Skip to content

iommu: apple-dart: adopt T8132 internal display DARTs by default on 26.6.2 firmware - #2189

Open
erikwb wants to merge 1 commit into
aurora-silicon:integration/aurorafrom
erikwb:j713/display-dart-default
Open

erikwb wants to merge 1 commit into
aurora-silicon:integration/aurorafrom
erikwb:j713/display-dart-default

Conversation

@erikwb

@erikwb erikwb commented Nov 18, 2026 •

Copy link
Copy Markdown
Contributor

Problem

On T8132 (M4) Macs running 26.6.2 firmware, a freshly installed system comes up on simpledrm only: no native display, and the GPU never starts. The kernel adopted the internal display DARTs that iBoot leaves live (0x30a340000 for DCP SID23 and 0x30a300000 for scanout/PIODMA SID0/SID16) only when the command line set apple_dart.j713_dcp=1 apple_dart.j713_display=1, and no distribution sets them by default (J713 report: #1704; J604 behaves the same; distro workarounds in aurora-silicon/alarm-aurora#307 and aurora-silicon/fedora-aurora#348).

Change

  • Scope (26.6.2 firmware only): Gate internal display DART adoption in apple_dart_probe() directly on the "dcp" devicetree node instead of apple_t8132_firmware_path() or opt-in module parameters: require that "dcp" is available, compatible with "apple,t8132-dcp", and carries apple,firmware-version <26 6 2> via apple_dart_dcp_firmware(). Because Aura does not yet publish an internal DCP handoff property (and per §1 chicken-and-egg the loader side must land first), apple_dart_dcp_firmware() stays strictly as a temporary hardware-safety guard matching apple_dart_dcp_validate(), with the Aura handoff property and kernel transition tracked in kboot: T8132: publish an internal DCP handoff property so the kernel can drop the 26.6.2 firmware-version check aura#462 and T8132 display: key the DART and DCP firmware-layout choices on each DCP's handoff, not on version/address/board checks #2621: the retained-DART path assumes the inherited table and DMA layout of the 26.6.2 DCP firmware, so any other firmware (including the 27.2 fleet stub, which is out of scope for this PR) or disabled DCP leaves both internal display DARTs untouched (-ENODEV) without pinning tables, reserving the SID23 window, or resetting the live DARTs.
  • Remove module parameters & why the early bring-up watchdog precondition no longer applies: Delete the runtime-writable apple_dart.j713_dcp and apple_dart.j713_display module parameters and their -ENODEV opt-out branch. Those parameters originated in initial READY bring-up (52d846f6b8d3), when probing a live locked SID23 DART risked resetting or corrupting hardware state and wedging the SoC unless userspace armed a watchdog first. That precondition no longer applies to the adopted path:
    • Probe performs zero DART register or page-table writes: dart->preserve_display skips apple_dart_hw_reset() and apple_dart_drain_stale_exceptions(), and apple_dart_dcp_import() validates the locked T8110 registers (TCR, ENABLE_STREAMS, TTBR) and inherited root/leaf PTEs read-only against ADT reserved-memory bounds and excluded Linux RAM, refusing any mismatch or instability without mutating hardware.
    • Probe refuses adoption (-EIO) if any fault is already pending or masked (apple_dart_fault_pending() / DART_T8110_ERROR_MASK), attaches the shared fault IRQ only once both display DARTs are admitted, quarantines the faulting DART's borrowed streams on fault without issuing FLUSH_UNLOCK, and quarantines both display peers before releasing the shared IRQ.
    • If DART adoption is refused, drm/apple refuses DCP probe on the missing IOMMU group (beaae8e35b23), keeping simpledrm active without faulting.
  • Update the APPLE_DART_J713_DCP Kconfig prompt and help text to describe retained T8132 display DART support rather than an opt-in bring-up profile to leave disabled.
  • Adopting the display DARTs holds the display kexec blocker for the rest of the boot as before; on T8132 Macs the managed radio DART also holds its own kexec blocker.
  • Affects T8132 (M4) Macs with internal display DARTs (0x30a300000, 0x30a340000) on 26.6.2 firmware. External display DARTs (j713_dcpext, j713_dispext), 27.2 boots, and non-T8132 Macs are unchanged.

Testing

At head 0f35583d8414 (rebased onto integration/aurora a9626a1487ad; code diff identical to 552b5c89f5e9):

  • arm64 LLVM=1 W=1 build of drivers/iommu/apple-dart.o with CONFIG_APPLE_DART=y + CONFIG_APPLE_DART_J713_DCP=y (0 errors, 0 warnings).
  • tools/testing/m4-dcp-errors/module_guard.py: PASS.
  • Host suites (run-host-tests.py): hardware-defaults-tests (1/1 PASS) and dart suites (16/16 PASS).
  • scripts/checkpatch.pl --git origin/integration/aurora..HEAD: 0 errors, 0 warnings.

Hardware runs:

  • J715 (Mac16,13, MacBook Air 15-inch M4, OS 26.6.2, system 27.2, tested by @kazuhidet at 552b5c89, identical code diff):
    • integration/aurora 0abc2f385 + this PR's change (stage 1 -> aura, default command line with no apple_dart.* args and /sys/module/apple_dart/parameters/ absent): PASS (one boot).
    • Both internal display DARTs adopted (30a340000 "Preserving DCP SID23 root ...; no writes", 30a300000 "Preserving DCP SID0/SID16 root ..." and "Inherited SID0/SID16 frontend admitted"), 4 external display DARTs untouched, 0 DART faults.
    • DCP / appledrm 26.6.2: "KMS scanout activated" at 5.46 s (2880x1864 on eDP-1), GPU aurora-g16 READY at 5.96 s, sddm + desktop + keyboard/trackpad normal; failed units and kernel errors unchanged vs baseline (no SELinux on install; s2idle/resume and 27.2 OS firmware not covered).
  • Prior hardware runs (earlier head 3f59f0e1cbd9, default command line on 26.6.2):
    • J713 (Mac16,12, 26.6.2): SID23 root preserved with no writes, appledrm bound the DCP, GPU came up (renderD128), desktop session started.
    • J604 (MacBook Pro 14", 26.6.2, tested by @aeries0x): appledrm bound eDP-1 (3024x1964), GPU probed (Apple M4 (G16G B0)), off-screen glmark2 passed, 0 DART faults.

Tested-by: @kazuhidet

Still to run

  • J713 (26.6.2) boot at this head with the default command line (including 0 SELinux denials, 0 failed units, 0 new kernel errors).
  • s2idle/resume with the adopted internal display DARTs on integration/aurora (with 870eb638eb67).
  • J713/J715 27.2 boot log confirmation showing both internal display DARTs left untouched (Leaving inherited display DART untouched) and simpledrm intact.
  • J773g / J623 / J624 on 26.6.2 if dcp is enabled by the loader.

Assisted-by: LLM (Anthropic:Claude, Google:Gemini 4 Neon, OpenAI:GPT-6.2 Sol)

@aeries0x

Copy link
Copy Markdown
Contributor

Tested on a J604 (MacBook Pro 14", M4, DCP firmware 26.6.2) at 3f59f0e1cbd9 with my J604 config and a local revert of #1888 to keep the speakers off.

With no apple_dart parameters: the SID23 root was preserved with no writes, appledrm bound the DCP, and eDP-1 came up at 3024x1964. The GPU probed with renderD128 present, Vulkan reports Apple M4 (G16G B0), and an off-screen glmark2 run passed. After about 10 minutes there were no DART faults and no watchdog resets.

With apple_dart.j713_dcp=0 apple_dart.j713_display=0: both internal display DARTs logged "Leaving inherited display DART untouched", and the Mac stayed on simpledrm with no render node, matching the previous behavior.

@RyanTheTide

Copy link
Copy Markdown
Contributor

No-regression row from a 13" M4 MacBook Air (J713, Mac16,12) on the macOS 27.2 (26B123) stub. This covers the DART side only, not the display or GPU.

Setup: kernel 7.3.0-bj713+ = integration/aurora 0abc2f3 + this PR (3f59f0e) + linux#2205, booted with the image's default command line (no apple_dart.* parameters). Stage 1 is aura 3d517e63, Stage 2 c0a50c00.

Result: 6 boots (one first boot plus 5 warm reboots). Every boot had 0 apple-dart faults or errors in dmesg and logged "Retaining the live T8132 MTP bypass". The MTP keyboard and multi-touch devices are present, and NVMe, Wi-Fi and SEP are up. SELinux Enforcing, 0 failed units.

Not covered: the display and GPU paths this PR is mainly for. On 27.2 the internal DCP is still refused (firmware-27 path needs apple,bw-scratch from aura) and the GPU isn't handed over by aura, so neither got as far as DART here. We'll post again when the 27.2 DCP and G16 work reaches the J713.

Assisted-by: LLM (Claude:Opus 5.5)

@kazuhidet

kazuhidet commented Nov 21, 2026 •

Copy link
Copy Markdown

J715 row for this change (MacBook Air 15-inch, M4, Mac16,13; requested in #2663).

Build

Result: pass

Check Result
DCP DART 30a340000 J713 display DART version 0202, locked 1; SID23 root preserved, no writes
Frontend DART 30a300000 SID0 and SID16 preserved; Inherited SID0/SID16 frontend admitted
External display DARTs Leaving inherited external display DART untouched (as expected)
DCP firmware 26.6.2 retained DCP backend prepared → IOMFB startup complete ... 2880x1864 → boot scanout preserved → KMS scanout activated
appledrm fb0: appledrmfb frame buffer device (card1); backlight apple-panel-bl present
GPU aurora-g16: firmware identity accepted, 10 cores, READY; card0 / renderD128
Session the SDDM greeter and the Hyprland desktop showed normally on the internal panel
DART fault handlers no faults
Kernel errors only unrelated ones: avd probe -2, and the two external DCPs that are not handed over

/sys/module/apple_dart/parameters/* are mode 0600, so the values weren't read as a regular user; the kernel log above shows both DARTs admitted.

@RyanTheTide

Copy link
Copy Markdown
Contributor

Blockers: 2.

  1. With the defaults on, the internal display DARTs are borrowed whenever apple_t8132_firmware_path() holds (apple-dart.c:5363). On the J713/J715 that is true on any firmware (t8132.h:15-19, 52-55). apple_dart_dcp_prepare()/apple_dart_frontend_prepare() import the inherited tables with the 26.6.2 layout and check no version. The only firmware check, apple_dart_dcp_firmware(), runs at DCP attach. On the 27.2 image (and after drm/apple, iommu: apple-dart: drive the M4 MacBook Air's internal panel on macOS 27.2 (v27 backend) #2590 without the firmware-27 clocks) a default boot now pins both DARTs' tables, takes the display kexec blocker and reserves the SID23 window, and then the DCP is refused. Before this change the same boot left both DARTs untouched. Please check the internal DCP node's apple,firmware-version at DART probe, as the dcpext DARTs already do with apple,t8132-dcpext-firmware-ready, and leave the DART untouched otherwise. The board/OS-version helper is then no longer needed in j713_display.

  2. apple_dart.j713_dcp / apple_dart.j713_display stay as runtime-writable (0600) opt-outs. Their names say J713 but they cover the J715/J604/J773g/iMacs too, and a write affects a deferred re-probe. With drm/apple, iommu: apple-dart: drive the M4 MacBook Air's internal panel on macOS 27.2 (v27 backend) #2590 applied they no longer leave the 27.2 DARTs untouched. Please remove them along with the -ENODEV branch at :5393-5398. Once adoption is gated on a validated handoff, no opt-out is needed.

Also: the commit message says kexec and kdump are unavailable "unless both are set to 0". The radio DART (apple,managed-sids in t8132-jxxx.dtsi) already takes its own kexec blocker on every M4 Mac (apple-dart.c:5979), so turning the display parameters off does not bring kexec back. Please reword. The APPLE_DART_J713_DCP help text still describes a bring-up profile to leave disabled.

Rows still needed: J713 on 26.6.2 at this head with the default command line; s2idle/resume with the adopted DARTs (integration now carries 870eb63 on that path); and the 27.2 J713/J715 probe log lines for both internal display DARTs.

@erikwb
erikwb force-pushed the j713/display-dart-default branch from 3f59f0e to 28ae172 Compare November 22, 2026 00:37
@erikwb

erikwb commented Nov 22, 2026

Copy link
Copy Markdown
Contributor Author

Note

🤖 Gemini 4 Neon responding on behalf of erikwb

Addressed all review items in 28ae17213fee (rebased onto integration/aurora):

  • Internal display DART probe gate: j713_display no longer uses apple_t8132_firmware_path(). apple_dart_probe() now looks up the "dcp" devicetree node and verifies it is available, compatible with "apple,t8132-dcp", and carries apple,firmware-version <26 6 2> via apple_dart_dcp_firmware(). Any other firmware (including 27.2) or disabled DCP leaves both internal display DARTs untouched (-ENODEV) without pinning tables, taking the display kexec blocker, or reserving the SID23 window.
  • Removed module parameters: Deleted the runtime-writable apple_dart.j713_dcp and apple_dart.j713_display parameters and their -ENODEV opt-out branch.
  • Kconfig & commit message: Updated the APPLE_DART_J713_DCP prompt/help text so it no longer describes an opt-in bring-up profile to leave disabled, and reworded the kexec blocker note to reflect that the managed radio DART already holds its own kexec blocker on T8132 Macs.

Verified with arm64 W=1 builds of drivers/iommu/apple-dart.o (CONFIG_APPLE_DART=y/CONFIG_APPLE_DART_J713_DCP=y and CONFIG_APPLE_DART=m), module_guard (DMA guard + Kconfig matrix), and the apple-dart host test suites. The requested hardware rows (J713 on 26.6.2 at this head, s2idle/resume with adopted DARTs, and 27.2 J713/J715 probe log lines) are tracked under Still to run in the PR description.

@RyanTheTide

Copy link
Copy Markdown
Contributor

Re-checked at 28ae172.

  • Adoption gate: fixed. apple-dart.c:5379-5394 now leaves both internal display DARTs untouched unless the dcp node is available, apple,t8132-dcp, and carries apple,firmware-version 26.6.2, the same shape as the dcpext apple,t8132-dcpext-firmware-ready check below it. A 27.2 boot or a loader that does not fill the property no longer pins the tables or reserves the SID23 window.
  • Module parameters: fixed, both removed with their -ENODEV branch, and nothing else in the tree still refers to them.
  • Commit message (kexec) and Kconfig help: fixed.

Nit: the dev_err_probe at apple-dart.c:5393 still says "retained J713 display"; it applies to every M4 Mac.

Still needed at this head, as the body lists: a J713 26.6.2 boot with the default command line (native display and GPU), s2idle/resume with the adopted DARTs, and the 27.2 J713/J715 probe lines showing both internal display DARTs left untouched.

Assisted-by: LLM (Claude:Opus 5.5)

@RyanTheTide

Copy link
Copy Markdown
Contributor

Merge-train adversarial review (MERGETRAIN-LINUX2) at 28ae172: BLOCKERS.

Not merging until the blockers below are fixed or answered; the details have file:line and a failure case.

Full review

#2189 iommu: apple-dart: bring up the M4 internal display and GPU without boot parameters @28ae1721

VERDICT: BLOCKERS

  1. apple-dart.c:5379-5386 + apple-dart-dcp.h:1041: the default-on decision is a firmware-version gate (apple,firmware-version == <26 6 2>) on the default boot path. That breaks the "no fw-version gates when a DT/ADT property could carry it" rule.
    • The same function already shows the alternative: the external DCP DARTs are gated on a loader handoff property, apple,t8132-dcpext-firmware-ready (apple-dart.c:5401, 5419). The internal DCP should get the equivalent loader-set handoff property instead of a version compare.
    • Concrete failure: the fleet firmware stub pins 27.2. On every T8132 Mac booted on 27.2 (J713/J715 included) this PR changes nothing: no native display, no GPU, on the default cmdline. So the PR title ("bring up the M4 internal display and GPU without boot parameters") is false on the fleet firmware.
    • The validate() gate on the same version is older, so this moves an existing gate rather than adding a new one. Even so, it becomes the gate that decides the default, and it is not a capability.
    • If the orchestrator accepts the gate as hardware safety (the inherited 26.6.2 layout), downgrade this to a nit. Then the PR title and body must say "26.6.2 only".

Nits:
2. The removed module params were documented as "userspace must arm its watchdog before probing SID23". Adoption now runs unconditionally at boot with no watchdog. If adoption wedges the SoC (the reason that guard existed), the machine hangs at boot with no recovery. The PR should say why the watchdog precondition no longer applies.
3. apple-dart.c:5365: instances are still identified by hardcoded MMIO base (res->start == 0x30a300000/0x30a340000). This is older code, but the PR rewrites the line; a DT property or compatible on the two nodes would carry it.
4. The Kconfig help says "Say Y" but there is no default y. Only arch/arm64/configs/asahi.config sets it. Distros that do not use that fragment get the -EOPNOTSUPP probe failure on 26.6.2 Macs.
5. It now silently ignores apple_dart.j713_dcp=1 apple_dart.j713_display=1, which alarm-aurora#307 / fedora-aurora#348 add. Clean those workarounds up when this merges.

Missing rows:

  • J713 on 26.6.2 at this head on the default cmdline. The PR lists it as still to run; all hardware rows are from the older head 3f59f0e.
  • J773g (Mac mini, no internal panel) and J623/J624 (iMac) with dcp enabled by the loader on 26.6.2. Default adoption now reaches them and none were tested.
  • J713/J715 on 27.2: confirm "Leaving inherited display DART untouched" and that the system still boots to simpledrm.
  • s2idle/resume with the adopted internal DARTs on the tip (with 870eb63).

@DjDeveloperr DjDeveloperr left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Review at 28ae17213fee: changes requested.

CI is green (7/7), commit attribution is clean (Signed-off-by: only), removing the j713_dcp/j713_display module parameters is the right direction, and the Kconfig/kexec text updates look good.

Blockers:

  1. Firmware-version gate on the default boot path (§1) & scope mismatch (drivers/iommu/apple-dart.c:5379-5386, drivers/iommu/apple-dart-dcp.h:1041): apple_dart_probe() gates default internal display DART adoption on apple_dart_dcp_firmware(owner) (apple,firmware-version == <26 6 2>), right above j713_dcpext/j713_dispext which already gate on a loader-set DT handoff property (apple,t8132-dcpext-firmware-ready). Per pre-PR checklist §1:
    • Either replace the version compare with a loader-set handoff/capability property (landing the Aura side first with 26.6.2 fallback per §1 chicken-and-egg), or
    • If apple_dart_dcp_firmware() is kept strictly as a temporary hardware-safety guard matching apple_dart_dcp_validate() until the loader property lands, narrow the PR title and commit/body to 26.6.2 only — on the fleet's 27.2 firmware stub this change leaves both internal display DARTs untouched and brings up neither display nor GPU.
  2. Watchdog precondition rationale (§3): The removed module parameters existed because "userspace must arm its watchdog before probing SID23". Removing them makes SID23 adoption unconditional at early boot; please document in the commit message why the watchdog precondition no longer applies (so a regression doesn't silently wedge boot).

Non-blocking nits:

  • drivers/iommu/apple-dart.c:5389: dev_err_probe message still says "retained J713 display" rather than T8132.
  • drivers/iommu/Kconfig: APPLE_DART_J713_DCP help says "Say Y" without default y (relies on asahi.config).

Hardware rows still needed at the final head (§5):

  • All existing J713/J604/J715 runs are from the earlier 3f59f0e1 head before the of_find_node_by_path("dcp") probe-gate rewrite and param removal. Please run on this head: J713 26.6.2 default-cmdline boot (with 0 SELinux denials, 0 failed units, 0 new kernel errors), s2idle/resume with adopted DARTs on integration/aurora (with 870eb638), and 27.2 J713/J715 probe logs confirming "Leaving inherited display DART untouched".

Assisted-by: LLM (Gemini 4 Neon)

@erikwb erikwb changed the title iommu: apple-dart: bring up the M4 internal display and GPU without boot parameters iommu: apple-dart: adopt T8132 internal display DARTs by default on 26.6.2 firmware Nov 22, 2026
@erikwb
erikwb force-pushed the j713/display-dart-default branch from 28ae172 to 552b5c8 Compare November 22, 2026 07:47
@erikwb

erikwb commented Nov 22, 2026

Copy link
Copy Markdown
Contributor Author

Note

🤖 Gemini 4 Neon responding on behalf of erikwb

Addressed both blockers (and the dev_err_probe message nit) in 552b5c89f5e9 and updated the PR title and description:

  1. Scope narrowed to 26.6.2 firmware & safety-guard rationale (§1): Checked Aura's dt_set_m4_dcp_handoff() (src/kboot.c) — unlike dcpext (apple,t8132-dcpext-firmware-ready), Aura does not yet publish an internal DCP handoff property. Per §1 (land the loader side first), apple_dart_dcp_firmware() stays strictly as a temporary hardware-safety guard matching apple_dart_dcp_validate(), and the PR title, commit message, and PR description are now explicitly scoped to 26.6.2 firmware (noting that 27.2 leaves both internal display DARTs untouched and is out of scope here).
  2. Watchdog precondition rationale (§3): Documented in the commit message and PR body why the early READY bring-up (52d846f6b8d3) userspace-watchdog requirement no longer applies:
    • Probe performs zero DART register or PTE writes (preserve_display skips apple_dart_hw_reset() and apple_dart_drain_stale_exceptions(), and apple_dart_dcp_import() validates locked T8110 registers and inherited root/leaf PTEs read-only against ADT reserved-memory bounds and excluded Linux RAM).
    • Probe refuses adoption (-EIO) if any fault is already pending or masked, attaches the shared fault IRQ only once both display DARTs pass validation, quarantines the faulting DART's borrowed streams on fault without issuing FLUSH_UNLOCK, and quarantines both display peers before releasing the shared IRQ.
    • If DART adoption is refused, drm/apple refuses DCP probe on the missing IOMMU group (beaae8e35b23), keeping simpledrm safely active.
  3. Diagnostic string nit: Updated dev_err_probe() at drivers/iommu/apple-dart.c:5393 from "retained J713 display" to "retained T8132 display".

Verified with arm64 LLVM=1 W=1 builds of drivers/iommu/apple-dart.o (CONFIG_APPLE_DART=y+J713_DCP=y 0 warnings; =m/=n clean aside from pre-existing unused warnings), tools/testing/m4-dcp-errors/module_guard.py (DMA guard + 5/5 Kconfig cases PASS), and run-host-tests.py (hardware-defaults, 16/16 dart suites, j613-mtp-s2idle, bcm4377-radio-sids all PASS). Hardware rows for this head remain tracked under Still to run.

@RyanTheTide

Copy link
Copy Markdown
Contributor

Re-read at 552b5c8. Scoping this to the 26.6.2 firmware and keeping apple_dart_dcp_firmware() as a guard until aura publishes an internal-DCP handoff property is a reasonable order (loader side first). Since it's a firmware-version check, please open the follow-up (aura publishes the property, the kernel keys on it, the version check goes) and link it from the commit message, so the guard doesn't become permanent. The watchdog rationale in the message reads right. Our earlier LGTM stands; the J713 26.6.2 boot on this head (internal panel and GPU up, s2idle) is the row from our side.

Assisted-by: LLM (Claude:Opus 5.5)

@kazuhidet

Copy link
Copy Markdown

J715 row at 552b5c89 (MacBook Air 15-inch, M4, Mac16,13). It replaces my earlier row, which was at 3f59f0e1.

Build

Result: pass (one boot)

Check Result
dcp alias node /soc/dcp@30ae00000, apple,t8132-dcp, okay, apple,firmware-version = 26 6 2 (set by aura)
Internal display DARTs both adopted: 30a340000 "Preserving DCP SID23 root …; no writes", 30a300000 "Preserving DCP SID0/SID16 root …" and "Inherited SID0/SID16 frontend admitted". No "Leaving inherited display DART untouched" for either
External display DARTs (4) "Leaving inherited external display DART untouched", as before
DART faults 0
DCP / appledrm firmware 26.6.2, "KMS scanout activated" at 5.46 s, 2880x1864 on eDP-1
GPU aurora-g16 READY at 5.96 s
Session sddm active, desktop normal, keyboard and trackpad normal
Failed units 2, both local to this install (j713-owner-assets, m4-j715-backlight). They fail the same way without this PR
Kernel errors the same set as without this PR: avd probe -2, hid-generic … no listeners, and external DCP not handed over on both dcpext
SELinux not applicable (no SELinux on this install)

Not covered:

  • s2idle/resume with the adopted DARTs.
  • 27.2 OS firmware. This J715 stays on 26.6.2, so I can't give the 27.2 probe lines.

I can't explain one line yet. At 21 s the DCP logs brightness 0 and IOMFB: power on -> off. The record stops there, and the screen was normal in use. That is probably the session's idle blanking, but I haven't confirmed it.

…6.6.2 firmware

On T8132 (M4) Macs running 26.6.2 firmware, the internal display DARTs
that iBoot leaves live (the DCP's SID23 at 0x30a340000 and the
scanout/PIODMA frontend at 0x30a300000) were adopted only when the
command line set apple_dart.j713_dcp=1 and apple_dart.j713_display=1.
No distribution sets them by default, so an installed 26.6.2 system came
up on simpledrm: appledrm never bound the DCP, and the GPU never
started.

Gate internal display DART adoption directly on the internal DCP
devicetree node instead of the board/OS-version helper or opt-in module
parameters: require that the "dcp" node is available, compatible with
"apple,t8132-dcp", and carries apple,firmware-version 26.6.2 via
apple_dart_dcp_firmware(). That check stays strictly as a hardware-safety
guard matching apple_dart_dcp_validate() until a loader-published
internal DCP handoff property lands in Aura first and the kernel keys on
it (tracked in aurora-silicon#2621): the
retained-DART path assumes the inherited table and DMA layout of the
26.6.2 DCP firmware, so any other firmware (including 27.2, which is out
of scope here) or disabled DCP leaves both internal display DARTs
untouched (-ENODEV) without pinning their tables, reserving the SID23
window, or resetting the live DARTs.

Remove the runtime-writable apple_dart.j713_dcp and
apple_dart.j713_display module parameters, and update the
APPLE_DART_J713_DCP Kconfig help text to reflect that it is no longer
an opt-in bring-up profile. Those parameters originally guarded early
READY bring-up with a requirement that userspace arm a watchdog before
probing SID23, when an unvalidated probe risked resetting or corrupting
a live locked display DART and wedging the SoC. That precondition no
longer applies to the adopted path:
- probe performs no DART register or page-table writes: preserve_display
  skips apple_dart_hw_reset() and apple_dart_drain_stale_exceptions(),
  and apple_dart_dcp_import() validates the locked T8110 registers
  (TCR, ENABLE_STREAMS, TTBR) and inherited root/leaf PTEs read-only
  against ADT reserved-memory bounds and excluded Linux RAM, refusing
  any mismatch or instability without mutating hardware;
- probe refuses adoption (-EIO) if any display DART fault is already
  pending or masked, attaches the shared fault IRQ only once both
  display DARTs pass validation, quarantines the faulting DART's
  borrowed streams on fault without issuing FLUSH_UNLOCK, and
  quarantines both display peers before releasing the shared IRQ;
- a refused DART leaves DCP probe refused (via the IOMMU-group check in
  drm/apple) so simpledrm keeps the panel safely.

Adopting the display DARTs holds the display kexec blocker for the rest
of the boot, as it already did with the parameters set; on T8132 Macs
the managed radio DART also holds its own kexec blocker.

Link: aurora-silicon/aura#462
Link: aurora-silicon#2621
Signed-off-by: Erik Bourget <erik@ebourget.net>
@erikwb
erikwb force-pushed the j713/display-dart-default branch from 552b5c8 to 0f35583 Compare November 22, 2026 17:05
@erikwb

erikwb commented Nov 22, 2026

Copy link
Copy Markdown
Contributor Author

Note

🤖 Gemini 4 Neon responding on behalf of erikwb

Rebased onto integration/aurora (a9626a1487ad) and linked the follow-up issues in 0f35583d8414 (code diff identical to 552b5c89f5e9):

  • Follow-up links: Opened https://github.com/aurora-silicon/aura/issues/462 (Aura publishing an internal DCP handoff property on T8132) and linked both aurora-silicon/aura#462 and aurora-silicon/linux#2621 in the commit message and PR body to track removing apple_dart_dcp_firmware() once the loader property lands and the kernel keys on it.
  • Rebase & verification: Rebased cleanly onto integration/aurora a9626a1487ad (resolving the apple_dart_board() context), verified arm64 LLVM=1 W=1 build of drivers/iommu/apple-dart.o (0 errors, 0 warnings), tools/testing/m4-dcp-errors/module_guard.py (PASS), run-host-tests.py (hardware-defaults 1/1, dart 16/16 PASS), and scripts/checkpatch.pl (0 errors, 0 warnings).
  • J715 hardware row: Thanks @kazuhidet for the J715 (Mac16,13, 26.6.2) PASS row at 552b5c89 — added it and your Tested-by: credit to the PR body.

J713 26.6.2 boot + s2idle/resume and 27.2 untouched-DART confirmation remain under Still to run.

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

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants