Skip to content

device/lynx: add the display overlay - #829

Open
cx3301J wants to merge 2 commits into
derekross:mainfrom
cx3301J:lynx-display-overlay
Open

cx3301J wants to merge 2 commits into
derekross:mainfrom
cx3301J:lynx-display-overlay

Conversation

@cx3301J

@cx3301J cx3301J commented Nov 15, 2026

Copy link
Copy Markdown
Contributor

The kernel's lynx DTS now leaves DSIM0, DECON0, the display clocks and the GPU disabled. This overlay turns them on for a phone with the S6E3FC5 panel, so a board without it keeps the bootloader framebuffer.

Part of linux-omarchy-mobile#202.

@derekross

Copy link
Copy Markdown
Owner

Thank you, @cx3301J. An opt-in overlay with the nodes disabled in the kernel is just what I asked for, and the file compiles and applies cleanly to #202's 7a device tree. I checked it with that PR's build.

Before it can go in:

  1. It waits for linux-omarchy-mobile#202. At device/tegu MODEM-PLAN, AGENTS: GNSS gets a fix on the 9a #202's current head the DECON doesn't bind on the 7a (details there), so there's nothing for this overlay to turn on yet. On today's published kernel it would enable DECON0 without its scanout pool, plus the GPU. Could you say at the top of the file that it needs a kernel with device/tegu MODEM-PLAN, AGENTS: GNSS gets a fix on the 9a #202?
  2. The GPU. device/tegu MODEM-PLAN, AGENTS: GNSS gets a fix on the 9a #202 currently enables the GPU in the device tree, so the GPU fragment here does nothing, and the commit message says the opposite. Whichever way device/tegu MODEM-PLAN, AGENTS: GNSS gets a fix on the 9a #202 ends up, please make the two agree.
  3. A short NOTES entry in device/lynx/NOTES.md, like the Pixel 7's MFC overlay got: what the overlay enables, which kernel it needs, how to apply it (it isn't in DTB_OVERLAYS, so it's applied by hand), and your cold-boot run from device-lynx 0.2.3: persist the DSIM transmit enable #438 with the kernel it ran on. device/lynx: clarify historical Wi-Fi source and safety evidence #822 also edits that file; whichever lands second rebases.

Two small things: the file is executable (mode 755), while every other overlay is 644. And CONTRIBUTING §7 asks for SPDX-License-Identifier: MIT on new files in this repository, bring-up overlays included.

@cx3301J
cx3301J force-pushed the lynx-display-overlay branch 2 times, most recently from 8046a5c to b57ac89 Compare November 16, 2026 13:11
The base DTS ships DSIM0, DECON0 and the display clocks disabled. This
overlay turns them on for a phone with the S6E3FC5 panel.

Signed-off-by: cx3301J <cx3301J@users.noreply.github.com>
Assisted-by: DeepSeek
@cx3301J
cx3301J force-pushed the lynx-display-overlay branch from b57ac89 to f94f424 Compare November 17, 2026 13:01
@cx3301J

cx3301J commented Nov 17, 2026

Copy link
Copy Markdown
Contributor Author

Rebased onto main c462357 and fixed the stray closing brace that was left when the GPU fragment was removed. Head 94f424.

Assisted-by: DeepSeek

@derekross

Copy link
Copy Markdown
Owner

Thank you, @cx3301J. The GPU fragment is gone and the commit message matches the file now, which settles that point: #202's board file turns the GPU on itself.

The red check isn't yours. It is a timing case in the shell package's power-key test (omamobile-keys, "an unconnected connector's Off does not end the wait"), which this PR doesn't touch; #438, pushed in the same second on the same base, passed it. I'll re-run the job.

I built the 7a's device tree from today's published kernel and from #202's head and applied the overlay to both. On #202 it does what it should: DSIM0 with the S6E3FC5 panel, DECON0 with its pool and the boot framebuffer, and the two display clock controllers. On today's published kernel it also applies, and there it turns on DECON0 with no scanout pool and DSIM0 with no panel under it. So the points from last time still stand:

  1. The header. Please say at the top of the file that it needs a kernel with linux-omarchy-mobile#202, that it isn't in DTB_OVERLAYS and is applied by hand, and that it goes with the package's display start from device-lynx 0.2.3: persist the DSIM transmit enable #438.
  2. A short entry in device/lynx/NOTES.md: what the overlay enables, the kernel it needs, how you apply it, and your 11-16 cold boot with its kernel release (7.3.0-rc6-omamobile-push1-g61e15599a196). I compared that commit with device/tegu MODEM-PLAN, AGENTS: GNSS gets a fix on the 9a #202's current head: only the panel's LHBM unlock and one binding line differ, so the run stands for the display path. device/lynx: clarify historical Wi-Fi source and safety evidence #822 edits the same file; whichever lands second rebases.
  3. The file mode is still 755; the other overlays are 644. And please use SPDX-License-Identifier: MIT, as CONTRIBUTING §7 asks for new files in this repository.

Order of merging: #438 first (its unit does nothing until the DECON node is enabled), then #202, then this one.

@cx3301J cx3301J changed the title device/lynx: add the display and GPU overlay device/lynx: add the display overlay Nov 18, 2026
The 7a's bootloader leaves 0x1cf8 and 0x2a82306d in DSIM0 at 60 Hz;
derekross#181's Pixel 7 row is the same pair.

Assisted-by: DeepSeek
Signed-off-by: cx3301J <cx3301J@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants