mbox series

[0/3] Add x1e Dell Inpsiron 14p

Message ID 20250424-qcom-linux-arm64-for-6-16-dell-inspiron14p-v1-0-ace76b31d024@linaro.org
Headers show
Series Add x1e Dell Inpsiron 14p | expand

Message

Bryan O'Donoghue April 23, 2025, 11:53 p.m. UTC
Add in a dtsi for the Dell Inspiron 14p.

I'm currently using this as a daily driver including sending this series
from. Its reasonably stable on 6.15-rcX.

The first two patches can be applied without dependency, the final patch
relies on patches staged for -next in the media tree to be applied,
presently.

https://gitlab.freedesktop.org/linux-media/media-committers/-/commit/2ab7f87a7f4bf392e3836a2600f115a1baa1415c
https://lore.kernel.org/linux-media/20250407-b4-media-comitters-next-25-03-13-ov02e10-v4-0-211e3e6fae90@linaro.org/

Working for me included in this series:

- USB
  Both Type-C USB ports
  Type-A port
  Fingerprint reader
- WiFi
- Bluetooth
- RGB Camera
- Toucpad, keyboard
- Display

Working for me but not included:
- Audio jack
- Iris

Not working:
- Speaker output
  Still WiP haven't touched this in a while

- Battery
  Dell has its own Embedded Controller likely from the x86 platform reused
  on Qcom which we need to reverse engineer or get some information on to
  make faster progress on.

- cDSP
  Haven't tried to bring this up.

- EVA
  No driver haven't tried it.

- Bugs
  Occasionally when resuming I get a fencing error with hyperlock and
  freedreno, TBH it looks like something in the GPU or SMMU according to
  Rob Clark: https://pastebin.com/AWjCyaap

  Ath12k has been splatting for me up to 6.14 when removing a SSID.
  I switched on ath12k debug when going to 6.15 and predictably couldn't
  reproduce the bug, either because the timings have changed as a result
  of Heisenbugging or because a fix has been slipped into ath12k.

  Other than those two I'm pretty happy with this system.

  The DTS is based on Aleksandrs Vinarskis XPS, Lenovo T14s and Qcom CRD.
 

Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
---
Bryan O'Donoghue (3):
      dt-bindings: arm: qcom: Add Dell Inspiron 14 Plus 7441
      arm64: dts: qcom: Add support for X1E80100 Dell Inspirion 14 Plus 7441
      arm64: dts: qcom: x1e80100-dell-inspiron14-7441: Switch on CAMSS RGB sensor

 Documentation/devicetree/bindings/arm/qcom.yaml    |    1 +
 arch/arm64/boot/dts/qcom/Makefile                  |    1 +
 .../qcom/x1e80100-dell-inspirion-14-plus-7441.dts  | 1490 ++++++++++++++++++++
 3 files changed, 1492 insertions(+)
---
base-commit: f7570505263aff2b63142f0d68fa607cd60eb060
change-id: 20250417-qcom-linux-arm64-for-6-16-dell-inspiron14p-ed68cd65ebad

Best regards,

Comments

Rob Clark April 24, 2025, 3:42 p.m. UTC | #1
On Wed, Apr 23, 2025 at 4:53 PM Bryan O'Donoghue
<bryan.odonoghue@linaro.org> wrote:
>
> Add in a dtsi for the Dell Inspiron 14p.
>
> I'm currently using this as a daily driver including sending this series
> from. Its reasonably stable on 6.15-rcX.
>
> The first two patches can be applied without dependency, the final patch
> relies on patches staged for -next in the media tree to be applied,
> presently.
>
> https://gitlab.freedesktop.org/linux-media/media-committers/-/commit/2ab7f87a7f4bf392e3836a2600f115a1baa1415c
> https://lore.kernel.org/linux-media/20250407-b4-media-comitters-next-25-03-13-ov02e10-v4-0-211e3e6fae90@linaro.org/
>
> Working for me included in this series:
>
> - USB
>   Both Type-C USB ports
>   Type-A port
>   Fingerprint reader
> - WiFi
> - Bluetooth
> - RGB Camera
> - Toucpad, keyboard
> - Display
>
> Working for me but not included:
> - Audio jack
> - Iris
>
> Not working:
> - Speaker output
>   Still WiP haven't touched this in a while
>
> - Battery
>   Dell has its own Embedded Controller likely from the x86 platform reused
>   on Qcom which we need to reverse engineer or get some information on to
>   make faster progress on.
>
> - cDSP
>   Haven't tried to bring this up.
>
> - EVA
>   No driver haven't tried it.
>
> - Bugs
>   Occasionally when resuming I get a fencing error with hyperlock and
>   freedreno, TBH it looks like something in the GPU or SMMU according to
>   Rob Clark: https://pastebin.com/AWjCyaap

Did you have the same issues on v6.14 or earlier?

The smmu pm usage count underflow is concerning, the smmu pm should be
tied to the GPU device via device_link_add().  I've never seen
something similar, but I'm still on v6.14.

Could you try reverting commit 0a679336dc17 ("iommu/arm-smmu: Set rpm
auto_suspend once during probe")

BR,
-R

>
>   Ath12k has been splatting for me up to 6.14 when removing a SSID.
>   I switched on ath12k debug when going to 6.15 and predictably couldn't
>   reproduce the bug, either because the timings have changed as a result
>   of Heisenbugging or because a fix has been slipped into ath12k.
>
>   Other than those two I'm pretty happy with this system.
>
>   The DTS is based on Aleksandrs Vinarskis XPS, Lenovo T14s and Qcom CRD.
>
>
> Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
> ---
> Bryan O'Donoghue (3):
>       dt-bindings: arm: qcom: Add Dell Inspiron 14 Plus 7441
>       arm64: dts: qcom: Add support for X1E80100 Dell Inspirion 14 Plus 7441
>       arm64: dts: qcom: x1e80100-dell-inspiron14-7441: Switch on CAMSS RGB sensor
>
>  Documentation/devicetree/bindings/arm/qcom.yaml    |    1 +
>  arch/arm64/boot/dts/qcom/Makefile                  |    1 +
>  .../qcom/x1e80100-dell-inspirion-14-plus-7441.dts  | 1490 ++++++++++++++++++++
>  3 files changed, 1492 insertions(+)
> ---
> base-commit: f7570505263aff2b63142f0d68fa607cd60eb060
> change-id: 20250417-qcom-linux-arm64-for-6-16-dell-inspiron14p-ed68cd65ebad
>
> Best regards,
> --
> Bryan O'Donoghue <bryan.odonoghue@linaro.org>
>
>
Bryan O'Donoghue April 24, 2025, 3:45 p.m. UTC | #2
On 24/04/2025 16:42, Rob Clark wrote:
>> - Bugs
>>    Occasionally when resuming I get a fencing error with hyperlock and
>>    freedreno, TBH it looks like something in the GPU or SMMU according to
>>    Rob Clark:https://pastebin.com/AWjCyaap
> Did you have the same issues on v6.14 or earlier?
> 
> The smmu pm usage count underflow is concerning, the smmu pm should be
> tied to the GPU device via device_link_add().  I've never seen
> something similar, but I'm still on v6.14.
> 
> Could you try reverting commit 0a679336dc17 ("iommu/arm-smmu: Set rpm
> auto_suspend once during probe")
> 
> BR,
> -R

6.15-rc1

I've seen this bug exactly twice. NP to revert but not so sure with only 
a sample set of two, I could say the change fixes.

Anyway, I'll revert and if I see the bug again, we know its not that.

---
bod