ARM: shmobile: sh-eth pins in DT for armadillo800eva

r8a7740 SoCs have an integrated gigabit ethernet MAC. This patch adds two pin groups: base for 100mbps and additional pins for the gigabit mode.

Signed-off-by Guennadi Liakhovetski <[email protected]>

---
 drivers/pinctrl/sh-pfc/pfc-r8a7740.c |   31 +++++++++++++++++++++++++++++++
1 files changed, 31 insertions(+), 0 deletions(-)

(quoted diff) +0 -0

b/drivers/pinctrl/sh-pfc/pfc-r8a7740.c index 4c92af8..d7f69dd 100644

drivers/pinctrl/sh-pfc/pfc-r8a7740.c +25 -0

@@ -2052,6 +2052,29 @@ static const unsigned int mmc0_ctrl_1_mux[] = {
 	MMC1_CMD_PORT104_MARK, MMC1_CLK_PORT103_MARK,
 };
 
+static const unsigned int eth_base_pins[] = {
+	203, 204, 205, 206, 207,
+	183, 184,
+	185, 186, 187, 188,
+	171, 170, 169, 168,
+	161, 163, 164, 174,
+};
+static const unsigned int eth_base_mux[] = {
+	ET_TX_ER_MARK, ET_RX_ER_MARK, ET_CRS_MARK, ET_MDC_MARK, ET_MDIO_MARK,
+	ET_TX_EN_MARK, ET_TX_CLK_MARK,
+	ET_ERXD0_MARK, ET_ERXD1_MARK, ET_ERXD2_MARK, ET_ERXD3_MARK,
+	ET_ETXD0_MARK, ET_ETXD1_MARK, ET_ETXD2_MARK, ET_ETXD3_MARK,
+	ET_RX_DV_MARK, ET_COL_MARK, ET_PHY_INT_MARK, ET_RX_CLK_MARK,
+};
+static const unsigned int eth_gbit_pins[] = {
+	189, 190, 191, 192,
+	167, 166, 173, 172,
+};
+static const unsigned int eth_gbit_mux[] = {
+	ET_ERXD4_MARK, ET_ERXD5_MARK, ET_ERXD6_MARK, ET_ERXD7_MARK,
+	ET_ETXD4_MARK, ET_ETXD5_MARK, ET_ETXD6_MARK, ET_ETXD7_MARK,
+};
+
 static const struct sh_pfc_pin_group pinmux_groups[] = {
 	SH_PFC_PIN_GROUP(lcd0_data8),
 	SH_PFC_PIN_GROUP(lcd0_data9),

@@ -2100,6 +2123,8 @@ static const struct sh_pfc_pin_group pinmux_groups[] =
 { SH_PFC_PIN_GROUP(mmc0_data4_1),
 	SH_PFC_PIN_GROUP(mmc0_data8_1),
 	SH_PFC_PIN_GROUP(mmc0_ctrl_1),
+	SH_PFC_PIN_GROUP(eth_base),
+	SH_PFC_PIN_GROUP(eth_gbit),
 };
 
 static const char * const lcd0_groups[] = {

drivers/pinctrl/sh-pfc/pfc-r8a7740.c +1 -0

 static const struct sh_pfc_function pinmux_functions[] = {
 	SH_PFC_FUNCTION(lcd0),
 	SH_PFC_FUNCTION(lcd1),

@@ -2174,6 +2204,7 @@ static const struct sh_pfc_function pinmux_functions[]
 = { SH_PFC_FUNCTION(sdhi1),
 	SH_PFC_FUNCTION(sdhi2),
 	SH_PFC_FUNCTION(mmc0),
+	SH_PFC_FUNCTION(eth),
 };
 
 #define PINMUX_FN_BASE	ARRAY_SIZE(pinmux_pins)

If an ethernet PHY can be reset by a GPIO, it can be specified in DT. Add a binding and code to parse it, request the GPIO and take the PHY out of reset.

Signed-off-by Guennadi Liakhovetski <[email protected]>

Cc devicetree-discuss at lists.ozlabs.org

Cc netdev at vger.kernel.org

---
 Documentation/devicetree/bindings/net/sh_ether.txt |    2 ++
 drivers/net/ethernet/renesas/sh_eth.c              |    9 ++++++++-
 2 files changed, 10 insertions(+), 1 deletions(-)

(quoted diff) +0 -0

b/Documentation/devicetree/bindings/net/sh_ether.txt index c11e45d..edaf683 100644

Documentation/devicetree/bindings/net/sh_ether.txt +1 -0

 - sh-eth,edmac-endian:          String, endian of sh_eth dmac.
 - sh-eth,register-type:         String, register type of sh_eth.
                                 Please select "gigabit", "fast-sh4" or

@@ -37,6 +38,7 @@ Example (armadillo800eva):
 		reg = <0xe9a00000 0x800>, <0xe9a01800 0x800>;
 		interrupts = <0x500>;
 		phy-mode = "mii";
+		phy-reset-gpios = <&gpio 18 0>;
 		sh-eth,edmac-endian = "little";
 		sh-eth,register-type = "gigabit";
 		sh-eth,phy-id = <0>;

(quoted diff) +0 -0

b/drivers/net/ethernet/renesas/sh_eth.c index 1f64848..06035a2 100644

drivers/net/ethernet/renesas/sh_eth.c +4 -1

@@ -20,6 +20,7 @@
  *  the file called "COPYING".
  */
 
+#include <linux/gpio.h>
 #include <linux/init.h>
 #include <linux/module.h>
 #include <linux/kernel.h>

@@ -33,6 +34,7 @@
 #include <linux/netdevice.h>
 #include <linux/of.h>
 #include <linux/of_device.h>
+#include <linux/of_gpio.h>
 #include <linux/of_platform.h>
 #include <linux/of_address.h>
 #include <linux/of_irq.h>

@@ -2376,10 +2378,11 @@ sh_eth_of_get_register_type(struct device_node *np)
 static struct sh_eth_plat_data *
 sh_eth_parse_dt(struct device *dev, struct net_device *ndev)
 {
-	int ret;
+	int ret, gpio;
 	const char *of_str;
 	struct device_node *np = dev->of_node;
 	struct sh_eth_plat_data *pdata;
+	enum of_gpio_flags flags;
 
 	pdata = devm_kzalloc(dev, sizeof(struct sh_eth_plat_data),
 					GFP_KERNEL);

drivers/net/ethernet/renesas/sh_eth.c +1 -0

+
 #ifdef CONFIG_OF_NET
 	if (!is_valid_ether_addr(ndev->dev_addr)) {
 		const char *macaddr = of_get_mac_address(np);

Hi Guennadi,

Quoted 18 lines · On Thursday 24 January 2013 17:07:32 Guennadi Liakhovetski wrote:

If an ethernet PHY can be reset by a GPIO, it can be specified in DT. Add a binding and code to parse it, request the GPIO and take the PHY out of reset.

Signed-off-by Guennadi Liakhovetski <[email protected]>

Cc devicetree-discuss at lists.ozlabs.org

Cc netdev at vger.kernel.org

---
 Documentation/devicetree/bindings/net/sh_ether.txt |    2 ++
 drivers/net/ethernet/renesas/sh_eth.c              |    9 ++++++++-
 2 files changed, 10 insertions(+), 1 deletions(-)

(quoted diff) +0 -0

b/Documentation/devicetree/bindings/net/sh_ether.txt index c11e45d..edaf683 100644

Documentation/devicetree/bindings/net/sh_ether.txt Required properties: · line 15

Quoted 4 lines

Documentation/devicetree/bindings/net/sh_ether.txt +1 -0

@@ -12,6 +12,7 @@ Required properties:
 - interrupts:                   Interrupt mapping for the sh_eth interrupt
                                 sources (vector id).
 - phy-mode:              String, operation mode of the PHY interface.
+- phy-reset-gpios:              PHY reset GPIO tuple

If you can't have more than one GPIO here, what about calling it phy-reset- gpio ?

From Documentation/devicetree/bindings/gpio/gpio.txt:

Quoted 48 lines

Documentation/devicetree/bindings/net/sh_ether.txt +1 -0

 - sh-eth,edmac-endian:          String, endian of sh_eth dmac.
 - sh-eth,register-type:         String, register type of sh_eth.
                                 Please select "gigabit", "fast-sh4" or

@@ -37,6 +38,7 @@ Example (armadillo800eva):
 		reg = <0xe9a00000 0x800>, <0xe9a01800 0x800>;
 		interrupts = <0x500>;
 		phy-mode = "mii";
+		phy-reset-gpios = <&gpio 18 0>;
 		sh-eth,edmac-endian = "little";
 		sh-eth,register-type = "gigabit";
 		sh-eth,phy-id = <0>;

(quoted diff) +0 -0

b/drivers/net/ethernet/renesas/sh_eth.c index 1f64848..06035a2 100644

drivers/net/ethernet/renesas/sh_eth.c +4 -1

@@ -20,6 +20,7 @@
  *  the file called "COPYING".
  */
 
+#include <linux/gpio.h>
 #include <linux/init.h>
 #include <linux/module.h>
 #include <linux/kernel.h>

@@ -33,6 +34,7 @@
 #include <linux/netdevice.h>
 #include <linux/of.h>
 #include <linux/of_device.h>
+#include <linux/of_gpio.h>
 #include <linux/of_platform.h>
 #include <linux/of_address.h>
 #include <linux/of_irq.h>

@@ -2376,10 +2378,11 @@ sh_eth_of_get_register_type(struct device_node *np)
 static struct sh_eth_plat_data *
 sh_eth_parse_dt(struct device *dev, struct net_device *ndev)
 {
-	int ret;
+	int ret, gpio;
 	const char *of_str;
 	struct device_node *np = dev->of_node;
 	struct sh_eth_plat_data *pdata;
+	enum of_gpio_flags flags;
 
 	pdata = devm_kzalloc(dev, sizeof(struct sh_eth_plat_data),
 					GFP_KERNEL);

@@ -2420,6 +2423,10 @@ sh_eth_parse_dt(struct device *dev, struct net_device
 *ndev) else
 		pdata->needs_init = 0;
 

drivers/net/ethernet/renesas/sh_eth.c sh_eth_parse_dt(struct device *dev, struct net_device *ndev) · line 2428

Quoted 3 lines

drivers/net/ethernet/renesas/sh_eth.c +3 -0

+	gpio = of_get_named_gpio_flags(np, "phy-reset-gpios", 0, &flags);
+	if (gpio_is_valid(gpio) && !devm_gpio_request(dev, gpio, NULL))
+		gpio_direction_output(gpio, !!(flags & OF_GPIO_ACTIVE_LOW));

You could use devm_gpio_request_one() here.

Is there no need to reset the phy at runtime ?

Quoted 4 lines

drivers/net/ethernet/renesas/sh_eth.c +1 -0

+
 #ifdef CONFIG_OF_NET
 	if (!is_valid_ether_addr(ndev->dev_addr)) {
 		const char *macaddr = of_get_mac_address(np);

Signature · 3 lines

Regards,

Laurent Pinchart

Quoted 1 line

Is there no need to reset the phy at runtime ?

No idea, I'm not developing the driver, I'm just porting one specific feature from one API to another with no functional changes (apart from postponing setting the GPIO).

Quoted 2 lines · Guennadi Liakhovetski

Quoted 1 line

Is there no need to reset the phy at runtime ?

Quoted 3 lines

No idea, I'm not developing the driver, I'm just porting one specific feature from one API to another with no functional changes (apart from postponing setting the GPIO).

Generally Linux relies on resetting the phy via the inband MDIO method, which is what Linux does. It is pretty much never required to reset via the hard pin - but you do need to generate a robust reset edge on the reset pin once after power up.

Quoted 13 lines · Laurent Pinchart

Quoted 12 lines · On Thursday 24 January 2013 17:07:32 Guennadi Liakhovetski wrote:

If an ethernet PHY can be reset by a GPIO, it can be specified in DT. Add a binding and code to parse it, request the GPIO and take the PHY out of reset.

Signed-off-by Guennadi Liakhovetski <[email protected]>

Cc devicetree-discuss at lists.ozlabs.org

Cc netdev at vger.kernel.org

---

 Documentation/devicetree/bindings/net/sh_ether.txt |    2 ++
 drivers/net/ethernet/renesas/sh_eth.c              |    9 ++++++++-
 2 files changed, 10 insertions(+), 1 deletions(-)

Quoted 4 lines

Quoted 4 lines

(quoted diff) +0 -0

b/drivers/net/ethernet/renesas/sh_eth.c index 1f64848..06035a2 100644

drivers/net/ethernet/renesas/sh_eth.c +0 -0

Quoted 10 lines

drivers/net/ethernet/renesas/sh_eth.c sh_eth_parse_dt(struct device *dev, struct net_device *ndev) · line 2428

Quoted 8 lines

drivers/net/ethernet/renesas/sh_eth.c +0 -0

@@ -2420,6 +2423,10 @@ sh_eth_parse_dt(struct device *dev, struct

net_device *ndev) else

drivers/net/ethernet/renesas/sh_eth.c +3 -0

 		pdata->needs_init = 0;
 
+	gpio = of_get_named_gpio_flags(np, "phy-reset-gpios", 0, &flags);
+	if (gpio_is_valid(gpio) && !devm_gpio_request(dev, gpio, NULL))
+		gpio_direction_output(gpio, !!(flags & OF_GPIO_ACTIVE_LOW));

You could use devm_gpio_request_one() here.

Yes, but then the flag would look uglier, something like

devm_gpio_request_one(dev, gpio, flags &
OF_GPIO_ACTIVE_LOW ? GPIOF_OUT_INIT_HIGH :
GPIOF_OUT_INIT_LOW);

Does it really look like an improvement? :)

This patch series gets rid of gpio_request()-style ethernet pin configuration on armadillo800eva in reference implementation.

Guennadi Liakhovetski (3):
pinctrl: add ethernet pin groups to r8a7740
ethernet: add a PHY reset GPIO DT binding to sh_eth
ARM: shmobile: move pin configuration on armadillo800eva-reference to
DT

 Documentation/devicetree/bindings/net/sh_ether.txt |    2 +
 .../boot/dts/r8a7740-armadillo800eva-reference.dts |    7 ++++
 .../board-armadillo800eva-reference.c              |   24 ---------------
 drivers/net/ethernet/renesas/sh_eth.c              |    9 +++++-
 drivers/pinctrl/sh-pfc/pfc-r8a7740.c               |   31 ++++++++++++++++++++
 5 files changed, 48 insertions(+), 25 deletions(-)

Signature · 8 lines

1.7.2.5

Thanks
Guennadi
---
Guennadi Liakhovetski, Ph.D.
Freelance Open-Source Software Developer
http://www.open-technology.de/

Quoted 2 lines · Guennadi Liakhovetski

This patch series gets rid of gpio_request()-style ethernet pin configuration on armadillo800eva in reference implementation.

Hi Guennadi,

these changes seem to be reasonable to me.

Are there any dependencies for the sh_eth patch?
I assume this will be handled by David Miller through the net-next tree. Are there any dependencies? The last time I checked the DT bindings for sh_eth had not been merged.

For the remaining two patches, which I assume will go through my renesas tree:
* Are there any dependencies that aren't satisfied by the of-intc branch?

* Could you get some Acks. At least from Laurent?

Quoted 16 lines

Guennadi Liakhovetski (3):
pinctrl: add ethernet pin groups to r8a7740
ethernet: add a PHY reset GPIO DT binding to sh_eth
ARM: shmobile: move pin configuration on armadillo800eva-reference to
DT

 Documentation/devicetree/bindings/net/sh_ether.txt |    2 +
 .../boot/dts/r8a7740-armadillo800eva-reference.dts |    7 ++++
 .../board-armadillo800eva-reference.c              |   24 ---------------
 drivers/net/ethernet/renesas/sh_eth.c              |    9 +++++-
 drivers/pinctrl/sh-pfc/pfc-r8a7740.c               |   31 ++++++++++++++++++++
 5 files changed, 48 insertions(+), 25 deletions(-)

Signature · 1 line

1.7.2.5

Quoted 7 lines

Thanks
Guennadi
---
Guennadi Liakhovetski, Ph.D.
Freelance Open-Source Software Developer
http://www.open-technology.de/

Hi Simon

Quoted 4 lines · Simon Horman

Quoted 2 lines · Guennadi Liakhovetski

This patch series gets rid of gpio_request()-style ethernet pin configuration on armadillo800eva in reference implementation.

Quoted 8 lines

Hi Guennadi,

these changes seem to be reasonable to me.

Are there any dependencies for the sh_eth patch?
I assume this will be handled by David Miller through the net-next tree. Are there any dependencies? The last time I checked the DT bindings for sh_eth had not been merged.

Obviously, it can only be applied, if the Documentation/devicetree/bindings/net/sh_ether.txt file and the sh_eth_parse_dt() function exist. Also, if there are no objections against the new phy-reset-gpios DT property. Otherwise there are no dependencies - as long as the phy-reset-gpios property isn't found in DT, the patch doesn't affect the driver.

Quoted 3 lines

For the remaining two patches, which I assume will go through my renesas tree:
* Are there any dependencies that aren't satisfied by the of-intc branch?

AFAICS, that your branch doesn't contain Laurent's pinctrl patches, which are needed for patch 1 to apply and for patch 3 to make sense. My earlier MMC DT / pinctrl patches aren't required for these patches to function, but these patches won't apply cleanly without them, since they touch the same code fragments. So, it would be easier to merge them in the order of submission.

Quoted 1 line

* Could you get some Acks. At least from Laurent?

Sure, let's give reviewers some more time :)

Quoted 14 lines · Guennadi Liakhovetski

Quoted 4 lines · Simon Horman

Quoted 2 lines · Guennadi Liakhovetski

This patch series gets rid of gpio_request()-style ethernet pin configuration on armadillo800eva in reference implementation.

Quoted 8 lines

Hi Guennadi,

these changes seem to be reasonable to me.

Are there any dependencies for the sh_eth patch?
I assume this will be handled by David Miller through the net-next tree. Are there any dependencies? The last time I checked the DT bindings for sh_eth had not been merged.

Quoted 6 lines

Obviously, it can only be applied, if the Documentation/devicetree/bindings/net/sh_ether.txt file and the sh_eth_parse_dt() function exist. Also, if there are no objections against the new phy-reset-gpios DT property. Otherwise there are no dependencies - as long as the phy-reset-gpios property isn't found in DT, the patch doesn't affect the driver.

Thanks, I'm slightly concerned that the other patch(es) relating to Documentation/devicetree/bindings/net/sh_ether.txt have gone missing in action.

Do you have an interest in chasing them down or would you like me to?

Quoted 14 lines

Quoted 4 lines

For the remaining two patches, which I assume will go through my renesas tree:
* Are there any dependencies that aren't satisfied by the of-intc branch?

AFAICS, that your branch doesn't contain Laurent's pinctrl patches, which are needed for patch 1 to apply and for patch 3 to make sense. My earlier MMC DT / pinctrl patches aren't required for these patches to function, but these patches won't apply cleanly without them, since they touch the same code fragments. So, it would be easier to merge them in the order of submission.

Quoted 1 line

* Could you get some Acks. At least from Laurent?

Quoted 1 line

Sure, let's give reviewers some more time :)

Indeed.

Laurent, if there are patches ready for me to take into the renesas tree please let me know. I'm reluctant to add any more pinmux changes for 3.9. But if a topic branch would help let me know.

Hi Simon,

Quoted 44 lines · On Friday 25 January 2013 18:05:44 Simon Horman wrote:

Quoted 14 lines · Guennadi Liakhovetski

Quoted 4 lines · Simon Horman

Quoted 2 lines · Guennadi Liakhovetski

This patch series gets rid of gpio_request()-style ethernet pin configuration on armadillo800eva in reference implementation.

Quoted 8 lines

Hi Guennadi,

these changes seem to be reasonable to me.

Are there any dependencies for the sh_eth patch?
I assume this will be handled by David Miller through the net-next tree. Are there any dependencies? The last time I checked the DT bindings for sh_eth had not been merged.

Quoted 6 lines

Obviously, it can only be applied, if the Documentation/devicetree/bindings/net/sh_ether.txt file and the sh_eth_parse_dt() function exist. Also, if there are no objections against the new phy-reset-gpios DT property. Otherwise there are no dependencies - as long as the phy-reset-gpios property isn't found in DT, the patch doesn't affect the driver.

Thanks, I'm slightly concerned that the other patch(es) relating to Documentation/devicetree/bindings/net/sh_ether.txt have gone missing in action.

Do you have an interest in chasing them down or would you like me to?

Quoted 15 lines

Quoted 4 lines

For the remaining two patches, which I assume will go through my renesas tree:
* Are there any dependencies that aren't satisfied by the of-intc branch?

AFAICS, that your branch doesn't contain Laurent's pinctrl patches, which are needed for patch 1 to apply and for patch 3 to make sense. My earlier MMC DT / pinctrl patches aren't required for these patches to function, but these patches won't apply cleanly without them, since they touch the same code fragments. So, it would be easier to merge them in the order of submission.

Quoted 1 line

* Could you get some Acks. At least from Laurent?

Sure, let's give reviewers some more time :)

Quoted 5 lines

Indeed.

Laurent, if there are patches ready for me to take into the renesas tree please let me know. I'm reluctant to add any more pinmux changes for 3.9. But if a topic branch would help let me know.

I'd like the gpio_request_one() patches to go to v3.9 if possible. The other pinctrl patches will need to wait until v3.10 I'm afraid.