I fully agree. I really like some of the DisplayLink products (e.g. ASUS MBR16AC). However, driver support for Linux is completely inconsistent. As a consequence, I don't buy DisplayLink products for my company.

Your "Ubuntu" driver package works inconsistently, fails every time there is a system update to the kernel or the X server, and performs badly. It's clearly not a priority for your development team, who likely have orders to focus their efforts on the larger market-share platforms. The open-source community is packed with talented developers who will work for free to help you deliver a more polished product, providing a better experience for your users, and ultimately resulting in more product sales.

Hugely agree, I'm very tired of having to re-hack this together after every kernel update / upgrade. Keeping things as a manual upgrade is so far from the current world of Linux, and the agile CI/CD processes.

Agreed. If the compression technology is too proprietary to be released, then make 2 separate modules, one proprietary to load the compression algorithms, one open source so that kernel developers can easily port the module and update with each new kernel, while the compression algorithm remains closed source and proprietary. That way is best of both worlds.

Derrick Rose: there's no talk of "multiple kernels" but of getting the support to the official, mainline kernel.

I would think having it as a module would be more supportive over multiple kernels. This way you dont have to recompile a kernel for each distro just compile the module for the kernel WHy NVIDIA, AMD and VMWARE do it.

Is there a chance displaylink could try to add the driver into mainline kernel (except the binary bits ofcourse), this way more distros and users could use it without problems, also community would probably help with development and maintenance.