Skip to content

arm64: dts: exynos: bluejay: describe the disabled PCIe Wi-Fi path - #230

Merged
derekross merged 2 commits into
derekross:omamobilefrom
latentoperator:omamobile-bluejay-wifi-dt
Nov 16, 2026
Merged

derekross merged 2 commits into
derekross:omamobilefrom
latentoperator:omamobile-bluejay-wifi-dt

Conversation

@latentoperator

@latentoperator latentoperator commented Nov 9, 2026 •

Copy link
Copy Markdown
Contributor

Describe the BCM4389 PCI identity/fallback and disabled Bluejay PCIe Wi-Fi hardware in two commits. WL_REG_ON is a disabled fixed regulator consumed through PCI power control; endpoint properties request restricted DMA, the qualified packet limit and provisioned ASCII MAC handling. The DMA pool and all new hardware nodes stay disabled.

#227 and #228 have merged. This PR is now reduced to its two own commits on current omamobile (317d87dc2), as requested. The board file retains both includes in the requested order: gs101-bluejay-input.dtsi, then gs101-bluejay-wifi.dtsi. Existing human author sign-offs and assistance trailers are retained.

Validation

  • Final reduced head: 9effb0347. The binding change is unchanged; the DT patch differs from the prior submission only in include context/order needed alongside the merged input description.
  • Full local kernel PR gate passes against the current base: two author sign-offs, policy, strict checkpatch with no errors, all ten affected Google device trees with W=1 and no new dtbs_check warnings, and dt_binding_check for the touched binding. One new-file MAINTAINERS reminder remains.
  • No C files are changed. The new head has not been booted; this rebase does not add runtime qualification. Fresh published-head CI 43782285260 passed on 9effb03478c77ebd97c851bafc0c38b2a41dcfb4.

Saved staged runtime evidence

The November 12–13 Pixel 6a run used signed integration 720a495233734e4328a09c33c095d72a8291e1c4, kernel 7.3.0-rc6-omamobile-bluejay-review-20261111, with matching Image, stage DTs and modules. This is an integration result containing the submitted changes.

The overlays enabled the disabled descriptions in order: w0 pool only; w1 pool plus WLAN regulator, PHY and PCIe with the endpoint unbound; w2 the matched endpoint policy and provisioned station firmware. The pool was 64 MiB at 0xc4000000; PCIe linked on try 1 to 14e4:4441; the driver verified restricted DMA below 4 GiB, and managed-station DHCP plus interface-bound HTTPS returned 200. The link-stage overlay enabled regulator-wlan. Submitted base nodes remain disabled.

These were attended headless stages with storage guards, a PID-1-owned 20-second watchdog, host logging and fresh SSH postchecks. They do not establish reference-device regression, suspend, endurance or installer qualification. AUTO_WRITE=no and masked automatic boot-success remain in force.

The unchanged tested integration is preserved on bringup/bluejay-reviewed-tested-20261111. The matching mobile overlay source follow-up omarchy-mobile #777 has merged. Its source comparison matches the saved trial DTs; no boot of the newly published overlay file is claimed.

No firmware, private captures, credentials, handset identifiers or calibration data are submitted.

@derekross

Copy link
Copy Markdown
Owner

Thanks, Chris — everything new is disabled and the shared dtsi only gains labels, which is what I'd hoped. This one simply waits on #227; once that's in, please rebase so only your two commits remain, and we'll let CI run the binding checks.

@latentoperator
latentoperator force-pushed the omamobile-bluejay-wifi-dt branch from a453f29 to 8b6ab24 Compare November 9, 2026 18:35
@derekross

Copy link
Copy Markdown
Owner

Thanks, Chris. This now carries #227 and #228 underneath, so it merges after both. Once they're in, please rebase so only your two commits remain, and CI will run the binding check.

@latentoperator
latentoperator force-pushed the omamobile-bluejay-wifi-dt branch from 8b6ab24 to 4cb0149 Compare November 11, 2026 18:48
@latentoperator

Copy link
Copy Markdown
Contributor Author

November 13 staged DT runtime result:

Tested signed integration: 720a495233734e4328a09c33c095d72a8291e1c4, based on fc7086a68fcaaed8960d09f1bdf6c51c9926c759; uname -r = 7.3.0-rc6-omamobile-bluejay-review-20261111, with its matching Image, stage DTs and modules. This is an integration result containing the submitted changes, rather than a boot of each PR in isolation.

The staged overlays enabled the submitted disabled descriptions in order: w0 pool only; w1 pool plus WLAN regulator, PHY and PCIe with the endpoint unbound; w2 the matched endpoint policy and provisioned station firmware. The pool was 64 MiB at 0xc4000000; PCIe linked on try 1 to 14e4:4441; the driver verified restricted DMA below 4 GiB, and managed-station DHCP plus interface-bound HTTPS returned 200. The link-stage overlay explicitly enabled regulator-wlan as required. Submitted base nodes remain disabled.

The merge order is unchanged: #227 and #228 first, then rebase this PR to its two own commits and rerun CI. No dependency rebase or source change was made for this report. These were attended headless stages with storage guards, a PID-1-owned 20-second watchdog, host logging and fresh SSH postchecks. They do not qualify the revised desktop/GPU path, native panel modesetting, suspend, endurance or a release installer. The historical graphical desktop has been restored. No kernel commit or DCO metadata changed; AUTO_WRITE=no and masked automatic boot-success remain in force.

@derekross

Copy link
Copy Markdown
Owner

Thank you, Chris. The staged run (pool, then the link, then the station) is the order I'd want, and everything new in this tree is disabled. The two labels in gs101-pixel-common.dtsi change nothing for the Pixel 6 and 6 Pro: their DTBs differ only by the new symbols. dt-validate passes on gs101-bluejay.dtb with this head's schemas, and the base's schemas reject the new nodes as unknown, so the binding commits are doing their job. The plan stands: #227 and #228 first, then a rebase to your two commits and a fresh CI run.

One thing for the omarchy-mobile side: main's device/bluejay/overlays/wifi-controller.dtso enables the PHY and the controller but not regulator-wlan. Your description says the link overlay must enable it, and your report says the one you ran did. Could you open a small omarchy-mobile PR with the overlays exactly as they ran? Then the next run comes from the tree. And, as on the others, please push 720a4952… so the run can be tied to these heads.

@latentoperator

Copy link
Copy Markdown
Contributor Author

The mobile overlay follow-up #777 has merged, and the unchanged tested integration 720a4952 is published on the fork. That closes the overlay/source-publication requests.

This PR remains draft until #227 and #228 land. Then I will rebase it to its two own commits, retain gs101-bluejay-input.dtsi before gs101-bluejay-wifi.dtsi as requested on #229, and rerun the binding/device-tree gates and CI. I have not rewritten the dependency stack or claimed the newly published overlay file itself was booted.

Allow the BCM4389 PCI identity together with the existing fullmac
fallback. Other compatible forms retain their existing schema.

Assisted-by: LLM
Signed-off-by: Chris Gray <chrisgray70@gmail.com>
Describe the GS101 PCIe1/PHY path, restricted DMA pool and BCM4389
endpoint. All new hardware and the WL_REG_ON regulator remain disabled.
PCI power control consumes the root-port regulator and its 200 ms delay;
the host no longer owns an endpoint enable GPIO. The endpoint requests
the confined DMA path, packet limit and provisioned ASCII MAC explicitly.

Depends on the GS101 PCIe/PHY and brcmfmac policy topics. The current
source is not the exact tested development stack and has not been
booted. A future gated overlay must also enable the WL_REG_ON regulator.

Assisted-by: LLM
Signed-off-by: Chris Gray <chrisgray70@gmail.com>
@latentoperator
latentoperator force-pushed the omamobile-bluejay-wifi-dt branch from 4cb0149 to 9effb03 Compare November 16, 2026 00:45
@latentoperator
latentoperator marked this pull request as ready for review November 16, 2026 00:51
@derekross

Copy link
Copy Markdown
Owner

Thank you, Chris. Everything I asked for is done: #227 and #228 are in, the branch is down to your two commits, CI is green on this head (with dtbs_check and dt_binding_check), #777 has the overlays as they ran, and 720a4952 is on your fork. I fetched it and compared. gs101-bluejay-wifi.dtsi, the binding and gs101-pixel-common.dtsi are the same blobs as in the tested integration, and only the include order differs in gs101-bluejay.dts. The bluejay DTB from this head on today's omamobile decompiles identical to the one built from 720a4952, so your staged runs stand for exactly this tree. The Pixel 6 and 6 Pro DTBs only gain the two labels (and the phandles that come with them), and the 7 Pro's is unchanged.

One note for the record: the only line that is live on bluejay before any overlay is &soc's wider ranges. That's the same 2 GiB window gs201 and zuma use, and nothing enabled sits in it, so I'm happy with it.

If you'd like to, you could drop the "Depends on …" and "has not been booted" sentences from the DTS commit message before I merge, since both are out of date now. Otherwise I'll merge it as it is.

@derekross
derekross merged commit ab5b6b1 into derekross:omamobile Nov 16, 2026
1 check passed
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