pmp v1 implementation by dd-dreams · Pull Request #576 · AsahiLinux/linux

dd-dreams · 2026-08-23T18:52:19Z

No description provided.

Apple Silicon based laptop use SPI as transport for HID. Add support for SPI-based HID devices and and Apple keyboard and trackpad devices. Intel based laptops using the keyboard input driver applespi use the same HID over SPI protocol and can be supported later. This requires SPI keyboard/mouse HID types since Apple's intenal keyboards/trackpads use the same product id. Signed-off-by: Janne Grunau <j@jannau.net>

Apple M2* chips have an embedded MTP processor that handles all HID functions, and does not go over a traditional bus like SPI. The devices still have real IDs, so add them here. Signed-off-by: Hector Martin <marcan@marcan.st>

Signed-off-by: Hector Martin <marcan@marcan.st>

This maximum is arbitrary. Recent Apple devices have some vendor-defined reports with 16384 here which fail to parse without this, so let's bump it to that. This value is used as follows: report->size += parser->global.report_size * parser->global.report_count; [...] /* Total size check: Allow for possible report index byte */ if (report->size > (max_buffer_size - 1) << 3) { hid_err(parser->device, "report is too long\n"); return -1; } All of these fields are unsigned integers, and report_count is bounded by HID_MAX_USAGES (12288). Therefore, as long as the respective maximums do not overflow an unsigned integer (let's say a signed integer just in case), we're safe. This holds for 16384. Signed-off-by: Hector Martin <marcan@marcan.st>

Apple MacBook keyboards started using HID over SPI in 2015. With the addition of the SPI HID transport they can be supported by this driver. Support all product ids over with the Apple SPI vendor id for now. The Macbook Pro (M1, 13-inch, 2020) uses the same function key mapping as other Macbook Pros with touchbar and dedicated ESC key. Apple silicon Macbooks use the same function key mapping as the 2021 and later Magic Keyboards. Signed-off-by: Janne Grunau <j@jannau.net>

We use BUS_HOST for MTP HID subdevices Signed-off-by: Hector Martin <marcan@marcan.st>

This mode is added to ease adding new xkeyboard configs for Apple silicon Macbook keyboards. The existing ones have strange quirks [1] and as the keyboard sends a key code for the 'fn' there is desire to use it as additional modifier [2]. [1]: https://pagure.io/fedora-asahi/remix-bugs/issue/17 [2]: https://asahilinux.org/docs/project/help-wanted/ (Keyboard layout cleanup) Signed-off-by: Janne Grunau <j@jannau.net>

Signed-off-by: Janne Grunau <j@jannau.net>

Will be used for supporting MacBook trackpads connected via SPI. Signed-off-by: Janne Grunau <j@jannau.net>

The trackpads in Macbooks beginning in 2015 are HID devices connected over SPI. On Intel Macbooks they are currently supported by applespi.c. This chang adds support for the trackpads on Apple Silicon Macbooks starting in late 2020. They use a new HID over SPI transport driver. The touch report format differs from USB/BT Magic Trackpads. It is the same format as the type 4 format supported by bcm5974.c. Signed-off-by: Janne Grunau <j@jannau.net>

Apple M2 devices expose the multi-touch device over the HID over DockChannel transport, which we represent as the HOST bus type. The report format is the same, except the legacy mouse header is gone and there is no enable request needed. Signed-off-by: Hector Martin <marcan@marcan.st>

The trackpad has to request multi touch reports during resume. Signed-off-by: Janne Grunau <j@jannau.net>

On at least some SPI devices (e.g. recent Apple Silicon machines), the Broadcom touch controller is prone to crashing. When this happens, the STM eventually notices and resets it. It then notifies the driver via HID report 0x60, and the driver needs to re-enable MT mode to make things work again. This poses an additional issue: the hidinput core will close the low-level transport while the device is closed, which can cause us to miss a reset notification. To fix this, override the input open/close callbacks and send the MT enable every time the HID device is opened, instead of only once on probe. This should increase general robustness, even if the reset mechanism doesn't work for some reason, so it's worth doing it for USB devices too. MTP devices are exempt since they do not require the MT enable at all. Signed-off-by: Hector Martin <marcan@marcan.st>

For SPI/MTP trackpads, query the dimensions via HID report instead of hardcoding values. TODO: Does this work for the USB/BT devices? Maybe we can get rid of the hardcoded sizes everywhere? Signed-off-by: Hector Martin <marcan@marcan.st>

Keyboard and trackpad of Apple Sillicon SoCs (M1, M1 Pro/Max) laptops are are HID devices connected via SPI. This is the same protocol as implemented by applespi.c. It was not noticed that protocol is a transport for HID. Adding support for ACPI based Intel MacBooks will be done in a separate commit. How HID is mapped in this protocol is not yet fully understood. Microsoft has a specification for HID over SPI [1] incompatible with the transport protocol used by Apple. [1] https://docs.microsoft.com/en-us/windows-hardware/drivers/hid/hid-over-spi Contains "HID: transport: spi: apple: Increase receive buffer size" The SPI receive buffer is passed directly to hid_input_report() if it contains a complete report. It is then passed to hid_report_raw_event() which computes the expected report size and memsets the "missing trailing data up to HID_MAX_BUFFER_SIZE (16K) or hid_ll_driver.max_buffer_size (if set) to zero. Co-developed-by: Hector Martin <marcan@marcan.st> Signed-off-by: Hector Martin <marcan@marcan.st> Signed-off-by: Janne Grunau <j@jannau.net>

DockChannel is a simple FIFO interface used to communicate between SoC blocks. Add a driver that represents the shared interrupt controller for the DockChannel block, and then exposes probe and data transfer functions that child device drivers can use to instantiate individual FIFOs. Signed-off-by: Hector Martin <marcan@marcan.st>

Apple M2 devices have an MTP coprocessor embedded in the SoC that handles HID for the integrated touchpad/keyboard, and communicates over the DockChannel interface. This driver implements this new interface. Signed-off-by: Hector Martin <marcan@marcan.st>

This driver can be used for coprocessors that do some background task or communicate out-of-band, and do not do any mailbox I/O beyond the standard RTKit initialization. Signed-off-by: Hector Martin <marcan@marcan.st>

Certain Broadcom bluetooth chips (bcm4377/bcm4378/bcm438) need ACL streams carrying audio to be set as "high priority" using a vendor specific command to prevent 10-ish second-long dropouts whenever something does a device scan. This patch sends the command when the socket priority is set to TC_PRIO_INTERACTIVE, as BlueZ does for audio. Signed-off-by: Sasha Finkelstein <fnkl.kernel@gmail.com>

The current approach of silently disabling all rust drivers if the toolchain is missing results in users that try to compile their own kernels getting a "successful" build and then being confused about where did their drivers go. In comparison, missing openssl results in a build failure, not a disappearance of everything that depends on it. This also means that allyesconfig will depend on rust, but since the rust experiment concluded with "rust is here to stay", i believe that allyesconfig should be building rust drivers too. Signed-off-by: Sasha Finkelstein <k@chaosmail.tech>

Signed-off-by: Janne Grunau <j@jannau.net>

Apple M3 Pro and Max devices are using 'gp00' keys for GPIO in addition to 'gP00' keys. Add a second compatible to handle this keys with an additional macsmc-gpio instance. Signed-off-by: Janne Grunau <j@jannau.net>

Add support for SMC GPIO keys with a lower letter 'p' via the "apple,smc-low-gpio" compatible. This adds support for a second macsmc-gpio controller using 'gp00' keys. These keys are used on Apple M3 Pro and Max MacBooks in the controller for keyboard and trackpad and for the built-in DisplayPort to HDMI converter. Signed-off-by: Janne Grunau <j@jannau.net>

Apple M3 Pro and Max devices are using 'gp00' keys for GPIO in addition to 'gP00' keys. These keys are handled by an additional macsmc-gpio instance using the "apple,smc-low-gpio" compatible. Signed-off-by: Janne Grunau <j@jannau.net>

Signed-off-by: dd-dreams <80887265+dd-dreams@users.noreply.github.com>

jannau

please clean your commits up. At least each commit should compile and each drivers should get their own commit. The dts changes need to be separate as well to avoid conflicts between bits/* branches

jannau · 2026-08-29T21:31:53Z


 	aliases {
 		gpu = &gpu;
+		pmp = &pmp;


I think for downstream pmp v1 should be optional as pmp v2. Please use the same #ifdefs as in the other device trees.
For upstream this requires select APPLE_PMP_V1 if PM in arch/arm64/Kconfig.platforms. For PMPv2 select APPLE_PMP_REPORT if PM is missing

jannau · 2026-08-29T21:34:36Z

 			msi-parent = <&pcie0>;
 			msi-ranges = <&aic AIC_IRQ 704 IRQ_TYPE_EDGE_RISING 32>;

-


nit: unrelated, is this an issue in the downstream or upstream dts. If it's downstream please submit fixup comment for this

jannau · 2026-08-29T23:11:47Z

+#[pin_data]
+struct PmpData {
+    dev: ARef<device::Device>,
+    pmp_mmio: Pin<KBox<Devres<IoMem<PMP_MMIO_SIZE>>>>,


please try to the iomem out of PmpData. This makes lifetime handling (coming in v7.2) much easier. At least for asc_mmio this should be no problem. If you rebase onto https://github.com/AsahiLinux/linux/tree/bits/250-aop start_cpu() can be simplified to

asc_mmio.update(ASC_CPU_CONTROL, |r| r.with_const_cpu_run::<1>());

directly inside probe()

is pmp_mmio used at all?

jannau and others added 19 commits June 20, 2026 17:01

HID: add HOST vendor/device IDs for Apple MTP devices

42c2f66

Apple M2* chips have an embedded MTP processor that handles all HID functions, and does not go over a traditional bus like SPI. The devices still have real IDs, so add them here. Signed-off-by: Hector Martin <marcan@marcan.st>

HID: core: Handle HOST bus type when announcing devices

63c2994

Signed-off-by: Hector Martin <marcan@marcan.st>

HID: apple: Bind to HOST devices for MTP

4cdbfc2

We use BUS_HOST for MTP HID subdevices Signed-off-by: Hector Martin <marcan@marcan.st>

HID: magicmouse: use a define of the max number of touch contacts

604082f

Signed-off-by: Janne Grunau <j@jannau.net>

HID: magicmouse: use struct input_mt_pos for X/Y

d45b52d

Signed-off-by: Janne Grunau <j@jannau.net>

HID: magicmouse: use ops function pointers for input functionality

0091e01

Will be used for supporting MacBook trackpads connected via SPI. Signed-off-by: Janne Grunau <j@jannau.net>

HID: magicmouse: Add .reset_resume for SPI trackpads

fdc798c

The trackpad has to request multi touch reports during resume. Signed-off-by: Janne Grunau <j@jannau.net>

soc: apple: Add RTKit helper driver

f77f12b

This driver can be used for coprocessors that do some background task or communicate out-of-band, and do not do any mailbox I/O beyond the standard RTKit initialization. Signed-off-by: Hector Martin <marcan@marcan.st>

dd-dreams force-pushed the pmp-v1-impl branch 6 times, most recently from 9653038 to ae66f0a Compare August 24, 2026 08:29

WhatAmISupposedToPutHere reviewed Aug 24, 2026

View reviewed changes

dd-dreams force-pushed the pmp-v1-impl branch from 7c29c60 to 9a61fdd Compare August 25, 2026 16:13

dd-dreams requested a review from WhatAmISupposedToPutHere August 25, 2026 16:14

dd-dreams force-pushed the pmp-v1-impl branch 2 times, most recently from 4582060 to 61a4ba0 Compare August 25, 2026 16:16

jannau and others added 11 commits August 26, 2026 10:46

Merge branch 'bits/220-tso' into asahi-wip

b26a3dc

Merge branch 'bits/240-isp' into asahi-wip

f842b53

Merge branch 'bits/250-aop' into asahi-wip

29f250d

Merge branch 'bits/260-spmi-typec' into asahi-wip

017fe87

Merge branch 'bits/500-backports' into asahi-wip

6a5fd7b

driver-core: Add error message to device_links_missing_supplier WARN()

31dd270

Signed-off-by: Janne Grunau <j@jannau.net>

dt-bindings: gpio: apple,smc: Add compatible for 'gp00' keys

d3950e5

Apple M3 Pro and Max devices are using 'gp00' keys for GPIO in addition to 'gP00' keys. Add a second compatible to handle this keys with an additional macsmc-gpio instance. Signed-off-by: Janne Grunau <j@jannau.net>

jannau force-pushed the asahi-wip branch from 77cb8f2 to b02aba8 Compare August 26, 2026 09:11

dd-dreams force-pushed the pmp-v1-impl branch 3 times, most recently from 50fd00e to 0408eee Compare August 26, 2026 18:50

dd-dreams requested a review from WhatAmISupposedToPutHere August 26, 2026 18:51

WhatAmISupposedToPutHere requested changes Aug 26, 2026

View reviewed changes

Comment thread Documentation/devicetree/bindings/power/apple,pmp-v1-report-entry.yaml Outdated

Comment thread Documentation/devicetree/bindings/power/apple,pmp-v1.yaml Outdated

dd-dreams force-pushed the pmp-v1-impl branch from 0408eee to 28300e1 Compare August 26, 2026 19:13

dd-dreams added 7 commits August 26, 2026 22:18

add PMPv1 driver

58545df

Signed-off-by: dd-dreams <80887265+dd-dreams@users.noreply.github.com>

add pmp-v1 pd docs

231eac6

Signed-off-by: dd-dreams <80887265+dd-dreams@users.noreply.github.com>

remove git leftovers

3bb6ff0

Signed-off-by: dd-dreams <80887265+dd-dreams@users.noreply.github.com>

fix build and address feedback

4bde67e

Signed-off-by: dd-dreams <80887265+dd-dreams@users.noreply.github.com>

update dt-bindings for pmp v1

93c9296

Signed-off-by: dd-dreams <80887265+dd-dreams@users.noreply.github.com>

split dt-binding for pmp-v1

4df058e

Signed-off-by: dd-dreams <80887265+dd-dreams@users.noreply.github.com>

change to const and remove redundant lines

9b60733

Signed-off-by: dd-dreams <80887265+dd-dreams@users.noreply.github.com>

dd-dreams force-pushed the pmp-v1-impl branch from 28300e1 to 9b60733 Compare August 26, 2026 19:21

dd-dreams requested a review from WhatAmISupposedToPutHere August 26, 2026 19:21

WhatAmISupposedToPutHere approved these changes Aug 26, 2026

View reviewed changes

jannau requested changes Aug 29, 2026

View reviewed changes

jannau force-pushed the asahi-wip branch from ccce11a to ca9a850 Compare August 30, 2026 09:57

Navigation Menu

Uh oh!

pmp v1 implementation - #576

pmp v1 implementation#576
dd-dreams wants to merge 54 commits into
AsahiLinux:asahi-wipfrom
dd-dreams:pmp-v1-impl

dd-dreams commented Aug 23, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

jannau left a comment

Uh oh!

jannau Aug 29, 2026

Uh oh!

jannau Aug 29, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

jannau Aug 29, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

		msi-parent = <&pcie0>;
		msi-ranges = <&aic AIC_IRQ 704 IRQ_TYPE_EDGE_RISING 32>;

Uh oh!

Conversation

dd-dreams commented Aug 23, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

jannau left a comment

Choose a reason for hiding this comment

Uh oh!

jannau Aug 29, 2026

Choose a reason for hiding this comment

Uh oh!

jannau Aug 29, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

jannau Aug 29, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants