Skip to content

drm/asahi: g15: M3 Pro: pin the cd05 and 4118 firmware TEXT, and decline early probes with -ENODEV - #2146

Open
scottjones wants to merge 5 commits into
aurora-silicon:integration/aurorafrom
scottjones:upstream/m3pro-g15-firmware-text
Open

scottjones wants to merge 5 commits into
aurora-silicon:integration/aurorafrom
scottjones:upstream/m3pro-g15-firmware-text

Conversation

@scottjones

Copy link
Copy Markdown
Contributor

Summary

Two fixes for M3 Pro (T6030) Macs.

J514S and J516S boot chains on 14.7 can load T6030 firmware whose TEXT digest is not in the admission table, so probe refuses it and there is no render node. This pins the cd05 digest (J514S, J516S) and the 4118 digest (J514S, J516S), each tied to its board, and refreshes the golden pins.

Before activation the G15 probe returned -EPROBE_DEFER. If the real probe later failed for good, the GPU stayed on the deferred-probe list and anything waiting on that list kept waiting. It now returns -ENODEV; activate() already reprobes unbound devices.

 g15 probe, before MODULE_READY
-  return -EPROBE_DEFER
+  return -ENODEV   (activate() reprobes)

Evidence

  • Built on integration/aurora a7dd26e6d60e with the 7.3 package config, LLVM 18 and Rust 1.85:
    • make W=1 drivers/gpu/drm/asahi/: builds
    • W=1: no warnings the base does not have
    • run-host-tests.py: every tools/asahi test is registered
    • tools/asahi host tests (control and mutation runs): g15-maintained-handoff-tests 2/2, g15-probe-init-tests 2/2, g15-zero-words-tests 2/2
    • Stacked with the other 12 PRs on the same base, the full kernel, modules and DTBs build and package (linux-aurora 7.3.0-0testastip2-1-ARCH, 137 DTBs).
  • Booted as part of the full set: a kernel built from integration/aurora a7dd26e6 plus all 13 branches of this set (before the three small fix-up commits) booted on an M3 MacBook Air (J613), a Neo (J700) and an M1 Pro with no regressions against their baselines (on the M1 Pro, only the checks already known to differ: GPU name and DRAM OPP), and a 25-minute GPU soak was clean on all three. No M3 Pro, M3 Max, J504, J615 or M4 Pro ran this kernel.
  • The cd05 and 4118 digests come from owner reports in G15: T6030 (M3 Pro) GPU firmware TEXT varies between Macs on the same boot chain; two J514S builds are refused omacom/linux-aurora#247 (issue); with the prefix override those Macs reached firmware READY with a render node. The -ENODEV change is drm/asahi: g15: decline probes before activation with -ENODEV omacom/linux-aurora#470.

Merge Danger

Door: two-way

Blast Radius: M3 Pro

The firmware rows only admit the named digests on the named boards. The probe change only affects the G15 path before activation.

🤖 Generated with Claude Code

javanoo6 and others added 5 commits November 17, 2026 06:22
The J514S 14.7 boot chain can load a T6030 firmware build whose full TEXT
digest is not in the admission table. Probe refuses it and no GPU render
node is created.

Desktop use with the explicit prefix override reaches firmware READY and
reduces Hyprland CPU use from several cores to single-digit percentages.
Pin the complete measured digest only for J514S, with its segment sizes
and marker, rather than requiring a persistent bring-up override.

Exercise the production firmware table and exact lookup with the override
disabled. Reject other boards, altered digest suffixes, sizes and markers.

Link: omacom/linux-aurora#247
Signed-off-by: Nikita Konkov <95355711+javanoo6@users.noreply.github.com>
Signed-off-by: Scott Jones <scottajones@gmail.com>
The J514S cd05 FirmwareBuild row adds 15 lines to the T6030 firmware
table, which changes the pinned hw.rs file and t6030 module digests.
With that row removed both digests are their previous values, so the
T6030 profile itself is unchanged. Refresh the two pins and record why.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: Scott Jones <scottajones@gmail.com>
Some 16-inch M3 Pro (J516S) units on the 14.7 boot chain load a B0 GPU
firmware image whose TEXT digest differs from the pinned J516S row:
TEXT 0x5c000, DATA 0x114000, GKTS marker 0x58e88 and TEXT sha256
cd05af6a828b52e4c2fed20086d3d18a7db42e03a8fd6f8354f9f06e0a61363e. The
build is refused with "loaded GFX TEXT identity mismatch", so there is
no render node and the compositor falls back to llvmpipe. The GPU only
starts with the g15_accept_text_prefix bring-up override.

Add this build as a J516S-only row. Rows still pin on the machine they
name; nothing changes on the J514S or on other machines. Extend the
maintained-handoff host test to check the new row and refresh the
zero-words golden pins for hw.rs and the T6030 module, which change by
exactly this one row.

Measured on a J516S (apple,j516s apple,t6030; os-fw 14.7, system-fw
15.5, iBoot2 iBoot-10151.140.19.700.2) with linux-aurora
7.2.11.aurora14-1 and the override: the admission line reports the
identical sizes, marker and digest, and the GPU reaches READY with
renderD128 present.

Link: omacom/linux-aurora#247
Link: omacom/omarchy#17739
Signed-off-by: joel cristobal <joelcristobal97@gmail.com>
Signed-off-by: Scott Jones <scottajones@gmail.com>
Some 14-inch (J514S) and 16-inch (J516S) M3 Pro units on the 14.7 boot
chain load a third B0 GPU firmware image: TEXT 0x5c000, DATA 0x114000,
GKTS marker 0x58e88 and TEXT sha256
411833a7ded773feeac4a181552e3ff5582eaf4435f2dd8f620c1ed0b31df1b0. No
row pins it, so probe fails with "loaded GFX TEXT identity mismatch",
there is no render node and the compositor falls back to llvmpipe.

Three machines have reported it in issue omacom/linux-aurora#247, all on the 14.7 stub with
system firmware 26.6.2: a J514S and a J516S (both iBoot-10151.140.19.700.2,
mBoot-18000.161.10), which logged the identical sizes, marker and digest,
and a second 14-inch M3 Pro. With g15_accept_text_prefix=0x411833a7ded773fe
all three reached firmware READY with renderD128 present. The boot
chain matches the J514S cd05 reports, so the chain alone does not
select the image.

Add one row per board, J514S then J516S, so each pin still names the
machine it was measured on; nothing changes for other machines. Extend
the maintained-handoff host test: both boards pin to their own row with
the override disabled, other boards are refused, and an altered digest
suffix, TEXT or DATA size or marker is refused. Refresh the zero-words
golden pins for hw.rs and the T6030 module; with the two rows removed
both are their previous values.

Link: omacom/linux-aurora#247
Assisted-by: Claude:claude-opus-5-5
Signed-off-by: Scott Jones <scottajones@gmail.com>
Before the module's final activate() step, the G15 probe returned
-EPROBE_DEFER. That put the GPU on the driver core's deferred-probe list.
activate() then runs the real probe through driver_attach(), but if that
probe fails for good, for example because the firmware is rejected, the
driver core does not take the device off the list. It stays in
/sys/kernel/debug/devices_deferred until the deferred-probe work next
runs, so anything that waits on that list keeps waiting.

Return -ENODEV instead. activate() already reprobes every unbound device
once MODULE_READY is set, and a device added later sees it set, so
nothing needs the deferred list to retry. The display and PMP waits
inside the real probe still return -EPROBE_DEFER.

g15-probe-init-tests.py now checks that every probe before activation
returns -ENODEV, and adds a mutant that defers again.

Assisted-by: Claude:claude-opus-5-5
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: Scott Jones <scottajones@gmail.com>

(cherry picked from commit 6447458068e3 on omacom/linux-aurora, omacom/linux-aurora#470)
[7.3 base: the probe-init test also runs with the G16 registration here, so
 its pre-activation probe count follows the registration order; the new
 mutant is added to this base's expected-failure table.]
@RyanTheTide

Copy link
Copy Markdown
Contributor

The production boot stub pins at macOS 27.2 (26B123), so please retarget this to 27.2. This pins T6030 firmware TEXT that 14.7 boot chains load; please pin the images the 27.2 stub loads on the J514S and J516S.

@scottjones

Copy link
Copy Markdown
Contributor Author

Thanks Ryan. Before I retarget these I want to check I'm reading it right, because as far as I can tell the installer doesn't put 27.2 on any of these boards yet.

At aurora-installer 99f180f, the 26B123 pin has "boards": {}. Its only entries are the experimental j714c, j714s, j716c, j716s, j775c, j775d and j873s, all not_hardware_qualified. The release still ships supported_fw 13.5 and 14.8.3, and the per-device minimums put these PRs on:

PR Boards Installer minimum today
this one J514S, J516S 14.8.3
#2149 J613, J615 14.8.3
#2150 J616s 26.6.2
#2152 v14.7 DCP the M3 boards above, 14.8.3
aurora-silicon/aura#292 J475c 13.4

The "14.7" chains are the 14.8.3 stub: it reports the same iBoot string as 14.7.

The startup firmware doc says a Mac below macOS 27.2 keeps its older entry, and that moving an existing install goes through a slot, since M1 to M3 stub containers are 2.5 GB and a 26.0+ stub needs 6 GB. The slot executor isn't wired yet.

So retargeting these to 27.2 only would drop GPU, camera or display support for current installs on these boards, and those installs can't move to 27.2 until slots land and the owner updates macOS.

Did you mean adding 27.2 support alongside the shipped stubs once slots land, or is there a policy change I've missed? I'm happy to add 27.2 paths in follow-ups once there's a way to test them.

@RyanTheTide

Copy link
Copy Markdown
Contributor

Thanks for tracking down the J514S/J516S images and for the -EPROBE_DEFER fix, which looks right.

On the pins: drivers/gpu/drm/asahi/g15/hw.rs:182,198,214,230 add each of the two TEXT digests (cd05 and 4118) twice, once per board, with machine: Some("apple,j514s") or Some("apple,j516s"). The digest identifies the firmware, and that is what admission needs. Binding the same image to the boards it was seen on means a third T6030 board loading the identical image would be refused.

We select by a hardware or firmware property rather than by the list of machines it was tested on. Could each digest be pinned once for T6030 (machine: None), unless an image really is board-specific? If one is, a short comment saying why would cover it.

This sits alongside the earlier retarget-to-27.2 question, so the two could be handled together.

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.

4 participants