Skip to contentSkip to conversation

usb: typec: tcpm: set correct data role for non-DRD

PATCH In review Jun Li 2 participants 5 messages Dec 27, 2019 – Jan 21, 2020

Subsystems (3)
lore
Share & tools
lore

Jun Li

author
#0 lore raw

PATCH 2/2 usb: typec: tcpm: set correct data role for non-DRD

From: Li Jun <[email protected]>

Since the typec port data role is separated from power role, so check the port data capability when setting data role.

---
 drivers/usb/typec/tcpm/tcpm.c | 24 +++++++++++++++++-------
 1 file changed, 17 insertions(+), 7 deletions(-)
drivers/usb/typec/tcpm/tcpm.c +17 -7
index 56fc356..1f0d82e 100644
@@ -780,7 +780,7 @@ static int tcpm_set_roles(struct tcpm_port *port, bool attached,
 			  enum typec_role role, enum typec_data_role data)
 {
 	enum typec_orientation orientation;
-	enum usb_role usb_role;
+	enum usb_role usb_role = USB_ROLE_NONE;
 	int ret;
 
 	if (port->polarity == TYPEC_POLARITY_CC1)
@@ -788,10 +788,20 @@ static int tcpm_set_roles(struct tcpm_port *port, bool attached,
 	else
 		orientation = TYPEC_ORIENTATION_REVERSE;
 
-	if (data == TYPEC_HOST)
-		usb_role = USB_ROLE_HOST;
-	else
-		usb_role = USB_ROLE_DEVICE;
+	if (port->typec_caps.data == TYPEC_PORT_DRD) {
+		if (data == TYPEC_HOST)
+			usb_role = USB_ROLE_HOST;
+		else
+			usb_role = USB_ROLE_DEVICE;
+	} else if (port->typec_caps.data == TYPEC_PORT_DFP) {
+		if (data == TYPEC_HOST)
+			usb_role = USB_ROLE_HOST;
+		data = TYPEC_HOST;
+	} else {
+		if (data == TYPEC_DEVICE)
+			usb_role = USB_ROLE_DEVICE;
+		data = TYPEC_DEVICE;
+	}
 
 	ret = tcpm_mux_set(port, TYPEC_STATE_USB, usb_role, orientation);
 	if (ret < 0)
@@ -1817,7 +1827,7 @@ static void tcpm_pd_ctrl_request(struct tcpm_port *port,
 		tcpm_set_state(port, SOFT_RESET, 0);
 		break;
 	case PD_CTRL_DR_SWAP:
-		if (port->port_type != TYPEC_PORT_DRP) {
+		if (port->typec_caps.data != TYPEC_PORT_DRD) {
 			tcpm_queue_message(port, PD_MSG_CTRL_REJECT);
 			break;
 		}
@@ -3969,7 +3979,7 @@ static int tcpm_dr_set(struct typec_port *p, enum typec_data_role data)
 	mutex_lock(&port->swap_lock);
 	mutex_lock(&port->lock);
 
-	if (port->port_type != TYPEC_PORT_DRP) {
+	if (port->typec_caps.data != TYPEC_PORT_DRD) {
 		ret = -EINVAL;
 		goto port_unlock;
 	}
Signature · 1 line
2.7.4
maintainer
#1 lore raw

Hi Jun,

Where's the 1/2 of this series?

Quoted 30 lines · Jun Li

From: Li Jun <[email protected]>

Since the typec port data role is separated from power role, so check the port data capability when setting data role.

---
 drivers/usb/typec/tcpm/tcpm.c | 24 +++++++++++++++++-------
 1 file changed, 17 insertions(+), 7 deletions(-)
drivers/usb/typec/tcpm/tcpm.c +1 -4
index 56fc356..1f0d82e 100644
@@ -780,7 +780,7 @@ static int tcpm_set_roles(struct tcpm_port *port, bool attached,
 			  enum typec_role role, enum typec_data_role data)
 {
 	enum typec_orientation orientation;
-	enum usb_role usb_role;
+	enum usb_role usb_role = USB_ROLE_NONE;
 	int ret;
 
 	if (port->polarity == TYPEC_POLARITY_CC1)
@@ -788,10 +788,20 @@ static int tcpm_set_roles(struct tcpm_port *port, bool attached,
 	else
 		orientation = TYPEC_ORIENTATION_REVERSE;
 
-	if (data == TYPEC_HOST)
-		usb_role = USB_ROLE_HOST;
-	else

drivers/usb/typec/tcpm/tcpm.c static int tcpm_set_roles(struct tcpm_port *port, bool attached, · line 799

Quoted 10 lines
drivers/usb/typec/tcpm/tcpm.c +9 -1
-		usb_role = USB_ROLE_DEVICE;
+	if (port->typec_caps.data == TYPEC_PORT_DRD) {
+		if (data == TYPEC_HOST)
+			usb_role = USB_ROLE_HOST;
+		else
+			usb_role = USB_ROLE_DEVICE;
+	} else if (port->typec_caps.data == TYPEC_PORT_DFP) {
+		if (data == TYPEC_HOST)
+			usb_role = USB_ROLE_HOST;
+		data = TYPEC_HOST;

So if data != host, tcpc is told that data == host, but the mux is set to USB_ROLE_NONE. So why tcpc needs to think the role is host in that case?

Shouldn't this function actually return error if the port is DFP only, and TYPEC_DEVICE is requested?

Quoted 28 lines
drivers/usb/typec/tcpm/tcpm.c +7 -2
+	} else {
+		if (data == TYPEC_DEVICE)
+			usb_role = USB_ROLE_DEVICE;
+		data = TYPEC_DEVICE;
+	}
 
 	ret = tcpm_mux_set(port, TYPEC_STATE_USB, usb_role, orientation);
 	if (ret < 0)
@@ -1817,7 +1827,7 @@ static void tcpm_pd_ctrl_request(struct tcpm_port *port,
 		tcpm_set_state(port, SOFT_RESET, 0);
 		break;
 	case PD_CTRL_DR_SWAP:
-		if (port->port_type != TYPEC_PORT_DRP) {
+		if (port->typec_caps.data != TYPEC_PORT_DRD) {
 			tcpm_queue_message(port, PD_MSG_CTRL_REJECT);
 			break;
 		}
@@ -3969,7 +3979,7 @@ static int tcpm_dr_set(struct typec_port *p, enum typec_data_role data)
 	mutex_lock(&port->swap_lock);
 	mutex_lock(&port->lock);
 
-	if (port->port_type != TYPEC_PORT_DRP) {
+	if (port->typec_caps.data != TYPEC_PORT_DRD) {
 		ret = -EINVAL;
 		goto port_unlock;
 	}
Signature · 1 line
2.7.4

thanks,

Signature · 1 line
heikki
author
#2 lore raw
Quoted 10 lines

-----Original Message-----
From: Heikki Krogerus <[email protected]>
Sent: 2020年1月9日 19:52
To: Jun Li <[email protected]>
Cc: [email protected]; [email protected]; dl-linux-imx <[email protected]>; [email protected]
Subject: Re: [PATCH 2/2] usb: typec: tcpm: set correct data role for non-DRD

Quoted 3 lines

Hi Jun,

Where's the 1/2 of this series?

1/2 patch is for controller driver change, so not TO or CC you in mail list. Will pay attention this to avoid confuse.

Quoted 29 lines · Jun Li

drivers/usb/typec/tcpm/tcpm.c static int tcpm_set_roles(struct tcpm_port *port,

Quoted 15 lines

From: Li Jun <[email protected]>

Since the typec port data role is separated from power role, so check the port data capability when setting data role.

---
 drivers/usb/typec/tcpm/tcpm.c | 24 +++++++++++++++++-------
 1 file changed, 17 insertions(+), 7 deletions(-)
(quoted diff) +0 -0

b/drivers/usb/typec/tcpm/tcpm.c index 56fc356..1f0d82e 100644

drivers/usb/typec/tcpm/tcpm.c +0 -0
@@ -780,7 +780,7 @@ static int tcpm_set_roles(struct tcpm_port *port,

bool attached,

Quoted 11 lines
drivers/usb/typec/tcpm/tcpm.c +1 -1
 			  enum typec_role role, enum typec_data_role data)  {
 	enum typec_orientation orientation;
-	enum usb_role usb_role;
+	enum usb_role usb_role = USB_ROLE_NONE;
 	int ret;
 
 	if (port->polarity == TYPEC_POLARITY_CC1) @@ -788,10 +788,20 @@
 static int tcpm_set_roles(struct tcpm_port *port, bool attached,
drivers/usb/typec/tcpm/tcpm.c +0 -0
 	else
 		orientation = TYPEC_ORIENTATION_REVERSE;
Quoted 16 lines

drivers/usb/typec/tcpm/tcpm.c static int tcpm_set_roles(struct tcpm_port *port, bool attached, · line 799

Quoted 13 lines
drivers/usb/typec/tcpm/tcpm.c +9 -4
-	if (data == TYPEC_HOST)
-		usb_role = USB_ROLE_HOST;
-	else
-		usb_role = USB_ROLE_DEVICE;
+	if (port->typec_caps.data == TYPEC_PORT_DRD) {
+		if (data == TYPEC_HOST)
+			usb_role = USB_ROLE_HOST;
+		else
+			usb_role = USB_ROLE_DEVICE;
+	} else if (port->typec_caps.data == TYPEC_PORT_DFP) {
+		if (data == TYPEC_HOST)
+			usb_role = USB_ROLE_HOST;
+		data = TYPEC_HOST;

So if data != host, tcpc is told that data == host, but the mux is set to USB_ROLE_NONE. So why tcpc needs to think the role is host in that case?

enum usb_role {
USB_ROLE_NONE,
USB_ROLE_HOST,
USB_ROLE_DEVICE,
};

enum typec_data_role {
TYPEC_DEVICE,
TYPEC_HOST,
};

If the port only support DFP(host), I think we should never tell tcpc the data is TYPEC_DEVICE, so TYPEC_HOST.

Quoted 3 lines

Shouldn't this function actually return error if the port is DFP only, and TYPEC_DEVICE is requested?

Current TCPM use one API to set both power and data role, doesn't consider the case of dual power role but single data role. Return error in tcpm_set_roles() may lose the setting for power role, I think the current change is use correct data role value when call to tcpm_set_roles().
For simple, I didn't change the caller places of tcpm_set_roles(), so just override the data and usb_role to be reasonable value here.

thanks
Jun

Quoted 36 lines

drivers/usb/typec/tcpm/tcpm.c static void tcpm_pd_ctrl_request(struct

Quoted 10 lines
drivers/usb/typec/tcpm/tcpm.c +5 -0
+	} else {
+		if (data == TYPEC_DEVICE)
+			usb_role = USB_ROLE_DEVICE;
+		data = TYPEC_DEVICE;
+	}
 
 	ret = tcpm_mux_set(port, TYPEC_STATE_USB, usb_role, orientation);
 	if (ret < 0)
@@ -1817,7 +1827,7 @@ static void tcpm_pd_ctrl_request(struct

tcpm_port *port,

drivers/usb/typec/tcpm/tcpm.c static int tcpm_dr_set(struct typec_port *p,

Quoted 9 lines
drivers/usb/typec/tcpm/tcpm.c +1 -1
 		tcpm_set_state(port, SOFT_RESET, 0);
 		break;
 	case PD_CTRL_DR_SWAP:
-		if (port->port_type != TYPEC_PORT_DRP) {
+		if (port->typec_caps.data != TYPEC_PORT_DRD) {
 			tcpm_queue_message(port, PD_MSG_CTRL_REJECT);
 			break;
 		}
@@ -3969,7 +3979,7 @@ static int tcpm_dr_set(struct typec_port *p,

enum typec_data_role data)

Quoted 10 lines
drivers/usb/typec/tcpm/tcpm.c +1 -2
 	mutex_lock(&port->swap_lock);
 	mutex_lock(&port->lock);
 
-	if (port->port_type != TYPEC_PORT_DRP) {
+	if (port->typec_caps.data != TYPEC_PORT_DRD) {
 		ret = -EINVAL;
 		goto port_unlock;
 	}
--

2.7.4

thanks,

--
heikki

maintainer
#3 lore raw

Hi Jun,

Quoted 67 lines · Jun Li
Quoted 11 lines

-----Original Message-----
From: Heikki Krogerus <[email protected]>
Sent: 2020年1月9日 19:52
To: Jun Li <[email protected]>
Cc: [email protected]; [email protected]; dl-linux-imx <[email protected]>; [email protected]
Subject: Re: [PATCH 2/2] usb: typec: tcpm: set correct data role for non-DRD

Hi Jun,

Where's the 1/2 of this series?

1/2 patch is for controller driver change, so not TO or CC you in mail list. Will pay attention this to avoid confuse.

Quoted 29 lines · Jun Li

drivers/usb/typec/tcpm/tcpm.c static int tcpm_set_roles(struct tcpm_port *port,

Quoted 15 lines

From: Li Jun <[email protected]>

Since the typec port data role is separated from power role, so check the port data capability when setting data role.

---
 drivers/usb/typec/tcpm/tcpm.c | 24 +++++++++++++++++-------
 1 file changed, 17 insertions(+), 7 deletions(-)
(quoted diff) +0 -0

b/drivers/usb/typec/tcpm/tcpm.c index 56fc356..1f0d82e 100644

drivers/usb/typec/tcpm/tcpm.c +0 -0
@@ -780,7 +780,7 @@ static int tcpm_set_roles(struct tcpm_port *port,

bool attached,

Quoted 11 lines
drivers/usb/typec/tcpm/tcpm.c +1 -1
 			  enum typec_role role, enum typec_data_role data)  {
 	enum typec_orientation orientation;
-	enum usb_role usb_role;
+	enum usb_role usb_role = USB_ROLE_NONE;
 	int ret;
 
 	if (port->polarity == TYPEC_POLARITY_CC1) @@ -788,10 +788,20 @@
 static int tcpm_set_roles(struct tcpm_port *port, bool attached,
drivers/usb/typec/tcpm/tcpm.c +0 -0
 	else
 		orientation = TYPEC_ORIENTATION_REVERSE;
Quoted 16 lines

drivers/usb/typec/tcpm/tcpm.c static int tcpm_set_roles(struct tcpm_port *port, bool attached, · line 799

Quoted 13 lines
drivers/usb/typec/tcpm/tcpm.c +9 -4
-	if (data == TYPEC_HOST)
-		usb_role = USB_ROLE_HOST;
-	else
-		usb_role = USB_ROLE_DEVICE;
+	if (port->typec_caps.data == TYPEC_PORT_DRD) {
+		if (data == TYPEC_HOST)
+			usb_role = USB_ROLE_HOST;
+		else
+			usb_role = USB_ROLE_DEVICE;
+	} else if (port->typec_caps.data == TYPEC_PORT_DFP) {
+		if (data == TYPEC_HOST)
+			usb_role = USB_ROLE_HOST;
+		data = TYPEC_HOST;

So if data != host, tcpc is told that data == host, but the mux is set to USB_ROLE_NONE. So why tcpc needs to think the role is host in that case?

enum usb_role {
USB_ROLE_NONE,
USB_ROLE_HOST,
USB_ROLE_DEVICE,
};

Quoted 7 lines

enum typec_data_role {
TYPEC_DEVICE,
TYPEC_HOST,
};

If the port only support DFP(host), I think we should never tell tcpc the data is TYPEC_DEVICE, so TYPEC_HOST.

But we should also not have to "lie" and force the role into something that works.

Quoted 9 lines
Quoted 2 lines

Shouldn't this function actually return error if the port is DFP only, and TYPEC_DEVICE is requested?

Current TCPM use one API to set both power and data role, doesn't consider the case of dual power role but single data role. Return error in tcpm_set_roles() may lose the setting for power role, I think the current change is use correct data role value when call to tcpm_set_roles().
For simple, I didn't change the caller places of tcpm_set_roles(), so just override the data and usb_role to be reasonable value here.

I think the correct thing to do would be to fix the places where the function is called, and here return error if the unsupported role is attempted. I hate to be picky, but this looks like a workaround that may potentially hide real issues in the code.

Quoted 23 lines
Quoted 23 lines

drivers/usb/typec/tcpm/tcpm.c static void tcpm_pd_ctrl_request(struct

Quoted 9 lines
drivers/usb/typec/tcpm/tcpm.c +5 -0
+	} else {
+		if (data == TYPEC_DEVICE)
+			usb_role = USB_ROLE_DEVICE;
+		data = TYPEC_DEVICE;
+	}
 
 	ret = tcpm_mux_set(port, TYPEC_STATE_USB, usb_role, orientation);
 	if (ret < 0)
@@ -1817,7 +1827,7 @@ static void tcpm_pd_ctrl_request(struct

tcpm_port *port,

drivers/usb/typec/tcpm/tcpm.c static int tcpm_dr_set(struct typec_port *p,

Quoted 9 lines
drivers/usb/typec/tcpm/tcpm.c +1 -1
 		tcpm_set_state(port, SOFT_RESET, 0);
 		break;
 	case PD_CTRL_DR_SWAP:
-		if (port->port_type != TYPEC_PORT_DRP) {
+		if (port->typec_caps.data != TYPEC_PORT_DRD) {
 			tcpm_queue_message(port, PD_MSG_CTRL_REJECT);
 			break;
 		}
@@ -3969,7 +3979,7 @@ static int tcpm_dr_set(struct typec_port *p,

enum typec_data_role data)

Quoted 3 lines
drivers/usb/typec/tcpm/tcpm.c +0 -0
 	mutex_lock(&port->swap_lock);
 	mutex_lock(&port->lock);
 
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
drivers/usb/typec/tcpm/tcpm.c +1 -2
-	if (port->port_type != TYPEC_PORT_DRP) {
+	if (port->typec_caps.data != TYPEC_PORT_DRD) {
 		ret = -EINVAL;
 		goto port_unlock;
 	}
--

2.7.4

thanks,

Signature · 1 line
heikki
author
#4 lore raw

Hi Heikki,

Quoted 100 lines

-----Original Message-----
From: Heikki Krogerus <[email protected]>
Sent: 2020年1月20日 21:51
To: Jun Li <[email protected]>
Cc: [email protected]; [email protected]; dl-linux-imx <[email protected]>; [email protected]
Subject: Re: [PATCH 2/2] usb: typec: tcpm: set correct data role for non-DRD

Hi Jun,

Quoted 35 lines · Jun Li
Quoted 9 lines

-----Original Message-----
From: Heikki Krogerus <[email protected]>
Sent: 2020年1月9日 19:52
To: Jun Li <[email protected]>
Cc: [email protected]; [email protected]; dl-linux-imx <[email protected]>; [email protected]
Subject: Re: [PATCH 2/2] usb: typec: tcpm: set correct data role for non-DRD

Quoted 3 lines

Hi Jun,

Where's the 1/2 of this series?

1/2 patch is for controller driver change, so not TO or CC you in mail list. Will pay attention this to avoid confuse.

Quoted 19 lines · Jun Li
Quoted 15 lines

From: Li Jun <[email protected]>

Since the typec port data role is separated from power role, so check the port data capability when setting data role.

---
 drivers/usb/typec/tcpm/tcpm.c | 24 +++++++++++++++++-------
 1 file changed, 17 insertions(+), 7 deletions(-)
(quoted diff) +0 -0

b/drivers/usb/typec/tcpm/tcpm.c index 56fc356..1f0d82e 100644

drivers/usb/typec/tcpm/tcpm.c static int tcpm_set_roles(struct tcpm_port · line 780

Quoted 1 line
drivers/usb/typec/tcpm/tcpm.c +0 -0
@@ -780,7 +780,7 @@ static int tcpm_set_roles(struct tcpm_port
 *port,

bool attached,

Quoted 7 lines
Quoted 7 lines

drivers/usb/typec/tcpm/tcpm.c static int tcpm_set_roles(struct tcpm_port *port, bool attached,

Quoted 7 lines
drivers/usb/typec/tcpm/tcpm.c +1 -1
 			  enum typec_role role, enum typec_data_role data)  {
 	enum typec_orientation orientation;
-	enum usb_role usb_role;
+	enum usb_role usb_role = USB_ROLE_NONE;
 	int ret;
 
 	if (port->polarity == TYPEC_POLARITY_CC1) @@ -788,10 +788,20

@@

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

static int tcpm_set_roles(struct tcpm_port *port, bool attached,

drivers/usb/typec/tcpm/tcpm.c +0 -0
 	else
 		orientation = TYPEC_ORIENTATION_REVERSE;
Quoted 16 lines
Quoted 8 lines
Quoted 8 lines
drivers/usb/typec/tcpm/tcpm.c +4 -4
-	if (data == TYPEC_HOST)
-		usb_role = USB_ROLE_HOST;
-	else
-		usb_role = USB_ROLE_DEVICE;
+	if (port->typec_caps.data == TYPEC_PORT_DRD) {
+		if (data == TYPEC_HOST)
+			usb_role = USB_ROLE_HOST;
+		else
Quoted 8 lines

drivers/usb/typec/tcpm/tcpm.c static int tcpm_set_roles(struct tcpm_port *port, bool attached, · line 799

Quoted 5 lines
drivers/usb/typec/tcpm/tcpm.c +5 -0
+			usb_role = USB_ROLE_DEVICE;
+	} else if (port->typec_caps.data == TYPEC_PORT_DFP) {
+		if (data == TYPEC_HOST)
+			usb_role = USB_ROLE_HOST;
+		data = TYPEC_HOST;

So if data != host, tcpc is told that data == host, but the mux is set to USB_ROLE_NONE. So why tcpc needs to think the role is host in

that case?

Quoted 7 lines

enum usb_role {
USB_ROLE_NONE,
USB_ROLE_HOST,
USB_ROLE_DEVICE,
};

Quoted 7 lines

enum typec_data_role {
TYPEC_DEVICE,
TYPEC_HOST,
};

If the port only support DFP(host), I think we should never tell tcpc the data is TYPEC_DEVICE, so TYPEC_HOST.

But we should also not have to "lie" and force the role into something that works.

Quoted 7 lines
Quoted 2 lines

Shouldn't this function actually return error if the port is DFP only, and TYPEC_DEVICE is requested?

Current TCPM use one API to set both power and data role, doesn't consider the case of dual power role but single data role. Return error in tcpm_set_roles() may lose the setting for power role, I think the current change is use correct data role value when call to

Quoted 8 lines

tcpm_set_roles().

Quoted 2 lines

For simple, I didn't change the caller places of tcpm_set_roles(), so just override the data and usb_role to be reasonable value here.

I think the correct thing to do would be to fix the places where the function is called, and here return error if the unsupported role is attempted. I hate to be picky, but this looks like a workaround that may potentially hide real issues in the code.

OK, understood, I will send v2.

Thanks
Li Jun

Quoted 38 lines
Quoted 9 lines
Quoted 9 lines

drivers/usb/typec/tcpm/tcpm.c static int tcpm_set_roles(struct tcpm_port *port, bool attached,

Quoted 9 lines
drivers/usb/typec/tcpm/tcpm.c +5 -0
+	} else {
+		if (data == TYPEC_DEVICE)
+			usb_role = USB_ROLE_DEVICE;
+		data = TYPEC_DEVICE;
+	}
 
 	ret = tcpm_mux_set(port, TYPEC_STATE_USB, usb_role,

orientation);

Quoted 16 lines
Quoted 16 lines

drivers/usb/typec/tcpm/tcpm.c static void tcpm_pd_ctrl_request(struct

Quoted 2 lines
drivers/usb/typec/tcpm/tcpm.c +0 -0
 	if (ret < 0)
@@ -1817,7 +1827,7 @@ static void tcpm_pd_ctrl_request(struct

tcpm_port *port,

drivers/usb/typec/tcpm/tcpm.c static int tcpm_dr_set(struct typec_port *p,

Quoted 9 lines
drivers/usb/typec/tcpm/tcpm.c +1 -1
 		tcpm_set_state(port, SOFT_RESET, 0);
 		break;
 	case PD_CTRL_DR_SWAP:
-		if (port->port_type != TYPEC_PORT_DRP) {
+		if (port->typec_caps.data != TYPEC_PORT_DRD) {
 			tcpm_queue_message(port, PD_MSG_CTRL_REJECT);
 			break;
 		}
@@ -3969,7 +3979,7 @@ static int tcpm_dr_set(struct typec_port *p,

enum typec_data_role data)

Quoted 3 lines
drivers/usb/typec/tcpm/tcpm.c +0 -0
 	mutex_lock(&port->swap_lock);
 	mutex_lock(&port->lock);
 
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
drivers/usb/typec/tcpm/tcpm.c +1 -2
-	if (port->port_type != TYPEC_PORT_DRP) {
+	if (port->typec_caps.data != TYPEC_PORT_DRD) {
 		ret = -EINVAL;
 		goto port_unlock;
 	}
--

2.7.4

thanks,

--
heikki