[1] https://lists.archlinux.org/pipermail/arch-general/2017-Febr...
But, in the case of running things in containers, you can stop exploits of user namespaces through seccomp filters that block unshare(CLONE_NEWUSER) -- Docker does this by default.
Red Hat created a Kubernetes compatible set of tools for running Docker compatible OCI containers called CRI-O https://cri-o.io/ with RHEL/Centos 7.7 and 8+ you can run containers as a regular user: https://www.redhat.com/en/blog/preview-running-containers-wi... using their tools.
The main thing about it is cgroups are disabled and it requires userspace networking.
Here's a write-up on it: https://engineering.docker.com/2019/02/experimenting-with-ro...
An unprivileged-user docker daemon would be limited to either communicate with an isolated network namespace on the parent side or do userspace forwarding of network traffic. Or it would require a privileged helper for the network parts.
All of that being said, this bug is caused by container runtimes trusting the rootfs too much. This is something I've been trying to improve but it's definitely not a trivial problem (lots of things require trusting the container processes in specific ways due to limitations in the corresponding kernel APIs -- though I am working on fixing those too).