[PATCH v2] phy: qcom: qmp-combo: Leave the USB PCS alone in DP-only mode

[PATCH v2] phy: qcom: qmp-combo: Leave the USB PCS alone in DP-only mode

[Joonhoe Kim]

In DP-only mode (DisplayPort pin assignments C and E) all four lanes
belong to the DisplayPort PHY and the USB PCS stays powered down.
Three paths still touch the USB PCS in this state:

- The USB controller can exit and re-init its PHY in DP-only mode, e.g.
  across a system suspend that loses power. qmp_combo_usb_init() then
  programs the USB serdes, lane and PCS tables over the DisplayPort
  configuration and times out waiting for PHYSTATUS, and
  qmp_combo_usb_exit() resets and powers down a PCS that is not in use.
- qmp_combo_typec_switch_set() powers the USB PCS off and on when the
  orientation changes in DP-only mode, e.g. when a DP-only sink is
  plugged in the other way round while the mux is still in DP-only
  mode.
- qmp_combo_reconfigure_phy() powers the USB PCS off again when the
  port leaves DP-only mode.

Skip the USB PCS in these paths in DP-only mode. Keep the
qmp_combo_com_init()/qmp_combo_com_exit() reference and usb_init_count,
so that qmp_combo_reconfigure_phy() powers the USB PCS up when the port
leaves DP-only mode.

On a Lenovo Y700 Gen 5 (TB323FU, Kaanapali) with a USB-C monitor
in pin assignment C, resume from s2idle fails with "phy initialization
timed-out", the dwc3 resume returns -110 and the following xHCI resume
hangs the system.

Fixes: 896277138c13 ("phy: qcom: qmp-combo: register a typec mux to change the QMPPHY_MODE")
Fixes: 89563a152973 ("phy: qualcomm: qmp-combo: Add preliminary USB4 support")
Signed-off-by: Joonhoe Kim <26rote@gmail.com>
---
Changes in v2:
- Also skip the USB PCS in qmp_combo_typec_switch_set() and in the
  power-off of qmp_combo_reconfigure_phy() in DP-only mode, as the
  Sashiko review of v1 pointed out [1]. Since 89563a152973 the mux keeps
  usb_init_count when it enters DP-only mode, so these paths see a
  non-zero count there.
- v1: https://lore.kernel.org/all/20261022165219.1368-1-26rote@gmail.com/

Tested on v7.3 with the qmp-combo mux changed to keep usb_init_count in
DP-only mode as in phy-next, on a Lenovo Y700 Gen 5 (out-of-tree board
support) with a USB-C monitor in pin assignment C: s2idle resume, a
replug in the other orientation while the mux was still in DP-only
mode (qmp_combo_typec_switch_set() then left the USB PCS alone), and a
USB 3 stick afterwards, which enumerated at SuperSpeed. Kernels without
89563a152973, where the mux drops usb_init_count in DP-only mode, need a
variant that does not count the USB user there either.

[1] https://lore.kernel.org/all/sashiko-outbox-181846@kernel.org/

 drivers/phy/qualcomm/phy-qcom-qmp-combo.c | 28 +++++++++++++++--------
 1 file changed, 18 insertions(+), 10 deletions(-)

diff --git a/drivers/phy/qualcomm/phy-qcom-qmp-combo.c b/drivers/phy/qualcomm/phy-qcom-qmp-combo.c
index 187f0cd11f6c..f1b565852caf 100644
--- a/drivers/phy/qualcomm/phy-qcom-qmp-combo.c
+++ b/drivers/phy/qualcomm/phy-qcom-qmp-combo.c
@@ -4459,10 +4459,12 @@ static int qmp_combo_usb_init(struct phy *phy)
 	if (ret)
 		return ret;
 
-	ret = qmp_combo_usb_power_on(phy);
-	if (ret) {
-		qmp_combo_com_exit(qmp, false);
-		return ret;
+	if (qmp->qmpphy_mode != QMPPHY_MODE_DP_ONLY) {
+		ret = qmp_combo_usb_power_on(phy);
+		if (ret) {
+			qmp_combo_com_exit(qmp, false);
+			return ret;
+		}
 	}
 
 	qmp->usb_init_count++;
@@ -4484,9 +4486,11 @@ static int qmp_combo_usb_exit(struct phy *phy)
 		return 0;
 	}
 
-	ret = qmp_combo_usb_power_off(phy);
-	if (ret)
-		return ret;
+	if (qmp->qmpphy_mode != QMPPHY_MODE_DP_ONLY) {
+		ret = qmp_combo_usb_power_off(phy);
+		if (ret)
+			return ret;
+	}
 
 	ret = qmp_combo_com_exit(qmp, false);
 	if (ret)
@@ -4511,7 +4515,8 @@ static int qmp_combo_reconfigure_phy(struct qmp_combo *qmp, enum qmpphy_mode new
 	dev_dbg(qmp->dev, "qmp_combo_reconfigure_phy: switching from qmpphy mode %d to %d\n",
 		qmp->qmpphy_mode, new_mode);
 
-	if (qmp->usb_init_count || qmp->usb4_init_count)
+	if ((qmp->usb_init_count || qmp->usb4_init_count) &&
+	    qmp->qmpphy_mode != QMPPHY_MODE_DP_ONLY)
 		qmp_combo_usb_power_off(qmp->usb_phy);
 
 	if (qmp->dp_init_count)
@@ -5087,12 +5092,15 @@ static int qmp_combo_typec_switch_set(struct typec_switch_dev *sw,
 	qmp->orientation = orientation;
 
 	if (qmp->init_count) {
-		if (qmp->usb_init_count || qmp->usb4_init_count)
+		bool usb = (qmp->usb_init_count || qmp->usb4_init_count) &&
+			   qmp->qmpphy_mode != QMPPHY_MODE_DP_ONLY;
+
+		if (usb)
 			qmp_combo_usb_power_off(qmp->usb_phy);
 		qmp_combo_com_exit(qmp, true);
 
 		qmp_combo_com_init(qmp, true);
-		if (qmp->usb_init_count || qmp->usb4_init_count)
+		if (usb)
 			qmp_combo_usb_power_on(qmp->usb_phy);
 		if (qmp->dp_init_count)
 			cfg->dp_aux_init(qmp);

base-commit: 2efb7a11f1aa4d2dc285fe2680cca22cb9f72f64
-- 
2.43.0

Re: [PATCH v2] phy: qcom: qmp-combo: Leave the USB PCS alone in DP-only mode

[Abel Vesa]

On 26-10-23 21:35:12, Joonhoe Kim wrote:
> In DP-only mode (DisplayPort pin assignments C and E) all four lanes
> belong to the DisplayPort PHY and the USB PCS stays powered down.
> Three paths still touch the USB PCS in this state:
> 
> - The USB controller can exit and re-init its PHY in DP-only mode, e.g.
>   across a system suspend that loses power. qmp_combo_usb_init() then
>   programs the USB serdes, lane and PCS tables over the DisplayPort
>   configuration and times out waiting for PHYSTATUS, and
>   qmp_combo_usb_exit() resets and powers down a PCS that is not in use.
> - qmp_combo_typec_switch_set() powers the USB PCS off and on when the
>   orientation changes in DP-only mode, e.g. when a DP-only sink is
>   plugged in the other way round while the mux is still in DP-only
>   mode.
> - qmp_combo_reconfigure_phy() powers the USB PCS off again when the
>   port leaves DP-only mode.
> 
> Skip the USB PCS in these paths in DP-only mode. Keep the
> qmp_combo_com_init()/qmp_combo_com_exit() reference and usb_init_count,
> so that qmp_combo_reconfigure_phy() powers the USB PCS up when the port
> leaves DP-only mode.
> 
> On a Lenovo Y700 Gen 5 (TB323FU, Kaanapali) with a USB-C monitor
> in pin assignment C, resume from s2idle fails with "phy initialization
> timed-out", the dwc3 resume returns -110 and the following xHCI resume
> hangs the system.
> 
> Fixes: 896277138c13 ("phy: qcom: qmp-combo: register a typec mux to change the QMPPHY_MODE")
> Fixes: 89563a152973 ("phy: qualcomm: qmp-combo: Add preliminary USB4 support")
> Signed-off-by: Joonhoe Kim <26rote@gmail.com>

Reviewed-by: Abel Vesa <abel.vesa@oss.qualcomm.com>

Re: [PATCH v2] phy: qcom: qmp-combo: Leave the USB PCS alone in DP-only mode

[Manivannan Sadhasivam]

On 2026-10-23 21:35 +0900, Joonhoe Kim wrote:
> In DP-only mode (DisplayPort pin assignments C and E) all four lanes
> belong to the DisplayPort PHY and the USB PCS stays powered down.
> Three paths still touch the USB PCS in this state:
> 
> - The USB controller can exit and re-init its PHY in DP-only mode, e.g.
>   across a system suspend that loses power. qmp_combo_usb_init() then
>   programs the USB serdes, lane and PCS tables over the DisplayPort
>   configuration and times out waiting for PHYSTATUS, and
>   qmp_combo_usb_exit() resets and powers down a PCS that is not in use.
> - qmp_combo_typec_switch_set() powers the USB PCS off and on when the
>   orientation changes in DP-only mode, e.g. when a DP-only sink is
>   plugged in the other way round while the mux is still in DP-only
>   mode.
> - qmp_combo_reconfigure_phy() powers the USB PCS off again when the
>   port leaves DP-only mode.
> 
> Skip the USB PCS in these paths in DP-only mode. Keep the
> qmp_combo_com_init()/qmp_combo_com_exit() reference and usb_init_count,
> so that qmp_combo_reconfigure_phy() powers the USB PCS up when the port
> leaves DP-only mode.
> 
> On a Lenovo Y700 Gen 5 (TB323FU, Kaanapali) with a USB-C monitor
> in pin assignment C, resume from s2idle fails with "phy initialization
> timed-out", the dwc3 resume returns -110 and the following xHCI resume
> hangs the system.
> 
> Fixes: 896277138c13 ("phy: qcom: qmp-combo: register a typec mux to change the QMPPHY_MODE")
> Fixes: 89563a152973 ("phy: qualcomm: qmp-combo: Add preliminary USB4 support")
> Signed-off-by: Joonhoe Kim <26rote@gmail.com>

Acked-by: Manivannan Sadhasivam <manivannan.sadhasivam@oss.qualcomm.com>

- Mani

Re: [PATCH v2] phy: qcom: qmp-combo: Leave the USB PCS alone in DP-only mode

[Konrad Dybcio]

On 10/23/26 2:35 PM, Joonhoe Kim wrote:
> In DP-only mode (DisplayPort pin assignments C and E) all four lanes
> belong to the DisplayPort PHY and the USB PCS stays powered down.
> Three paths still touch the USB PCS in this state:

[...]


> @@ -4459,10 +4459,12 @@ static int qmp_combo_usb_init(struct phy *phy)
>  	if (ret)
>  		return ret;
>  
> -	ret = qmp_combo_usb_power_on(phy);
> -	if (ret) {
> -		qmp_combo_com_exit(qmp, false);
> -		return ret;
> +	if (qmp->qmpphy_mode != QMPPHY_MODE_DP_ONLY) {
> +		ret = qmp_combo_usb_power_on(phy);
> +		if (ret) {
> +			qmp_combo_com_exit(qmp, false);
> +			return ret;
> +		}
>  	}

The USB sub-block provides the PIPE clock to the DWC3 controller.
If we somehow enter DP altmode before DWC3 gets to probe, or if for
any reason we need to access that controller (maybe debugfs reads),
I think this will fail.

Konrad

Re: [PATCH v2] phy: qcom: qmp-combo: Leave the USB PCS alone in DP-only mode

[Konrad Dybcio]

On 10/23/26 2:35 PM, Joonhoe Kim wrote:
> In DP-only mode (DisplayPort pin assignments C and E) all four lanes
> belong to the DisplayPort PHY and the USB PCS stays powered down.
> Three paths still touch the USB PCS in this state:
> 
> - The USB controller can exit and re-init its PHY in DP-only mode, e.g.
>   across a system suspend that loses power. qmp_combo_usb_init() then
>   programs the USB serdes, lane and PCS tables over the DisplayPort
>   configuration and times out waiting for PHYSTATUS, and
>   qmp_combo_usb_exit() resets and powers down a PCS that is not in use.
> - qmp_combo_typec_switch_set() powers the USB PCS off and on when the
>   orientation changes in DP-only mode, e.g. when a DP-only sink is
>   plugged in the other way round while the mux is still in DP-only
>   mode.
> - qmp_combo_reconfigure_phy() powers the USB PCS off again when the
>   port leaves DP-only mode.
> 
> Skip the USB PCS in these paths in DP-only mode. Keep the
> qmp_combo_com_init()/qmp_combo_com_exit() reference and usb_init_count,
> so that qmp_combo_reconfigure_phy() powers the USB PCS up when the port
> leaves DP-only mode.
> 
> On a Lenovo Y700 Gen 5 (TB323FU, Kaanapali) with a USB-C monitor
> in pin assignment C, resume from s2idle fails with "phy initialization
> timed-out", the dwc3 resume returns -110 and the following xHCI resume
> hangs the system.

This sounds like a gap in the DWC3 driver, we should bail out of r/w
any registers if resume failed

Konrad