Hi Shakthi,
- POSIX environment.
I am just wondering how POSIX compliance, could possibly bring the concept of translators (as in Hurd) to Genode, or probably that exists already? Please clarify.
my proposal was not related to the POSIX environment topic but the mentioned tar-based ROM service could certainly be useful in such a context.
With POSIX enviroment, we mean an execution environment that enables us to port existing plain Unix applications in an easy way to Genode. For example, to make bash, gcc, and vim usable on Genode using a light- weight execution container rather than a full-blown Linux instance (e.g., OKLinux). We are not going the route to enhance Genode's base mechanisms to make the framework POSIX compatible nor are we planning to apply the everything-is-a-file metaphor to Genode. Our API facilitates access to objects (using capabilities as references) rather than operating on files. We will also not support functionality such as 'fork' natively on Genode because Genode provides a different mechanism of process creation and resource accounting.
However, a Genode subsystem may provide such Unix-like abstractions to its children without giving the children the ability to perform native Genode operations (very similar to how L4Linux manages the system calls of its Linux user-level processes). Such a container would enable us to reuse existing Unix software in one or multiple (isolated) POSIX execution containers. The design of such a container is work in progress.
| But as I said, please let us know more about your technical interests. | If you like to focus more on device drivers or other topics, there's | plenty of work to do as well. .-) --
Sure. This sounds good too. Do we have a list of required device drivers, or protocol stacks?
You asked for small tasks, so I haven't mentioned protocol-stack related work. But because you just asked, it would be very cool to make the BSD TCP/IP stack available to Genode. This TCP/IP stack seems to be very portable because it is used in a number of non-BSD projects such as QNX. But porting it is certainly not a small task. ;-)
Do you have particular interests in the one or the other feature of Genode or do you have a concrete application of Genode in mind? If yes, it would certainly make sense to gear your work towards a goal that is useful for you personally.
Norman