[PATCH v1] usb: typec: tcpm: Wait for vbus discharge to VSAFE0V before toggling

Subsystems: the rest, usb subsystem, usb typec class, usb typec port controller drivers

STALE2041d

5 messages, 2 authors, 2021-02-19 · open the first message on its own page

[PATCH v1] usb: typec: tcpm: Wait for vbus discharge to VSAFE0V before toggling

From: Badhri Jagan Sridharan <badhri@google.com>
Date: 2021-02-18 11:34:33

When vbus auto discharge is enabled, TCPM can sometimes be faster than
the TCPC i.e. TCPM can go ahead and move the port to unattached state
(involves disabling vbus auto discharge) before TCPC could effectively
discharge vbus to VSAFE0V. This leaves vbus with residual charge and
increases the decay time which prevents tsafe0v from being met.
This change introduces a new state VBUS_DISCHARGE where the TCPM waits
for a maximum of tSafe0V(max) for vbus to discharge to VSAFE0V before
transitioning to unattached state and re-enable toggling. If vbus
discharges to vsafe0v sooner, then, transition to unattached state
happens right away.

Also, while in SNK_READY, when auto discharge is enabled, drive
disconnect based on vbus turning off instead of Rp disappearing on
CC pins. Rp disappearing on CC pins is almost instanteous compared
to vbus decay.

Signed-off-by: Badhri Jagan Sridharan <badhri@google.com>
---
 drivers/usb/typec/tcpm/tcpm.c | 60 +++++++++++++++++++++++++++++++----
 1 file changed, 53 insertions(+), 7 deletions(-)
diff --git a/drivers/usb/typec/tcpm/tcpm.c b/drivers/usb/typec/tcpm/tcpm.c
index be0b6469dd3d..0ed71725980f 100644
--- a/drivers/usb/typec/tcpm/tcpm.c
+++ b/drivers/usb/typec/tcpm/tcpm.c
@@ -62,6 +62,8 @@
 	S(SNK_TRANSITION_SINK_VBUS),		\
 	S(SNK_READY),				\
 						\
+	S(VBUS_DISCHARGE),			\
+						\
 	S(ACC_UNATTACHED),			\
 	S(DEBUG_ACC_ATTACHED),			\
 	S(AUDIO_ACC_ATTACHED),			\
@@ -438,6 +440,9 @@ struct tcpm_port {
 	enum tcpm_ams next_ams;
 	bool in_ams;
 
+	/* Auto vbus discharge state */
+	bool auto_vbus_discharge_enabled;
+
 #ifdef CONFIG_DEBUG_FS
 	struct dentry *dentry;
 	struct mutex logbuffer_lock;	/* log buffer access lock */
@@ -3413,6 +3418,8 @@ static int tcpm_src_attach(struct tcpm_port *port)
 	if (port->tcpc->enable_auto_vbus_discharge) {
 		ret = port->tcpc->enable_auto_vbus_discharge(port->tcpc, true);
 		tcpm_log_force(port, "enable vbus discharge ret:%d", ret);
+		if (!ret)
+			port->auto_vbus_discharge_enabled = true;
 	}
 
 	ret = tcpm_set_roles(port, true, TYPEC_SOURCE, tcpm_data_role_for_source(port));
@@ -3495,6 +3502,8 @@ static void tcpm_reset_port(struct tcpm_port *port)
 	if (port->tcpc->enable_auto_vbus_discharge) {
 		ret = port->tcpc->enable_auto_vbus_discharge(port->tcpc, false);
 		tcpm_log_force(port, "Disable vbus discharge ret:%d", ret);
+		if (!ret)
+			port->auto_vbus_discharge_enabled = false;
 	}
 	port->in_ams = false;
 	port->ams = NONE_AMS;
@@ -3568,6 +3577,8 @@ static int tcpm_snk_attach(struct tcpm_port *port)
 		tcpm_set_auto_vbus_discharge_threshold(port, TYPEC_PWR_MODE_USB, false, VSAFE5V);
 		ret = port->tcpc->enable_auto_vbus_discharge(port->tcpc, true);
 		tcpm_log_force(port, "enable vbus discharge ret:%d", ret);
+		if (!ret)
+			port->auto_vbus_discharge_enabled = true;
 	}
 
 	ret = tcpm_set_roles(port, true, TYPEC_SINK, tcpm_data_role_for_sink(port));
@@ -3684,6 +3695,12 @@ static void run_state_machine(struct tcpm_port *port)
 	switch (port->state) {
 	case TOGGLING:
 		break;
+	case VBUS_DISCHARGE:
+		if (port->port_type == TYPEC_PORT_SRC)
+			tcpm_set_state(port, SRC_UNATTACHED, PD_T_SAFE_0V);
+		else
+			tcpm_set_state(port, SNK_UNATTACHED, PD_T_SAFE_0V);
+		break;
 	/* SRC states */
 	case SRC_UNATTACHED:
 		if (!port->non_pd_role_swap)
@@ -4669,7 +4686,9 @@ static void _tcpm_cc_change(struct tcpm_port *port, enum typec_cc_status cc1,
 	case SRC_READY:
 		if (tcpm_port_is_disconnected(port) ||
 		    !tcpm_port_is_source(port)) {
-			if (port->port_type == TYPEC_PORT_SRC)
+			if (port->auto_vbus_discharge_enabled && !port->vbus_vsafe0v)
+				tcpm_set_state(port, VBUS_DISCHARGE, 0);
+			else if (port->port_type == TYPEC_PORT_SRC)
 				tcpm_set_state(port, SRC_UNATTACHED, 0);
 			else
 				tcpm_set_state(port, SNK_UNATTACHED, 0);
@@ -4703,7 +4722,18 @@ static void _tcpm_cc_change(struct tcpm_port *port, enum typec_cc_status cc1,
 			tcpm_set_state(port, SNK_DEBOUNCED, 0);
 		break;
 	case SNK_READY:
-		if (tcpm_port_is_disconnected(port))
+		/*
+		 * When set_auto_vbus_discharge_threshold is enabled, CC pins go
+		 * away before vbus decays to disconnect threshold. Allow
+		 * disconnect to be driven by vbus disconnect when auto vbus
+		 * discharge is enabled.
+		 *
+		 * EXIT condition is based primarily on vbus disconnect and CC is secondary.
+		 * "A port that has entered into USB PD communications with the Source and
+		 * has seen the CC voltage exceed vRd-USB may monitor the CC pin to detect
+		 * cable disconnect in addition to monitoring VBUS.
+		 */
+		if (!port->auto_vbus_discharge_enabled && tcpm_port_is_disconnected(port))
 			tcpm_set_state(port, unattached_state(port), 0);
 		else if (!port->pd_capable &&
 			 (cc1 != old_cc1 || cc2 != old_cc2))
@@ -4803,9 +4833,16 @@ static void _tcpm_cc_change(struct tcpm_port *port, enum typec_cc_status cc1,
 		 */
 		break;
 
+	case VBUS_DISCHARGE:
+		/* Do nothing. Waiting for vsafe0v signal */
+		break;
 	default:
-		if (tcpm_port_is_disconnected(port))
-			tcpm_set_state(port, unattached_state(port), 0);
+		if (tcpm_port_is_disconnected(port)) {
+			if (port->auto_vbus_discharge_enabled && !port->vbus_vsafe0v)
+				tcpm_set_state(port, VBUS_DISCHARGE, 0);
+			else
+				tcpm_set_state(port, unattached_state(port), 0);
+		}
 		break;
 	}
 }
@@ -4988,9 +5025,12 @@ static void _tcpm_pd_vbus_off(struct tcpm_port *port)
 		break;
 
 	default:
-		if (port->pwr_role == TYPEC_SINK &&
-		    port->attached)
-			tcpm_set_state(port, SNK_UNATTACHED, 0);
+		if (port->pwr_role == TYPEC_SINK && port->attached) {
+			if (port->auto_vbus_discharge_enabled && !port->vbus_vsafe0v)
+				tcpm_set_state(port, VBUS_DISCHARGE, 0);
+			else
+				tcpm_set_state(port, SNK_UNATTACHED, 0);
+		}
 		break;
 	}
 }
@@ -5012,6 +5052,12 @@ static void _tcpm_pd_vbus_vsafe0v(struct tcpm_port *port)
 			tcpm_set_state(port, tcpm_try_snk(port) ? SNK_TRY : SRC_ATTACHED,
 				       PD_T_CC_DEBOUNCE);
 		break;
+	case VBUS_DISCHARGE:
+		if (port->port_type == TYPEC_PORT_SRC)
+			tcpm_set_state(port, SRC_UNATTACHED, 0);
+		else
+			tcpm_set_state(port, SNK_UNATTACHED, 0);
+		break;
 	default:
 		break;
 	}
-- 
2.30.0.617.g56c4b15f3c-goog

Re: [PATCH v1] usb: typec: tcpm: Wait for vbus discharge to VSAFE0V before toggling

From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Date: 2021-02-18 11:30:54

On Thu, Feb 18, 2021 at 02:02:43AM -0800, Badhri Jagan Sridharan wrote:
When vbus auto discharge is enabled, TCPM can sometimes be faster than
the TCPC i.e. TCPM can go ahead and move the port to unattached state
(involves disabling vbus auto discharge) before TCPC could effectively
discharge vbus to VSAFE0V. This leaves vbus with residual charge and
increases the decay time which prevents tsafe0v from being met.
This change introduces a new state VBUS_DISCHARGE where the TCPM waits
for a maximum of tSafe0V(max) for vbus to discharge to VSAFE0V before
transitioning to unattached state and re-enable toggling. If vbus
discharges to vsafe0v sooner, then, transition to unattached state
happens right away.

Also, while in SNK_READY, when auto discharge is enabled, drive
disconnect based on vbus turning off instead of Rp disappearing on
CC pins. Rp disappearing on CC pins is almost instanteous compared
to vbus decay.

Signed-off-by: Badhri Jagan Sridharan <badhri@google.com>
---
 drivers/usb/typec/tcpm/tcpm.c | 60 +++++++++++++++++++++++++++++++----
 1 file changed, 53 insertions(+), 7 deletions(-)
As this seems to be a bugfix, what commit does it fix?  Should it go to
stable kernels?  If so, how far back?

And as this is the merge window, I can't do anything with this until
5.12-rc1 is out, so be prepared for the delay...

thanks,

greg k-h

Re: [PATCH v1] usb: typec: tcpm: Wait for vbus discharge to VSAFE0V before toggling

From: Badhri Jagan Sridharan <badhri@google.com>
Date: 2021-02-18 12:24:52

Hi Greg,

This patch is a bug fix for the following patch which was introduced in 5.11.

commit f321a02caebdd0c56e167610cda2fa148cd96e8b
Author: Badhri Jagan Sridharan [off-list ref]
Date:   Wed Oct 28 23:31:35 2020 -0700

    usb: typec: tcpm: Implement enabling Auto Discharge disconnect support

    TCPCI spec allows TCPC hardware to autonomously discharge the vbus
    capacitance upon disconnect. The expectation is that the TCPM enables
    AutoDischargeDisconnect while entering SNK/SRC_ATTACHED states. Hardware
    then automously discharges vbus when the vbus falls below a certain
    threshold i.e. VBUS_SINK_DISCONNECT_THRESHOLD.

    Apart from enabling the vbus discharge circuit, AutoDischargeDisconnect
    is also used a flag to move TCPCI based TCPC implementations into
    Attached.Snk/Attached.Src state as mentioned in
    Figure 4-15. TCPC State Diagram before a Connection of the
    USB Type-C Port Controller Interface Specification.
    In such TCPC implementations, setting AutoDischargeDisconnect would
    prevent TCPC into entering "Connection_Invalid" state as well.

    Signed-off-by: Badhri Jagan Sridharan [off-list ref]
    Reviewed-by: Heikki Krogerus [off-list ref]
    Link: https://lore.kernel.org/r/20201029063138.1429760-8-badhri@google.com
    Signed-off-by: Greg Kroah-Hartman [off-list ref]

Thanks,
Badhri

On Thu, Feb 18, 2021 at 2:10 AM Greg Kroah-Hartman
[off-list ref] wrote:
On Thu, Feb 18, 2021 at 02:02:43AM -0800, Badhri Jagan Sridharan wrote:
quoted
When vbus auto discharge is enabled, TCPM can sometimes be faster than
the TCPC i.e. TCPM can go ahead and move the port to unattached state
(involves disabling vbus auto discharge) before TCPC could effectively
discharge vbus to VSAFE0V. This leaves vbus with residual charge and
increases the decay time which prevents tsafe0v from being met.
This change introduces a new state VBUS_DISCHARGE where the TCPM waits
for a maximum of tSafe0V(max) for vbus to discharge to VSAFE0V before
transitioning to unattached state and re-enable toggling. If vbus
discharges to vsafe0v sooner, then, transition to unattached state
happens right away.

Also, while in SNK_READY, when auto discharge is enabled, drive
disconnect based on vbus turning off instead of Rp disappearing on
CC pins. Rp disappearing on CC pins is almost instanteous compared
to vbus decay.

Signed-off-by: Badhri Jagan Sridharan <badhri@google.com>
---
 drivers/usb/typec/tcpm/tcpm.c | 60 +++++++++++++++++++++++++++++++----
 1 file changed, 53 insertions(+), 7 deletions(-)
As this seems to be a bugfix, what commit does it fix?  Should it go to
stable kernels?  If so, how far back?

And as this is the merge window, I can't do anything with this until
5.12-rc1 is out, so be prepared for the delay...

thanks,

greg k-h

Re: [PATCH v1] usb: typec: tcpm: Wait for vbus discharge to VSAFE0V before toggling

From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Date: 2021-02-18 12:37:14

On Thu, Feb 18, 2021 at 02:38:45AM -0800, Badhri Jagan Sridharan wrote:
Hi Greg,

This patch is a bug fix for the following patch which was introduced in 5.11.

commit f321a02caebdd0c56e167610cda2fa148cd96e8b
Author: Badhri Jagan Sridharan [off-list ref]
Date:   Wed Oct 28 23:31:35 2020 -0700

    usb: typec: tcpm: Implement enabling Auto Discharge disconnect support

    TCPCI spec allows TCPC hardware to autonomously discharge the vbus
    capacitance upon disconnect. The expectation is that the TCPM enables
    AutoDischargeDisconnect while entering SNK/SRC_ATTACHED states. Hardware
    then automously discharges vbus when the vbus falls below a certain
    threshold i.e. VBUS_SINK_DISCONNECT_THRESHOLD.

    Apart from enabling the vbus discharge circuit, AutoDischargeDisconnect
    is also used a flag to move TCPCI based TCPC implementations into
    Attached.Snk/Attached.Src state as mentioned in
    Figure 4-15. TCPC State Diagram before a Connection of the
    USB Type-C Port Controller Interface Specification.
    In such TCPC implementations, setting AutoDischargeDisconnect would
    prevent TCPC into entering "Connection_Invalid" state as well.

    Signed-off-by: Badhri Jagan Sridharan [off-list ref]
    Reviewed-by: Heikki Krogerus [off-list ref]
    Link: https://lore.kernel.org/r/20201029063138.1429760-8-badhri@google.com
    Signed-off-by: Greg Kroah-Hartman [off-list ref]
Great, then can you resend the patch and add a proper Fixes: tag, along
with a cc: stable as well?

thanks,

greg k-h

Re: [PATCH v1] usb: typec: tcpm: Wait for vbus discharge to VSAFE0V before toggling

From: Badhri Jagan Sridharan <badhri@google.com>
Date: 2021-02-19 09:13:22

Done. Just sent out the following patch and CCed stable@vger.kernel.org as well.

[PATCH v2] usb: typec: tcpm: Wait for vbus discharge to VSAFE0V before toggling

Thanks,
Badhri


On Thu, Feb 18, 2021 at 2:42 AM Greg Kroah-Hartman
[off-list ref] wrote:
On Thu, Feb 18, 2021 at 02:38:45AM -0800, Badhri Jagan Sridharan wrote:
quoted
Hi Greg,

This patch is a bug fix for the following patch which was introduced in 5.11.

commit f321a02caebdd0c56e167610cda2fa148cd96e8b
Author: Badhri Jagan Sridharan [off-list ref]
Date:   Wed Oct 28 23:31:35 2020 -0700

    usb: typec: tcpm: Implement enabling Auto Discharge disconnect support

    TCPCI spec allows TCPC hardware to autonomously discharge the vbus
    capacitance upon disconnect. The expectation is that the TCPM enables
    AutoDischargeDisconnect while entering SNK/SRC_ATTACHED states. Hardware
    then automously discharges vbus when the vbus falls below a certain
    threshold i.e. VBUS_SINK_DISCONNECT_THRESHOLD.

    Apart from enabling the vbus discharge circuit, AutoDischargeDisconnect
    is also used a flag to move TCPCI based TCPC implementations into
    Attached.Snk/Attached.Src state as mentioned in
    Figure 4-15. TCPC State Diagram before a Connection of the
    USB Type-C Port Controller Interface Specification.
    In such TCPC implementations, setting AutoDischargeDisconnect would
    prevent TCPC into entering "Connection_Invalid" state as well.

    Signed-off-by: Badhri Jagan Sridharan [off-list ref]
    Reviewed-by: Heikki Krogerus [off-list ref]
    Link: https://lore.kernel.org/r/20201029063138.1429760-8-badhri@google.com
    Signed-off-by: Greg Kroah-Hartman [off-list ref]
Great, then can you resend the patch and add a proper Fixes: tag, along
with a cc: stable as well?

thanks,

greg k-h
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help