Will we be switching to the ce-no updates for now?

[asiaon123]

Currently, a significant amount of work has been done based on kernel version 4.9; however, CE has already begun supporting the "ce-no" variant of DoVi. Will there be a transition to this?

[cpm-code]

I guess unlikely, here are some thoughts:

Quick Summary & Recommendation

For most users, I recommend using the mainline CoreELEC version. Thanks to the breakthrough for FEL support and its consumer-friendly VS10 implementation, it already covers the needs of most consumer very well.

Moving the Fork to the New Kernel

Porting this fork to the new kernel appears to be a significant undertaking — likely only realistic with substantial AI assistance and a large number of tokens. Given rising costs, the timing is poor for me to absorb this as a hobby project.
Others in the community may be better positioned to take this on. There are now other active repositories in this space that are more likely (and better suited) to pursue the migration.

Potential Quick-Win Un-blockers

If the new kernel already handles the following well, the port could be significantly easier:
• Correct A/V sync for bitstream passthrough — especially if the kernel manages timing natively, removing the need for active feedback/adjustments in Kodi.
• PCM audio handling — if the kernel correctly supports all PCM layouts over HDMI, this would eliminate the need for custom rework.
Feedback and real-world testing on these two areas from the wider community would be very valuable.

Direction & Suitability of My Fork

My fork was primarily a technical exploration tool — focused on understanding limitations, pushing edge cases, and improving specific technical areas. It was never intended as a polished “do-it-all” solution.

As development continued, it diverged further because:

• Changes were made to the kernel and Kodi code to clarify behavior and enable specific features from a clean technical perspective.
• This approach inevitably breaks or deprioritizes features I don’t personally use.

Key limitations:
• Heavily tied to the s922x SoC (requires rework for broader hardware support).
• Large amounts of code specific to the 4.9 kernel may not be needed (or may cause issues) on 5.x.
• Significant re-work required for areas where the new kernel has changed (e.g. Dolby Vision code).

Major Areas of Work (if porting)

Here are the notable technical elements that would need attention:
• PCM Audio — Full support for all speaker layouts.
• Bitstream A/V sync — LAV rework for MAT + base/sync learning for other formats, work for DialNorm.
• HDR PGS handling.
• Startup Audio/Video — Clean playback initialization.
• HEVC decoding & frame pacing improvements (across kernel and Kodi layers).
• UI/OSD values for existing data (flowing from RPU parsing, etc.).
• Speaker layout representation work.
• Dolby Vision processing — tighter unification and refactor of BDMV + FEL support. HDR10+ conversion, original VSVDB work, DV modes, CMv4.0 etc.

Plus lots of more minor items.

[asiaon123]

In other words, does this mean that adapting Kodi 22 based on CE-NG is currently easier?

[cpm-code]

Yes; adapting CE-NG to Kodi 22 and aarch64 is what is already done here, it supports what is required for this narrow version and features to work.

If was looking longer term then the most sensible approach would be make it CE-NO given FEL is working. and have as much as possible in CE upstream. My take is with AI assistance that may now be in the realms of possible, just needs time and tokens and lots of testing and agreement on feature sets etc.

I would not look here, but elsewhere for what comes next and maybe something like that to emerge over time.

[bike3097]

Now the official NO22 build already supports FEL, so the CPM project with 4.9.269 kernel is no longer needed for the time being.

[cpm-code]

Yes please use that one, thx.