Skip to content

T8142 DisplayPort: J813 fixes (crossbar config, USB without the DP banks, recovery fixture) - #2

Merged
m0d16l14n1 merged 3 commits into
m0d16l14n1:j815/t8142-dp-altmode-wipfrom
sithglan:thomas/t8142-dp-usb-fallback
Nov 16, 2026
Merged

m0d16l14n1 merged 3 commits into
m0d16l14n1:j815/t8142-dp-altmode-wipfrom
sithglan:thomas/t8142-dp-usb-fallback

Conversation

@sithglan

Copy link
Copy Markdown

Three fixes on top of j815/t8142-dp-altmode-wip (aurora-silicon#1806, head e4bdeb8), found while running it on a J813 under the macOS 27.2 stub, with Linux at EL2.

J813 result

The external display works on a J813 (Mac17,3), macOS 27.2 (26B123) stub, Linux at EL2:

  • The kernel is the PR's 9 display commits on our J813 tree, with the first two fixes below.
  • The loader is aura with 608e17b8, 91e2ffbe and b97ef3a8 from your aura branch, enrolled.
  • Both /chosen/aurora,dcpext{0,1}-handoff read okay, and both processors probe.
  • An EIZO EV2780 behind a dock on the rear port (dock in DP Alt Mode, pin assignment D, its USB3 hub at 10000M) is card1-DP-2 at 2560x1440@59.95, next to eDP-1, under sway. The dock's USB keeps running.
  • DP audio is not in that tree; I have not tried it.

What this adds

  1. selftests: apple-dcp: give the recovery fixture the per-profile audio endpoint.
    • The DP audio commit makes afk_compact_capable() call dcp_av_audio_endpoint().
    • The retained recovery control actual-queue-quiesce compiles that function without the helper, so on e4bdeb8 it fails to build and apple-dcp's service-rebind fails.
    • The fixture now gets hw.av_audio_endpoint and the helper, as it already does for the M4 check.
  2. phy: apple: atc: T8142: Serve a DP Alt Mode partner's USB without the DisplayPort calibration.
    • Without the loader's LIOA and common-config banks ("DisplayPort bank calibration missing; DP disabled"), the mode switch refuses every DP mode.
    • A DP Alt Mode partner then loses its USB too. Under a loader without the bank export, the dock's port got phy init failed --> -95, and 40a280000.usb never probed. The kernel before the T8142 DisplayPort sequence ran that dock's hub at 10000M.
    • The DP modes now map to the USB the partner keeps, as T8132 already does: USB2 next to four lanes, USB3 next to two.
  3. arm64: configs: j813: Build the display crossbar.
    • j813.config has no MUX_APPLE_DPXBAR.
    • Without it, the apple,t8142-display-crossbar nodes never bind. Both dcpext then wait until the deferred probe timeout and end "external display processor unusable (probe failed: -110)", although the loader handed them over.
    • This was the last thing missing on our J813.

Testing

  • tools/asahi/run-host-tests.py: 323/323 on this branch (on e4bdeb8 apple-dcp fails).
  • atc-usb3-tests has new cases for the T8142 mapping, with and without the banks. On the PHY code without fix 2, they fail.
  • allnoconfig + j813.config resolves to CONFIG_MUX_APPLE_DPXBAR=y. The Image and the J813 DTB build (LLVM 22). atc.c builds with W=1 and no warnings. checkpatch --strict is clean.
  • Fixes 2 and 3 come from the J813 runs above: the -110 without the crossbar, and the lost dock USB under the old loader. The display result above runs with fix 3. Fix 2 changes nothing when the banks are present, and I have not yet booted the J813 under the old loader with it.

cc @Acelogic @DjDeveloperr @RyanTheTide @LiamRay10

… endpoint

"drm/apple: T8142 dcpext: take the audio interface on dcpav-interface-ep"
makes afk_compact_capable() compare the endpoint with
dcp_av_audio_endpoint(). The retained recovery control
actual-queue-quiesce compiles that function without the helper, so it
no longer built and service-rebind failed.

When afk_compact_capable() uses it, give the fixture hw.av_audio_endpoint
and the helper from dcp-internal.h, as it already does for the M4 check.

Signed-off-by: Thomas Glanzmann <thomas@glanzmann.de>
… DisplayPort calibration

On T8142 the mode switch refuses every mode with DP AUX while the boot
loader's LIOA and common configuration banks are missing ("DisplayPort
bank calibration missing; DP disabled"). A partner in DP Alt Mode then
loses its USB as well: on a J813 with a dock in pin assignment D, under
a loader without the bank export, the PHY init fails with -EOPNOTSUPP
and the port's dwc3 never probes, where the kernel before the T8142
DisplayPort sequence ran the dock's hub at 10 Gb/s.

Map the DP modes to the USB the partner keeps, as T8132 does: USB2 next
to four DP lanes (pin assignments C and E), USB3 next to two (D).
atc-usb3-tests covers the mapping with and without the banks.

Signed-off-by: Thomas Glanzmann <thomas@glanzmann.de>
Every external display route of the T8142 Airs goes through the display
crossbars (apple,t8142-display-crossbar). Without their driver both
dcpext processors wait for the crossbars until the deferred probe
timeout and end as unusable pipes ("external display processor unusable
(probe failed: -110)"), so the j813.config kernel shows no external
display although the boot loader hands both processors over.

Signed-off-by: Thomas Glanzmann <thomas@glanzmann.de>
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