Skip to content

drm/asahi: admit the 26B123 image a second 40-core J616C loads - #4

Open
Rex50527 wants to merge 3 commits into
astaff:t6041/j616c-fw27-26b131from
Rex50527:t6041/j616c-26b123-40core-image
Open

Rex50527 wants to merge 3 commits into
astaff:t6041/j616c-fw27-26b131from
Rex50527:t6041/j616c-26b123-40core-image

Conversation

@Rex50527

Copy link
Copy Markdown

For aurora-silicon#1954: one commit on top of 7e4fac710431 that lists the 26B123 GPU image a second 40-core J616c loads. Without it, aurora-silicon#1954 boots and the DCPs modeset on this Mac, but the GPU firmware is refused, so there is no render node and the screens stay black under a compositor.

It changes only BUILD_26B123.images in drivers/gpu/drm/asahi/hw/t6041.rs: one more FirmwareBuild with the same sizes and entropy offsets as yours and no boot-filled ranges. The digest repeated across boots, and the GPU comes up with the unchanged V26B123 variant and BASE config. Details are in the commit message.

One thing I saw that this commit doesn't touch: about 4.5 minutes after boot the driver logged

aurora 28a600000.gpu: G17 Primary Statistics ring stopped: EIO

and quarantined the statistics ring. Rendering carried on, and nothing else from the GPU appeared in the kernel log. I haven't traced which check in consume_report failed. It may be a 26B123 report-format difference, so I'm mentioning it in case you see it too.

🤖 Generated with Claude Code

astaff and others added 3 commits November 18, 2026 11:56
Add support for the T6041 (M4 Max, 40-core) GPU firmware shipped with
macOS 27.2 (26B123, iBoot mBoot-20457.41.3):

- Add the g17::fw::initdata::v26b123 wire layouts for the updated
  interface version, 60-slot register descriptor table and separate
  GPTBAT MMIO mapping, 0x2850-byte hardware-data object, 0x6c48-byte
  cold power configuration, cold power-state limiter table, 19-row AON
  sensor table, shifted region-C prefix (+4 bytes past +0x38), and
  relocated main-configuration power pointer (+0x481).
- Upgrade the T6041 25G83/25F84 runtime graph into distinct retained
  allocations when Variant::v26b123 is active while keeping all 25G83
  and 25F84 paths unchanged.
- Emit opcode 0x13 without incrementing the host power-assert tally for
  scheduler-state release on 26B123 (matching M5), as 0x14 is repurposed
  in 26B123.
- Register the normalized T6041 26B123 loaded-image identity in
  hw::t6041::BUILDS and cover it in the fwid and event-record unit tests.

Signed-off-by: Artyom Astafurov <artyom.astafurov@gmail.com>
…27.2

This change enables internal and external displays on J616c (MacBook Pro
16-inch, M4 Max) with macOS 27.2 (26B123).

- Add apple_t6041_j616c_dcp_v27_2() and apple_t6041_j616c_dcpext_v27_2()
  pinned to UUID 964EDE5B-8F78-38C2-A727-519CC9F0AF27.
- Enable the internal display chain and external HDMI and USB-C chains
  in display-gate on J616c 27.2.
- Keep external display chains off on unapproved firmware-27 builds.
- Select DCP_FIRMWARE_V_27_0 for J616c 27.2 internal and external DCPs
  after checking the v27 method table.
- Use 41-bit firmware DMA, live-session adoption, two DPTX ports, and
  sink metadata EDID on T6041 v27 dcpext.
- Build the two-entry disp0 clock table when apple,clock-ids is absent,
  and allow zero clocks on T6041 dcpext.
- Enable desktop Type-C route power, PHY handling, HDMI rail delay, and
  crossbar link-up on T6041 v27 dcpext.

Existing 26.5.2 and 26.6.2 desktop display support stays unchanged. This
change adds display-gate KUnit and apple-dcp selftests.

Signed-off-by: Artyom Astafurov <artyom.astafurov@gmail.com>
iBoot fills part of the loaded GPU firmware TEXT for each Mac, so every
admitted T6041 image is a measurement. A MacBook Pro 16-inch M4 Max with
the 40-core GPU (J616C) on the macOS 27.2 (26B123) stub loads the same
26B123 build as the image listed (TEXT 0x64000, DATA 0xcc000, entropy
0x5d88c/0x5d96f), but its image does not match it. The driver refuses
the firmware and never initializes the GPU, so no render node is created
and compositors have no renderer:

  aurora 28a600000.gpu: unsupported GPU firmware build: TEXT 0x64000
    DATA 0xcc000 entropy 0x5d88c/0x5d96f sha256 df0338a09b6c8dfd...
  aurora 28a600000.gpu: Unsupported GPU firmware image (reported
    macOS version Some([27, 2, 0]))

Measured as the listed image is, with only the entropy zeroed, this
image is

  df0338a09b6c8dfdb48fcb05b8ebbc8059581a46c925584c6024c9279459fc5d

on every boot measured: the driver logged this digest on one boot
without this change and admitted it on the next with it. On 25G83 the
same Mac's image needed the header load field and the record block
zeroed as well; on 26B123 it does not.

List it next to the existing image. It runs with the build's V26B123
variant and the BASE configuration, unchanged.

Tested on a J616C with the 40-core GPU on 7.3.0 from 54e846c (this
branch's previous head; drivers/gpu/drm/asahi is the same as here) plus
this change and a local entry for this Mac's 25G83 image, which 26B123
boots do not use:
the image is admitted ("Apple GFX 26B123 for T6041 (normalized loaded
image, 40-core J616C)"), the driver initializes the GPU and creates
renderD128, and Hyprland renders on the GPU on the internal panel.

Assisted-by: Claude:claude-opus-5-5
Signed-off-by: Rex50527 <48970558+Rex50527@users.noreply.github.com>
@astaff
astaff force-pushed the t6041/j616c-fw27-26b131 branch from 7e4fac7 to e538e57 Compare November 19, 2026 17:53
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