit logs (and the DT).

OK.

Quoted 13 lines
Quoted 13 lines

Supported features (+ means tested):
+ One Cortex-A7 CPU core (or four with experimental U-Boot PSCI patches)
+ AXP808 power management chip
+ OZ80120 voltage regulator
+ Serial console port (internal)
+ eMMC and SD card slot
+ USB 2.0 host ports on on-board USB hub
+ SATA port on on-board SATA-to-USB bridge *
+ IEEE 802.11 a/b/g/n/ac SDIO Wifi
+ Real-time clock
+ LEDs
- IR receiver for remote control

Quoted 6 lines
Quoted 4 lines

* Only shows up when a SATA device is connected. Also, if a power source is connected to the USB 3.0 connector across power cycles (e.g. FEL boot), the bridge may not properly reset and not show up on the USB bus. The vendor U-Boot performs some unknown magic which resets the bridge.

Is that magic at the USB level, or specific to the bridge itself?

I don't know, but since the other USB port connected to the hub comes up fine, I'm assuming it's something to do with the SATA-to-USB bridge.

Quoted 5 lines
Quoted 5 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +0 -0
new file mode 100644
index 0000000..f5496d2

[...]

Quoted 4 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 97

Quoted 1 line
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +1 -0
+	/* USB 3.0 standard-A receptacle. For now, only Vbus is supported. */

I'm not sure what you mean by "only VBUS is supported"? Is there any other signal?

I'm just trying to say that at the moment, all the connector will do is to supply power to a connected device (for charging batteries or whatever).

Quoted 9 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 104

Quoted 7 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +7 -0
+	reg_usb0_vbus: regulator-usb0-vbus {
+		compatible = "regulator-fixed";
+		regulator-name = "usb0-vbus";
+		regulator-min-microvolt = <5000000>;
+		regulator-max-microvolt = <5000000>;
+		gpio = <&pio 7 15 GPIO_ACTIVE_HIGH>;	/* PH15 */
+		enable-active-high;

This is redundant with the GPIO flag

No. I have now tested without "enable-active-high" and then Vbus stays off. This is also in line with the binding documentation:

"- enable-active-high: Polarity of GPIO is Active high
If this property is missing, the default assumed is Active low."

It also seems to me that the way this is implemented in Linux, the GPIO polarity flag is ignored.

Quoted 4 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts

Quoted 1 line
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +1 -0
+		regulator-always-on;

And it shouldn't be always on. The USB driver will enable it if needs be.

Yes, we just don't have a driver for the A80 USB 3.0 port yet.

Quoted 5 lines
Quoted 5 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +5 -0
+	/*
+	 * A GL850G hub with two external USB connectors is connected
+	 * to ehci0. Each has a Vbus regulator controlled by a GPIO:
+	 * PL7 for port 1, closest to the 12 V power connector, and
+	 * PL8 for port 2, next to the SD card slot.
Quoted 8 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 117

Quoted 5 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +5 -0
+	 * Because regulator-fixed doesn't support a GPIO list, and
+	 * allwinner,sun9i-a80-usb-phy doesn't support more than one
+	 * supply, we have to use regulator-always-on on usb1-2-vbus.
+	 * Note that the GPIO pins also need cldo1 to be enabled.
+	 */

What is the source of those regulators connected then? Some PMIC regulator? AC-IN?

These two regulators are supplied from a fixed 12 V to 5 V DC-DC converter which can't be controlled in any way.

Quoted 12 lines
Quoted 12 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +12 -0
+	reg_usb1_1_vbus: regulator-usb1-1-vbus {
+		compatible = "regulator-fixed";
+		regulator-name = "usb1-1-vbus";
+		regulator-min-microvolt = <5000000>;
+		regulator-max-microvolt = <5000000>;
+		gpio = <&r_pio 0 7 GPIO_ACTIVE_HIGH>;	/* PL7 */
+		enable-active-high;
+	};
+
+	reg_usb1_2_vbus: regulator-usb1-2-vbus {
+		compatible = "regulator-fixed";
+		regulator-name = "usb1-2-vbus";
Quoted 8 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 134

Quoted 5 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +5 -0
+		regulator-min-microvolt = <5000000>;
+		regulator-max-microvolt = <5000000>;
+		gpio = <&r_pio 0 8 GPIO_ACTIVE_HIGH>;	/* PL8 */
+		enable-active-high;
+		regulator-always-on;

Same comment about always on. If the driver needs fixing to grab an additional regulator, fix it, but this shouldn't be left that way.

I agree, but a lot of work is missing to get to that point. To get the USB ports up and running, you need to enable quite a few regulators:
1. VDD09-USB for the controller internals.
2. VCC33-USB for the PHY.
3. Vbus for each connector belonging to the controller.
4. Pin group supply for the GPIO pins controlling each of the Vbus supplies.

Nobody seems to have thought about the first and the last two supplies when the bindings were written. In this case, 1, 2 and 4 will be connected to always-on regulators anyway. Getting support for multiple Vbus supplies probably needs something along the lines of what was discussed with your coupled regulator patches:

https://lists.infradead.org/pipermail/linux-arm-kernel/2016-January/398862.html
https://lists.infradead.org/pipermail/linux-arm-kernel/2016-February/405401.html
https://lists.infradead.org/pipermail/linux-arm-kernel/2016-November/466207.html

Quoted 19 lines
Quoted 19 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +19 -0
+	/* OZ80120 voltage regulator for the four Cortex-A15 cores. */
+	reg_vdd_cpub: regulator-vdd-cpub {
+		compatible = "regulator-gpio";
+
+		regulator-always-on;
+		regulator-min-microvolt = < 800000>;
+		regulator-max-microvolt = <1100000>;
+		regulator-name = "vdd-cpub";
+
+		enable-gpio = <&r_pio 0 2 GPIO_ACTIVE_HIGH>;	/* PL2 */
+		enable-active-high;
+		gpios = <&r_pio 0 3 GPIO_ACTIVE_HIGH>,		/* PL3 */
+			<&r_pio 0 4 GPIO_ACTIVE_HIGH>,		/* PL4 */
+			<&r_pio 0 5 GPIO_ACTIVE_HIGH>;		/* PL5 */
+
+		gpios-states = <1 0 0>;
+		states = <	 750000 0x7
+				 800000 0x3
+				 850000 0x5
Quoted 8 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 160

Quoted 5 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +5 -0
+				 900000 0x1
+				 950000 0x6
+				1000000 0x2
+				1100000 0x4
+				1200000 0x0>;

You're listing a minimum state of 750mv, yet your minimum voltage is 800mV.

Yes. The regulator is capable of outputting the listed voltages in the 750 mV to 1200 mV range, but the permissible range of the consumer is only 800 mV to 1100 mV according to the A80 data sheet. Likewise, the AXP806 regulators support a larger voltage range than that of the connected consumers.

Quoted 3 lines
Quoted 1 line

+/* Ampak AP6335 IEEE 802.11 a/b/g/n/ac "Wifi". */

Why "wifi" ? It's not implementing true wifi?

I put quotes around it because the proper spelling is Wi-Fi, but since that is a trademark, most people use the term Wifi capitalised in one way or another (or sometimes not). Grepping the other dts files, it seems customary to leave out the quotes, so I'll do the same.

Quoted 10 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 195

Quoted 7 lines

+&mmc1 {

arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +6 -0
+	bus-width = <4>;
+	non-removable;
+	pinctrl-names = "default";
+	pinctrl-0 = <&mmc1_pins>;
+	vmmc-supply = <&reg_cldo3>;	/* See cldo2,cldo3 note. */
+	vqmmc-supply = <&reg_aldo2>;

So it's able to support 1.2 or 1.8V IO modes? Surely you want to enable those modes here to.

Thanks for the hint. I thought the driver would automatically detect that. The following modes are supported according to the data sheet:

* DS: Default speed up to 25MHz (3.3V signaling).
* HS: High speed up to 50MHz (3.3V signaling).
* SDR12: SDR up to 25MHz (1.8V signaling).
* SDR25: SDR up to 50MHz (1.8V signaling).
* SDR50: SDR up to 100MHz (1.8V signaling).
* SDR104: SDR up to 208MHz (1.8V signaling).
* DDR50: DDR up to 50MHz (1.8V signaling).

Unfortunately, when I enable the 1.8 V modes, or even just SDR-12, the device fails to initialise with error -110 (-ETIMEDOUT). I will take a look at that, but no promises.

Quoted 11 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 253

Quoted 9 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +9 -0
+	pmic at 745 {
+		compatible = "x-powers,axp808", "x-powers,axp806";
+		reg = <0x745>;
+		interrupt-parent = <&nmi_intc>;
+		interrupts = <0 IRQ_TYPE_LEVEL_LOW>;
+		interrupt-controller;
+		#interrupt-cells = <1>;
+
+		swin-supply = <&reg_dcdce>;

Please use an incude for that PMIC.

I don't understand. What would that include file contain? Very few lines would be shared between .dts files. You could save < 8 lines total if you #include the 5 lines that are common between the Cubieboard4, the Optimus and this board.

Quoted 5 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 257

Quoted 3 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +3 -0
+		/* In comments: Initial setup from vendor sys_config.fex. */
+		regulators {
+			/* 3.0 V (enabled). */

This might be disabled though.

Indeed it is currently disabled by the kernel during startup, as there are no consumers listed in the device tree. I don't know what that regulator supplies, and it is on my to-do list for probing the next time that I have the heat sink off the board.

Quoted 9 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 270

Quoted 6 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +6 -0
+			/* 1.8 V (enabled). */
+			reg_aldo2: aldo2 {
+				regulator-boot-on;
+				regulator-min-microvolt = <1800000>;
+				regulator-max-microvolt = <3600000>;
+				regulator-name = "vcc-pg-pm-wifi+btio-audio";

Usually, there is simpler names available on the schematics, or at least simpler names we can come up with.

I don't have schematics as it's not a developer board. V1.1 of the PCB, which I have, doesn't even have the component names silkscreened onto it.

Signature · 1 line
Rask Ingemann Lambertsen
#15 lore raw

Re: [PATCH v6 5/5] ARM: dts: sun9i: Initial support for the Sunchip CX-A99 board

Quoted 200 lines · Rask Ingemann Lambertsen
Quoted 6 lines · Maxime Ripard

Hi,

Quoted 3 lines · Rask Ingemann Lambertsen

The Suncip CX-A99 board is found in at least four brands of media players. It features an Allwinner A80 ARM SoC and is found in two models:

Quoted 7 lines
Quoted 5 lines

1) 2 GiB DDR3 DRAM and 16 GB eMMC
2) 4 GiB DDR3 DRAM and 32 GB eMMC

For details of the board, see the linux-sunxi page <URL:https://linux-sunxi.org/Sunchip_CX-A99>.

Please don't put URLs in commit logs (and the DT).

OK.

Quoted 13 lines
Quoted 13 lines

Supported features (+ means tested):
+ One Cortex-A7 CPU core (or four with experimental U-Boot PSCI patches)
+ AXP808 power management chip
+ OZ80120 voltage regulator
+ Serial console port (internal)
+ eMMC and SD card slot
+ USB 2.0 host ports on on-board USB hub
+ SATA port on on-board SATA-to-USB bridge *
+ IEEE 802.11 a/b/g/n/ac SDIO Wifi
+ Real-time clock
+ LEDs
- IR receiver for remote control

Quoted 6 lines
Quoted 4 lines

* Only shows up when a SATA device is connected. Also, if a power source is connected to the USB 3.0 connector across power cycles (e.g. FEL boot), the bridge may not properly reset and not show up on the USB bus. The vendor U-Boot performs some unknown magic which resets the bridge.

Is that magic at the USB level, or specific to the bridge itself?

I don't know, but since the other USB port connected to the hub comes up fine, I'm assuming it's something to do with the SATA-to-USB bridge.

Quoted 5 lines
Quoted 5 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +0 -0
new file mode 100644
index 0000000..f5496d2

[...]

Quoted 4 lines
Quoted 1 line

+ /* USB 3.0 standard-A receptacle. For now, only Vbus is supported. */

I'm not sure what you mean by "only VBUS is supported"? Is there any other signal?

I'm just trying to say that at the moment, all the connector will do is to supply power to a connected device (for charging batteries or whatever).

Quoted 9 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 104

Quoted 7 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +7 -0
+   reg_usb0_vbus: regulator-usb0-vbus {
+           compatible = "regulator-fixed";
+           regulator-name = "usb0-vbus";
+           regulator-min-microvolt = <5000000>;
+           regulator-max-microvolt = <5000000>;
+           gpio = <&pio 7 15 GPIO_ACTIVE_HIGH>;    /* PH15 */
+           enable-active-high;

This is redundant with the GPIO flag

No. I have now tested without "enable-active-high" and then Vbus stays off. This is also in line with the binding documentation:

"- enable-active-high: Polarity of GPIO is Active high
If this property is missing, the default assumed is Active low."

It also seems to me that the way this is implemented in Linux, the GPIO polarity flag is ignored.

Quoted 4 lines
Quoted 1 line

+ regulator-always-on;

And it shouldn't be always on. The USB driver will enable it if needs be.

Yes, we just don't have a driver for the A80 USB 3.0 port yet.

Quoted 11 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 117

Quoted 10 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +10 -0
+   /*
+    * A GL850G hub with two external USB connectors is connected
+    * to ehci0. Each has a Vbus regulator controlled by a GPIO:
+    * PL7 for port 1, closest to the 12 V power connector, and
+    * PL8 for port 2, next to the SD card slot.
+    * Because regulator-fixed doesn't support a GPIO list, and
+    * allwinner,sun9i-a80-usb-phy doesn't support more than one
+    * supply, we have to use regulator-always-on on usb1-2-vbus.
+    * Note that the GPIO pins also need cldo1 to be enabled.
+    */
Quoted 2 lines

What is the source of those regulators connected then? Some PMIC regulator? AC-IN?

These two regulators are supplied from a fixed 12 V to 5 V DC-DC converter which can't be controlled in any way.

Quoted 18 lines
Quoted 7 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +7 -0
+   reg_usb1_1_vbus: regulator-usb1-1-vbus {
+           compatible = "regulator-fixed";
+           regulator-name = "usb1-1-vbus";
+           regulator-min-microvolt = <5000000>;
+           regulator-max-microvolt = <5000000>;
+           gpio = <&r_pio 0 7 GPIO_ACTIVE_HIGH>;   /* PL7 */
+           enable-active-high;

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 134

Quoted 10 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +10 -0
+   };
+
+   reg_usb1_2_vbus: regulator-usb1-2-vbus {
+           compatible = "regulator-fixed";
+           regulator-name = "usb1-2-vbus";
+           regulator-min-microvolt = <5000000>;
+           regulator-max-microvolt = <5000000>;
+           gpio = <&r_pio 0 8 GPIO_ACTIVE_HIGH>;   /* PL8 */
+           enable-active-high;
+           regulator-always-on;
Quoted 2 lines

Same comment about always on. If the driver needs fixing to grab an additional regulator, fix it, but this shouldn't be left that way.

I agree, but a lot of work is missing to get to that point. To get the USB ports up and running, you need to enable quite a few regulators:
1. VDD09-USB for the controller internals.
2. VCC33-USB for the PHY.
3. Vbus for each connector belonging to the controller.
4. Pin group supply for the GPIO pins controlling each of the Vbus supplies.

Nobody seems to have thought about the first and the last two supplies when the bindings were written. In this case, 1, 2 and 4 will be connected to always-on regulators anyway. Getting support for multiple Vbus supplies probably needs something along the lines of what was discussed with your coupled regulator patches:

https://lists.infradead.org/pipermail/linux-arm-kernel/2016-January/398862.html
https://lists.infradead.org/pipermail/linux-arm-kernel/2016-February/405401.html
https://lists.infradead.org/pipermail/linux-arm-kernel/2016-November/466207.html

Quoted 19 lines
Quoted 19 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +19 -0
+   /* OZ80120 voltage regulator for the four Cortex-A15 cores. */
+   reg_vdd_cpub: regulator-vdd-cpub {
+           compatible = "regulator-gpio";
+
+           regulator-always-on;
+           regulator-min-microvolt = < 800000>;
+           regulator-max-microvolt = <1100000>;
+           regulator-name = "vdd-cpub";
+
+           enable-gpio = <&r_pio 0 2 GPIO_ACTIVE_HIGH>;    /* PL2 */
+           enable-active-high;
+           gpios = <&r_pio 0 3 GPIO_ACTIVE_HIGH>,          /* PL3 */
+                   <&r_pio 0 4 GPIO_ACTIVE_HIGH>,          /* PL4 */
+                   <&r_pio 0 5 GPIO_ACTIVE_HIGH>;          /* PL5 */
+
+           gpios-states = <1 0 0>;
+           states = <       750000 0x7
+                            800000 0x3
+                            850000 0x5
Quoted 8 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 160

Quoted 5 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +5 -0
+                            900000 0x1
+                            950000 0x6
+                           1000000 0x2
+                           1100000 0x4
+                           1200000 0x0>;

You're listing a minimum state of 750mv, yet your minimum voltage is 800mV.

Yes. The regulator is capable of outputting the listed voltages in the 750 mV to 1200 mV range, but the permissible range of the consumer is only 800 mV to 1100 mV according to the A80 data sheet. Likewise, the AXP806 regulators support a larger voltage range than that of the connected consumers.

Quoted 3 lines
Quoted 1 line

+/* Ampak AP6335 IEEE 802.11 a/b/g/n/ac "Wifi". */

Why "wifi" ? It's not implementing true wifi?

I put quotes around it because the proper spelling is Wi-Fi, but since that is a trademark, most people use the term Wifi capitalised in one way or another (or sometimes not). Grepping the other dts files, it seems customary to leave out the quotes, so I'll do the same.

Quoted 10 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 195

Quoted 7 lines

+&mmc1 {

arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +6 -0
+   bus-width = <4>;
+   non-removable;
+   pinctrl-names = "default";
+   pinctrl-0 = <&mmc1_pins>;
+   vmmc-supply = <&reg_cldo3>;     /* See cldo2,cldo3 note. */
+   vqmmc-supply = <&reg_aldo2>;

So it's able to support 1.2 or 1.8V IO modes? Surely you want to enable those modes here to.

Thanks for the hint. I thought the driver would automatically detect that. The following modes are supported according to the data sheet:

* DS: Default speed up to 25MHz (3.3V signaling).
* HS: High speed up to 50MHz (3.3V signaling).
* SDR12: SDR up to 25MHz (1.8V signaling).
* SDR25: SDR up to 50MHz (1.8V signaling).
* SDR50: SDR up to 100MHz (1.8V signaling).
* SDR104: SDR up to 208MHz (1.8V signaling).
* DDR50: DDR up to 50MHz (1.8V signaling).

Quoted 3 lines

Unfortunately, when I enable the 1.8 V modes, or even just SDR-12, the device fails to initialise with error -110 (-ETIMEDOUT). I will take a look at that, but no promises.

The kernel doesn't support lower I/O voltages for the A80 at the moment. The pin controller has voltage level controls that need to match the external regulators supplying it power.

Rockchip has a similar thing called "Rockchip IO Domain". Look for the ROCKCHIP_IODOMAIN Kconfig symbol. My plan is to do something similar, as a subdevice to the pin controller.

We cannot have it as part of the pin controller as that leads to a circular dependency for the R_PIO, thus blocking device probing for the rest of the system.

However this is somewhat low on my todo list as it works fine without.

Regards
ChenYu

Quoted 44 lines
Quoted 11 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 253

Quoted 10 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +9 -0
+   pmic at 745 {
+           compatible = "x-powers,axp808", "x-powers,axp806";
+           reg = <0x745>;
+           interrupt-parent = <&nmi_intc>;
+           interrupts = <0 IRQ_TYPE_LEVEL_LOW>;
+           interrupt-controller;
+           #interrupt-cells = <1>;
+
+           swin-supply = <&reg_dcdce>;
Quoted 1 line

Please use an incude for that PMIC.

I don't understand. What would that include file contain? Very few lines would be shared between .dts files. You could save < 8 lines total if you #include the 5 lines that are common between the Cubieboard4, the Optimus and this board.

Quoted 5 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 257

Quoted 3 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +3 -0
+           /* In comments: Initial setup from vendor sys_config.fex. */
+           regulators {
+                   /* 3.0 V (enabled). */

This might be disabled though.

Indeed it is currently disabled by the kernel during startup, as there are no consumers listed in the device tree. I don't know what that regulator supplies, and it is on my to-do list for probing the next time that I have the heat sink off the board.

Quoted 9 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 270

Quoted 6 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +6 -0
+                   /* 1.8 V (enabled). */
+                   reg_aldo2: aldo2 {
+                           regulator-boot-on;
+                           regulator-min-microvolt = <1800000>;
+                           regulator-max-microvolt = <3600000>;
+                           regulator-name = "vcc-pg-pm-wifi+btio-audio";

Usually, there is simpler names available on the schematics, or at least simpler names we can come up with.

I don't have schematics as it's not a developer board. V1.1 of the PCB, which I have, doesn't even have the component names silkscreened onto it.

--
Rask Ingemann Lambertsen

#16 lore raw

Re: [PATCH v6 5/5] ARM: dts: sun9i: Initial support for the Sunchip CX-A99 board

In reply to Rask Ingemann Lambertsen
Quoted 18 lines · Rask Ingemann Lambertsen
Quoted 18 lines
Quoted 17 lines

Supported features (+ means tested):
+ One Cortex-A7 CPU core (or four with experimental U-Boot PSCI patches)
+ AXP808 power management chip
+ OZ80120 voltage regulator
+ Serial console port (internal)
+ eMMC and SD card slot
+ USB 2.0 host ports on on-board USB hub
+ SATA port on on-board SATA-to-USB bridge *
+ IEEE 802.11 a/b/g/n/ac SDIO Wifi
+ Real-time clock
+ LEDs
- IR receiver for remote control

* Only shows up when a SATA device is connected. Also, if a power source is connected to the USB 3.0 connector across power cycles (e.g. FEL boot), the bridge may not properly reset and not show up on the USB bus. The vendor U-Boot performs some unknown magic which resets the bridge.

Quoted 4 lines
Quoted 1 line

Is that magic at the USB level, or specific to the bridge itself?

I don't know, but since the other USB port connected to the hub comes up fine, I'm assuming it's something to do with the SATA-to-USB bridge.

But where is that magic written to then? the USB controller registers, or is it an USB packet?

Quoted 6 lines
Quoted 5 lines
Quoted 5 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +0 -0
new file mode 100644
index 0000000..f5496d2

[...]

Quoted 7 lines
Quoted 4 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 97

Quoted 1 line
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +1 -0
+	/* USB 3.0 standard-A receptacle. For now, only Vbus is supported. */

I'm not sure what you mean by "only VBUS is supported"? Is there any other signal?

I'm just trying to say that at the moment, all the connector will do is to supply power to a connected device (for charging batteries or whatever).

So the USB controller itself is non-functionnal then?

Quoted 10 lines
Quoted 9 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 104

Quoted 7 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +7 -0
+	reg_usb0_vbus: regulator-usb0-vbus {
+		compatible = "regulator-fixed";
+		regulator-name = "usb0-vbus";
+		regulator-min-microvolt = <5000000>;
+		regulator-max-microvolt = <5000000>;
+		gpio = <&pio 7 15 GPIO_ACTIVE_HIGH>;	/* PH15 */
+		enable-active-high;

This is redundant with the GPIO flag

Quoted 8 lines

No. I have now tested without "enable-active-high" and then Vbus stays off. This is also in line with the binding documentation:

"- enable-active-high: Polarity of GPIO is Active high
If this property is missing, the default assumed is Active low."

It also seems to me that the way this is implemented in Linux, the GPIO polarity flag is ignored.

It works just fine if the driver uses the gpiod API. If that driver doesn't, then that driver needs some fixing :)

Quoted 6 lines
Quoted 4 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts

Quoted 1 line
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +1 -0
+		regulator-always-on;

And it shouldn't be always on. The USB driver will enable it if needs be.

Yes, we just don't have a driver for the A80 USB 3.0 port yet.

Aaah, so that's what you meant. Leave it out of the DT then.

Quoted 13 lines
Quoted 4 lines
Quoted 4 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +4 -0
+	 * A GL850G hub with two external USB connectors is connected
+	 * to ehci0. Each has a Vbus regulator controlled by a GPIO:
+	 * PL7 for port 1, closest to the 12 V power connector, and
+	 * PL8 for port 2, next to the SD card slot.
Quoted 8 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 117

Quoted 5 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +5 -0
+	 * Because regulator-fixed doesn't support a GPIO list, and
+	 * allwinner,sun9i-a80-usb-phy doesn't support more than one
+	 * supply, we have to use regulator-always-on on usb1-2-vbus.
+	 * Note that the GPIO pins also need cldo1 to be enabled.
+	 */

What is the source of those regulators connected then? Some PMIC regulator? AC-IN?

Quoted 2 lines

These two regulators are supplied from a fixed 12 V to 5 V DC-DC converter which can't be controlled in any way.

Ok.

Quoted 21 lines
Quoted 12 lines
Quoted 12 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +12 -0
+	reg_usb1_1_vbus: regulator-usb1-1-vbus {
+		compatible = "regulator-fixed";
+		regulator-name = "usb1-1-vbus";
+		regulator-min-microvolt = <5000000>;
+		regulator-max-microvolt = <5000000>;
+		gpio = <&r_pio 0 7 GPIO_ACTIVE_HIGH>;	/* PL7 */
+		enable-active-high;
+	};
+
+	reg_usb1_2_vbus: regulator-usb1-2-vbus {
+		compatible = "regulator-fixed";
+		regulator-name = "usb1-2-vbus";
Quoted 8 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 134

Quoted 5 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +5 -0
+		regulator-min-microvolt = <5000000>;
+		regulator-max-microvolt = <5000000>;
+		gpio = <&r_pio 0 8 GPIO_ACTIVE_HIGH>;	/* PL8 */
+		enable-active-high;
+		regulator-always-on;

Same comment about always on. If the driver needs fixing to grab an additional regulator, fix it, but this shouldn't be left that way.

Quoted 3 lines

I agree, but a lot of work is missing to get to that point. To get the USB ports up and running, you need to enable quite a few regulators:
1. VDD09-USB for the controller internals.

This one has never been needed in any SoC before. You should definitely add a regulator property to the USB controller node then, and add a regulator_get to the driver. Ideally based on the compatible.

Quoted 1 line

2. VCC33-USB for the PHY.

This one will be taken by the PHY driver, through a new, optional regulator.

Quoted 1 line

3. Vbus for each connector belonging to the controller.

These should be taken through the usb*_vbus-supply properties

Quoted 2 lines

4. Pin group supply for the GPIO pins controlling each of the Vbus supplies.

And we haven't had any case of a board where those were tied to regulators that were not in themselves essential to the system, which is why no one really bothered to add them. If they have dedicated regulators on your board, then it's time to add those extra properties.

Quoted 9 lines

Nobody seems to have thought about the first and the last two supplies when the bindings were written. In this case, 1, 2 and 4 will be connected to always-on regulators anyway. Getting support for multiple Vbus supplies probably needs something along the lines of what was discussed with your coupled regulator patches:

https://lists.infradead.org/pipermail/linux-arm-kernel/2016-January/398862.html
https://lists.infradead.org/pipermail/linux-arm-kernel/2016-February/405401.html
https://lists.infradead.org/pipermail/linux-arm-kernel/2016-November/466207.html

This is not really the point, always-on is a hack. We can use it if needed to workaround weird cases like this one, but in trivial cases like the one you exposed above, this really should be modelled properly from the very start.

Quoted 28 lines
Quoted 19 lines
Quoted 19 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +19 -0
+	/* OZ80120 voltage regulator for the four Cortex-A15 cores. */
+	reg_vdd_cpub: regulator-vdd-cpub {
+		compatible = "regulator-gpio";
+
+		regulator-always-on;
+		regulator-min-microvolt = < 800000>;
+		regulator-max-microvolt = <1100000>;
+		regulator-name = "vdd-cpub";
+
+		enable-gpio = <&r_pio 0 2 GPIO_ACTIVE_HIGH>;	/* PL2 */
+		enable-active-high;
+		gpios = <&r_pio 0 3 GPIO_ACTIVE_HIGH>,		/* PL3 */
+			<&r_pio 0 4 GPIO_ACTIVE_HIGH>,		/* PL4 */
+			<&r_pio 0 5 GPIO_ACTIVE_HIGH>;		/* PL5 */
+
+		gpios-states = <1 0 0>;
+		states = <	 750000 0x7
+				 800000 0x3
+				 850000 0x5
Quoted 8 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 160

Quoted 5 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +5 -0
+				 900000 0x1
+				 950000 0x6
+				1000000 0x2
+				1100000 0x4
+				1200000 0x0>;

You're listing a minimum state of 750mv, yet your minimum voltage is 800mV.

Quoted 5 lines

Yes. The regulator is capable of outputting the listed voltages in the 750 mV to 1200 mV range, but the permissible range of the consumer is only 800 mV to 1100 mV according to the A80 data sheet. Likewise, the AXP806 regulators support a larger voltage range than that of the connected consumers.

Adding a comment stating that would be nice then.

Quoted 22 lines
Quoted 10 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 195

Quoted 7 lines

+&mmc1 {

arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +6 -0
+	bus-width = <4>;
+	non-removable;
+	pinctrl-names = "default";
+	pinctrl-0 = <&mmc1_pins>;
+	vmmc-supply = <&reg_cldo3>;	/* See cldo2,cldo3 note. */
+	vqmmc-supply = <&reg_aldo2>;

So it's able to support 1.2 or 1.8V IO modes? Surely you want to enable those modes here to.

Thanks for the hint. I thought the driver would automatically detect that. The following modes are supported according to the data sheet:

* DS: Default speed up to 25MHz (3.3V signaling).
* HS: High speed up to 50MHz (3.3V signaling).
* SDR12: SDR up to 25MHz (1.8V signaling).
* SDR25: SDR up to 50MHz (1.8V signaling).
* SDR50: SDR up to 100MHz (1.8V signaling).
* SDR104: SDR up to 208MHz (1.8V signaling).
* DDR50: DDR up to 50MHz (1.8V signaling).

Quoted 3 lines

Unfortunately, when I enable the 1.8 V modes, or even just SDR-12, the device fails to initialise with error -110 (-ETIMEDOUT). I will take a look at that, but no promises.

Ok.

Quoted 12 lines
Quoted 11 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 253

Quoted 9 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +9 -0
+	pmic at 745 {
+		compatible = "x-powers,axp808", "x-powers,axp806";
+		reg = <0x745>;
+		interrupt-parent = <&nmi_intc>;
+		interrupts = <0 IRQ_TYPE_LEVEL_LOW>;
+		interrupt-controller;
+		#interrupt-cells = <1>;
+
+		swin-supply = <&reg_dcdce>;

Please use an incude for that PMIC.

Quoted 4 lines

I don't understand. What would that include file contain? Very few lines would be shared between .dts files. You could save < 8 lines total if you #include the 5 lines that are common between the Cubieboard4, the Optimus and this board.

You also have to take into account that the PMIC is barely supported at the moment, you'll also need to define the GPIO controllers, the various power supplies, adc, etc. nodes that are going (hopefully to be added) in the future.

If the include is shared, once it is enabled, every one benefits from it. Otherwise, you have to duplicate that change across all DTs. This is much likely to error, harder to maintain, and you end up in a situation where none of the boards really support the same feature set while being based on the same SoC / PMIC combination.

Quoted 10 lines
Quoted 5 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 257

Quoted 3 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +3 -0
+		/* In comments: Initial setup from vendor sys_config.fex. */
+		regulators {
+			/* 3.0 V (enabled). */

This might be disabled though.

Indeed it is currently disabled by the kernel during startup, as there are no consumers listed in the device tree. I don't know what that regulator supplies, and it is on my to-do list for probing the next time that I have the heat sink off the board.

I don't think those comments are needed actually. some are misleading, like the one above, and some are likely to change at some point, like the one below.

Quoted 10 lines
Quoted 9 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 270

Quoted 6 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +6 -0
+			/* 1.8 V (enabled). */
+			reg_aldo2: aldo2 {
+				regulator-boot-on;
+				regulator-min-microvolt = <1800000>;
+				regulator-max-microvolt = <3600000>;
+				regulator-name = "vcc-pg-pm-wifi+btio-audio";

Usually, there is simpler names available on the schematics, or at least simpler names we can come up with.

Quoted 2 lines

I don't have schematics as it's not a developer board. V1.1 of the PCB, which I have, doesn't even have the component names silkscreened onto it.

Ok.

Thanks,
Maxime

Signature · 10 lines
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 801 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20170216/7a6ae092/attachment-0001.sig>
#10 lore raw

Re: [PATCH v6 5/5] ARM: dts: sun9i: Initial support for the Sunchip CX-A99 board

In reply to Rask Ingemann Lambertsen
Quoted 26 lines · Rask Ingemann Lambertsen
Quoted 5 lines · Chen-Yu Tsai
Quoted 3 lines · Maxime Ripard

Hi,

On Thu, Feb 09, 2017 at 12:34:06AM +0100, Rask Ingemann Lambertsen wrote:

[...]

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +0 -0
new file mode 100644
index 0000000..f5496d2

[...]

Quoted 5 lines
Quoted 2 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 246

Quoted 2 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +2 -0
+     pmic at 745 {
+             compatible = "x-powers,axp808", "x-powers,axp806";

As you mentioned elsewhere, they are not really compatible. You should drop the latter compatible.

The compatibility is good enough that it works fine with the driver that went into kernel 4.9 and 4.10. Using this device tree file, I have built and booted 4.9.9 and 4.10-rc7 kernels as per the instructions here;
https://linux-sunxi.org/Sunchip_CX-A99#Linux_kernel
Likewise any other AXP806 driver which doesn't touch the register at address 0xff will work fine.

Quoted 3 lines

So, the disadvantage of removing the "x-powers,axp806" compatible is that it breaks on two kernel versions where it works fine with the compatible. What is the advantage of removing the "x-powers,axp806" compatible?

This works fine for the features you tested, which are the regulators. All the other features of the PMIC are untested, and might or might not change, so you really don't know about the overall capability of the PMIC with another model.

Maxime

Signature · 10 lines
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 801 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20170216/e378788b/attachment-0001.sig>
#11 lore raw

AXP808 vs. AXP806 debugged, no difference? (Was: [PATCH v6 5/5] ARM: dts: sun9i: Initial support for the Sunchip CX-A99 board)

In reply to Rask Ingemann Lambertsen

In summary: I'll propose a new property "extended-address" or so to the existing "x-powers,axp806" compatible, because the axp808 does seem to have the register at 0xff after all, it just needs a value of 0 instead of the 16 which drivers/mfd/axp20x.c as of recently is now writing unconditionally to 0xff. A new "x-powers,axp808" compatible and new AXP808_ID is therefore not needed. This should make for a simpler and smaller patch.

Quoted 9 lines · Rask Ingemann Lambertsen

As to touching the register at 0xff, it would be interesting to know more precisely what goes wrong. The symptom is these messages in the log:

[    3.209573] axp20x-rsb sunxi-rsb-745: AXP20x variant AXP806 found
[    3.210153] axp20x-rsb sunxi-rsb-745: Failed to set masks in 0x40: -5
[    3.210178] axp20x-rsb sunxi-rsb-745: failed to add irq chip: -5
[    3.210306] axp20x-rsb: probe of sunxi-rsb-745 failed with error -5

I haven't yet looked further into what happens at the RSB bus level.

With the patch below applied for debugging purposes and using the "xpowers,axp806" compatible, I get:

[    3.209955] axp20x-rsb sunxi-rsb-745: AXP20x variant AXP806 found
[    3.210065] axp20x-rsb sunxi-rsb-745: Chip ID read 1 = 0x60
[    3.210119] axp20x-rsb sunxi-rsb-745: Reg addr ext = 0x0
[    3.210171] axp20x-rsb sunxi-rsb-745: Bus addr ext = 0x0
[    3.210381] axp20x-rsb sunxi-rsb-745: Chip ID read 2 = 0x60
(16 written to register 0xff here)
[    3.210570] sunxi-rsb 8003400.i2c: RSB transfer data error
[    3.210596] sunxi-rsb 8003400.i2c: RSB error reading dev 58 register 3 length 1.
[    3.210625] axp20x-rsb sunxi-rsb-745: Chip ID read 3 = 0xffffffff
(0 written to register 0xff here)
[    3.210828] axp20x-rsb sunxi-rsb-745: Chip ID read 4 = 0x60

I.e. register 0xff (Reg addr ext) and 0xfe (Bus addr ext) exist with a default of 0. After writing 16 to register 0xff, the chip ID (register 3) can not be read, but writing 0 to register 0xff makes the AXP808 respond again and the board boots up as normal.

Also, can someone with an AXP806 confirm that it does use the same chip ID of 0x60 as the AXP808?

drivers/mfd/axp20x.c +31 -0
index 125b470..db557f3 100644
@@ -910,9 +910,40 @@ int axp20x_device_probe(struct axp20x_dev *axp20x)
 	 * address.
 	 */
 	if (axp20x->variant == AXP806_ID)
+	{
+		int dummy;
+
+		dummy = -1;
+		regmap_read(axp20x->regmap, AXP806_CHIP_ID, &dummy);
+		dev_info(axp20x->dev, "Chip ID read 1 = 0x%x\n", dummy);
+
+		regmap_read(axp20x->regmap, AXP806_REG_ADDR_EXT, &dummy);
+		dev_info(axp20x->dev, "Reg addr ext = 0x%x\n", dummy);
+		regmap_read(axp20x->regmap, AXP806_BUS_ADDR_EXT, &dummy);
+		dev_info(axp20x->dev, "Bus addr ext = 0x%x\n", dummy);
+
+		regmap_reinit_cache(axp20x->regmap, axp20x->regmap_cfg);
+		dummy = -1;
+		regmap_read(axp20x->regmap, AXP806_CHIP_ID, &dummy);
+		dev_info(axp20x->dev, "Chip ID read 2 = 0x%x\n", dummy);
+
 		regmap_write(axp20x->regmap, AXP806_REG_ADDR_EXT,
 			     AXP806_REG_ADDR_EXT_ADDR_SLAVE_MODE);
 
+		regmap_reinit_cache(axp20x->regmap, axp20x->regmap_cfg);
+		dummy = -1;
+		regmap_read(axp20x->regmap, AXP806_CHIP_ID, &dummy);
+		dev_info(axp20x->dev, "Chip ID read 3 = 0x%x\n", dummy);
+
+		regmap_write(axp20x->regmap, AXP806_REG_ADDR_EXT,
+			     0);
+
+		regmap_reinit_cache(axp20x->regmap, axp20x->regmap_cfg);
+		dummy = -1;
+		regmap_read(axp20x->regmap, AXP806_CHIP_ID, &dummy);
+		dev_info(axp20x->dev, "Chip ID read 4 = 0x%x\n", dummy);
+	}
+
 	ret = regmap_add_irq_chip(axp20x->regmap, axp20x->irq,
 			  IRQF_ONESHOT | IRQF_SHARED | axp20x->irq_flags,
 			   -1, axp20x->regmap_irq_chip, &axp20x->regmap_irqc);
drivers/bus/sunxi-rsb.c +8 -0
index 795c9d9..85e4926 100644
@@ -349,6 +349,10 @@ static int sunxi_rsb_read(struct sunxi_rsb *rsb, u8 rtaddr, u8 addr,
 unlock:
 	mutex_unlock(&rsb->lock);
 
+	if (ret)
+		dev_dbg(rsb->dev, "RSB error reading dev %u register %u length %u.\n",
+			rtaddr, addr, (unsigned int) len);
+
 	return ret;
 }
 
@@ -386,6 +390,10 @@ static int sunxi_rsb_write(struct sunxi_rsb *rsb, u8 rtaddr, u8 addr,
 
 	mutex_unlock(&rsb->lock);
 
+	if (ret)
+		dev_dbg(rsb->dev, "RSB error writing dev %u register %u length %u.\n",
+			rtaddr, addr, (unsigned int) len);
+
 	return ret;
 }
 
drivers/bus/Makefile +1 -0
index cc6364b..4b81226 100644
@@ -24,3 +24,4 @@ obj-$(CONFIG_UNIPHIER_SYSTEM_BUS)	+= uniphier-system-bus.o
 obj-$(CONFIG_VEXPRESS_CONFIG)	+= vexpress-config.o
 
 obj-$(CONFIG_DA8XX_MSTPRI)	+= da8xx-mstpri.o
+CFLAGS_sunxi-rsb.o		+= -DDEBUG -DEBUG
Signature · 1 line
Rask Ingemann Lambertsen
#12 lore raw

Re: AXP808 vs. AXP806 debugged, no difference? (Was: [PATCH v6 5/5] ARM: dts: sun9i: Initial support for the Sunchip CX-A99 board)

Quoted 6 lines · Rask Ingemann Lambertsen

In summary: I'll propose a new property "extended-address" or so to the existing "x-powers,axp806" compatible, because the axp808 does seem to have the register at 0xff after all, it just needs a value of 0 instead of the 16 which drivers/mfd/axp20x.c as of recently is now writing unconditionally to 0xff. A new "x-powers,axp808" compatible and new AXP808_ID is therefore not needed. This should make for a simpler and smaller patch.

It might be the same core as the AXP806, though without a datasheet to actually confirm it, it might be best to stick to a new compatible.

However, we could add a new property "x-powers,master-mode". If the driver sees it, it would clear bit 4, instead of setting it. The hardware bit is latched from the external pin, so this would be per board.

Regards
ChenYu

Quoted 117 lines · Rask Ingemann Lambertsen
Quoted 9 lines

As to touching the register at 0xff, it would be interesting to know more precisely what goes wrong. The symptom is these messages in the log:

[    3.209573] axp20x-rsb sunxi-rsb-745: AXP20x variant AXP806 found
[    3.210153] axp20x-rsb sunxi-rsb-745: Failed to set masks in 0x40: -5
[    3.210178] axp20x-rsb sunxi-rsb-745: failed to add irq chip: -5
[    3.210306] axp20x-rsb: probe of sunxi-rsb-745 failed with error -5

I haven't yet looked further into what happens at the RSB bus level.

With the patch below applied for debugging purposes and using the "xpowers,axp806" compatible, I get:

[    3.209955] axp20x-rsb sunxi-rsb-745: AXP20x variant AXP806 found
[    3.210065] axp20x-rsb sunxi-rsb-745: Chip ID read 1 = 0x60
[    3.210119] axp20x-rsb sunxi-rsb-745: Reg addr ext = 0x0
[    3.210171] axp20x-rsb sunxi-rsb-745: Bus addr ext = 0x0
[    3.210381] axp20x-rsb sunxi-rsb-745: Chip ID read 2 = 0x60
(16 written to register 0xff here)
[    3.210570] sunxi-rsb 8003400.i2c: RSB transfer data error
[    3.210596] sunxi-rsb 8003400.i2c: RSB error reading dev 58 register 3 length 1.
[    3.210625] axp20x-rsb sunxi-rsb-745: Chip ID read 3 = 0xffffffff
(0 written to register 0xff here)
[    3.210828] axp20x-rsb sunxi-rsb-745: Chip ID read 4 = 0x60

I.e. register 0xff (Reg addr ext) and 0xfe (Bus addr ext) exist with a default of 0. After writing 16 to register 0xff, the chip ID (register 3) can not be read, but writing 0 to register 0xff makes the AXP808 respond again and the board boots up as normal.

Also, can someone with an AXP806 confirm that it does use the same chip ID of 0x60 as the AXP808?

drivers/mfd/axp20x.c +31 -0
index 125b470..db557f3 100644
@@ -910,9 +910,40 @@ int axp20x_device_probe(struct axp20x_dev *axp20x)
         * address.
         */
        if (axp20x->variant == AXP806_ID)
+       {
+               int dummy;
+
+               dummy = -1;
+               regmap_read(axp20x->regmap, AXP806_CHIP_ID, &dummy);
+               dev_info(axp20x->dev, "Chip ID read 1 = 0x%x\n", dummy);
+
+               regmap_read(axp20x->regmap, AXP806_REG_ADDR_EXT, &dummy);
+               dev_info(axp20x->dev, "Reg addr ext = 0x%x\n", dummy);
+               regmap_read(axp20x->regmap, AXP806_BUS_ADDR_EXT, &dummy);
+               dev_info(axp20x->dev, "Bus addr ext = 0x%x\n", dummy);
+
+               regmap_reinit_cache(axp20x->regmap, axp20x->regmap_cfg);
+               dummy = -1;
+               regmap_read(axp20x->regmap, AXP806_CHIP_ID, &dummy);
+               dev_info(axp20x->dev, "Chip ID read 2 = 0x%x\n", dummy);
+
                regmap_write(axp20x->regmap, AXP806_REG_ADDR_EXT,
                             AXP806_REG_ADDR_EXT_ADDR_SLAVE_MODE);
 
+               regmap_reinit_cache(axp20x->regmap, axp20x->regmap_cfg);
+               dummy = -1;
+               regmap_read(axp20x->regmap, AXP806_CHIP_ID, &dummy);
+               dev_info(axp20x->dev, "Chip ID read 3 = 0x%x\n", dummy);
+
+               regmap_write(axp20x->regmap, AXP806_REG_ADDR_EXT,
+                            0);
+
+               regmap_reinit_cache(axp20x->regmap, axp20x->regmap_cfg);
+               dummy = -1;
+               regmap_read(axp20x->regmap, AXP806_CHIP_ID, &dummy);
+               dev_info(axp20x->dev, "Chip ID read 4 = 0x%x\n", dummy);
+       }
+
        ret = regmap_add_irq_chip(axp20x->regmap, axp20x->irq,
                          IRQF_ONESHOT | IRQF_SHARED | axp20x->irq_flags,
                           -1, axp20x->regmap_irq_chip, &axp20x->regmap_irqc);
drivers/bus/sunxi-rsb.c +8 -0
index 795c9d9..85e4926 100644
@@ -349,6 +349,10 @@ static int sunxi_rsb_read(struct sunxi_rsb *rsb, u8 rtaddr, u8 addr,
 unlock:
        mutex_unlock(&rsb->lock);
 
+       if (ret)
+               dev_dbg(rsb->dev, "RSB error reading dev %u register %u length %u.\n",
+                       rtaddr, addr, (unsigned int) len);
+
        return ret;
 }
 
@@ -386,6 +390,10 @@ static int sunxi_rsb_write(struct sunxi_rsb *rsb, u8 rtaddr, u8 addr,
 
        mutex_unlock(&rsb->lock);
 
+       if (ret)
+               dev_dbg(rsb->dev, "RSB error writing dev %u register %u length %u.\n",
+                       rtaddr, addr, (unsigned int) len);
+
        return ret;
 }
 
drivers/bus/Makefile +1 -1
index cc6364b..4b81226 100644
@@ -24,3 +24,4 @@ obj-$(CONFIG_UNIPHIER_SYSTEM_BUS)     += uniphier-system-bus.o
 obj-$(CONFIG_VEXPRESS_CONFIG)  += vexpress-config.o
 
 obj-$(CONFIG_DA8XX_MSTPRI)     += da8xx-mstpri.o
+CFLAGS_sunxi-rsb.o             += -DDEBUG -DEBUG
 
--

Rask Ingemann Lambertsen

#13 lore raw

Re: AXP808 vs. AXP806 debugged, no difference? (Was: [PATCH v6 5/5] ARM: dts: sun9i: Initial support for the Sunchip CX-A99 board)

Quoted 11 lines · Chen-Yu Tsai
Quoted 6 lines · Rask Ingemann Lambertsen

In summary: I'll propose a new property "extended-address" or so to the existing "x-powers,axp806" compatible, because the axp808 does seem to have the register at 0xff after all, it just needs a value of 0 instead of the 16 which drivers/mfd/axp20x.c as of recently is now writing unconditionally to 0xff. A new "x-powers,axp808" compatible and new AXP808_ID is therefore not needed. This should make for a simpler and smaller patch.

It might be the same core as the AXP806, though without a datasheet to actually confirm it, it might be best to stick to a new compatible.

Yes, an extra compatible is cheap. Also, it occurs to me that either could turn out to have features that the other one doesn't.

FWIW, both are 56-pin devices going by board photos. I'd say they differ in the power-on voltages. On the AXP808, I've measured (in FEL mode, assuming that the BROM doesn't have any PMIC code at all):

aldo1 3.0 V
aldo2 off
aldo3 off
bldo1 1.8 V
bldo2 0.9 V (here the AXP806 gives 1.8 V for DLL and PLL)
bldo3 off
bldo4 ? (haven't found it yet)
cldo1 3.3 V
cldo2 off
cldo3 off
dcdca 0.9 V
dcdcb 1.5 V
dcdcc 0.9 V
dcdcd 0.9 V
dcdce 3.3 V (AXP806 set to 2.1 V in Optimus and Cubieboard4 dts)
sw0 ? (haven't found it yet)

It would also explain why they were given the same chip ID, as from driver point of view, they behave in the same way. The Allwinner kernel tree also uses the same driver (confusingly called axp15) for both.

Quoted 3 lines

However, we could add a new property "x-powers,master-mode". If the driver sees it, it would clear bit 4, instead of setting it. The hardware bit is latched from the external pin, so this would be per board.

Patch coming.

Signature · 1 line
Rask Ingemann Lambertsen
#17 lore raw

Re: [PATCH v6 5/5] ARM: dts: sun9i: Initial support for the Sunchip CX-A99 board

In reply to Maxime Ripard
Quoted 2 lines · Maxime Ripard
Quoted 1 line · Rask Ingemann Lambertsen
Quoted 1 line
Quoted 1 line

Supported features (+ means tested):

[...]

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

+ SATA port on on-board SATA-to-USB bridge *

[...]

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 4 lines

* Only shows up when a SATA device is connected. Also, if a power source is connected to the USB 3.0 connector across power cycles (e.g. FEL boot), the bridge may not properly reset and not show up on the USB bus. The vendor U-Boot performs some unknown magic which resets the bridge.

Quoted 7 lines
Quoted 4 lines
Quoted 1 line

Is that magic at the USB level, or specific to the bridge itself?

I don't know, but since the other USB port connected to the hub comes up fine, I'm assuming it's something to do with the SATA-to-USB bridge.

But where is that magic written to then? the USB controller registers, or is it an USB packet?

I don't know. Like I said, unknown magic. I haven't found the source code to the U-Boot version that the board uses.

For all I know, it could be a bug in the Linux USB stack. I've always had to unplug and replug my USB keyboards after Linux has booted, otherwise they are not detected. If the keyboard was connected to a hub, Linux doesn't detect that port on the hub, so the hub must be unplugged and replugged. If it's an internal hub, tough luck.

Quoted 9 lines
Quoted 9 lines
Quoted 9 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 104

Quoted 7 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +7 -0
+	reg_usb0_vbus: regulator-usb0-vbus {
+		compatible = "regulator-fixed";
+		regulator-name = "usb0-vbus";
+		regulator-min-microvolt = <5000000>;
+		regulator-max-microvolt = <5000000>;
+		gpio = <&pio 7 15 GPIO_ACTIVE_HIGH>;	/* PH15 */
+		enable-active-high;

This is redundant with the GPIO flag

The GPIO flag is overridden by the regulator core, which makes it misleading. I will remove the GPIO flag and add a comment there that GPIO flags are not supported. This is to prevent someone from thinking that

gpio = <... GPIO_ACTIVE_HIGH>;
enable-active-high;

and

gpio = <... GPIO_ACTIVE_LOW>;

are equivalent. It would be an easy mistake to make, because in mmc nodes,

cd-gpios = <... GPIO_ACTIVE_HIGH>;

and

cd-gpios = <... GPIO_ACTIVE_LOW>;
cd-inverted;

_are_ equivalent.

Quoted 11 lines
Quoted 8 lines

No. I have now tested without "enable-active-high" and then Vbus stays off. This is also in line with the binding documentation:

"- enable-active-high: Polarity of GPIO is Active high
If this property is missing, the default assumed is Active low."

It also seems to me that the way this is implemented in Linux, the GPIO polarity flag is ignored.

It works just fine if the driver uses the gpiod API. If that driver doesn't, then that driver needs some fixing :)

Take it up with the maintainer of the regulator core, then. :)

Quoted 1 line

From of_get_fixed_voltage_config() in drivers/regulator/fixed.c:

[...]

	config->gpio = of_get_named_gpio(np, "gpio", 0);
[...]
	config->enable_high = of_property_read_bool(np, "enable-active-high");
	config->gpio_is_open_drain = of_property_read_bool(np,
							   "gpio-open-drain");

[...]

Quoted 1 line

From reg_fixed_voltage_probe() in drivers/regulator/fixed.c:

[...]

		config = of_get_fixed_voltage_config(&pdev->dev,
						     &drvdata->desc);
[...]
	if (gpio_is_valid(config->gpio)) {
		cfg.ena_gpio = config->gpio;
		if (pdev->dev.of_node)
			cfg.ena_gpio_initialized = true;
	}
	cfg.ena_gpio_invert = !config->enable_high;
	if (config->enabled_at_boot) {
		if (config->enable_high)
			cfg.ena_gpio_flags |= GPIOF_OUT_INIT_HIGH;
		else
			cfg.ena_gpio_flags |= GPIOF_OUT_INIT_LOW;
	} else {
		if (config->enable_high)
			cfg.ena_gpio_flags |= GPIOF_OUT_INIT_LOW;
		else
			cfg.ena_gpio_flags |= GPIOF_OUT_INIT_HIGH;
	}
	if (config->gpio_is_open_drain)
		cfg.ena_gpio_flags |= GPIOF_OPEN_DRAIN;

[...]

Quoted 1 line

From regulator_ena_gpio_request() in drivers/regulator/core.c:

[...]

	gpiod = gpio_to_desc(config->ena_gpio);
[...]
	ret = gpio_request_one(config->ena_gpio,
				GPIOF_DIR_OUT | config->ena_gpio_flags,
				rdev_get_name(rdev));
[...]
	pin->gpiod = gpiod;
	pin->ena_gpio_invert = config->ena_gpio_invert;

And finally, gpio_request_one() in drivers/gpio/gpiolib-legacy.c:

int gpio_request_one(unsigned gpio, unsigned long flags, const char *label)

{
	struct gpio_desc *desc;
	int err;

	desc = gpio_to_desc(gpio);
[...]
	if (flags & GPIOF_OPEN_DRAIN)
		set_bit(FLAG_OPEN_DRAIN, &desc->flags);

	if (flags & GPIOF_OPEN_SOURCE)
		set_bit(FLAG_OPEN_SOURCE, &desc->flags);

	if (flags & GPIOF_ACTIVE_LOW)
		set_bit(FLAG_ACTIVE_LOW, &desc->flags);

	if (flags & GPIOF_DIR_IN)
		err = gpiod_direction_input(desc);
	else
		err = gpiod_direction_output_raw(desc,
				(flags & GPIOF_INIT_HIGH) ? 1 : 0);

[...]

Quoted 8 lines
Quoted 6 lines
Quoted 4 lines

Review

Quoted 1 line
(quoted diff) +1 -0
+		regulator-always-on;

And it shouldn't be always on. The USB driver will enable it if needs be.

Yes, we just don't have a driver for the A80 USB 3.0 port yet.

Aaah, so that's what you meant. Leave it out of the DT then.

OK.

Quoted 10 lines
Quoted 8 lines
Quoted 2 lines

You're listing a minimum state of 750mv, yet your minimum voltage is 800mV.

Yes. The regulator is capable of outputting the listed voltages in the 750 mV to 1200 mV range, but the permissible range of the consumer is only 800 mV to 1100 mV according to the A80 data sheet. Likewise, the AXP806 regulators support a larger voltage range than that of the connected consumers.

Adding a comment stating that would be nice then.

OK.

Quoted 22 lines
Quoted 12 lines
Quoted 11 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 253

Quoted 9 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +9 -0
+	pmic at 745 {
+		compatible = "x-powers,axp808", "x-powers,axp806";
+		reg = <0x745>;
+		interrupt-parent = <&nmi_intc>;
+		interrupts = <0 IRQ_TYPE_LEVEL_LOW>;
+		interrupt-controller;
+		#interrupt-cells = <1>;
+
+		swin-supply = <&reg_dcdce>;

Please use an incude for that PMIC.

Quoted 4 lines

I don't understand. What would that include file contain? Very few lines would be shared between .dts files. You could save < 8 lines total if you #include the 5 lines that are common between the Cubieboard4, the Optimus and this board.

You also have to take into account that the PMIC is barely supported at the moment, you'll also need to define the GPIO controllers, the various power supplies, adc, etc. nodes that are going (hopefully to be added) in the future.

Quoted 5 lines

If the include is shared, once it is enabled, every one benefits from it. Otherwise, you have to duplicate that change across all DTs. This is much likely to error, harder to maintain, and you end up in a situation where none of the boards really support the same feature set while being based on the same SoC / PMIC combination.

Fair enough, I'll add an include file the next time around. Like you mentioned youself, though, we will only really know how compatible the two PMICs are when we have the code to support the other features (or a data sheet/user's manual is found). I'm hopeful based on what has been found so far.

Signature · 1 line
Rask Ingemann Lambertsen
#18 lore raw

Re: [PATCH v6 5/5] ARM: dts: sun9i: Initial support for the Sunchip CX-A99 board

Quoted 13 lines · Rask Ingemann Lambertsen
Quoted 2 lines · Maxime Ripard
Quoted 1 line · Rask Ingemann Lambertsen
Quoted 1 line
Quoted 1 line

Supported features (+ means tested):

[...]

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

+ SATA port on on-board SATA-to-USB bridge *

[...]

Quoted 7 lines
Quoted 7 lines
Quoted 6 lines
Quoted 4 lines

* Only shows up when a SATA device is connected. Also, if a power source is connected to the USB 3.0 connector across power cycles (e.g. FEL boot), the bridge may not properly reset and not show up on the USB bus. The vendor U-Boot performs some unknown magic which resets the bridge.

Is that magic at the USB level, or specific to the bridge itself?

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

I don't know, but since the other USB port connected to the hub comes up fine, I'm assuming it's something to do with the SATA-to-USB bridge.

But where is that magic written to then? the USB controller registers, or is it an USB packet?

I don't know. Like I said, unknown magic. I haven't found the source code to the U-Boot version that the board uses.

Then how do you know there's a magic in the first place?

Quoted 5 lines

For all I know, it could be a bug in the Linux USB stack. I've always had to unplug and replug my USB keyboards after Linux has booted, otherwise they are not detected. If the keyboard was connected to a hub, Linux doesn't detect that port on the hub, so the hub must be unplugged and replugged. If it's an internal hub, tough luck.

I have a laptop that certainly doesn't show that behaviour, so a bug in the USB stack seems weird.

Quoted 23 lines
Quoted 9 lines
Quoted 9 lines
Quoted 9 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 104

Quoted 7 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +7 -0
+	reg_usb0_vbus: regulator-usb0-vbus {
+		compatible = "regulator-fixed";
+		regulator-name = "usb0-vbus";
+		regulator-min-microvolt = <5000000>;
+		regulator-max-microvolt = <5000000>;
+		gpio = <&pio 7 15 GPIO_ACTIVE_HIGH>;	/* PH15 */
+		enable-active-high;

This is redundant with the GPIO flag

The GPIO flag is overridden by the regulator core, which makes it misleading. I will remove the GPIO flag and add a comment there that GPIO flags are not supported. This is to prevent someone from thinking that

gpio = <... GPIO_ACTIVE_HIGH>;
enable-active-high;

and

gpio = <... GPIO_ACTIVE_LOW>;

are equivalent. It would be an easy mistake to make, because in mmc nodes,

Quoted 8 lines

cd-gpios = <... GPIO_ACTIVE_HIGH>;

and

cd-gpios = <... GPIO_ACTIVE_LOW>;
cd-inverted;

_are_ equivalent.

And your point is? The gpio flags are now properly passed through to their respective consumers, which was not the case before. Everything works, without the workaround that we had to use before. Why would you want to use anything else than the property that just works without any modification and with the minimal amount of "code" ?

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines

No. I have now tested without "enable-active-high" and then Vbus stays off. This is also in line with the binding documentation:

"- enable-active-high: Polarity of GPIO is Active high
If this property is missing, the default assumed is Active low."

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

It also seems to me that the way this is implemented in Linux, the GPIO polarity flag is ignored.

It works just fine if the driver uses the gpiod API. If that driver doesn't, then that driver needs some fixing :)

Take it up with the maintainer of the regulator core, then. :)

Or you can also submit a patch fixing it.

Quoted 123 lines

From of_get_fixed_voltage_config() in drivers/regulator/fixed.c:
[...]

	config->gpio = of_get_named_gpio(np, "gpio", 0);
[...]
	config->enable_high = of_property_read_bool(np, "enable-active-high");
	config->gpio_is_open_drain = of_property_read_bool(np,
							   "gpio-open-drain");

[...]

From reg_fixed_voltage_probe() in drivers/regulator/fixed.c:
[...]

		config = of_get_fixed_voltage_config(&pdev->dev,
						     &drvdata->desc);
[...]
	if (gpio_is_valid(config->gpio)) {
		cfg.ena_gpio = config->gpio;
		if (pdev->dev.of_node)
			cfg.ena_gpio_initialized = true;
	}
	cfg.ena_gpio_invert = !config->enable_high;
	if (config->enabled_at_boot) {
		if (config->enable_high)
			cfg.ena_gpio_flags |= GPIOF_OUT_INIT_HIGH;
		else
			cfg.ena_gpio_flags |= GPIOF_OUT_INIT_LOW;
	} else {
		if (config->enable_high)
			cfg.ena_gpio_flags |= GPIOF_OUT_INIT_LOW;
		else
			cfg.ena_gpio_flags |= GPIOF_OUT_INIT_HIGH;
	}
	if (config->gpio_is_open_drain)
		cfg.ena_gpio_flags |= GPIOF_OPEN_DRAIN;

[...]

From regulator_ena_gpio_request() in drivers/regulator/core.c:
[...]

	gpiod = gpio_to_desc(config->ena_gpio);
[...]
	ret = gpio_request_one(config->ena_gpio,
				GPIOF_DIR_OUT | config->ena_gpio_flags,
				rdev_get_name(rdev));
[...]
	pin->gpiod = gpiod;
	pin->ena_gpio_invert = config->ena_gpio_invert;

And finally, gpio_request_one() in drivers/gpio/gpiolib-legacy.c:

int gpio_request_one(unsigned gpio, unsigned long flags, const char *label)

{
	struct gpio_desc *desc;
	int err;

	desc = gpio_to_desc(gpio);
[...]
	if (flags & GPIOF_OPEN_DRAIN)
		set_bit(FLAG_OPEN_DRAIN, &desc->flags);

	if (flags & GPIOF_OPEN_SOURCE)
		set_bit(FLAG_OPEN_SOURCE, &desc->flags);

	if (flags & GPIOF_ACTIVE_LOW)
		set_bit(FLAG_ACTIVE_LOW, &desc->flags);

	if (flags & GPIOF_DIR_IN)
		err = gpiod_direction_input(desc);
	else
		err = gpiod_direction_output_raw(desc,
				(flags & GPIOF_INIT_HIGH) ? 1 : 0);

[...]

Quoted 8 lines
Quoted 6 lines
Quoted 4 lines

Review

Quoted 1 line
(quoted diff) +1 -0
+		regulator-always-on;

And it shouldn't be always on. The USB driver will enable it if needs be.

Yes, we just don't have a driver for the A80 USB 3.0 port yet.

Aaah, so that's what you meant. Leave it out of the DT then.

OK.

Quoted 10 lines
Quoted 8 lines
Quoted 2 lines

You're listing a minimum state of 750mv, yet your minimum voltage is 800mV.

Yes. The regulator is capable of outputting the listed voltages in the 750 mV to 1200 mV range, but the permissible range of the consumer is only 800 mV to 1100 mV according to the A80 data sheet. Likewise, the AXP806 regulators support a larger voltage range than that of the connected consumers.

Adding a comment stating that would be nice then.

OK.

Quoted 22 lines
Quoted 12 lines
Quoted 11 lines

arch/arm/boot/dts/sun9i-a80-cx-a99.dts line 253

Quoted 9 lines
arch/arm/boot/dts/sun9i-a80-cx-a99.dts new file +9 -0
+	pmic at 745 {
+		compatible = "x-powers,axp808", "x-powers,axp806";
+		reg = <0x745>;
+		interrupt-parent = <&nmi_intc>;
+		interrupts = <0 IRQ_TYPE_LEVEL_LOW>;
+		interrupt-controller;
+		#interrupt-cells = <1>;
+
+		swin-supply = <&reg_dcdce>;

Please use an incude for that PMIC.

Quoted 4 lines

I don't understand. What would that include file contain? Very few lines would be shared between .dts files. You could save < 8 lines total if you #include the 5 lines that are common between the Cubieboard4, the Optimus and this board.

You also have to take into account that the PMIC is barely supported at the moment, you'll also need to define the GPIO controllers, the various power supplies, adc, etc. nodes that are going (hopefully to be added) in the future.

Quoted 5 lines

If the include is shared, once it is enabled, every one benefits from it. Otherwise, you have to duplicate that change across all DTs. This is much likely to error, harder to maintain, and you end up in a situation where none of the boards really support the same feature set while being based on the same SoC / PMIC combination.

Quoted 5 lines

Fair enough, I'll add an include file the next time around. Like you mentioned youself, though, we will only really know how compatible the two PMICs are when we have the code to support the other features (or a data sheet/user's manual is found). I'm hopeful based on what has been found so far.

Adding a new compatible early on is cheap. Dealing with a device that was supposed to be fully compatible but turned out not to be is really painful, so yeah, it's better to just add a new compatible from the very beginning, even if we'll always behave in the exact same way.

Maxime

Signature · 10 lines
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 801 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20170221/4fc5f4b0/attachment.sig>