[PATCH v4 5/5] arm64: dts: qcom: Add Lenovo ThinkBook 16 G7 QOY device tree

[Rob Clark]

On Mon, May 26, 2025 at 1:36 AM Dmitry Baryshkov
<dmitry.baryshkov at oss.qualcomm.com> wrote:
>
> On Sun, May 25, 2025 at 09:43:36PM +0200, Aleksandrs Vinarskis wrote:
> > On Sun, 25 May 2025 at 15:33, Dmitry Baryshkov
> > <dmitry.baryshkov at oss.qualcomm.com> wrote:
> > >
> > > On Sat, May 24, 2025 at 07:58:13PM +0200, Aleksandrs Vinarskis wrote:
> > > > On Sat, 24 May 2025 at 17:33, Dmitry Baryshkov
> > > > <dmitry.baryshkov at oss.qualcomm.com> wrote:
> > > > >
> > > > > On Sat, May 24, 2025 at 01:48:40PM +0200, Jens Glathe via B4 Relay wrote:
> > > > > > From: Jens Glathe <jens.glathe at oldschoolsolutions.biz>
> > > > > >
> > > > > > Device tree for the Lenovo Thinkbook 16 G7 QOY
> > > > > >
> > > > > > The Laptop is a Snapdragon X1 / X1 Plus (Purwa) based device [1].
> > > > > >
> > > > > > Supported features:
> > > > > >
> > > > > > - USB type-c and type-a ports
> > > > > > - Keyboard
> > > > > > - Touchpad (all that are described in the dsdt)
> > > > > > - Touchscreen (described in the dsdt, no known SKUss)
> > > > > > - Display including PWM backlight control
> > > > > > - PCIe devices
> > > > > > - nvme
> > > > > > - SDHC card reader
> > > > > > - ath12k WCN7850 Wifi and Bluetooth
> > > > > > - ADSP and CDSP
> > > > > > - GPIO keys (Lid switch)
> > > > > > - Sound via internal speakers / DMIC / USB / headphone jack
> > > > > > - DP Altmode with 2 lanes (as all of these still do)
> > > > > > - Integrated fingerprint reader (FPC)
> > > > > > - Integrated UVC camera
> > > > > >
> > > > > > Not supported yet:
> > > > > >
> > > > > > - HDMI port.
> > > > > > - EC and some fn hotkeys.
> > > > > >
> > > > > > Limited support yet:
> > > > > >
> > > > > > - SDHC card reader is based on the on-chip sdhc_2 controller, but the driver from
> > > > > > the Snapdragon Dev Kit is only a partial match. It can do normal slow sd cards,
> > > > > > but not UHS-I (SD104) and UHS-II.
> > > > > >
> > > > > > - The GPU is not yet supported. Graphics is only software rendered.
> > > > > >
> > > > > > This work was done without any schematics or non-public knowledge of the device.
> > > > > > So, it is based on the existing x1e device trees, dsdt analysis, using HWInfo
> > > > > > ARM64, and pure guesswork. It has been confirmed, however, that the device really
> > > > > > has 4 NXP PTN3222 eUSB2 repeaters, one of which doesn't have a reset GPIO (eusb5
> > > > > > @43).
> > > > > >
> > > > > > Signed-off-by: Jens Glathe <jens.glathe at oldschoolsolutions.biz>
> > > > > > Co-developed by: Aleksandrs Vinarskis <alex.vinarskis at gmail.com>
> > > > > > ---

[snip]

> > > > > I think that you wrote that GPU isn't supported (yet).
> > > >
> > > > GPU is not supported on SoC level (x1p42100/purwa), not device
> > > > specific. The idea I suggested to Jens was to mimic Asus Zenbook A14
> > > > as per discussion abou purwa GPU and purwa firmware in general [1]: to
> > > > already include correct filename and firmware path (since these are
> > > > known) such that once SoC lands the GPU support laptops with that SoC
> > > > will gain GPU support 'automatically'. As nothing consumes this
> > > > property (on purwa) just yet, adding it already does not cause any
> > > > harm.
> > >
> > >
> > > Not quite. A14 specified the firmware name, but didn't enable the GPU.
> > > This patch actually enables the GPU device. Which means, that we are
> > > going to probe a driver on an unsupported device, using invalid GPU ids.
> > >
> >
> > A14 also enables the GPU, just in the .dsti, not in the purwa .dts.
> > Though now looking at it, perhaps it's a bit misleading.
>
> It should be enabled in the board.dts, not in the interim dtsi.
>
> > Yes you are right, it does give GPU-related error in dmesg, which
> > appear to be otherwise harmless. Perhaps because `x1p42100.dtsi`
> > deletes the compatible of the GPU, hence no driver is probing it?
> >
> > ```
> > [    3.085697] msm_dpu ae01000.display-controller: no GPU device was found
> > ```
> >
> > Otherwise, if you still say GPU should be disabled regardless, shall I
> > also disable it on A14's purwa variant?
>
> Yes, please. Don't set okay status for the GPU until it gets enabled.

Drive-by: Shouldn't the dtb describe the hw and not the state of the
linux kernel's support for the hw?  Ie. if bad things happen if we
describe hw which is missing driver support, shouldn't we fix that in
the driver.

(In the case of the GPU there is the slight wrinkle that we don't have
a gpu-id yet so there is no compatible in the dtb yet.)

BR,
-R