You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
J700 now has suspend-to-idle reports with RTC, keyboard and trackpad wake, a rear-port external display up to 1920×1080 at 180 Hz, a working camera and Touch ID matches after resume. Each result belongs to the development kernel its PR names. A completed public install, standby power, speaker playback and sustained thermal safety remain unqualified.
Refreshed October 16, 2026. Current public branches, PR states, follow-up reports and source were checked. No new hardware test was run for this update; the results below are those reported in the linked PRs, on the kernels they name. Source integration, individual subsystem tests and a complete laptop qualification are separate milestones.
What changed · October 16, 2026
Suspend-to-idle:linux #222 keeps NVMe, GPU and Wi-Fi powered and enables the PMU RTC alarm. With linux #234, linux #235 and linux #236, ten 70-second RTC cycles kept every tested subsystem working with 293–297 ms device resume, and keyboard and trackpad now wake the machine. linux #298 fixes s2idle timekeeping. USB wake (linux #304) is open and standby power is unmeasured.
External display: the rear-port DP Alt Mode path no longer sits behind a module parameter since linux #293. Its R69 build drove a monitor at 1920×1080 and 180 Hz in both cable orientations, with the EDID name and clamshell use, and linux #313 brings it back after s2idle. HDR, DSC, VRR and DP audio are unsupported; the front port is USB 2.0 only.
Touch ID:linux #303 reports 2/2 attended matches after resume and three enrolled prints kept through more than 100 s2idle cycles. linux #325 registered the sensor on 104/104 cold boots, where 7 of 16 had failed.
Radios and audio:linux #323 passed 20 s2idle cycles without a Wi-Fi failure and linux #266 kept 3589/3595 pings over an idle hour. High-quality microphone capture now works (linux #237, linux #239); speaker playback is measured only in open asahi-audio #7.
Boot chain and packaging:m1n1 #2 and U-Boot #1 are now merged into their aurora-wip branches, and linux #301 avoids a boot SError under macOS 27 system firmware. The Neo installer runs from macOS 27 and its DMG (omarchy-mac #1), but no completed public install is reported; aurora_j700_defconfig (linux #306) is open.
How to read this tracker
Marker
Meaning
🟢 Reported on hardware
A linked report identifies J700 and the tested feature. Its exact revisions and limits still apply.
🔵 Published code
Public implementation exists; this alone does not establish hardware operation.
🟠 Partial
Integration, qualification or functionality remains incomplete.
⚪ Not established
No qualifying J700 result was found in this audit; this does not mean unsupported.
🔴 Known gap
A public report identifies missing or failing functionality.
Results on another board, a private integration, an older implementation or an open branch do not automatically qualify the public J700 stack. “Merged” below refers to Aurora's repository, not mainline Linux or a distribution release.
Includes the J700 suspend, external-display, SEP, ISP, DART, power, GPU and radio series merged through October 16. The default branch is aurora-stable, which does not carry this work. Merged linux #274 fragments reproduce the tested configuration; the single defconfig is still open.
m1n1 J700
m1n1 #2 merged October 10 into aurora-wip and main; m1n1 #32 display-handoff validation merged into aurora-wip only (tip f3ac615eb137)
The handoff validation reports host tests only; no J700 boot of its exact head is reported. External-display coprocessor handover (m1n1 #26 and an earlier copy) closed without merge.
U-Boot J700
U-Boot #1 merged October 11 into aurora-wip (0cb6fed6a105), which also carries the U-Boot #5 NVMe write-cache flush
The merged PR reports two J700 boots and one 30 s s2idle. Built-in keyboard and trackpad capture in U-Boot is still pending.
Local firmware, unit calibration and device identity are required. They are not supplied by the kernel PRs; do not substitute another machine's inputs.
Installer and packages
omarchy-mac #1 (merged to feat/neo) runs the installer from macOS 27.0 and its DMG. DMG sealing (omarchy-mac #4) and a package kernel pin to a6cfa4dd are open
The installer's full install reached Recovery; no completed public install is reported. The pin's Neo check booted 99b5f912 (3/3 s2idle), not the packaged build.
Hardware status
🧩 Core platform
Hardware / SoC block
Public implementation
Hardware evidence / limits
Sources
Boot and kernel handoff
🔵 m1n1 and U-Boot J700 work merged; #301 leaves the firmware-owned disp_cpu power state untouched
🟢 KDE on the earlier RAM-overlay stack; internal ESP → GRUB → Linux reported; the 2026-10-15 release kernels were one-shot GRUB boots. #301 booted to the desktop under macOS 27.0 system firmware (26.6.2 OS firmware), using a test-only hook until a rebuilt Stage 2 DTB ships. 🟠 No completed public install
🟢 Earlier USB, radio and AVD DMA results stand. #221: the display hand-off fault (3 of 52 earlier boots) was absent in 5/5 boots. #232: 1,785 freed audio buffers stayed intact over 1,572 cycles. #237: 54,392 capture cycles in 31 min, 0 TLB timeouts. 🟠 Open #304 changes DART resume on every Apple SoC and awaits M1 Pro/M2 runs
🔵 Driver and handoff merged; U-Boot now flushes its NVMe write cache
🟢 Internal-storage boot; NVMe live after s2idle (#222). #320: 5/5 RTC s2idle cycles on each of 4 release kernels, 363–407 fsynced MiB intact per cycle. 🟠 Inherited live ANS session required; reset-under-I/O unqualified
🔵 Merged; #214 keeps an idle E cluster at or above 1464 MHz under the power allocator
🟢 4044/2424 MHz maxima, uncapped in the release power check (#317). #214 cut the worst idle E-core burst from 114 to 61 ms under the performance governor. #55 reports capacity on its own branch
🔵 WFI only, by design (linux #216); linux #298 adds an s2idle-eligible WFI entry
🟢 #216: plain WFI power-gates the E cores and second P core in 94–99.9% of 1 kHz samples; the boot P core never gates. The CPU power-down CYC_OVRD write is an undefined instruction on T8140, which explains the earlier hang. #298 (contributor, base 3b2bd35a8591): 3/3 RTC sleeps accounted 19.8–20.4 s of 20 s, against 3/3 failures before. No deep CPU power-down
🔵 Native H17P DCP, 250 ms backlight fade (linux #263) and hand-off DART fix integrated
🟢 GNOME 2408×1506@60, 2–525 nits, 20/20 DPMS. #263: the fade settles in 264–279 ms and returns at the same level after a 20 s s2idle (revision not stated). Panel and backlight return after s2idle (#222, #313). 🟠 DCP firmware-crash recovery unqualified
🔵 G17 driver in aurora-wip (the closed #171 tree, landed as direct commits) plus #305–#309; open linux #200/linux #201 stay on neo-gpu
🟢 #171's own Mesa/kernel pair: Vulkan CTS 9,961 pass, 0 fail. #307: no desktop loss over 3 boots (3/3 before); #308: a stalled ray traversal no longer crashes the firmware (3/3 before); #305: no queue exhaustion past 150 failed clients. Mesa agx3b: sparse 2,553 pass / 0 fail, RT pipeline 2,504 pass / 3 fail. 🟠 Needs Mesa agx3b (mesa #1 to main is open). No task shaders; tessellation-stage ray queries fail; completion-stamp and 8 GB memory-pressure limits remain; no game or Steam qualification
🔵 Rear-port H17P external path, unconditional since linux #293
🟢 #293 R69 build with U-Boot 0cb6fed6: Acer QG270 S3 through a dock at 1920×1080 180 Hz 8-bit SDR, EDID name, both cable orientations, clamshell and one attended suspend batch. #313: eDP plus 1080p180 back after RTC s2idle in 6/6 cycles, 3 runs. 🟠 Only single attended replugs reported; two sinks untested. 🔴 HDR, DSC, VRR and DP audio unsupported; front port is USB 2.0 only
🟠 Merged PRs report no speaker playback. Open asahi-audio #7 measured +3 dB make-up gain with a −0.5 dBFS limiter: 0 clipped samples, no speakerguardd reduction over 60 s. Headphone jack: start volume only. Bluetooth audio is separate
🔵 AOP audio, run-state counting (linux #239), capture stop sync (#259) and UCM on the AOP card
🟢 #237: 100/100 high-quality array captures carry fresh data. #239: 85,313 capture cycles in 31 min with 0 EIO (41 before). #38's earlier high-quality capture failure did not recur in these runs. 🟠 Lid-closed microphone hiding is open
🟢 1080p NV12 at 30.01 fps; 30 fps held at 0 lux (15 fps before). #229 (tip 90c6f8c7564b): 3 idle and 2 mid-capture s2idle cycles resumed, captures completing 3600 and 2100 frames. #319: 301 frames in 10 s, 30.00 fps after s2idle. 🟠 Firmware starts once per boot, so a rebind fails until reboot; about 30–50 mW extra idle draw
🔵 Port, PHY and DART support; linux #236 keeps the fixed-hub controller running across sleep
🟢 Earlier connector, SSD and PD results stand. #236: device resume 2822 → 293–297 ms with no re-enumeration. 🟠 USB wake is open linux #304; its own kernel passed attended USB-keyboard wake, including clamshell. USB3 at both connectors unqualified
🔵 Automatic one-shot bootstrap, retained across s2idle
🟢 #204: both functions enumerate in about 1.4 s; the Wi-Fi function stays powered through s2idle (#222). Arena retained until reset; removal, kexec and hibernation unsupported
🔵 Fullmac plus merged robustness fixes (#261, #266, #323 and others)
🟢 Channel 149 at 40 MHz: TCP about 183 Mbit/s up and 187 down on development kernels, 65.8–106.8 up and 90.8–111.9 down on the release candidate; 3589/3595 idle pings in an hour (#266). #323: 20 s2idle cycles, 0 failures. 🟠 DFS joins, 80/160 MHz, WPA3, roaming, AP and module reload unqualified; throughput gaps during scans; 2.4 GHz loss after several s2idle cycles
🔵 Modular transport, start after Wi-Fi, controller reset and suspend fixes
🟢 Power and scan (17–34 devices); linux #230: 10 s2idle cycles with 0 Bluetooth errors; 3/3 forced controller resets recovered (#262). 🟠 No current PR tests pairing, audio profiles, reconnect after resume or Bluetooth wake; earlier AirPods/AAC results are historical
🔵 J700 input merged; linux #235 makes them wake sources
🟢 Input operator-confirmed in the #33 campaign. #235 (attended): Shift, click and touch each woke s2idle; with the lid closed for about 90 s it stayed asleep. The first contact of a waking slide is dropped. U-Boot preboot keyboard is separate
🔵 Integrated; linux #234 keeps the GPIO domain on for the enclave
🟢 #42: boot, RNG, trusted keys. #234: after each of 10 × 70 s s2idle cycles, hwrng, trusted-key seal/load and a Touch ID verify start worked (before: wedged after any sleep of 30 s or more). #326 refuses a damaged host-state store; its one boot left the real store byte-identical
🟢 #303 on the 2026-10-15 release stack: 2/2 attended matches after resume; 3 enrolled prints kept through more than 100 s2idle cycles over 6 boots; 20 cancelled verifies with 0 EBUSY. #325: 104/104 cold boots registered (7 of 16 failed on unmodified release7). 🟠 Fresh enrolment on a public install not reported; lid-closed hiding (aurora-sep-userspace #1) is open; SELinux packaging (#329) was tested on M1 Pro/M2 only
🔵 Merged; #317 keeps a failed sensor's hwmon channel slot
🟢 RTC read/write and reboot persistence; rtc0 is SMC and rtc1 the PMU wake alarm (#222); 84 temperature channels in the release check. Some labels remain guesses
🔵 Battery driver merged; UPower now keeps the kernel's time estimates
🟢 Full/100% readings; the UPower change ran 70 s on a full battery. Charging/discharging transitions and positive time estimates still need a board report
🔵 IPA, PMP firmware limits, #214 cooling maps and the #219 DCS limit on Tp0t (95 °C control)
🟢 #219: without the DCS limit a performance-governor GL/Vulkan load reached 104 °C in 36 s; with it, peaks were 97–98 °C and the 10-minute composite score fell about 1–2%. 🟠 Battery-only, long-duration and hotplug/teardown thermal safety remain open
🔵 s2idle (linux #222); wake from RTC, power button, lid open, keyboard and trackpad
🟢 See the suspend table below; #235 also reports an attended wake on opening the lid. 🟠 USB wake open in #304; Bluetooth wake not implemented; standby power unmeasured; hibernation refused
aurora-wipac895820e + polish series (rel-g257ca7bed, -g3faecd4f7)
DCP 6/6 × 3 runs; USB 5/5 per run plus 10; storage 5/5
eDP + 1080p180 back with 0 DCP errors; lsusb steady at 13; Touch ID, camera, audio and NVMe checks pass
Standby power was not measured; NVMe, GPU, Wi-Fi and ISP domains stay powered (the ISP alone adds an estimated 30–50 mW). Bluetooth cannot wake the machine and USB wake needs open #304.
Limits: only 1920×1080 at 60, 120 and 180 Hz was tested. Five-fold hotplug, two sinks and plugging in during a modeset are untested; VRR is refused. The front-port notifier (linux #334) is open and not exercised on hardware.
Boot chain: the R69 run names U-Boot but not its m1n1 revision. Initramfs bring-up (aurora-neo-platform-config #2 and an Omarchy twin) is open and tested offline only.
Neo video-decode qualification
Physical J700/T8140, kernel 7.1.12-aurora-neo-existing-wifi-20261001-r2+, matching AVD module and added AVD/DART/power-domain wiring, with a temporary RAM overlay. Firmware was replaced and the idle driver normally unloaded/reloaded; this was not a new whole-system cold-boot qualification.
Test
Hardware runs
Exact hardware/software frame matches
H.264 360p/1080p, normal download and vaGetImage
4
120
HEVC 360p, 8/10-bit, both download paths
4
120
VP9 360p, 8/10-bit, both download paths
4
120
AV1 360p, 8/10-bit, both download paths
4
120
H.264/HEVC/VP9 720p, repeated full decode and 1s/3s seeks
18
1,260
Total
34
1,740
Every run required hardware/backend markers and byte comparison against software decoding. Both guarded groups finished idle without timeout, wedge, decoder error or kernel fault. KDE, Wi-Fi HTTPS and powered Bluetooth passed afterward. Other SoCs, concurrent clients, suspend/resume and application-level playback are outside this qualification.
VA-API backend: libva-v4l2_request 5b5046cbda689, with FFmpeg n9.0.1 and libva 2.24. The framebuffer setup used /dev/dri/card0 explicitly.
Evidence archive retained locally, not public: neo-avd-aurora-evidence.tar.gz, SHA-256 dbd939da3b20e8c06d85d6e12538cadd559a15a18d665c287597678d672faa04 (11 input clips, scripts, frame hashes, logs and guard records). Public results and attribution are in #154.
The decoder foundation and #45 fixes credit sofus, Aaron (aquarat), and iconidentify. Neo bring-up credits DJ, Ryan Murray and Ace. Original author/co-author history and license notices are retained; Aurora firmware includes DJ's changes and preserves the inherited MIT notices.
Radio qualification and remaining work
The tested Wi-Fi path is fullmac. This tracker does not claim that the proposed mt76/softmac path works. Locally installed firmware and the target unit's calibration/identity were used; no blobs, credentials or unit-specific data are included in the PRs.
Coexistence test
Gateway replies
HTTPS transfers
Result
5 GHz Wi-Fi with Bluetooth discovery/HCI power cycling
54/54
18/18
No kernel/driver faults
2.4 GHz Wi-Fi with Bluetooth discovery/HCI power cycling
53/54
18/18
One missed ping; not a zero-loss result
Earlier Bluetooth-only testing confirmed AirPods reconnect and user-audible AAC output after enabling PipeWire/WirePlumber. Headset audio during simultaneous Wi-Fi traffic was not tested on the combined build. The late replay disconnect was user-initiated; that replay is not counted as a complete stability run.
Current radio boundary:linux #205, linux #206 and linux #207 merged on October 8 and report AU → US → AU country changes with 20/20 pings and 0.5 s address-change reconnects on 7.1.12-aurora-j700-fu1+. Later fixes add signal-threshold events (no background scans in 180 s, #261), one TX doorbell per burst (#266) and safety fixes (#318, #323). Firmware, regulatory data and calibration remain local inputs. Removal, kexec, hibernation, DFS joins, WPA3/required MFP, roaming, AP/P2P, pairing, SCO, LE Audio and long-duration stability are not qualified. The coexistence table is historical, not a rerun of the current head.
Audit scope and public references
This refresh covers public PR states, bodies and comments, current branches and relevant source. Earlier hardware results keep their original revisions; no physical test was run for this update.
A18 Pro catalogue: its October 1 Linux summary predates the radio, decode, suspend and display work above; hardware identity is not a support guarantee.
Reply with one feature or failure per report, tied to exact board and software revisions.
Board: J700
SoC: T8140
Model / configuration: MacBook Neo
Feature or SoC block:
Kernel repository / branch / commit:
m1n1 / U-Boot / Mesa commits, where relevant:
Firmware source / revision / hash:
Steps to reproduce:
Expected result:
Observed result and repeat count:
Public log or artifact URL:
Artifact SHA-256 and command to repeat the check:
Known limits / recovery steps:
Keep credentials, per-unit calibration and device identities out of public reports. A maintainer can update the table when the linked evidence establishes the result for this board.
One hardware thread for MacBook Neo in the A18 family. Keep reports tied to J700/T8140 and exact software revisions.
Current source cross-check: the T8140 G17 match, the s2idle WFI entry, J700 RTC aliases, the memory-cache driver, J700 config fragments and the removed external-display parameter were checked in the pinned tree; aurora_j700_defconfig and tools/aurora-wake are not in it. Original merge ancestry is recorded separately below.
Audit baseline:f53454ba0da1; public source and PR reports checked October 16, 2026. No new physical test was performed for this update. Working-tree integration does not imply an aurora-stable release.
Merge is in the pinned aurora-wip history · 72f3265bfc55
PR state, source integration and hardware qualification are separate. A closed PR is not necessarily merged. Public reports describe their original test kernels, including contributor branches. Repository administrators can edit this opening post.
hardwareBoard-specific hardware evidence and SoC indexessoc:T8140Hardware discussions for T8140 (A18 Pro)
1 participant
Heading
Bold
Italic
Quote
Code
Link
Numbered list
Unordered list
Task list
Attach files
Mention
Reference
Menu
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
MacBook Neo · A18 family
🧭 Hardware directory › A18 category › J700 · T8140
T8140J700Important
J700 now has suspend-to-idle reports with RTC, keyboard and trackpad wake, a rear-port external display up to 1920×1080 at 180 Hz, a working camera and Touch ID matches after resume. Each result belongs to the development kernel its PR names. A completed public install, standby power, speaker playback and sustained thermal safety remain unqualified.
Refreshed October 16, 2026. Current public branches, PR states, follow-up reports and source were checked. No new hardware test was run for this update; the results below are those reported in the linked PRs, on the kernels they name. Source integration, individual subsystem tests and a complete laptop qualification are separate milestones.
What changed · October 16, 2026
aurora-wipas direct commits; drm/asahi: recycle settled quarantined G17 compute queues #305 and drm/asahi: g17: replay render groups firmware retired without executing #307–drm/asahi: g17: ray queries in render stages #309 add recovery fixes and render-stage ray queries. linux #215 turns on the system-level cache, about doubling GL throughput. Ray tracing, mesh shaders and sparse resources (mesa #12, mesa #22) are on Mesa branchagx3b, not Mesamain.aurora-wipbranches, and linux #301 avoids a boot SError under macOS 27 system firmware. The Neo installer runs from macOS 27 and its DMG (omarchy-mac #1), but no completed public install is reported;aurora_j700_defconfig(linux #306) is open.How to read this tracker
Results on another board, a private integration, an older implementation or an open branch do not automatically qualify the public J700 stack. “Merged” below refers to Aurora's repository, not mainline Linux or a distribution release.
Source and integration status
linux/aurora-wip@f53454ba0da1aurora-stable, which does not carry this work. Merged linux #274 fragments reproduce the tested configuration; the single defconfig is still open.aurora-wipandmain; m1n1 #32 display-handoff validation merged intoaurora-wiponly (tipf3ac615eb137)aurora-wip(0cb6fed6a105), which also carries the U-Boot #5 NVMe write-cache flushavd-fwat9cbde5c7d993avd-fw-v5-t0.binand install it separately. This is the firmware used for the latest Neo decode qualification.feat/neo) runs the installer from macOS 27.0 and its DMG. DMG sealing (omarchy-mac #4) and a package kernel pin toa6cfa4ddare open99b5f912(3/3 s2idle), not the packaged build.Hardware status
🧩 Core platform
disp_cpupower state untouchedCYC_OVRDwrite is an undefined instruction on T8140, which explains the earlier hang. #298 (contributor, base3b2bd35a8591): 3/3 RTC sleeps accounted 19.8–20.4 s of 20 s, against 3/3 failures before. No deep CPU power-down🎨 Display and media
aurora-wip(the closed #171 tree, landed as direct commits) plus #305–#309; open linux #200/linux #201 stay onneo-gpuagx3b: sparse 2,553 pass / 0 fail, RT pipeline 2,504 pass / 3 fail. 🟠 Needs Mesaagx3b(mesa #1 tomainis open). No task shaders; tessellation-stage ray queries fail; completion-stamp and 8 GB memory-pressure limits remain; no game or Steam qualification0cb6fed6: Acer QG270 S3 through a dock at 1920×1080 180 Hz 8-bit SDR, EDID name, both cable orientations, clamshell and one attended suspend batch. #313: eDP plus 1080p180 back after RTC s2idle in 6/6 cycles, 3 runs. 🟠 Only single attended replugs reported; two sinks untested. 🔴 HDR, DSC, VRR and DP audio unsupported; front port is USB 2.0 only90c6f8c7564b): 3 idle and 2 mid-capture s2idle cycles resumed, captures completing 3600 and 2100 frames. #319: 301 frames in 10 s, 30.00 fps after s2idle. 🟠 Firmware starts once per boot, so a rebind fails until reboot; about 30–50 mW extra idle draw🔌 I/O and security
🔋 Power and system readiness
rtc0is SMC andrtc1the PMU wake alarm (#222); 84 temperature channels in the release check. Some labels remain guessesTp0t(95 °C control)Suspend-to-idle reports
aurora-wip3fecb279810d+ series90c6f8c7564b+ ISP series3b2bd35a8591CLOCK_MONOTONICwifisafe-g5cbf717b9aurora-wipac895820e+ polish series (rel-g257ca7bed,-g3faecd4f7)Standby power was not measured; NVMe, GPU, Wi-Fi and ISP domains stay powered (the ISP alone adds an estimated 30–50 mW). Bluetooth cannot wake the machine and USB wake needs open #304.
External display (DP Alt Mode)
Neo video-decode qualification
Physical J700/T8140, kernel
7.1.12-aurora-neo-existing-wifi-20261001-r2+, matching AVD module and added AVD/DART/power-domain wiring, with a temporary RAM overlay. Firmware was replaced and the idle driver normally unloaded/reloaded; this was not a new whole-system cold-boot qualification.vaGetImageEvery run required hardware/backend markers and byte comparison against software decoding. Both guarded groups finished idle without timeout, wedge, decoder error or kernel fault. KDE, Wi-Fi HTTPS and powered Bluetooth passed afterward. Other SoCs, concurrent clients, suspend/resume and application-level playback are outside this qualification.
avd-fw9cbde5c7d993844bf086318974958b11076600a0, using its Meson build. Installavd-fw-v5-t0.binat/lib/firmware/apple/avd-fw-v5-t0.bin.d9f9e699dd63df51a3f3a44eee5141657d0a55d91d87039158e5edfe511ed63e.5b5046cbda689, with FFmpeg n9.0.1 and libva 2.24. The framebuffer setup used/dev/dri/card0explicitly.neo-avd-aurora-evidence.tar.gz, SHA-256dbd939da3b20e8c06d85d6e12538cadd559a15a18d665c287597678d672faa04(11 input clips, scripts, frame hashes, logs and guard records). Public results and attribution are in #154.The decoder foundation and #45 fixes credit sofus, Aaron (aquarat), and iconidentify. Neo bring-up credits DJ, Ryan Murray and Ace. Original author/co-author history and license notices are retained; Aurora firmware includes DJ's changes and preserves the inherited MIT notices.
Radio qualification and remaining work
The tested Wi-Fi path is fullmac. This tracker does not claim that the proposed mt76/softmac path works. Locally installed firmware and the target unit's calibration/identity were used; no blobs, credentials or unit-specific data are included in the PRs.
Earlier Bluetooth-only testing confirmed AirPods reconnect and user-audible AAC output after enabling PipeWire/WirePlumber. Headset audio during simultaneous Wi-Fi traffic was not tested on the combined build. The late replay disconnect was user-initiated; that replay is not counted as a complete stability run.
Current radio boundary: linux #205, linux #206 and linux #207 merged on October 8 and report AU → US → AU country changes with 20/20 pings and 0.5 s address-change reconnects on
7.1.12-aurora-j700-fu1+. Later fixes add signal-threshold events (no background scans in 180 s, #261), one TX doorbell per burst (#266) and safety fixes (#318, #323). Firmware, regulatory data and calibration remain local inputs. Removal, kexec, hibernation, DFS joins, WPA3/required MFP, roaming, AP/P2P, pairing, SCO, LE Audio and long-duration stability are not qualified. The coexistence table is historical, not a rerun of the current head.Audit scope and public references
This refresh covers public PR states, bodies and comments, current branches and relevant source. Earlier hardware results keep their original revisions; no physical test was run for this update.
🧪 Add a reproducible hardware report
Reply with one feature or failure per report, tied to exact board and software revisions.
Keep credentials, per-unit calibration and device identities out of public reports. A maintainer can update the table when the linked evidence establishes the result for this board.
One hardware thread for MacBook Neo in the A18 family. Keep reports tied to J700/T8140 and exact software revisions.
Current source cross-check: the T8140 G17 match, the s2idle WFI entry, J700 RTC aliases, the memory-cache driver, J700 config fragments and the removed external-display parameter were checked in the pinned tree;
aurora_j700_defconfigandtools/aurora-wakeare not in it. Original merge ancestry is recorded separately below.Audit baseline: f53454ba0da1; public source and PR reports checked October 16, 2026. No new physical test was performed for this update. Working-tree integration does not imply an
aurora-stablerelease.PR state and integration audit · October 16, 2026
mainccb95b45db2cmain1fe563501e7bmainbbd6aa9057afaurora-wipaurora-wiphistory ·566b52139598aurora-wipaurora-wiphistory ·7693112b7caaaurora-wipaurora-wiphistory ·1c9b7b5caa6caurora-wipaurora-wiphistory ·671ad267cb37aurora-wipaurora-wiphistory ·57c662cd1eb1aurora-wipaurora-wiphistory ·ed4fd3a576b4aurora-wipaurora-wiphistory ·16ed49a8919aaurora-wipaurora-wiphistory ·4fdff5cdd810aurora-wipaurora-wiphistory ·355ec1c1e481aurora-wipaurora-wiphistory ·92257036d7b8aurora-wipaurora-wiphistory ·112ce3f771bbaurora-wipaurora-wiphistory ·c301767aedf6aurora-wipaurora-wiphistory ·44ed24a040e0aurora-wipdbc3e87b1c5aaurora-wipaurora-wiphistory ·7f5fe6f3f874aurora-wipaurora-wiphistory ·c9a55efdff8faurora-wipaurora-wiphistory ·026b4131cf05aurora-wipaurora-wiphistory ·510ac224fe96aurora-wipaurora-wiphistory ·45acf2644990aurora-wipaurora-wiphistory ·414632095d26aurora-wipaurora-wiphistory ·918c0ff6270caurora-wipaurora-wiphistory ·e60f880ec635codex/neo-pcie-root-bringup1aac0a61a944aurora-wip53d1304629a6aurora-wipaurora-wiphistory ·b8fef4b5b0b5aurora-wipg17/andhw/driver trees are identical toaurora-wipat27178c4e, so the code landed through direct commits ·d128b5e48465aurora-wipaurora-wiphistory ·8b60159e5cb3aurora-wipaurora-wiphistory ·1472fb29a8b9aurora-wipaurora-wiphistory ·936ad7d7d7a4aurora-wipaurora-wiphistory ·d8ad51f6b8c8aurora-wipaurora-wiphistory ·d83758d6c37caurora-wipaurora-wiphistory ·5de43d36cd52neo-gpua6f8817ebbf6neo-gpu66db357bd614aurora-wipaurora-wiphistory ·63a914f4e0b0aurora-wipaurora-wiphistory ·df7801c436caaurora-wipaurora-wiphistory ·35c7d87921c0aurora-wipaurora-wiphistory ·ef4ae11a6545aurora-wipaurora-wiphistory ·64fd324d3345aurora-wipaurora-wiphistory ·2f19bd0c526eaurora-wipaurora-wiphistory ·f9ee7e25bd2daurora-wipaurora-wiphistory ·4f3a64e7baa9aurora-wipaurora-wiphistory ·d41b89bdd071aurora-wipaurora-wiphistory ·f49c93d94e9eaurora-wipaurora-wiphistory ·3fecb279810daurora-wipaurora-wiphistory ·ccedac0c6a30aurora-wipaurora-wiphistory ·18837942fa8daurora-wipaurora-wiphistory ·f215db074fc2aurora-wipaurora-wiphistory ·f01a59d7cc8aaurora-wipaurora-wiphistory ·1702229de78faurora-wipaurora-wiphistory ·db29a8dd8ee0aurora-wipaurora-wiphistory ·a4975426272faurora-wipaurora-wiphistory ·c2e653555074aurora-wipaurora-wiphistory ·c679c8a75a89aurora-wipaurora-wiphistory ·259ee1d76704aurora-wipaurora-wiphistory ·8f0900acde8aaurora-wipaurora-wiphistory ·17e4f0f47acdaurora-wipaurora-wiphistory ·803274a7ca61aurora-wipaurora-wiphistory ·cf6198d365e9aurora-wipaurora-wiphistory ·d588f760941eaurora-wipaurora-wiphistory ·1b6f65959749aurora-wipaurora-wiphistory ·5f6c917f2c82aurora-wipaurora-wiphistory ·a490adab6fc3aurora-wip94ae820025f1aurora-wip4152bb971d6eaurora-wipaurora-wiphistory ·5cf6579fa592aurora-wipaurora-wiphistory ·8d777a4df56caurora-wipaurora-wiphistory ·97075ac83279aurora-wip600b09505597aurora-wipaurora-wiphistory ·a5f49f9eff3aaurora-wipc7b545d062a5aurora-wipaurora-wiphistory ·97196e94cb45maindba15302aa05agx3bf2c0805dc5a2agx3bcc5cb4e13947feat/neoabde4b6cc99bfeat/neo58c4d631b015aurora-wipaurora-wiphistory ·15204343ccd6j700aurora-wiphistory ·72f3265bfc55PR state, source integration and hardware qualification are separate. A closed PR is not necessarily merged. Public reports describe their original test kernels, including contributor branches. Repository administrators can edit this opening post.
All reactions