Repository navigation
Merge integration/aurora 43c790fe into aurora-wip - #1093
Merged
Merged
Conversation
…ssword After a few failed matches the enclave answers every MATCH_RESULT with status 0x1, and on macOS that is the point where it asks for the password. Here the driver reported it as a refused enclave step, libfprint as "the secure enclave refused a step", and Touch ID failed until a reboot, because nothing ever showed the enclave the secret again. Tell the lockout apart. A MATCH_RESULT of 0x1 now ends the verify with its own status, SEP_BIO_STATUS_LOCKOUT (8), and every verify after it is refused at once, without a capture, so the user is not asked to touch a sensor whose match would be refused anyway. The log says how many matches failed first. Re-arm it the way macOS clears it with the passcode. A new root-only node, /dev/sep-bio-rearm, takes SEP_BIO_LOCKOUT_STATE and SEP_BIO_REARM. fprintd holds /dev/sep-bio open, so the lockout needs a node of its own. REARM needs CAP_SYS_ADMIN and does nothing without a lockout; with one, it sets up the SCRD match credential again (SKS verifies the stored keybag secret into a new context) under the lock that verify, enrolment and deletion hold, and clears the lockout only when that succeeded. The driver never does this on its own: re-establishing the credential on the enclave's refusal would undo the lockout. Userspace calls it only after the user's password has been verified (pam_aurora_touchid and aurora-touchid-rearm, packaged with libfprint's aurora driver). Keep the lockout across a restart. The enclave forgets it (the lockout record restores as EmptyTolerated), and macOS covers that by asking for the passcode after every restart. Linux may offer Touch ID at the login screen, so a restart would hand out a fresh set of attempts. The driver records the lockout in /var/lib/aurora-sep-touchid-lockout.bin and keeps refusing after the next boot until a password re-arms it. A missing or unreadable record is no lockout: only root can remove it, and root can restart anyway. The interface version stays 5: the verify poll is unchanged, and older libfprint builds report status 8 as an unknown driver status. Assisted-by: LLM Signed-off-by: Ryan Murray <ryan@aurorasilicon.org>
Re-arming after a password set up a new SCRD match credential, but the enclave kept refusing matches: MATCH_RESULT still answered 0x1. The enclave ends a biometric lockout when the user's key bag is unlocked with its secret. Setting up a match credential does not do it. So REARM now follows the same order as a passcode unlock. It unlocks the designated bag with the stored secret, as activation already does before the restore. Then it sets up a new SCRD match context, and then saves the enclave's lockout record so that a restart does not restore the record from before the lockout. The lockout is cleared only when the unlock and the credential both succeed. The boot path's special unlock moves into a helper that both paths share. Assisted-by: LLM Signed-off-by: Ryan Murray <ryan@aurorasilicon.org>
Provision a Linux-owned KEK/VEK record pair on the first password re-arm, persist it atomically in the host-state store, and require an option-2 FV_UNWRAP_VEK before refreshing SCRD and saving lockout state. Carry the already-verified password through the root-only re-arm ioctl and wipe the request and returned opaque key. Add KUnit coverage for the three request layouts and persistent pair format. Assisted-by: LLM
On the MacBook Neo the password re-arm provisions its Linux-owned FV pair: FV_NEW_KEK (0x42) and FV_NEW_VEK (0x40) both answer status 0, and the option-2 FV_UNWRAP_VEK (0x41) of the new pair answers status 0 too, with a 0x98-byte reply image. That reply carries the same 64-byte opaque key as a volume unwrap. The re-arm decoder required a 32-byte key, read from the shorter macOS reply, so it threw away a successful unwrap as EMSGSIZE. The pair was never committed, every password re-provisioned a new one, and the caller reported only that the stored secret had not unwrapped the pair. Check the reply's shape instead of a fixed length: version 0, one complete key blob of 1 to 64 bytes, flags bit 0 and nothing after. The key is never copied out of the reply buffer, which is wiped on drop. Log each failure stage: the SKS status of a refused unwrap, a missing reply, the lengths and flags of an unusable reply (never key bytes), the errno seen by the sbio caller, and a missing or unreadable identity bag. A stored pair that the enclave refuses, or one that no longer decodes, is replaced by a newly proved pair instead of failing for good. Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com> Assisted-by: LLM Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Add CONFIG_EXECMEM_CONFINE: an optional backing mode for execmem ranges flagged EXECMEM_CONFINE. Such ranges keep their virtual placement (so a module stays within branch range of the kernel) but draw their physical backing pages from a single fixed window registered by the architecture via execmem_confine_init(), rather than from arbitrary buddy pages. This lets a platform keep every run-time executable physical page (modules, BPF, kprobes, ftrace) inside a hardware-enforceable code region such as an Apple-silicon CTXR window, which is a prerequisite for running the kernel at that privilege level. A confined allocation reserves its virtual area with __get_vm_area_node() inside the range, pulls a physically-contiguous chunk from a gen_pool over the window, and maps it with vmap_pages_range(); the backing pages are published on the vm_struct so set_memory_rox() can mirror permission changes onto their linear alias. If the pool is exhausted the allocation fails, with no buddy fallback, mirroring a hardware execute-range fault. A verifier sweeps for any executable mapping that resolves outside the window (the page-table sweep itself is provided per-architecture); when the window is enforcing a violation is fatal. "ctxr_exec_selftest" adds a boot-time positive/negative self-check. Signed-off-by: Diljit Singh <hey@djdev.me>
On T8140/M4-class Apple-silicon machines the boot firmware arms a per-CPU
code text execute range (CTXR) on the secondary cores: at the privileged
exception level an instruction may only be fetched from inside a fixed
physical window. To run the kernel there rather than dropping it to EL1,
every executable physical page must live inside one window.
Add CONFIG_ARM64_CTXR_EXECMEM, the arm64 end of the EXECMEM_CONFINE
contract:
- ctxr_exec_reserve(), from arm64_memblock_init(), learns the window
from /chosen/aurora,ctxr-exec-window (emitted by the EL2 monitor) or
the ctxr_exec_window= command line (for VM bring-up), checks that the
kernel image fits, and reserves it from memblock.
- execmem_arch_setup() carves the non-image remainder into a gen_pool,
registers it with the confinement core, and flags the executable
ranges (module text, BPF, kprobes and, by inheritance, ftrace) with
EXECMEM_CONFINE. Module data is left on ordinary vmalloc backing.
- execmem_confine_arch_verify() provides the page-table sweep used by
the verifier: every privileged-executable leaf outside the window is
reported, standing in for the hardware fault when testing in a VM.
The option is inert unless a window is supplied at boot, so it is enabled
in asahi.config and is a no-op on machines that do not need it.
Signed-off-by: Diljit Singh <hey@djdev.me>
…ndow Build the executable pool only from memory that was free before the window was reserved. Snapshot for_each_free_mem_range() over the window just before memblock_reserve(), so anything already reserved at that point -- the image, the device tree blob, an initrd, /reserved-memory and memreserve entries, EFI runtime regions and the UEFI memory map -- and any hole or no-map range stays out of the pool. Previously the pool was the window minus the image only, so a lowest-first executable allocation could overwrite a live DTB, an initrd or EFI runtime memory that a loader had placed inside the window. When no base is given and anchoring the window on the image would run past the end of DRAM (a loader placed the image within the window size of the top), slide the window down to end at DRAM's end, still covering the image, instead of panicking. List the pool's ranges in /sys/kernel/debug/execmem_confine. Signed-off-by: Diljit Singh <hey@djdev.me>
The kernel image and the code of every module take about 66 MiB, and RAM in the window that the pool doesn't use is lost to the rest of the system. 128 MiB holds them with room to spare. Signed-off-by: Diljit Singh <hey@djdev.me>
Picks up the bindgen fix (#992), without which the asahi profile does not build. Signed-off-by: Diljit Singh <hey@djdev.me>
Kernel text inside an armed CTXR window can't execute at EL1 on a core the window is enforced on, so the usual EL1 detour on entry faults. Once the boot CPU has found an armed window, cores entering after it switch to VHE in init_kernel_el and never run at EL1. Signed-off-by: Diljit Singh <hey@djdev.me>
Kernel text inside an armed CTXR window can't execute at EL1 on a core the window is enforced on, and the window isn't known when the boot CPU first enters. Apple cores are VHE only, so with CONFIG_ARM64_CTXR_EXECMEM they now switch to VHE in init_kernel_el on every entry, boot CPU included, and never run at EL1. This replaces the flag the boot CPU set for later cores. Signed-off-by: Diljit Singh <hey@djdev.me>
The profile had no dm-crypt, so a kernel built from it can't unlock an encrypted root. Add it with the crypto interfaces LUKS uses, and the modules an Aurora install's services load. Signed-off-by: Diljit Singh <hey@djdev.me>
The vDSO's code is part of the kernel image, which sits in the CTXR window, and EL0 can't execute there on a core the window is enforced on. A process calling into the vDSO on such a core takes a permission fault that the kernel finds nothing to fix, and retries it forever. Map a copy from ordinary memory into userspace instead. Signed-off-by: Diljit Singh <hey@djdev.me>
A core that enforces the CTXR window doesn't let EL0 execute from it either. The fetch takes a permission fault that the page tables don't explain, so the kernel retries it forever. The window is reserved, but free_reserved_area() can still release parts of it: __init memory, which is part of the image, and an initrd placed in the window. Those pages reach the page allocator and can become user pages. Keep them reserved, and have the verifier count any window page that isn't reserved as a violation. Signed-off-by: Diljit Singh <hey@djdev.me>
With CONFIG_ARM64_CTXR_EXECMEM every Apple core switched to VHE in init_kernel_el, whether or not it has CTXR and whether or not a window was supplied, so a kernel built with the option took a different entry path on M1 to M3. Only M4-generation and later cores have CTXR. Their part numbers start at 0x50, and every earlier core's is below that, so limit the direct switch to them. Earlier cores drop to EL1 and come back through finalise_el2 as before. Signed-off-by: Diljit Singh <hey@djdev.me>
The copy was made whenever CONFIG_ARM64_CTXR_EXECMEM was on. With no window there is nothing to keep the vDSO out of, so map the image's pages as before. Signed-off-by: Diljit Singh <hey@djdev.me>
…t run EFI runtime code runs at EL2 on whichever core efi_rts_wq is on. On a core that enforces the window, code outside it faults, and the fault handler dies with "execution of user memory" before the EFI fixup runs. So when a window is enforcing and any EFI_RUNTIME_SERVICES_CODE region is outside it, warn once and turn runtime services off for that boot. Misaligned code counts too: on a 16K kernel it makes the misaligned regions after it executable, wherever they are, and the verifier would then panic on them. The window is read from this file rather than through execmem_confine_*(), since execmem_init() is a later core_initcall than arm64_efi_rt_init(). Have the verifier sweep efi_mm as well, so its report covers the runtime mappings that stay on. Signed-off-by: Diljit Singh <hey@djdev.me>
On a core enforcing the window, an EL0 fetch from a window page takes a level 3 permission fault that the page tables don't explain. handle_mm_fault() finds nothing to fix, so the task retries forever: it stays runnable, with no panic and no hung-task report. The zero page stays in the window, so a program jumping into fresh executable memory before writing it hits this. When a window is enforcing and such a fault lands on a present, executable mapping of a window page, log the task and kill it. Every other fault takes the path it took before. Signed-off-by: Diljit Singh <hey@djdev.me>
…window The relocation code runs from an ordinary page, outside the window, and a crash dump can start on a core enforcing it, where it would hang during a panic. Fail the load cleanly instead, with one message, so nothing finds out only when it needs the second kernel. Signed-off-by: Diljit Singh <hey@djdev.me>
A kernel in a window runs at EL2 with no monitor below it, so the bootloader is what supplies the window. Say so, and document the entry path. Signed-off-by: Diljit Singh <hey@djdev.me>
ctxr_exec_enforce= was only read along with ctxr_exec_window=, so it did nothing to a window from the device tree. strstr() matched inside other words and took the first use of an option. A value that didn't parse gave a default window, and a window too small for the image panicked a machine that would otherwise boot. Use early_param(), which runs before arm64_memblock_init(). The last use of an option wins, and ctxr_exec_enforce= applies to a window from either place. A value that doesn't parse is ignored with a warning. So is a window from the command line that can't work, which leaves the device tree's window if there is one. A window from the device tree that can't work is still fatal when enforcing, since the kernel can't run on the confined cores outside it. Signed-off-by: Diljit Singh <hey@djdev.me>
The pool leaves out whatever was reserved before the window, but two paths still took window pages to the page allocator, and the verifier then panicked at boot. A reusable CMA region inside the window goes to the page allocator from cma_activate_area(). Don't activate a CMA area that overlaps the window, and keep its pages reserved. A reserved-memory node with a "reg" whose init fails is freed back to memblock from unflatten_device_tree(), after the window was reserved. That leaves a hole for memblock_free_all() to release, so reserve the window again before that. Record the window when it's reserved rather than when the pool is set up, because execmem_init() and CMA activation are both core_initcalls. Signed-off-by: Diljit Singh <hey@djdev.me>
__vmalloc_node_range() unpoisons a VM_ALLOC area once it's mapped and registers it with kmemleak, and vfree() unregisters it. A confined allocation goes around both, so on a debug kernel KASAN sees the area as poisoned and kmemleak doesn't know it. Do the same here. Signed-off-by: Diljit Singh <hey@djdev.me>
Describe the /chosen properties a bootloader uses to give the kernel its executable window. Signed-off-by: Diljit Singh <hey@djdev.me>
Add the command-line options to kernel-parameters.txt and describe how they treat bad values. Say that a confined machine loses CPU hotplug and kexec, and that hibernate is untested under a window. aura doesn't emit aurora,ctxr-exec-enforce, so it's no longer shown as part of what the bootloader emits. The Kconfig help said to say N if unsure while asahi.config turns the option on, so say who wants it. Signed-off-by: Diljit Singh <hey@djdev.me>
…r not Whether a core can fetch from outside the window is decided by the hardware, not by aurora,ctxr-exec-enforce or ctxr_exec_enforce=. Those only decide whether a verifier finding is fatal. EFI runtime services, kexec and the EL0 fetch check keyed on enforcement, so a confined machine booted with enforcement off, now possible from the command line, would run EFI runtime code from efi_rts_wq on a confined core, allow a kexec that hangs, and leave a task faulting forever on the zero page. Key all three on the window being active. Assisted-by: LLM Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com>
empty_zero_page is in the image's .bss, so it is inside the window. Any anonymous mapping that is read before it is written maps it, including PROT_EXEC ones, userfaultfd's UFFDIO_ZEROPAGE and KSM's zero pages, and KVM can map it into a guest's stage 2 for untouched guest RAM. EL0 can't execute from the window on a confined core, so an unprivileged program that jumps into such memory is killed there and gets SIGILL anywhere else, and a guest fetching from it may fault to EL2 again and again. With a window reserved, allocate the zero pages from memblock, outside the window, and clear their MTE tags as cpu_enable_mte() does for empty_zero_page. The fault-path kill stays as a backstop. Assisted-by: LLM Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com>
The direct entry for Apple parts from 0x50 sets E2H without checking ID_AA64MMFR1_EL1.VH. A guest whose virtual EL2 has no VHE, such as a nested KVM guest on an M4 host that sees the host's MIDR, would hang in early boot. Fall back to the usual path when VH is 0. Assisted-by: LLM Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com>
arm64: dts: apple: t6040-j616s: Describe the USB3 path of the Type-C ports
Every Wi-Fi disconnect before system sleep logged "brcmf_msgbuf_delete_flowring: timed out waiting for txstatus", one or two per sleep cycle on BCM4387 Macs (M1 Pro, M2). A kprobe on brcmf_msgbuf_delete_flowring() reading each flowring's outstanding_tx showed two cases, 10 timeouts in 12 sleep cycles: - the count is positive: frames were queued to the AP when the link went down, and the firmware completes them only when it flushes the ring for the delete request, so the wait before that request cannot succeed; - the count is negative: after a timeout the count is set to zero, the late completions then take it below zero, and nothing resets it when the flowid is used again, so the next delete on that flowid waits out its full timeout as well. Reset the count when a ring is created, never decrement it below zero, and report the timeout at debug level with the count: it is expected when the peer is gone, and the ring is deleted either way. Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com> Assisted-by: LLM
…tstanding wifi: brcmfmac: msgbuf: keep the flowring tx count from going negative
drm/asahi: g17: log the MMU fault banks on recovery
drm/asahi: g17: cheaper submits and a deeper compute queue
arm64: ctxr-execmem: follow-ups from the Neo run and a second review
…arms it On the MacBook Neo the enclave refused the third Touch ID attempt after two misses (MATCH_RESULT status 0x1 while the user lock state stayed clear), and no password step ended it: only a restart of the enclave did. A trace of a macOS 26.6.2 Touch ID prompt on the same machine (three wrong touches, the password in the same dialog, then a new prompt with an enrolled finger) shows macOS never reaches that refusal. Its three-try limit is LocalAuthentication's own and the enclave lock state stays clear. The request shapes match Linux; the per-touch sequence does not. Match the way macOS does: - Match in the LocalAuthentication class (match flags 0x88). The login window's keybag-unlock class (0x89) is not needed on Linux. - One match operation (OPERATION_DEVICE_LIST) and one unlock evaluation per prompt, i.e. per open of the Touch ID device, which fprintd claims for one PAM authentication. MATCH_CANCEL only when the prompt ends: on a match, at the retry limit, or when the next prompt, a password re-arm, an enrolment or a suspend ends it. - After a no-match, relay the further frames of the same touch (0x65, each followed by RETRY_CAPTURE_INFO 0x3e), then MATCH_RETRY_FINALIZE (0x0c) before MATCH_POLICY, so the touch counts as one result. - Save the enclave lockout record after every answered result. - Enforce the three-attempt limit on the host: after three failed matches in a row Touch ID refuses without a capture until a password authentication re-arms it, recorded across a restart. Also carried here from the test series: the password re-arm verifies the designated key bag in password mode and sets up the cold-match credential again; matches run under an unlock evaluation whose context is sent as the credential blob; the enclave's background operation results are collected and the catacombs saved after a template update (user, then master); the sensor stays registered between verifies; and MATCH_POLICY is read after every result. Tested on the MacBook Neo (J700) with a finger, twice: three misses in one sudo prompt each answered as a plain no-match, the host limit asked for the password, and the enrolled finger matched right after it without a restart. No-finger lifecycle gate, SELinux enforcing, 0 AVC. Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com> Assisted-by: LLM
soc: apple: sep: Touch ID lockout that a password re-arms, matched like a macOS prompt (tested on the MacBook Neo with a finger, twice).
The interrupt-controller on Apple silicon systems is initialized early via IRQCHIP_DECLARE. The early initialization does not include the usual automatic handling of simple "power-domains" and leaves its power-domain unused. This may result in genpd powering this critical power-domain down resulting in a mostly dead system after "PM: genpd: Disabling unused power domains". Signed-off-by: Janne Grunau <j@jannau.net> (cherry picked from commit 74768a25d95a1ce174c9e5084416f37edb69a93e) Signed-off-by: Diljit Singh <hey@djdev.me>
Add WiFi, Bluetooth, SD-Card reader, ethernet NIC and USB xHCI device nodes for M3 Ultra Mac Studio. The bootloader will fill MAC addresses for the network devices, add Wifi and Bluetooth calibration data and set "apple,antenna-sku". Signed-off-by: Janne Grunau <j@jannau.net> (cherry picked from commit 4aa4ffb27752f05082ec4366496c46c04731d5fa) Signed-off-by: Diljit Singh <hey@djdev.me>
The power-state ps_disp_cpu must not be power-gate as the DCP coprocessor will not wake up end probably needs to be reinitialized. It is unknown how coprocessors started by Apple's bootloader are booted. The power-state dependencies for the display differ on t6031 from all other M1/M2/M3 SoCs. This likely contributed to adding the "apple,min-state" to the wrong power-state. Add a comment explaing this differences. Signed-off-by: Janne Grunau <j@jannau.net> (cherry picked from commit 3267df96fb667aa4a48fe59870a999f9908d67e9) [Diljit Singh: keep "apple,min-state" on ps_disp_fe as well, which this tree already sets; only add it to ps_disp_cpu and the comment] Signed-off-by: Diljit Singh <hey@djdev.me>
The J575D Wi-Fi node is new in this tree. Mark it with brcm,ccode-map-trivial, as commit 7d9a9ae ("arm64: dts: apple: pass the country to the Wi-Fi firmware") did for the other Apple Broadcom Wi-Fi nodes, so brcmfmac passes the regulatory domain to the firmware. Signed-off-by: Diljit Singh <hey@djdev.me>
Signed-off-by: sofus <sofus.c@icloud.com> Signed-off-by: Janne Grunau <j@jannau.net> (cherry picked from commit b22e6f33935d4a5746ddadf69ea8d651870a8e88) [Diljit Singh: drop the J613 and J615 hunks; this tree already has their panel nodes and boot framebuffer links] Signed-off-by: Diljit Singh <hey@djdev.me>
The J434 describes its Ethernet controller, whose MAC address the loader
fills in, but unlike the J433 it has no ethernet0 alias. Add it.
Ported from asahi-wip commit f12cfb01a49c ("arm64: dts: apple: t8122: Add
PCIe device nodes") by Alyssa Milburn; the rest of that commit is
already in this tree.
Signed-off-by: Diljit Singh <hey@djdev.me>
Like ps_disp_cpu, the power states of the two external display
coprocessors should not be power-gated: a power-gated DCP coprocessor
does not wake up again without being reinitialized. The M1, M2 and
T8122 trees already floor their dispext CPU power states at 4.
The display gate still raises these states to active before it enables
an external display processor.
Ported from asahi-wip commit e7d06d083305 ("arm64: dts: apple: Describe T6030
display hardware") by Ramzi Abdiche; that commit's display nodes are
not taken, as this tree describes the T6030 display its own way.
Signed-off-by: Diljit Singh <hey@djdev.me>
With the panel's physical size, simpledrm reports the real DPI of the
internal display instead of assuming 96, so early users pick the same
HiDPI scale as the display driver does later. The J613 and J615 boot
framebuffers already link their panels this way.
Label the existing J514 and J516 panel nodes on the T6030 and T6031
boards and link each boot framebuffer to its panel.
post-init-providers keeps the framebuffer from waiting for the panel,
which has no driver of its own.
The T6034 boards have no DCP or panel node in this tree, so they are
left as they are.
Ported from asahi-wip commit 5d53e284e6b3 ("arm64: dts: apple: t603x-j514-j516:
Add panel nodes and use DCP") by sofus; only the framebuffer link is
taken, as this tree already has its own panel nodes.
Signed-off-by: Diljit Singh <hey@djdev.me>
Add device nodes for display hardware and prepare for using dcp on the integrated display. Since dipslay output is not supported on the M3 Ultra remove the "display-subsystem" node as well. Signed-off-by: sofus <sofus.c@icloud.com> Signed-off-by: Janne Grunau <j@jannau.net> (cherry picked from commit d8fd894089b117a68184b9c9b4f323ebcd1b0c6e) [Diljit Singh: take only the four external display processors on each die (DCP, DARTs and mailbox, all disabled), their clocks and the pmgr_dcp node; this tree describes the internal display in t6031-dcp.dtsi, so its nodes, the display subsystem and the T6032 deletions are left out. Add "ranges;" to each dcpext node, as every DCP node here that sets #address-cells has.] Signed-off-by: Diljit Singh <hey@djdev.me>
The M3 Pro and Max MacBook Pros carry a BCM4388, as the T6032 and every M4 board here already say. Use its Wi-Fi (14e4:4434) and Bluetooth (14e4:5f72) IDs instead of the BCM4387 ones. PCI binds by ID and the node by its address, so nothing changes at runtime. Signed-off-by: Diljit Singh <hey@djdev.me>
arm64: dts: apple: M3: port what asahi-wip has and Aurora lacks
Brings aurora-wip up to the tested release branch: CTXR execmem (#985, #1079), KFIX speaker sleep and resume fixes (#1077, #1080), JPEG v2 (#1072), Jonathan's G17 and USB work (#1073-#1076), brcmfmac flowring count (#1081), ANE H16 startup restart (#1066) and the Touch ID per-prompt match (#904). Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com> Assisted-by: LLM
The PM notifier queues tb_dp_pm_reopen_work on system_unbound_wq. That
queue is deprecated in this tree, so the first resume on every Mac with
a Type-C port prints a kernel warning:
workqueue: work func tb_dp_pm_reopen_fn [thunderbolt] enqueued on
deprecated workqueue. Use system_{percpu|dfl}_wq instead.
system_dfl_wq is the unbound queue's replacement and keeps the same
behaviour: the work only takes the system sleep lock and clears a flag,
so it needs no CPU affinity.
Signed-off-by: Ryan Murray <81541646+RyanTheTide@users.noreply.github.com>
Assisted-by: LLM
The J616C device tree doesn't include the fan hwmon nodes the other 16-inch MacBook Pros have, so its two fans have no sensors or targets. Add them, with manual targets as on the J614s and J616s. Signed-off-by: AxxzyWasTaken <axxzytte@gmail.com>
Hardware-proven on the Neo by lane DT2 (Stage 2 DT trial, test merge of integration 43c790f + 1034 + 1043, DTB 1572d939, FLEET4 reg.sh no regression). W=1 t8140-j700.dtb 5 = base 5; DTB diff = the new disabled spi3/fpwm0/SPMI nodes and the J700 CPU zone limits. Overnight merge train 11-07.
Hardware-proven on the Neo by lane DT2 (Stage 2 DT trial, test merge of integration 43c790f + 1034 + 1043, DTB 1572d939, FLEET4 reg.sh no regression). W=1 t8140-j700.dtb 5 = base 5; DTB diff = the new disabled spi3/fpwm0/SPMI nodes and the J700 CPU zone limits. Overnight merge train 11-07.
arm64: dts: apple: t6041-j616c: describe the fans
…l-wq Merge train 11-07: FLEET4 one-shot 7.3.0-aurora-f4tbwq1 on the M2 (integration 43c790f + this, same tree), full silent suite vs fleet22: the tb_dp_pm_reopen_fn 'deprecated workqueue' warning is gone after 12 resumes, DP-1 + eDP kept, kerr 0, 0 AVC, v2 12/12 regrant. Host-tests CI green. One-line queue_work target change.
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 43c790f with one merge commit (5c6e6be, tree b3e93c21), so aurora-wip again contains the tested release branch. The merge is clean: aurora-wip's own extra content (the J813/J873 bring-up series) is untouched.
What integration adds since the last sync (#1071):
Gates (buildbox, the exact merge tree): fleet22 config Image + modules + dtbs build, run-host-tests, apple-dcp host selftests (fleet22 baseline 86/86). aurora-wip is the WIP branch, so build and host tests are its gate; everything integration brings is already hardware-qualified as fleet22 on all four Macs.
🤖 Generated with Claude Code