Skip to content

omamobile-waydroid 0.1.4, docs: Waydroid on the OnePlus 7T Pro's Adreno - #935

Open
johannSo wants to merge 1 commit into
derekross:mainfrom
johannSo:hotdog-waydroid
Open

johannSo wants to merge 1 commit into
derekross:mainfrom
johannSo:hotdog-waydroid

Conversation

@johannSo

Copy link
Copy Markdown

What and why

omamobile-waydroid 0.1.4, plus a how-to, docs/ANDROID_APPS.md.

The check accepted only a panthor render node. On the OnePlus 7T Pro (#931) the render node is renderD128 on msm_dpu, with an Adreno 640 bound and freedreno in Waydroid's vendor image. So the check reported no GPU, and the container wouldn't start.

Pixels unchanged:

  • The check still looks for panthor on every render node first, with the same output lines.
  • Only where no node is panthor does it take the first render node (the one Waydroid binds). It accepts msm or msm_dpu when an Adreno is bound, then checks the node's numbers against lxc-omamobile.conf as it does for panthor.
  • omamobile-waydroid mesa (and init, which calls it) skips the Mali-G710 patch only on msm. Every other phone is patched as before.

docs/ANDROID_APPS.md covers running Android apps in Waydroid on each phone, from install to troubleshooting.

Device and build

  • Phone: OnePlus 7T Pro (hotdog)
  • Kernel release (uname -r): 7.3.0-rc6-omamobile-push1-g7539ca3e06f6
  • Packages changed, with versions: omamobile-waydroid 0.1.4

Checks (no phone needed)

  • tools/pr-check origin/main passes
  • tests/packages: omamobile-waydroid test/all passes
  • Version bumped and in the commit subject
  • Commits have area prefixes and bodies that say why; AI help is a Co-Authored-By: trailer

On the phone

  • I ran the port's build of this check on my own phone (2026-11-05, recorded in the doc):
    • the check printed "renderD128 is msm_dpu with an Adreno bound";
    • Android boots on the Adreno and reaches the internet through ufw;
    • it can't reach block devices or debugfs;
    • F-Droid installs and opens from the launcher.
  • What I did not test:
    • this version's reordering, which keeps the panthor-first search, on a phone;
    • any Pixel: the doc says "not yet tried on the Pixel 7 Pro".
  • Nothing in this change writes outside slot B and userdata, weakens the storage guard, or reads registers from userspace

🤖 Generated with Claude Code

The check accepted only a panthor render node, so on the OnePlus 7T Pro
(renderD128 on msm_dpu, an Adreno 640 bound, freedreno in Waydroid's
vendor image) it reported no GPU and the container would not start.

The check still looks for panthor on every render node first, exactly as
before, with the same lines. Only where none is panthor, it takes the
first render node, which Waydroid binds, and accepts msm or msm_dpu when
an Adreno is bound; its numbers are then checked against
lxc-omamobile.conf as panthor's are. "omamobile-waydroid mesa" (and init,
which runs it) skips the Mali-G710 patch only on msm; every other phone
is patched as before.

docs/ANDROID_APPS.md: how to run Android apps in Waydroid, per phone,
from the install to troubleshooting. Its status is the OnePlus 7T Pro's
run of 2026-11-05 with the port's build of this check (Android boots on
the Adreno, reaches the internet through ufw, can't reach block devices
or debugfs; F-Droid installs and opens from the launcher); not yet tried
on the Pixel 7 Pro.

Checked read-only on the 7T Pro with the port's check: "renderD128 is
msm_dpu with an Adreno bound". test/all passes.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.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.

1 participant