Update nRF Connect and McuX SDKs

[benedekkupper]

Update the nRF Connect and McuX SDKs to the latest available version. UHK60 is affected, but no meaningful changes expected. UHK80 shall be more carefully tested, with focus on BLE and also USB.

• UHK60 tested
• UHK80 BLE left-right tested
• UHK80 BLE right-dongle tested
• UHK80 BLE HOGP tested
• UHK80 USB tested
• UHK80 new mcuboot FW update tested

Closes #1043

[Copilot]

Pull request overview

Updates the firmware’s west manifests to newer nRF Connect SDK and MCUXpresso SDK versions, along with a set of compatibility adjustments (BLE advertising params, build system tweaks, and patch set updates) to keep builds working across UHK60/UHK80 targets.

Changes:

• Bump nRF Connect SDK to v3.2.4 and MCUX manifests to v26.03.00, adjusting imported module allowlists accordingly.
• Update UHK80 board configuration (runner args, Kconfig defaults, PM hook stubs) and BLE advertising parameters for newer Zephyr/NCS APIs.
• Refresh patch set by updating the sysbuild CMakePresets patch and dropping multiple no-longer-needed patches.

Reviewed changes

Copilot reviewed 15 out of 15 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
west_nrfsdk.yml	Updates NCS revision and imported module allowlist to match the new SDK layout.
west_mcuxsdk.yml	Updates MCUX manifest revision.
right/src/macros/command_hash.gperf	Suppresses a GCC warning during gperf-generated compilation.
right/CMakeLists.txt	Adjusts MCUX build configuration to work in a west workspace.
patches/zephyr/0003-sysbuild-support-application-CMakePresets.json-files.patch	Updates the sysbuild patch to pass APP_DIR via environment and adds sysbuild CMakePresets.json.
patches/zephyr/0001-udc-nrf-disable-SOF-interrupts.patch	Removed patch (presumably upstreamed/obsolete with the SDK bump).
patches/zephyr/0001-drivers-udc-eliminate-C-compiler-warnings.patch	Removed patch (presumably upstreamed/obsolete with the SDK bump).
patches/sdk-nrf/0001-Remove-mcuboot-build-warnings.patch	Removed patch (presumably upstreamed/obsolete with the SDK bump).
patches/mcuboot/0003-zephyr-cmake-fails-early-and-with-meaningful-message.patch	Removed patch (presumably upstreamed/obsolete with the SDK bump).
patches/core/0003-cmake-add-ihex-binary-output-option.patch	Removed patch (presumably upstreamed/obsolete with the SDK bump).
device/src/keyboard/charger.c	Removes an unused SoC header include.
device/src/bt_advertise.c	Updates advertising parameter macros/options to newer Zephyr/NCS API.
boards/ugl/uhk-80/board.cmake	Adds conditional nrfutil ext-mem runner arg and adjusts include ordering.
boards/ugl/uhk-80/board.c	Adds weak PM hook stubs (likely required by new PM defaults).
boards/ugl/uhk-80/Kconfig.defconfig	Enables PM/stack protection defaults and updates BT controller-related defaults.

[mondalaci]

https://github.com/UltimateHackingKeyboard/firmware/actions/runs/24426227631 bricked my UHK 80. Please test it.

[benedekkupper]

@kareltucek please give this PR a try. I had to disable enableDataLengthExtension to the HID host, as when it was enabled, the BT connection establishment to Windows would hang halfway, the UHK80 crashing. The source code says that this was required to make pairing work on Linux, can you check if this now works without it?

[kareltucek]

@benedekkupper 😭.

My bluetooth testing says:

• android pairing randomly fails.
• linux pairing works the first time, but if I unpair and try to repair, it fails. I have to go through the whole rfkill reset, and maybe uhk reset. Didn't test very thoroughly.

What is worse, when trying to switch between connections, I am getting this crash. (Got it at least 5 times in the last 10 minutes.):

ASSERTION FAIL @ WEST_TOPDIR/zephyr/subsys/bluetooth/host/hci_core.c:504
<err> os: r0/a1:  0x00000003  r1/a2:  0x200102b0  r2/a3:  0x00000000
�<inf> udc_nrf: Preinit
[00:00:00.001,892] <inf> udc_nrf: Preinit

Some logs from failing linux pairings:

Connected: 98:5F:41:D2:92:3A (public)
Security failed: 98:5F:41:D2:92:3A (public) level 1 reason: Authentication requirements (4)
LE data len updated: TX (len: 251 time: 2120) RX (len: 251 time: 2120)
uhk80:right$ <inf> StateSync: Batteries: 77% 100% (3983 / 4100 mV; 4127 / 4120 mV)
[00:07:32.067,901] <inf> StateSync: Batteries: 77% 100% (3983 / 4100 mV; 4127 / 4120 mV)
<inf> bas: BAS Notifications enabled
[00:07:32.807,434] <inf> bas: BAS Notifications enabled
<inf> Bt: MTU exchanging n/a (Tux, 98:5f:41:d2:92:3a)
[00:07:32.807,952] <inf> Bt: MTU exchanging n/a (Tux, 98:5f:41:d2:92:3a)
<inf> Bt: Connected n/a (Tux, 98:5f:41:d2:92:3a), 7 3
[00:07:32.808,349] <inf> Bt: Connected n/a (Tux, 98:5f:41:d2:92:3a), 7 3
<wrn> Bt: Bt security failed: n/a (Tux, 98:5f:41:d2:92:3a), level 1, err 4, disconnecting
[00:07:33.013,183] <wrn> Bt: Bt security failed: n/a (Tux, 98:5f:41:d2:92:3a), level 1, err 4, disconnecting
Disconnected: 98:5F:41:D2:92:3A (public) (reason 0x05)
uhk80:right$ <wrn> Bt: MTU exchange failed for n/a (Tux, 98:5f:41:d2:92:3a) (err 14)
[00:07:33.193,420] <wrn> Bt: MTU exchange failed for n/a (Tux, 98:5f:41:d2:92:3a) (err 14)
<inf> bas: BAS Notifications disabled
[00:07:33.193,542] <inf> bas: BAS Notifications disabled
<inf> Bt: Disconnected from n/a (Tux, 98:5f:41:d2:92:3a), reason 5
[00:07:33.194,183] <inf> Bt: Disconnected from n/a (Tux, 98:5f:41:d2:92:3a), reason 5
<inf> Bt: BtManager: Scheduling scan/adv in 200ms. (disconnected)
[00:07:33.194,274] <inf> Bt: BtManager: Scheduling scan/adv in 200ms. (disconnected)
<inf> Bt: BtManager: Start 'advertising' (0).
[00:07:33.395,233] <inf> Bt: BtManager: Start 'advertising' (0).
<inf> Bt: Adv: 'HID "and NUS"' started
[00:07:33.398,010] <inf> Bt: Adv: 'HID "and NUS"' started



<wrn> bt_hci_core: opcode 0x2027 status 0x12
[00:08:34.286,987] <wrn> bt_hci_core: opcode 0x2027 status 0x12
<err> bt_id: Failed to add IRK to controller
[00:08:34.286,987] <err> bt_id: Failed to add IRK to controller
<inf> Notify: Notification: Ble slot 7 unpaired.
[00:08:34.288,208] <inf> Notify: Notification: Ble slot 7 unpaired.
Connected: 98:5F:41:D2:92:3A (public)
Security failed: 98:5F:41:D2:92:3A (public) level 1 reason: Authentication requirements (4)
LE data len updated: TX (len: 251 time: 2120) RX (len: 251 time: 2120)
Disconnected: 98:5F:41:D2:92:3A (public) (reason 0x05)
uhk80:right$ <inf> Bt: MTU exchanging n/a (Tux, 98:5f:41:d2:92:3a)
[00:08:37.784,851] <inf> Bt: MTU exchanging n/a (Tux, 98:5f:41:d2:92:3a)
<inf> Bt: Connected n/a (Tux, 98:5f:41:d2:92:3a), 7 3
[00:08:37.785,247] <inf> Bt: Connected n/a (Tux, 98:5f:41:d2:92:3a), 7 3
<wrn> Bt: Bt security failed: n/a (Tux, 98:5f:41:d2:92:3a), level 1, err 4, disconnecting
[00:08:38.002,777] <wrn> Bt: Bt security failed: n/a (Tux, 98:5f:41:d2:92:3a), level 1, err 4, disconnecting
<wrn> Bt: MTU exchange failed for n/a (Tux, 98:5f:41:d2:92:3a) (err 14)
[00:08:38.171,752] <wrn> Bt: MTU exchange failed for n/a (Tux, 98:5f:41:d2:92:3a) (err 14)
<inf> Bt: Disconnected from n/a (Tux, 98:5f:41:d2:92:3a), reason 5
[00:08:38.172,454] <inf> Bt: Disconnected from n/a (Tux, 98:5f:41:d2:92:3a), reason 5
<inf> Bt: BtManager: Scheduling scan/adv in 100ms. (disconnected)
[00:08:38.172,485] <inf> Bt: BtManager: Scheduling scan/adv in 100ms. (disconnected)
<inf> Bt: BtManager: Start 'advertising' (0).
[00:08:38.272,338] <inf> Bt: BtManager: Start 'advertising' (0).
<inf> Bt: Adv: 'HID' started
[00:08:38.275,329] <inf> Bt: Adv: 'HID' started




[bluetoothctl]> pair FA:3E:B0:AD:3E:E0
Attempting to pair with FA:3E:B0:AD:3E:E0
[CHG] Device FA:3E:B0:AD:3E:E0 Connected: yes
Failed to pair: org.bluez.Error.ConnectionAttemptFailed
[SIGNAL] org.bluez.Device1.Disconnected org.bluez.Reason.Unknown Unspecified
[CHG] Device FA:3E:B0:AD:3E:E0 Connected: no
[bluetoothctl]>

I don't have energy right now for further digging. Sorry!

[benedekkupper]

Walking through the nordic BLE course has given me some clues (that the release notes didn't) on what could have changed in between:

1. The BLE advertisements are now always one-off, automatically turned off when a connection is established. This means we need to re-examine the logic when to (re)enable the advertisements.
2. There is a hint in the exercise that bluetooth API shouldn't be called directly from bluetooth callback context. This is a probable cause of all the hangups and subsequent BT issues we have been seeing (as enableDataLengthExtension is called from connected callback).

@kareltucek can you check these points?
This also means that the adffec5 commit is definitely the wrong path and should be reverted.

[kareltucek]

The BLE advertisements are now always one-off, automatically turned off when a connection is established. This means we need to re-examine the logic when to (re)enable the advertisements.

This has always been so as far as I remember, at least for those advertising schemes that we were using.

that bluetooth API shouldn't be called directly from bluetooth callback context

Argh. Right.

Pushing a fix to use the system workqueue.

So far, my uhk behaves.

[kareltucek]

• I think uhk80 is getting stuck in the bootloader. It works after flashing via west flash, but once a reenumerate command is issued, or uhk80 is powercycled, it remains "bricked" until next west flash.

• Oob pairing is failing. Will look into this once the above is resolved.

[kareltucek]

Nevermind. Both are probably unrelated to this issue.

[kareltucek]

UHK80 new mcuboot FW update tested

Any specific pointers on what needs to be tested?

(Clearly, the timeout, agent usb flashing. Anything else?)

[benedekkupper]

UHK80 new mcuboot FW update tested

Any specific pointers on what needs to be tested?

(Clearly, the timeout, agent usb flashing. Anything else?)

Nothing specifically, beyond flashing the whole image with the debug probe, and then running a FW update through agent.

[kareltucek]

Encountered when testing:

• uhk paired to a ble linux host
• at first, everything was working, then, after a few minutes and maybe a few host switches, sending reports stopped working, producing the 2 error code:

<wrn> UsbReports: Send failed: No such file or directory

• reconnecting or re-pairing the ble connection didn't help
• rebooting uhk did help

Apart from that, everything seems to work.

[benedekkupper]

Encountered when testing:

• uhk paired to a ble linux host

• at first, everything was working, then, after a few minutes and maybe a few host switches, sending reports stopped working, producing the 2 error code:

<wrn> UsbReports: Send failed: No such file or directory

I'm clueless from this info, the code paths I found that can give this error are static, so it should either always or never happen. Maybe you can try to reproduce it with the debugger, once you have it, step through the sending call stack and see what condition produces this error. You might need to watch the data element's memory location that triggered the error for write events to find what corrupts it.