Advice regarding kernel_modules vs hal_modules

@kevindpatino (or others), I’m looking for some advice on the kernel modules and hal modules.
I’m currently experimenting with the Google Assistant implementation using the kernel modules. I got the assistant to work, so that’s great!

Here’s what I am aiming for:
I plan to use several Creators throughout the house in order to build a “Jarvis-like system”. I want to use the Google Assistant + Snowboy to have custom wakeword-activated speech assistant.
One of the systems will have a touch-screen (livingroom) and a pi-camera (for motion detection and face recognition), the others will be just a Pi+Creator.
Of course I want to use everloop for pretty led displays, and I also want to use Zwave, Zigbee and a 433Mhz module to control some 433mhz devices. Also I’ll probably use the IR and NFC.

I understand that at the moment the Google Assistant implementation uses the Kernel modules, which is incompatible with the HAL modules. Considering all the stuff I described above, I’m now unsure how to proceed. For what I want to do (camera, zwave, zigbee, etc) I’d say that using the HAL or Core would be best, but this is incompatible with Google Assistant at the moment. What are your plans for the (near) future? Should I focus on using the HAL and will you make that compatible with the audio-stuff the Google Assistant needs? Or should I focus on working with the kernel modules and will you add support for all the other stuff I need?

1 Like

In my opinion fokus the kernel modules, as everloop (character device), sensors (libiio), gpio (standard pi, currently without pwm) are currently working with the kernel modules.

They have already started rewriting the hal-drivers, as i will do for my python-hal-driver. if they have rewritten the hal-driver the malos / matrix-os will work also with the kernel modules.

1 Like