Skip to content

j514s (MacBook Pro 14" M3 Pro): M3 handoff works; Dell U4919DW over USB-C gets no DP lanes on any port, HDMI works #455

Open

Description

@desaperados

Machine: apple,j514s / apple,t6030, MacBook Pro (14-inch, M3 Pro, Nov 2023)
Firmware: asahi,os-fw-version = 14.7, asahi,system-fw-version = 26.5.2 (iboot1 mBoot-18000.121.3, iboot2 iBoot-10151.140.19.700.2, stub 14.8.3 / 23J220)
Build: aurora-2026.10.17.20, linux-aurora 7.1.12.aurora2-20261017.20, m1n1-aurora 1.6.1.aurora17-5, mesa-m3 26.1.4.m3.2-5, mesa-m3-native-pro 26.1.4.nativepro.20261013-10
m1n1 guard: did not fire. No m1n1-oslog-overlap node, three dcp-oslog regions
Touch ID: not supported on M3 (expected)
Thunderbolt: no dock
M3 path: handoff with --m3-handoff, m3-mode = handoff t6030-12
Display: Dell U4919DW (5120x1440@60 ultrawide, USB-C upstream with DP alt mode, a built-in USB hub and power delivery)

This Mac was installed with the Asahi installer plus omarchy-mac-setup, so stage 1 is clean v1.6.1. This is the same situation as #47, #138 and #360, and PR #361 would cover it. I made three local edits to this release's installer, disclosed here:

  1. M3_STAGE1_VERSIONS="v1.6.1-dirty v1.6.1", the allowlist edit from M3 Pro clean stage1: qualify explicit handoff admission on the supported stub #47.
  2. In package_database_check(), skip hyprland|hyprtoolkit|hyprland-guiutils, as in j514s (MacBook Pro 14" M3 Pro): M3 handoff works #138.
  3. In package_database_check install, skip rows pacman already marks [ignored]. Here that was brave-bin: a newer version exists only in the omarchy-mac [omarchy] repo, which is Usage = Sync, so omarchy update never clears it and the installer always stopped with "1 package upgrade(s) pending".

What worked

  • The m1n1 handoff boots to GRUB, then the desktop on the built-in panel: eDP-1 3024x1964@120 (74 notch rows shown, from Omarchy's appledrm show_notch=1).
  • GPU: mesa-m3-session.state gives decision=active reason=active, with render node /dev/dri/renderD128 and display card2.
  • HDMI works: the Dell U4919DW comes up on HDMI-A-1 at 5120x1440@59.98 through dcpext0 (HDMI: forwarding converter service requests (HBR3 link)). The DCP logs Realtime Bandwidth setup failed, vid-clock-to-disp-clock-factor read failed and Modeset done, but pipe not enabled, but the picture is fine.
  • USB-C power delivery from the Dell charges the Mac, and the Dell's USB hub enumerates (at USB 2.0 only, see below).

What did not

The Dell U4919DW over USB-C gets no picture on any of the three USB-C ports (atc0, atc1 and atc2). The same cable and monitor work over USB-C from an M1 Air on Asahi. Every attempt follows the same pattern on dcpext1 (route=borrowed): the session opens, the link configuration times out after about 2 s, and after five tries:

apple-dcp 2d6c00000.dcp: allocated Type-C DPTX PHY 0
apple-dcp 2d6c00000.dcp: dcp_dptx_connect_session(port=0) die=0 atc=0 dfp_port=0 tunnel=0 typec=1 route=borrowed conn_type=20 connected=0
apple-dcp 2d6c00000.dcp: DPTXPort: GET_SUPPORTS_HPD 1 usb4=0
apple-dcp 2d6c00000.dcp: dcp_dptx_connect: timed out waiting for port 0 link configuration
apple-dcp 2d6c00000.dcp: DPTXPort: SET_LINK_RATE 0x0
apple-dcp 2d6c00000.dcp: DPTXPort: DEACTIVATE
...
apple-dcp 2d6c00000.dcp: DPTXPort: INACTIVE_SINK_DETECTED (keep waiting for lanes)
apple-dcp 2d6c00000.dcp: Type-C DPTX reconnect failed after 5 retries: -110

At boot (cable attached on atc2), firmware also logs nifiedPipeline.cpp:8189: HPD not present, skipping mode export. The same connected=0 / link-configuration timeout appears for atc=2, atc=1 and atc=0.

Other observations that may help:

  • The Dell's hub (TI 0451:8442, plus 0451:8242 and 0451:82ff HID) only enumerates on the USB 2.0 bus. Nothing appears at SuperSpeed.
  • /sys/class/typec/portN-partner/identity/* are all 0x00000000, usb_power_delivery_revision is 0.0, and the partner has no alternate-mode entries. Each port reports port*.0 svid=ff01 vdo=0x00000c02 active=yes.
  • The Dell's OSD hides "USB-C Prioritization" (2-lane vs 4-lane DP) unless it already has a signal, so I couldn't test the other lane split from the M3.

#138 had an LG UltraFine 4K working on atc0 of a J514S, so this may be specific to a monitor with a hub and PD (lane split, pin assignment D vs C?), rather than to the port.

Not tested

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions