Skip to content

Release Notes

Yocto mx95-yocto-walnascar-6.12.49_2.2.0-v1.0

Release Notes

Based on release Yocto: Poky 5.2.4, BSP: NXP 6.12.49_2.2.0, Linux: lf-6.12.y
Release tag mx95-yocto-walnascar-6.12.49_2.2.0-v1.0
Release manifest imx-6.12.49-2.2.0.xml
Date 01/26/2026
Supported platforms DART-MX95
SOM revision v1.2
File System build system Yocto Walnascar
Recovery SD card link mx95__yocto-walnascar-6.12.49_2.2.0-v1.0__android-15.0.0_2.0.0-v1.1.wic.zst
Recovery SD card SHA256 98737719b19a600eca0683aec6cda47642cc1821f152a492e25cdac57ff0ff67
Cortex M7 U-Boot support MCUXPRESSO_25.06.0_V1.0_DART-MX95
Cortex M7 Linux remoteproc support MCUXPRESSO_25.06.0_V1.0_DART-MX95

Relevant git Repositories

Source Repository Branch Commit ID
System Manager https://github.com/varigit/imx-sm lf-6.12.49-2.2.0_var01 a2a6857910ab15be3fe00990595e0dce3579e442
U-Boot https://github.com/varigit/uboot-imx lf_v2025.04_6.12.49-2.2.0_var01 8e974a6f5ab077bfe6ff06d6f37c40c344f1e7a6
Kernel https://github.com/varigit/linux-imx lf-6.12.y_6.12.49-2.2.0_var01 8d60ff1e3dfb19f130274f5af955cb3a9cef3db8
meta-variscite-bsp-imx https://github.com/varigit/meta-variscite-bsp-imx walnascar_6.12.49-2.2.0_var01 5551fa88ddb7d4d36e6e6b3310c7bc3557f93218
meta-variscite-bsp-common https://github.com/varigit/meta-variscite-bsp-common walnascar_6.12.49-2.2.0_var01 1432b399ffa62bb1d9fc1586436539c8fae3c9b2
meta-variscite-sdk-imx https://github.com/varigit/meta-variscite-sdk-imx walnascar_6.12.49-2.2.0_var01 67a4ebf0e9363e95b8db31b8dfe01917c7064452
meta-variscite-sdk-common https://github.com/varigit/meta-variscite-sdk-common walnascar_6.12.49-2.2.0_var01 9be948d8c62b5e526425850281b2c208faca3d35

After downloading the image, visit our Recovery SD card guide to learn how to program a recovery SD card.

Change Log

Change List Description
Release 1.0
Initial Yocto Walnascar 6.12.49_2.2.0 release Based on NXP Yocto Walnascar 6.12.49_2.2.0 and NXP Linux Kernel lf-6.12.y
Known Issues
Tensorflow build fetch error Due to possible checksum changes in the files hosted at https://gitlab.arm.com/kleidi/kleidiai/-/archive/, TensorFlow Lite may no longer be built from source. Unless these files are already present in your local download cache, the build will fail. As a workaround, the Machine Learning package must be removed from the build. Add the following to the conf/local.conf:
For the image fsl-image-gui/fsl-image-gui-chromium:
RDEPENDS:packagegroup-variscite-imx-ml:mx9-nxp-bsp:remove = "packagegroup-imx-ml"
For the image fsl-image-qt6, in addition:
RDEPENDS:packagegroup-imx-gopoint:remove = "imx-nnstreamer-examples"
Inaccurate timing on audio playback Audio Playback can fail or may be perceived as timing inaccurate, especially when different sample rates are used and the audio device is accessed directly. As a workaround, it is recommended to either use the default sample rate (48kHz), start playback twice in a row, or use the mixer instead of direct access. The underlying reason is a faulty synchronization of the clock settings, which can result in a mismatch between the SAI and the codec clock.
Note
Allow using only one camera: Due to a limitation in the current kernel implementation of MIPI-CSI control chain, only the cameras actually present must be enabled. In order to use a single camera, only that specific camera can be enabled, while the other and related chain items must be disabled.