[PATCH 0/4 v4] usb: dwc2: fix the usb host for host port at RK32

[PATCH 0/4 v4] usb: dwc2: fix the usb host for host port at RK32

[Randy Li]

Hello All:
 This version would move reset into phy. I wanted to implement a hard
code version but I meet some problem with ioremap reset controller,
in order to keep my word, I offer this version this time.

Randy Li (4):
  phy: Add reset callback
  phy: rockchip-usb: use rockchip_usb_phy_reset to reset phy during
    wakeup
  usb: dwc2: assert phy reset when waking up in rk3288 platform
  ARM: dts: rockchip: Point rk3288 dwc2 usb at the full PHY reset

 arch/arm/boot/dts/rk3288.dtsi  |  5 +++++
 drivers/phy/phy-rockchip-usb.c | 20 ++++++++++++++++++++
 drivers/usb/dwc2/core_intr.c   |  9 +++++++++
 include/linux/phy/phy.h        |  2 ++
 4 files changed, 36 insertions(+)

-- 
2.7.4

[PATCH 1/4] phy: Add reset callback

[Randy Li]

The only use for this is for solving a hardware design problem in
usb of Rockchip RK3288.

Signed-off-by: Randy Li <ayaka@soulik.info>
---
 include/linux/phy/phy.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/include/linux/phy/phy.h b/include/linux/phy/phy.h
index f08b672..4d34607 100644
--- a/include/linux/phy/phy.h
+++ b/include/linux/phy/phy.h
@@ -36,6 +36,7 @@ enum phy_mode {
  * @power_on: powering on the phy
  * @power_off: powering off the phy
  * @set_mode: set the mode of the phy
+ * @reset: reseting the phy
  * @owner: the module owner containing the ops
  */
 struct phy_ops {
@@ -44,6 +45,7 @@ struct phy_ops {
 	int	(*power_on)(struct phy *phy);
 	int	(*power_off)(struct phy *phy);
 	int	(*set_mode)(struct phy *phy, enum phy_mode mode);
+	int	(*reset)(struct phy *phy);
 	struct module *owner;
 };
 
-- 
2.7.4

[PATCH 2/4] phy: rockchip-usb: use rockchip_usb_phy_reset to reset phy during wakeup

[Randy Li]

It is a hardware bug in RK3288, the only way to solve it is to
reset the phy.

Signed-off-by: Randy Li <ayaka@soulik.info>
---
 drivers/phy/phy-rockchip-usb.c | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/drivers/phy/phy-rockchip-usb.c b/drivers/phy/phy-rockchip-usb.c
index 2a7381f..734987f 100644
--- a/drivers/phy/phy-rockchip-usb.c
+++ b/drivers/phy/phy-rockchip-usb.c
@@ -29,6 +29,7 @@
 #include <linux/reset.h>
 #include <linux/regmap.h>
 #include <linux/mfd/syscon.h>
+#include <linux/delay.h>
 
 static int enable_usb_uart;
 
@@ -64,6 +65,7 @@ struct rockchip_usb_phy {
 	struct clk_hw	clk480m_hw;
 	struct phy	*phy;
 	bool		uart_enabled;
+	struct reset_control *reset;
 };
 
 static int rockchip_usb_phy_power(struct rockchip_usb_phy *phy,
@@ -144,9 +146,23 @@ static int rockchip_usb_phy_power_on(struct phy *_phy)
 	return clk_prepare_enable(phy->clk480m);
 }
 
+static int rockchip_usb_phy_reset(struct phy *_phy)
+{
+	struct rockchip_usb_phy *phy = phy_get_drvdata(_phy);
+
+	if (phy->reset) {
+		reset_control_assert(phy->reset);
+		udelay(10);
+		reset_control_deassert(phy->reset);
+	}
+
+	return 0;
+}
+
 static const struct phy_ops ops = {
 	.power_on	= rockchip_usb_phy_power_on,
 	.power_off	= rockchip_usb_phy_power_off,
+	.reset		= rockchip_usb_phy_reset,
 	.owner		= THIS_MODULE,
 };
 
@@ -185,6 +201,10 @@ static int rockchip_usb_phy_init(struct rockchip_usb_phy_base *base,
 		return -EINVAL;
 	}
 
+	rk_phy->reset = of_reset_control_get(child, "phy-reset");
+	if (IS_ERR(rk_phy->reset))
+		rk_phy->reset = NULL;
+
 	rk_phy->reg_offset = reg_offset;
 
 	rk_phy->clk = of_clk_get_by_name(child, "phyclk");
-- 
2.7.4

[PATCH 3/4] usb: dwc2: assert phy reset when waking up in rk3288 platform

[Randy Li]

On the rk3288 USB host-only port (the one that's not the OTG-enabled
port) the PHY can get into a bad state when a wakeup is asserted (not
just a wakeup from full system suspend but also a wakeup from
autosuspend).

We can get the PHY out of its bad state by asserting its "port reset",
but unfortunately that seems to assert a reset onto the USB bus so it
could confuse things if we don't actually deenumerate / reenumerate the
device.

We can also get the PHY out of its bad state by fully resetting it using
the reset from the CRU (clock reset unit) in chip, which does a more full
reset.  The CRU-based reset appears to actually cause devices on the bus
to be removed and reinserted, which fixes the problem (albeit in a hacky
way).

It's unfortunate that we need to do a full re-enumeration of devices at
wakeup time, but this is better than alternative of letting the bus get
wedged.

The original patches came from Doug Anderson <dianders@chromium.org>.

Signed-off-by: Randy Li <ayaka@soulik.info>
---
 drivers/usb/dwc2/core_intr.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/drivers/usb/dwc2/core_intr.c b/drivers/usb/dwc2/core_intr.c
index d85c5c9..c31c3fe 100644
--- a/drivers/usb/dwc2/core_intr.c
+++ b/drivers/usb/dwc2/core_intr.c
@@ -345,6 +345,7 @@ static void dwc2_handle_session_req_intr(struct dwc2_hsotg *hsotg)
 static void dwc2_handle_wakeup_detected_intr(struct dwc2_hsotg *hsotg)
 {
 	int ret;
+	struct device_node *np = hsotg->dev->of_node;
 
 	/* Clear interrupt */
 	dwc2_writel(GINTSTS_WKUPINT, hsotg->regs + GINTSTS);
@@ -379,6 +380,14 @@ static void dwc2_handle_wakeup_detected_intr(struct dwc2_hsotg *hsotg)
 			/* Restart the Phy Clock */
 			pcgcctl &= ~PCGCTL_STOPPCLK;
 			dwc2_writel(pcgcctl, hsotg->regs + PCGCTL);
+
+			/* It is a quirk in Rockchip RK3288, causing by
+			 * a hardware bug. This will propagate out and
+			 * eventually we'll re-enumerate the device. 
+			 * Not great but the best we can do */
+			if (of_device_is_compatible(np, "rockchip,rk3288-usb"))
+				hsotg->phy->ops->reset(hsotg->phy);
+
 			mod_timer(&hsotg->wkp_timer,
 				  jiffies + msecs_to_jiffies(71));
 		} else {
-- 
2.7.4

[PATCH 4/4] ARM: dts: rockchip: Point rk3288 dwc2 usb at the full PHY reset

[Randy Li]

The "host1" port (AKA the dwc2 port that isn't the OTG port) on rk3288
has a hardware errata that causes everything to get confused when we get
a remote wakeup.  We'll use the reset that's in the CRU to reset the
port when it's in a bad state.

Note that we add the reset to both dwc2 controllers even though only one
has the errata in case we find some other use for this reset that's
unrelated to the current hardware errata.  Only the host port gets the
quirk property, though.

This patch came from Doug Anderson <dianders@chromium.org> originally.

Signed-off-by: Randy Li <ayaka@soulik.info>
---
 arch/arm/boot/dts/rk3288.dtsi | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/arch/arm/boot/dts/rk3288.dtsi b/arch/arm/boot/dts/rk3288.dtsi
index 48ca4e4..646f49d 100644
--- a/arch/arm/boot/dts/rk3288.dtsi
+++ b/arch/arm/boot/dts/rk3288.dtsi
@@ -855,6 +855,8 @@
 				clocks = <&cru SCLK_OTGPHY0>;
 				clock-names = "phyclk";
 				#clock-cells = <0>;
+				resets = <&cru SRST_USBOTG_PHY>;
+				reset-names = "phy-reset";
 			};
 
 			usbphy1: usb-phy@334 {
@@ -871,6 +873,9 @@
 				clocks = <&cru SCLK_OTGPHY2>;
 				clock-names = "phyclk";
 				#clock-cells = <0>;
+				resets = <&cru SRST_USBHOST1_PHY>;
+				reset-names = "phy-reset";
+
 			};
 		};
 	};
-- 
2.7.4

Re: [PATCH 3/4] usb: dwc2: assert phy reset when waking up in rk3288 platform

[Sergei Shtylyov]

Hello.

On 08/20/2016 11:32 PM, Randy Li wrote:

> On the rk3288 USB host-only port (the one that's not the OTG-enabled
> port) the PHY can get into a bad state when a wakeup is asserted (not
> just a wakeup from full system suspend but also a wakeup from
> autosuspend).
>
> We can get the PHY out of its bad state by asserting its "port reset",
> but unfortunately that seems to assert a reset onto the USB bus so it
> could confuse things if we don't actually deenumerate / reenumerate the
> device.
>
> We can also get the PHY out of its bad state by fully resetting it using
> the reset from the CRU (clock reset unit) in chip, which does a more full
> reset.  The CRU-based reset appears to actually cause devices on the bus
> to be removed and reinserted, which fixes the problem (albeit in a hacky
> way).
>
> It's unfortunate that we need to do a full re-enumeration of devices at
> wakeup time, but this is better than alternative of letting the bus get
> wedged.
>
> The original patches came from Doug Anderson <dianders@chromium.org>.
>
> Signed-off-by: Randy Li <ayaka@soulik.info>
> ---
>  drivers/usb/dwc2/core_intr.c | 9 +++++++++
>  1 file changed, 9 insertions(+)
>
> diff --git a/drivers/usb/dwc2/core_intr.c b/drivers/usb/dwc2/core_intr.c
> index d85c5c9..c31c3fe 100644
> --- a/drivers/usb/dwc2/core_intr.c
> +++ b/drivers/usb/dwc2/core_intr.c
[...]
> @@ -379,6 +380,14 @@ static void dwc2_handle_wakeup_detected_intr(struct dwc2_hsotg *hsotg)
>  			/* Restart the Phy Clock */
>  			pcgcctl &= ~PCGCTL_STOPPCLK;
>  			dwc2_writel(pcgcctl, hsotg->regs + PCGCTL);
> +
> +			/* It is a quirk in Rockchip RK3288, causing by
> +			 * a hardware bug. This will propagate out and
> +			 * eventually we'll re-enumerate the device.
> +			 * Not great but the best we can do */

   The preferred multi-line comment style is this:

/*
  * bla
  * bla
  */

    Yours is close. :-)

[...]

MBR, Sergei

[PATCH 0/4 v5] usb: dwc2: fix the usb host for host port at RK32

[Randy Li]

Hello All:
 This version would move reset into phy. I wanted to implement a hard
code version but I meet some problem with ioremap reset controller,
in order to keep my word, I offer this version this time.

Changelog:
v5
  A few modification at style, add the missing doc in the last 
  commit.
v4
  1. Adding the reset callback in struct phy_ops.
  2. Moving the reset into phy rockchip usb.
  3. Trying to call a reset when dwc2 wakeup in rk3288.
v3
 Rebased from previous commit.

Randy Li (4):
  phy: Add reset callback
  phy: rockchip-usb: use rockchip_usb_phy_reset to reset phy during
    wakeup
  usb: dwc2: assert phy reset when waking up in rk3288 platform
  ARM: dts: rockchip: Point rk3288 dwc2 usb at the full PHY reset

 arch/arm/boot/dts/rk3288.dtsi  |  5 +++++
 drivers/phy/phy-rockchip-usb.c | 20 ++++++++++++++++++++
 drivers/usb/dwc2/core_intr.c   |  9 +++++++++
 include/linux/phy/phy.h        |  2 ++
 4 files changed, 36 insertions(+)

-- 
2.7.4

[PATCH 1/4] phy: Add reset callback

[Randy Li]

The only use for this is for solving a hardware design problem in
usb of Rockchip RK3288.

Signed-off-by: Randy Li <ayaka@soulik.info>
---
 include/linux/phy/phy.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/include/linux/phy/phy.h b/include/linux/phy/phy.h
index f08b672..4d34607 100644
--- a/include/linux/phy/phy.h
+++ b/include/linux/phy/phy.h
@@ -36,6 +36,7 @@ enum phy_mode {
  * @power_on: powering on the phy
  * @power_off: powering off the phy
  * @set_mode: set the mode of the phy
+ * @reset: reseting the phy
  * @owner: the module owner containing the ops
  */
 struct phy_ops {
@@ -44,6 +45,7 @@ struct phy_ops {
 	int	(*power_on)(struct phy *phy);
 	int	(*power_off)(struct phy *phy);
 	int	(*set_mode)(struct phy *phy, enum phy_mode mode);
+	int	(*reset)(struct phy *phy);
 	struct module *owner;
 };
 
-- 
2.7.4

[PATCH 2/4] phy: rockchip-usb: use rockchip_usb_phy_reset to reset phy during wakeup

[Randy Li]

It is a hardware bug in RK3288, the only way to solve it is to
reset the phy.

Signed-off-by: Randy Li <ayaka@soulik.info>
---
 .../devicetree/bindings/phy/rockchip-usb-phy.txt     |  3 +++
 drivers/phy/phy-rockchip-usb.c                       | 20 ++++++++++++++++++++
 2 files changed, 23 insertions(+)

diff --git a/Documentation/devicetree/bindings/phy/rockchip-usb-phy.txt b/Documentation/devicetree/bindings/phy/rockchip-usb-phy.txt
index cc6be96..57dc388 100644
--- a/Documentation/devicetree/bindings/phy/rockchip-usb-phy.txt
+++ b/Documentation/devicetree/bindings/phy/rockchip-usb-phy.txt
@@ -27,6 +27,9 @@ Optional Properties:
 - clocks : phandle + clock specifier for the phy clocks
 - clock-names: string, clock name, must be "phyclk"
 - #clock-cells: for users of the phy-pll, should be 0
+- reset-names: Only allow the following entries:
+ - phy-reset
+- resets: Must contain an entry for each entry in reset-names.
 
 Example:
 
diff --git a/drivers/phy/phy-rockchip-usb.c b/drivers/phy/phy-rockchip-usb.c
index 2a7381f..734987f 100644
--- a/drivers/phy/phy-rockchip-usb.c
+++ b/drivers/phy/phy-rockchip-usb.c
@@ -29,6 +29,7 @@
 #include <linux/reset.h>
 #include <linux/regmap.h>
 #include <linux/mfd/syscon.h>
+#include <linux/delay.h>
 
 static int enable_usb_uart;
 
@@ -64,6 +65,7 @@ struct rockchip_usb_phy {
 	struct clk_hw	clk480m_hw;
 	struct phy	*phy;
 	bool		uart_enabled;
+	struct reset_control *reset;
 };
 
 static int rockchip_usb_phy_power(struct rockchip_usb_phy *phy,
@@ -144,9 +146,23 @@ static int rockchip_usb_phy_power_on(struct phy *_phy)
 	return clk_prepare_enable(phy->clk480m);
 }
 
+static int rockchip_usb_phy_reset(struct phy *_phy)
+{
+	struct rockchip_usb_phy *phy = phy_get_drvdata(_phy);
+
+	if (phy->reset) {
+		reset_control_assert(phy->reset);
+		udelay(10);
+		reset_control_deassert(phy->reset);
+	}
+
+	return 0;
+}
+
 static const struct phy_ops ops = {
 	.power_on	= rockchip_usb_phy_power_on,
 	.power_off	= rockchip_usb_phy_power_off,
+	.reset		= rockchip_usb_phy_reset,
 	.owner		= THIS_MODULE,
 };
 
@@ -185,6 +201,10 @@ static int rockchip_usb_phy_init(struct rockchip_usb_phy_base *base,
 		return -EINVAL;
 	}
 
+	rk_phy->reset = of_reset_control_get(child, "phy-reset");
+	if (IS_ERR(rk_phy->reset))
+		rk_phy->reset = NULL;
+
 	rk_phy->reg_offset = reg_offset;
 
 	rk_phy->clk = of_clk_get_by_name(child, "phyclk");
-- 
2.7.4

[PATCH 3/4] usb: dwc2: assert phy reset when waking up in rk3288 platform

[Randy Li]

On the rk3288 USB host-only port (the one that's not the OTG-enabled
port) the PHY can get into a bad state when a wakeup is asserted (not
just a wakeup from full system suspend but also a wakeup from
autosuspend).

We can get the PHY out of its bad state by asserting its "port reset",
but unfortunately that seems to assert a reset onto the USB bus so it
could confuse things if we don't actually deenumerate / reenumerate the
device.

We can also get the PHY out of its bad state by fully resetting it using
the reset from the CRU (clock reset unit) in chip, which does a more full
reset.  The CRU-based reset appears to actually cause devices on the bus
to be removed and reinserted, which fixes the problem (albeit in a hacky
way).

It's unfortunate that we need to do a full re-enumeration of devices at
wakeup time, but this is better than alternative of letting the bus get
wedged.

The original patches came from Doug Anderson <dianders@chromium.org>.

Signed-off-by: Randy Li <ayaka@soulik.info>
---
 drivers/usb/dwc2/core_intr.c | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/drivers/usb/dwc2/core_intr.c b/drivers/usb/dwc2/core_intr.c
index d85c5c9..f57c48a 100644
--- a/drivers/usb/dwc2/core_intr.c
+++ b/drivers/usb/dwc2/core_intr.c
@@ -345,6 +345,7 @@ static void dwc2_handle_session_req_intr(struct dwc2_hsotg *hsotg)
 static void dwc2_handle_wakeup_detected_intr(struct dwc2_hsotg *hsotg)
 {
 	int ret;
+	struct device_node *np = hsotg->dev->of_node;
 
 	/* Clear interrupt */
 	dwc2_writel(GINTSTS_WKUPINT, hsotg->regs + GINTSTS);
@@ -379,6 +380,16 @@ static void dwc2_handle_wakeup_detected_intr(struct dwc2_hsotg *hsotg)
 			/* Restart the Phy Clock */
 			pcgcctl &= ~PCGCTL_STOPPCLK;
 			dwc2_writel(pcgcctl, hsotg->regs + PCGCTL);
+
+			/* 
+			 * It is a quirk in Rockchip RK3288, causing by
+			 * a hardware bug. This will propagate out and
+			 * eventually we'll re-enumerate the device. 
+			 * Not great but the best we can do 
+			 */
+			if (of_device_is_compatible(np, "rockchip,rk3288-usb"))
+				hsotg->phy->ops->reset(hsotg->phy);
+
 			mod_timer(&hsotg->wkp_timer,
 				  jiffies + msecs_to_jiffies(71));
 		} else {
-- 
2.7.4

[PATCH 4/4] ARM: dts: rockchip: Point rk3288 dwc2 usb at the full PHY reset

[Randy Li]

The "host1" port (AKA the dwc2 port that isn't the OTG port) on rk3288
has a hardware errata that causes everything to get confused when we get
a remote wakeup.  We'll use the reset that's in the CRU to reset the
port when it's in a bad state.

Note that we add the reset to both dwc2 controllers even though only one
has the errata in case we find some other use for this reset that's
unrelated to the current hardware errata.  Only the host port gets the
quirk property, though.

This patch came from Doug Anderson <dianders@chromium.org> originally.

Signed-off-by: Randy Li <ayaka@soulik.info>
---
 arch/arm/boot/dts/rk3288.dtsi | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/arch/arm/boot/dts/rk3288.dtsi b/arch/arm/boot/dts/rk3288.dtsi
index 48ca4e4..646f49d 100644
--- a/arch/arm/boot/dts/rk3288.dtsi
+++ b/arch/arm/boot/dts/rk3288.dtsi
@@ -855,6 +855,8 @@
 				clocks = <&cru SCLK_OTGPHY0>;
 				clock-names = "phyclk";
 				#clock-cells = <0>;
+				resets = <&cru SRST_USBOTG_PHY>;
+				reset-names = "phy-reset";
 			};
 
 			usbphy1: usb-phy@334 {
@@ -871,6 +873,9 @@
 				clocks = <&cru SCLK_OTGPHY2>;
 				clock-names = "phyclk";
 				#clock-cells = <0>;
+				resets = <&cru SRST_USBHOST1_PHY>;
+				reset-names = "phy-reset";
+
 			};
 		};
 	};
-- 
2.7.4

Re: [PATCH 4/4] ARM: dts: rockchip: Point rk3288 dwc2 usb at the full PHY reset

[Sergei Shtylyov]

Hello.

On 8/21/2016 10:56 AM, Randy Li wrote:

> The "host1" port (AKA the dwc2 port that isn't the OTG port) on rk3288
> has a hardware errata that causes everything to get confused when we get
> a remote wakeup.  We'll use the reset that's in the CRU to reset the
> port when it's in a bad state.
>
> Note that we add the reset to both dwc2 controllers even though only one
> has the errata in case we find some other use for this reset that's
> unrelated to the current hardware errata.  Only the host port gets the
> quirk property, though.
>
> This patch came from Doug Anderson <dianders@chromium.org> originally.

    Was it signed off by him? Don't you need to keep his authorship via th 
"From:" tag?

> Signed-off-by: Randy Li <ayaka@soulik.info>
> ---
>  arch/arm/boot/dts/rk3288.dtsi | 5 +++++
>  1 file changed, 5 insertions(+)
>
> diff --git a/arch/arm/boot/dts/rk3288.dtsi b/arch/arm/boot/dts/rk3288.dtsi
> index 48ca4e4..646f49d 100644
> --- a/arch/arm/boot/dts/rk3288.dtsi
> +++ b/arch/arm/boot/dts/rk3288.dtsi
[...]
> @@ -871,6 +873,9 @@
>  				clocks = <&cru SCLK_OTGPHY2>;
>  				clock-names = "phyclk";
>  				#clock-cells = <0>;
> +				resets = <&cru SRST_USBHOST1_PHY>;
> +				reset-names = "phy-reset";
> +

    Don't need empty line here.

>  			};
>  		};
>  	};

MBR , Sergei

Re: [PATCH 4/4] ARM: dts: rockchip: Point rk3288 dwc2 usb at the full PHY reset

[ayaka]

On 08/21/2016 07:06 PM, Sergei Shtylyov wrote:
> Hello.
>
> On 8/21/2016 10:56 AM, Randy Li wrote:
>
>> The "host1" port (AKA the dwc2 port that isn't the OTG port) on rk3288
>> has a hardware errata that causes everything to get confused when we get
>> a remote wakeup.  We'll use the reset that's in the CRU to reset the
>> port when it's in a bad state.
>>
>> Note that we add the reset to both dwc2 controllers even though only one
>> has the errata in case we find some other use for this reset that's
>> unrelated to the current hardware errata.  Only the host port gets the
>> quirk property, though.
>>
>> This patch came from Doug Anderson <dianders@chromium.org> originally.
>
>    Was it signed off by him? Don't you need to keep his authorship via 
> th "From:" tag?
I would rather the idea comes from him, but the implementation is 
different. I don't which tag would be
better? If the "From:" would be the best choice. I would resend the last 
two patches.
>> Signed-off-by: Randy Li <ayaka@soulik.info>
>> ---
>>  arch/arm/boot/dts/rk3288.dtsi | 5 +++++
>>  1 file changed, 5 insertions(+)
>>
>> diff --git a/arch/arm/boot/dts/rk3288.dtsi 
>> b/arch/arm/boot/dts/rk3288.dtsi
>> index 48ca4e4..646f49d 100644
>> --- a/arch/arm/boot/dts/rk3288.dtsi
>> +++ b/arch/arm/boot/dts/rk3288.dtsi
> [...]
>> @@ -871,6 +873,9 @@
>>                  clocks = <&cru SCLK_OTGPHY2>;
>>                  clock-names = "phyclk";
>>                  #clock-cells = <0>;
>> +                resets = <&cru SRST_USBHOST1_PHY>;
>> +                reset-names = "phy-reset";
>> +
>
>    Don't need empty line here.
I would be removed
>
>>              };
>>          };
>>      };
>
> MBR , Sergei

[RESEND PATCH 3/4] usb: dwc2: assert phy reset when waking up in rk3288 platform

[Randy Li]

On the rk3288 USB host-only port (the one that's not the OTG-enabled
port) the PHY can get into a bad state when a wakeup is asserted (not
just a wakeup from full system suspend but also a wakeup from
autosuspend).

We can get the PHY out of its bad state by asserting its "port reset",
but unfortunately that seems to assert a reset onto the USB bus so it
could confuse things if we don't actually deenumerate / reenumerate the
device.

We can also get the PHY out of its bad state by fully resetting it using
the reset from the CRU (clock reset unit) in chip, which does a more full
reset.  The CRU-based reset appears to actually cause devices on the bus
to be removed and reinserted, which fixes the problem (albeit in a hacky
way).

It's unfortunate that we need to do a full re-enumeration of devices at
wakeup time, but this is better than alternative of letting the bus get
wedged.

Signed-off-by: Randy Li <ayaka@soulik.info>
---
 drivers/usb/dwc2/core_intr.c | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/drivers/usb/dwc2/core_intr.c b/drivers/usb/dwc2/core_intr.c
index d85c5c9..f57c48a 100644
--- a/drivers/usb/dwc2/core_intr.c
+++ b/drivers/usb/dwc2/core_intr.c
@@ -345,6 +345,7 @@ static void dwc2_handle_session_req_intr(struct dwc2_hsotg *hsotg)
 static void dwc2_handle_wakeup_detected_intr(struct dwc2_hsotg *hsotg)
 {
 	int ret;
+	struct device_node *np = hsotg->dev->of_node;
 
 	/* Clear interrupt */
 	dwc2_writel(GINTSTS_WKUPINT, hsotg->regs + GINTSTS);
@@ -379,6 +380,16 @@ static void dwc2_handle_wakeup_detected_intr(struct dwc2_hsotg *hsotg)
 			/* Restart the Phy Clock */
 			pcgcctl &= ~PCGCTL_STOPPCLK;
 			dwc2_writel(pcgcctl, hsotg->regs + PCGCTL);
+
+			/* 
+			 * It is a quirk in Rockchip RK3288, causing by
+			 * a hardware bug. This will propagate out and
+			 * eventually we'll re-enumerate the device. 
+			 * Not great but the best we can do 
+			 */
+			if (of_device_is_compatible(np, "rockchip,rk3288-usb"))
+				hsotg->phy->ops->reset(hsotg->phy);
+
 			mod_timer(&hsotg->wkp_timer,
 				  jiffies + msecs_to_jiffies(71));
 		} else {
-- 
2.7.4

[RESEND PATCH 4/4] ARM: dts: rockchip: Point rk3288 dwc2 usb at the full PHY reset

[Randy Li]

The "host1" port (AKA the dwc2 port that isn't the OTG port) on rk3288
has a hardware errata that causes everything to get confused when we get
a remote wakeup.  We'll use the reset that's in the CRU to reset the
port when it's in a bad state.

Note that we add the reset to both dwc2 controllers even though only one
has the errata in case we find some other use for this reset that's
unrelated to the current hardware errata.  Only the host port gets the
quirk property, though.

Signed-off-by: Randy Li <ayaka@soulik.info>
---
 arch/arm/boot/dts/rk3288.dtsi | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/arch/arm/boot/dts/rk3288.dtsi b/arch/arm/boot/dts/rk3288.dtsi
index 48ca4e4..de54604 100644
--- a/arch/arm/boot/dts/rk3288.dtsi
+++ b/arch/arm/boot/dts/rk3288.dtsi
@@ -855,6 +855,8 @@
 				clocks = <&cru SCLK_OTGPHY0>;
 				clock-names = "phyclk";
 				#clock-cells = <0>;
+				resets = <&cru SRST_USBOTG_PHY>;
+				reset-names = "phy-reset";
 			};
 
 			usbphy1: usb-phy@334 {
@@ -871,6 +873,8 @@
 				clocks = <&cru SCLK_OTGPHY2>;
 				clock-names = "phyclk";
 				#clock-cells = <0>;
+				resets = <&cru SRST_USBHOST1_PHY>;
+				reset-names = "phy-reset";
 			};
 		};
 	};
-- 
2.7.4

Re: [PATCH 4/4] ARM: dts: rockchip: Point rk3288 dwc2 usb at the full PHY reset

[Sergei Shtylyov]

Hello.

On 8/21/2016 3:19 PM, ayaka wrote:

>>> The "host1" port (AKA the dwc2 port that isn't the OTG port) on rk3288
>>> has a hardware errata that causes everything to get confused when we get
>>> a remote wakeup.  We'll use the reset that's in the CRU to reset the
>>> port when it's in a bad state.
>>>
>>> Note that we add the reset to both dwc2 controllers even though only one
>>> has the errata in case we find some other use for this reset that's
>>> unrelated to the current hardware errata.  Only the host port gets the
>>> quirk property, though.
>>>
>>> This patch came from Doug Anderson <dianders@chromium.org> originally.
>>
>>    Was it signed off by him? Don't you need to keep his authorship via th
>> "From:" tag?

> I would rather the idea comes from him, but the implementation is different. I

    Then write exactly that. Because now it sounds like the patch itself was 
originated from him.

> don't which tag would be
> better? If the "From:" would be the best choice. I would resend the last two
> patches.

    There's also "Suggested-by:" tag which might fit here.

>>> Signed-off-by: Randy Li <ayaka@soulik.info>
[...]

MBR, Sergei

Re: [PATCH 2/4] phy: rockchip-usb: use rockchip_usb_phy_reset to reset phy during wakeup

[Kishon Vijay Abraham I]

Hi,

On Sunday 21 August 2016 02:02 AM, Randy Li wrote:
> It is a hardware bug in RK3288, the only way to solve it is to
> reset the phy.
> 
> Signed-off-by: Randy Li <ayaka@soulik.info>
> ---
>  drivers/phy/phy-rockchip-usb.c | 20 ++++++++++++++++++++
>  1 file changed, 20 insertions(+)
> 
> diff --git a/drivers/phy/phy-rockchip-usb.c b/drivers/phy/phy-rockchip-usb.c
> index 2a7381f..734987f 100644
> --- a/drivers/phy/phy-rockchip-usb.c
> +++ b/drivers/phy/phy-rockchip-usb.c
> @@ -29,6 +29,7 @@
>  #include <linux/reset.h>
>  #include <linux/regmap.h>
>  #include <linux/mfd/syscon.h>
> +#include <linux/delay.h>
>  
>  static int enable_usb_uart;
>  
> @@ -64,6 +65,7 @@ struct rockchip_usb_phy {
>  	struct clk_hw	clk480m_hw;
>  	struct phy	*phy;
>  	bool		uart_enabled;
> +	struct reset_control *reset;
>  };
>  
>  static int rockchip_usb_phy_power(struct rockchip_usb_phy *phy,
> @@ -144,9 +146,23 @@ static int rockchip_usb_phy_power_on(struct phy *_phy)
>  	return clk_prepare_enable(phy->clk480m);
>  }
>  
> +static int rockchip_usb_phy_reset(struct phy *_phy)
> +{
> +	struct rockchip_usb_phy *phy = phy_get_drvdata(_phy);
> +
> +	if (phy->reset) {
> +		reset_control_assert(phy->reset);
> +		udelay(10);
> +		reset_control_deassert(phy->reset);
> +	}
> +
> +	return 0;
> +}
> +
>  static const struct phy_ops ops = {
>  	.power_on	= rockchip_usb_phy_power_on,
>  	.power_off	= rockchip_usb_phy_power_off,
> +	.reset		= rockchip_usb_phy_reset,

why not just reuse the .init ops? reset can be done during initialization right?

Thanks
Kishon

Re: [PATCH 2/4] phy: rockchip-usb: use rockchip_usb_phy_reset to reset phy during wakeup

[Heiko Stübner]

Am Montag, 22. August 2016, 17:17:41 schrieb Kishon Vijay Abraham I:
> Hi,
> 
> On Sunday 21 August 2016 02:02 AM, Randy Li wrote:
> > It is a hardware bug in RK3288, the only way to solve it is to
> > reset the phy.
> > 
> > Signed-off-by: Randy Li <ayaka@soulik.info>
> > ---
> > 
> >  drivers/phy/phy-rockchip-usb.c | 20 ++++++++++++++++++++
> >  1 file changed, 20 insertions(+)
> > 
> > diff --git a/drivers/phy/phy-rockchip-usb.c
> > b/drivers/phy/phy-rockchip-usb.c index 2a7381f..734987f 100644
> > --- a/drivers/phy/phy-rockchip-usb.c
> > +++ b/drivers/phy/phy-rockchip-usb.c
> > @@ -29,6 +29,7 @@
> > 
> >  #include <linux/reset.h>
> >  #include <linux/regmap.h>
> >  #include <linux/mfd/syscon.h>
> > 
> > +#include <linux/delay.h>
> > 
> >  static int enable_usb_uart;
> > 
> > @@ -64,6 +65,7 @@ struct rockchip_usb_phy {
> > 
> >  	struct clk_hw	clk480m_hw;
> >  	struct phy	*phy;
> >  	bool		uart_enabled;
> > 
> > +	struct reset_control *reset;
> > 
> >  };
> >  
> >  static int rockchip_usb_phy_power(struct rockchip_usb_phy *phy,
> > 
> > @@ -144,9 +146,23 @@ static int rockchip_usb_phy_power_on(struct phy
> > *_phy)
> > 
> >  	return clk_prepare_enable(phy->clk480m);
> >  
> >  }
> > 
> > +static int rockchip_usb_phy_reset(struct phy *_phy)
> > +{
> > +	struct rockchip_usb_phy *phy = phy_get_drvdata(_phy);
> > +
> > +	if (phy->reset) {
> > +		reset_control_assert(phy->reset);
> > +		udelay(10);
> > +		reset_control_deassert(phy->reset);
> > +	}
> > +
> > +	return 0;
> > +}
> > +
> > 
> >  static const struct phy_ops ops = {
> >  
> >  	.power_on	= rockchip_usb_phy_power_on,
> >  	.power_off	= rockchip_usb_phy_power_off,
> > 
> > +	.reset		= rockchip_usb_phy_reset,
> 
> why not just reuse the .init ops? reset can be done during initialization
> right?

The naming of power_on + power_off and init + exit probably suggests that they 
are supposed to be used in pairs. (aka module_init + module_exit and probably 
more)

But in fact I've seen different combinations so far (phy_init + phy_power_on 
... phy_power_off + phy_exit but also phy_power_on + phy_init ... phy_exit + 
phy_power_off), so I guess the semantics are not that strictly defined.

Re: [PATCH 2/4] phy: rockchip-usb: use rockchip_usb_phy_reset to reset phy during wakeup

[Rob Herring]

On Sun, Aug 21, 2016 at 03:56:43PM +0800, Randy Li wrote:
> It is a hardware bug in RK3288, the only way to solve it is to
> reset the phy.
> 
> Signed-off-by: Randy Li <ayaka@soulik.info>
> ---
>  .../devicetree/bindings/phy/rockchip-usb-phy.txt     |  3 +++
>  drivers/phy/phy-rockchip-usb.c                       | 20 ++++++++++++++++++++
>  2 files changed, 23 insertions(+)

Acked-by: Rob Herring <robh@kernel.org>

Re: [RESEND PATCH 3/4] usb: dwc2: assert phy reset when waking up in rk3288 platform

[John Youn]

On 8/21/2016 12:32 PM, Randy Li wrote:
> On the rk3288 USB host-only port (the one that's not the OTG-enabled
> port) the PHY can get into a bad state when a wakeup is asserted (not
> just a wakeup from full system suspend but also a wakeup from
> autosuspend).
> 
> We can get the PHY out of its bad state by asserting its "port reset",
> but unfortunately that seems to assert a reset onto the USB bus so it
> could confuse things if we don't actually deenumerate / reenumerate the
> device.
> 
> We can also get the PHY out of its bad state by fully resetting it using
> the reset from the CRU (clock reset unit) in chip, which does a more full
> reset.  The CRU-based reset appears to actually cause devices on the bus
> to be removed and reinserted, which fixes the problem (albeit in a hacky
> way).
> 
> It's unfortunate that we need to do a full re-enumeration of devices at
> wakeup time, but this is better than alternative of letting the bus get
> wedged.
> 
> Signed-off-by: Randy Li <ayaka@soulik.info>
> ---
>  drivers/usb/dwc2/core_intr.c | 11 +++++++++++
>  1 file changed, 11 insertions(+)
> 
> diff --git a/drivers/usb/dwc2/core_intr.c b/drivers/usb/dwc2/core_intr.c
> index d85c5c9..f57c48a 100644
> --- a/drivers/usb/dwc2/core_intr.c
> +++ b/drivers/usb/dwc2/core_intr.c
> @@ -345,6 +345,7 @@ static void dwc2_handle_session_req_intr(struct dwc2_hsotg *hsotg)
>  static void dwc2_handle_wakeup_detected_intr(struct dwc2_hsotg *hsotg)
>  {
>  	int ret;
> +	struct device_node *np = hsotg->dev->of_node;
>  
>  	/* Clear interrupt */
>  	dwc2_writel(GINTSTS_WKUPINT, hsotg->regs + GINTSTS);
> @@ -379,6 +380,16 @@ static void dwc2_handle_wakeup_detected_intr(struct dwc2_hsotg *hsotg)
>  			/* Restart the Phy Clock */
>  			pcgcctl &= ~PCGCTL_STOPPCLK;
>  			dwc2_writel(pcgcctl, hsotg->regs + PCGCTL);
> +
> +			/* 
> +			 * It is a quirk in Rockchip RK3288, causing by
> +			 * a hardware bug. This will propagate out and
> +			 * eventually we'll re-enumerate the device. 
> +			 * Not great but the best we can do 

Remove the trailing whitespaces in this comment.

Run scripts/checkpatch.pl to catch these.

> +			 */
> +			if (of_device_is_compatible(np, "rockchip,rk3288-usb"))
> +				hsotg->phy->ops->reset(hsotg->phy);
> +

You should probably check for NULL before calling the reset()
callback.

Also, I'd rather see a generic quirk property that you set for your
platform.

Something like "phy_reset_on_wakeup_quirk".

Also, try to preserve the version tag in your subject for all the
patches so that we can easily identify the latest version of the
series, like:

[PATCH v5 3/4] ...

And, typically "RESEND" means there are no code change.

Regards,
John

Re: [RESEND PATCH 3/4] usb: dwc2: assert phy reset when waking up in rk3288 platform

[Randy Li]

On 08/24/2016 04:46 AM, John Youn wrote:
> On 8/21/2016 12:32 PM, Randy Li wrote:
>> On the rk3288 USB host-only port (the one that's not the OTG-enabled
>> port) the PHY can get into a bad state when a wakeup is asserted (not
>> just a wakeup from full system suspend but also a wakeup from
>> autosuspend).
>>
>> We can get the PHY out of its bad state by asserting its "port reset",
>> but unfortunately that seems to assert a reset onto the USB bus so it
>> could confuse things if we don't actually deenumerate / reenumerate the
>> device.
>>
>> We can also get the PHY out of its bad state by fully resetting it using
>> the reset from the CRU (clock reset unit) in chip, which does a more full
>> reset.  The CRU-based reset appears to actually cause devices on the bus
>> to be removed and reinserted, which fixes the problem (albeit in a hacky
>> way).
>>
>> It's unfortunate that we need to do a full re-enumeration of devices at
>> wakeup time, but this is better than alternative of letting the bus get
>> wedged.
>>
>> Signed-off-by: Randy Li <ayaka@soulik.info>
>> ---
>>  drivers/usb/dwc2/core_intr.c | 11 +++++++++++
>>  1 file changed, 11 insertions(+)
>>
>> diff --git a/drivers/usb/dwc2/core_intr.c b/drivers/usb/dwc2/core_intr.c
>> index d85c5c9..f57c48a 100644
>> --- a/drivers/usb/dwc2/core_intr.c
>> +++ b/drivers/usb/dwc2/core_intr.c
>> @@ -345,6 +345,7 @@ static void dwc2_handle_session_req_intr(struct dwc2_hsotg *hsotg)
>>  static void dwc2_handle_wakeup_detected_intr(struct dwc2_hsotg *hsotg)
>>  {
>>  	int ret;
>> +	struct device_node *np = hsotg->dev->of_node;
>>
>>  	/* Clear interrupt */
>>  	dwc2_writel(GINTSTS_WKUPINT, hsotg->regs + GINTSTS);
>> @@ -379,6 +380,16 @@ static void dwc2_handle_wakeup_detected_intr(struct dwc2_hsotg *hsotg)
>>  			/* Restart the Phy Clock */
>>  			pcgcctl &= ~PCGCTL_STOPPCLK;
>>  			dwc2_writel(pcgcctl, hsotg->regs + PCGCTL);
>> +
>> +			/*
>> +			 * It is a quirk in Rockchip RK3288, causing by
>> +			 * a hardware bug. This will propagate out and
>> +			 * eventually we'll re-enumerate the device.
>> +			 * Not great but the best we can do
>
> Remove the trailing whitespaces in this comment.
>
> Run scripts/checkpatch.pl to catch these.
I see.
>
>> +			 */
>> +			if (of_device_is_compatible(np, "rockchip,rk3288-usb"))
>> +				hsotg->phy->ops->reset(hsotg->phy);
>> +
>
> You should probably check for NULL before calling the reset()
> callback.
Sure.
>
> Also, I'd rather see a generic quirk property that you set for your
> platform.
>
> Something like "phy_reset_on_wakeup_quirk".
But Rob Herring want me to implied by the SoC specific compatible 
string. I agree with him. It is certainly bug in RK3288 platform.
It is no found no the other platform.
>
> Also, try to preserve the version tag in your subject for all the
> patches so that we can easily identify the latest version of the
> series, like:
>
> [PATCH v5 3/4] ...
>
> And, typically "RESEND" means there are no code change.
I see, I just make up my mind whether code style is new version
before. I won't again.

Thank you very much.
>
> Regards,
> John
>
> _______________________________________________
> Linux-rockchip mailing list
> Linux-rockchip@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-rockchip
>

-- 
Randy Li
The third produce department

Re: [RESEND PATCH 3/4] usb: dwc2: assert phy reset when waking up in rk3288 platform

[John Youn]

On 8/24/2016 1:54 AM, Randy Li wrote:
> 
> 
> On 08/24/2016 04:46 AM, John Youn wrote:
>> On 8/21/2016 12:32 PM, Randy Li wrote:
>>> On the rk3288 USB host-only port (the one that's not the OTG-enabled
>>> port) the PHY can get into a bad state when a wakeup is asserted (not
>>> just a wakeup from full system suspend but also a wakeup from
>>> autosuspend).
>>>
>>> We can get the PHY out of its bad state by asserting its "port reset",
>>> but unfortunately that seems to assert a reset onto the USB bus so it
>>> could confuse things if we don't actually deenumerate / reenumerate the
>>> device.
>>>
>>> We can also get the PHY out of its bad state by fully resetting it using
>>> the reset from the CRU (clock reset unit) in chip, which does a more full
>>> reset.  The CRU-based reset appears to actually cause devices on the bus
>>> to be removed and reinserted, which fixes the problem (albeit in a hacky
>>> way).
>>>
>>> It's unfortunate that we need to do a full re-enumeration of devices at
>>> wakeup time, but this is better than alternative of letting the bus get
>>> wedged.
>>>
>>> Signed-off-by: Randy Li <ayaka@soulik.info>
>>> ---
>>>  drivers/usb/dwc2/core_intr.c | 11 +++++++++++
>>>  1 file changed, 11 insertions(+)
>>>
>>> diff --git a/drivers/usb/dwc2/core_intr.c b/drivers/usb/dwc2/core_intr.c
>>> index d85c5c9..f57c48a 100644
>>> --- a/drivers/usb/dwc2/core_intr.c
>>> +++ b/drivers/usb/dwc2/core_intr.c
>>> @@ -345,6 +345,7 @@ static void dwc2_handle_session_req_intr(struct dwc2_hsotg *hsotg)
>>>  static void dwc2_handle_wakeup_detected_intr(struct dwc2_hsotg *hsotg)
>>>  {
>>>  	int ret;
>>> +	struct device_node *np = hsotg->dev->of_node;
>>>
>>>  	/* Clear interrupt */
>>>  	dwc2_writel(GINTSTS_WKUPINT, hsotg->regs + GINTSTS);
>>> @@ -379,6 +380,16 @@ static void dwc2_handle_wakeup_detected_intr(struct dwc2_hsotg *hsotg)
>>>  			/* Restart the Phy Clock */
>>>  			pcgcctl &= ~PCGCTL_STOPPCLK;
>>>  			dwc2_writel(pcgcctl, hsotg->regs + PCGCTL);
>>> +
>>> +			/*
>>> +			 * It is a quirk in Rockchip RK3288, causing by
>>> +			 * a hardware bug. This will propagate out and
>>> +			 * eventually we'll re-enumerate the device.
>>> +			 * Not great but the best we can do
>>
>> Remove the trailing whitespaces in this comment.
>>
>> Run scripts/checkpatch.pl to catch these.
> I see.
>>
>>> +			 */
>>> +			if (of_device_is_compatible(np, "rockchip,rk3288-usb"))
>>> +				hsotg->phy->ops->reset(hsotg->phy);
>>> +
>>
>> You should probably check for NULL before calling the reset()
>> callback.
> Sure.
>>
>> Also, I'd rather see a generic quirk property that you set for your
>> platform.
>>
>> Something like "phy_reset_on_wakeup_quirk".
> But Rob Herring want me to implied by the SoC specific compatible 
> string. I agree with him. It is certainly bug in RK3288 platform.
> It is no found no the other platform.

Ok, I missed that before.

Based on the drivers I'm familiar with (like dwc3), you would typically add a
"quirk" anyways.

Felipe,

Do you have some policy or preference on this?

Regards,
John

Re: [RESEND PATCH 3/4] usb: dwc2: assert phy reset when waking up in rk3288 platform

[Felipe Balbi]

[-- Attachment #1: Type: text/plain, Size: 1146 bytes --]


Hi,

John Youn <John.Youn@synopsys.com> writes:

[...]

>>>> +			 */
>>>> +			if (of_device_is_compatible(np, "rockchip,rk3288-usb"))
>>>> +				hsotg->phy->ops->reset(hsotg->phy);
>>>> +
>>>
>>> You should probably check for NULL before calling the reset()
>>> callback.
>> Sure.
>>>
>>> Also, I'd rather see a generic quirk property that you set for your
>>> platform.
>>>
>>> Something like "phy_reset_on_wakeup_quirk".
>> But Rob Herring want me to implied by the SoC specific compatible 
>> string. I agree with him. It is certainly bug in RK3288 platform.
>> It is no found no the other platform.
>
> Ok, I missed that before.
>
> Based on the drivers I'm familiar with (like dwc3), you would
> typically add a "quirk" anyways.
>
> Felipe,
>
> Do you have some policy or preference on this?

if it's not a dwc2-generic feature, then let's do it via compatible
flag, sure. What we don't want is for things like:

if (is_compatible('synopsys') || is_compatible('rockchip') ||
	is_compatible('foobar') ... )

For that, we'd be better of adding a generic quirk flag which several
can use.

-- 
balbi

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 800 bytes --]

Re: [PATCH 2/4] phy: rockchip-usb: use rockchip_usb_phy_reset to reset phy during wakeup

[Kishon Vijay Abraham I]

Hi,

On Monday 22 August 2016 06:30 PM, Heiko Stübner wrote:
> Am Montag, 22. August 2016, 17:17:41 schrieb Kishon Vijay Abraham I:
>> Hi,
>>
>> On Sunday 21 August 2016 02:02 AM, Randy Li wrote:
>>> It is a hardware bug in RK3288, the only way to solve it is to
>>> reset the phy.
>>>
>>> Signed-off-by: Randy Li <ayaka@soulik.info>
>>> ---
>>>
>>>  drivers/phy/phy-rockchip-usb.c | 20 ++++++++++++++++++++
>>>  1 file changed, 20 insertions(+)
>>>
>>> diff --git a/drivers/phy/phy-rockchip-usb.c
>>> b/drivers/phy/phy-rockchip-usb.c index 2a7381f..734987f 100644
>>> --- a/drivers/phy/phy-rockchip-usb.c
>>> +++ b/drivers/phy/phy-rockchip-usb.c
>>> @@ -29,6 +29,7 @@
>>>
>>>  #include <linux/reset.h>
>>>  #include <linux/regmap.h>
>>>  #include <linux/mfd/syscon.h>
>>>
>>> +#include <linux/delay.h>
>>>
>>>  static int enable_usb_uart;
>>>
>>> @@ -64,6 +65,7 @@ struct rockchip_usb_phy {
>>>
>>>  	struct clk_hw	clk480m_hw;
>>>  	struct phy	*phy;
>>>  	bool		uart_enabled;
>>>
>>> +	struct reset_control *reset;
>>>
>>>  };
>>>  
>>>  static int rockchip_usb_phy_power(struct rockchip_usb_phy *phy,
>>>
>>> @@ -144,9 +146,23 @@ static int rockchip_usb_phy_power_on(struct phy
>>> *_phy)
>>>
>>>  	return clk_prepare_enable(phy->clk480m);
>>>  
>>>  }
>>>
>>> +static int rockchip_usb_phy_reset(struct phy *_phy)
>>> +{
>>> +	struct rockchip_usb_phy *phy = phy_get_drvdata(_phy);
>>> +
>>> +	if (phy->reset) {
>>> +		reset_control_assert(phy->reset);
>>> +		udelay(10);
>>> +		reset_control_deassert(phy->reset);
>>> +	}
>>> +
>>> +	return 0;
>>> +}
>>> +
>>>
>>>  static const struct phy_ops ops = {
>>>  
>>>  	.power_on	= rockchip_usb_phy_power_on,
>>>  	.power_off	= rockchip_usb_phy_power_off,
>>>
>>> +	.reset		= rockchip_usb_phy_reset,
>>
>> why not just reuse the .init ops? reset can be done during initialization
>> right?
> 
> The naming of power_on + power_off and init + exit probably suggests that they 
> are supposed to be used in pairs. (aka module_init + module_exit and probably 
> more)
> 
> But in fact I've seen different combinations so far (phy_init + phy_power_on 
> ... phy_power_off + phy_exit but also phy_power_on + phy_init ... phy_exit + 
> phy_power_off), so I guess the semantics are not that strictly defined.

yeah, it got difficult to maintain the semantics once the number of phy users
increased. Thought it was better than bloating the framework with platform
specific callbacks.

Thanks
Kishon

Re: [RESEND PATCH 3/4] usb: dwc2: assert phy reset when waking up in rk3288 platform

[John Youn]

On 8/29/2016 12:51 AM, Felipe Balbi wrote:
> 
> Hi,
> 
> John Youn <John.Youn@synopsys.com> writes:
> 
> [...]
> 
>>>>> +			 */
>>>>> +			if (of_device_is_compatible(np, "rockchip,rk3288-usb"))
>>>>> +				hsotg->phy->ops->reset(hsotg->phy);
>>>>> +
>>>>
>>>> You should probably check for NULL before calling the reset()
>>>> callback.
>>> Sure.
>>>>
>>>> Also, I'd rather see a generic quirk property that you set for your
>>>> platform.
>>>>
>>>> Something like "phy_reset_on_wakeup_quirk".
>>> But Rob Herring want me to implied by the SoC specific compatible 
>>> string. I agree with him. It is certainly bug in RK3288 platform.
>>> It is no found no the other platform.
>>
>> Ok, I missed that before.
>>
>> Based on the drivers I'm familiar with (like dwc3), you would
>> typically add a "quirk" anyways.
>>
>> Felipe,
>>
>> Do you have some policy or preference on this?
> 
> if it's not a dwc2-generic feature, then let's do it via compatible
> flag, sure. What we don't want is for things like:
> 
> if (is_compatible('synopsys') || is_compatible('rockchip') ||
> 	is_compatible('foobar') ... )
> 
> For that, we'd be better of adding a generic quirk flag which several
> can use.
> 

Alright sounds reasonable.

Randy, could you respin with the other feedback? There's no need to
add a quirk.

Regards,
John