On 05/06/2026 14:14, David Heidelberg via B4 Relay wrote:
Note: WIP tag added, as not everything from the previous review round has
been addressed.

# Short summary

This patch series extends the Qualcomm CAMSS (Camera Subsystem),
including CSID and CSIPHY components, to support C-PHY mode configuration.

# Background and motivation

Modern smartphone cameras increasingly rely on MIPI C-PHY rather than
D-PHY, thanks to its higher data throughput and signal efficiency.
As a result, many OEMs adopt C-PHY interfaces for main (rear) cameras on
Qualcomm-based devices.

Until now, mainline Linux lacked C-PHY configuration support for Qualcomm
chipsets, preventing bring-up of primary camera sensors on several
Snapdragon platforms. This series closes that gap.

  - Introduces C-PHY configuration support for the CAMSS driver stack,
    covering both CSID and CSIPHY blocks.
  - Successfully enables C-PHY operation on the Snapdragon 845 platform.
  - Tested on OnePlus 6 and 6T phones running mainline Linux,
    using the Sony IMX519 main camera sensor.
  - The new configuration allows other chipsets versionsto enable C-PHY by
    simply adding corresponding sensor driver support and csiphy
    initialization data, following the example set for sdm845.

With this patch series, mainline Linux gains working C-PHY support for
Snapdragon 845, paving the way for improved main camera functionality
across many Qualcomm-based devices. The groundwork also simplifies
future enablement efforts for additional SoCs and sensors.

Until merged, the series will be also available at:
   https://codeberg.org/sdm845/linux/commits/branch/b4/qcom-cphy

Signed-off-by: David Heidelberg <[email protected]>
---
Changes in v7:
- Move logic out of switch. (Frank)
- Change csid cphy knob to phy_sel using enum. (Bryan, Frank)
- Link to v6: 
https://lore.kernel.org/r/[email protected]
So I was asking you previously about a dts to support this in some way in effect to prove it works.

I had a meeting with some qcom guys on Friday and the informed me they have based work on top of yours, which is inference enough that it works.

Some Tested-by: from others interested in this work, would be greatly appreciated at this stage.

---
bod

Reply via email to