removing L4Linux

Stefan Kalkowski stefan.kalkowski at ...1...
Tue Jan 3 08:45:46 CET 2017


Hi folks,

since the very first project that used L4Linux on top of Genode, almost
nobody cared much about it. The main developers of Genode do not make
use of it, and I do not know of any project currently depending on it.
When it comes to OS ABI compatibility, we make use of VMMs like
VirtualBox or Seoul. For device driver support, the project also do not
follow the L4Linux approach, but drivers are built by ourselves or we
port them using the device-driver environment (DDE) approach.

On the other hand, L4Linux often stands in the way when e.g.,
modernizing the Genode API like with recent issue #1987. It also kept me
from upgrading to newer Fiasco.OC kernel versions in the past, because
of the dependencies of L4Linux with regard to the kernel. In contrast to
L4Linux, I know that people are using Genode with Fiasco.OC and would
profit from newer versions. I also do so, e.g., when using Fiasco.OC's
debugger on real hardware. If maintaining Genode/Fiasco.OC gets easier
by removing the burden of L4Linux I'm for it.

Nevertheless, if somebody is still actively using it, we should find a
solution, e.g., by sharing the maintenance effort, or moving the current
state into the Genode world repository. I have opened an issue accordingly:

  https://github.com/genodelabs/genode/issues/2216

If you have a future interest in L4Linux on top of Genode, please
contribute to the discussion in the issue tracker.

Best regards
Stefan

-- 
Stefan Kalkowski
Genode Labs

https://github.com/skalk ยท http://genode.org/




More information about the users mailing list