Repository navigation
Merge integration/aurora ad21b1b0 into aurora-wip - #1175
Merged
Merged
Conversation
Read the boot-populated /product artwork-display-gamut only when the paired version-1 primary calibration matrix has a valid bounded shape. Retain only the standard sRGB or Display P3 target: DCP already owns the per-panel matrix at its factory location, so never submit it as a DRM CTM. Use the raw target for H17P and desktop firmware while keeping complete firmware chromaticities authoritative. Replace the short-lived board-DT gamut fallback and cover P3, sRGB, bounds, provenance, firmware precedence and both publication paths with synthetic fixtures. Assisted-by: LLM Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com> (cherry picked from commit 3f0394c)
The J713 keeps its SoC-level wiring in its own board files. Put it in shared files for the other M4 Macs instead of copying it per board: - t8132-board.dtsi: the loader handoffs (AOP, PMP, MCC, PMC), CPU DVFS and deep idle, the boot NOR and its NVRAM partition, the SMC (RTC, reboot, hwmon), the PMU's NVMEM cells, the SEP mailbox, the radio port with its managed streams, and ATC0/ATC1 in host mode; - t8132-isp.dtsi: the ISP firmware placeholders and power domains, and the 12 MP front camera's modes; - t8132-apcie-port2.dtsi: the built-in PCIe controller's third port and its DART (the J604 SD reader, the J773g Ethernet), disabled until the managed radio contract admits a second port; - t8132-gpu.dtsi, t8132-dcp.dtsi, t8132-dcpext.dtsi: the G16 GPU and the internal and external display coprocessors, disabled until a board's firmware and panel handoff are qualified; a running display DART is then never touched. Values are from the J713, J604 and J773g ADTs, which agree. The J713 does not include these files yet, so its tree is unchanged. Signed-off-by: Diljit Singh <hey@djdev.me>
The M4 has four ATCs, but t8132.dtsi describes two. Add the third and fourth ATC PHYs, their display crossbars, DWC3 controllers and DARTs, and the second SPMI bus for port controllers (the J604's right-hand port), all disabled, in a file of their own that boards include. Addresses and interrupts are from the J604 and J773g ADTs (atc-phy2/3, usb-drd2/3, dart-usb2/3, atc2/3-dpxbar, nub-spmi-a1). Signed-off-by: Diljit Singh <hey@djdev.me>
The J713's Wi-Fi/Bluetooth reg-on is the SMC's gP13 written with a fixed
template, not with the generic output command. The J604 and J773g ADTs
have the same /amfm function-reg_on, pKW4('gP13', 0x800000), so give
them the same path. The M4 iMacs (J623, J624) have no ADT yet and are
assumed to match.
The selftest compiles the board list and keeps the J713 in it.
Signed-off-by: Diljit Singh <hey@djdev.me>
Add the J604, J623, J624 and J773g to the boot NOR table with the J713's controller, NVRAM window and part. The J713 and J773g ADTs place anvram at 0x700000 for 1 MiB; the J604's 27.0 ADT dump omits the offset and the iMacs have no ADT yet, so those take the standard window. A different part fails the JEDEC check, so no program or erase is admitted there. Signed-off-by: Diljit Singh <hey@djdev.me>
On the J713 a T8132 DWC3 enters host mode before core init and gets the CIO setup before xHCI starts. The J604, J623, J624 and J773g have the same ATCs and controllers, so give them the same sequence. The selftest compiles the board list and keeps the J713 in it. Signed-off-by: Diljit Singh <hey@djdev.me>
The J713's radio port runs a managed contract: Linux adopts the trained link and owns only the radio's two ordinary streams. The J604 and J773g ADTs map the same two streams on the same port (dart-apcie0 mapper-apcie0-wlan and -bt, pci-bridge0), so admit the profile there. The M4 iMacs (J623, J624) have no ADT yet and are assumed to match. Signed-off-by: Diljit Singh <hey@djdev.me>
The managed radio DART is admitted only with its board's profile. Admit it on the J604, J623, J624 and J773g, whose radio ports match the J713's (see the matching PCIe change). Signed-off-by: Diljit Singh <hey@djdev.me>
iBoot leaves the MTP running with its DART streams in bypass, and on the J713 Linux keeps that bypass instead of resetting the DART under the running firmware. The J604 has the same MTP (ADT dockchannel-mtp and mtp nodes as the J713's), so retain it there too. The log line now names the SoC; the radio selftest finds its probe slice by it. Signed-off-by: Diljit Singh <hey@djdev.me>
On the J713 the helper takes over the MTP session iBoot leaves running instead of restarting it. The J604 has the same MTP, so preserve its session too. The legacy loader contract stays J713-only. Signed-off-by: Diljit Singh <hey@djdev.me>
Add the M4 MacBook Pro 14-inch (J604) speaker and jack cards, and the M4 iMacs' (J623, J624) jack cards. Signed-off-by: Diljit Singh <hey@djdev.me>
The J604 has the J614s's six SN012776 amps in the same AOP-fed frame and chassis, and like the J614s its speakers have had no bring-up and have no protection profile: it takes the J614s configuration, which never binds the speaker link. Its jack is the J616s's AOP-fed CS42L84. The M4 iMacs (J623, J624) get jack-only cards on the same AOP back-end; their speakers are not described yet. Signed-off-by: Diljit Singh <hey@djdev.me>
Describe the MacBook Pro 14-inch (M4, 2024) like the J713, from the J604 ADT (macOS 27.0): - the shared T8132 wiring: NVMe, SMC, PMU, AOP, PMP, CPU DVFS and deep idle, the boot NOR's NVRAM, the radio port (board type "guam"); - three USB-C ports: port controllers hpm0 and hpm1 on nub-spmi-a0 and hpm3 on nub-spmi-a1, ATC0, ATC1 and ATC3 in USB2/USB3 host mode; - the keyboard and trackpad (MTP, reset lines gP1c and gP1d), the keyboard backlight (FPWM1) and Touch ID (Mesa on spi2); - the six SN012776 speaker amps in ADT provider order with their sense slots and amp-tx-zd-config idle masks, held disabled with their card until the speakers' bring-up, and the CS42L84 jack as its own card; - the front camera, the fans and SMC sensors, and the panel's size for simpledrm; - the SD card reader (GL9755, power gP19) on the PCIe port 2, disabled with that port. The GPU, the display coprocessors, USB4 and HDMI (ATC2 DisplayPort to the PCON at i2c0 0x08) are described or noted but stay off until their drivers admit this board. Signed-off-by: Diljit Singh <hey@djdev.me>
Describe the Mac mini (M4, 2024) like the J713, from the J773g ADT (macOS 26.6.2): - the shared T8132 wiring: NVMe, SMC, PMU, AOP, PMP, CPU DVFS and deep idle, the boot NOR's NVRAM, the radio port (board type "sakhalin"); - the three rear USB-C ports: port controllers hpm0, hpm1 and hpm3 on nub-spmi-a0, ATC0, ATC1 and ATC3 in USB2/USB3 host mode; - the two front USB-C ports: the hub behind ATC2 at high speed, since the PHY only enters USB3 mode from a Type-C mode switch; - the SN012776 speaker amp (disabled) and the CS42L84 jack codec; - the fan and SMC sensors; - the Gigabit Ethernet on PCIe port 2, disabled with that port. There is no sound card yet: the AOP audio component always runs a microphone stream, and this Mac has neither microphones nor speaker sense. The HDMI port stays on dispext0 as iBoot left it. Signed-off-by: Diljit Singh <hey@djdev.me>
Describe the iMac (24-inch, 2x USB-C, M4, 2024) like the J713, with the parts it shares with the 4-port model in t8132-j623-j624.dtsi. There is no ADT for these boards: the layout follows the M3 iMacs and the T8132 wiring the J713, J604 and J773g ADTs share, and every value from neither is marked as guessed: - the jack codec's interrupt (AP GPIO 98) and inverted tip sense; - the AOP's jack DMA channels; - the port controllers at nub-spmi-a0 0xa and 0xc; - the Ethernet on PCIe port 2 (disabled with that port); - the 12 MP camera's modes, as on the other M4 Macs. The radio's board type is left to the boot loader, and the speakers are not described until their amps are known. Signed-off-by: Diljit Singh <hey@djdev.me>
Describe the iMac (24-inch, 4x USB-C, M4, 2024): the shared iMac parts and the third and fourth USB-C ports on ATC3 and ATC2, in USB2/USB3 host mode. With no ADT, their port controllers are guessed: ATC3's at nub-spmi-a0 0x8 as on the Mac mini, ATC2's at nub-spmi-a1 0xc as the MacBook Pro's extra port. Signed-off-by: Diljit Singh <hey@djdev.me>
The 14-core M3 Max is the T6031 die with fewer cores enabled, and its J514m and J516m MacBook Pros share the J514/J516 chassis and panels. Unlike the T6031 boards they had no DCP, display DARTs, mailbox, display subsystem or panel node, so simpledrm assumed 96 dpi for their boot framebuffer. asahi-wip describes the display on T6034 too. Include t6031-dcp.dtsi in both boards, add the J514/J516 panel nodes with the values the T6030 and T6031 boards use, and link the boot framebuffer to its panel as they do. Every display node stays disabled: the display gate only acts on apple,t6031 machines. Signed-off-by: Diljit Singh <hey@djdev.me>
The M3 GPU thermal limit needs the SMC die temperature zone, which macsmc-hwmon builds from apple,soc-die-keys: on every T8122 board, and together with the PMP hotspot on the J514S and J516S. Only the J613, J516S, J514C and J516C described the keys, so the J433, J434, J504, J615, J514S, J514M and J516M had no die zone. The keys are the SoC's, so give each board the list its SoC already has: - T8122: move the J613 list and its thermal pressure key to t8122-jxxx.dtsi, which every T8122 board includes; - J514S: the J516S list (T6030); - J514M, J516M: the T6031 list, the same die. Keys the firmware lacks are skipped, and a missing pressure key is ignored with a warning, so a key that is absent on one of these boards costs nothing. The J613 output is unchanged. Signed-off-by: Diljit Singh <hey@djdev.me>
The boot loader finds the internal keyboard through the "keyboard" alias and fills in its hid-country-code and apple,keyboard-layout-id from the ADT; the M2 and M4 MacBooks have the alias. The M3 MacBook Air and Pro boards had the keyboard node but no alias, so every M3 keyboard reported country code 0 and hid-apple never applied its ISO key swap. Add the alias to t603x-j514-j516.dtsi for the six J514 and J516 boards, and to the J504, J613 and J615. Signed-off-by: Diljit Singh <hey@djdev.me>
The J613 got Type-C display routes for dcpext in 2e05454 ("arm64: dts: apple: t8122-j613: Wire dcpext to both USB-C ports"); the J615, the 15-inch MacBook Air on the same SoC, had none, so it had no external display. The routes depend only on SoC wiring that both Airs share: t8122-usbpd-spmi.dtsi connects the left-back port to ATC0 and the left-front port to ATC1 on both, dcpext sits on display crossbar ufp1 port 0 (state 2) of either ATC, and its DPTX target is the ATC index. Give the J615 the J613's dcpext alias, routes and connector endpoints. As on the J613, dcpext, its routes, DARTs, mailbox and both crossbars stay disabled; the display gate enables them only after the boot loader has marked the dcpext firmware ready and the internal display was handed over. Not yet tested on a J615. Signed-off-by: Diljit Singh <hey@djdev.me>
The M3 iMacs have microphones and an ambient light sensor behind the AOP, as the M1 iMacs (J456, J457) and every other T8122 Mac do, but their device trees left the AOP, its mailbox, DART and ADMAC disabled, so neither was available. Enable them and name the chassis for the AOP audio card, as on the M1 iMacs. The AOP and its services are the SoC's, so nothing here is board-specific beyond the chassis name; not yet tested on an M3 iMac. Signed-off-by: Diljit Singh <hey@djdev.me>
The M1 iMacs are the only fan-cooled M1 and M2 Macs without fan hwmon sensors: the Mac minis include hwmon-mini.dtsi, the MacBook Pros and Mac Studios hwmon-fan.dtsi or hwmon-fan-dual.dtsi. The 4-port J456 has two fans and the 2-port J457 one; the counts are from the product spec, not an ADT. The F0/F1 keys are the same on every Apple silicon Mac in this tree, and macsmc-hwmon skips a fan whose keys the firmware lacks, so a wrong count only hides or drops a channel. Signed-off-by: Diljit Singh <hey@djdev.me>
apple,soc-die-keys lists SMC keys of the SoC, not of the board, but it was repeated in each board file: the same T6030 list in the J514S and J516S, the same T6031 list in the J514C, J516C, J514M and J516M, and the T8122 list in t8122-jxxx.dtsi beside the SoC's own CPU temperature channels. Move each list to its SoC: t8122.dtsi gets it in smc_hwmon next to those channels, t6030.dtsi and t6031.dtsi as an override at the end. t6031.dtsi rather than t6031-die0.dtsi, which the T6032 shares. The T6034 boards include t6031.dtsi, so they keep their list. No change in any dtb. Signed-off-by: Diljit Singh <hey@djdev.me>
The J613 and J615 were two full board files that differed only in the panel, the Wi-Fi/BT board type, the speakers and their sound card, the AOP chassis name, the firmware names and the J613's thermal zones. The keyboard backlight, AOP, jack codec, MTP, camera, Touch ID, hwmon and the dcpext Type-C routes were copied in both. Move the shared parts to t8122-j613-j615.dtsi, as t600x-j314-j316.dtsi does for the M1 Pro/Max MacBook Pros, and label the MTP multi-touch node mtp_mt so the boards set its firmware the way the J514/J516 do. Drop the boards' framebuffer panel links: t8122-jxxx.dtsi already sets the same ones. No change in any dtb other than property order; dtc -s output of every board is identical. Signed-off-by: Diljit Singh <hey@djdev.me>
The two M3 iMacs differ only in the model, the panel compatible, the Wi-Fi/BT board type, the AOP chassis name, the sound card name and the J434's second PCIe port. Everything else (Ethernet, AOP, the jack codec and its card, the camera and the fans) was copied in both. Move the shared parts to t8122-j433-j434.dtsi and label the sound node so the boards only name the card, as the J514/J516 boards do. No change in any dtb other than property order; dtc -s output of every board is identical. Signed-off-by: Diljit Singh <hey@djdev.me>
Several T8122 comments were paragraphs: measurement procedures, curve fits, boot loader and gate hand-off narratives and crash addresses, up to 27 lines above a node. That history is in the commits that added those nodes. Cut each to the facts a reader of the node needs, in a line or two, and put the text of the 25G83 comments that started on the /* line on lines of their own. Use the mtp_mt label in the 25G83 boards. No change in any dtb. Signed-off-by: Diljit Singh <hey@djdev.me>
The T6030, T6031 and T6034 additions carried the same kind of paragraphs as the T8122 ones: boot loader function names, gate command-line behaviour, register maps and IRQ orders spelt out, and lines past 100 columns. That history is in the commits that added the nodes. Cut each to the facts a reader of the node needs, give t6030-gpu.dtsi the header the other GPU files have, move the T6030 GPU's status to the end of the node and drop two stray blank lines. No change in any dtb other than property order; dtc -s output of every board is identical. Signed-off-by: Diljit Singh <hey@djdev.me>
t8122-jxxx.dtsi said its GPU memory nodes were filled by an opt-in diagnostic exporter; the boot loader fills and enables them when it hands the GPU over, as t8122-gpu.dtsi says. A "To be filled by loader" comment after the last of them described nothing. t603x-j514s-j516s-usb4.dtsi described itself as loader aliases only; most of it is the Type-C display routes and dcpext1's audio. No change in any dtb. Signed-off-by: Diljit Singh <hey@djdev.me>
Add resource descriptions and bindings for unclaimed media, memory, security, trace, PCIe DMA and DP input interfaces on T8122, T6030 and T6031. Preserve disabled status for these interfaces and their translation banks. Describe the six memory channels on the harvested T6034 die. Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com> Assisted-by: LLM
Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com> Assisted-by: LLM
Assisted-by: LLM Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com>
Reuse the generic radio PMU and Stockholm SPMI bindings for the populated followers. Both endpoints and their parent controllers remain disabled. Assisted-by: LLM Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com>
Keep coverage manifests and their describe-only DTS prerequisite in sync; the standalone DTS change is reviewed separately before the coverage checker lands. Assisted-by: LLM Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com>
T6040 USB4 resource description from linux#1052. Only DT and binding description content is reused; no PCI or Thunderbolt driver changes are included. Keep the tunnel containers explicitly disabled along with their physical child devices. Assisted-by: LLM Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com>
Normalize the aliases in the description from linux#1052 to the lowercase, digits and hyphens required by the core schema. Node labels and all resource references stay the same; every router and PCIe tunnel remains disabled. Assisted-by: LLM Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com>
Assisted-by: LLM Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com>
Use live J616s and Neo ADTs and live M1 Pro and M2 IODeviceTrees. Remove pruned template instances and runtime IOKit enumeration mirrors. Resolve every retained record to a DT node or an exact reasoned waiver for a virtual vote, firmware service view, or logical relationship. Keep the generator and source audits private. Assisted-by: LLM Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com>
Merge train 11-08: J6AM's MSR try2 one-shot Stage 2 trial on the J616s (fleet27) ran exactly these four status flips: scaler probed (hardware version 21), NV12/P010 <-> ARGB/XRGB within 1 LSB at 1.3-1.9 ms per 1080p job, no DART faults, runtime suspend and rmmod clean. The built j616s DTB differs from integration by these four lines only. dt-checks (0 new) and host-tests green.
Retain the integration scaler enable and the disabled retimer descriptions. No new idle-state links or device enables are introduced relative to integration. Assisted-by: LLM Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com>
Preserve the integration scaler status and the separate disabled fleet resource descriptions. Assisted-by: LLM Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com>
Assisted-by: LLM Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com>
Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com> Assisted-by: LLM
Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com> Assisted-by: LLM
Preserve the published integration merge and all local live inventory work. The source tree is identical to the preceding validated candidate; this is a history-only merge for a fast-forward push. Assisted-by: LLM Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com>
Merge train 11-08: DT2/DT3 ready (apple-dt-check 0 new). We have no T602x hardware, so the gate is no runtime change. On the merged tree the new nodes are all disabled, and the only enabled-set difference is the main PMIC's node name, spmi pmic@f -> pmic@b. Its content is identical (apple,maverick-pmic, reg 0x0b, the same nvmem cells), so the unit address now matches reg. W=1 equals base on j414s/j416c/j475d/j180d. CI dt-checks + host-tests green.
The T6040 AVD is IO-coherent in both directions, measured the same way as on T8112 on a J616s: with every cache operation skipped and one picture in flight, CPU-dirtied and CPU-cached CAPTURE buffers and CPU-written bitstream buffers queued without a clean all decode bit-exact. T6041 uses the same decoder block and node. On M4 Pro the CPU already reads the uncached mapping fast (about 50 GB/s for 1080p frames), so this changes no measured speed. It describes the hardware correctly, leaves the DART mappings cacheable and needs no cache maintenance. J616s, same boot, avd-bench, 16 pictures in flight, 3 passes, every frame checked against software decode (stock -> dma-coherent -> stock): h264 1080p 1865-1893 / 1866-1891 / 1865-1894 fps hevc 1080p 2584-2610 / 2563-2594 / 2582-2603 fps h264 2160p 479-482 / 478-480 / 478-479 fps hevc 2160p 689-699 / 689-698 / 689-698 fps vp9 1080p 874-879 / 874-879 / 874-878 fps av1 1080p 2278-2308 / 2285-2320 / 2294-2315 fps Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com> Assisted-by: LLM
Retain the describe-only fleet additions on the #1149 integration tip. No driver changes, new enables or CPU idle-state links are added relative to integration. Assisted-by: LLM Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com>
Assisted-by: LLM Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com>
Assisted-by: LLM Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com>
Merge train 11-08: DT3/t-5680 ready (apple-dt-check 0 new / 0 removed across 122 DTBs, 19/19 bindings ok). No T8132 hardware here; on the merged tree j713 and j715 gain 50 nodes each, all disabled, with no status changes and W=1 equal to base. CI green.
Merge train 11-08: DT3/t-5680 ready (apple-dt-check 0 new / 0 removed, 15/15 bindings ok). No T6032 hardware here; on the merged tree j575d gains 81 nodes, all disabled, with no status changes and W=1 equal to base. Clean on the tip after #1131. CI green.
Merge train 11-08 (ORCH). Neo (GPUX): refused-queue spins during the 2 s wait 2.0M -> 12, one-shot clean, Enforcing, 0 AVC. J616s (test kernel 7.3.0-aurora-mtrl3 = integration 95e30d5 + this + #1147 + #1105, one-shot, FLEET4 reg.sh vs fleet27): kerr 0, kwarn classes = base, dEQP-VK timeline_semaphore 3835 Pass = base, 0 GPU faults, gpubench and glmark within fleet27's own spread. Reviewed: a refused creator now holds nothing, and render_fits(open + 1) is the same bound as before. Host-tests CI green.
…power Merge train 11-08 (ORCH). Neo (BOOTLOG2): 4/4 boots, no regression. J616s (mtrl3 one-shot with this, FLEET4 reg.sh vs fleet27): kerr 0, AOP-fed paths (v2 audio grant, sleeps x12 regrant, ANE) equal to base. It waits at most 1 s, and only when the endpoint map's last fragment is late. CI green.
Conflict in apple,avd.yaml: integration moved the AVD properties into per-compatible if/then branches. Take integration's file and add dma-coherent to the branch that covers the T8112 and T604x decoders. Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com> Assisted-by: LLM
…burn-down-1111-follow-up-fleet-mani Merge train 11-08: DT3 ok (0 new warnings, 7/7 bindings ok). On the merged tree the fleet DTBs gain only disabled nodes (j616s +33, j316s +13, j413 +17, j700 none), with W=1 equal to base. CI dt-checks + host-tests green.
…verage-tool-ci-check-every-adt-devi Merge train 11-08: DT3 ok. Adds the ADT-vs-DT inventory coverage ratchet to dt-checks CI (scripts/dtc/apple-adt-coverage.py with tests, per-board coverage lists, waivers). On the merged tree the j616s, j316s, j413 and j700 DTBs gain no nodes and nothing changes status, and W=1 equals base. Clean on the tip. CI dt-checks + host-tests green.
Merge train 11-08: AVX2 proved the AVD IO-coherent on the M2 (T8112) and the J616s (T6040) with avd-bench and a cache-skipping ioctl shim. Dirty CPU lines are replaced, stale lines aren't read, uncleaned bitstreams decode, and every row is bit-exact. A same-boot stock/coherent/stock module A/B gives M2 decode+readback 318 -> 1718 fps (1080p) and 81 -> 442 fps (4K), J616s equal, decode alone unchanged, 0 DART faults, 0 AVC. Merged up to integration 634b402 (the binding moved into per-compatible branches; dma-coherent sits in the T8112/T604x one). W=1 j413/j616s equal base. CI dt-checks + host-tests green.
Brings aurora-wip up to the 11-08 morning integration tip: DT descriptions for every Mac family (#1094, #1097, #1111, #1131, #1132, #1148, #1149, #1155, #1156, #1157, #1172), the J616s scaler (#1168), the panel primaries from boot data (#1105), AVD macOS command parity (#1116), AVD dma-coherent (#1169), the M2 PMU RTC alarm (#1151), G17 admission without spinning (#1146) and the AOP map wait (#1147). Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com> Assisted-by: LLM
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Brings aurora-wip up to integration/aurora ad21b1b with one clean merge commit (e78e14a, tree 02763b01).
Integration added since #1159: device-tree descriptions for every Mac family, the J616s scaler, panel primaries from boot data, AVD macOS command parity and dma-coherent, the M2 PMU RTC alarm, G17 admission without spinning and the AOP endpoint-map wait. Each one went through the 11-08 merge train.
Gate (buildbox, exact merge tree): fleet config Image + modules + dtbs, run-host-tests, apple-dcp host selftests. Results are posted before merging.
🤖 Generated with Claude Code