Re: [PATCH v2 2/3] arm64: dts: qcom: Add HP EliteBook X G2q 14 AI

[Konrad Dybcio]

On 6/28/26 6:29 PM, Jason Pettit wrote:
> On Sun, 28 Jun 2026 16:02:24 +0300, Dmitry Baryshkov wrote:
>> This is very surprising. Can you check, is it actually the DWC3 or is it
>> some power supply that is being toggled by the USB controller? In the
>> former case it would mean that we miss some bits of PHY programming and
>> still depend on the USB host / USB part of the PHY.
> 
> I tried to check this. I am fairly new to this, so I mostly poked at it on
> the hardware. Here is what I did and what I saw, and you can probably read
> more into it than I can.
> 
> 1. First, with &usb_2 (the dwc3, HS-only) and &usb_2_hsphy enabled, HDMI
>    works. I looked at the power state on that boot to see if the dwc3 turns
>    on anything the DP side might also need. As far as I can tell they are
>    separate. The combo PHY, the dwc3, and the DP controller are each in a
>    different power domain, and the PHY's regulators show up under the PHY,
>    not the dwc3:
> 
>      gcc_usb_2_phy_gdsc    on
>          88e1000.phy                       active   (the combo PHY)
>      gcc_usb30_tert_gdsc   on
>          a000000.usb                       active   (the dwc3)
>      mmcx                  on
>          af64000.displayport-controller    active   (mdss_dp2)
> 
>      vreg_l2f_e0_0p94 -> 88e1000.phy-refgen
>      vreg_l3f_e0_0p91 -> 88e1000.phy-vdda-pll
>      vreg_l4h_e0_1p2  -> 88e1000.phy-vdda-phy
> 
>    So I did not find a rail or GDSC that the dwc3 owns and the PHY needs.
> 
> 2. Then I removed &usb_2 and &usb_2_hsphy, rebuilt, and rebooted. mdss_dp2
>    still binds:
> 
>      msm_dpu ae01000.display-controller: bound af64000.displayport-controller
> 
>    and HDMI-A-1 still reads connected, so the PHY seems powered and the AUX
>    side is working, but there is no picture on the monitor.

We had this odd symptom on the Yoga Slim 7x - but the USB controllers
were enabled.. although we had some issues with USB/UCSI itself..

Bjorn recently discovered that the eDP PHY support for Glymur was
incomplete. Maybe we have some holes in the combophy driver too..

At a glance, this seems to be different vs the reference data, could
you give it a spin?

diff --git a/drivers/phy/qualcomm/phy-qcom-qmp-combo.c b/drivers/phy/qualcomm/phy-qcom-qmp-combo.c
index cdcfad2e86b1..cbb685299d73 100644
--- a/drivers/phy/qualcomm/phy-qcom-qmp-combo.c
+++ b/drivers/phy/qualcomm/phy-qcom-qmp-combo.c
@@ -1675,7 +1675,7 @@ static const struct qmp_phy_init_tbl qmp_v8_dp_serdes_tbl_hbr3[] = {
        QMP_PHY_INIT_CFG(QSERDES_V8_USB43_COM_SYS_CLK_CTRL, 0x02),
        QMP_PHY_INIT_CFG(QSERDES_V8_USB43_COM_SYSCLK_BUF_ENABLE, 0x06),
        QMP_PHY_INIT_CFG(QSERDES_V8_USB43_COM_PLL_IVCO, 0x07),
-       QMP_PHY_INIT_CFG(QSERDES_V8_USB43_COM_SYSCLK_EN_SEL, 0x04),
+       QMP_PHY_INIT_CFG(QSERDES_V8_USB43_COM_SYSCLK_EN_SEL, 0x3b),
        QMP_PHY_INIT_CFG(QSERDES_V8_USB43_COM_VCO_TUNE_CTRL, 0x00),
        QMP_PHY_INIT_CFG(QSERDES_V8_USB43_COM_VCO_TUNE_MAP, 0x00),
        QMP_PHY_INIT_CFG(QSERDES_V8_USB43_COM_CLK_SELECT, 0x30),

Konrad