usb: dwc3: Support wake-up from USB suspend.

The current USB gadget driver behaviour is to stop the controller and disconnect from the bus during System sleep.

The 'snps,gadget-keep-connect-sys-sleep' property can be used to change this behaviour and keep the controller active and connected to the bus during System sleep. This is useful for applications that want to enter a low power state when USB is suspended but remain connected so they can resume activity on USB resume.

This feature introduces a new constraint if Gadget driver is connected to USB host: i.e. the gadget must be in USB suspend state to allow a System sleep as we cannot process any USB transactions when in System sleep.

The system hardware is responsible to detect the end of USB suspend and wake up the system so we can begin processing the USB transactions as soon as possible.

Cc [email protected]

Signed-off-by Roger Quadros <[email protected]>

---
 Documentation/devicetree/bindings/usb/snps,dwc3.yaml | 5 +++++
 1 file changed, 5 insertions(+)

The current USB gadget driver behaviour is to stop the controller and disconnect from the bus during System sleep.

The 'snps,gadget-keep-connect-sys-sleep' property can be used to change this behaviour and keep the controller active and connected to the bus during System sleep. This is useful for applications that want to enter a low power state when USB is suspended but remain connected so they can resume activity on USB resume.

This feature introduces a new constraint if Gadget driver is connected to USB host: i.e. the gadget must be in USB suspend state to allow a System sleep as we cannot process any USB transactions when in System sleep.

The system hardware is responsible to detect the end of USB suspend and wake up the system so we can begin processing the USB transactions as soon as possible.

Cc [email protected]

Signed-off-by Roger Quadros <[email protected]>

---
 Documentation/devicetree/bindings/usb/snps,dwc3.yaml | 5 +++++
 1 file changed, 5 insertions(+)

Documentation/devicetree/bindings/usb/snps,dwc3.yaml +0 -0

index be36956af53b..1ce8008e7fef 100644

Documentation/devicetree/bindings/usb/snps,dwc3.yaml +1 -0

+
   snps,hird-threshold:
     description: HIRD threshold
     $ref: /schemas/types.yaml#/definitions/uint8

Signature · 1 line

2.34.1

Quoted 2 lines · Roger Quadros

The current USB gadget driver behaviour is to stop the controller and disconnect from the bus during System sleep.

What's USB gadget? ;)

Quoted 12 lines

The 'snps,gadget-keep-connect-sys-sleep' property can be used to change this behaviour and keep the controller active and connected to the bus during System sleep. This is useful for applications that want to enter a low power state when USB is suspended but remain connected so they can resume activity on USB resume.

This feature introduces a new constraint if Gadget driver is connected to USB host: i.e. the gadget must be in USB suspend state to allow a System sleep as we cannot process any USB transactions when in System sleep.

Quoted 3 lines

The system hardware is responsible to detect the end of USB suspend and wake up the system so we can begin processing the USB transactions as soon as possible.

Sounds like something the user/OS would want to choose rather than fixed by your board's firmware.

Is this somehow DWC3 specific? If not, why a DWC3 specific property?

Quoted 16 lines

Cc [email protected]

Signed-off-by Roger Quadros <[email protected]>

---
 Documentation/devicetree/bindings/usb/snps,dwc3.yaml | 5 +++++
 1 file changed, 5 insertions(+)

Documentation/devicetree/bindings/usb/snps,dwc3.yaml +0 -0

index be36956af53b..1ce8008e7fef 100644
@@ -262,6 +262,11 @@ properties:
       asserts utmi_sleep_n.
     type: boolean
 

Documentation/devicetree/bindings/usb/snps,dwc3.yaml properties: · line 268

Quoted 4 lines

Documentation/devicetree/bindings/usb/snps,dwc3.yaml +4 -0

+  snps,gadget-keep-connect-sys-sleep:
+    description:
+      If True then gadget driver will not disconnect during system sleep.
+      System sleep will not be allowed if gadget is not already in USB suspend.

'gadget' is a Linuxism.

Quoted 8 lines

Documentation/devicetree/bindings/usb/snps,dwc3.yaml +1 -0

+
   snps,hird-threshold:
     description: HIRD threshold
     $ref: /schemas/types.yaml#/definitions/uint8

Signature · 1 line

2.34.1

Implement 'snps,gadget-keep-connect-sys-sleep' property.

Do not stop the gadget controller and disconnect if this property is present and we are connected to a USB Host.

Prevent System sleep if Gadget is not in USB suspend.

Signed-off-by Roger Quadros <[email protected]>

---
 drivers/usb/dwc3/core.c   | 25 +++++++++++++++++++------
 drivers/usb/dwc3/core.h   |  2 ++
 drivers/usb/dwc3/gadget.c | 25 +++++++++++++++++++++++--
 3 files changed, 44 insertions(+), 8 deletions(-)

drivers/usb/dwc3/core.c +19 -6

index 476b63618511..a47bbaa27302 100644
@@ -1575,6 +1575,9 @@ static void dwc3_get_properties(struct dwc3 *dwc)
 	dwc->dis_split_quirk = device_property_read_bool(dev,
 				"snps,dis-split-quirk");
 
+	dwc->gadget_keep_connect_sys_sleep = device_property_read_bool(dev,
+				"snps,gadget-keep-connect-sys-sleep");
+
 	dwc->lpm_nyet_threshold = lpm_nyet_threshold;
 	dwc->tx_de_emphasis = tx_de_emphasis;
 

@@ -2027,14 +2030,20 @@ static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
 {
 	unsigned long	flags;
 	u32 reg;
+	int ret;
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
 		if (pm_runtime_suspended(dwc->dev))
 			break;
-		dwc3_gadget_suspend(dwc);
+		ret = dwc3_gadget_suspend(dwc);
+		if (ret) {
+			dev_err(dwc->dev, "gadget not suspended: %d\n", ret);
+			return ret;
+		}
 		synchronize_irq(dwc->irq_gadget);
-		dwc3_core_exit(dwc);
+		if(!dwc->gadget_keep_connect_sys_sleep)
+			dwc3_core_exit(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:
 		if (!PMSG_IS_AUTO(msg) && !device_may_wakeup(dwc->dev)) {

@@ -2088,11 +2097,15 @@ static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg)
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
-		ret = dwc3_core_init_for_resume(dwc);
-		if (ret)
-			return ret;
+		if (!dwc->gadget_keep_connect_sys_sleep)
+		{
+			ret = dwc3_core_init_for_resume(dwc);
+			if (ret)
+				return ret;
+
+			dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
+		}
 
-		dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
 		dwc3_gadget_resume(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:

drivers/usb/dwc3/core.h +2 -0

index 582ebd9cf9c2..f84bac815bed 100644
@@ -1328,6 +1328,8 @@ struct dwc3 {
 	unsigned		dis_split_quirk:1;
 	unsigned		async_callbacks:1;
 
+	unsigned		gadget_keep_connect_sys_sleep:1;
+
 	u16			imod_interval;
 
 	int			max_cfg_eps;

drivers/usb/dwc3/gadget.c +1 -0

index 3c63fa97a680..8062e44f63f6 100644
@@ -4572,12 +4572,23 @@ void dwc3_gadget_exit(struct dwc3 *dwc)
 int dwc3_gadget_suspend(struct dwc3 *dwc)
 {
 	unsigned long flags;
+	int link_state;
 
 	if (!dwc->gadget_driver)
 		return 0;

drivers/usb/dwc3/gadget.c +14 -1

+		return 0;
+	}
 
+	dwc3_gadget_run_stop(dwc, false, false);
 	spin_lock_irqsave(&dwc->lock, flags);
 	dwc3_disconnect_gadget(dwc);
 	__dwc3_gadget_stop(dwc);

@@ -4588,11 +4599,21 @@ int dwc3_gadget_suspend(struct dwc3 *dwc)
 
 int dwc3_gadget_resume(struct dwc3 *dwc)
 {
-	int			ret;
+	int ret;
+	irqreturn_t irq_t;
 
 	if (!dwc->gadget_driver || !dwc->softconnect)
 		return 0;
 
+	if (dwc->gadget_keep_connect_sys_sleep && dwc->connected) {
+		dwc3_gadget_enable_irq(dwc);
+		/* check pending events */
+		irq_t = dwc3_check_event_buf(dwc->ev_buf);
+		if (irq_t == IRQ_WAKE_THREAD)
+			dwc3_process_event_buf(dwc->ev_buf);
+		return 0;
+	}
+
 	ret = __dwc3_gadget_start(dwc);
 	if (ret < 0)
 		goto err0;

Signature · 1 line

2.34.1

Hi,

Quoted 97 lines · Roger Quadros

Implement 'snps,gadget-keep-connect-sys-sleep' property.

Do not stop the gadget controller and disconnect if this property is present and we are connected to a USB Host.

Prevent System sleep if Gadget is not in USB suspend.

Signed-off-by Roger Quadros <[email protected]>

---
 drivers/usb/dwc3/core.c   | 25 +++++++++++++++++++------
 drivers/usb/dwc3/core.h   |  2 ++
 drivers/usb/dwc3/gadget.c | 25 +++++++++++++++++++++++--
 3 files changed, 44 insertions(+), 8 deletions(-)

drivers/usb/dwc3/core.c +19 -6

index 476b63618511..a47bbaa27302 100644
@@ -1575,6 +1575,9 @@ static void dwc3_get_properties(struct dwc3 *dwc)
 	dwc->dis_split_quirk = device_property_read_bool(dev,
 				"snps,dis-split-quirk");
 
+	dwc->gadget_keep_connect_sys_sleep = device_property_read_bool(dev,
+				"snps,gadget-keep-connect-sys-sleep");
+
 	dwc->lpm_nyet_threshold = lpm_nyet_threshold;
 	dwc->tx_de_emphasis = tx_de_emphasis;
 

@@ -2027,14 +2030,20 @@ static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
 {
 	unsigned long	flags;
 	u32 reg;
+	int ret;
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
 		if (pm_runtime_suspended(dwc->dev))
 			break;
-		dwc3_gadget_suspend(dwc);
+		ret = dwc3_gadget_suspend(dwc);
+		if (ret) {
+			dev_err(dwc->dev, "gadget not suspended: %d\n", ret);
+			return ret;
+		}
 		synchronize_irq(dwc->irq_gadget);
-		dwc3_core_exit(dwc);
+		if(!dwc->gadget_keep_connect_sys_sleep)
+			dwc3_core_exit(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:
 		if (!PMSG_IS_AUTO(msg) && !device_may_wakeup(dwc->dev)) {

@@ -2088,11 +2097,15 @@ static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg)
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
-		ret = dwc3_core_init_for_resume(dwc);
-		if (ret)
-			return ret;
+		if (!dwc->gadget_keep_connect_sys_sleep)
+		{
+			ret = dwc3_core_init_for_resume(dwc);
+			if (ret)
+				return ret;
+
+			dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
+		}
 
-		dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
 		dwc3_gadget_resume(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:

drivers/usb/dwc3/core.h +2 -0

index 582ebd9cf9c2..f84bac815bed 100644
@@ -1328,6 +1328,8 @@ struct dwc3 {
 	unsigned		dis_split_quirk:1;
 	unsigned		async_callbacks:1;
 
+	unsigned		gadget_keep_connect_sys_sleep:1;
+
 	u16			imod_interval;
 
 	int			max_cfg_eps;

drivers/usb/dwc3/gadget.c +1 -0

index 3c63fa97a680..8062e44f63f6 100644
@@ -4572,12 +4572,23 @@ void dwc3_gadget_exit(struct dwc3 *dwc)
 int dwc3_gadget_suspend(struct dwc3 *dwc)
 {
 	unsigned long flags;
+	int link_state;
 
 	if (!dwc->gadget_driver)
 		return 0;
 

drivers/usb/dwc3/gadget.c void dwc3_gadget_exit(struct dwc3 *dwc) · line 4587

Quoted 9 lines

drivers/usb/dwc3/gadget.c +8 -1

-	dwc3_gadget_run_stop(dwc, false, false);
+	if (dwc->gadget_keep_connect_sys_sleep && dwc->connected) {
+		link_state = dwc3_gadget_get_link_state(dwc);
+		/* Prevent PM Sleep if not in U3/L2 */
+		if (link_state != DWC3_LINK_STATE_U3)
+			return -EBUSY;
+
+		/* don't stop/disconnect */
+		dwc3_gadget_disable_irq(dwc);

We shouldn't disable event interrupt here. What will happen if the

device is disconnected and reconnect to the host while the device is still in system suspend? The host would not be able to communicate with the device then.

BR,
Thinh

Quoted 32 lines

drivers/usb/dwc3/gadget.c +14 -1

+		return 0;
+	}
 
+	dwc3_gadget_run_stop(dwc, false, false);
 	spin_lock_irqsave(&dwc->lock, flags);
 	dwc3_disconnect_gadget(dwc);
 	__dwc3_gadget_stop(dwc);

@@ -4588,11 +4599,21 @@ int dwc3_gadget_suspend(struct dwc3 *dwc)
 
 int dwc3_gadget_resume(struct dwc3 *dwc)
 {
-	int			ret;
+	int ret;
+	irqreturn_t irq_t;
 
 	if (!dwc->gadget_driver || !dwc->softconnect)
 		return 0;
 
+	if (dwc->gadget_keep_connect_sys_sleep && dwc->connected) {
+		dwc3_gadget_enable_irq(dwc);
+		/* check pending events */
+		irq_t = dwc3_check_event_buf(dwc->ev_buf);
+		if (irq_t == IRQ_WAKE_THREAD)
+			dwc3_process_event_buf(dwc->ev_buf);
+		return 0;
+	}
+
 	ret = __dwc3_gadget_start(dwc);
 	if (ret < 0)
 		goto err0;

Signature · 1 line

2.34.1

Hi Thinh,

Quoted 100 lines · Thinh Nguyen

Hi,

Quoted 97 lines · Roger Quadros

Implement 'snps,gadget-keep-connect-sys-sleep' property.

Do not stop the gadget controller and disconnect if this property is present and we are connected to a USB Host.

Prevent System sleep if Gadget is not in USB suspend.

Signed-off-by Roger Quadros <[email protected]>

---
 drivers/usb/dwc3/core.c   | 25 +++++++++++++++++++------
 drivers/usb/dwc3/core.h   |  2 ++
 drivers/usb/dwc3/gadget.c | 25 +++++++++++++++++++++++--
 3 files changed, 44 insertions(+), 8 deletions(-)

drivers/usb/dwc3/core.c +19 -6

index 476b63618511..a47bbaa27302 100644
@@ -1575,6 +1575,9 @@ static void dwc3_get_properties(struct dwc3 *dwc)
 	dwc->dis_split_quirk = device_property_read_bool(dev,
 				"snps,dis-split-quirk");
 
+	dwc->gadget_keep_connect_sys_sleep = device_property_read_bool(dev,
+				"snps,gadget-keep-connect-sys-sleep");
+
 	dwc->lpm_nyet_threshold = lpm_nyet_threshold;
 	dwc->tx_de_emphasis = tx_de_emphasis;
 

@@ -2027,14 +2030,20 @@ static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
 {
 	unsigned long	flags;
 	u32 reg;
+	int ret;
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
 		if (pm_runtime_suspended(dwc->dev))
 			break;
-		dwc3_gadget_suspend(dwc);
+		ret = dwc3_gadget_suspend(dwc);
+		if (ret) {
+			dev_err(dwc->dev, "gadget not suspended: %d\n", ret);
+			return ret;
+		}
 		synchronize_irq(dwc->irq_gadget);
-		dwc3_core_exit(dwc);
+		if(!dwc->gadget_keep_connect_sys_sleep)
+			dwc3_core_exit(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:
 		if (!PMSG_IS_AUTO(msg) && !device_may_wakeup(dwc->dev)) {

@@ -2088,11 +2097,15 @@ static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg)
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
-		ret = dwc3_core_init_for_resume(dwc);
-		if (ret)
-			return ret;
+		if (!dwc->gadget_keep_connect_sys_sleep)
+		{
+			ret = dwc3_core_init_for_resume(dwc);
+			if (ret)
+				return ret;
+
+			dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
+		}
 
-		dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
 		dwc3_gadget_resume(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:

drivers/usb/dwc3/core.h +2 -0

index 582ebd9cf9c2..f84bac815bed 100644
@@ -1328,6 +1328,8 @@ struct dwc3 {
 	unsigned		dis_split_quirk:1;
 	unsigned		async_callbacks:1;
 
+	unsigned		gadget_keep_connect_sys_sleep:1;
+
 	u16			imod_interval;
 
 	int			max_cfg_eps;

drivers/usb/dwc3/gadget.c +1 -0

index 3c63fa97a680..8062e44f63f6 100644
@@ -4572,12 +4572,23 @@ void dwc3_gadget_exit(struct dwc3 *dwc)
 int dwc3_gadget_suspend(struct dwc3 *dwc)
 {
 	unsigned long flags;
+	int link_state;
 
 	if (!dwc->gadget_driver)
 		return 0;
 

Quoted 11 lines

drivers/usb/dwc3/gadget.c void dwc3_gadget_exit(struct dwc3 *dwc) · line 4587

Quoted 9 lines

drivers/usb/dwc3/gadget.c +8 -1

-	dwc3_gadget_run_stop(dwc, false, false);
+	if (dwc->gadget_keep_connect_sys_sleep && dwc->connected) {
+		link_state = dwc3_gadget_get_link_state(dwc);
+		/* Prevent PM Sleep if not in U3/L2 */
+		if (link_state != DWC3_LINK_STATE_U3)
+			return -EBUSY;
+
+		/* don't stop/disconnect */
+		dwc3_gadget_disable_irq(dwc);

We shouldn't disable event interrupt here. What will happen if the

Due to some reason, if I don't disable the event interrupts here then after USB resume the USB controller is malfunctioning.
It no longer responds to any requests from Host.

Quoted 4 lines

device is disconnected and reconnect to the host while the device is still in system suspend? The host would not be able to communicate with the device then.

In the TI platform, The system is woken up on any VBUS/linestate change and in dwc3_gadget_resume we enable the events again and check for pending events. Is it pointless to check for pending events there?

Quoted 115 lines · Roger Quadros

Hi Thinh,

Quoted 100 lines · Thinh Nguyen

Hi,

Quoted 97 lines · Roger Quadros

Implement 'snps,gadget-keep-connect-sys-sleep' property.

Do not stop the gadget controller and disconnect if this property is present and we are connected to a USB Host.

Prevent System sleep if Gadget is not in USB suspend.

Signed-off-by Roger Quadros <[email protected]>

---
 drivers/usb/dwc3/core.c   | 25 +++++++++++++++++++------
 drivers/usb/dwc3/core.h   |  2 ++
 drivers/usb/dwc3/gadget.c | 25 +++++++++++++++++++++++--
 3 files changed, 44 insertions(+), 8 deletions(-)

drivers/usb/dwc3/core.c +19 -6

index 476b63618511..a47bbaa27302 100644
@@ -1575,6 +1575,9 @@ static void dwc3_get_properties(struct dwc3 *dwc)
 	dwc->dis_split_quirk = device_property_read_bool(dev,
 				"snps,dis-split-quirk");
 
+	dwc->gadget_keep_connect_sys_sleep = device_property_read_bool(dev,
+				"snps,gadget-keep-connect-sys-sleep");
+
 	dwc->lpm_nyet_threshold = lpm_nyet_threshold;
 	dwc->tx_de_emphasis = tx_de_emphasis;
 

@@ -2027,14 +2030,20 @@ static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
 {
 	unsigned long	flags;
 	u32 reg;
+	int ret;
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
 		if (pm_runtime_suspended(dwc->dev))
 			break;
-		dwc3_gadget_suspend(dwc);
+		ret = dwc3_gadget_suspend(dwc);
+		if (ret) {
+			dev_err(dwc->dev, "gadget not suspended: %d\n", ret);
+			return ret;
+		}
 		synchronize_irq(dwc->irq_gadget);
-		dwc3_core_exit(dwc);
+		if(!dwc->gadget_keep_connect_sys_sleep)
+			dwc3_core_exit(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:
 		if (!PMSG_IS_AUTO(msg) && !device_may_wakeup(dwc->dev)) {

@@ -2088,11 +2097,15 @@ static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg)
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
-		ret = dwc3_core_init_for_resume(dwc);
-		if (ret)
-			return ret;
+		if (!dwc->gadget_keep_connect_sys_sleep)
+		{
+			ret = dwc3_core_init_for_resume(dwc);
+			if (ret)
+				return ret;
+
+			dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
+		}
 
-		dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
 		dwc3_gadget_resume(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:

drivers/usb/dwc3/core.h +2 -0

index 582ebd9cf9c2..f84bac815bed 100644
@@ -1328,6 +1328,8 @@ struct dwc3 {
 	unsigned		dis_split_quirk:1;
 	unsigned		async_callbacks:1;
 
+	unsigned		gadget_keep_connect_sys_sleep:1;
+
 	u16			imod_interval;
 
 	int			max_cfg_eps;

drivers/usb/dwc3/gadget.c +1 -0

index 3c63fa97a680..8062e44f63f6 100644
@@ -4572,12 +4572,23 @@ void dwc3_gadget_exit(struct dwc3 *dwc)
 int dwc3_gadget_suspend(struct dwc3 *dwc)
 {
 	unsigned long flags;
+	int link_state;
 
 	if (!dwc->gadget_driver)
 		return 0;
 

Quoted 11 lines

drivers/usb/dwc3/gadget.c void dwc3_gadget_exit(struct dwc3 *dwc) · line 4587

Quoted 9 lines

drivers/usb/dwc3/gadget.c +8 -1

-	dwc3_gadget_run_stop(dwc, false, false);
+	if (dwc->gadget_keep_connect_sys_sleep && dwc->connected) {
+		link_state = dwc3_gadget_get_link_state(dwc);
+		/* Prevent PM Sleep if not in U3/L2 */
+		if (link_state != DWC3_LINK_STATE_U3)
+			return -EBUSY;
+
+		/* don't stop/disconnect */
+		dwc3_gadget_disable_irq(dwc);

We shouldn't disable event interrupt here. What will happen if the

Quoted 3 lines

Due to some reason, if I don't disable the event interrupts here then after USB resume the USB controller is malfunctioning.
It no longer responds to any requests from Host.

You should look into this. These events are important as they can tell whether the host initiates resume.

Quoted 9 lines

Quoted 4 lines

device is disconnected and reconnect to the host while the device is still in system suspend? The host would not be able to communicate with the device then.

In the TI platform, The system is woken up on any VBUS/linestate change and in dwc3_gadget_resume we enable the events again and check for pending events. Is it pointless to check for pending events there?

It seems fragile for the implementation to be dependent on platform specific feature right?

Also, what will happen in a typical case when the host puts the device in suspend and initiates resume while the device is in system suspend (and stay in suspend over a period of time)? There is no VBUS change. There will be problem if host detects no response from device in time.

Don't we need these events to wakeup the device?

Quoted 136 lines · Thinh Nguyen

Quoted 115 lines · Roger Quadros

Hi Thinh,

Quoted 100 lines · Thinh Nguyen

Hi,

Quoted 97 lines · Roger Quadros

Implement 'snps,gadget-keep-connect-sys-sleep' property.

Do not stop the gadget controller and disconnect if this property is present and we are connected to a USB Host.

Prevent System sleep if Gadget is not in USB suspend.

Signed-off-by Roger Quadros <[email protected]>

---
 drivers/usb/dwc3/core.c   | 25 +++++++++++++++++++------
 drivers/usb/dwc3/core.h   |  2 ++
 drivers/usb/dwc3/gadget.c | 25 +++++++++++++++++++++++--
 3 files changed, 44 insertions(+), 8 deletions(-)

drivers/usb/dwc3/core.c +19 -6

index 476b63618511..a47bbaa27302 100644
@@ -1575,6 +1575,9 @@ static void dwc3_get_properties(struct dwc3 *dwc)
 	dwc->dis_split_quirk = device_property_read_bool(dev,
 				"snps,dis-split-quirk");
 
+	dwc->gadget_keep_connect_sys_sleep = device_property_read_bool(dev,
+				"snps,gadget-keep-connect-sys-sleep");
+
 	dwc->lpm_nyet_threshold = lpm_nyet_threshold;
 	dwc->tx_de_emphasis = tx_de_emphasis;
 

@@ -2027,14 +2030,20 @@ static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
 {
 	unsigned long	flags;
 	u32 reg;
+	int ret;
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
 		if (pm_runtime_suspended(dwc->dev))
 			break;
-		dwc3_gadget_suspend(dwc);
+		ret = dwc3_gadget_suspend(dwc);
+		if (ret) {
+			dev_err(dwc->dev, "gadget not suspended: %d\n", ret);
+			return ret;
+		}
 		synchronize_irq(dwc->irq_gadget);
-		dwc3_core_exit(dwc);
+		if(!dwc->gadget_keep_connect_sys_sleep)
+			dwc3_core_exit(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:
 		if (!PMSG_IS_AUTO(msg) && !device_may_wakeup(dwc->dev)) {

@@ -2088,11 +2097,15 @@ static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg)
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
-		ret = dwc3_core_init_for_resume(dwc);
-		if (ret)
-			return ret;
+		if (!dwc->gadget_keep_connect_sys_sleep)
+		{
+			ret = dwc3_core_init_for_resume(dwc);
+			if (ret)
+				return ret;
+
+			dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
+		}
 
-		dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
 		dwc3_gadget_resume(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:

drivers/usb/dwc3/core.h +2 -0

index 582ebd9cf9c2..f84bac815bed 100644
@@ -1328,6 +1328,8 @@ struct dwc3 {
 	unsigned		dis_split_quirk:1;
 	unsigned		async_callbacks:1;
 
+	unsigned		gadget_keep_connect_sys_sleep:1;
+
 	u16			imod_interval;
 
 	int			max_cfg_eps;

drivers/usb/dwc3/gadget.c +1 -0

index 3c63fa97a680..8062e44f63f6 100644
@@ -4572,12 +4572,23 @@ void dwc3_gadget_exit(struct dwc3 *dwc)
 int dwc3_gadget_suspend(struct dwc3 *dwc)
 {
 	unsigned long flags;
+	int link_state;
 
 	if (!dwc->gadget_driver)
 		return 0;
 

Quoted 11 lines

drivers/usb/dwc3/gadget.c void dwc3_gadget_exit(struct dwc3 *dwc) · line 4587

Quoted 9 lines

drivers/usb/dwc3/gadget.c +8 -1

-	dwc3_gadget_run_stop(dwc, false, false);
+	if (dwc->gadget_keep_connect_sys_sleep && dwc->connected) {
+		link_state = dwc3_gadget_get_link_state(dwc);
+		/* Prevent PM Sleep if not in U3/L2 */
+		if (link_state != DWC3_LINK_STATE_U3)
+			return -EBUSY;
+
+		/* don't stop/disconnect */
+		dwc3_gadget_disable_irq(dwc);

We shouldn't disable event interrupt here. What will happen if the

Quoted 3 lines

Due to some reason, if I don't disable the event interrupts here then after USB resume the USB controller is malfunctioning.
It no longer responds to any requests from Host.

You should look into this. These events are important as they can tell whether the host initiates resume.

Quoted 9 lines

Quoted 4 lines

device is disconnected and reconnect to the host while the device is still in system suspend? The host would not be able to communicate with the device then.

In the TI platform, The system is woken up on any VBUS/linestate change and in dwc3_gadget_resume we enable the events again and check for pending events. Is it pointless to check for pending events there?

It seems fragile for the implementation to be dependent on platform specific feature right?

Quoted 6 lines

Also, what will happen in a typical case when the host puts the device in suspend and initiates resume while the device is in system suspend (and stay in suspend over a period of time)? There is no VBUS change. There will be problem if host detects no response from device in time.

Don't we need these events to wakeup the device?

Quoted 1 line

We may not be able to suspend everything in system suspend for this case. I'm thinking of treating these events as if they are PME to wakeup the device, but they are not the same. It may not be simple to handle this. The lower layers may need to stay awake for the dwc3 to handle these events. Hm... it gets a bit complicated.

Quoted 142 lines · Thinh Nguyen

Quoted 141 lines · Thinh Nguyen

Quoted 115 lines · Roger Quadros

Hi Thinh,

Quoted 100 lines · Thinh Nguyen

Hi,

Quoted 97 lines · Roger Quadros

Implement 'snps,gadget-keep-connect-sys-sleep' property.

Do not stop the gadget controller and disconnect if this property is present and we are connected to a USB Host.

Prevent System sleep if Gadget is not in USB suspend.

Signed-off-by Roger Quadros <[email protected]>

---
 drivers/usb/dwc3/core.c   | 25 +++++++++++++++++++------
 drivers/usb/dwc3/core.h   |  2 ++
 drivers/usb/dwc3/gadget.c | 25 +++++++++++++++++++++++--
 3 files changed, 44 insertions(+), 8 deletions(-)

drivers/usb/dwc3/core.c +19 -6

index 476b63618511..a47bbaa27302 100644
@@ -1575,6 +1575,9 @@ static void dwc3_get_properties(struct dwc3 *dwc)
 	dwc->dis_split_quirk = device_property_read_bool(dev,
 				"snps,dis-split-quirk");
 
+	dwc->gadget_keep_connect_sys_sleep = device_property_read_bool(dev,
+				"snps,gadget-keep-connect-sys-sleep");
+
 	dwc->lpm_nyet_threshold = lpm_nyet_threshold;
 	dwc->tx_de_emphasis = tx_de_emphasis;
 

@@ -2027,14 +2030,20 @@ static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
 {
 	unsigned long	flags;
 	u32 reg;
+	int ret;
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
 		if (pm_runtime_suspended(dwc->dev))
 			break;
-		dwc3_gadget_suspend(dwc);
+		ret = dwc3_gadget_suspend(dwc);
+		if (ret) {
+			dev_err(dwc->dev, "gadget not suspended: %d\n", ret);
+			return ret;
+		}
 		synchronize_irq(dwc->irq_gadget);
-		dwc3_core_exit(dwc);
+		if(!dwc->gadget_keep_connect_sys_sleep)
+			dwc3_core_exit(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:
 		if (!PMSG_IS_AUTO(msg) && !device_may_wakeup(dwc->dev)) {

@@ -2088,11 +2097,15 @@ static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg)
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
-		ret = dwc3_core_init_for_resume(dwc);
-		if (ret)
-			return ret;
+		if (!dwc->gadget_keep_connect_sys_sleep)
+		{
+			ret = dwc3_core_init_for_resume(dwc);
+			if (ret)
+				return ret;
+
+			dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
+		}
 
-		dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
 		dwc3_gadget_resume(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:

drivers/usb/dwc3/core.h +2 -0

index 582ebd9cf9c2..f84bac815bed 100644
@@ -1328,6 +1328,8 @@ struct dwc3 {
 	unsigned		dis_split_quirk:1;
 	unsigned		async_callbacks:1;
 
+	unsigned		gadget_keep_connect_sys_sleep:1;
+
 	u16			imod_interval;
 
 	int			max_cfg_eps;

drivers/usb/dwc3/gadget.c +1 -0

index 3c63fa97a680..8062e44f63f6 100644
@@ -4572,12 +4572,23 @@ void dwc3_gadget_exit(struct dwc3 *dwc)
 int dwc3_gadget_suspend(struct dwc3 *dwc)
 {
 	unsigned long flags;
+	int link_state;
 
 	if (!dwc->gadget_driver)
 		return 0;
 

Quoted 11 lines

drivers/usb/dwc3/gadget.c void dwc3_gadget_exit(struct dwc3 *dwc) · line 4587

Quoted 9 lines

drivers/usb/dwc3/gadget.c +8 -1

-	dwc3_gadget_run_stop(dwc, false, false);
+	if (dwc->gadget_keep_connect_sys_sleep && dwc->connected) {
+		link_state = dwc3_gadget_get_link_state(dwc);
+		/* Prevent PM Sleep if not in U3/L2 */
+		if (link_state != DWC3_LINK_STATE_U3)
+			return -EBUSY;
+
+		/* don't stop/disconnect */
+		dwc3_gadget_disable_irq(dwc);

We shouldn't disable event interrupt here. What will happen if the

Quoted 3 lines

Due to some reason, if I don't disable the event interrupts here then after USB resume the USB controller is malfunctioning.
It no longer responds to any requests from Host.

You should look into this. These events are important as they can tell whether the host initiates resume.

Quoted 9 lines

Quoted 4 lines

device is disconnected and reconnect to the host while the device is still in system suspend? The host would not be able to communicate with the device then.

In the TI platform, The system is woken up on any VBUS/linestate change and in dwc3_gadget_resume we enable the events again and check for pending events. Is it pointless to check for pending events there?

It seems fragile for the implementation to be dependent on platform specific feature right?

Also, what will happen in a typical case when the host puts the device in suspend and initiates resume while the device is in system suspend (and stay in suspend over a period of time)? There is no VBUS change. There will be problem if host detects no response from device in time.

Quoted 1 line

Quoted 1 line

Don't we need these events to wakeup the device?

That's why the TI implementation has line-state change detection to detect a USB resume. We are doing a out-of-band wake-up. The wake up events are configured in the wrapper driver (dwc3-am62.c).

Do you know of any dwc3 implementation that uses in-band mechanism to wake up the System. i.e. it relies on events enabled in DEVTEN register?

Quoted 7 lines

Quoted 1 line

We may not be able to suspend everything in system suspend for this case. I'm thinking of treating these events as if they are PME to wakeup the device, but they are not the same. It may not be simple to handle this. The lower layers may need to stay awake for the dwc3 to handle these events. Hm... it gets a bit complicated.

As we are going into suspend, we are not really in a position to handle any (DEVTEN) events till we have fully resumed.
So yes, we need to rely on platform specific implementation to wake the System on any USB event.

Quoted 143 lines · Roger Quadros

Quoted 142 lines · Thinh Nguyen

Quoted 141 lines · Thinh Nguyen

Quoted 115 lines · Roger Quadros

Hi Thinh,

Quoted 100 lines · Thinh Nguyen

Hi,

Quoted 97 lines · Roger Quadros

Implement 'snps,gadget-keep-connect-sys-sleep' property.

Do not stop the gadget controller and disconnect if this property is present and we are connected to a USB Host.

Prevent System sleep if Gadget is not in USB suspend.

Signed-off-by Roger Quadros <[email protected]>

---
 drivers/usb/dwc3/core.c   | 25 +++++++++++++++++++------
 drivers/usb/dwc3/core.h   |  2 ++
 drivers/usb/dwc3/gadget.c | 25 +++++++++++++++++++++++--
 3 files changed, 44 insertions(+), 8 deletions(-)

drivers/usb/dwc3/core.c +19 -6

index 476b63618511..a47bbaa27302 100644
@@ -1575,6 +1575,9 @@ static void dwc3_get_properties(struct dwc3 *dwc)
 	dwc->dis_split_quirk = device_property_read_bool(dev,
 				"snps,dis-split-quirk");
 
+	dwc->gadget_keep_connect_sys_sleep = device_property_read_bool(dev,
+				"snps,gadget-keep-connect-sys-sleep");
+
 	dwc->lpm_nyet_threshold = lpm_nyet_threshold;
 	dwc->tx_de_emphasis = tx_de_emphasis;
 

@@ -2027,14 +2030,20 @@ static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
 {
 	unsigned long	flags;
 	u32 reg;
+	int ret;
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
 		if (pm_runtime_suspended(dwc->dev))
 			break;
-		dwc3_gadget_suspend(dwc);
+		ret = dwc3_gadget_suspend(dwc);
+		if (ret) {
+			dev_err(dwc->dev, "gadget not suspended: %d\n", ret);
+			return ret;
+		}
 		synchronize_irq(dwc->irq_gadget);
-		dwc3_core_exit(dwc);
+		if(!dwc->gadget_keep_connect_sys_sleep)
+			dwc3_core_exit(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:
 		if (!PMSG_IS_AUTO(msg) && !device_may_wakeup(dwc->dev)) {

@@ -2088,11 +2097,15 @@ static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg)
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
-		ret = dwc3_core_init_for_resume(dwc);
-		if (ret)
-			return ret;
+		if (!dwc->gadget_keep_connect_sys_sleep)
+		{
+			ret = dwc3_core_init_for_resume(dwc);
+			if (ret)
+				return ret;
+
+			dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
+		}
 
-		dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
 		dwc3_gadget_resume(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:

drivers/usb/dwc3/core.h +2 -0

index 582ebd9cf9c2..f84bac815bed 100644
@@ -1328,6 +1328,8 @@ struct dwc3 {
 	unsigned		dis_split_quirk:1;
 	unsigned		async_callbacks:1;
 
+	unsigned		gadget_keep_connect_sys_sleep:1;
+
 	u16			imod_interval;
 
 	int			max_cfg_eps;

drivers/usb/dwc3/gadget.c +1 -0

index 3c63fa97a680..8062e44f63f6 100644
@@ -4572,12 +4572,23 @@ void dwc3_gadget_exit(struct dwc3 *dwc)
 int dwc3_gadget_suspend(struct dwc3 *dwc)
 {
 	unsigned long flags;
+	int link_state;
 
 	if (!dwc->gadget_driver)
 		return 0;
 

Quoted 11 lines

drivers/usb/dwc3/gadget.c void dwc3_gadget_exit(struct dwc3 *dwc) · line 4587

Quoted 9 lines

drivers/usb/dwc3/gadget.c +8 -1

-	dwc3_gadget_run_stop(dwc, false, false);
+	if (dwc->gadget_keep_connect_sys_sleep && dwc->connected) {
+		link_state = dwc3_gadget_get_link_state(dwc);
+		/* Prevent PM Sleep if not in U3/L2 */
+		if (link_state != DWC3_LINK_STATE_U3)
+			return -EBUSY;
+
+		/* don't stop/disconnect */
+		dwc3_gadget_disable_irq(dwc);

We shouldn't disable event interrupt here. What will happen if the

Quoted 3 lines

Due to some reason, if I don't disable the event interrupts here then after USB resume the USB controller is malfunctioning.
It no longer responds to any requests from Host.

You should look into this. These events are important as they can tell whether the host initiates resume.

Quoted 9 lines

Quoted 4 lines

device is disconnected and reconnect to the host while the device is still in system suspend? The host would not be able to communicate with the device then.

In the TI platform, The system is woken up on any VBUS/linestate change and in dwc3_gadget_resume we enable the events again and check for pending events. Is it pointless to check for pending events there?

It seems fragile for the implementation to be dependent on platform specific feature right?

Also, what will happen in a typical case when the host puts the device in suspend and initiates resume while the device is in system suspend (and stay in suspend over a period of time)? There is no VBUS change. There will be problem if host detects no response from device in time.

Quoted 9 lines

Quoted 1 line

Quoted 1 line

Don't we need these events to wakeup the device?

That's why the TI implementation has line-state change detection to detect a USB resume. We are doing a out-of-band wake-up. The wake up events are configured in the wrapper driver (dwc3-am62.c).

Do you know of any dwc3 implementation that uses in-band mechanism to wake up the System. i.e. it relies on events enabled in DEVTEN register?

We rely on PME. The PME is generated from the PMU of the usb controller when it detects a resume. If your platform supports hibernation and if the resume signal is connected to the lower layer power manager of your device, then you can wakeup the system one level at a time. For example, if your device is a pci device, that wakeup signal would tie to the pci power manager, waking up the pci layer before waking up the core of the usb controller. That's how the host wakes up the host system (e.g. from remote wakeup). For this to work, we expect something similar on the device side.

Quoted 8 lines

Quoted 7 lines

Quoted 1 line

We may not be able to suspend everything in system suspend for this case. I'm thinking of treating these events as if they are PME to wakeup the device, but they are not the same. It may not be simple to handle this. The lower layers may need to stay awake for the dwc3 to handle these events. Hm... it gets a bit complicated.

Quoted 5 lines

As we are going into suspend, we are not really in a position to handle any (DEVTEN) events till we have fully resumed.
So yes, we need to rely on platform specific implementation to wake the System on any USB event.

You may be able to detect vbus change through the connector controller. However, the usb controller is the one that detects host resume. What platform specific implementation do you have outside of the usb controller do you have to get around that?

I'm not sure if your platform supports hibernation or if the PME signal on your platform can wakeup the system, but currently dwc3 driver doesn't handle hibernation (device side). If there's no hibernation, there's no PME.

Quoted 183 lines · Thinh Nguyen

Quoted 143 lines · Roger Quadros

Quoted 142 lines · Thinh Nguyen

Quoted 141 lines · Thinh Nguyen

Quoted 115 lines · Roger Quadros

Hi Thinh,

Quoted 100 lines · Thinh Nguyen

Hi,

Quoted 97 lines · Roger Quadros

Implement 'snps,gadget-keep-connect-sys-sleep' property.

Do not stop the gadget controller and disconnect if this property is present and we are connected to a USB Host.

Prevent System sleep if Gadget is not in USB suspend.

Signed-off-by Roger Quadros <[email protected]>

---
 drivers/usb/dwc3/core.c   | 25 +++++++++++++++++++------
 drivers/usb/dwc3/core.h   |  2 ++
 drivers/usb/dwc3/gadget.c | 25 +++++++++++++++++++++++--
 3 files changed, 44 insertions(+), 8 deletions(-)

drivers/usb/dwc3/core.c +19 -6

index 476b63618511..a47bbaa27302 100644
@@ -1575,6 +1575,9 @@ static void dwc3_get_properties(struct dwc3 *dwc)
 	dwc->dis_split_quirk = device_property_read_bool(dev,
 				"snps,dis-split-quirk");
 
+	dwc->gadget_keep_connect_sys_sleep = device_property_read_bool(dev,
+				"snps,gadget-keep-connect-sys-sleep");
+
 	dwc->lpm_nyet_threshold = lpm_nyet_threshold;
 	dwc->tx_de_emphasis = tx_de_emphasis;
 

@@ -2027,14 +2030,20 @@ static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
 {
 	unsigned long	flags;
 	u32 reg;
+	int ret;
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
 		if (pm_runtime_suspended(dwc->dev))
 			break;
-		dwc3_gadget_suspend(dwc);
+		ret = dwc3_gadget_suspend(dwc);
+		if (ret) {
+			dev_err(dwc->dev, "gadget not suspended: %d\n", ret);
+			return ret;
+		}
 		synchronize_irq(dwc->irq_gadget);
-		dwc3_core_exit(dwc);
+		if(!dwc->gadget_keep_connect_sys_sleep)
+			dwc3_core_exit(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:
 		if (!PMSG_IS_AUTO(msg) && !device_may_wakeup(dwc->dev)) {

@@ -2088,11 +2097,15 @@ static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg)
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
-		ret = dwc3_core_init_for_resume(dwc);
-		if (ret)
-			return ret;
+		if (!dwc->gadget_keep_connect_sys_sleep)
+		{
+			ret = dwc3_core_init_for_resume(dwc);
+			if (ret)
+				return ret;
+
+			dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
+		}
 
-		dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
 		dwc3_gadget_resume(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:

drivers/usb/dwc3/core.h +2 -0

index 582ebd9cf9c2..f84bac815bed 100644
@@ -1328,6 +1328,8 @@ struct dwc3 {
 	unsigned		dis_split_quirk:1;
 	unsigned		async_callbacks:1;
 
+	unsigned		gadget_keep_connect_sys_sleep:1;
+
 	u16			imod_interval;
 
 	int			max_cfg_eps;

drivers/usb/dwc3/gadget.c +1 -0

index 3c63fa97a680..8062e44f63f6 100644
@@ -4572,12 +4572,23 @@ void dwc3_gadget_exit(struct dwc3 *dwc)
 int dwc3_gadget_suspend(struct dwc3 *dwc)
 {
 	unsigned long flags;
+	int link_state;
 
 	if (!dwc->gadget_driver)
 		return 0;
 

Quoted 11 lines

drivers/usb/dwc3/gadget.c void dwc3_gadget_exit(struct dwc3 *dwc) · line 4587

Quoted 9 lines

drivers/usb/dwc3/gadget.c +8 -1

-	dwc3_gadget_run_stop(dwc, false, false);
+	if (dwc->gadget_keep_connect_sys_sleep && dwc->connected) {
+		link_state = dwc3_gadget_get_link_state(dwc);
+		/* Prevent PM Sleep if not in U3/L2 */
+		if (link_state != DWC3_LINK_STATE_U3)
+			return -EBUSY;
+
+		/* don't stop/disconnect */
+		dwc3_gadget_disable_irq(dwc);

We shouldn't disable event interrupt here. What will happen if the

Quoted 3 lines

Due to some reason, if I don't disable the event interrupts here then after USB resume the USB controller is malfunctioning.
It no longer responds to any requests from Host.

You should look into this. These events are important as they can tell whether the host initiates resume.

Quoted 9 lines

Quoted 4 lines

device is disconnected and reconnect to the host while the device is still in system suspend? The host would not be able to communicate with the device then.

In the TI platform, The system is woken up on any VBUS/linestate change and in dwc3_gadget_resume we enable the events again and check for pending events. Is it pointless to check for pending events there?

It seems fragile for the implementation to be dependent on platform specific feature right?

Also, what will happen in a typical case when the host puts the device in suspend and initiates resume while the device is in system suspend (and stay in suspend over a period of time)? There is no VBUS change. There will be problem if host detects no response from device in time.

Quoted 9 lines

Quoted 1 line

Quoted 1 line

Don't we need these events to wakeup the device?

That's why the TI implementation has line-state change detection to detect a USB resume. We are doing a out-of-band wake-up. The wake up events are configured in the wrapper driver (dwc3-am62.c).

Do you know of any dwc3 implementation that uses in-band mechanism to wake up the System. i.e. it relies on events enabled in DEVTEN register?

We rely on PME. The PME is generated from the PMU of the usb controller when it detects a resume. If your platform supports hibernation and if the resume signal is connected to the lower layer power manager of your device, then you can wakeup the system one level at a time. For example, if your device is a pci device, that wakeup signal would tie to the pci power manager, waking up the pci layer before waking up the core of the usb controller. That's how the host wakes up the host system (e.g. from remote wakeup). For this to work, we expect something similar on the device side.

Quoted 8 lines

Quoted 7 lines

Quoted 1 line

We may not be able to suspend everything in system suspend for this case. I'm thinking of treating these events as if they are PME to wakeup the device, but they are not the same. It may not be simple to handle this. The lower layers may need to stay awake for the dwc3 to handle these events. Hm... it gets a bit complicated.

Quoted 5 lines

As we are going into suspend, we are not really in a position to handle any (DEVTEN) events till we have fully resumed.
So yes, we need to rely on platform specific implementation to wake the System on any USB event.

You may be able to detect vbus change through the connector controller. However, the usb controller is the one that detects host resume. What platform specific implementation do you have outside of the usb controller do you have to get around that?

Quoted 4 lines

I'm not sure if your platform supports hibernation or if the PME signal on your platform can wakeup the system, but currently dwc3 driver doesn't handle hibernation (device side). If there's no hibernation, there's no PME.

Quoted 1 line

Actually, I think the dwc3 core is still on during system suspend for you right? Then I think we can use the wakeup event to wakeup system suspend on host resume? You can ignore about PME in this case. You may need to look into what needs stay awake to allow for handling of the dwc3 event.

Quoted 184 lines · Thinh Nguyen

Quoted 183 lines · Thinh Nguyen

Quoted 143 lines · Roger Quadros

Quoted 142 lines · Thinh Nguyen

Quoted 141 lines · Thinh Nguyen

Quoted 115 lines · Roger Quadros

Hi Thinh,

Quoted 100 lines · Thinh Nguyen

Hi,

Quoted 97 lines · Roger Quadros

Implement 'snps,gadget-keep-connect-sys-sleep' property.

Do not stop the gadget controller and disconnect if this property is present and we are connected to a USB Host.

Prevent System sleep if Gadget is not in USB suspend.

Signed-off-by Roger Quadros <[email protected]>

---
 drivers/usb/dwc3/core.c   | 25 +++++++++++++++++++------
 drivers/usb/dwc3/core.h   |  2 ++
 drivers/usb/dwc3/gadget.c | 25 +++++++++++++++++++++++--
 3 files changed, 44 insertions(+), 8 deletions(-)

drivers/usb/dwc3/core.c +19 -6

index 476b63618511..a47bbaa27302 100644
@@ -1575,6 +1575,9 @@ static void dwc3_get_properties(struct dwc3 *dwc)
 	dwc->dis_split_quirk = device_property_read_bool(dev,
 				"snps,dis-split-quirk");
 
+	dwc->gadget_keep_connect_sys_sleep = device_property_read_bool(dev,
+				"snps,gadget-keep-connect-sys-sleep");
+
 	dwc->lpm_nyet_threshold = lpm_nyet_threshold;
 	dwc->tx_de_emphasis = tx_de_emphasis;
 

@@ -2027,14 +2030,20 @@ static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
 {
 	unsigned long	flags;
 	u32 reg;
+	int ret;
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
 		if (pm_runtime_suspended(dwc->dev))
 			break;
-		dwc3_gadget_suspend(dwc);
+		ret = dwc3_gadget_suspend(dwc);
+		if (ret) {
+			dev_err(dwc->dev, "gadget not suspended: %d\n", ret);
+			return ret;
+		}
 		synchronize_irq(dwc->irq_gadget);
-		dwc3_core_exit(dwc);
+		if(!dwc->gadget_keep_connect_sys_sleep)
+			dwc3_core_exit(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:
 		if (!PMSG_IS_AUTO(msg) && !device_may_wakeup(dwc->dev)) {

@@ -2088,11 +2097,15 @@ static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg)
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
-		ret = dwc3_core_init_for_resume(dwc);
-		if (ret)
-			return ret;
+		if (!dwc->gadget_keep_connect_sys_sleep)
+		{
+			ret = dwc3_core_init_for_resume(dwc);
+			if (ret)
+				return ret;
+
+			dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
+		}
 
-		dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
 		dwc3_gadget_resume(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:

drivers/usb/dwc3/core.h +2 -0

index 582ebd9cf9c2..f84bac815bed 100644
@@ -1328,6 +1328,8 @@ struct dwc3 {
 	unsigned		dis_split_quirk:1;
 	unsigned		async_callbacks:1;
 
+	unsigned		gadget_keep_connect_sys_sleep:1;
+
 	u16			imod_interval;
 
 	int			max_cfg_eps;

drivers/usb/dwc3/gadget.c +1 -0

index 3c63fa97a680..8062e44f63f6 100644
@@ -4572,12 +4572,23 @@ void dwc3_gadget_exit(struct dwc3 *dwc)
 int dwc3_gadget_suspend(struct dwc3 *dwc)
 {
 	unsigned long flags;
+	int link_state;
 
 	if (!dwc->gadget_driver)
 		return 0;
 

Quoted 11 lines

drivers/usb/dwc3/gadget.c void dwc3_gadget_exit(struct dwc3 *dwc) · line 4587

Quoted 9 lines

drivers/usb/dwc3/gadget.c +8 -1

-	dwc3_gadget_run_stop(dwc, false, false);
+	if (dwc->gadget_keep_connect_sys_sleep && dwc->connected) {
+		link_state = dwc3_gadget_get_link_state(dwc);
+		/* Prevent PM Sleep if not in U3/L2 */
+		if (link_state != DWC3_LINK_STATE_U3)
+			return -EBUSY;
+
+		/* don't stop/disconnect */
+		dwc3_gadget_disable_irq(dwc);

We shouldn't disable event interrupt here. What will happen if the

Quoted 3 lines

Due to some reason, if I don't disable the event interrupts here then after USB resume the USB controller is malfunctioning.
It no longer responds to any requests from Host.

You should look into this. These events are important as they can tell whether the host initiates resume.

Quoted 9 lines

Quoted 4 lines

device is disconnected and reconnect to the host while the device is still in system suspend? The host would not be able to communicate with the device then.

In the TI platform, The system is woken up on any VBUS/linestate change and in dwc3_gadget_resume we enable the events again and check for pending events. Is it pointless to check for pending events there?

It seems fragile for the implementation to be dependent on platform specific feature right?

Also, what will happen in a typical case when the host puts the device in suspend and initiates resume while the device is in system suspend (and stay in suspend over a period of time)? There is no VBUS change. There will be problem if host detects no response from device in time.

Quoted 9 lines

Quoted 1 line

Quoted 1 line

Don't we need these events to wakeup the device?

That's why the TI implementation has line-state change detection to detect a USB resume. We are doing a out-of-band wake-up. The wake up events are configured in the wrapper driver (dwc3-am62.c).

Do you know of any dwc3 implementation that uses in-band mechanism to wake up the System. i.e. it relies on events enabled in DEVTEN register?

We rely on PME. The PME is generated from the PMU of the usb controller when it detects a resume. If your platform supports hibernation and if the resume signal is connected to the lower layer power manager of your device, then you can wakeup the system one level at a time. For example, if your device is a pci device, that wakeup signal would tie to the pci power manager, waking up the pci layer before waking up the core of the usb controller. That's how the host wakes up the host system (e.g. from remote wakeup). For this to work, we expect something similar on the device side.

Quoted 8 lines

Quoted 7 lines

Quoted 1 line

We may not be able to suspend everything in system suspend for this case. I'm thinking of treating these events as if they are PME to wakeup the device, but they are not the same. It may not be simple to handle this. The lower layers may need to stay awake for the dwc3 to handle these events. Hm... it gets a bit complicated.

Quoted 5 lines

As we are going into suspend, we are not really in a position to handle any (DEVTEN) events till we have fully resumed.
So yes, we need to rely on platform specific implementation to wake the System on any USB event.

You may be able to detect vbus change through the connector controller. However, the usb controller is the one that detects host resume. What platform specific implementation do you have outside of the usb controller do you have to get around that?

Quoted 4 lines

Quoted 4 lines

I'm not sure if your platform supports hibernation or if the PME signal on your platform can wakeup the system, but currently dwc3 driver doesn't handle hibernation (device side). If there's no hibernation, there's no PME.

No, in this TI SoC, hibernation feature is not supported in the dwc3 core.

Quoted 2 lines

Quoted 1 line

Quoted 5 lines

Actually, I think the dwc3 core is still on during system suspend for you right? Then I think we can use the wakeup event to wakeup system suspend on host resume? You can ignore about PME in this case. You may need to look into what needs stay awake to allow for handling of the dwc3 event.

But in SoC deep-sleep state, all clocks to the dwc3 core are stopped. So I'm not sure if dwc3 events will work.

Quoted 196 lines · Roger Quadros

Quoted 184 lines · Thinh Nguyen

Quoted 183 lines · Thinh Nguyen

Quoted 143 lines · Roger Quadros

Quoted 142 lines · Thinh Nguyen

Quoted 141 lines · Thinh Nguyen

Quoted 115 lines · Roger Quadros

Hi Thinh,

Quoted 100 lines · Thinh Nguyen

Hi,

Quoted 97 lines · Roger Quadros

Implement 'snps,gadget-keep-connect-sys-sleep' property.

Do not stop the gadget controller and disconnect if this property is present and we are connected to a USB Host.

Prevent System sleep if Gadget is not in USB suspend.

Signed-off-by Roger Quadros <[email protected]>

---
 drivers/usb/dwc3/core.c   | 25 +++++++++++++++++++------
 drivers/usb/dwc3/core.h   |  2 ++
 drivers/usb/dwc3/gadget.c | 25 +++++++++++++++++++++++--
 3 files changed, 44 insertions(+), 8 deletions(-)

drivers/usb/dwc3/core.c +19 -6

index 476b63618511..a47bbaa27302 100644
@@ -1575,6 +1575,9 @@ static void dwc3_get_properties(struct dwc3 *dwc)
 	dwc->dis_split_quirk = device_property_read_bool(dev,
 				"snps,dis-split-quirk");
 
+	dwc->gadget_keep_connect_sys_sleep = device_property_read_bool(dev,
+				"snps,gadget-keep-connect-sys-sleep");
+
 	dwc->lpm_nyet_threshold = lpm_nyet_threshold;
 	dwc->tx_de_emphasis = tx_de_emphasis;
 

@@ -2027,14 +2030,20 @@ static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
 {
 	unsigned long	flags;
 	u32 reg;
+	int ret;
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
 		if (pm_runtime_suspended(dwc->dev))
 			break;
-		dwc3_gadget_suspend(dwc);
+		ret = dwc3_gadget_suspend(dwc);
+		if (ret) {
+			dev_err(dwc->dev, "gadget not suspended: %d\n", ret);
+			return ret;
+		}
 		synchronize_irq(dwc->irq_gadget);
-		dwc3_core_exit(dwc);
+		if(!dwc->gadget_keep_connect_sys_sleep)
+			dwc3_core_exit(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:
 		if (!PMSG_IS_AUTO(msg) && !device_may_wakeup(dwc->dev)) {

@@ -2088,11 +2097,15 @@ static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg)
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
-		ret = dwc3_core_init_for_resume(dwc);
-		if (ret)
-			return ret;
+		if (!dwc->gadget_keep_connect_sys_sleep)
+		{
+			ret = dwc3_core_init_for_resume(dwc);
+			if (ret)
+				return ret;
+
+			dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
+		}
 
-		dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
 		dwc3_gadget_resume(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:

drivers/usb/dwc3/core.h +2 -0

index 582ebd9cf9c2..f84bac815bed 100644
@@ -1328,6 +1328,8 @@ struct dwc3 {
 	unsigned		dis_split_quirk:1;
 	unsigned		async_callbacks:1;
 
+	unsigned		gadget_keep_connect_sys_sleep:1;
+
 	u16			imod_interval;
 
 	int			max_cfg_eps;

drivers/usb/dwc3/gadget.c +1 -0

index 3c63fa97a680..8062e44f63f6 100644
@@ -4572,12 +4572,23 @@ void dwc3_gadget_exit(struct dwc3 *dwc)
 int dwc3_gadget_suspend(struct dwc3 *dwc)
 {
 	unsigned long flags;
+	int link_state;
 
 	if (!dwc->gadget_driver)
 		return 0;
 

Quoted 11 lines

drivers/usb/dwc3/gadget.c void dwc3_gadget_exit(struct dwc3 *dwc) · line 4587

Quoted 9 lines

drivers/usb/dwc3/gadget.c +8 -1

-	dwc3_gadget_run_stop(dwc, false, false);
+	if (dwc->gadget_keep_connect_sys_sleep && dwc->connected) {
+		link_state = dwc3_gadget_get_link_state(dwc);
+		/* Prevent PM Sleep if not in U3/L2 */
+		if (link_state != DWC3_LINK_STATE_U3)
+			return -EBUSY;
+
+		/* don't stop/disconnect */
+		dwc3_gadget_disable_irq(dwc);

We shouldn't disable event interrupt here. What will happen if the

Quoted 3 lines

Due to some reason, if I don't disable the event interrupts here then after USB resume the USB controller is malfunctioning.
It no longer responds to any requests from Host.

You should look into this. These events are important as they can tell whether the host initiates resume.

Quoted 9 lines

Quoted 4 lines

device is disconnected and reconnect to the host while the device is still in system suspend? The host would not be able to communicate with the device then.

In the TI platform, The system is woken up on any VBUS/linestate change and in dwc3_gadget_resume we enable the events again and check for pending events. Is it pointless to check for pending events there?

It seems fragile for the implementation to be dependent on platform specific feature right?

Also, what will happen in a typical case when the host puts the device in suspend and initiates resume while the device is in system suspend (and stay in suspend over a period of time)? There is no VBUS change. There will be problem if host detects no response from device in time.

Quoted 9 lines

Quoted 1 line

Quoted 1 line

Don't we need these events to wakeup the device?

That's why the TI implementation has line-state change detection to detect a USB resume. We are doing a out-of-band wake-up. The wake up events are configured in the wrapper driver (dwc3-am62.c).

Do you know of any dwc3 implementation that uses in-band mechanism to wake up the System. i.e. it relies on events enabled in DEVTEN register?

We rely on PME. The PME is generated from the PMU of the usb controller when it detects a resume. If your platform supports hibernation and if the resume signal is connected to the lower layer power manager of your device, then you can wakeup the system one level at a time. For example, if your device is a pci device, that wakeup signal would tie to the pci power manager, waking up the pci layer before waking up the core of the usb controller. That's how the host wakes up the host system (e.g. from remote wakeup). For this to work, we expect something similar on the device side.

Quoted 8 lines

Quoted 7 lines

Quoted 1 line

We may not be able to suspend everything in system suspend for this case. I'm thinking of treating these events as if they are PME to wakeup the device, but they are not the same. It may not be simple to handle this. The lower layers may need to stay awake for the dwc3 to handle these events. Hm... it gets a bit complicated.

Quoted 5 lines

As we are going into suspend, we are not really in a position to handle any (DEVTEN) events till we have fully resumed.
So yes, we need to rely on platform specific implementation to wake the System on any USB event.

You may be able to detect vbus change through the connector controller. However, the usb controller is the one that detects host resume. What platform specific implementation do you have outside of the usb controller do you have to get around that?

Quoted 4 lines

Quoted 4 lines

I'm not sure if your platform supports hibernation or if the PME signal on your platform can wakeup the system, but currently dwc3 driver doesn't handle hibernation (device side). If there's no hibernation, there's no PME.

No, in this TI SoC, hibernation feature is not supported in the dwc3 core.

Quoted 2 lines

Quoted 1 line

Quoted 9 lines

Quoted 5 lines

Actually, I think the dwc3 core is still on during system suspend for you right? Then I think we can use the wakeup event to wakeup system suspend on host resume? You can ignore about PME in this case. You may need to look into what needs stay awake to allow for handling of the dwc3 event.

But in SoC deep-sleep state, all clocks to the dwc3 core are stopped. So I'm not sure if dwc3 events will work.

Right, you need to keep those clocks running to detect host resume. There's still some power saving through the dwc3 controller's handling in suspend. You may have some limited power saving from other suspended devices on your setup. However, I don't think we can expect the platform to go into deep-sleep and also handle host resume.

Hi,

Quoted 207 lines · Thinh Nguyen

Quoted 205 lines · Roger Quadros

Quoted 184 lines · Thinh Nguyen

Quoted 183 lines · Thinh Nguyen

Quoted 143 lines · Roger Quadros

Quoted 142 lines · Thinh Nguyen

Quoted 141 lines · Thinh Nguyen

Quoted 115 lines · Roger Quadros

Hi Thinh,

Quoted 100 lines · Thinh Nguyen

Hi,

Quoted 97 lines · Roger Quadros

Implement 'snps,gadget-keep-connect-sys-sleep' property.

Do not stop the gadget controller and disconnect if this property is present and we are connected to a USB Host.

Prevent System sleep if Gadget is not in USB suspend.

Signed-off-by Roger Quadros <[email protected]>

---
 drivers/usb/dwc3/core.c   | 25 +++++++++++++++++++------
 drivers/usb/dwc3/core.h   |  2 ++
 drivers/usb/dwc3/gadget.c | 25 +++++++++++++++++++++++--
 3 files changed, 44 insertions(+), 8 deletions(-)

drivers/usb/dwc3/core.c +19 -6

index 476b63618511..a47bbaa27302 100644
@@ -1575,6 +1575,9 @@ static void dwc3_get_properties(struct dwc3 *dwc)
 	dwc->dis_split_quirk = device_property_read_bool(dev,
 				"snps,dis-split-quirk");
 
+	dwc->gadget_keep_connect_sys_sleep = device_property_read_bool(dev,
+				"snps,gadget-keep-connect-sys-sleep");
+
 	dwc->lpm_nyet_threshold = lpm_nyet_threshold;
 	dwc->tx_de_emphasis = tx_de_emphasis;
 

@@ -2027,14 +2030,20 @@ static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
 {
 	unsigned long	flags;
 	u32 reg;
+	int ret;
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
 		if (pm_runtime_suspended(dwc->dev))
 			break;
-		dwc3_gadget_suspend(dwc);
+		ret = dwc3_gadget_suspend(dwc);
+		if (ret) {
+			dev_err(dwc->dev, "gadget not suspended: %d\n", ret);
+			return ret;
+		}
 		synchronize_irq(dwc->irq_gadget);
-		dwc3_core_exit(dwc);
+		if(!dwc->gadget_keep_connect_sys_sleep)
+			dwc3_core_exit(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:
 		if (!PMSG_IS_AUTO(msg) && !device_may_wakeup(dwc->dev)) {

@@ -2088,11 +2097,15 @@ static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg)
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
-		ret = dwc3_core_init_for_resume(dwc);
-		if (ret)
-			return ret;
+		if (!dwc->gadget_keep_connect_sys_sleep)
+		{
+			ret = dwc3_core_init_for_resume(dwc);
+			if (ret)
+				return ret;
+
+			dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
+		}
 
-		dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
 		dwc3_gadget_resume(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:

drivers/usb/dwc3/core.h +2 -0

index 582ebd9cf9c2..f84bac815bed 100644
@@ -1328,6 +1328,8 @@ struct dwc3 {
 	unsigned		dis_split_quirk:1;
 	unsigned		async_callbacks:1;
 
+	unsigned		gadget_keep_connect_sys_sleep:1;
+
 	u16			imod_interval;
 
 	int			max_cfg_eps;

drivers/usb/dwc3/gadget.c +1 -0

index 3c63fa97a680..8062e44f63f6 100644
@@ -4572,12 +4572,23 @@ void dwc3_gadget_exit(struct dwc3 *dwc)
 int dwc3_gadget_suspend(struct dwc3 *dwc)
 {
 	unsigned long flags;
+	int link_state;
 
 	if (!dwc->gadget_driver)
 		return 0;
 

Quoted 11 lines

drivers/usb/dwc3/gadget.c void dwc3_gadget_exit(struct dwc3 *dwc) · line 4587

Quoted 9 lines

drivers/usb/dwc3/gadget.c +8 -1

-	dwc3_gadget_run_stop(dwc, false, false);
+	if (dwc->gadget_keep_connect_sys_sleep && dwc->connected) {
+		link_state = dwc3_gadget_get_link_state(dwc);
+		/* Prevent PM Sleep if not in U3/L2 */
+		if (link_state != DWC3_LINK_STATE_U3)
+			return -EBUSY;
+
+		/* don't stop/disconnect */
+		dwc3_gadget_disable_irq(dwc);

We shouldn't disable event interrupt here. What will happen if the

Quoted 3 lines

Due to some reason, if I don't disable the event interrupts here then after USB resume the USB controller is malfunctioning.
It no longer responds to any requests from Host.

You should look into this. These events are important as they can tell whether the host initiates resume.

Quoted 9 lines

Quoted 4 lines

device is disconnected and reconnect to the host while the device is still in system suspend? The host would not be able to communicate with the device then.

In the TI platform, The system is woken up on any VBUS/linestate change and in dwc3_gadget_resume we enable the events again and check for pending events. Is it pointless to check for pending events there?

It seems fragile for the implementation to be dependent on platform specific feature right?

Also, what will happen in a typical case when the host puts the device in suspend and initiates resume while the device is in system suspend (and stay in suspend over a period of time)? There is no VBUS change. There will be problem if host detects no response from device in time.

Quoted 9 lines

Quoted 1 line

Quoted 1 line

Don't we need these events to wakeup the device?

That's why the TI implementation has line-state change detection to detect a USB resume. We are doing a out-of-band wake-up. The wake up events are configured in the wrapper driver (dwc3-am62.c).

Do you know of any dwc3 implementation that uses in-band mechanism to wake up the System. i.e. it relies on events enabled in DEVTEN register?

We rely on PME. The PME is generated from the PMU of the usb controller when it detects a resume. If your platform supports hibernation and if the resume signal is connected to the lower layer power manager of your device, then you can wakeup the system one level at a time. For example, if your device is a pci device, that wakeup signal would tie to the pci power manager, waking up the pci layer before waking up the core of the usb controller. That's how the host wakes up the host system (e.g. from remote wakeup). For this to work, we expect something similar on the device side.

Quoted 8 lines

Quoted 7 lines

Quoted 1 line

We may not be able to suspend everything in system suspend for this case. I'm thinking of treating these events as if they are PME to wakeup the device, but they are not the same. It may not be simple to handle this. The lower layers may need to stay awake for the dwc3 to handle these events. Hm... it gets a bit complicated.

Quoted 5 lines

As we are going into suspend, we are not really in a position to handle any (DEVTEN) events till we have fully resumed.
So yes, we need to rely on platform specific implementation to wake the System on any USB event.

You may be able to detect vbus change through the connector controller. However, the usb controller is the one that detects host resume. What platform specific implementation do you have outside of the usb controller do you have to get around that?

Quoted 4 lines

Quoted 4 lines

I'm not sure if your platform supports hibernation or if the PME signal on your platform can wakeup the system, but currently dwc3 driver doesn't handle hibernation (device side). If there's no hibernation, there's no PME.

No, in this TI SoC, hibernation feature is not supported in the dwc3 core.

Quoted 7 lines

Quoted 1 line

Actually, I think the dwc3 core is still on during system suspend for you right? Then I think we can use the wakeup event to wakeup system suspend on host resume? You can ignore about PME in this case. You may need to look into what needs stay awake to allow for handling of the dwc3 event.

But in SoC deep-sleep state, all clocks to the dwc3 core are stopped. So I'm not sure if dwc3 events will work.

Quoted 5 lines

Right, you need to keep those clocks running to detect host resume. There's still some power saving through the dwc3 controller's handling in suspend. You may have some limited power saving from other suspended devices on your setup. However, I don't think we can expect the platform to go into deep-sleep and also handle host resume.

Why not? if the PHY can detect the host resume and wake up the SoC it will work right?

Quoted 210 lines · Roger Quadros

Hi,

Quoted 207 lines · Thinh Nguyen

Quoted 205 lines · Roger Quadros

Quoted 184 lines · Thinh Nguyen

Quoted 183 lines · Thinh Nguyen

Quoted 143 lines · Roger Quadros

Quoted 142 lines · Thinh Nguyen

Quoted 141 lines · Thinh Nguyen

Quoted 115 lines · Roger Quadros

Hi Thinh,

Quoted 100 lines · Thinh Nguyen

Hi,

Quoted 97 lines · Roger Quadros

Implement 'snps,gadget-keep-connect-sys-sleep' property.

Do not stop the gadget controller and disconnect if this property is present and we are connected to a USB Host.

Prevent System sleep if Gadget is not in USB suspend.

Signed-off-by Roger Quadros <[email protected]>

---
 drivers/usb/dwc3/core.c   | 25 +++++++++++++++++++------
 drivers/usb/dwc3/core.h   |  2 ++
 drivers/usb/dwc3/gadget.c | 25 +++++++++++++++++++++++--
 3 files changed, 44 insertions(+), 8 deletions(-)

drivers/usb/dwc3/core.c +19 -6

index 476b63618511..a47bbaa27302 100644
@@ -1575,6 +1575,9 @@ static void dwc3_get_properties(struct dwc3 *dwc)
 	dwc->dis_split_quirk = device_property_read_bool(dev,
 				"snps,dis-split-quirk");
 
+	dwc->gadget_keep_connect_sys_sleep = device_property_read_bool(dev,
+				"snps,gadget-keep-connect-sys-sleep");
+
 	dwc->lpm_nyet_threshold = lpm_nyet_threshold;
 	dwc->tx_de_emphasis = tx_de_emphasis;
 

@@ -2027,14 +2030,20 @@ static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
 {
 	unsigned long	flags;
 	u32 reg;
+	int ret;
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
 		if (pm_runtime_suspended(dwc->dev))
 			break;
-		dwc3_gadget_suspend(dwc);
+		ret = dwc3_gadget_suspend(dwc);
+		if (ret) {
+			dev_err(dwc->dev, "gadget not suspended: %d\n", ret);
+			return ret;
+		}
 		synchronize_irq(dwc->irq_gadget);
-		dwc3_core_exit(dwc);
+		if(!dwc->gadget_keep_connect_sys_sleep)
+			dwc3_core_exit(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:
 		if (!PMSG_IS_AUTO(msg) && !device_may_wakeup(dwc->dev)) {

@@ -2088,11 +2097,15 @@ static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg)
 
 	switch (dwc->current_dr_role) {
 	case DWC3_GCTL_PRTCAP_DEVICE:
-		ret = dwc3_core_init_for_resume(dwc);
-		if (ret)
-			return ret;
+		if (!dwc->gadget_keep_connect_sys_sleep)
+		{
+			ret = dwc3_core_init_for_resume(dwc);
+			if (ret)
+				return ret;
+
+			dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
+		}
 
-		dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
 		dwc3_gadget_resume(dwc);
 		break;
 	case DWC3_GCTL_PRTCAP_HOST:

drivers/usb/dwc3/core.h +2 -0

index 582ebd9cf9c2..f84bac815bed 100644
@@ -1328,6 +1328,8 @@ struct dwc3 {
 	unsigned		dis_split_quirk:1;
 	unsigned		async_callbacks:1;
 
+	unsigned		gadget_keep_connect_sys_sleep:1;
+
 	u16			imod_interval;
 
 	int			max_cfg_eps;

drivers/usb/dwc3/gadget.c +1 -0

index 3c63fa97a680..8062e44f63f6 100644
@@ -4572,12 +4572,23 @@ void dwc3_gadget_exit(struct dwc3 *dwc)
 int dwc3_gadget_suspend(struct dwc3 *dwc)
 {
 	unsigned long flags;
+	int link_state;
 
 	if (!dwc->gadget_driver)
 		return 0;
 

Quoted 11 lines

drivers/usb/dwc3/gadget.c void dwc3_gadget_exit(struct dwc3 *dwc) · line 4587

Quoted 9 lines

drivers/usb/dwc3/gadget.c +8 -1

-	dwc3_gadget_run_stop(dwc, false, false);
+	if (dwc->gadget_keep_connect_sys_sleep && dwc->connected) {
+		link_state = dwc3_gadget_get_link_state(dwc);
+		/* Prevent PM Sleep if not in U3/L2 */
+		if (link_state != DWC3_LINK_STATE_U3)
+			return -EBUSY;
+
+		/* don't stop/disconnect */
+		dwc3_gadget_disable_irq(dwc);

We shouldn't disable event interrupt here. What will happen if the

Quoted 3 lines

Due to some reason, if I don't disable the event interrupts here then after USB resume the USB controller is malfunctioning.
It no longer responds to any requests from Host.

You should look into this. These events are important as they can tell whether the host initiates resume.

Quoted 9 lines

Quoted 4 lines

device is disconnected and reconnect to the host while the device is still in system suspend? The host would not be able to communicate with the device then.

In the TI platform, The system is woken up on any VBUS/linestate change and in dwc3_gadget_resume we enable the events again and check for pending events. Is it pointless to check for pending events there?

It seems fragile for the implementation to be dependent on platform specific feature right?

Also, what will happen in a typical case when the host puts the device in suspend and initiates resume while the device is in system suspend (and stay in suspend over a period of time)? There is no VBUS change. There will be problem if host detects no response from device in time.

Quoted 9 lines

Quoted 1 line

Quoted 1 line

Don't we need these events to wakeup the device?

That's why the TI implementation has line-state change detection to detect a USB resume. We are doing a out-of-band wake-up. The wake up events are configured in the wrapper driver (dwc3-am62.c).

Do you know of any dwc3 implementation that uses in-band mechanism to wake up the System. i.e. it relies on events enabled in DEVTEN register?

We rely on PME. The PME is generated from the PMU of the usb controller when it detects a resume. If your platform supports hibernation and if the resume signal is connected to the lower layer power manager of your device, then you can wakeup the system one level at a time. For example, if your device is a pci device, that wakeup signal would tie to the pci power manager, waking up the pci layer before waking up the core of the usb controller. That's how the host wakes up the host system (e.g. from remote wakeup). For this to work, we expect something similar on the device side.

Quoted 8 lines

Quoted 7 lines

Quoted 1 line

We may not be able to suspend everything in system suspend for this case. I'm thinking of treating these events as if they are PME to wakeup the device, but they are not the same. It may not be simple to handle this. The lower layers may need to stay awake for the dwc3 to handle these events. Hm... it gets a bit complicated.

Quoted 5 lines

As we are going into suspend, we are not really in a position to handle any (DEVTEN) events till we have fully resumed.
So yes, we need to rely on platform specific implementation to wake the System on any USB event.

You may be able to detect vbus change through the connector controller. However, the usb controller is the one that detects host resume. What platform specific implementation do you have outside of the usb controller do you have to get around that?

Quoted 4 lines

Quoted 4 lines

I'm not sure if your platform supports hibernation or if the PME signal on your platform can wakeup the system, but currently dwc3 driver doesn't handle hibernation (device side). If there's no hibernation, there's no PME.

No, in this TI SoC, hibernation feature is not supported in the dwc3 core.

Quoted 7 lines

Quoted 1 line

Actually, I think the dwc3 core is still on during system suspend for you right? Then I think we can use the wakeup event to wakeup system suspend on host resume? You can ignore about PME in this case. You may need to look into what needs stay awake to allow for handling of the dwc3 event.

But in SoC deep-sleep state, all clocks to the dwc3 core are stopped. So I'm not sure if dwc3 events will work.

Quoted 9 lines

Quoted 5 lines

Right, you need to keep those clocks running to detect host resume. There's still some power saving through the dwc3 controller's handling in suspend. You may have some limited power saving from other suspended devices on your setup. However, I don't think we can expect the platform to go into deep-sleep and also handle host resume.

Why not? if the PHY can detect the host resume and wake up the SoC it will work right?

Hm... I supposed it may be possible. But it may need some unconventional design? The dwc3 controller is currently registered to the phy. For that to work, your phy needs to be able to talk to both the dwc3 controller and some other controller (equivalent to dwc3 PMU) that manages power/interrupt. The dwc3 controller would need to relinquish control to this other phy controller on suspend. The phy driver would then be able to assert interrupt waking up the system on resume sigal detection, which in turn relinquish control to the dwc3 controller. All of this has to work while the phy signaling remains synchronized with the dwc3 controller.

From the patches you sent, I don't see the changes necesssary for this to work. If there is something that I'm missing, please also note it or add it here to the series.

Quoted 221 lines · Thinh Nguyen

Quoted 210 lines · Roger Quadros

Hi,

Quoted 207 lines · Thinh Nguyen

Quoted 205 lines · Roger Quadros

Quoted 184 lines · Thinh Nguyen

Quoted 183 lines · Thinh Nguyen

Quoted 143 lines · Roger Quadros

Quoted 142 lines · Thinh Nguyen

Quoted 141 lines · Thinh Nguyen

Quoted 115 lines · Roger Quadros

Hi Thinh,

Quoted 100 lines · Thinh Nguyen

Hi,

Quoted 97 lines · Roger Quadros

Implement 'snps,gadget-keep-connect-sys-sleep' property.

Do not stop the gadget controller and disconnect if this
property is present and we are connected to a USB Host.

Prevent System sleep if Gadget is not in USB suspend.

Signed-off-by: Roger Quadros <[email protected]>
---
drivers/usb/dwc3/core.c | 25 +++++++++++++++++++------
drivers/usb/dwc3/core.h | 2 ++
drivers/usb/dwc3/gadget.c | 25 +++++++++++++++++++++++--
3 files changed, 44 insertions(+), 8 deletions(-)

diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c
index 476b63618511..a47bbaa27302 100644
--- a/drivers/usb/dwc3/core.c
+++ b/drivers/usb/dwc3/core.c
@@ -1575,6 +1575,9 @@ static void dwc3_get_properties(struct dwc3 *dwc)
dwc->dis_split_quirk = device_property_read_bool(dev,
"snps,dis-split-quirk");

+ dwc->gadget_keep_connect_sys_sleep = device_property_read_bool(dev,
+ "snps,gadget-keep-connect-sys-sleep");
+
dwc->lpm_nyet_threshold = lpm_nyet_threshold;
dwc->tx_de_emphasis = tx_de_emphasis;

@@ -2027,14 +2030,20 @@ static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
{
unsigned long flags;
u32 reg;
+ int ret;

switch (dwc->current_dr_role) {
case DWC3_GCTL_PRTCAP_DEVICE:
if (pm_runtime_suspended(dwc->dev))
break;
- dwc3_gadget_suspend(dwc);
+ ret = dwc3_gadget_suspend(dwc);
+ if (ret) {
+ dev_err(dwc->dev, "gadget not suspended: %d\n", ret);
+ return ret;
+ }
synchronize_irq(dwc->irq_gadget);
- dwc3_core_exit(dwc);
+ if(!dwc->gadget_keep_connect_sys_sleep)
+ dwc3_core_exit(dwc);
break;
case DWC3_GCTL_PRTCAP_HOST:
if (!PMSG_IS_AUTO(msg) && !device_may_wakeup(dwc->dev)) {
@@ -2088,11 +2097,15 @@ static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg)

switch (dwc->current_dr_role) {
case DWC3_GCTL_PRTCAP_DEVICE:
- ret = dwc3_core_init_for_resume(dwc);
- if (ret)
- return ret;
+ if (!dwc->gadget_keep_connect_sys_sleep)
+ {
+ ret = dwc3_core_init_for_resume(dwc);
+ if (ret)
+ return ret;
+
+ dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
+ }

- dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
dwc3_gadget_resume(dwc);
break;
case DWC3_GCTL_PRTCAP_HOST:
diff --git a/drivers/usb/dwc3/core.h b/drivers/usb/dwc3/core.h
index 582ebd9cf9c2..f84bac815bed 100644
--- a/drivers/usb/dwc3/core.h
+++ b/drivers/usb/dwc3/core.h
@@ -1328,6 +1328,8 @@ struct dwc3 {
unsigned dis_split_quirk:1;
unsigned async_callbacks:1;

+ unsigned gadget_keep_connect_sys_sleep:1;
+
u16 imod_interval;

int max_cfg_eps;
diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
index 3c63fa97a680..8062e44f63f6 100644
--- a/drivers/usb/dwc3/gadget.c
+++ b/drivers/usb/dwc3/gadget.c
@@ -4572,12 +4572,23 @@ void dwc3_gadget_exit(struct dwc3 *dwc)
int dwc3_gadget_suspend(struct dwc3 *dwc)
{
unsigned long flags;
+ int link_state;

if (!dwc->gadget_driver)
return 0;

Quoted 11 lines

Quoted 9 lines

- dwc3_gadget_run_stop(dwc, false, false);
+ if (dwc->gadget_keep_connect_sys_sleep && dwc->connected) {
+ link_state = dwc3_gadget_get_link_state(dwc);
+ /* Prevent PM Sleep if not in U3/L2 */
+ if (link_state != DWC3_LINK_STATE_U3)
+ return -EBUSY;
+
+ /* don't stop/disconnect */
+ dwc3_gadget_disable_irq(dwc);

We shouldn't disable event interrupt here. What will happen if the

Quoted 3 lines

Due to some reason, if I don't disable the event interrupts here then after USB resume the USB controller is malfunctioning.
It no longer responds to any requests from Host.

You should look into this. These events are important as they can tell whether the host initiates resume.

Quoted 9 lines

Quoted 4 lines

device is disconnected and reconnect to the host while the device is still in system suspend? The host would not be able to communicate with the device then.

In the TI platform, The system is woken up on any VBUS/linestate change and in dwc3_gadget_resume we enable the events again and check for pending events. Is it pointless to check for pending events there?

It seems fragile for the implementation to be dependent on platform specific feature right?

Also, what will happen in a typical case when the host puts the device in suspend and initiates resume while the device is in system suspend (and stay in suspend over a period of time)? There is no VBUS change. There will be problem if host detects no response from device in time.

Quoted 9 lines

Quoted 1 line

Quoted 1 line

Don't we need these events to wakeup the device?

That's why the TI implementation has line-state change detection to detect a USB resume. We are doing a out-of-band wake-up. The wake up events are configured in the wrapper driver (dwc3-am62.c).

Do you know of any dwc3 implementation that uses in-band mechanism to wake up the System. i.e. it relies on events enabled in DEVTEN register?

We rely on PME. The PME is generated from the PMU of the usb controller when it detects a resume. If your platform supports hibernation and if the resume signal is connected to the lower layer power manager of your device, then you can wakeup the system one level at a time. For example, if your device is a pci device, that wakeup signal would tie to the pci power manager, waking up the pci layer before waking up the core of the usb controller. That's how the host wakes up the host system (e.g. from remote wakeup). For this to work, we expect something similar on the device side.

Quoted 8 lines

Quoted 7 lines

Quoted 1 line

We may not be able to suspend everything in system suspend for this case. I'm thinking of treating these events as if they are PME to wakeup the device, but they are not the same. It may not be simple to handle this. The lower layers may need to stay awake for the dwc3 to handle these events. Hm... it gets a bit complicated.

Quoted 5 lines

As we are going into suspend, we are not really in a position to handle any (DEVTEN) events till we have fully resumed.
So yes, we need to rely on platform specific implementation to wake the System on any USB event.

You may be able to detect vbus change through the connector controller. However, the usb controller is the one that detects host resume. What platform specific implementation do you have outside of the usb controller do you have to get around that?

Quoted 4 lines

Quoted 4 lines

I'm not sure if your platform supports hibernation or if the PME signal on your platform can wakeup the system, but currently dwc3 driver doesn't handle hibernation (device side). If there's no hibernation, there's no PME.

No, in this TI SoC, hibernation feature is not supported in the dwc3 core.

Quoted 7 lines

Quoted 1 line

Actually, I think the dwc3 core is still on during system suspend for you right? Then I think we can use the wakeup event to wakeup system suspend on host resume? You can ignore about PME in this case. You may need to look into what needs stay awake to allow for handling of the dwc3 event.

But in SoC deep-sleep state, all clocks to the dwc3 core are stopped. So I'm not sure if dwc3 events will work.

Quoted 9 lines

Quoted 5 lines

Right, you need to keep those clocks running to detect host resume. There's still some power saving through the dwc3 controller's handling in suspend. You may have some limited power saving from other suspended devices on your setup. However, I don't think we can expect the platform to go into deep-sleep and also handle host resume.

Why not? if the PHY can detect the host resume and wake up the SoC it will work right?

Quoted 10 lines

Hm... I supposed it may be possible. But it may need some unconventional design? The dwc3 controller is currently registered to the phy. For that to work, your phy needs to be able to talk to both the dwc3 controller and some other controller (equivalent to dwc3 PMU) that manages power/interrupt. The dwc3 controller would need to relinquish control to this other phy controller on suspend. The phy driver would then be able to assert interrupt waking up the system on resume sigal detection, which in turn relinquish control to the dwc3 controller. All of this has to work while the phy signaling remains synchronized with the dwc3 controller.

My understanding is that all this is taken care by PHY integration design with DWC3 core on the TI SoC.

Quoted 4 lines

From the patches you sent, I don't see the changes necesssary for this to work. If there is something that I'm missing, please also note it or add it here to the series.

There is nothing more as the details are taken care by PHY logic and necessary integration with DWC3.

For the PHY wake-up programming details you have already checked this series [1].

[1] - https://urldefense.com/v3/__https://lore.kernel.org/all/[email protected]/__;!!A4F2R9G_pg!ayqpaWwnWIO7SKktk4kI2EJLwDTIv2nYYCgBGHUlt56KTzeYkDEdq2Q5ZvZFsuCWIlcXGET230YY5J3y_sHV$

Quoted 248 lines · Roger Quadros

Quoted 221 lines · Thinh Nguyen

Quoted 210 lines · Roger Quadros

Hi,

Quoted 207 lines · Thinh Nguyen

Quoted 205 lines · Roger Quadros

Quoted 184 lines · Thinh Nguyen

Quoted 183 lines · Thinh Nguyen

Quoted 143 lines · Roger Quadros

Quoted 142 lines · Thinh Nguyen

Quoted 141 lines · Thinh Nguyen

Quoted 115 lines · Roger Quadros

Hi Thinh,

Quoted 111 lines · Thinh Nguyen

Hi,

On Mon, Mar 20, 2023, Roger Quadros wrote:
Implement 'snps,gadget-keep-connect-sys-sleep' property.

Do not stop the gadget controller and disconnect if this
property is present and we are connected to a USB Host.

Prevent System sleep if Gadget is not in USB suspend.

Signed-off-by: Roger Quadros <[email protected]>
---
drivers/usb/dwc3/core.c | 25 +++++++++++++++++++------
drivers/usb/dwc3/core.h | 2 ++
drivers/usb/dwc3/gadget.c | 25 +++++++++++++++++++++++--
3 files changed, 44 insertions(+), 8 deletions(-)

diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c
index 476b63618511..a47bbaa27302 100644
--- a/drivers/usb/dwc3/core.c
+++ b/drivers/usb/dwc3/core.c
@@ -1575,6 +1575,9 @@ static void dwc3_get_properties(struct dwc3 *dwc)
dwc->dis_split_quirk = device_property_read_bool(dev,
"snps,dis-split-quirk");

+ dwc->gadget_keep_connect_sys_sleep = device_property_read_bool(dev,
+ "snps,gadget-keep-connect-sys-sleep");
+
dwc->lpm_nyet_threshold = lpm_nyet_threshold;
dwc->tx_de_emphasis = tx_de_emphasis;

@@ -2027,14 +2030,20 @@ static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
{
unsigned long flags;
u32 reg;
+ int ret;

switch (dwc->current_dr_role) {
case DWC3_GCTL_PRTCAP_DEVICE:
if (pm_runtime_suspended(dwc->dev))
break;
- dwc3_gadget_suspend(dwc);
+ ret = dwc3_gadget_suspend(dwc);
+ if (ret) {
+ dev_err(dwc->dev, "gadget not suspended: %d\n", ret);
+ return ret;
+ }
synchronize_irq(dwc->irq_gadget);
- dwc3_core_exit(dwc);
+ if(!dwc->gadget_keep_connect_sys_sleep)
+ dwc3_core_exit(dwc);
break;
case DWC3_GCTL_PRTCAP_HOST:
if (!PMSG_IS_AUTO(msg) && !device_may_wakeup(dwc->dev)) {
@@ -2088,11 +2097,15 @@ static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg)

switch (dwc->current_dr_role) {
case DWC3_GCTL_PRTCAP_DEVICE:
- ret = dwc3_core_init_for_resume(dwc);
- if (ret)
- return ret;
+ if (!dwc->gadget_keep_connect_sys_sleep)
+ {
+ ret = dwc3_core_init_for_resume(dwc);
+ if (ret)
+ return ret;
+
+ dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
+ }

- dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
dwc3_gadget_resume(dwc);
break;
case DWC3_GCTL_PRTCAP_HOST:
diff --git a/drivers/usb/dwc3/core.h b/drivers/usb/dwc3/core.h
index 582ebd9cf9c2..f84bac815bed 100644
--- a/drivers/usb/dwc3/core.h
+++ b/drivers/usb/dwc3/core.h
@@ -1328,6 +1328,8 @@ struct dwc3 {
unsigned dis_split_quirk:1;
unsigned async_callbacks:1;

+ unsigned gadget_keep_connect_sys_sleep:1;
+
u16 imod_interval;

int max_cfg_eps;
diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
index 3c63fa97a680..8062e44f63f6 100644
--- a/drivers/usb/dwc3/gadget.c
+++ b/drivers/usb/dwc3/gadget.c
@@ -4572,12 +4572,23 @@ void dwc3_gadget_exit(struct dwc3 *dwc)
int dwc3_gadget_suspend(struct dwc3 *dwc)
{
unsigned long flags;
+ int link_state;

if (!dwc->gadget_driver)
return 0;

- dwc3_gadget_run_stop(dwc, false, false);
+ if (dwc->gadget_keep_connect_sys_sleep && dwc->connected) {
+ link_state = dwc3_gadget_get_link_state(dwc);
+ /* Prevent PM Sleep if not in U3/L2 */
+ if (link_state != DWC3_LINK_STATE_U3)
+ return -EBUSY;
+
+ /* don't stop/disconnect */
+ dwc3_gadget_disable_irq(dwc);

We shouldn't disable event interrupt here. What will happen if the

Quoted 3 lines

Due to some reason, if I don't disable the event interrupts here then after USB resume the USB controller is malfunctioning.
It no longer responds to any requests from Host.

You should look into this. These events are important as they can tell whether the host initiates resume.

Quoted 9 lines

Quoted 4 lines

device is disconnected and reconnect to the host while the device is
still in system suspend? The host would not be able to communicate with
the device then.

In the TI platform, The system is woken up on any VBUS/linestate change and in dwc3_gadget_resume we enable the events again and check for pending events. Is it pointless to check for pending events there?

It seems fragile for the implementation to be dependent on platform specific feature right?

Also, what will happen in a typical case when the host puts the device in suspend and initiates resume while the device is in system suspend (and stay in suspend over a period of time)? There is no VBUS change. There will be problem if host detects no response from device in time.

Quoted 9 lines

Quoted 1 line

Quoted 1 line

Don't we need these events to wakeup the device?

That's why the TI implementation has line-state change detection to detect a USB resume. We are doing a out-of-band wake-up. The wake up events are configured in the wrapper driver (dwc3-am62.c).

Do you know of any dwc3 implementation that uses in-band mechanism to wake up the System. i.e. it relies on events enabled in DEVTEN register?

We rely on PME. The PME is generated from the PMU of the usb controller when it detects a resume. If your platform supports hibernation and if the resume signal is connected to the lower layer power manager of your device, then you can wakeup the system one level at a time. For example, if your device is a pci device, that wakeup signal would tie to the pci power manager, waking up the pci layer before waking up the core of the usb controller. That's how the host wakes up the host system (e.g. from remote wakeup). For this to work, we expect something similar on the device side.

Quoted 8 lines

Quoted 7 lines

Quoted 1 line

We may not be able to suspend everything in system suspend for this case. I'm thinking of treating these events as if they are PME to wakeup the device, but they are not the same. It may not be simple to handle this. The lower layers may need to stay awake for the dwc3 to handle these events. Hm... it gets a bit complicated.

Quoted 5 lines

As we are going into suspend, we are not really in a position to handle any (DEVTEN) events till we have fully resumed.
So yes, we need to rely on platform specific implementation to wake the System on any USB event.

You may be able to detect vbus change through the connector controller. However, the usb controller is the one that detects host resume. What platform specific implementation do you have outside of the usb controller do you have to get around that?

Quoted 4 lines

Quoted 4 lines

I'm not sure if your platform supports hibernation or if the PME signal on your platform can wakeup the system, but currently dwc3 driver doesn't handle hibernation (device side). If there's no hibernation, there's no PME.

No, in this TI SoC, hibernation feature is not supported in the dwc3 core.

Quoted 7 lines

Quoted 1 line

Actually, I think the dwc3 core is still on during system suspend for you right? Then I think we can use the wakeup event to wakeup system suspend on host resume? You can ignore about PME in this case. You may need to look into what needs stay awake to allow for handling of the dwc3 event.

But in SoC deep-sleep state, all clocks to the dwc3 core are stopped. So I'm not sure if dwc3 events will work.

Quoted 9 lines

Quoted 5 lines

Right, you need to keep those clocks running to detect host resume. There's still some power saving through the dwc3 controller's handling in suspend. You may have some limited power saving from other suspended devices on your setup. However, I don't think we can expect the platform to go into deep-sleep and also handle host resume.

Why not? if the PHY can detect the host resume and wake up the SoC it will work right?

Quoted 10 lines

Hm... I supposed it may be possible. But it may need some unconventional design? The dwc3 controller is currently registered to the phy. For that to work, your phy needs to be able to talk to both the dwc3 controller and some other controller (equivalent to dwc3 PMU) that manages power/interrupt. The dwc3 controller would need to relinquish control to this other phy controller on suspend. The phy driver would then be able to assert interrupt waking up the system on resume sigal detection, which in turn relinquish control to the dwc3 controller. All of this has to work while the phy signaling remains synchronized with the dwc3 controller.

My understanding is that all this is taken care by PHY integration design with DWC3 core on the TI SoC.

Quoted 4 lines

From the patches you sent, I don't see the changes necesssary for this to work. If there is something that I'm missing, please also note it or add it here to the series.

There is nothing more as the details are taken care by PHY logic and necessary integration with DWC3.

For the PHY wake-up programming details you have already checked this series [1].

Quoted 2 lines

[1] - https://urldefense.com/v3/__https://lore.kernel.org/all/[email protected]/__;!!A4F2R9G_pg!ayqpaWwnWIO7SKktk4kI2EJLwDTIv2nYYCgBGHUlt56KTzeYkDEdq2Q5ZvZFsuCWIlcXGET230YY5J3y_sHV$

I may have misunderstood your platform implementation. My understanding is that it can only detect VBUS and that it can only resume on VBUS valid.

Does the "LINESTATE" here gets asserted if say there's a LFPS detection?

Quoted 252 lines · Thinh Nguyen

Quoted 250 lines · Roger Quadros

Quoted 221 lines · Thinh Nguyen

Quoted 210 lines · Roger Quadros

Hi,

Quoted 207 lines · Thinh Nguyen

Quoted 205 lines · Roger Quadros

Quoted 184 lines · Thinh Nguyen

Quoted 183 lines · Thinh Nguyen

Quoted 143 lines · Roger Quadros

Quoted 142 lines · Thinh Nguyen

Quoted 141 lines · Thinh Nguyen

Quoted 118 lines · Roger Quadros

Hi Thinh,

On 20/03/2023 20:52, Thinh Nguyen wrote:
Hi,

On Mon, Mar 20, 2023, Roger Quadros wrote:
Implement 'snps,gadget-keep-connect-sys-sleep' property.

Do not stop the gadget controller and disconnect if this
property is present and we are connected to a USB Host.

Prevent System sleep if Gadget is not in USB suspend.

Signed-off-by: Roger Quadros <[email protected]>
---
drivers/usb/dwc3/core.c | 25 +++++++++++++++++++------
drivers/usb/dwc3/core.h | 2 ++
drivers/usb/dwc3/gadget.c | 25 +++++++++++++++++++++++--
3 files changed, 44 insertions(+), 8 deletions(-)

diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c
index 476b63618511..a47bbaa27302 100644
--- a/drivers/usb/dwc3/core.c
+++ b/drivers/usb/dwc3/core.c
@@ -1575,6 +1575,9 @@ static void dwc3_get_properties(struct dwc3 *dwc)
dwc->dis_split_quirk = device_property_read_bool(dev,
"snps,dis-split-quirk");

+ dwc->gadget_keep_connect_sys_sleep = device_property_read_bool(dev,
+ "snps,gadget-keep-connect-sys-sleep");
+
dwc->lpm_nyet_threshold = lpm_nyet_threshold;
dwc->tx_de_emphasis = tx_de_emphasis;

@@ -2027,14 +2030,20 @@ static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
{
unsigned long flags;
u32 reg;
+ int ret;

switch (dwc->current_dr_role) {
case DWC3_GCTL_PRTCAP_DEVICE:
if (pm_runtime_suspended(dwc->dev))
break;
- dwc3_gadget_suspend(dwc);
+ ret = dwc3_gadget_suspend(dwc);
+ if (ret) {
+ dev_err(dwc->dev, "gadget not suspended: %d\n", ret);
+ return ret;
+ }
synchronize_irq(dwc->irq_gadget);
- dwc3_core_exit(dwc);
+ if(!dwc->gadget_keep_connect_sys_sleep)
+ dwc3_core_exit(dwc);
break;
case DWC3_GCTL_PRTCAP_HOST:
if (!PMSG_IS_AUTO(msg) && !device_may_wakeup(dwc->dev)) {
@@ -2088,11 +2097,15 @@ static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg)

switch (dwc->current_dr_role) {
case DWC3_GCTL_PRTCAP_DEVICE:
- ret = dwc3_core_init_for_resume(dwc);
- if (ret)
- return ret;
+ if (!dwc->gadget_keep_connect_sys_sleep)
+ {
+ ret = dwc3_core_init_for_resume(dwc);
+ if (ret)
+ return ret;
+
+ dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
+ }

- dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
dwc3_gadget_resume(dwc);
break;
case DWC3_GCTL_PRTCAP_HOST:
diff --git a/drivers/usb/dwc3/core.h b/drivers/usb/dwc3/core.h
index 582ebd9cf9c2..f84bac815bed 100644
--- a/drivers/usb/dwc3/core.h
+++ b/drivers/usb/dwc3/core.h
@@ -1328,6 +1328,8 @@ struct dwc3 {
unsigned dis_split_quirk:1;
unsigned async_callbacks:1;

+ unsigned gadget_keep_connect_sys_sleep:1;
+
u16 imod_interval;

int max_cfg_eps;
diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
index 3c63fa97a680..8062e44f63f6 100644
--- a/drivers/usb/dwc3/gadget.c
+++ b/drivers/usb/dwc3/gadget.c
@@ -4572,12 +4572,23 @@ void dwc3_gadget_exit(struct dwc3 *dwc)
int dwc3_gadget_suspend(struct dwc3 *dwc)
{
unsigned long flags;
+ int link_state;

if (!dwc->gadget_driver)
return 0;

- dwc3_gadget_run_stop(dwc, false, false);
+ if (dwc->gadget_keep_connect_sys_sleep && dwc->connected) {
+ link_state = dwc3_gadget_get_link_state(dwc);
+ /* Prevent PM Sleep if not in U3/L2 */
+ if (link_state != DWC3_LINK_STATE_U3)
+ return -EBUSY;
+
+ /* don't stop/disconnect */
+ dwc3_gadget_disable_irq(dwc);

We shouldn't disable event interrupt here. What will happen if the

Due to some reason, if I don't disable the event interrupts here then
after USB resume the USB controller is malfunctioning.
It no longer responds to any requests from Host.

You should look into this. These events are important as they can tell whether the host initiates resume.

Quoted 9 lines

device is disconnected and reconnect to the host while the device is
still in system suspend? The host would not be able to communicate with
the device then.

In the TI platform, The system is woken up on any VBUS/linestate change
and in dwc3_gadget_resume we enable the events again and check for pending
events. Is it pointless to check for pending events there?

It seems fragile for the implementation to be dependent on platform specific feature right?

Also, what will happen in a typical case when the host puts the device in suspend and initiates resume while the device is in system suspend (and stay in suspend over a period of time)? There is no VBUS change. There will be problem if host detects no response from device in time.

Quoted 9 lines

Quoted 1 line

Quoted 1 line

Don't we need these events to wakeup the device?

That's why the TI implementation has line-state change detection to detect a USB resume. We are doing a out-of-band wake-up. The wake up events are configured in the wrapper driver (dwc3-am62.c).

Do you know of any dwc3 implementation that uses in-band mechanism to wake up the System. i.e. it relies on events enabled in DEVTEN register?

We rely on PME. The PME is generated from the PMU of the usb controller when it detects a resume. If your platform supports hibernation and if the resume signal is connected to the lower layer power manager of your device, then you can wakeup the system one level at a time. For example, if your device is a pci device, that wakeup signal would tie to the pci power manager, waking up the pci layer before waking up the core of the usb controller. That's how the host wakes up the host system (e.g. from remote wakeup). For this to work, we expect something similar on the device side.

Quoted 8 lines

Quoted 7 lines

Quoted 1 line

We may not be able to suspend everything in system suspend for this case. I'm thinking of treating these events as if they are PME to wakeup the device, but they are not the same. It may not be simple to handle this. The lower layers may need to stay awake for the dwc3 to handle these events. Hm... it gets a bit complicated.

Quoted 5 lines

As we are going into suspend, we are not really in a position to handle any (DEVTEN) events till we have fully resumed.
So yes, we need to rely on platform specific implementation to wake the System on any USB event.

You may be able to detect vbus change through the connector controller. However, the usb controller is the one that detects host resume. What platform specific implementation do you have outside of the usb controller do you have to get around that?

Quoted 4 lines

Quoted 4 lines

I'm not sure if your platform supports hibernation or if the PME signal on your platform can wakeup the system, but currently dwc3 driver doesn't handle hibernation (device side). If there's no hibernation, there's no PME.

No, in this TI SoC, hibernation feature is not supported in the dwc3 core.

Quoted 7 lines

Quoted 1 line

Actually, I think the dwc3 core is still on during system suspend for you right? Then I think we can use the wakeup event to wakeup system suspend on host resume? You can ignore about PME in this case. You may need to look into what needs stay awake to allow for handling of the dwc3 event.

But in SoC deep-sleep state, all clocks to the dwc3 core are stopped. So I'm not sure if dwc3 events will work.

Quoted 9 lines

Quoted 5 lines

Right, you need to keep those clocks running to detect host resume. There's still some power saving through the dwc3 controller's handling in suspend. You may have some limited power saving from other suspended devices on your setup. However, I don't think we can expect the platform to go into deep-sleep and also handle host resume.

Why not? if the PHY can detect the host resume and wake up the SoC it will work right?

Quoted 10 lines

Hm... I supposed it may be possible. But it may need some unconventional design? The dwc3 controller is currently registered to the phy. For that to work, your phy needs to be able to talk to both the dwc3 controller and some other controller (equivalent to dwc3 PMU) that manages power/interrupt. The dwc3 controller would need to relinquish control to this other phy controller on suspend. The phy driver would then be able to assert interrupt waking up the system on resume sigal detection, which in turn relinquish control to the dwc3 controller. All of this has to work while the phy signaling remains synchronized with the dwc3 controller.

My understanding is that all this is taken care by PHY integration design with DWC3 core on the TI SoC.

Quoted 4 lines

From the patches you sent, I don't see the changes necesssary for this to work. If there is something that I'm missing, please also note it or add it here to the series.

There is nothing more as the details are taken care by PHY logic and necessary integration with DWC3.

For the PHY wake-up programming details you have already checked this series [1].

[1] - https://urldefense.com/v3/__https://lore.kernel.org/all/[email protected]/__;!!A4F2R9G_pg!ayqpaWwnWIO7SKktk4kI2EJLwDTIv2nYYCgBGHUlt56KTzeYkDEdq2Q5ZvZFsuCWIlcXGET230YY5J3y_sHV$

Quoted 5 lines

I may have misunderstood your platform implementation. My understanding is that it can only detect VBUS and that it can only resume on VBUS valid.

Does the "LINESTATE" here gets asserted if say there's a LFPS detection?

Yes. The wake up logic on the SoC is snooping the UTMI lines from the PHY and on any change it can detect and wake up the SoC.

Quoted 259 lines · Roger Quadros

Quoted 256 lines · Thinh Nguyen

Quoted 243 lines · Roger Quadros

Quoted 221 lines · Thinh Nguyen

Quoted 210 lines · Roger Quadros

Hi,

Quoted 207 lines · Thinh Nguyen

Quoted 205 lines · Roger Quadros

Quoted 184 lines · Thinh Nguyen

Quoted 183 lines · Thinh Nguyen

Quoted 143 lines · Roger Quadros

Quoted 142 lines · Thinh Nguyen

Quoted 141 lines · Thinh Nguyen

On Tue, Mar 21, 2023, Roger Quadros wrote:
Hi Thinh,

On 20/03/2023 20:52, Thinh Nguyen wrote:
Hi,

On Mon, Mar 20, 2023, Roger Quadros wrote:
Implement 'snps,gadget-keep-connect-sys-sleep' property.

Do not stop the gadget controller and disconnect if this
property is present and we are connected to a USB Host.

Prevent System sleep if Gadget is not in USB suspend.

Signed-off-by: Roger Quadros <[email protected]>
---
drivers/usb/dwc3/core.c | 25 +++++++++++++++++++------
drivers/usb/dwc3/core.h | 2 ++
drivers/usb/dwc3/gadget.c | 25 +++++++++++++++++++++++--
3 files changed, 44 insertions(+), 8 deletions(-)

diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c
index 476b63618511..a47bbaa27302 100644
--- a/drivers/usb/dwc3/core.c
+++ b/drivers/usb/dwc3/core.c
@@ -1575,6 +1575,9 @@ static void dwc3_get_properties(struct dwc3 *dwc)
dwc->dis_split_quirk = device_property_read_bool(dev,
"snps,dis-split-quirk");

+ dwc->gadget_keep_connect_sys_sleep = device_property_read_bool(dev,
+ "snps,gadget-keep-connect-sys-sleep");
+
dwc->lpm_nyet_threshold = lpm_nyet_threshold;
dwc->tx_de_emphasis = tx_de_emphasis;

@@ -2027,14 +2030,20 @@ static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
{
unsigned long flags;
u32 reg;
+ int ret;

switch (dwc->current_dr_role) {
case DWC3_GCTL_PRTCAP_DEVICE:
if (pm_runtime_suspended(dwc->dev))
break;
- dwc3_gadget_suspend(dwc);
+ ret = dwc3_gadget_suspend(dwc);
+ if (ret) {
+ dev_err(dwc->dev, "gadget not suspended: %d\n", ret);
+ return ret;
+ }
synchronize_irq(dwc->irq_gadget);
- dwc3_core_exit(dwc);
+ if(!dwc->gadget_keep_connect_sys_sleep)
+ dwc3_core_exit(dwc);
break;
case DWC3_GCTL_PRTCAP_HOST:
if (!PMSG_IS_AUTO(msg) && !device_may_wakeup(dwc->dev)) {
@@ -2088,11 +2097,15 @@ static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg)

switch (dwc->current_dr_role) {
case DWC3_GCTL_PRTCAP_DEVICE:
- ret = dwc3_core_init_for_resume(dwc);
- if (ret)
- return ret;
+ if (!dwc->gadget_keep_connect_sys_sleep)
+ {
+ ret = dwc3_core_init_for_resume(dwc);
+ if (ret)
+ return ret;
+
+ dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
+ }

- dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE);
dwc3_gadget_resume(dwc);
break;
case DWC3_GCTL_PRTCAP_HOST:
diff --git a/drivers/usb/dwc3/core.h b/drivers/usb/dwc3/core.h
index 582ebd9cf9c2..f84bac815bed 100644
--- a/drivers/usb/dwc3/core.h
+++ b/drivers/usb/dwc3/core.h
@@ -1328,6 +1328,8 @@ struct dwc3 {
unsigned dis_split_quirk:1;
unsigned async_callbacks:1;

+ unsigned gadget_keep_connect_sys_sleep:1;
+
u16 imod_interval;

int max_cfg_eps;
diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
index 3c63fa97a680..8062e44f63f6 100644
--- a/drivers/usb/dwc3/gadget.c
+++ b/drivers/usb/dwc3/gadget.c
@@ -4572,12 +4572,23 @@ void dwc3_gadget_exit(struct dwc3 *dwc)
int dwc3_gadget_suspend(struct dwc3 *dwc)
{
unsigned long flags;
+ int link_state;

if (!dwc->gadget_driver)
return 0;

- dwc3_gadget_run_stop(dwc, false, false);
+ if (dwc->gadget_keep_connect_sys_sleep && dwc->connected) {
+ link_state = dwc3_gadget_get_link_state(dwc);
+ /* Prevent PM Sleep if not in U3/L2 */
+ if (link_state != DWC3_LINK_STATE_U3)
+ return -EBUSY;
+
+ /* don't stop/disconnect */
+ dwc3_gadget_disable_irq(dwc);

We shouldn't disable event interrupt here. What will happen if the

Due to some reason, if I don't disable the event interrupts here then
after USB resume the USB controller is malfunctioning.
It no longer responds to any requests from Host.

You should look into this. These events are important as they can tell
whether the host initiates resume.

device is disconnected and reconnect to the host while the device is
still in system suspend? The host would not be able to communicate with
the device then.

In the TI platform, The system is woken up on any VBUS/linestate change
and in dwc3_gadget_resume we enable the events again and check for pending
events. Is it pointless to check for pending events there?

It seems fragile for the implementation to be dependent on platform
specific feature right?

Also, what will happen in a typical case when the host puts the device
in suspend and initiates resume while the device is in system suspend
(and stay in suspend over a period of time)? There is no VBUS change.
There will be problem if host detects no response from device in time.

Quoted 9 lines

Quoted 1 line

Quoted 1 line

Don't we need these events to wakeup the device?

That's why the TI implementation has line-state change detection to detect a USB resume. We are doing a out-of-band wake-up. The wake up events are configured in the wrapper driver (dwc3-am62.c).

Do you know of any dwc3 implementation that uses in-band mechanism to wake up the System. i.e. it relies on events enabled in DEVTEN register?

We rely on PME. The PME is generated from the PMU of the usb controller when it detects a resume. If your platform supports hibernation and if the resume signal is connected to the lower layer power manager of your device, then you can wakeup the system one level at a time. For example, if your device is a pci device, that wakeup signal would tie to the pci power manager, waking up the pci layer before waking up the core of the usb controller. That's how the host wakes up the host system (e.g. from remote wakeup). For this to work, we expect something similar on the device side.

Quoted 8 lines

Quoted 7 lines

Quoted 1 line

We may not be able to suspend everything in system suspend for this case. I'm thinking of treating these events as if they are PME to wakeup the device, but they are not the same. It may not be simple to handle this. The lower layers may need to stay awake for the dwc3 to handle these events. Hm... it gets a bit complicated.

Quoted 5 lines

As we are going into suspend, we are not really in a position to handle any (DEVTEN) events till we have fully resumed.
So yes, we need to rely on platform specific implementation to wake the System on any USB event.

You may be able to detect vbus change through the connector controller. However, the usb controller is the one that detects host resume. What platform specific implementation do you have outside of the usb controller do you have to get around that?

Quoted 4 lines

Quoted 4 lines

I'm not sure if your platform supports hibernation or if the PME signal on your platform can wakeup the system, but currently dwc3 driver doesn't handle hibernation (device side). If there's no hibernation, there's no PME.

No, in this TI SoC, hibernation feature is not supported in the dwc3 core.

Quoted 7 lines

Quoted 1 line

Actually, I think the dwc3 core is still on during system suspend for you right? Then I think we can use the wakeup event to wakeup system suspend on host resume? You can ignore about PME in this case. You may need to look into what needs stay awake to allow for handling of the dwc3 event.

But in SoC deep-sleep state, all clocks to the dwc3 core are stopped. So I'm not sure if dwc3 events will work.

Quoted 9 lines

Quoted 5 lines

Right, you need to keep those clocks running to detect host resume. There's still some power saving through the dwc3 controller's handling in suspend. You may have some limited power saving from other suspended devices on your setup. However, I don't think we can expect the platform to go into deep-sleep and also handle host resume.

Why not? if the PHY can detect the host resume and wake up the SoC it will work right?

Quoted 10 lines

Hm... I supposed it may be possible. But it may need some unconventional design? The dwc3 controller is currently registered to the phy. For that to work, your phy needs to be able to talk to both the dwc3 controller and some other controller (equivalent to dwc3 PMU) that manages power/interrupt. The dwc3 controller would need to relinquish control to this other phy controller on suspend. The phy driver would then be able to assert interrupt waking up the system on resume sigal detection, which in turn relinquish control to the dwc3 controller. All of this has to work while the phy signaling remains synchronized with the dwc3 controller.

My understanding is that all this is taken care by PHY integration design with DWC3 core on the TI SoC.

Quoted 4 lines

From the patches you sent, I don't see the changes necesssary for this to work. If there is something that I'm missing, please also note it or add it here to the series.

Quoted 7 lines

There is nothing more as the details are taken care by PHY logic and necessary integration with DWC3.

For the PHY wake-up programming details you have already checked this series [1].

[1] - https://urldefense.com/v3/__https://lore.kernel.org/all/[email protected]/__;!!A4F2R9G_pg!ayqpaWwnWIO7SKktk4kI2EJLwDTIv2nYYCgBGHUlt56KTzeYkDEdq2Q5ZvZFsuCWIlcXGET230YY5J3y_sHV$

I may have misunderstood your platform implementation. My understanding is that it can only detect VBUS and that it can only resume on VBUS valid.

Quoted 5 lines

Quoted 1 line

Does the "LINESTATE" here gets asserted if say there's a LFPS detection?

Yes. The wake up logic on the SoC is snooping the UTMI lines from the PHY and on any change it can detect and wake up the SoC.

Are you referring to the utmi_linestate signal? Isn't that for usb2 speed only? Does your platform support usb3 speed?

Quoted 11 lines · Thinh Nguyen

Quoted 10 lines · Roger Quadros

Quoted 6 lines · Thinh Nguyen

I may have misunderstood your platform implementation. My understanding is that it can only detect VBUS and that it can only resume on VBUS valid.

Does the "LINESTATE" here gets asserted if say there's a LFPS detection?

Quoted 6 lines

Quoted 3 lines

Yes. The wake up logic on the SoC is snooping the UTMI lines from the PHY and on any change it can detect and wake up the SoC.

Are you referring to the utmi_linestate signal? Isn't that for usb2 speed only? Does your platform support usb3 speed?

The wake-up on deepSleep feature is only supported for USB2 on this particular SoC.

Quoted 18 lines · Roger Quadros

Quoted 15 lines · Thinh Nguyen

Quoted 13 lines · Roger Quadros

Quoted 6 lines · Thinh Nguyen

I may have misunderstood your platform implementation. My understanding is that it can only detect VBUS and that it can only resume on VBUS valid.

Does the "LINESTATE" here gets asserted if say there's a LFPS detection?

Yes. The wake up logic on the SoC is snooping the UTMI lines from the PHY and on any change it can detect and wake up the SoC.

Quoted 5 lines

Quoted 2 lines

Are you referring to the utmi_linestate signal? Isn't that for usb2 speed only? Does your platform support usb3 speed?

The wake-up on deepSleep feature is only supported for USB2 on this particular SoC.

I mean can your platform operate in usb3 speed. If that's the case, then how do you plan to handle it here.

Also, when you tested this in highspeed, did you observe successful resume? Or did the host have to perform a port reset?