On Wed, Nov 07, 2001 at 07:19:23PM +0000, Patrick Strasser wrote: > I think a sub-Hurd is the Hurd's way of doing this at the moment. What > can it do for us in this direction,
It can already do everything. It runs its own filesystems, its own proc server, its own pfinet server, and has its own user id space. > and more important: what is a > sub-Hurd not possible to do? I am not aware of any limitation. A sub Hurd is just a second Hurd running in parallel with the other one. Currently, the console is directed to the terminal running the sub-Hurd, and it is too difficult to make two Hurds communicate with each other. > How does resource sharing work? Well, resource sharing is not implemented at all currently. That doesn't mean it is impossible or overly difficult to add. > What about security with a sub-Hurd? What about it? The Hurds don't know about each other. There are no ports from the one Hurd accessible to the others (well, I should say almost no ports, but it's the user starting the sub-Hurd who eventually will be able to decide). In particular, no privileges are leaked. > What about networking? Currently only possible with an extra NIC per sub Hurd. A virtual bridge is not implemented. > Are there real perspectives for running the Hurd in parallel to any > known OS? The Hurd can run in parallel to itself. If it can run in parallel to other OSs depends if those OSs can run on Mach, and if you have a bootloader for them that runs in the Hurd. > Hopefully this mail gives some new useful > aspects for further development of the Hurd. Well, the possibilities to do such things (and other things), and the actual functionality at least partly, was built into the Hurd from ground up by design. It can be improved a lot by adding the missing bits and pieces, and by adding some more good design for things like passing ports from and to sub-Hurds (like, you can't use shadowfs for the root fs of the sub-Hurd yet, you need a real partition or disk image file). But adding such functionality is something that moves in the direction the Hurd already has taken, and not against it :), so it should be more fun to hack it (entirely in user space as well, btw). Thanks, Marcus -- `Rhubarb is no Egyptian god.' Debian http://www.debian.org [EMAIL PROTECTED] Marcus Brinkmann GNU http://www.gnu.org [EMAIL PROTECTED] [EMAIL PROTECTED] http://www.marcus-brinkmann.de _______________________________________________ Bug-hurd mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/bug-hurd
