Skip to content

PCI: exynos: add staged GS101 Wi-Fi controller and PHY support - #227

Merged
derekross merged 3 commits into
derekross:omamobilefrom
latentoperator:omamobile-gs101-pcie-bluejay
Nov 15, 2026
Merged

derekross merged 3 commits into
derekross:omamobilefrom
latentoperator:omamobile-gs101-pcie-bluejay

Conversation

@latentoperator

@latentoperator latentoperator commented Nov 9, 2026 •

Copy link
Copy Markdown
Contributor

Add GS101 controller and PHY behavior through per-compatible match data, keeping the existing GS201 defaults. Endpoint power is supplied through PCI power control and the DT regulator rather than a host-driver enable GPIO.

The GS101 match data now explains its single link attempt: retry power cycles and failure-path link-history reads are not qualified on this controller. This suppresses the final failure dump; it does not claim the existing history-control setup writes are disabled. The PHY retains the imported GPL calibration attribution and the controller's analog-supply handling.

The requested revised-stack Pixel 6a run has passed in the signed integration (runtime section below). The exact tested integration is published below; this is ready for maintainer review.

Runtime update — November 12–13, 2026

November 13 fresh Pixel 6a 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.

  • w0 first verified the enabled 64 MiB restricted pool at 0xc4000000, with the other seven experimental hardware nodes disabled.
  • On a separate w1 boot, enabled the WLAN regulator, PHY and PCIe path while leaving touch and the Wi-Fi endpoint driver disabled. The revised PHY and controller modules were loaded once under logging/watchdog gates; both bound, the link came up on try 1, and PCI enumerated Broadcom 14e4:4441. The endpoint stayed unbound and no Wi-Fi firmware was loaded.
  • A fresh SSH postcheck passed. The subsequent w2 boot independently repeated the PCIe proof before the station trial described in wifi: brcmfmac: confine the Bluejay BCM4389 bring-up #228.

This supplies the requested fresh 6a run of the revised controller/PHY path. The exact tested integration is available for review; no reference-device regression or failure/retry-power-cycle claim is inferred. 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.

Prior source/build validation

Validation on current upstream base fc7086a: independent source review found no remaining concrete blocker. The common integration passes policy, strict checkpatch without errors, W=1 builds of seven touched C files, ten Google device trees with no new dtbs_check warnings and four binding checks. New-file coverage warnings and one continuation-alignment check are disclosed. A full Image, Bluejay DTB and 1,828 modules also build for 7.3.0-rc6-omamobile-bluejay-review-20261111; matching dependency indexes and module resolution pass. The full build reports two format warnings in unchanged AoC ALSA code.

The human certified the final patches under DCO on November 11. Independent checks verify every author sign-off, all 22 commit objects including integration copies, and identical source trees and exact diffs before/after certification. Signing changed metadata only; the frozen source-build evidence remains applicable. Fresh published-head CI passed on f90429400883782ae50e47905ebf56dff3a954a2 (sign-offs, checkpatch, W=1 and device-tree/binding gates as applicable).

The original tested source c934e31 and known-good phone setup remain preserved. AUTO_WRITE=no, the automatic boot-success mask and the GS101 slot-marking hold remain in force. No firmware, private captures, credentials, device identifiers or calibration data are submitted. Development/source status is documented in Bluejay UPSTREAM, with historical bring-up isolated in mobile #565.

Published tested integration — November 14, 2026

The exact signed integration cited in the November 12–13 runtime report is now available on the fork's bringup/bluejay-reviewed-tested-20261111 branch, at 720a495233734e4328a09c33c095d72a8291e1c4. The commit, source tree and existing human sign-offs are unchanged; this publishes the source for inspection and adds no new runtime qualification.

@derekross

Copy link
Copy Markdown
Owner

Thank you, Chris — the gs101 variant is cleanly gated and I checked that the Pixel 7 Pro's path is unchanged. Before it leaves draft: could the endpoint power go through pci-pwrctrl as on gs201, rather than an enable GPIO in the host driver? And since #178 extends the same driver and binding for the 9a, per-compatible match data instead of a bool would let both land without fighting; I'll help sort the order.

@latentoperator
latentoperator force-pushed the omamobile-gs101-pcie-bluejay branch from 0afd3ac to 2e83129 Compare November 9, 2026 18:35
@derekross

Copy link
Copy Markdown
Owner

Thank you, Chris. Match data and pwrctrl are exactly the shape I hoped for, and the Pixel 7 Pro's controller keeps its defaults. Could you add a sentence on why gs101 gets a single link try and skips the history dump? I'd also like one boot of this rebased version on your 6a before it leaves draft.

Add the GS101 compatible and its two analog supplies without changing
the current zuma/zumapro alternatives. Forbid the unqualified coherent
DMA and CMU diagnostic path on GS101. Endpoint enable GPIOs belong to
the regulator and PCI power-control description, not the host binding.

Assisted-by: LLM
Signed-off-by: Chris Gray <chrisgray70@gmail.com>
Use the ordered GS101 CAL writes and its PCS, timing and power-down
sequence. Leave the existing GS201 path selected on its compatible.
The earlier development path ran on Bluejay; the current rebased source
has not been booted. Keep the TCXO diagnostic unsupported for GS101.

Link: https://android.googlesource.com/kernel/google-modules/soc/gs/+/204078758737776959ecc85c527a00fd2cabb62f/drivers/pci/controller/dwc-whi/pcie-exynos-gs101-rc-cal.c

Assisted-by: LLM
Signed-off-by: Chris Gray <chrisgray70@gmail.com>
Extend the current SoC match data for GS101 supply votes, inaccessible
CMU diagnostics and bounded link attempts. The already merged 9a
match-data paths are retained. Endpoint power uses the existing PCI
power-control sequence, including its error and teardown handling;
remove the proposed direct WLAN GPIO ownership from the host driver.

The companion disabled Bluejay DT supplies a fixed WL_REG_ON regulator
to the root port. That reworked sequencing is not hardware-qualified.

Assisted-by: LLM
Signed-off-by: Chris Gray <chrisgray70@gmail.com>
@latentoperator
latentoperator force-pushed the omamobile-gs101-pcie-bluejay branch from 2e83129 to f904294 Compare November 11, 2026 18:48
@latentoperator

Copy link
Copy Markdown
Contributor Author

November 13 fresh Pixel 6a 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.

  • w0 first verified the enabled 64 MiB restricted pool at 0xc4000000, with the other seven experimental hardware nodes disabled.
  • On a separate w1 boot, enabled the WLAN regulator, PHY and PCIe path while leaving touch and the Wi-Fi endpoint driver disabled. The revised PHY and controller modules were loaded once under logging/watchdog gates; both bound, the link came up on try 1, and PCI enumerated Broadcom 14e4:4441. The endpoint stayed unbound and no Wi-Fi firmware was loaded.
  • A fresh SSH postcheck passed. The subsequent w2 boot independently repeated the PCIe proof before the station trial described in wifi: brcmfmac: confine the Bluejay BCM4389 bring-up #228.

This supplies the requested fresh 6a run of the revised controller/PHY path. Leaving draft for your review of the evidence; no reference-device regression or failure/retry-power-cycle claim is inferred. 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 match data, the comment on the single link try, and pwrctrl for the endpoint (in #230's tree) are all as I'd hoped. Our gate passes on today's omamobile (2db2bc3), and CI is green. I also checked the Pixel 7 Pro side two ways. By reading, every new field is zero in gs201_pcie_data and the PHY's gs101 path needs google,gs101-pcie-phy. And I built pcie-gs201.o and phy-gs201-pcie.o before and after and compared them function by function: only the functions that contain or inline the new checks change. The gs101 PMU offset (0x3ec4, PHY_CTRL_PCIE_GEN4_1) is in gs101's PMU access table, and the PHY takes the SMC regmap, so that write lands.

One request before it leaves draft. Your 11-13 run is from 720a495233734e4328a09c33c095d72a8291e1c4, which isn't on GitHub (a fetch says "not our ref"). Could you push it to a branch on your fork? Then I can match its files to this head by blob, and the 6a run counts for this PR. After that this is ready, and it can land before or after #228, since they don't touch each other's files.

@latentoperator

Copy link
Copy Markdown
Contributor Author

Closing the source-publication request: the exact tested signed integration is available on bringup/bluejay-reviewed-tested-20261111 at 720a495233734e4328a09c33c095d72a8291e1c4. The commit, source tree and existing human sign-offs are unchanged. The PR description carries the same link; this is a pointer to the previously reported 6a run, not a new per-PR boot claim.

@derekross

Copy link
Copy Markdown
Owner

Thank you, Chris. With 720a4952 on your fork I could tie the 6a run to this PR. All four files this PR touches (both bindings, pcie-gs201.c and phy-gs201-pcie.c) have the same blobs there as at this head, so your w1 link run is evidence for exactly this code. Our gate passes on today's omamobile (d49d68c), and none of these files has changed on omamobile since CI ran, so the Pixel 7 Pro compare from my last review still holds.

I'll merge it. Your drafts #254 and #255 touch the same parts of pcie-gs201.c (the includes, the match-data struct and the match table). Whichever lands second only needs both sides kept.

@derekross
derekross merged commit 03f52a4 into derekross:omamobile Nov 15, 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