Skip to contentSkip to conversation

Add the Allwinner A31/A31s PWM driver

PATCH Superseded [email protected] 2 participants 9 messages Feb 15, 2017 – Feb 21, 2017 4 messages not in the archive

Subsystems (3)

v3 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

v3 cover Add the Allwinner A31/A31s PWM driver

From: Siarhei Volkau <[email protected]>

Main goal is to add support of the sun6i PWM variant.

This is the third version of the sun6i PWM patchset. Difference between v2 and v3:
- regmap API used for access to PWM registers.

Thanks to Maxime Ripard for pointing me to the regmap API.

Siarhei Volkau (4):
pwm: sunxi: Code switched to regmap API instead of iomem.
pwm: sunxi: Selectable prescaler table for support sun6i.
pwm: sunxi: Add support the Allwinner A31 PWM.
ARM: dts: sun6i: Add the PWM block to the A31/A31s.

 .../devicetree/bindings/pwm/pwm-sun4i.txt          |   3 +-
 arch/arm/boot/dts/sun6i-a31.dtsi                   |   8 +
 drivers/pwm/Kconfig                                |   2 +-
 drivers/pwm/pwm-sun4i.c                            | 359 ++++++++++++++++-----
 4 files changed, 294 insertions(+), 78 deletions(-)

--
2.4.11

A message that is not in the archive (sent off-list or not yet received): 1487189167-32486-1-git-send-email-lis8215-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org

#2 lore raw

PATCH v3 1/4 pwm: sunxi: Code switched to regmap API instead of iomem.

From: Siarhei Volkau <[email protected]>

sun6i PWM has different register map in comparison to sun4i compatible SoCs. But bit map of the registers and behavior are very similar.

This patch introduces a uniform way to access PWM registers.

---
 drivers/pwm/Kconfig     |   2 +-
 drivers/pwm/pwm-sun4i.c | 263 ++++++++++++++++++++++++++++++++++--------------
 2 files changed, 191 insertions(+), 74 deletions(-)
drivers/pwm/Kconfig +1 -1
index 2d0cfaa..6b4dc1a 100644
@@ -416,7 +416,7 @@ config PWM_STMPE
 config PWM_SUN4I
 	tristate "Allwinner PWM support"
 	depends on ARCH_SUNXI || COMPILE_TEST
-	depends on HAS_IOMEM && COMMON_CLK
+	depends on REGMAP_MMIO && COMMON_CLK
 	help
 	  Generic PWM framework driver for Allwinner SoCs.
 
drivers/pwm/pwm-sun4i.c +190 -73

428 lines not on this page · raw message

--
2.4.11

A message that is not in the archive (sent off-list or not yet received): 1487189167-32486-2-git-send-email-lis8215-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org

#4 lore raw

Re: [PATCH v3 1/4] pwm: sunxi: Code switched to regmap API instead of iomem.

Hi,

Quoted 8 lines · On Wed, Feb 15, 2017 at 11:06:04PM +0300…

From: Siarhei Volkau <[email protected]>

sun6i PWM has different register map in comparison to sun4i compatible SoCs. But bit map of the registers and behavior are very similar.

This patch introduces a uniform way to access PWM registers.

In order to be easier to review (and bisect if needed), could you split that patch into one to convert to the regmap API, without any change but to replace the sun4i_pwm_readl/sun4i_pwm_writel calls by their regmap equivalent, and then convert to the regmap fields the registers that need to?

Quoted 85 lines
---
 drivers/pwm/Kconfig     |   2 +-
 drivers/pwm/pwm-sun4i.c | 263 ++++++++++++++++++++++++++++++++++--------------
 2 files changed, 191 insertions(+), 74 deletions(-)
drivers/pwm/Kconfig +1 -1
index 2d0cfaa..6b4dc1a 100644
@@ -416,7 +416,7 @@ config PWM_STMPE
 config PWM_SUN4I
 	tristate "Allwinner PWM support"
 	depends on ARCH_SUNXI || COMPILE_TEST
-	depends on HAS_IOMEM && COMMON_CLK
+	depends on REGMAP_MMIO && COMMON_CLK
 	help
 	  Generic PWM framework driver for Allwinner SoCs.
 
drivers/pwm/pwm-sun4i.c +32 -11
index b0803f6..7291000 100644
@@ -9,7 +9,7 @@
 #include <linux/bitops.h>
 #include <linux/clk.h>
 #include <linux/err.h>
-#include <linux/io.h>
+#include <linux/regmap.h>
 #include <linux/module.h>
 #include <linux/of.h>
 #include <linux/of_device.h>
@@ -26,25 +26,56 @@
 #define PWM_CH_PRD(ch)		(PWM_CH_PRD_BASE + PWM_CH_PRD_OFFSET * (ch))
 
 #define PWMCH_OFFSET		15
-#define PWM_PRESCAL_MASK	GENMASK(3, 0)
-#define PWM_PRESCAL_OFF		0
-#define PWM_EN			BIT(4)
-#define PWM_ACT_STATE		BIT(5)
-#define PWM_CLK_GATING		BIT(6)
-#define PWM_MODE		BIT(7)
-#define PWM_PULSE		BIT(8)
-#define PWM_BYPASS		BIT(9)
+
+#define PWM_PRESCAL_LSB		0
+#define PWM_PRESCAL_MSB		3
+#define PWM_PRESCAL_MASK	GENMASK(PWM_PRESCAL_MSB - PWM_PRESCAL_LSB, 0)
+
+#define PWM_EN_BIT		4
+#define PWM_ACT_STATE_BIT	5
+#define PWM_CLK_GATING_BIT	6
+#define PWM_MODE_BIT		7
+#define PWM_PULSE_BIT		8
+#define PWM_BYPASS_BIT		9
 
 #define PWM_RDY_BASE		28
 #define PWM_RDY_OFFSET		1
-#define PWM_RDY(ch)		BIT(PWM_RDY_BASE + PWM_RDY_OFFSET * (ch))
+#define PWM_RDY_BIT(ch)		(PWM_RDY_BASE + PWM_RDY_OFFSET * (ch))
 
 #define PWM_PRD(prd)		(((prd) - 1) << 16)
 #define PWM_PRD_MASK		GENMASK(15, 0)
 
 #define PWM_DTY_MASK		GENMASK(15, 0)
 
-#define BIT_CH(bit, chan)	((bit) << ((chan) * PWMCH_OFFSET))
+#define BIT_CH(bit, chan)	((bit) + ((chan) * PWMCH_OFFSET))
+
+#define FIELD_PRESCALER		0
+#define FIELD_POLARITY		1
+#define FIELD_CLK_GATING	2
+#define FIELD_READY		3
+#define NUM_FIELDS		4
+
+#define MAX_CHANNELS		2
+
+#define SUN4I_REGMAP_FIELDS(chan) {\
+	[FIELD_PRESCALER] = \
+		REG_FIELD(PWM_CTRL_REG, \
+			  BIT_CH(PWM_PRESCAL_LSB, chan), \
+			  BIT_CH(PWM_PRESCAL_MSB, chan)), \
+	[FIELD_POLARITY] = \
+		REG_FIELD(PWM_CTRL_REG, \
+			  BIT_CH(PWM_ACT_STATE_BIT, chan), \
+			  BIT_CH(PWM_ACT_STATE_BIT, chan)), \

drivers/pwm/pwm-sun4i.c line 78

Quoted 10 lines
drivers/pwm/pwm-sun4i.c +10 -0
+	[FIELD_CLK_GATING] = \
+		REG_FIELD(PWM_CTRL_REG, \
+			  BIT_CH(PWM_CLK_GATING_BIT, chan), \
+			  BIT_CH(PWM_CLK_GATING_BIT, chan)), \
+	[FIELD_READY] = \
+		REG_FIELD(PWM_CTRL_REG, \
+			  PWM_RDY_BIT(chan), \
+			  PWM_RDY_BIT(chan)), \
+}
+

This is not correct, unfortunately. If someone calls that macro with chan++ or chan--, it will be modified four times instead of one as the caller would expect.

Thanks!
Maxime

Signature · 3 lines
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com
#5 lore raw

Re: [PATCH v3 1/4] pwm: sunxi: Code switched to regmap API instead of iomem.

Hello,

Quoted 5 lines

In order to be easier to review (and bisect if needed), could you split that patch into one to convert to the regmap API, without any change but to replace the sun4i_pwm_readl/sun4i_pwm_writel calls by their regmap equivalent, and then convert to the regmap fields the registers that need to?

Looks reasonable, will be done shortly.

Quoted 15 lines
Quoted 15 lines
drivers/pwm/pwm-sun4i.c +15 -0
+#define SUN4I_REGMAP_FIELDS(chan) {\
+     [FIELD_PRESCALER] = \
+             REG_FIELD(PWM_CTRL_REG, \
+                       BIT_CH(PWM_PRESCAL_LSB, chan), \
+                       BIT_CH(PWM_PRESCAL_MSB, chan)), \
+     [FIELD_POLARITY] = \
+             REG_FIELD(PWM_CTRL_REG, \
+                       BIT_CH(PWM_ACT_STATE_BIT, chan), \
+                       BIT_CH(PWM_ACT_STATE_BIT, chan)), \
+     [FIELD_CLK_GATING] = \
+             REG_FIELD(PWM_CTRL_REG, \
+                       BIT_CH(PWM_CLK_GATING_BIT, chan), \
+                       BIT_CH(PWM_CLK_GATING_BIT, chan)), \
+     [FIELD_READY] = \
+             REG_FIELD(PWM_CTRL_REG, \
Quoted 8 lines

drivers/pwm/pwm-sun4i.c line 78

Quoted 4 lines
drivers/pwm/pwm-sun4i.c +4 -0
+                       PWM_RDY_BIT(chan), \
+                       PWM_RDY_BIT(chan)), \
+}
+

This is not correct, unfortunately. If someone calls that macro with chan++ or chan--, it will be modified four times instead of one as the caller would expect.

Ok, i will refactor that.
I think you should know about existence of unsafe macros (like MIN MAX or ABS) into existing linux kernel tree, this macros can be easily refactored into the inline functions. Please tell me your point of view on that things.

Thanks,
Siarhei.

A message that is not in the archive (sent off-list or not yet received): CAKNVLfa-doz_6oARF5Eg1EbKFuita34Pg2x6Jutjzt9fshUEwQ-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org

#7 lore raw

Re: [PATCH v3 1/4] pwm: sunxi: Code switched to regmap API instead of iomem.

Hi,

Quoted 24 lines · Siarhei Volkau
Quoted 15 lines
Quoted 15 lines
drivers/pwm/pwm-sun4i.c +15 -0
+#define SUN4I_REGMAP_FIELDS(chan) {\
+     [FIELD_PRESCALER] = \
+             REG_FIELD(PWM_CTRL_REG, \
+                       BIT_CH(PWM_PRESCAL_LSB, chan), \
+                       BIT_CH(PWM_PRESCAL_MSB, chan)), \
+     [FIELD_POLARITY] = \
+             REG_FIELD(PWM_CTRL_REG, \
+                       BIT_CH(PWM_ACT_STATE_BIT, chan), \
+                       BIT_CH(PWM_ACT_STATE_BIT, chan)), \
+     [FIELD_CLK_GATING] = \
+             REG_FIELD(PWM_CTRL_REG, \
+                       BIT_CH(PWM_CLK_GATING_BIT, chan), \
+                       BIT_CH(PWM_CLK_GATING_BIT, chan)), \
+     [FIELD_READY] = \
+             REG_FIELD(PWM_CTRL_REG, \
Quoted 8 lines

drivers/pwm/pwm-sun4i.c line 78

Quoted 4 lines
drivers/pwm/pwm-sun4i.c +4 -0
+                       PWM_RDY_BIT(chan), \
+                       PWM_RDY_BIT(chan)), \
+}
+

This is not correct, unfortunately. If someone calls that macro with chan++ or chan--, it will be modified four times instead of one as the caller would expect.

Quoted 5 lines

Ok, i will refactor that.
I think you should know about existence of unsafe macros (like MIN MAX or ABS) into existing linux kernel tree, this macros can be easily refactored into the inline functions. Please tell me your point of view on that things.

This might be bad, but we also have functions for those, which should behave properly. Anyway, I don't really expect you to fix the whole kernel before taking that patch in :)

Maxime

Signature · 3 lines
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com
#8 lore raw

PATCH v3 2/4 pwm: sunxi: Selectable prescaler table for support sun6i.

In reply to a message not in the archive

From: Siarhei Volkau <[email protected]>

This patch not introduce new features, just prepare code for adding sun6i PWM driver in next commit.

A31 SoC have a different set of prescalers than sun4i compatible ASoCs, but its position and count in the control register are the same.

This patch make the table of prescalers customizable.

---
 drivers/pwm/pwm-sun4i.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
drivers/pwm/pwm-sun4i.c +8 -1
index 7291000..b4706af 100644
@@ -77,7 +77,7 @@
 }
 
 
-static const u32 prescaler_table[] = {
+static const u32 sun4i_prescaler_table[] = {
 	120,
 	180,
 	240,
@@ -107,6 +107,7 @@ struct sun4i_pwm_data {
 	bool has_prescaler_bypass;
 	bool has_rdy;
 	unsigned int npwm;
+	const u32 *prescaler_table;
 	const struct sunxi_pwmch_reg_info *chan_info;
 };
 
@@ -129,6 +130,7 @@ static int sun4i_pwm_config(struct pwm_chip *chip, struct pwm_device *pwm,
 {
 	struct sun4i_pwm_chip *sun4i_pwm = to_sun4i_pwm_chip(chip);
 	struct regmap_field **chan_fields = sun4i_pwm->fields[pwm->hwpwm];
+	const u32 *prescaler_table = sun4i_pwm->data->prescaler_table;
 	u32 prd, dty, busy, clk_gate;
 	u64 clk_rate, div = 0;
 	unsigned int prescaler = 0;
@@ -328,6 +330,7 @@ static const struct sun4i_pwm_data sun4i_pwm_data_a10 = {
 	.has_prescaler_bypass = false,
 	.has_rdy = false,
 	.npwm = 2,
+	.prescaler_table = sun4i_prescaler_table,
 	.chan_info = sun4i_field_info,
 };
 
@@ -335,6 +338,7 @@ static const struct sun4i_pwm_data sun4i_pwm_data_a10s = {
 	.has_prescaler_bypass = true,
 	.has_rdy = true,
 	.npwm = 2,
+	.prescaler_table = sun4i_prescaler_table,
 	.chan_info = sun4i_field_info,
 };
 
@@ -342,6 +346,7 @@ static const struct sun4i_pwm_data sun4i_pwm_data_a13 = {
 	.has_prescaler_bypass = true,
 	.has_rdy = true,
 	.npwm = 1,
+	.prescaler_table = sun4i_prescaler_table,
 	.chan_info = sun4i_field_info,
 };
 
@@ -349,6 +354,7 @@ static const struct sun4i_pwm_data sun4i_pwm_data_a20 = {
 	.has_prescaler_bypass = true,
 	.has_rdy = true,
 	.npwm = 2,
+	.prescaler_table = sun4i_prescaler_table,
 	.chan_info = sun4i_field_info,
 };
 
@@ -356,6 +362,7 @@ static const struct sun4i_pwm_data sun4i_pwm_data_h3 = {
 	.has_prescaler_bypass = true,
 	.has_rdy = true,
 	.npwm = 1,
+	.prescaler_table = sun4i_prescaler_table,
 	.chan_info = sun4i_field_info,
 };
 

--
2.4.11

A message that is not in the archive (sent off-list or not yet received): 1487189167-32486-3-git-send-email-lis8215-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org

Acked-by
#10 lore raw

Re: [PATCH v3 2/4] pwm: sunxi: Selectable prescaler table for support sun6i.

Quoted 5 lines · On Wed, Feb 15, 2017 at 11:06:05PM +0300…

From: Siarhei Volkau <[email protected]>

This patch not introduce new features, just prepare code for adding sun6i PWM driver in next commit.

Quoted 7 lines

A31 SoC have a different set of prescalers than sun4i compatible ASoCs, but its position and count in the control register are the same.

This patch make the table of prescalers customizable.

Thanks,
Maxime

Signature · 3 lines
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com
#11 lore raw

PATCH v3 3/4 pwm: sunxi: Add support the Allwinner A31 PWM.

In reply to a message not in the archive

From: Siarhei Volkau <[email protected]>

This patch introduce the sun6i PWM driver itself:
- sun6i channels configuration,
- sun6i prescaler table,
- DT bindings for A31 SoC,
- documentation update,
- module description and author update.

---
 .../devicetree/bindings/pwm/pwm-sun4i.txt          |  3 +-
 drivers/pwm/pwm-sun4i.c                            | 89 +++++++++++++++++++++-
 2 files changed, 88 insertions(+), 4 deletions(-)
Documentation/devicetree/bindings/pwm/pwm-sun4i.txt +2 -1
index f1cbeef..109b997 100644
@@ -1,10 +1,11 @@
-Allwinner sun4i and sun7i SoC PWM controller
+Allwinner sunxi SoC PWM controller
 
 Required properties:
   - compatible: should be one of:
     - "allwinner,sun4i-a10-pwm"
     - "allwinner,sun5i-a10s-pwm"
     - "allwinner,sun5i-a13-pwm"
+    - "allwinner,sun6i-a31-pwm"
     - "allwinner,sun7i-a20-pwm"
     - "allwinner,sun8i-h3-pwm"
   - reg: physical base address and length of the controller's registers
drivers/pwm/pwm-sun4i.c +86 -3

141 lines not on this page · raw message

--
2.4.11

#12 lore raw

PATCH v3 4/4 ARM: dts: sun6i: Add the PWM block to the A31/A31s.

In reply to a message not in the archive

From: Siarhei Volkau <[email protected]>

---
 arch/arm/boot/dts/sun6i-a31.dtsi | 8 ++++++++
 1 file changed, 8 insertions(+)
arch/arm/boot/dts/sun6i-a31.dtsi +8 -0
index ee1eb6d..fcba129 100644
@@ -627,6 +627,14 @@
 			status = "disabled";
 		};
 
+		pwm: pwm@01c21400 {
+			compatible = "allwinner,sun6i-a31-pwm";
+			reg = <0x01c21400 0x400>;
+			clocks = <&osc24M>;
+			#pwm-cells = <3>;
+			status = "disabled";
+		};
+
 		lradc: lradc@01c22800 {
 			compatible = "allwinner,sun4i-a10-lradc-keys";
 			reg = <0x01c22800 0x100>;
Signature · 1 line
2.4.11