[PATCH v3 0/3] usb: typec: ps883x: fixes for older Thunderbolt 4 / USB4 docks

[PATCH v3 0/3] usb: typec: ps883x: fixes for older Thunderbolt 4 / USB4 docks

[Jens Glathe via B4 Relay]

On Qualcomm X1E80100 / X1P42100 platforms using the Parade PS883x
retimer, connecting USB4-capable docks such as the Lenovo 40B0 via a
regular Type-C cable (Type-C fallback mode) often results in working
USB but no DisplayPort output.

This series addresses the issue with three driver-only changes:

- Support the legacy TYPEC_DP_STATE_F request (deprecated since DP Alt
  Mode 1.0b) that docks such as the 40B0 and SSK SC220 still send.
  Handle it in the existing DP altmode path in ps883x_set(). F is 
  programmed like D; E is unchanged.

- Add a post-configuration delay after writing the connection-status
  registers so the analog front-end can settle. This makes DP Alt Mode
  hotplug reliable on the 40B0.

- On platforms whose USB4 / DP-tunneling stack is not ready yet, reject
  TYPEC_MODE_USB4 with -EOPNOTSUPP. That forces the Type-C stack to
  fall back to USB3 + DP Alt Mode. The quirk is a machine-compatible
  table inside the driver, not a DT property.

This is a temporary workaround until full USB4 DP tunneling support is
available in the X1E USB4 controller and qmp-combo PHY stack.

As an additional observation, the same dock with a Type-C cable works
well on Thinkpad X13s, Thinkbook 16 G7 QOY and Ideapad 5 14Q8X9, which
do not use this retimer.

To: Heikki Krogerus <heikki.krogerus@linux.intel.com>
To: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: Abel Vesa <abelvesa@kernel.org>
To: Bjorn Andersson <andersson@kernel.org>
To: Konrad Dybcio <konradybcio@kernel.org>
To: Sebastian Reichel <sre@kernel.org>
Cc: linux-usb@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: linux-arm-msm@vger.kernel.org
Cc: stable@vger.kernel.org
Cc: Dr. David Alan Gilbert <dave@treblig.org>
Cc: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>

Assisted-by: Grok(xAI):4.6
Signed-off-by: Jens Glathe <jens.glathe@oldschoolsolutions.biz>
---
Changes in v3:
- reduced the dp logic changein ps883x_set() to only adding 
  TYPEC_DP_STATE_F handling grouped with TYPEC_DP_STATE_D.
- picked up Tested-by tag from David, thanks for the re-test.
- Link to v2: https://lore.kernel.org/r/20260914-ps883x-disable-usb4-v2-0-523c32b4e4d7@oldschoolsolutions.biz

Changes in v2:
- Drop the parade,disable-usb4 DT property, bindings and board DTS
  changes (Sebastian Reichel, Krzysztof Kozlowski). DT describes
  hardware; incomplete software support is not a board property.
- Disable USB4 from the driver via of_machine_is_compatible() for
  qcom,x1e80100, qcom,x1p42100, qcom,hamoa and qcom,purwa. The quirk
  stays kernel-contained and can be dropped later without ABI.
- Keep DP altmode handling in the existing ps883x_set() path instead
  of extracting a helper (Konrad Dybcio). Treat STATE_F like STATE_D
  and STATE_E like STATE_C, matching qcom-qmp-combo.
- Replace usleep_range(20000, 30000) with fsleep(30000) and document
  why the delay is required (Heikki Krogerus).
- Drop the per-event "USB4 disabled via DT property" log; print once
  at probe when the quirk is active.
- Link to v1: https://lore.kernel.org/r/20260718-ps883x-disable-usb4-v1-0-cec86d0b909e@oldschoolsolutions.biz

---
Jens Glathe (3):
      usb: typec: mux: ps883x: support TYPEC_DP_STATE_F
      usb: typec: mux: ps883x: add a delay after writing config regs
      usb: typec: mux: ps883x: disable USB4 on incomplete USB4 platforms

 drivers/usb/typec/mux/ps883x.c | 48 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 48 insertions(+)
---
base-commit: 601057b3eec13d1f37e5b45dd1410facd7b1bdeb
change-id: 20260718-ps883x-disable-usb4-19b0fc13b5ab

Best regards,
-- 
Jens Glathe <jens.glathe@oldschoolsolutions.biz>

[PATCH v3 1/3] usb: typec: mux: ps883x: support TYPEC_DP_STATE_F

[Jens Glathe via B4 Relay]

From: Jens Glathe <jens.glathe@oldschoolsolutions.biz>

Several docks, including the Lenovo 40B0 and SSK SC220, still send the
legacy TYPEC_DP_STATE_F mode request (deprecated since DisplayPort Alt
Mode spec version 1.0b).

Group it with TYPEC_DP_STATE_D.

Assisted-by: Grok(xAI):4.6
Signed-off-by: Jens Glathe <jens.glathe@oldschoolsolutions.biz>
---
 drivers/usb/typec/mux/ps883x.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/usb/typec/mux/ps883x.c b/drivers/usb/typec/mux/ps883x.c
index 64e0a61b776a1..a5696aef054ee 100644
--- a/drivers/usb/typec/mux/ps883x.c
+++ b/drivers/usb/typec/mux/ps883x.c
@@ -207,6 +207,7 @@ static int ps883x_set(struct ps883x_retimer *retimer, struct typec_retimer_state
 
 			switch (state->mode)  {
 			case TYPEC_DP_STATE_D:
+			case TYPEC_DP_STATE_F:
 				cfg0 |= CONN_STATUS_0_USB_3_1_CONNECTED;
 				fallthrough;
 			case TYPEC_DP_STATE_C:

-- 
2.53.0

[PATCH v3 2/3] usb: typec: mux: ps883x: add a delay after writing config regs

[Jens Glathe via B4 Relay]

From: Jens Glathe <jens.glathe@oldschoolsolutions.biz>

The connection-status registers need time to propagate through the
analog front-end (PLLs, lane training) after they are written. A 30 ms
delay after ps883x_configure() brings DisplayPort Alt Mode hotplug on
the Lenovo 40B0 to a reliable success rate.

Tested with T14s G6 and IdeaCentre Mini with the 40B0 dock on the HDMI
port with a Type-C host cable. Qualcomm firmware that programs this
retimer from an MCU uses a similar post-config delay.

Assisted-by: Grok(xAI):4.6
Signed-off-by: Jens Glathe <jens.glathe@oldschoolsolutions.biz>
---
 drivers/usb/typec/mux/ps883x.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/drivers/usb/typec/mux/ps883x.c b/drivers/usb/typec/mux/ps883x.c
index a5696aef054ee..15aee1d613b1a 100644
--- a/drivers/usb/typec/mux/ps883x.c
+++ b/drivers/usb/typec/mux/ps883x.c
@@ -7,6 +7,7 @@
 
 #include <drm/bridge/aux-bridge.h>
 #include <linux/clk.h>
+#include <linux/delay.h>
 #include <linux/gpio/consumer.h>
 #include <linux/i2c.h>
 #include <linux/kernel.h>
@@ -184,6 +185,15 @@ static int ps883x_configure(struct ps883x_retimer *retimer, int cfg0,
 		return ret;
 	}
 
+	/*
+	 * The retimer needs time after the connection-status registers are
+	 * written for the analog front-end (PLLs, lane training) to settle.
+	 * Without this, DisplayPort Alt Mode hotplug is unreliable on some
+	 * docks (e.g. Lenovo 40B0). Qualcomm firmware on platforms that
+	 * program this chip from an MCU uses a similar post-config delay.
+	 */
+	fsleep(30000);
+
 	return 0;
 }
 

-- 
2.53.0

[PATCH v3 3/3] usb: typec: mux: ps883x: disable USB4 on incomplete USB4 platforms

[Jens Glathe via B4 Relay]

From: Jens Glathe <jens.glathe@oldschoolsolutions.biz>

On Qualcomm X1E80100 / X1P42100 platforms using the Parade PS883x
retimer, hotplugging USB4-capable docks such as the Lenovo 40B0 can
result in working USB but no DisplayPort output.

When the dock negotiates USB4, the retimer receives TYPEC_MODE_USB4
and forwards it via typec_mux_set(). The qmp-combo PHY then selects
USB3-only because no classic DP altmode SVID is present in the state,
leaving the DP transmitter and AUX channel disabled.

Reject USB4 with -EOPNOTSUPP on platforms whose USB4 / DP-tunneling
stack is not ready yet. The Type-C stack then falls back to USB3 + DP
Alt Mode. DP altmode configuration continues to use the existing
ps883x_set() path.

Use a machine-compatible table rather than a DT property so the quirk
stays inside the kernel and can be removed later without creating ABI.

Link: https://patch.msgid.link/20260312101431.2375709-1-krishna.kurapati@oss.qualcomm.com
Assisted-by: Grok(xAI):4.6
Tested-by: Dr. David Alan Gilbert <dave@treblig.org>
Signed-off-by: Jens Glathe <jens.glathe@oldschoolsolutions.biz>
---
 drivers/usb/typec/mux/ps883x.c | 37 +++++++++++++++++++++++++++++++++++++
 1 file changed, 37 insertions(+)

diff --git a/drivers/usb/typec/mux/ps883x.c b/drivers/usb/typec/mux/ps883x.c
index 15aee1d613b1a..2ec57cfe748a9 100644
--- a/drivers/usb/typec/mux/ps883x.c
+++ b/drivers/usb/typec/mux/ps883x.c
@@ -13,6 +13,7 @@
 #include <linux/kernel.h>
 #include <linux/module.h>
 #include <linux/mutex.h>
+#include <linux/of.h>
 #include <linux/regmap.h>
 #include <linux/regulator/consumer.h>
 #include <linux/usb/pd.h>
@@ -42,6 +43,22 @@
 #define CONN_STATUS_2_TBT_UNIDIR_LSRX_ACT_LT	BIT(4)
 #define CONN_STATUS_2_USB4_CONNECTED		BIT(7)
 
+/*
+ * Platforms where the USB4 / DP-tunneling stack is not ready yet. Rejecting
+ * USB4 here lets the Type-C stack fall back to USB3 + DP Alt Mode instead of
+ * negotiating USB4 and then failing to drive DisplayPort.
+ *
+ * This is a temporary, kernel-contained quirk (not DT ABI). Drop the entries
+ * once the corresponding USB4 support is complete.
+ */
+static const char * const ps883x_disable_usb4_compats[] = {
+	"qcom,x1e80100",
+	"qcom,x1p42100",
+	"qcom,hamoa",
+	"qcom,purwa",
+	NULL,
+};
+
 struct ps883x_retimer {
 	struct i2c_client *client;
 	struct gpio_desc *reset_gpio;
@@ -63,8 +80,21 @@ struct ps883x_retimer {
 
 	enum typec_orientation orientation;
 	bool in_reset;
+	bool disable_usb4;
 };
 
+static bool ps883x_should_disable_usb4(void)
+{
+	const char * const *compat;
+
+	for (compat = ps883x_disable_usb4_compats; *compat; compat++) {
+		if (of_machine_is_compatible(*compat))
+			return true;
+	}
+
+	return false;
+}
+
 static int ps883x_enable_vregs(struct ps883x_retimer *retimer)
 {
 	struct device *dev = &retimer->client->dev;
@@ -260,6 +290,9 @@ static int ps883x_set(struct ps883x_retimer *retimer, struct typec_retimer_state
 			cfg0 |= CONN_STATUS_0_USB_3_1_CONNECTED;
 			break;
 		case TYPEC_MODE_USB4:
+			if (retimer->disable_usb4)
+				return -EOPNOTSUPP;
+
 			eudo_data = state->data;
 
 			cfg2 |= CONN_STATUS_2_USB4_CONNECTED;
@@ -389,6 +422,10 @@ static int ps883x_retimer_probe(struct i2c_client *client)
 
 	retimer->client = client;
 
+	retimer->disable_usb4 = ps883x_should_disable_usb4();
+	if (retimer->disable_usb4)
+		dev_info(dev, "USB4 disabled until platform USB4 support is complete\n");
+
 	mutex_init(&retimer->lock);
 
 	retimer->regmap = devm_regmap_init_i2c(client, &ps883x_retimer_regmap);

-- 
2.53.0

Re: [PATCH v3 1/3] usb: typec: mux: ps883x: support TYPEC_DP_STATE_F

[Heikki Krogerus]

On Wed, Sep 16, 2026 at 09:36:32PM +0200, Jens Glathe via B4 Relay wrote:
> From: Jens Glathe <jens.glathe@oldschoolsolutions.biz>
> 
> Several docks, including the Lenovo 40B0 and SSK SC220, still send the
> legacy TYPEC_DP_STATE_F mode request (deprecated since DisplayPort Alt
> Mode spec version 1.0b).
> 
> Group it with TYPEC_DP_STATE_D.
> 
> Assisted-by: Grok(xAI):4.6
> Signed-off-by: Jens Glathe <jens.glathe@oldschoolsolutions.biz>

Acked-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>

> ---
>  drivers/usb/typec/mux/ps883x.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/usb/typec/mux/ps883x.c b/drivers/usb/typec/mux/ps883x.c
> index 64e0a61b776a1..a5696aef054ee 100644
> --- a/drivers/usb/typec/mux/ps883x.c
> +++ b/drivers/usb/typec/mux/ps883x.c
> @@ -207,6 +207,7 @@ static int ps883x_set(struct ps883x_retimer *retimer, struct typec_retimer_state
>  
>  			switch (state->mode)  {
>  			case TYPEC_DP_STATE_D:
> +			case TYPEC_DP_STATE_F:
>  				cfg0 |= CONN_STATUS_0_USB_3_1_CONNECTED;
>  				fallthrough;
>  			case TYPEC_DP_STATE_C:
> 
> -- 
> 2.53.0
> 

-- 
heikki

Re: [PATCH v3 2/3] usb: typec: mux: ps883x: add a delay after writing config regs

[Heikki Krogerus]

On Wed, Sep 16, 2026 at 09:36:33PM +0200, Jens Glathe via B4 Relay wrote:
> From: Jens Glathe <jens.glathe@oldschoolsolutions.biz>
> 
> The connection-status registers need time to propagate through the
> analog front-end (PLLs, lane training) after they are written. A 30 ms
> delay after ps883x_configure() brings DisplayPort Alt Mode hotplug on
> the Lenovo 40B0 to a reliable success rate.
> 
> Tested with T14s G6 and IdeaCentre Mini with the 40B0 dock on the HDMI
> port with a Type-C host cable. Qualcomm firmware that programs this
> retimer from an MCU uses a similar post-config delay.
> 
> Assisted-by: Grok(xAI):4.6
> Signed-off-by: Jens Glathe <jens.glathe@oldschoolsolutions.biz>

Acked-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>

> ---
>  drivers/usb/typec/mux/ps883x.c | 10 ++++++++++
>  1 file changed, 10 insertions(+)
> 
> diff --git a/drivers/usb/typec/mux/ps883x.c b/drivers/usb/typec/mux/ps883x.c
> index a5696aef054ee..15aee1d613b1a 100644
> --- a/drivers/usb/typec/mux/ps883x.c
> +++ b/drivers/usb/typec/mux/ps883x.c
> @@ -7,6 +7,7 @@
>  
>  #include <drm/bridge/aux-bridge.h>
>  #include <linux/clk.h>
> +#include <linux/delay.h>
>  #include <linux/gpio/consumer.h>
>  #include <linux/i2c.h>
>  #include <linux/kernel.h>
> @@ -184,6 +185,15 @@ static int ps883x_configure(struct ps883x_retimer *retimer, int cfg0,
>  		return ret;
>  	}
>  
> +	/*
> +	 * The retimer needs time after the connection-status registers are
> +	 * written for the analog front-end (PLLs, lane training) to settle.
> +	 * Without this, DisplayPort Alt Mode hotplug is unreliable on some
> +	 * docks (e.g. Lenovo 40B0). Qualcomm firmware on platforms that
> +	 * program this chip from an MCU uses a similar post-config delay.
> +	 */
> +	fsleep(30000);
> +
>  	return 0;
>  }
>  
> 
> -- 
> 2.53.0
> 

-- 
heikki

Re: [PATCH v3 3/3] usb: typec: mux: ps883x: disable USB4 on incomplete USB4 platforms

[Heikki Krogerus]

On Wed, Sep 16, 2026 at 09:36:34PM +0200, Jens Glathe via B4 Relay wrote:
> From: Jens Glathe <jens.glathe@oldschoolsolutions.biz>
> 
> On Qualcomm X1E80100 / X1P42100 platforms using the Parade PS883x
> retimer, hotplugging USB4-capable docks such as the Lenovo 40B0 can
> result in working USB but no DisplayPort output.
> 
> When the dock negotiates USB4, the retimer receives TYPEC_MODE_USB4
> and forwards it via typec_mux_set(). The qmp-combo PHY then selects
> USB3-only because no classic DP altmode SVID is present in the state,
> leaving the DP transmitter and AUX channel disabled.
> 
> Reject USB4 with -EOPNOTSUPP on platforms whose USB4 / DP-tunneling
> stack is not ready yet. The Type-C stack then falls back to USB3 + DP
> Alt Mode. DP altmode configuration continues to use the existing
> ps883x_set() path.
> 
> Use a machine-compatible table rather than a DT property so the quirk
> stays inside the kernel and can be removed later without creating ABI.
> 
> Link: https://patch.msgid.link/20260312101431.2375709-1-krishna.kurapati@oss.qualcomm.com
> Assisted-by: Grok(xAI):4.6
> Tested-by: Dr. David Alan Gilbert <dave@treblig.org>
> Signed-off-by: Jens Glathe <jens.glathe@oldschoolsolutions.biz>

Acked-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>

> ---
>  drivers/usb/typec/mux/ps883x.c | 37 +++++++++++++++++++++++++++++++++++++
>  1 file changed, 37 insertions(+)
> 
> diff --git a/drivers/usb/typec/mux/ps883x.c b/drivers/usb/typec/mux/ps883x.c
> index 15aee1d613b1a..2ec57cfe748a9 100644
> --- a/drivers/usb/typec/mux/ps883x.c
> +++ b/drivers/usb/typec/mux/ps883x.c
> @@ -13,6 +13,7 @@
>  #include <linux/kernel.h>
>  #include <linux/module.h>
>  #include <linux/mutex.h>
> +#include <linux/of.h>
>  #include <linux/regmap.h>
>  #include <linux/regulator/consumer.h>
>  #include <linux/usb/pd.h>
> @@ -42,6 +43,22 @@
>  #define CONN_STATUS_2_TBT_UNIDIR_LSRX_ACT_LT	BIT(4)
>  #define CONN_STATUS_2_USB4_CONNECTED		BIT(7)
>  
> +/*
> + * Platforms where the USB4 / DP-tunneling stack is not ready yet. Rejecting
> + * USB4 here lets the Type-C stack fall back to USB3 + DP Alt Mode instead of
> + * negotiating USB4 and then failing to drive DisplayPort.
> + *
> + * This is a temporary, kernel-contained quirk (not DT ABI). Drop the entries
> + * once the corresponding USB4 support is complete.
> + */
> +static const char * const ps883x_disable_usb4_compats[] = {
> +	"qcom,x1e80100",
> +	"qcom,x1p42100",
> +	"qcom,hamoa",
> +	"qcom,purwa",
> +	NULL,
> +};
> +
>  struct ps883x_retimer {
>  	struct i2c_client *client;
>  	struct gpio_desc *reset_gpio;
> @@ -63,8 +80,21 @@ struct ps883x_retimer {
>  
>  	enum typec_orientation orientation;
>  	bool in_reset;
> +	bool disable_usb4;
>  };
>  
> +static bool ps883x_should_disable_usb4(void)
> +{
> +	const char * const *compat;
> +
> +	for (compat = ps883x_disable_usb4_compats; *compat; compat++) {
> +		if (of_machine_is_compatible(*compat))
> +			return true;
> +	}
> +
> +	return false;
> +}
> +
>  static int ps883x_enable_vregs(struct ps883x_retimer *retimer)
>  {
>  	struct device *dev = &retimer->client->dev;
> @@ -260,6 +290,9 @@ static int ps883x_set(struct ps883x_retimer *retimer, struct typec_retimer_state
>  			cfg0 |= CONN_STATUS_0_USB_3_1_CONNECTED;
>  			break;
>  		case TYPEC_MODE_USB4:
> +			if (retimer->disable_usb4)
> +				return -EOPNOTSUPP;
> +
>  			eudo_data = state->data;
>  
>  			cfg2 |= CONN_STATUS_2_USB4_CONNECTED;
> @@ -389,6 +422,10 @@ static int ps883x_retimer_probe(struct i2c_client *client)
>  
>  	retimer->client = client;
>  
> +	retimer->disable_usb4 = ps883x_should_disable_usb4();
> +	if (retimer->disable_usb4)
> +		dev_info(dev, "USB4 disabled until platform USB4 support is complete\n");
> +
>  	mutex_init(&retimer->lock);
>  
>  	retimer->regmap = devm_regmap_init_i2c(client, &ps883x_retimer_regmap);
> 
> -- 
> 2.53.0
> 

-- 
heikki