Skip to contentSkip to conversation

Add basic support for RTC on Allwinner H6 SoC

PATCH Applied [email protected] 4 participants 21 messages Aug 20, 2019 – Aug 24, 2019

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

v2 cover Add basic support for RTC on Allwinner H6 SoC

From: Ondrej Jirman <megous-5qf/[email protected]>

I went through the datasheets for H6 and H5, and compared the differences. RTCs are largely similar, but not entirely compatible. Incompatibilities are in details not yet implemented by the rtc driver though.

I also corrected the clock tree in H6 DTSI.

This patchset is necessary for implementing the WiFi/Bluetooth support on boards using H6 SoC.

There was some discussion previously of describing HOSC, DCXO and XO oscillators and clocks as part of RTC in DT, but I decided against it because it's not necessary, becuse information that would be provided as a part of DT can already be determined at runtime from RTC registers, so this woudn't add any value and would only introduce complications to the driver. See: https://patchwork.kernel.org/cover/10898083/

Please take a look.

Thank you and regards,
Ondrej Jirman

Changes in v2:
- bindings converted to yaml
- added reviewed by tags

Ondrej Jirman (3):
dt-bindings: Add compatible for H6 RTC
rtc: sun6i: Add support for H6 RTC
arm64: dts: sun50i-h6: Add support for RTC and fix the clock tree

 .../bindings/rtc/allwinner,sun6i-a31-rtc.yaml | 13 ++++++
 arch/arm64/boot/dts/allwinner/sun50i-h6.dtsi  | 30 +++++++-------
 drivers/rtc/rtc-sun6i.c                       | 40 ++++++++++++++++++-
 3 files changed, 67 insertions(+), 16 deletions(-)
Signature · 1 line
2.22.1
#1 lore raw

PATCH v2 1/3 dt-bindings: Add compatible for H6 RTC

From: Ondrej Jirman <megous-5qf/[email protected]>

RTC on H6 is similar to the one on H5 SoC, but incompatible in small details. See the driver for description of differences. For example H6 RTC needs to enable the external low speed oscillator. Add new compatible for this RTC.

---
 .../bindings/rtc/allwinner,sun6i-a31-rtc.yaml       | 13 +++++++++++++
 1 file changed, 13 insertions(+)
Documentation/devicetree/bindings/rtc/allwinner,sun6i-a31-rtc.yaml +13 -0
index 924622f39c44..d7a57ec4a640 100644
@@ -25,6 +25,7 @@ properties:
       - items:
           - const: allwinner,sun50i-a64-rtc
           - const: allwinner,sun8i-h3-rtc
+      - const: allwinner,sun50i-h6-rtc
 
   reg:
     maxItems: 1
@@ -92,6 +93,18 @@ allOf:
           minItems: 3
           maxItems: 3
 
+  - if:
+      properties:
+        compatible:
+          contains:
+            const: allwinner,sun50i-h6-rtc
+
+    then:
+      properties:
+        clock-output-names:
+          minItems: 3
+          maxItems: 3
+
   - if:
       properties:
         compatible:
Signature · 1 line
2.22.1
maintainer Applied
#2 lore raw

Re: [PATCH v2 1/3] dt-bindings: Add compatible for H6 RTC

Quoted 7 lines · On 20/08/2019 17:19:32+0200, megous-5qf/[email protected]…

From: Ondrej Jirman <megous-5qf/[email protected]>

RTC on H6 is similar to the one on H5 SoC, but incompatible in small details. See the driver for description of differences. For example H6 RTC needs to enable the external low speed oscillator. Add new compatible for this RTC.

Quoted 5 lines
---
 .../bindings/rtc/allwinner,sun6i-a31-rtc.yaml       | 13 +++++++++++++
 1 file changed, 13 insertions(+)

Applied, thanks.

Signature · 3 lines
Alexandre Belloni, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
#3 lore raw

PATCH v2 2/3 rtc: sun6i: Add support for H6 RTC

In reply to [email protected]

From: Ondrej Jirman <megous-5qf/[email protected]>

RTC on H6 is mostly the same as on H5 and H3. It has slight differences mostly in features that are not yet supported by this driver.

Some differences are already stated in the comments in existing code. One other difference is that H6 has extra bit in LOSC_CTRL_REG, called EXT_LOSC_EN to enable/disable external low speed crystal oscillator.

It also has bit EXT_LOSC_STA in LOSC_AUTO_SWT_STA_REG, to check whether external low speed oscillator is working correctly.

This patch adds support for enabling LOSC when necessary:

- during reparenting
- when probing the clock

H6 also has capacbility to automatically reparent RTC clock from external crystal oscillator, to internal RC oscillator, if external oscillator fails. This is enabled by default. Disable it during probe.

---
 drivers/rtc/rtc-sun6i.c | 40 ++++++++++++++++++++++++++++++++++++++--
 1 file changed, 38 insertions(+), 2 deletions(-)
drivers/rtc/rtc-sun6i.c +38 -2
index d50ee023b559..b0c3752bed3f 100644
@@ -32,9 +32,11 @@
 /* Control register */
 #define SUN6I_LOSC_CTRL				0x0000
 #define SUN6I_LOSC_CTRL_KEY			(0x16aa << 16)
+#define SUN6I_LOSC_CTRL_AUTO_SWT_BYPASS		BIT(15)
 #define SUN6I_LOSC_CTRL_ALM_DHMS_ACC		BIT(9)
 #define SUN6I_LOSC_CTRL_RTC_HMS_ACC		BIT(8)
 #define SUN6I_LOSC_CTRL_RTC_YMD_ACC		BIT(7)
+#define SUN6I_LOSC_CTRL_EXT_LOSC_EN		BIT(4)
 #define SUN6I_LOSC_CTRL_EXT_OSC			BIT(0)
 #define SUN6I_LOSC_CTRL_ACC_MASK		GENMASK(9, 7)
 
@@ -128,6 +130,8 @@ struct sun6i_rtc_clk_data {
 	unsigned int has_prescaler : 1;
 	unsigned int has_out_clk : 1;
 	unsigned int export_iosc : 1;
+	unsigned int has_losc_en : 1;
+	unsigned int has_auto_swt : 1;
 };
 
 struct sun6i_rtc_dev {
@@ -190,6 +194,10 @@ static int sun6i_rtc_osc_set_parent(struct clk_hw *hw, u8 index)
 	val &= ~SUN6I_LOSC_CTRL_EXT_OSC;
 	val |= SUN6I_LOSC_CTRL_KEY;
 	val |= index ? SUN6I_LOSC_CTRL_EXT_OSC : 0;
+	if (rtc->data->has_losc_en) {
+		val &= ~SUN6I_LOSC_CTRL_EXT_LOSC_EN;
+		val |= index ? SUN6I_LOSC_CTRL_EXT_LOSC_EN : 0;
+	}
 	writel(val, rtc->base + SUN6I_LOSC_CTRL);
 	spin_unlock_irqrestore(&rtc->lock, flags);
 
@@ -215,6 +223,7 @@ static void __init sun6i_rtc_clk_init(struct device_node *node,
 	const char *iosc_name = "rtc-int-osc";
 	const char *clkout_name = "osc32k-out";
 	const char *parents[2];
+	u32 reg;
 
 	rtc = kzalloc(sizeof(*rtc), GFP_KERNEL);
 	if (!rtc)
@@ -235,9 +244,18 @@ static void __init sun6i_rtc_clk_init(struct device_node *node,
 		goto err;
 	}
 
+	reg = SUN6I_LOSC_CTRL_KEY;
+	if (rtc->data->has_auto_swt) {
+		/* Bypass auto-switch to int osc, on ext losc failure */
+		reg |= SUN6I_LOSC_CTRL_AUTO_SWT_BYPASS;
+		writel(reg, rtc->base + SUN6I_LOSC_CTRL);
+	}
+
 	/* Switch to the external, more precise, oscillator */
-	writel(SUN6I_LOSC_CTRL_KEY | SUN6I_LOSC_CTRL_EXT_OSC,
-	       rtc->base + SUN6I_LOSC_CTRL);
+	reg |= SUN6I_LOSC_CTRL_EXT_OSC;
+	if (rtc->data->has_losc_en)
+		reg |= SUN6I_LOSC_CTRL_EXT_LOSC_EN;
+	writel(reg, rtc->base + SUN6I_LOSC_CTRL);
 
 	/* Yes, I know, this is ugly. */
 	sun6i_rtc = rtc;
@@ -345,6 +363,23 @@ CLK_OF_DECLARE_DRIVER(sun8i_h3_rtc_clk, "allwinner,sun8i-h3-rtc",
 CLK_OF_DECLARE_DRIVER(sun50i_h5_rtc_clk, "allwinner,sun50i-h5-rtc",
 		      sun8i_h3_rtc_clk_init);
 
+static const struct sun6i_rtc_clk_data sun50i_h6_rtc_data = {
+	.rc_osc_rate = 16000000,
+	.fixed_prescaler = 32,
+	.has_prescaler = 1,
+	.has_out_clk = 1,
+	.export_iosc = 1,
+	.has_losc_en = 1,
+	.has_auto_swt = 1,
+};
+
+static void __init sun50i_h6_rtc_clk_init(struct device_node *node)
+{
+	sun6i_rtc_clk_init(node, &sun50i_h6_rtc_data);
+}
+CLK_OF_DECLARE_DRIVER(sun50i_h6_rtc_clk, "allwinner,sun50i-h6-rtc",
+		      sun50i_h6_rtc_clk_init);
+
 static const struct sun6i_rtc_clk_data sun8i_v3_rtc_data = {
 	.rc_osc_rate = 32000,
 	.has_out_clk = 1,
@@ -675,6 +710,7 @@ static const struct of_device_id sun6i_rtc_dt_ids[] = {
 	{ .compatible = "allwinner,sun8i-r40-rtc" },
 	{ .compatible = "allwinner,sun8i-v3-rtc" },
 	{ .compatible = "allwinner,sun50i-h5-rtc" },
+	{ .compatible = "allwinner,sun50i-h6-rtc" },
 	{ /* sentinel */ },
 };
 MODULE_DEVICE_TABLE(of, sun6i_rtc_dt_ids);
Signature · 1 line
2.22.1
maintainer Applied
#4 lore raw

Re: [PATCH v2 2/3] rtc: sun6i: Add support for H6 RTC

Quoted 22 lines · On 20/08/2019 17:19:33+0200, [email protected] wrote:

From: Ondrej Jirman <[email protected]>

RTC on H6 is mostly the same as on H5 and H3. It has slight differences mostly in features that are not yet supported by this driver.

Some differences are already stated in the comments in existing code. One other difference is that H6 has extra bit in LOSC_CTRL_REG, called EXT_LOSC_EN to enable/disable external low speed crystal oscillator.

It also has bit EXT_LOSC_STA in LOSC_AUTO_SWT_STA_REG, to check whether external low speed oscillator is working correctly.

This patch adds support for enabling LOSC when necessary:

- during reparenting
- when probing the clock

H6 also has capacbility to automatically reparent RTC clock from external crystal oscillator, to internal RC oscillator, if external oscillator fails. This is enabled by default. Disable it during probe.

Quoted 6 lines
---
 drivers/rtc/rtc-sun6i.c | 40 ++++++++++++++++++++++++++++++++++++++--
 1 file changed, 38 insertions(+), 2 deletions(-)

Applied, thanks.

Signature · 3 lines
Alexandre Belloni, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
maintainer
#5 lore raw

Re: [linux-sunxi] [PATCH v2 2/3] rtc: sun6i: Add support for H6 RTC

In reply to [email protected]

Hi!

Dne torek, 20. avgust 2019 ob 17:19:33 CEST je megous-5qf/[email protected] napisal(a):

Quoted 33 lines

From: Ondrej Jirman <megous-5qf/[email protected]>

RTC on H6 is mostly the same as on H5 and H3. It has slight differences mostly in features that are not yet supported by this driver.

Some differences are already stated in the comments in existing code. One other difference is that H6 has extra bit in LOSC_CTRL_REG, called EXT_LOSC_EN to enable/disable external low speed crystal oscillator.

It also has bit EXT_LOSC_STA in LOSC_AUTO_SWT_STA_REG, to check whether external low speed oscillator is working correctly.

This patch adds support for enabling LOSC when necessary:

- during reparenting
- when probing the clock

H6 also has capacbility to automatically reparent RTC clock from external crystal oscillator, to internal RC oscillator, if external oscillator fails. This is enabled by default. Disable it during probe.

---
 drivers/rtc/rtc-sun6i.c | 40 ++++++++++++++++++++++++++++++++++++++--
 1 file changed, 38 insertions(+), 2 deletions(-)
drivers/rtc/rtc-sun6i.c +0 -0
index d50ee023b559..b0c3752bed3f 100644

drivers/rtc/rtc-sun6i.c line 35

Quoted 4 lines
drivers/rtc/rtc-sun6i.c +1 -0
@@ -32,9 +32,11 @@
 /* Control register */
 #define SUN6I_LOSC_CTRL				0x0000
 #define SUN6I_LOSC_CTRL_KEY			(0x16aa << 16)
+#define SUN6I_LOSC_CTRL_AUTO_SWT_BYPASS		BIT(15)

User manual says that above field is bit 14.

Best regards,
Jernej

Quoted 88 lines
drivers/rtc/rtc-sun6i.c +37 -2
 #define SUN6I_LOSC_CTRL_ALM_DHMS_ACC		BIT(9)
 #define SUN6I_LOSC_CTRL_RTC_HMS_ACC		BIT(8)
 #define SUN6I_LOSC_CTRL_RTC_YMD_ACC		BIT(7)
+#define SUN6I_LOSC_CTRL_EXT_LOSC_EN		BIT(4)
 #define SUN6I_LOSC_CTRL_EXT_OSC			BIT(0)
 #define SUN6I_LOSC_CTRL_ACC_MASK		GENMASK(9, 7)
 
@@ -128,6 +130,8 @@ struct sun6i_rtc_clk_data {
 	unsigned int has_prescaler : 1;
 	unsigned int has_out_clk : 1;
 	unsigned int export_iosc : 1;
+	unsigned int has_losc_en : 1;
+	unsigned int has_auto_swt : 1;
 };
 
 struct sun6i_rtc_dev {
@@ -190,6 +194,10 @@ static int sun6i_rtc_osc_set_parent(struct clk_hw *hw,
 u8 index) val &= ~SUN6I_LOSC_CTRL_EXT_OSC;
 	val |= SUN6I_LOSC_CTRL_KEY;
 	val |= index ? SUN6I_LOSC_CTRL_EXT_OSC : 0;
+	if (rtc->data->has_losc_en) {
+		val &= ~SUN6I_LOSC_CTRL_EXT_LOSC_EN;
+		val |= index ? SUN6I_LOSC_CTRL_EXT_LOSC_EN : 0;
+	}
 	writel(val, rtc->base + SUN6I_LOSC_CTRL);
 	spin_unlock_irqrestore(&rtc->lock, flags);
 
@@ -215,6 +223,7 @@ static void __init sun6i_rtc_clk_init(struct device_node
 *node, const char *iosc_name = "rtc-int-osc";
 	const char *clkout_name = "osc32k-out";
 	const char *parents[2];
+	u32 reg;
 
 	rtc = kzalloc(sizeof(*rtc), GFP_KERNEL);
 	if (!rtc)
@@ -235,9 +244,18 @@ static void __init sun6i_rtc_clk_init(struct
 device_node *node, goto err;
 	}
 
+	reg = SUN6I_LOSC_CTRL_KEY;
+	if (rtc->data->has_auto_swt) {
+		/* Bypass auto-switch to int osc, on ext losc failure */
+		reg |= SUN6I_LOSC_CTRL_AUTO_SWT_BYPASS;
+		writel(reg, rtc->base + SUN6I_LOSC_CTRL);
+	}
+
 	/* Switch to the external, more precise, oscillator */
-	writel(SUN6I_LOSC_CTRL_KEY | SUN6I_LOSC_CTRL_EXT_OSC,
-	       rtc->base + SUN6I_LOSC_CTRL);
+	reg |= SUN6I_LOSC_CTRL_EXT_OSC;
+	if (rtc->data->has_losc_en)
+		reg |= SUN6I_LOSC_CTRL_EXT_LOSC_EN;
+	writel(reg, rtc->base + SUN6I_LOSC_CTRL);
 
 	/* Yes, I know, this is ugly. */
 	sun6i_rtc = rtc;
@@ -345,6 +363,23 @@ CLK_OF_DECLARE_DRIVER(sun8i_h3_rtc_clk,
 "allwinner,sun8i-h3-rtc", CLK_OF_DECLARE_DRIVER(sun50i_h5_rtc_clk,
 "allwinner,sun50i-h5-rtc", sun8i_h3_rtc_clk_init);
 
+static const struct sun6i_rtc_clk_data sun50i_h6_rtc_data = {
+	.rc_osc_rate = 16000000,
+	.fixed_prescaler = 32,
+	.has_prescaler = 1,
+	.has_out_clk = 1,
+	.export_iosc = 1,
+	.has_losc_en = 1,
+	.has_auto_swt = 1,
+};
+
+static void __init sun50i_h6_rtc_clk_init(struct device_node *node)
+{
+	sun6i_rtc_clk_init(node, &sun50i_h6_rtc_data);
+}
+CLK_OF_DECLARE_DRIVER(sun50i_h6_rtc_clk, "allwinner,sun50i-h6-rtc",
+		      sun50i_h6_rtc_clk_init);
+
 static const struct sun6i_rtc_clk_data sun8i_v3_rtc_data = {
 	.rc_osc_rate = 32000,
 	.has_out_clk = 1,
@@ -675,6 +710,7 @@ static const struct of_device_id sun6i_rtc_dt_ids[] = {
 	{ .compatible = "allwinner,sun8i-r40-rtc" },
 	{ .compatible = "allwinner,sun8i-v3-rtc" },
 	{ .compatible = "allwinner,sun50i-h5-rtc" },
+	{ .compatible = "allwinner,sun50i-h6-rtc" },
 	{ /* sentinel */ },
 };
 MODULE_DEVICE_TABLE(of, sun6i_rtc_dt_ids);
#6 lore raw

Re: [linux-sunxi] [PATCH v2 2/3] rtc: sun6i: Add support for H6 RTC

Hi,

Quoted 31 lines · Jernej Škrabec

Hi!

Dne torek, 20. avgust 2019 ob 17:19:33 CEST je megous-5qf/[email protected] napisal(a):

Quoted 28 lines

From: Ondrej Jirman <megous-5qf/[email protected]>

RTC on H6 is mostly the same as on H5 and H3. It has slight differences mostly in features that are not yet supported by this driver.

Some differences are already stated in the comments in existing code. One other difference is that H6 has extra bit in LOSC_CTRL_REG, called EXT_LOSC_EN to enable/disable external low speed crystal oscillator.

It also has bit EXT_LOSC_STA in LOSC_AUTO_SWT_STA_REG, to check whether external low speed oscillator is working correctly.

This patch adds support for enabling LOSC when necessary:

- during reparenting
- when probing the clock

H6 also has capacbility to automatically reparent RTC clock from external crystal oscillator, to internal RC oscillator, if external oscillator fails. This is enabled by default. Disable it during probe.

---
 drivers/rtc/rtc-sun6i.c | 40 ++++++++++++++++++++++++++++++++++++++--
 1 file changed, 38 insertions(+), 2 deletions(-)
Quoted 11 lines

drivers/rtc/rtc-sun6i.c line 35

Quoted 9 lines
drivers/rtc/rtc-sun6i.c +1 -0
index d50ee023b559..b0c3752bed3f 100644
@@ -32,9 +32,11 @@
 /* Control register */
 #define SUN6I_LOSC_CTRL				0x0000
 #define SUN6I_LOSC_CTRL_KEY			(0x16aa << 16)
+#define SUN6I_LOSC_CTRL_AUTO_SWT_BYPASS		BIT(15)

User manual says that above field is bit 14.

See the previous discussion, this is from BSP.

regards,

	o.
Quoted 100 lines

Best regards,
Jernej

Quoted 81 lines
drivers/rtc/rtc-sun6i.c +36 -2
 #define SUN6I_LOSC_CTRL_ALM_DHMS_ACC		BIT(9)
 #define SUN6I_LOSC_CTRL_RTC_HMS_ACC		BIT(8)
 #define SUN6I_LOSC_CTRL_RTC_YMD_ACC		BIT(7)
+#define SUN6I_LOSC_CTRL_EXT_LOSC_EN		BIT(4)
 #define SUN6I_LOSC_CTRL_EXT_OSC			BIT(0)
 #define SUN6I_LOSC_CTRL_ACC_MASK		GENMASK(9, 7)
 
@@ -128,6 +130,8 @@ struct sun6i_rtc_clk_data {
 	unsigned int has_prescaler : 1;
 	unsigned int has_out_clk : 1;
 	unsigned int export_iosc : 1;
+	unsigned int has_losc_en : 1;
+	unsigned int has_auto_swt : 1;
 };
 
 struct sun6i_rtc_dev {
@@ -190,6 +194,10 @@ static int sun6i_rtc_osc_set_parent(struct clk_hw *hw,
 u8 index) val &= ~SUN6I_LOSC_CTRL_EXT_OSC;
 	val |= SUN6I_LOSC_CTRL_KEY;
 	val |= index ? SUN6I_LOSC_CTRL_EXT_OSC : 0;
+	if (rtc->data->has_losc_en) {
+		val &= ~SUN6I_LOSC_CTRL_EXT_LOSC_EN;
+		val |= index ? SUN6I_LOSC_CTRL_EXT_LOSC_EN : 0;
+	}
 	writel(val, rtc->base + SUN6I_LOSC_CTRL);
 	spin_unlock_irqrestore(&rtc->lock, flags);
 
@@ -215,6 +223,7 @@ static void __init sun6i_rtc_clk_init(struct device_node
 *node, const char *iosc_name = "rtc-int-osc";
 	const char *clkout_name = "osc32k-out";
 	const char *parents[2];
+	u32 reg;
 
 	rtc = kzalloc(sizeof(*rtc), GFP_KERNEL);
 	if (!rtc)
@@ -235,9 +244,18 @@ static void __init sun6i_rtc_clk_init(struct
 device_node *node, goto err;
 	}
 
+	reg = SUN6I_LOSC_CTRL_KEY;
+	if (rtc->data->has_auto_swt) {
+		/* Bypass auto-switch to int osc, on ext losc failure */
+		reg |= SUN6I_LOSC_CTRL_AUTO_SWT_BYPASS;
+		writel(reg, rtc->base + SUN6I_LOSC_CTRL);
+	}
+
 	/* Switch to the external, more precise, oscillator */
-	writel(SUN6I_LOSC_CTRL_KEY | SUN6I_LOSC_CTRL_EXT_OSC,
-	       rtc->base + SUN6I_LOSC_CTRL);
+	reg |= SUN6I_LOSC_CTRL_EXT_OSC;
+	if (rtc->data->has_losc_en)
+		reg |= SUN6I_LOSC_CTRL_EXT_LOSC_EN;
+	writel(reg, rtc->base + SUN6I_LOSC_CTRL);
 
 	/* Yes, I know, this is ugly. */
 	sun6i_rtc = rtc;
@@ -345,6 +363,23 @@ CLK_OF_DECLARE_DRIVER(sun8i_h3_rtc_clk,
 "allwinner,sun8i-h3-rtc", CLK_OF_DECLARE_DRIVER(sun50i_h5_rtc_clk,
 "allwinner,sun50i-h5-rtc", sun8i_h3_rtc_clk_init);
 
+static const struct sun6i_rtc_clk_data sun50i_h6_rtc_data = {
+	.rc_osc_rate = 16000000,
+	.fixed_prescaler = 32,
+	.has_prescaler = 1,
+	.has_out_clk = 1,
+	.export_iosc = 1,
+	.has_losc_en = 1,
+	.has_auto_swt = 1,
+};
+
+static void __init sun50i_h6_rtc_clk_init(struct device_node *node)
+{
+	sun6i_rtc_clk_init(node, &sun50i_h6_rtc_data);
+}
+CLK_OF_DECLARE_DRIVER(sun50i_h6_rtc_clk, "allwinner,sun50i-h6-rtc",
+		      sun50i_h6_rtc_clk_init);
+
 static const struct sun6i_rtc_clk_data sun8i_v3_rtc_data = {
 	.rc_osc_rate = 32000,
 	.has_out_clk = 1,

drivers/rtc/rtc-sun6i.c static const struct of_device_id sun6i_rtc_dt_ids[] = { · line 716

Quoted 7 lines
drivers/rtc/rtc-sun6i.c +1 -0
@@ -675,6 +710,7 @@ static const struct of_device_id sun6i_rtc_dt_ids[] = {
 	{ .compatible = "allwinner,sun8i-r40-rtc" },
 	{ .compatible = "allwinner,sun8i-v3-rtc" },
 	{ .compatible = "allwinner,sun50i-h5-rtc" },
+	{ .compatible = "allwinner,sun50i-h6-rtc" },
 	{ /* sentinel */ },
 };
 MODULE_DEVICE_TABLE(of, sun6i_rtc_dt_ids);

_______________________________________________
linux-arm-kernel mailing list linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Signature · 3 lines
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
To view this discussion on the web, visit https://groups.google.com/d/msgid/linux-sunxi/20190824124654.nqtlkam2gsaqsj77%40core.my.home.
maintainer
#7 lore raw

Re: [linux-sunxi] [PATCH v2 2/3] rtc: sun6i: Add support for H6 RTC

Dne sobota, 24. avgust 2019 ob 14:46:54 CEST je Ondřej Jirman napisal(a):

Quoted 6 lines

Hi,

Quoted 3 lines · Jernej Škrabec

Hi!

Dne torek, 20. avgust 2019 ob 17:19:33 CEST je megous-5qf/[email protected]

napisal(a):

Quoted 39 lines
Quoted 39 lines
Quoted 39 lines

From: Ondrej Jirman <megous-5qf/[email protected]>

RTC on H6 is mostly the same as on H5 and H3. It has slight differences mostly in features that are not yet supported by this driver.

Some differences are already stated in the comments in existing code. One other difference is that H6 has extra bit in LOSC_CTRL_REG, called EXT_LOSC_EN to enable/disable external low speed crystal oscillator.

It also has bit EXT_LOSC_STA in LOSC_AUTO_SWT_STA_REG, to check whether external low speed oscillator is working correctly.

This patch adds support for enabling LOSC when necessary:

- during reparenting
- when probing the clock

H6 also has capacbility to automatically reparent RTC clock from external crystal oscillator, to internal RC oscillator, if external oscillator fails. This is enabled by default. Disable it during probe.

---

 drivers/rtc/rtc-sun6i.c | 40 ++++++++++++++++++++++++++++++++++++++--
 1 file changed, 38 insertions(+), 2 deletions(-)
drivers/rtc/rtc-sun6i.c +0 -0
index d50ee023b559..b0c3752bed3f 100644
@@ -32,9 +32,11 @@
 
 /* Control register */
 #define SUN6I_LOSC_CTRL				0x0000
 #define SUN6I_LOSC_CTRL_KEY			(0x16aa << 16)
 
Quoted 5 lines
Quoted 3 lines

drivers/rtc/rtc-sun6i.c line 35

Quoted 1 line
drivers/rtc/rtc-sun6i.c +1 -0
+#define SUN6I_LOSC_CTRL_AUTO_SWT_BYPASS		BIT(15)

User manual says that above field is bit 14.

See the previous discussion, this is from BSP.

I have two versions of BSP (don't ask me which) which have this set as bit 14 and changing this to 14 actually solves all my problems with LOSC (no more issues with setting RTC and HDMI-CEC works now - it uses LOSC as parent) on Tanix TX6 box.

Best regards,
Jernej

Quoted 55 lines

regards,

	o.
Quoted 51 lines

Best regards,
Jernej

Quoted 48 lines
drivers/rtc/rtc-sun6i.c +1 -0
 #define SUN6I_LOSC_CTRL_ALM_DHMS_ACC		BIT(9)
 #define SUN6I_LOSC_CTRL_RTC_HMS_ACC		BIT(8)
 #define SUN6I_LOSC_CTRL_RTC_YMD_ACC		BIT(7)
 
+#define SUN6I_LOSC_CTRL_EXT_LOSC_EN		BIT(4)

#define SUN6I_LOSC_CTRL_EXT_OSC BIT(0)
#define SUN6I_LOSC_CTRL_ACC_MASK GENMASK(9, 7)

drivers/rtc/rtc-sun6i.c +2 -0
@@ -128,6 +130,8 @@ struct sun6i_rtc_clk_data {
 
 	unsigned int has_prescaler : 1;
 	unsigned int has_out_clk : 1;
 	unsigned int export_iosc : 1;
 
+	unsigned int has_losc_en : 1;
+	unsigned int has_auto_swt : 1;
 

};

struct sun6i_rtc_dev {

drivers/rtc/rtc-sun6i.c +4 -0
@@ -190,6 +194,10 @@ static int sun6i_rtc_osc_set_parent(struct clk_hw
 *hw,
 u8 index) val &= ~SUN6I_LOSC_CTRL_EXT_OSC;
 
 	val |= SUN6I_LOSC_CTRL_KEY;
 	val |= index ? SUN6I_LOSC_CTRL_EXT_OSC : 0;
 
+	if (rtc->data->has_losc_en) {
+		val &= ~SUN6I_LOSC_CTRL_EXT_LOSC_EN;
+		val |= index ? SUN6I_LOSC_CTRL_EXT_LOSC_EN : 0;
+	}
drivers/rtc/rtc-sun6i.c +1 -0
 	writel(val, rtc->base + SUN6I_LOSC_CTRL);
 	spin_unlock_irqrestore(&rtc->lock, flags);
@@ -215,6 +223,7 @@ static void __init sun6i_rtc_clk_init(struct
 device_node *node, const char *iosc_name = "rtc-int-osc";
 
 	const char *clkout_name = "osc32k-out";
 	const char *parents[2];
 
+	u32 reg;
 
drivers/rtc/rtc-sun6i.c +0 -0
 	rtc = kzalloc(sizeof(*rtc), GFP_KERNEL);
 	if (!rtc)
Quoted 8 lines
Quoted 8 lines

drivers/rtc/rtc-sun6i.c static void __init sun6i_rtc_clk_init(struct · line 250

Quoted 8 lines
drivers/rtc/rtc-sun6i.c +3 -0
@@ -235,9 +244,18 @@ static void __init sun6i_rtc_clk_init(struct
 device_node *node, goto err;
 
 	}
 
+	reg = SUN6I_LOSC_CTRL_KEY;
+	if (rtc->data->has_auto_swt) {
+		/* Bypass auto-switch to int osc, on ext losc failure 

*/

Quoted 62 lines
Quoted 62 lines
Quoted 53 lines
drivers/rtc/rtc-sun6i.c +8 -2
+		reg |= SUN6I_LOSC_CTRL_AUTO_SWT_BYPASS;
+		writel(reg, rtc->base + SUN6I_LOSC_CTRL);
+	}
+
 
 	/* Switch to the external, more precise, oscillator */
 
-	writel(SUN6I_LOSC_CTRL_KEY | SUN6I_LOSC_CTRL_EXT_OSC,
-	       rtc->base + SUN6I_LOSC_CTRL);
+	reg |= SUN6I_LOSC_CTRL_EXT_OSC;
+	if (rtc->data->has_losc_en)
+		reg |= SUN6I_LOSC_CTRL_EXT_LOSC_EN;
+	writel(reg, rtc->base + SUN6I_LOSC_CTRL);
drivers/rtc/rtc-sun6i.c +17 -0
 	/* Yes, I know, this is ugly. */
 	sun6i_rtc = rtc;
@@ -345,6 +363,23 @@ CLK_OF_DECLARE_DRIVER(sun8i_h3_rtc_clk,
 "allwinner,sun8i-h3-rtc", CLK_OF_DECLARE_DRIVER(sun50i_h5_rtc_clk,
 "allwinner,sun50i-h5-rtc", sun8i_h3_rtc_clk_init);
 
+static const struct sun6i_rtc_clk_data sun50i_h6_rtc_data = {
+	.rc_osc_rate = 16000000,
+	.fixed_prescaler = 32,
+	.has_prescaler = 1,
+	.has_out_clk = 1,
+	.export_iosc = 1,
+	.has_losc_en = 1,
+	.has_auto_swt = 1,
+};
+
+static void __init sun50i_h6_rtc_clk_init(struct device_node *node)
+{
+	sun6i_rtc_clk_init(node, &sun50i_h6_rtc_data);
+}
+CLK_OF_DECLARE_DRIVER(sun50i_h6_rtc_clk, "allwinner,sun50i-h6-rtc",
+		      sun50i_h6_rtc_clk_init);
+
 
 static const struct sun6i_rtc_clk_data sun8i_v3_rtc_data = {
 
drivers/rtc/rtc-sun6i.c +1 -0
 	.rc_osc_rate = 32000,
 	.has_out_clk = 1,
@@ -675,6 +710,7 @@ static const struct of_device_id sun6i_rtc_dt_ids[]
 = {
 
 	{ .compatible = "allwinner,sun8i-r40-rtc" },
 	{ .compatible = "allwinner,sun8i-v3-rtc" },
 	{ .compatible = "allwinner,sun50i-h5-rtc" },
 
+	{ .compatible = "allwinner,sun50i-h6-rtc" },
Quoted 4 lines

{ /* sentinel */ },

};
MODULE_DEVICE_TABLE(of, sun6i_rtc_dt_ids);

_______________________________________________
linux-arm-kernel mailing list linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Signature · 3 lines
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
To view this discussion on the web, visit https://groups.google.com/d/msgid/linux-sunxi/2544007.NTLiB2pbcT%40jernej-laptop.
#8 lore raw

Re: [linux-sunxi] [PATCH v2 2/3] rtc: sun6i: Add support for H6 RTC

Quoted 51 lines · Jernej Škrabec

Dne sobota, 24. avgust 2019 ob 14:46:54 CEST je Ondřej Jirman napisal(a):

Quoted 6 lines

Hi,

Quoted 3 lines · Jernej Škrabec

Hi!

Dne torek, 20. avgust 2019 ob 17:19:33 CEST je megous-5qf/[email protected]

napisal(a):

Quoted 43 lines
Quoted 42 lines
Quoted 34 lines

From: Ondrej Jirman <megous-5qf/[email protected]>

RTC on H6 is mostly the same as on H5 and H3. It has slight differences mostly in features that are not yet supported by this driver.

Some differences are already stated in the comments in existing code. One other difference is that H6 has extra bit in LOSC_CTRL_REG, called EXT_LOSC_EN to enable/disable external low speed crystal oscillator.

It also has bit EXT_LOSC_STA in LOSC_AUTO_SWT_STA_REG, to check whether external low speed oscillator is working correctly.

This patch adds support for enabling LOSC when necessary:

- during reparenting
- when probing the clock

H6 also has capacbility to automatically reparent RTC clock from external crystal oscillator, to internal RC oscillator, if external oscillator fails. This is enabled by default. Disable it during probe.

---

 drivers/rtc/rtc-sun6i.c | 40 ++++++++++++++++++++++++++++++++++++++--
 1 file changed, 38 insertions(+), 2 deletions(-)
drivers/rtc/rtc-sun6i.c +0 -0
index d50ee023b559..b0c3752bed3f 100644

drivers/rtc/rtc-sun6i.c line 37

Quoted 6 lines
drivers/rtc/rtc-sun6i.c +1 -0
@@ -32,9 +32,11 @@
 
 /* Control register */
 #define SUN6I_LOSC_CTRL				0x0000
 #define SUN6I_LOSC_CTRL_KEY			(0x16aa << 16)
 
+#define SUN6I_LOSC_CTRL_AUTO_SWT_BYPASS		BIT(15)

User manual says that above field is bit 14.

Quoted 6 lines
Quoted 1 line

See the previous discussion, this is from BSP.

I have two versions of BSP (don't ask me which) which have this set as bit 14 and changing this to 14 actually solves all my problems with LOSC (no more issues with setting RTC and HDMI-CEC works now - it uses LOSC as parent) on Tanix TX6 box.

Interesting. Is LOSC fed externally generated clock, or is it setup as a crystal oscillator on your board?

Anyway, see here:

https://megous.com/git/linux/tree/drivers/rtc/rtc-sunxi.h?h=h6-4.9-bsp#n649
https://megous.com/git/linux/tree/drivers/rtc/rtc-sunxi.c?h=h6-4.9-bsp#n652

It would be nice to know what's really happening.

My output is:

[    0.832252] sun6i-rtc 7000000.rtc: registered as rtc0
[    0.832257] sun6i-rtc 7000000.rtc: RTC enabled
[    1.728968] sun6i-rtc 7000000.rtc: setting system clock to 1970-01-01T00:00:07 UTC (7)

I think, you may have just enabled the auto switch feature, and running the clock from low precision RC oscillator with your patch.

The real issue probably is that the mainline driver is missing this:

https://megous.com/git/linux/tree/drivers/rtc/rtc-sunxi.c?h=h6-4.9-bsp#n650

regards,

	o.
Quoted 138 lines

Best regards,
Jernej

Quoted 55 lines

regards,

	o.
Quoted 51 lines

Best regards,
Jernej

Quoted 48 lines
drivers/rtc/rtc-sun6i.c +1 -0
 #define SUN6I_LOSC_CTRL_ALM_DHMS_ACC		BIT(9)
 #define SUN6I_LOSC_CTRL_RTC_HMS_ACC		BIT(8)
 #define SUN6I_LOSC_CTRL_RTC_YMD_ACC		BIT(7)
 
+#define SUN6I_LOSC_CTRL_EXT_LOSC_EN		BIT(4)

#define SUN6I_LOSC_CTRL_EXT_OSC BIT(0)
#define SUN6I_LOSC_CTRL_ACC_MASK GENMASK(9, 7)

drivers/rtc/rtc-sun6i.c +2 -0
@@ -128,6 +130,8 @@ struct sun6i_rtc_clk_data {
 
 	unsigned int has_prescaler : 1;
 	unsigned int has_out_clk : 1;
 	unsigned int export_iosc : 1;
 
+	unsigned int has_losc_en : 1;
+	unsigned int has_auto_swt : 1;
 

};

struct sun6i_rtc_dev {

drivers/rtc/rtc-sun6i.c +4 -0
@@ -190,6 +194,10 @@ static int sun6i_rtc_osc_set_parent(struct clk_hw
 *hw,
 u8 index) val &= ~SUN6I_LOSC_CTRL_EXT_OSC;
 
 	val |= SUN6I_LOSC_CTRL_KEY;
 	val |= index ? SUN6I_LOSC_CTRL_EXT_OSC : 0;
 
+	if (rtc->data->has_losc_en) {
+		val &= ~SUN6I_LOSC_CTRL_EXT_LOSC_EN;
+		val |= index ? SUN6I_LOSC_CTRL_EXT_LOSC_EN : 0;
+	}
drivers/rtc/rtc-sun6i.c +1 -0
 	writel(val, rtc->base + SUN6I_LOSC_CTRL);
 	spin_unlock_irqrestore(&rtc->lock, flags);
@@ -215,6 +223,7 @@ static void __init sun6i_rtc_clk_init(struct
 device_node *node, const char *iosc_name = "rtc-int-osc";
 
 	const char *clkout_name = "osc32k-out";
 	const char *parents[2];
 
+	u32 reg;
 
drivers/rtc/rtc-sun6i.c +0 -0
 	rtc = kzalloc(sizeof(*rtc), GFP_KERNEL);
 	if (!rtc)
Quoted 8 lines
Quoted 8 lines

drivers/rtc/rtc-sun6i.c static void __init sun6i_rtc_clk_init(struct · line 250

Quoted 8 lines
drivers/rtc/rtc-sun6i.c +3 -0
@@ -235,9 +244,18 @@ static void __init sun6i_rtc_clk_init(struct
 device_node *node, goto err;
 
 	}
 
+	reg = SUN6I_LOSC_CTRL_KEY;
+	if (rtc->data->has_auto_swt) {
+		/* Bypass auto-switch to int osc, on ext losc failure 

*/

Quoted 55 lines
Quoted 55 lines
Quoted 55 lines
drivers/rtc/rtc-sun6i.c +8 -2
+		reg |= SUN6I_LOSC_CTRL_AUTO_SWT_BYPASS;
+		writel(reg, rtc->base + SUN6I_LOSC_CTRL);
+	}
+
 
 	/* Switch to the external, more precise, oscillator */
 
-	writel(SUN6I_LOSC_CTRL_KEY | SUN6I_LOSC_CTRL_EXT_OSC,
-	       rtc->base + SUN6I_LOSC_CTRL);
+	reg |= SUN6I_LOSC_CTRL_EXT_OSC;
+	if (rtc->data->has_losc_en)
+		reg |= SUN6I_LOSC_CTRL_EXT_LOSC_EN;
+	writel(reg, rtc->base + SUN6I_LOSC_CTRL);
drivers/rtc/rtc-sun6i.c +17 -0
 	/* Yes, I know, this is ugly. */
 	sun6i_rtc = rtc;
@@ -345,6 +363,23 @@ CLK_OF_DECLARE_DRIVER(sun8i_h3_rtc_clk,
 "allwinner,sun8i-h3-rtc", CLK_OF_DECLARE_DRIVER(sun50i_h5_rtc_clk,
 "allwinner,sun50i-h5-rtc", sun8i_h3_rtc_clk_init);
 
+static const struct sun6i_rtc_clk_data sun50i_h6_rtc_data = {
+	.rc_osc_rate = 16000000,
+	.fixed_prescaler = 32,
+	.has_prescaler = 1,
+	.has_out_clk = 1,
+	.export_iosc = 1,
+	.has_losc_en = 1,
+	.has_auto_swt = 1,
+};
+
+static void __init sun50i_h6_rtc_clk_init(struct device_node *node)
+{
+	sun6i_rtc_clk_init(node, &sun50i_h6_rtc_data);
+}
+CLK_OF_DECLARE_DRIVER(sun50i_h6_rtc_clk, "allwinner,sun50i-h6-rtc",
+		      sun50i_h6_rtc_clk_init);
+
 
 static const struct sun6i_rtc_clk_data sun8i_v3_rtc_data = {
 
drivers/rtc/rtc-sun6i.c +1 -0
 	.rc_osc_rate = 32000,
 	.has_out_clk = 1,
@@ -675,6 +710,7 @@ static const struct of_device_id sun6i_rtc_dt_ids[]
 = {
 
 	{ .compatible = "allwinner,sun8i-r40-rtc" },
 	{ .compatible = "allwinner,sun8i-v3-rtc" },
 	{ .compatible = "allwinner,sun50i-h5-rtc" },
 
+	{ .compatible = "allwinner,sun50i-h6-rtc" },

{ /* sentinel */ },

Quoted 7 lines
Quoted 7 lines
Quoted 2 lines

};
MODULE_DEVICE_TABLE(of, sun6i_rtc_dt_ids);

_______________________________________________
linux-arm-kernel mailing list linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

_______________________________________________
linux-arm-kernel mailing list linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Signature · 3 lines
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
To view this discussion on the web, visit https://groups.google.com/d/msgid/linux-sunxi/20190824130544.kxypq3siv7kffidp%40core.my.home.
maintainer
#9 lore raw

Re: [linux-sunxi] [PATCH v2 2/3] rtc: sun6i: Add support for H6 RTC

Dne sobota, 24. avgust 2019 ob 15:05:44 CEST je Ondřej Jirman napisal(a):

Quoted 50 lines · Jernej Škrabec
Quoted 49 lines

Dne sobota, 24. avgust 2019 ob 14:46:54 CEST je Ondřej Jirman napisal(a):

Quoted 6 lines

Hi,

Quoted 3 lines · Jernej Škrabec

Hi!

Dne torek, 20. avgust 2019 ob 17:19:33 CEST je megous-5qf/[email protected]

napisal(a):

Quoted 40 lines
Quoted 40 lines
Quoted 40 lines

From: Ondrej Jirman <megous-5qf/[email protected]>

RTC on H6 is mostly the same as on H5 and H3. It has slight differences
mostly in features that are not yet supported by this driver.

Some differences are already stated in the comments in existing code.
One other difference is that H6 has extra bit in LOSC_CTRL_REG, called
EXT_LOSC_EN to enable/disable external low speed crystal oscillator.

It also has bit EXT_LOSC_STA in LOSC_AUTO_SWT_STA_REG, to check whether
external low speed oscillator is working correctly.

This patch adds support for enabling LOSC when necessary:

- during reparenting
- when probing the clock

H6 also has capacbility to automatically reparent RTC clock from external crystal oscillator, to internal RC oscillator, if external oscillator fails. This is enabled by default. Disable it during probe.

---

 drivers/rtc/rtc-sun6i.c | 40
 ++++++++++++++++++++++++++++++++++++++--
 1 file changed, 38 insertions(+), 2 deletions(-)
drivers/rtc/rtc-sun6i.c +0 -0
index d50ee023b559..b0c3752bed3f 100644
@@ -32,9 +32,11 @@
 
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

drivers/rtc/rtc-sun6i.c

Quoted 3 lines
drivers/rtc/rtc-sun6i.c +0 -0
 /* Control register */
 #define SUN6I_LOSC_CTRL				0x0000
 #define SUN6I_LOSC_CTRL_KEY			(0x16aa 

<< 16)

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

drivers/rtc/rtc-sun6i.c line 35

Quoted 2 lines
drivers/rtc/rtc-sun6i.c +1 -0
+#define SUN6I_LOSC_CTRL_AUTO_SWT_BYPASS		BIT(15)

User manual says that above field is bit 14.

See the previous discussion, this is from BSP.

Quoted 7 lines
Quoted 4 lines

I have two versions of BSP (don't ask me which) which have this set as bit 14 and changing this to 14 actually solves all my problems with LOSC (no more issues with setting RTC and HDMI-CEC works now - it uses LOSC as parent) on Tanix TX6 box.

Interesting. Is LOSC fed externally generated clock, or is it setup as a crystal oscillator on your board?

I really don't know, but here is DT: http://ix.io/1ThI

Quoted 5 lines

Anyway, see here:

https://megous.com/git/linux/tree/drivers/rtc/rtc-sunxi.h?h=h6-4.9-bsp#n649
https://megous.com/git/linux/tree/drivers/rtc/rtc-sunxi.c?h=h6-4.9-bsp#n652

Interesting, 4.9 BSP has additional bit definition, which is not documented in manual and 3.10 BSP to which I refer.

I was referring to 3.10 BSP, which uses only bit 14. I thought that you named it differently.

Quoted 9 lines

It would be nice to know what's really happening.

My output is:

[    0.832252] sun6i-rtc 7000000.rtc: registered as rtc0
[    0.832257] sun6i-rtc 7000000.rtc: RTC enabled
[    1.728968] sun6i-rtc 7000000.rtc: setting system clock to
1970-01-01T00:00:07 UTC (7)

With change, I get same output.

Quoted 3 lines

I think, you may have just enabled the auto switch feature, and running the clock from low precision RC oscillator with your patch.

True, now I think there is no external crystal, but I'm still not sure how to confirm that.

Quoted 5 lines

The real issue probably is that the mainline driver is missing this:

https://megous.com/git/linux/tree/drivers/rtc/rtc-sunxi.c?h=h6-4.9-bsp#n650

Not sure what you mean by that. ext vs. int source selection?

Best regards,
Jernej

Quoted 141 lines

regards,

	o.
Quoted 138 lines

Best regards,
Jernej

Quoted 56 lines

regards,

	o.
Quoted 52 lines

Best regards,
Jernej

Quoted 49 lines
drivers/rtc/rtc-sun6i.c +1 -0
 #define SUN6I_LOSC_CTRL_ALM_DHMS_ACC		BIT(9)
 #define SUN6I_LOSC_CTRL_RTC_HMS_ACC		BIT(8)
 #define SUN6I_LOSC_CTRL_RTC_YMD_ACC		BIT(7)
 
+#define SUN6I_LOSC_CTRL_EXT_LOSC_EN		BIT(4)
 

#define SUN6I_LOSC_CTRL_EXT_OSC BIT(0)
#define SUN6I_LOSC_CTRL_ACC_MASK GENMASK(9, 7)

drivers/rtc/rtc-sun6i.c +2 -0
@@ -128,6 +130,8 @@ struct sun6i_rtc_clk_data {
 
 	unsigned int has_prescaler : 1;
 	unsigned int has_out_clk : 1;
 	unsigned int export_iosc : 1;
 
+	unsigned int has_losc_en : 1;
+	unsigned int has_auto_swt : 1;
 

};

struct sun6i_rtc_dev {

drivers/rtc/rtc-sun6i.c +0 -0
@@ -190,6 +194,10 @@ static int sun6i_rtc_osc_set_parent(struct

clk_hw
*hw,
u8 index) val &= ~SUN6I_LOSC_CTRL_EXT_OSC;

drivers/rtc/rtc-sun6i.c +5 -0
 	val |= SUN6I_LOSC_CTRL_KEY;
 	val |= index ? SUN6I_LOSC_CTRL_EXT_OSC : 0;
 
+	if (rtc->data->has_losc_en) {
+		val &= ~SUN6I_LOSC_CTRL_EXT_LOSC_EN;
+		val |= index ? SUN6I_LOSC_CTRL_EXT_LOSC_EN : 0;
+	}
 
 	writel(val, rtc->base + SUN6I_LOSC_CTRL);
 	spin_unlock_irqrestore(&rtc->lock, flags);
@@ -215,6 +223,7 @@ static void __init sun6i_rtc_clk_init(struct
 device_node *node, const char *iosc_name = "rtc-int-osc";
 
 	const char *clkout_name = "osc32k-out";
 	const char *parents[2];
 
+	u32 reg;
 
drivers/rtc/rtc-sun6i.c +0 -0
 	rtc = kzalloc(sizeof(*rtc), GFP_KERNEL);
 	if (!rtc)
Quoted 8 lines
Quoted 8 lines

drivers/rtc/rtc-sun6i.c static void __init sun6i_rtc_clk_init(struct · line 250

Quoted 8 lines
drivers/rtc/rtc-sun6i.c +3 -0
@@ -235,9 +244,18 @@ static void __init sun6i_rtc_clk_init(struct
 device_node *node, goto err;
 
 	}
 
+	reg = SUN6I_LOSC_CTRL_KEY;
+	if (rtc->data->has_auto_swt) {
+		/* Bypass auto-switch to int osc, on ext losc failure

*/

Quoted 56 lines
Quoted 56 lines
Quoted 56 lines
drivers/rtc/rtc-sun6i.c +8 -2
+		reg |= SUN6I_LOSC_CTRL_AUTO_SWT_BYPASS;
+		writel(reg, rtc->base + SUN6I_LOSC_CTRL);
+	}
+
 
 	/* Switch to the external, more precise, oscillator */
 
-	writel(SUN6I_LOSC_CTRL_KEY | SUN6I_LOSC_CTRL_EXT_OSC,
-	       rtc->base + SUN6I_LOSC_CTRL);
+	reg |= SUN6I_LOSC_CTRL_EXT_OSC;
+	if (rtc->data->has_losc_en)
+		reg |= SUN6I_LOSC_CTRL_EXT_LOSC_EN;
+	writel(reg, rtc->base + SUN6I_LOSC_CTRL);
drivers/rtc/rtc-sun6i.c +17 -0
 	/* Yes, I know, this is ugly. */
 	sun6i_rtc = rtc;
@@ -345,6 +363,23 @@ CLK_OF_DECLARE_DRIVER(sun8i_h3_rtc_clk,
 "allwinner,sun8i-h3-rtc", CLK_OF_DECLARE_DRIVER(sun50i_h5_rtc_clk,
 "allwinner,sun50i-h5-rtc", sun8i_h3_rtc_clk_init);
 
+static const struct sun6i_rtc_clk_data sun50i_h6_rtc_data = {
+	.rc_osc_rate = 16000000,
+	.fixed_prescaler = 32,
+	.has_prescaler = 1,
+	.has_out_clk = 1,
+	.export_iosc = 1,
+	.has_losc_en = 1,
+	.has_auto_swt = 1,
+};
+
+static void __init sun50i_h6_rtc_clk_init(struct device_node *node)
+{
+	sun6i_rtc_clk_init(node, &sun50i_h6_rtc_data);
+}
+CLK_OF_DECLARE_DRIVER(sun50i_h6_rtc_clk, "allwinner,sun50i-h6-rtc",
+		      sun50i_h6_rtc_clk_init);
+
 
 static const struct sun6i_rtc_clk_data sun8i_v3_rtc_data = {
 
drivers/rtc/rtc-sun6i.c +0 -0
 	.rc_osc_rate = 32000,
 	.has_out_clk = 1,
@@ -675,6 +710,7 @@ static const struct of_device_id

sun6i_rtc_dt_ids[]
= {

drivers/rtc/rtc-sun6i.c +1 -0
 	{ .compatible = "allwinner,sun8i-r40-rtc" },
 	{ .compatible = "allwinner,sun8i-v3-rtc" },
 	{ .compatible = "allwinner,sun50i-h5-rtc" },
 
+	{ .compatible = "allwinner,sun50i-h6-rtc" },
 
 	{ /* sentinel */ },
Quoted 7 lines
Quoted 7 lines
Quoted 2 lines

};
MODULE_DEVICE_TABLE(of, sun6i_rtc_dt_ids);

_______________________________________________
linux-arm-kernel mailing list linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

_______________________________________________
linux-arm-kernel mailing list linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Signature · 3 lines
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
To view this discussion on the web, visit https://groups.google.com/d/msgid/linux-sunxi/10880274.nteIY8W56H%40jernej-laptop.
#10 lore raw

Re: [linux-sunxi] [PATCH v2 2/3] rtc: sun6i: Add support for H6 RTC

Quoted 94 lines · Jernej Škrabec

Dne sobota, 24. avgust 2019 ob 15:05:44 CEST je Ondřej Jirman napisal(a):

Quoted 50 lines · Jernej Škrabec
Quoted 49 lines

Dne sobota, 24. avgust 2019 ob 14:46:54 CEST je Ondřej Jirman napisal(a):

Quoted 6 lines

Hi,

Quoted 3 lines · Jernej Škrabec

Hi!

Dne torek, 20. avgust 2019 ob 17:19:33 CEST je megous-5qf/[email protected]

napisal(a):

Quoted 40 lines
Quoted 40 lines
Quoted 40 lines

From: Ondrej Jirman <megous-5qf/[email protected]>

RTC on H6 is mostly the same as on H5 and H3. It has slight differences
mostly in features that are not yet supported by this driver.

Some differences are already stated in the comments in existing code.
One other difference is that H6 has extra bit in LOSC_CTRL_REG, called
EXT_LOSC_EN to enable/disable external low speed crystal oscillator.

It also has bit EXT_LOSC_STA in LOSC_AUTO_SWT_STA_REG, to check whether
external low speed oscillator is working correctly.

This patch adds support for enabling LOSC when necessary:

- during reparenting
- when probing the clock

H6 also has capacbility to automatically reparent RTC clock from external crystal oscillator, to internal RC oscillator, if external oscillator fails. This is enabled by default. Disable it during probe.

---

 drivers/rtc/rtc-sun6i.c | 40
 ++++++++++++++++++++++++++++++++++++++--
 1 file changed, 38 insertions(+), 2 deletions(-)
drivers/rtc/rtc-sun6i.c +0 -0
index d50ee023b559..b0c3752bed3f 100644
@@ -32,9 +32,11 @@
 
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

drivers/rtc/rtc-sun6i.c

Quoted 3 lines
drivers/rtc/rtc-sun6i.c +0 -0
 /* Control register */
 #define SUN6I_LOSC_CTRL				0x0000
 #define SUN6I_LOSC_CTRL_KEY			(0x16aa 

<< 16)

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

drivers/rtc/rtc-sun6i.c line 35

Quoted 2 lines
drivers/rtc/rtc-sun6i.c +1 -0
+#define SUN6I_LOSC_CTRL_AUTO_SWT_BYPASS		BIT(15)

User manual says that above field is bit 14.

See the previous discussion, this is from BSP.

Quoted 7 lines
Quoted 4 lines

I have two versions of BSP (don't ask me which) which have this set as bit 14 and changing this to 14 actually solves all my problems with LOSC (no more issues with setting RTC and HDMI-CEC works now - it uses LOSC as parent) on Tanix TX6 box.

Interesting. Is LOSC fed externally generated clock, or is it setup as a crystal oscillator on your board?

I really don't know, but here is DT: http://ix.io/1ThI

Quoted 5 lines

Anyway, see here:

https://megous.com/git/linux/tree/drivers/rtc/rtc-sunxi.h?h=h6-4.9-bsp#n649
https://megous.com/git/linux/tree/drivers/rtc/rtc-sunxi.c?h=h6-4.9-bsp#n652

Interesting, 4.9 BSP has additional bit definition, which is not documented in manual and 3.10 BSP to which I refer.

I was referring to 3.10 BSP, which uses only bit 14. I thought that you named it differently.

Quoted 9 lines

It would be nice to know what's really happening.

My output is:

[    0.832252] sun6i-rtc 7000000.rtc: registered as rtc0
[    0.832257] sun6i-rtc 7000000.rtc: RTC enabled
[    1.728968] sun6i-rtc 7000000.rtc: setting system clock to
1970-01-01T00:00:07 UTC (7)
Quoted 8 lines

With change, I get same output.

Quoted 3 lines

I think, you may have just enabled the auto switch feature, and running the clock from low precision RC oscillator with your patch.

True, now I think there is no external crystal, but I'm still not sure how to confirm that.

Visually?

That would explain why it doesn't work for you. The mainline RTC driver disables auto-switch feature, and if your board doesn't have a crystal for LOSC, RTC will not generate a clock for the RTC.

H6's dtsi describes by default a situatiuon with external 32k crystal oscillator. See ext_osc32k node. That's incorrect for your board if it doesn't have the crystal. You need to fix this in the DTS for your board instead of patching the driver.

The driver has parent clock selection logic in case the LOSC crystal is not used.

Your patch enables automatic detection of LOSC failure and RTC changes clock to LOSC automatically, despite what's described in the DTS. That may fix the issue, but is not the correct solution.

Registers on my board look like this (external 32k osc is used) for reference:

LOSC_CTRL_REG[7000000]: 8011

	KEY_FIELD                      ???                  (0)
	LOSC_AUTO_SWT_BYPASS           EN                   (1)
	LOSC_AUTO_SWT_EN               DIS                  (0)
	EXT_LOSC_EN                    EN                   (1)
	EXT_LOSC_GSM                   LOW                  (0)
	BATTERY_DIR                    DISCHARGE            (0)
	LOSC_SRC_SEL                   EXT32k               (1)

LOSC_AUTO_SWT_STA_REG[7000004]: 1

	EXT_LOSC_STA                   OK                   (0)
	LOSC_AUTO_SWT_PEND             NOEFF                (0)
	LOSC_SRC_SEL_STA               EXT32K               (1)

regards,

	o.
Quoted 15 lines
Quoted 5 lines

The real issue probably is that the mainline driver is missing this:

https://megous.com/git/linux/tree/drivers/rtc/rtc-sunxi.c?h=h6-4.9-bsp#n650

Not sure what you mean by that. ext vs. int source selection?

Best regards,
Jernej

Quoted 2 lines

regards,

	o.
Signature · 3 lines
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
To view this discussion on the web, visit https://groups.google.com/d/msgid/linux-sunxi/20190824133057.rhpj3xuzr3vymdiy%40core.my.home.
maintainer
#11 lore raw

Re: [linux-sunxi] [PATCH v2 2/3] rtc: sun6i: Add support for H6 RTC

Dne sobota, 24. avgust 2019 ob 15:30:57 CEST je Ondřej Jirman napisal(a):

Quoted 4 lines · Jernej Škrabec
Quoted 3 lines

Dne sobota, 24. avgust 2019 ob 15:05:44 CEST je Ondřej Jirman napisal(a):

Quoted 2 lines · Jernej Škrabec
Quoted 1 line

Dne sobota, 24. avgust 2019 ob 14:46:54 CEST je Ondřej Jirman

napisal(a):

Quoted 44 lines
Quoted 44 lines
Quoted 44 lines
Quoted 44 lines
Quoted 6 lines

Hi,

Quoted 3 lines · Jernej Škrabec

Hi!

Dne torek, 20. avgust 2019 ob 17:19:33 CEST je megous-5qf/[email protected]

napisal(a):

Quoted 36 lines
Quoted 36 lines
Quoted 36 lines

From: Ondrej Jirman <megous-5qf/[email protected]>

RTC on H6 is mostly the same as on H5 and H3. It has slight differences
mostly in features that are not yet supported by this driver.

Some differences are already stated in the comments in existing code.
One other difference is that H6 has extra bit in LOSC_CTRL_REG, called
EXT_LOSC_EN to enable/disable external low speed crystal oscillator.

It also has bit EXT_LOSC_STA in LOSC_AUTO_SWT_STA_REG, to check whether
external low speed oscillator is working correctly.

This patch adds support for enabling LOSC when necessary:

- during reparenting
- when probing the clock

H6 also has capacbility to automatically reparent RTC clock from external crystal oscillator, to internal RC oscillator, if external
oscillator fails. This is enabled by default. Disable it during probe.

---

 drivers/rtc/rtc-sun6i.c | 40
 ++++++++++++++++++++++++++++++++++++++--
 1 file changed, 38 insertions(+), 2 deletions(-)
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

drivers/rtc/rtc-sun6i.c line 34

Quoted 8 lines
drivers/rtc/rtc-sun6i.c +0 -0
index d50ee023b559..b0c3752bed3f 100644
@@ -32,9 +32,11 @@
 
 /* Control register */
 #define SUN6I_LOSC_CTRL				

0x0000

Quoted 55 lines
Quoted 44 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

drivers/rtc/rtc-sun6i.c line 35

Quoted 1 line
drivers/rtc/rtc-sun6i.c +0 -0
 #define SUN6I_LOSC_CTRL_KEY			(0x16aa

<< 16)

Quoted 13 lines
Quoted 4 lines
Quoted 4 lines
Quoted 3 lines

drivers/rtc/rtc-sun6i.c line 36

Quoted 1 line
drivers/rtc/rtc-sun6i.c +1 -0
+#define SUN6I_LOSC_CTRL_AUTO_SWT_BYPASS		BIT(15)

User manual says that above field is bit 14.

Quoted 8 lines
Quoted 1 line

See the previous discussion, this is from BSP.

I have two versions of BSP (don't ask me which) which have this set as bit
14 and changing this to 14 actually solves all my problems with LOSC (no
more issues with setting RTC and HDMI-CEC works now - it uses LOSC as parent) on Tanix TX6 box.

Quoted 2 lines

Interesting. Is LOSC fed externally generated clock, or is it setup as a crystal oscillator on your board?

I really don't know, but here is DT: http://ix.io/1ThI

Quoted 6 lines

Anyway, see here:

https://megous.com/git/linux/tree/drivers/rtc/rtc-sunxi.h?h=h6-4.9-bsp#n
649
https://megous.com/git/linux/tree/drivers/rtc/rtc-sunxi.c?h=h6-4.9-bsp#n
652

Interesting, 4.9 BSP has additional bit definition, which is not documented in manual and 3.10 BSP to which I refer.

I was referring to 3.10 BSP, which uses only bit 14. I thought that you named it differently.

Quoted 8 lines

It would be nice to know what's really happening.

My output is:

[    0.832252] sun6i-rtc 7000000.rtc: registered as rtc0
[    0.832257] sun6i-rtc 7000000.rtc: RTC enabled
[    1.728968] sun6i-rtc 7000000.rtc: setting system clock to
1970-01-01T00:00:07 UTC (7)
Quoted 8 lines

With change, I get same output.

Quoted 3 lines

I think, you may have just enabled the auto switch feature, and running the
clock from low precision RC oscillator with your patch.

True, now I think there is no external crystal, but I'm still not sure how to confirm that.

Visually?

Quoted 8 lines

That would explain why it doesn't work for you. The mainline RTC driver disables auto-switch feature, and if your board doesn't have a crystal for LOSC, RTC will not generate a clock for the RTC.

H6's dtsi describes by default a situatiuon with external 32k crystal oscillator. See ext_osc32k node. That's incorrect for your board if it doesn't have the crystal. You need to fix this in the DTS for your board instead of patching the driver.

I see that reparenting is supported, but I'm not sure how to fix that in DT. Any suggestion?

Quoted 20 lines

The driver has parent clock selection logic in case the LOSC crystal is not used.

Your patch enables automatic detection of LOSC failure and RTC changes clock to LOSC automatically, despite what's described in the DTS. That may fix the issue, but is not the correct solution.

Registers on my board look like this (external 32k osc is used) for reference:

LOSC_CTRL_REG[7000000]: 8011

	KEY_FIELD                      ???                  (0)
	LOSC_AUTO_SWT_BYPASS           EN                   (1)
	LOSC_AUTO_SWT_EN               DIS                  (0)
	EXT_LOSC_EN                    EN                   (1)
	EXT_LOSC_GSM                   LOW                  (0)
	BATTERY_DIR                    DISCHARGE            (0)
	LOSC_SRC_SEL                   EXT32k               (1)
Quoted 5 lines

LOSC_AUTO_SWT_STA_REG[7000004]: 1

	EXT_LOSC_STA                   OK                   (0)
	LOSC_AUTO_SWT_PEND             NOEFF                (0)
	LOSC_SRC_SEL_STA               EXT32K               (1)

In my case LOSC_CTRL_REG has value 0x4010 and LOSC_AUTO_SWT_STA_REG has value 0x4, so there is issue with external crystal (it's missing) and RTC switched to internal one.

BTW, what's wrong with automatic switching? Why is it disabled?

Best regards,
Jernej

Quoted 18 lines

regards,

	o.
Quoted 15 lines
Quoted 4 lines

The real issue probably is that the mainline driver is missing this:

https://megous.com/git/linux/tree/drivers/rtc/rtc-sunxi.c?h=h6-4.9-bsp#n
650

Not sure what you mean by that. ext vs. int source selection?

Best regards,
Jernej

Quoted 3 lines

regards,

	o.
Signature · 3 lines
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
To view this discussion on the web, visit https://groups.google.com/d/msgid/linux-sunxi/1690798.2HKiRSsjat%40jernej-laptop.
#12 lore raw

Re: [linux-sunxi] [PATCH v2 2/3] rtc: sun6i: Add support for H6 RTC

Hello Jernej,

Quoted 6 lines · Jernej Škrabec
Quoted 6 lines

Visually?

That would explain why it doesn't work for you. The mainline RTC driver disables auto-switch feature, and if your board doesn't have a crystal for LOSC, RTC will not generate a clock for the RTC.

Quoted 7 lines
Quoted 4 lines

H6's dtsi describes by default a situatiuon with external 32k crystal oscillator. See ext_osc32k node. That's incorrect for your board if it doesn't have the crystal. You need to fix this in the DTS for your board instead of patching the driver.

I see that reparenting is supported, but I'm not sure how to fix that in DT. Any suggestion?

You may try removing the clocks property from rtc node.

Quoted 26 lines
Quoted 20 lines

The driver has parent clock selection logic in case the LOSC crystal is not used.

Your patch enables automatic detection of LOSC failure and RTC changes clock to LOSC automatically, despite what's described in the DTS. That may fix the issue, but is not the correct solution.

Registers on my board look like this (external 32k osc is used) for reference:

LOSC_CTRL_REG[7000000]: 8011

	KEY_FIELD                      ???                  (0)
	LOSC_AUTO_SWT_BYPASS           EN                   (1)
	LOSC_AUTO_SWT_EN               DIS                  (0)
	EXT_LOSC_EN                    EN                   (1)
	EXT_LOSC_GSM                   LOW                  (0)
	BATTERY_DIR                    DISCHARGE            (0)
	LOSC_SRC_SEL                   EXT32k               (1)
Quoted 5 lines

LOSC_AUTO_SWT_STA_REG[7000004]: 1

	EXT_LOSC_STA                   OK                   (0)
	LOSC_AUTO_SWT_PEND             NOEFF                (0)
	LOSC_SRC_SEL_STA               EXT32K               (1)
Quoted 5 lines

In my case LOSC_CTRL_REG has value 0x4010 and LOSC_AUTO_SWT_STA_REG has value 0x4, so there is issue with external crystal (it's missing) and RTC switched to internal one.

BTW, what's wrong with automatic switching? Why is it disabled?

It always was disabled on mainline (bit 14 was set to 0 even before my patch). H6 just probably has another extra undocummented bit, that's needed to disables it properly.

You probably don't want a glitch to switch your RTC from high-precision clock to a low precision one possibly without any indication in the userspace or a kernel log.

Regardless of all this, DTS needs to have a correct description of the HW, which means if RTC module is not connected to the 32.757kHz crystal/clock, clocks property should be empty.

regards,

	o.
Quoted 25 lines

Best regards,
Jernej

Quoted 18 lines

regards,

	o.
Quoted 15 lines
Quoted 4 lines

The real issue probably is that the mainline driver is missing this:

https://megous.com/git/linux/tree/drivers/rtc/rtc-sunxi.c?h=h6-4.9-bsp#n
650

Not sure what you mean by that. ext vs. int source selection?

Best regards,
Jernej

Quoted 3 lines

regards,

	o.
Signature · 3 lines
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
To view this discussion on the web, visit https://groups.google.com/d/msgid/linux-sunxi/20190824212746.a5pyilkrrvysjjbd%40core.my.home.
maintainer
#13 lore raw

Re: [linux-sunxi] [PATCH v2 2/3] rtc: sun6i: Add support for H6 RTC

Dne sobota, 24. avgust 2019 ob 23:27:46 CEST je Ondřej Jirman napisal(a):

Quoted 15 lines

Hello Jernej,

Quoted 12 lines · Jernej Škrabec
Quoted 11 lines

Visually?

That would explain why it doesn't work for you. The mainline RTC driver disables auto-switch feature, and if your board doesn't have a crystal for
LOSC, RTC will not generate a clock for the RTC.

H6's dtsi describes by default a situatiuon with external 32k crystal oscillator. See ext_osc32k node. That's incorrect for your board if it doesn't have the crystal. You need to fix this in the DTS for your board instead of patching the driver.

Quoted 4 lines
Quoted 2 lines

I see that reparenting is supported, but I'm not sure how to fix that in DT. Any suggestion?

You may try removing the clocks property from rtc node.

I don't think this would work:
https://elixir.bootlin.com/linux/latest/source/drivers/rtc/rtc-sun6i.c#L246

Quoted 38 lines
Quoted 28 lines
Quoted 22 lines

The driver has parent clock selection logic in case the LOSC crystal is not
used.

Your patch enables automatic detection of LOSC failure and RTC changes clock to LOSC automatically, despite what's described in the DTS. That may fix the issue, but is not the correct solution.

Registers on my board look like this (external 32k osc is used) for reference:

LOSC_CTRL_REG[7000000]: 8011

	KEY_FIELD                      ???                  (0)
	LOSC_AUTO_SWT_BYPASS           EN                   (1)
	LOSC_AUTO_SWT_EN               DIS                  (0)
	EXT_LOSC_EN                    EN                   (1)
	EXT_LOSC_GSM                   LOW                  (0)
	BATTERY_DIR                    DISCHARGE            (0)
	LOSC_SRC_SEL                   EXT32k               (1)
Quoted 5 lines

LOSC_AUTO_SWT_STA_REG[7000004]: 1

	EXT_LOSC_STA                   OK                   (0)
	LOSC_AUTO_SWT_PEND             NOEFF                (0)
	LOSC_SRC_SEL_STA               EXT32K               (1)
Quoted 5 lines

In my case LOSC_CTRL_REG has value 0x4010 and LOSC_AUTO_SWT_STA_REG has value 0x4, so there is issue with external crystal (it's missing) and RTC switched to internal one.

BTW, what's wrong with automatic switching? Why is it disabled?

It always was disabled on mainline (bit 14 was set to 0 even before my patch). H6 just probably has another extra undocummented bit, that's needed to disables it properly.

Quoted 7 lines

You probably don't want a glitch to switch your RTC from high-precision clock to a low precision one possibly without any indication in the userspace or a kernel log.

Regardless of all this, DTS needs to have a correct description of the HW, which means if RTC module is not connected to the 32.757kHz crystal/clock, clocks property should be empty.

If we are talking about correct HW description, then clock property should actually have possibility that two clocks are defined - one for internal RC (always present) and one external crystal (optional). In such case I could really just omit external clock and be done with it. But I'm not sure if such solution is acceptable at this point.

Best regards,
Jernej

Quoted 27 lines

regards,

	o.
Quoted 23 lines

Best regards,
Jernej

Quoted 20 lines

regards,

	o.
Quoted 16 lines
Quoted 5 lines

The real issue probably is that the mainline driver is missing this:

https://megous.com/git/linux/tree/drivers/rtc/rtc-sunxi.c?h=h6-4.9-b
sp#n
650

Not sure what you mean by that. ext vs. int source selection?

Best regards,
Jernej

Quoted 3 lines

regards,

	o.
Signature · 3 lines
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
To view this discussion on the web, visit https://groups.google.com/d/msgid/linux-sunxi/7913281.jYEbquIlsS%40jernej-laptop.
#14 lore raw

Re: [linux-sunxi] [PATCH v2 2/3] rtc: sun6i: Add support for H6 RTC

Quoted 16 lines · Jernej Škrabec

Dne sobota, 24. avgust 2019 ob 23:27:46 CEST je Ondřej Jirman napisal(a):

Quoted 15 lines

Hello Jernej,

Quoted 12 lines · Jernej Škrabec
Quoted 11 lines

Visually?

That would explain why it doesn't work for you. The mainline RTC driver disables auto-switch feature, and if your board doesn't have a crystal for
LOSC, RTC will not generate a clock for the RTC.

H6's dtsi describes by default a situatiuon with external 32k crystal oscillator. See ext_osc32k node. That's incorrect for your board if it doesn't have the crystal. You need to fix this in the DTS for your board instead of patching the driver.

Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

I see that reparenting is supported, but I'm not sure how to fix that in DT. Any suggestion?

You may try removing the clocks property from rtc node.

I don't think this would work:
https://elixir.bootlin.com/linux/latest/source/drivers/rtc/rtc-sun6i.c#L246

Well, I don't know. There has to be some way to make it work, since the code deals with it here:

https://elixir.bootlin.com/linux/latest/source/drivers/rtc/rtc-sun6i.c#L270

Number of parents for LOSC is calculated from the DT somehow. Maybne something to do with the #clock-cells property?

Sorry I can't be of more help here.

Quoted 42 lines
Quoted 42 lines
Quoted 28 lines
Quoted 22 lines

The driver has parent clock selection logic in case the LOSC crystal is not
used.

Your patch enables automatic detection of LOSC failure and RTC changes clock to LOSC automatically, despite what's described in the DTS. That may fix the issue, but is not the correct solution.

Registers on my board look like this (external 32k osc is used) for reference:

LOSC_CTRL_REG[7000000]: 8011

	KEY_FIELD                      ???                  (0)
	LOSC_AUTO_SWT_BYPASS           EN                   (1)
	LOSC_AUTO_SWT_EN               DIS                  (0)
	EXT_LOSC_EN                    EN                   (1)
	EXT_LOSC_GSM                   LOW                  (0)
	BATTERY_DIR                    DISCHARGE            (0)
	LOSC_SRC_SEL                   EXT32k               (1)
Quoted 5 lines

LOSC_AUTO_SWT_STA_REG[7000004]: 1

	EXT_LOSC_STA                   OK                   (0)
	LOSC_AUTO_SWT_PEND             NOEFF                (0)
	LOSC_SRC_SEL_STA               EXT32K               (1)
Quoted 5 lines

In my case LOSC_CTRL_REG has value 0x4010 and LOSC_AUTO_SWT_STA_REG has value 0x4, so there is issue with external crystal (it's missing) and RTC switched to internal one.

BTW, what's wrong with automatic switching? Why is it disabled?

It always was disabled on mainline (bit 14 was set to 0 even before my patch). H6 just probably has another extra undocummented bit, that's needed to disables it properly.

You probably don't want a glitch to switch your RTC from high-precision clock to a low precision one possibly without any indication in the userspace or a kernel log.

Quoted 8 lines
Quoted 3 lines

Regardless of all this, DTS needs to have a correct description of the HW, which means if RTC module is not connected to the 32.757kHz crystal/clock, clocks property should be empty.

If we are talking about correct HW description, then clock property should actually have possibility that two clocks are defined - one for internal RC (always present) and one external crystal (optional). In such case I could really just omit external clock and be done with it. But I'm not sure if such

Internal RC is thought to be part of the RTC module, so it's not defined as an input clock to the RTC module.

regards,
Ondrej

Quoted 35 lines

Best regards,
Jernej

Quoted 27 lines

regards,

	o.
Quoted 23 lines

Best regards,
Jernej

Quoted 20 lines

regards,

	o.
Quoted 16 lines
Quoted 5 lines

The real issue probably is that the mainline driver is missing this:

https://megous.com/git/linux/tree/drivers/rtc/rtc-sunxi.c?h=h6-4.9-b
sp#n
650

Not sure what you mean by that. ext vs. int source selection?

Best regards,
Jernej

Quoted 3 lines

regards,

	o.
#15 lore raw

PATCH v2 3/3 arm64: dts: sun50i-h6: Add support for RTC and fix the clock tree

In reply to [email protected]

From: Ondrej Jirman <megous-5qf/[email protected]>

This patch adds RTC node and fixes the clock properties and nodes to reflect the real clock tree.

The device nodes for the internal oscillator and osc32k are removed, as these clocks are now provided by the RTC device. Clock references are fixed accordingly, too.

---
 arch/arm64/boot/dts/allwinner/sun50i-h6.dtsi | 30 +++++++++++---------
 1 file changed, 16 insertions(+), 14 deletions(-)
arch/arm64/boot/dts/allwinner/sun50i-h6.dtsi +16 -14
index 67b732e34091..67f920e0fc33 100644
@@ -56,14 +56,6 @@
 		status = "disabled";
 	};
 
-	iosc: internal-osc-clk {
-		#clock-cells = <0>;
-		compatible = "fixed-clock";
-		clock-frequency = <16000000>;
-		clock-accuracy = <300000000>;
-		clock-output-names = "iosc";
-	};
-
 	osc24M: osc24M_clk {
 		#clock-cells = <0>;
 		compatible = "fixed-clock";
@@ -71,11 +63,11 @@
 		clock-output-names = "osc24M";
 	};
 
-	osc32k: osc32k_clk {
+	ext_osc32k: ext_osc32k_clk {
 		#clock-cells = <0>;
 		compatible = "fixed-clock";
 		clock-frequency = <32768>;
-		clock-output-names = "osc32k";
+		clock-output-names = "ext_osc32k";
 	};
 
 	psci {
@@ -197,7 +189,7 @@
 		ccu: clock@3001000 {
 			compatible = "allwinner,sun50i-h6-ccu";
 			reg = <0x03001000 0x1000>;
-			clocks = <&osc24M>, <&osc32k>, <&iosc>;
+			clocks = <&osc24M>, <&rtc 0>, <&rtc 2>;
 			clock-names = "hosc", "losc", "iosc";
 			#clock-cells = <1>;
 			#reset-cells = <1>;
@@ -236,7 +228,7 @@
 				     <GIC_SPI 53 IRQ_TYPE_LEVEL_HIGH>,
 				     <GIC_SPI 54 IRQ_TYPE_LEVEL_HIGH>,
 				     <GIC_SPI 59 IRQ_TYPE_LEVEL_HIGH>;
-			clocks = <&ccu CLK_APB1>, <&osc24M>, <&osc32k>;
+			clocks = <&ccu CLK_APB1>, <&osc24M>, <&rtc 0>;
 			clock-names = "apb", "hosc", "losc";
 			gpio-controller;
 			#gpio-cells = <3>;
@@ -710,10 +702,20 @@
 			};
 		};
 
+		rtc: rtc@7000000 {
+			compatible = "allwinner,sun50i-h6-rtc";
+			reg = <0x07000000 0x400>;
+			interrupts = <GIC_SPI 101 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 102 IRQ_TYPE_LEVEL_HIGH>;
+			clock-output-names = "osc32k", "osc32k-out", "iosc";
+			clocks = <&ext_osc32k>;
+			#clock-cells = <1>;
+		};
+
 		r_ccu: clock@7010000 {
 			compatible = "allwinner,sun50i-h6-r-ccu";
 			reg = <0x07010000 0x400>;
-			clocks = <&osc24M>, <&osc32k>, <&iosc>,
+			clocks = <&osc24M>, <&rtc 0>, <&rtc 2>,
 				 <&ccu CLK_PLL_PERIPH0>;
 			clock-names = "hosc", "losc", "iosc", "pll-periph";
 			#clock-cells = <1>;
@@ -741,7 +743,7 @@
 			reg = <0x07022000 0x400>;
 			interrupts = <GIC_SPI 105 IRQ_TYPE_LEVEL_HIGH>,
 				     <GIC_SPI 111 IRQ_TYPE_LEVEL_HIGH>;
-			clocks = <&r_ccu CLK_R_APB1>, <&osc24M>, <&osc32k>;
+			clocks = <&r_ccu CLK_R_APB1>, <&osc24M>, <&rtc 0>;
 			clock-names = "apb", "hosc", "losc";
 			gpio-controller;
 			#gpio-cells = <3>;
Signature · 1 line
2.22.1
#16 lore raw

Re: [PATCH v2 3/3] arm64: dts: sun50i-h6: Add support for RTC and fix the clock tree

Quoted 10 lines · On Tue, Aug 20, 2019 at 05:19:34PM +0200…

From: Ondrej Jirman <megous-5qf/[email protected]>

This patch adds RTC node and fixes the clock properties and nodes to reflect the real clock tree.

The device nodes for the internal oscillator and osc32k are removed, as these clocks are now provided by the RTC device. Clock references are fixed accordingly, too.

The prefix should be "arm64: dts: allwinner: h6:"

I've fixed it up and applied it.

Maxime

--
Maxime Ripard, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com

maintainer
#17 lore raw
In reply to [email protected]

Hi!

Dne torek, 20. avgust 2019 ob 17:19:31 CEST je megous-5qf/[email protected] napisal(a):

Quoted 18 lines

From: Ondrej Jirman <megous-5qf/[email protected]>

I went through the datasheets for H6 and H5, and compared the differences. RTCs are largely similar, but not entirely compatible. Incompatibilities are in details not yet implemented by the rtc driver though.

I also corrected the clock tree in H6 DTSI.

This patchset is necessary for implementing the WiFi/Bluetooth support on boards using H6 SoC.

There was some discussion previously of describing HOSC, DCXO and XO oscillators and clocks as part of RTC in DT, but I decided against it because it's not necessary, becuse information that would be provided as a part of DT can already be determined at runtime from RTC registers, so this woudn't add any value and would only introduce complications to the driver. See: https://patchwork.kernel.org/cover/10898083/

Quoted 5 lines

Please take a look.

Thank you and regards,
Ondrej Jirman

Sorry for a bit late test, but with your patches on Tanix TX6 box I get this in dmesg:

[   17.431742] sun6i-rtc 7000000.rtc: Failed to set rtc time.
[   20.439742] sun6i-rtc 7000000.rtc: rtc is still busy.
[   21.435744] sun6i-rtc 7000000.rtc: rtc is still busy.
[   24.055741] sun6i-rtc 7000000.rtc: rtc is still busy.
[   24.439752] sun6i-rtc 7000000.rtc: rtc is still busy.

Last line is repeated non-stop.

Any idea what could be wrong?

Best regards,
Jernej

maintainer
#18 lore raw

Dne sobota, 24. avgust 2019 ob 10:04:24 CEST je Jernej Škrabec napisal(a):

Quoted 37 lines

Hi!

Dne torek, 20. avgust 2019 ob 17:19:31 CEST je megous-5qf/[email protected] napisal(a):

Quoted 18 lines

From: Ondrej Jirman <megous-5qf/[email protected]>

I went through the datasheets for H6 and H5, and compared the differences. RTCs are largely similar, but not entirely compatible. Incompatibilities are in details not yet implemented by the rtc driver though.

I also corrected the clock tree in H6 DTSI.

This patchset is necessary for implementing the WiFi/Bluetooth support on boards using H6 SoC.

There was some discussion previously of describing HOSC, DCXO and XO oscillators and clocks as part of RTC in DT, but I decided against it because it's not necessary, becuse information that would be provided as a part of DT can already be determined at runtime from RTC registers, so this woudn't add any value and would only introduce complications to the driver. See: https://patchwork.kernel.org/cover/10898083/

Quoted 6 lines

Please take a look.

Thank you and regards,

Ondrej Jirman

Sorry for a bit late test, but with your patches on Tanix TX6 box I get this in dmesg:

[   17.431742] sun6i-rtc 7000000.rtc: Failed to set rtc time.
[   20.439742] sun6i-rtc 7000000.rtc: rtc is still busy.
[   21.435744] sun6i-rtc 7000000.rtc: rtc is still busy.
[   24.055741] sun6i-rtc 7000000.rtc: rtc is still busy.
[   24.439752] sun6i-rtc 7000000.rtc: rtc is still busy.
Quoted 3 lines

Last line is repeated non-stop.

Any idea what could be wrong?

Additional info - this is on kernel 5.2.6 with your patches applied.

Best regards,
Jernej

Signature · 3 lines
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
To view this discussion on the web, visit https://groups.google.com/d/msgid/linux-sunxi/5421621.t8Lore9UF7%40jernej-laptop.
#19 lore raw

Hi,

Quoted 38 lines · Jernej Škrabec

Dne sobota, 24. avgust 2019 ob 10:04:24 CEST je Jernej Škrabec napisal(a):

Quoted 37 lines

Hi!

Dne torek, 20. avgust 2019 ob 17:19:31 CEST je megous-5qf/[email protected] napisal(a):

Quoted 18 lines

From: Ondrej Jirman <megous-5qf/[email protected]>

I went through the datasheets for H6 and H5, and compared the differences. RTCs are largely similar, but not entirely compatible. Incompatibilities are in details not yet implemented by the rtc driver though.

I also corrected the clock tree in H6 DTSI.

This patchset is necessary for implementing the WiFi/Bluetooth support on boards using H6 SoC.

There was some discussion previously of describing HOSC, DCXO and XO oscillators and clocks as part of RTC in DT, but I decided against it because it's not necessary, becuse information that would be provided as a part of DT can already be determined at runtime from RTC registers, so this woudn't add any value and would only introduce complications to the driver. See: https://patchwork.kernel.org/cover/10898083/

Quoted 6 lines

Please take a look.

Thank you and regards,

Ondrej Jirman

Sorry for a bit late test, but with your patches on Tanix TX6 box I get this in dmesg:

[   17.431742] sun6i-rtc 7000000.rtc: Failed to set rtc time.
[   20.439742] sun6i-rtc 7000000.rtc: rtc is still busy.
[   21.435744] sun6i-rtc 7000000.rtc: rtc is still busy.
[   24.055741] sun6i-rtc 7000000.rtc: rtc is still busy.
[   24.439752] sun6i-rtc 7000000.rtc: rtc is still busy.
Quoted 5 lines
Quoted 3 lines

Last line is repeated non-stop.

Any idea what could be wrong?

Additional info - this is on kernel 5.2.6 with your patches applied.

Do you have schematics, or a FEX file for the board or any other info on how the RTC is set up on that board?

Can you dump the RTC register range?

regards,

	o.
Quoted 11 lines

Best regards,
Jernej

_______________________________________________
linux-arm-kernel mailing list linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Signature · 3 lines
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
To view this discussion on the web, visit https://groups.google.com/d/msgid/linux-sunxi/20190824125612.zq5qsay2wv62wykt%40core.my.home.
maintainer
#20 lore raw

Dne sobota, 24. avgust 2019 ob 14:56:12 CEST je Ondřej Jirman napisal(a):

Quoted 7 lines

Hi,

Quoted 4 lines · Jernej Škrabec

Dne sobota, 24. avgust 2019 ob 10:04:24 CEST je Jernej Škrabec napisal(a):

Quoted 3 lines

Hi!

Dne torek, 20. avgust 2019 ob 17:19:31 CEST je megous-5qf/[email protected]

napisal(a):

Quoted 39 lines
Quoted 39 lines
Quoted 39 lines
Quoted 21 lines

From: Ondrej Jirman <megous-5qf/[email protected]>

I went through the datasheets for H6 and H5, and compared the differences.
RTCs are largely similar, but not entirely compatible. Incompatibilities
are in details not yet implemented by the rtc driver though.

I also corrected the clock tree in H6 DTSI.

This patchset is necessary for implementing the WiFi/Bluetooth support on boards using H6 SoC.

There was some discussion previously of describing HOSC, DCXO and XO oscillators and clocks as part of RTC in DT, but I decided against it because it's not necessary, becuse information that would be provided as a part of DT can already be determined at runtime from RTC registers,
so this woudn't add any value and would only introduce complications to the driver. See: https://patchwork.kernel.org/cover/10898083/

Quoted 6 lines

Please take a look.

Thank you and regards,

Ondrej Jirman

Sorry for a bit late test, but with your patches on Tanix TX6 box I get this in dmesg:

[   17.431742] sun6i-rtc 7000000.rtc: Failed to set rtc time.
[   20.439742] sun6i-rtc 7000000.rtc: rtc is still busy.
[   21.435744] sun6i-rtc 7000000.rtc: rtc is still busy.
[   24.055741] sun6i-rtc 7000000.rtc: rtc is still busy.
[   24.439752] sun6i-rtc 7000000.rtc: rtc is still busy.

Last line is repeated non-stop.

Quoted 8 lines
Quoted 3 lines
Quoted 1 line

Any idea what could be wrong?

Additional info - this is on kernel 5.2.6 with your patches applied.

Do you have schematics, or a FEX file for the board or any other info on how the RTC is set up on that board?

Can you dump the RTC register range?

I have only Android DT, but as I already said in latest reply to patch 2, changing switch bypass to bit 14 instead of 15 solved all issues.

Best regards,
Jernej

Quoted 15 lines

regards,

	o.
Quoted 11 lines

Best regards,
Jernej

_______________________________________________
linux-arm-kernel mailing list linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Signature · 3 lines
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
To view this discussion on the web, visit https://groups.google.com/d/msgid/linux-sunxi/2660592.K6CN9dZ0QN%40jernej-laptop.