Matthew Dillon dillon at
Tue Feb 17 10:47:51 PST 2015

Running well in virtualized environments is important to us, but it isn't a
focus.  If FreeBSD makes sufficient progress with bhyve then it would
probably wind up getting ported to DFly eventually.  But I view VMs as
creating the same sort of continous catch-up work (in terms of what and how
well hardware gets emulated, and how much of the real hardware is
supported) that new chipsets create for driver developers.  I don't think
it's healthy for developers to have to spend their whole lives playing

What I would rather have is an OS core that is stable enough that it just
doesn't crash any more.  On the server side of the equation I think the
goal is possible.  Servers basically only have to deal with network and
disk controllers, not sound or video or other hardware, so the drivers used
by servers are relatively more simple and can slowly approach a bug-free
state.  Once an operating system reaches that point, there is no reason to
run the OS itself under a VM even in huge deployments.

I would rather that we focused on better fire-walling of services within
the infrastructure of the operating system rather than punting on it and
running each service in its own VM with its own copy of the operating


On Tue, Feb 17, 2015 at 9:10 AM, Luis Mendes <luislupe at> wrote:

> Now that DragonFlyBSD's kernel has concurrent and parallel features done
> as best as possible, main focus seems to be shifting to HAMMER2, am I
> correct?
> Can we expect to have DragonFlyBSD as a virtualization host (besides
> jails) after work in HAMMER2 is done?  What can we expect, xen as dom0,
> bhyve or other technology?  Is there some time horizon for this?
> Thanks for any info,
> Luis
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the Users mailing list